Che materia stai cercando?

EEAG report 2010 Appunti scolastici Premium

Dispensa al corso di Economia dell'integrazione europea della Prof.ssa Lilia Cavallari. Trattasi del rapporto 2010 dello European Economic Advisory Group, all'interno del quale sono affrontati i seguenti argomenti: situazione economica europea nel 2010, conseguenze della crisi finanziaria sul mercato della finanza, crescita del debito... Vedi di più

Esame di Economia aziendale docente Prof. L. Cavallari

Anteprima

ESTRATTO DOCUMENTO

Chapter 2

see how it correlates with the

Figure 2.3 trust these people have towards

banks and insurance companies.

The survey in fact also reports

information on how much peo-

ple trust banks and insurance

companies, which can be corre-

lated with their past experience

of financial fraud. This correla-

tion is shown in Figure 2.3 which

clearly documents that those

who were cheated more often in

the past 5 years tend today to

trust intermediaries less than

those who were cheated less

often or not cheated at all.

Hence one expects that similar

Figure 2.4a effects have been at work during

The geographical spread of Madoff´s victims the financial crisis as its unfolding

US victims concentration revealed the frauds to which in-

vestors were exposed. To test for

this effect we rely on the Financial

Trust Index Survey and merge the

data with the number of Madoff

victims in the area (either the zip

code district or the state) were the

investor lives and check how it

correlates with the level of trust of

these investors. The idea is that

in areas where the number of

Madoff victims is larger (for a

given population), Madoff ’s

fraud, and more generally finan-

cial frauds, are more salient, either

because chances of knowing di-

Figure 2.4b rectly or indirectly (through word

The geographical spread of Madoff´s victims of mouth) one of the victims are

Europe victims concentration higher or because, in places with

lots of victims, the local press

devotes a lot more attention and

coverage, which adds to that

devoted by the national press.

Hence, in these states the drop in

trust following Madoff scandal

should be more marked. Figu-

re 2.4a shows how spread out are

the Madoff victims and where

they were located in the US.

Figure 2.4b documents that the

victims of this fraud were present

also across Europe so that if it had

any effect on trust it also under-

59 EEAG Report 2010

Chapter 2 and actually may have behaved honestly. Indeed, when

mined that of European investors. Because of data the fraud comes to be known by many it tends to

availability we focus on the effect on the trust of US spread the suspicion to the whole financial industry

investors. Figure 2.5 shows in four different panels the leading, to a shared fall in trust, as happened during

correlation between the average level of trust investors the crisis. In other words, the emergence of Madoff’s

living in a state have towards banks, bankers, brokers fraud undermined the confidence in the whole financial

and the stock market, respectively, and the density of industry. Interpreting this popular sentiment Paul

Madoff victims in the state where they live. The figure Krugman in a New York Times column (December

shows clearly that trust towards banks, bankers and 2008) asked: “How different is what Wall Street did

brokers is lower where the number of Madoff victims from the Madoff affair? Well Madoff simply skipped a

is larger, while the salience of this fraud seems to have few steps, simply stealing his clients’ money rather than

little effect on the trust towards the stocks market, collecting big fees while exposing investors to risks they

which is consistent with the fact that Madoff was a did not understand” (NYT, December 2008). Obvious-

fund manager. There are three points to notice. First, ly, those who have been damaged the most are the inter-

these correlations show that a financial fraud not only mediaries or brokers that have always done their job

affects the trust of the direct victims of the fraud but it honestly. Quint Tatro, president, founder and manager

also affects the trust of those who, even if they have not of Tatro Capital, an investment management company,

directly suffered from the scandal, have come to know in a sorrowful letter wrote in January 2009: “A funny

about it, either because the information was publicised thing happened recently: Many new individuals simply

through the press or because they met a victim. This is have a hard time believing a traditional investment

more likely when the fraud is sizable and information manager from Kentucky didn’t ‘get killed’ along with

about it reached many investors, as it was the case dur- everyone else. I have now heard that at least 2 people,

ing the financial crisis with the Madoff case first, fol- when my firm was recommended to them, responded

lowed by the Sir Allen Stanford fraud and many other by asking whether we were ‘legit.’ One advisor, who

minor but diffused examples of deception and financial held half of a mutual client’s investment and will no

abuses that, because the topic was on demand, cap- longer be holding that half, went so far as to request the

tured the attention of the press. Second, not only the Schwab statements from the client verifying my per-

trust towards those who committed the fraud falls – a formance. I suspect the client didn’t amuse the advisor

specific banks or banker – but the drop in trust spills with this degrading request, but who knows. While

over to many other agents that are not directly most of my frustration can be pushed back onto Wall

involved, such as banks, bankers or brokers that may Street … I entirely believe that Bernard Madoff is

have no direct link with those who committed the fraud directly correlated with this

Figure 2.5 new rise in scepticism. So

now, in addition to battling

Madoff´s victims and loss of confidence Mr. Market on a daily basis, I

Effect of Madoff victims on trust on banks Effect of Madoff victims on trust on bankers

4

4 have to deal with charges of

3.5

3.5 untrustworthiness.”

bankers

banks 3

on 3 on

Trust Trust Third, the correlations shown

2.5

2.5 in Figure 2.5 only show the

2

2 effect on the level

differential

0 2 4 6 8 0 2 4 6 8

N. of Madoff victims (logs)

N. of Madoff victims (logs) of trust of the Madoff’s fraud

Level of trust in bankers Fitted values

Level of trust in banks Fitted values due to the fact that investors

Effect of Madoff victims on trust on brokers Effect of Madoff victims on trust on stock market

2.8 3 in different states were differ-

2.6 entially informed about it, for

market 2.5

brokers 2.4 example, because in states

stock

on 2.2 on with more victims local news-

Trust Trust 2

2 papers devoted more space

1.8 and for more time to it. This

1.5

0 2 4 6 8 0 2 4 6 8

N. of Madoff victims (logs)

N. of Madoff victims (logs) proves that the Madoff fraud

Level of trust in brokers Fitted values Level of trust in stock market Fitted values has lowered trust in financial

The figure shows the relation between the n. of Nadoff´s victim by State and the average value of trust

in banks, bankers, brokers and the stock market in the US after the October 2008 crisis. intermediaries, but it is likely

Effects: 1 sd in lncase lowers trust in banks by 1/3 of trust in banks standard deviation. to understate the effect since it

Effects: 1 sd in lncase lowers trust in bankers by 1/3 of trust in bankers standard deviation.

60

EEAG Report 2010 Chapter 2

impact on the working of financial markets in the

is unable to identify the drop in the average level trust coming years. But before illustrating these margins, it

of American investors after the discovery of Madoff’s is worth noticing that the decline in trust played a very

fraud: the latter could be first order. important role already during the crisis as those who

lost their trust towards their bank were also the first

Finally, to further strengthen the link between the fall in to withdraw cash from their deposits during the days

trust and the perceptions of opportunistic behaviour in following the collapse of Lehman Brothers.

financial markets brought to light by the crisis, we

examine the following question asked in the FTIS: “Do In ongoing research, Guiso, Sapienza and Zingales

you feel you have been cheated or misled by a bank in (2009) argue that differences in levels of trust across

the last year?” Respondents could answers “yes” or individuals can explain who starts a run on the bank

“no”. In unreported regressions we find that those who in a period of financial distress. Using data from the

have been cheated or misled by a bank over the year Trust Fi-nancial Index Survey they show that those

prior to the crisis report a lower level of trust towards investors that lost trust in banks and the financial sys-

banks and bankers. Furthermore, not only these people tem where the first to withdraw cash from their

lost confidence in the intermediary that cheated them accounts at the peak of the crisis – that is they started

(banks in this case) but also in other intermediaries and a bank run. Figure 2.6 documents this finding show-

markets such as brokers, mutual funds and the stock ing the correlation between the fraction of people that

market though by a somewhat smaller amount – a more run on the bank and the level of trust of the investors:

direct way of supporting Quint Tatro’s closing state- people that lost trust in their bank were more than

ment in the previous citation and showing its generali- four times more likely to run on the bank than those

ty. Thus misbehaviour by one intermediary triggers a who retained full trust, contributing to the spread of

loss of trust in the whole industry. In addition this

6 the panic. Guiso, Sapienza and Zingales (2009) report

spillover effect extends to trust in large corporations similar evidence for a sample of investors of a large

and even to trust towards other people in general, Italian bank. The interesting feature is that in this case

though the effect is much smaller. Insofar as trust they can look at the correlation between the decision

towards these entities also matters for transactions and to run and the level of trust well before the crisis.

trade, the loss of trust provoked by the crisis has affect- Those who used to trust less were also more likely to

ed the economy not only because investors have become take out their deposits, consistent with the idea that

more cautious in making their money available to the lack of trust makes a bank fragile and more exposed

financial industry but also because they have become to runs.

more reluctant to trade in general. This has acted as an

amplifier of the effect of the financial crisis on the econ- For the future, the drop in trust is likely to have perva-

omy. Finally it is interesting to note that these effects sive effects on investors’ reliance on financial markets

were obtained after controlling for a variety of charac- across various dimensions – one of the most impor-

teristics, in particular for an index of how angry tant legacies of the financial crisis. In particular the

investors were because of the crisis, reassuring that the

effect of the experience of decep-

tion on people’s trust does not Figure 2.6

reflect some other variable that

also may impinge on their trust.

How will the fall in trust affect

financial transactions?

The fall in trust is likely to affect

investors’ decisions on various

margins that may have a strong

6 Notice that also trust in the Fed drops; to

some extent this may look surprising, since

in principle the Fed’s response to the crisis in

terms of liquidity provision was “right”. But

people seem to think otherwise. One inter-

pretation is that they held the Fed responsi-

ble for not having done enough prior to the

crisis to prevent banks’ misbehaviour. 61 EEAG Report 2010

Chapter 2 fall in trust is likely to affect peo- Tabl e 2. 3

ple’s willingness to enter into any Chan g e inn willi ng nesss too i nv estt inn t hee stoc k mar kett

type of financial contract. This afterr t hee fi na nci all crisiss

should not be surprising since any United

Great France Italy Spain Germany States

Britain

financial transaction involves an % % % % % %

exchange of money today against Unweighted base 821 824 657 639 701 777

a promise of returning (more) More likely to

invest in stocks

money tomorrow. But the willing- and [EU: share,

ness to believe the promise and US: stock funds] 7 5 7 9 6 9

thus enter the transaction crucial- My attitude has

stayed the same 54 50 40 46 52 46

ly hinges on how much trust one Less likely to

has in the person that issues the invest in stock and

[EU: share, US:

promise. Below we examine some stock funds] 39 44 54 46 41 46

of the effects in greater detail. Answers to the question: “Compared with two years ago how has your attitude

to investing on the stock market changed, if at all?”

Source: September 2009 Financial Times/Harris Poll, Base: All EU adults in five

countries and US adults with savings/investments.

Trust and investment in risky

assets because the probabilities of the returns are intrinsical-

There is evidence that the level of trust affects investors’ ly uncertain (e.g. because they have a short history on

willingness to invest in stocks and, more generally, in which to estimate these probabilities) are more

risky assets. Stocks and risky assets lend themselves exposed to the risk of frauds and consequently are

more easily to opportunistic behaviour than simpler more easily placed among high-trust investors. When

securities. For instance Guiso et al. (2008) find that trust falls and becomes scarce one should see a decline

high-trust people are less likely to hold stocks in their in the demand for these instruments and an increase in

portfolio and conditional on holding, they invest lower the demand for simpler and more familiar securities.

shares in stocks. Since this finding is obtained using One of the consequences is that investors will revert to

variation in trust in a sample of Dutch investors, it can- instruments issued by national agents, perceived as

not be due to trust reflecting differences in the quality more “familiar” which become attractive as gener-

or effectiveness of legal enforcement (which is held con- alised trust vanishes. More generally, one of the conse-

stant) but rather the subjectively perceived probability quences of the crisis will be to shorten the distance

that people have of being cheated by the counterpart in between the investor and the issuer of the financial

a trade. This finding, which the same authors show instrument, thus reducing portfolio diversification and

holds in a sample of Italian investors, is consistent with amplifying a home bias. There is anecdotal evidence

the results of a recent Financial Times/Harris Poll that consistent with this idea. In some countries, in spite of

interviewed a sample of investors in the US and vari- the crisis, some banks – typically smaller, unsophisti-

ous European countries. It shows that in most coun- cated banks that in the past were not involved in the

tries people today have a lower propensity to invest in placement with their customers of structured securi-

stocks (Table 2.3). For instance, in Germany 41 percent ties and derivatives – have experienced a significant

report that today they are less ready to invest in stocks growth in deposits notwithstanding the crisis; on the

than before the crisis, and the percentages are similar in other hand, large, sophisticated banks that used to

other countries. Sapienza and Zingales (2008), using place complex securities have lost deposit market

the FTIS, find that those who plan to decrease their shares. One explanation is that investors revert back to

stock investments after the crisis are those who have the “familiar” for fear of being cheated by an interme-

less trust in financial markets and in particular the diary that deals with unfamiliar securities.

stock market. Thus, as a consequence of the fall in

trust, portfolios will likely be twisted markedly towards

safer securities and away from stocks. Trust and diversification across stocks and banks

One implication of the diminished trust is that

investors will form less diversified portfolios

Trust and investment in ambiguous securities because they will focus more on domestic assets.

Financial instruments that are more ambiguous either Guiso et al. (2008) show that this property is more

because of the complex nature of the contract or general and that investors that invest in stocks tend

62

EEAG Report 2010 Chapter 2

their own without any involve-

Figure 2.7 ment of the intermediary. All the

others relied on the intermediary

to a smaller or greater extent,

with 20 percent that delegated

either all decisions or a substan-

tial part to the intermediary. The

last two columns of Table 2.4

show the average level of trust

and the fraction of investors that

trusted the intermediary a lot. It

is clear that a fundamental ingre-

dient in the intensity of financial

delegation is the level of inves-

tors’ trust. Among those who rely

only on themselves when making

financial decisions 39.7 percent

to hold a more diversified stock portfolio when they trust the intermediary either substantially or a lot;

trust more. On the other hand, diminished trust among those who let the bank choose for them, the

towards intermediaries leads an investor to enter- share of those who trust a lot is 93 percent. Thus, the

tain multiple relations to diversify the risk of fall in trust should result in a marked decrease in del-

opportunistic behaviours by reducing exposure to egated investment. Since delegation is all the more

each one of them. We document this effect in necessary the more one invests in sophisticated securi-

Figure 2.7, which shows that in a sample of Italian ties, also through this channel there should be a move

investors, those who trust more (on a scale between towards simpler portfolios. These portfolios, however,

1 and 5 where 1 stands for very low trust and 5 for need not be necessarily better ones in the sense of pro-

very high trust) there is a strong negative correla- viding a higher return per unit of risk. Guiso and

tion between the level of trust and the number of Jappelli (2006) in fact find that investors who trust

bank relations an investor has. Both effects are more and delegate more are better diversified and are

costly: the first because one loses the benefits of able to attain more efficient portfolios.

diversification, the second because of the cost of

setting and maintaining multiple relations. Trust and the demand for insurance

Trust, demand for advice and delegation Though most of the literature has focused on the

effects of trust on investors portfolios, the fall in trust

Besides selling financial products, financial intermedi- involves all operators in the financial industry as shown

aries offer investors advice and information on how to in Table 2.1, including insurance companies. In fact,

allocate their financial wealth. Investors’ willingness since an insurance contract is itself a financial contract

to heed this advice depends on the trust they have in and as such is prone to the opportunistic behaviour of

the intermediary as much as their

decision to lend their savings to

the intermediary. One of the con- Tabl e 2. 4

sequences of the fall in trust is a Trustt an d d el eg atio n off fi na nci all d e cisionss

lower investors’ propensity to % share of

delegate financial decisions to the Average those trusting

level of a lot or

%

intermediary and to use his trust

Mode of making financial decisions responses substantially

advice. Table 2.4 shows the distri- I decide entirely on my own 12.0 2.98 39.7

I ask the bank to review my choice 30.4 3.92 82.4

bution of the extent of delega- I listen to my bank/advisor proposals but the

tion of financial decisions in a final word is always mine 37.7 3.88 78.3

UniCredit sample of Italian By and large I follow my bank/advisor 16.3 4.19 86.4

I let my bank/advisor decide everything 3.7 4.49 93.3

investors before the crisis. Only Source: UniCredit Italian Investors Survey, 2007 wave.

12 percent chose to decide on 63 EEAG Report 2010

Chapter 2 Box

x 2 .1

1 Prop osalss b y th e Fi na nci all S ta bility

y B oar d (FSB)) t o i mpro vee finan ci all re gul atio n

The predecessor of the Financial Stability Board (FSB) was the Financial Stability Forum (FSF), which was established by the G7

Finance Ministers and Central Bank Governors in 1997. The main idea was to create an institutional body that promotes coopera-

tion among national and international supervisory boards as well as international financial institutions to achieve more stability in

the financial system. Additionally it included representatives of the International Monetary Fund (IMF), the World Bank, the

Bank for International Settlements (BIS) and the OECD.

In November 2008 with the financial crisis in full swing, the G20 proposed extending the membership of the FSF by all those G20

countries that were not participating so far, and – at the same time – to broaden its mandate. This proposal was implemented at the

G20 Summit in April 2009 by founding the FSB, with Mario Draghi, Governor of the Banca d'Italia, being the first chairperson.

The mission of the FSB is to enhance stability by implementing strong regulatory and supervisory measures.

Already in April 2008 a report was produced which highlighted the main sources of the crisis and put forward concrete actions for

strengthening the financial system. A second report was published in April 2009 – with its major focus lying on reducing procyc-

licality and improving cross-border crisis management. Both reports constituted the basis for the Washington and London declara-

tions of the G20. Since then, the FSB is in charge of monitoring and co-ordinating the implementation of the action plan.

The main cornerstone of the FSB proposals is a less leveraged financial system in which all institutions have significantly higher

capital and liquidity reserves. However, in total there are nine building blocks that are addressed by the FSB:

1. Strengthening the global capital framework

A revised capital framework by the Basel Committee on Banking Supervision will become operative once the economic crisis is

overcome. Accordingly, minimum capital requirements will increase in their level and quality, and will be required to behave

countercyclically so that capital is accumulated during good times and may be used to overcome bad times. This step also in-

cludes the specification of a harmonised definition of capital in order to facilitate the comparability of institutions in different

countries.

2. Making global liquidity more robust

The financial crisis has shown that insufficient liquidity may have severe consequences even for banks that had a sound capital

basis. This problem is addressed by the Basel Committee by introducing a liquidity coverage ratio, thus creating a harmonised

framework that in particular is supposed to reduce cross-border liquidity shortages.

3. Reducing the moral hazard posed by systemically important institutions

A major source of instability was created by moral hazard due to “too big (or too complex) to fail”. Strengthening capital and

liquidity are steps in the right direction; however, further measures will be needed to overcome this problem. Until the end of

2010 measures to reduce systemic risk will be developed which – according to the FSB – may include actions to reduce the com-

plexity of group structures, specific additional capital requirements and promotion of stand-alone subsidiaries.

4. Strengthening accounting standards

In order to meet the objectives of convergence, transparency, and the mitigation of procyclicality, standard setters are required to

agree upon a single set of high quality global accounting standards. However, the International Accounting Standards Board

(IASB) and the US Financial Accounting Standards Board (FASB) are considering different accounting approaches that may lead

to significant differences in banks’ total assets. The FSB strongly encourages the IASB and FASB to cooperate with supervisors,

regulators and other constituents in order to converge and improve their accounting standards with respect to the required amount

of credit information, and the simplification of accounting principles for financial instruments.

5. Improving compensation practices

In order to improve the effectiveness of compensation policies, the FSB Principles for Sound Compensation Practices outline

private and official sector action. The principles need to be applied to significant financial institutions and systemically relevant

firms, and have to be implemented in all major financial centres in a fast and coordinated way. Constant and independent supervi-

sion ensures that all necessary improvements are made.

6. Expanding oversight of the financial system

It is necessary that not only the banking sector but the broader financial sector is subjected to appropriate oversight and regulation.

Such a broad framework of regulation should particularly take hedge funds and rating agencies into account.

7. Strengthening the robustness of the OTC derivatives market

The risk in the market for over-the-counter (OTC) derivatives has to be reduced. Therefore, international standards need to be

established that take full account of counterparty risks, the benefits of centrally cleared contracts and collateralisation. The regula-

tion should ensure that equivalent standards are met outside the banking sector.

8. Re-launching securitisation on a sound basis

The revival of securitisation markets is crucial for the provision of credit to the real economy, whereas the official sector is re-

quired to provide a framework that ensures discipline in the securitisation market. In 2010, the main goals for supervisors and

regulators are the establishment of rules for banks’ management and disclosures, and the alignment of incentives of issuers with

investors. If necessary, measures may be adjusted in order to reduce complexity and enhance transparency.

9. Adherence to international standards

In order to strengthen adherence to international regulatory and prudential standards, the FSB framework intends to facilitate the

provision of comprehensive and updated compliance information and to identify non-cooperative jurisdictions. An important

means for achieving these goals is the system of peer reviews among FSB members to assess the implementation of international

financial standards and to discuss additional steps.

Source: http://www.financialstabilityboard.org. 64

EEAG Report 2010 Chapter 2

the insurance company, the fall in Figure 2.8a

trust should also affect the de-

mand for insurance. Guiso et al.

(2008) find that, in the sample of

Dutch investors they examined,

individuals that trust less are less

likely to purchase insurance. In an

interesting paper that relies on a

field experiment in Indian villages,

Cole et al. (2009) show that over-

coming mistrust can result in a

significant increase in peasants’

adoption of insurance contracts

and Guiso and Schivardi (2009)

find that in a sample of small

businesses a critical factor limiting

entrepreneurs willingness to in-

sure their firm is mistrust towards

insurance companies. To sum up,

given the importance of trust in

all financial contracts, the fall in

trust towards all segments of the

financial industry will give rise to

a generalised flight from financial

trades and particularly deal from Figure 2.8b

those trades that are severely

exposed to opportunistic be-

haviour.

Rebuilding trust in finance

As illustrated, the fall in trust is

likely to have pervasive effects on

people’s willingness to enter into

financial contracts and can thus

hamper the process of financial

development. Insofar as it results

in a shift towards safer assets, it

will push up the equity premium

and make equity financing more

expensive. This may have conse-

quences for fast growing and

innovative firms that depend

more heavily on this type of

financing. Similarly, if the in-

creased mistrust results in a pref-

erence for instruments with

shorter maturity, it will raise the

cost of long-term financing, hampering projects with The regulatory approach

high-yields but longer maturities. Because of this it is One approach, so far the only one that has been fol-

important to understand how one can rebuild trust in lowed to raise trust, is to enhance the intensity of

financial markets and intermediaries. Here we will ex- financial regulation. This approach, shared by many

amine some avenues. 65 EEAG Report 2010

Chapter 2 governments particularly in Europe, has been the sub- would be capable of increasing the trustworthiness of

the intermediaries and because of this the trust of the

ject of several of the recent G20 meetings and of the investors is still to be proved. The evidence so far from

proposals for intervention that are being discussed at cross country correlations is that countries with

the Financial Stability Board (see Box 2.1). stronger regulation have lower average levels of trust,

not higher! (See Aghion et al. (2010); Pinotti (2008);

Needless to say, many of the regulatory proposals Carlin et al. (2009).)

that are under scrutiny go beyond the purpose of

rebuilding trust. Rather, they are justified by regula- From the viewpoint of individual investors and of

tory failures that have become manifest during the the regulation of their relation with financial inter-

financial crisis. In fact, the set of proposals under dis- mediaries, the closest proposal that can help rebuild

cussion is ample and heterogeneous and ranges from trust is the creation of a consumer protection

more stringent capital requirements to the establish- agency, as proposed by the Obama administration.

ment of new authorities for macro-prudential super- The agency would oversee consumer financial prod-

vision, the breaking up of banks into smaller units to ucts which have been regulated in the past but

deal with too-big-to-fail issues, to policies aimed at whose oversight was exposed as lax. Another initia-

lessening the impact of bank failures and the associ- tive that has been taken very recently is the creation

ated contagion risks through regulatory constraints in the US of a Financial Fraud Enforcement task

on connectedness. Many of these policies, assuming 7 Interestingly, as

force to combat financial crimes.

they will be finally adopted, will most likely affect the made clear by Attorney General Eric Holder, the

of the intermediaries and may

perceived solvency task force is intentionally created to address the fall

result in a lower likelihood of future crises. Some in trust induced by the scandals that have been

policies – such as the limitations that the Financial brought about by the financial crisis. He notes: “We

Stability Board proposed for implementation in face unprecedented challenges in responding to the

September 2009 on the structure of compensation of financial crisis that has gripped our economy for

top managers at financial institutions – may help to the past year. Mortgage, securities, and corporate

assuage investors anger for the losses suffered during fraud schemes have eroded the public’s confidence

the crisis and their indignation at the high level of in the nation’s financial markets and have led to a

compensation for top executives in the financial growing sentiment that Wall Street does not play by

industry, thought to be responsible for their losses. the same rules as Main Street. Unscrupulous exec-

But these measures are likely to have little impact on utives, Ponzi scheme operators, and common crim-

the trust investors have in financial intermediaries inals alike have targeted the pocketbooks and

and markets. Rather, it is the drop in trust that retirement accounts of middle class Americans, and

increases the demand for regulation and builds con- in many cases, devastated entire families’ futures.

sensus around it. In fact, those who mistrust banks We will not allow these actions to go unpunished,

and financial intermediaries tend to favour tighter which is why President Obama has established this

regulations. To show this link we rely on a set of Financial Fraud Enforcement Task Force to inves-

questions that have been asked in the FTIS on tigate and prosecute fraud and financial crime …

whether the respondent is supportive or not of This Task Force’s mission is not just to hold

tighter regulation of US financial intermediaries and accountable those who helped bring about the last

large corporations and whether he agrees on setting financial meltdown, but to prevent another melt-

caps on the compensation of top managers in finan- down from happening.”

cial corporations. Figure 2.8 shows the correlation

between the intensity of trust towards financial inter- Because these initiatives are both specifically target-

mediaries (Panel A) and bankers (Panel B) and the ed to protecting investors from abuses they may

support for regulation measured by the fraction of actually contribute to rebuilding trust. But there are

people that agrees with the policy. The fraction of also reasons to believe that by themselves, these

those supporting a more stringent regulation is high- interventions may have limited impact. Concerning

er among those whose trust has fallen during the cri- the consumer protection agency and more generally

sis than among those who continue to trust banks regulatory interventions, because they are imposed

and financial markets. from outside the industry perceives the costs of the

But causality here most likely runs from the fall in

trust to the demand for regulation. The latter, in turn, 7 See http://www.justice.gov/ag/speeches/2009/ag-speech-091117.html.

66

EEAG Report 2010 Chapter 2

regulations but not the benefits; hence financial playing partners manipulated to resemble themselves

they trusted them more than when the face of an

intermediaries will tend to circumvent their applica- unknown person was shown. Guiso and Schivardi

tion, with greater success the weaker the actual report evidence that is consistent with De Bruine’s

enforcement is. Since investors anticipate this, they (2002) findings. In a survey of a sample of small busi-

may not revise their trust priorities significantly. nessmen interviewees (280 overall) were asked to

Furthermore, sometimes financial regulation, even report, at the end of the face-to-face interview, their

when designed to protect investors, may be bother- judgment about the trustworthiness of businessman

some for them as well. Because of this and in order that they interviewed on a scale between 0 and 10

to limit the burden, they will be willing to tolerate (0 = totally untrustworthy, 10 = fully trustworthy).

intermediaries’ misapplications of these rules. A Interviewees also reported their opinion (again on a

good example is the recent set of norms imposed by scale between 0 and 10) on how much affinity they

the EU’s Markets in Financial Instruments Directive felt to the businessman (0 = no affinity, 10 = com-

(MiFID) to classify investors according to their abil- plete affinity). The data show two interesting facts:

ity to make financial decisions and their capacity to first, the more a person feels affinity the more he

bear financial risk. To achieve this classification trusts; second, while at low levels of affinity the level

banks can obtain information from their customers, of trust towards the businessman is highly variable, at

for example, by asking them to fill in specific ques- high levels of affinity one trusts fairly reliably. It is

tionnaires. But because the latter are costly to sub- reasonable to assume that one tends to trust people

mit, banks have all the incentives to minimise the that are not much different from oneself. This ten-

effort and propose minimal questionnaires, possibly dency to trust those who are similar is also true when

based on investors’ self-classification (so as to avoid similarity is measured along various dimensions,

any responsibility for misclassifications); since filling including cultural and genetic distance among people

in these forms is bothersome for investors too, they (Guiso et al. 2009). Thus, one possible strategy to

care little about the quality of the information that raise trust is to improve the match between investors

banks collect and will instead join them in minimis- and the manager of the relation at the intermediary,

ing the effort put in collecting the MiFID data. But for instance assigning a manager of the same gender

this contributes to the failure of MiFID objective: 8

and geographical origin to the investor. While this

limiting banks opportunistic behaviour by forcing may help in raising the average trust that investors

them to segment their clientele and restrain products have towards their bank/broker, it is unclear that it

that can be sold to unsophisticated and risk adverse helps raising the trust of those who lost it. To raise

investors. Anticipating this, people’s trust in banks is the latter one needs to set up mechanisms that signal

likely to change little. in a credible way that the intermediary has become

more trustworthy because, thanks to the mechanism,

there are weaker incentives to adopt predatory

An industry-based strategy actions towards the investors. Below we will discuss

some possible mechanisms.

Losing investors’ trust is very costly for the financial

industry. Since this is the case one would expect that

intermediaries have strong incentives to take actions A rating system that even the most (financially)

to re-build their reputation and re-gain the trust of illiterate investor can understand

their customers. Today one of the big questions that

any financial operator is confronted with is how to One possibility is to adopt a rating system aimed at

rebuild the trust of their investors. reducing the scope for exploiting conflicts of interests

that often arise in universal banks that manage the

Unfortunately there are no easy recipes on how A savings of the investors. The strategy followed so far

may convince B to re-consider his opinion about the by the regulators to control conflicts of interests is to

trustworthiness of A. The recent literature on trust

has shown evidence that B would trust A more if A is

“similar” to B in some dimension. In a well-known 8 Of course if matching according to similarity is an effective way to

raise trust, markets should be doing it already. If they already do,

trust game experiment, De Bruine (2002) reports the then this is not a relevant policy. If they do not it may be because this

type of matching entails costs that exceed the benefits, in which case

effects of a manipulation of facial resemblance on and the proposal would have little practical value. But it may also be

that they do not match according to similarity because they ignore

players’ willingness to trust the opponent. She finds its potential benefits. We cannot rule out this possibility; after all,

that when subjects were shown faces of ostensible research showing the trust effects of similarity is quite new.

67 EEAG Report 2010

Chapter 2 impose tighter disclosure requirements on the inter- the bank. Needless to say there could be many imple-

mediaries. Yet this strategy has proved to be faulty or mentation problems, including the fact that finding

insufficient. The main problem with disclosure is that independent and uncorrupted rating agencies may, as

it takes for granted that investors are able to under- the crisis has shown, not be a trivial issue. But the

stand what is disclosed and its implications in terms biggest problem, in our view, is initialising the

of incentives of the intermediary. However, the avail- process. If the exploitation of conflicts of interests

able evidence on poor levels of financial literacy and and misbehaviour more generally is a diffused prac-

knowledge of the majority of the investors in almost tice in the industry, then even a honest intermediary

all countries (see e.g. Lusardi and Mitchell (2009), (but still sensitive to short-run profitability) may find

Guiso and Jappelli (2008) Jappelli (2009)), even those it difficult to subject its bank to the “bank fairness

with high levels of achieved education, casts serious index” and give up a source of profit, as this may

doubts on the validity of the assumption. Relying on concede an advantage to its competitors. To put it

the loss of reputation as a deterrent for intermedi- differently, an outcome where low trustworthiness is

aries misbehaviour, and thus as a mechanism to raise pervasive may be stable. It may be unwise to play

trust in financial intermediaries, requires not only honestly when everyone else is cheating; if an inter-

that information about potential conflicts is made mediary does not cheat while all the others do, it

available but also that the investor has the ability to misses the upside. If it cheats when all the others do,

elaborate this information. For this to be the case one there is no downside as “cheating” becomes the pre-

has to make the disclosed information understand- dominant rule of behaviour and one cannot punish

able to the least experienced and financially knowl- the whole industry when all follow the same practice.

edgeable investor – i.e. to the typical customer of a Today it is perhaps easier to circumvent this problem

bank. One way of doing this is to rely on a third party given the greater value that rebuilding reputation has

to rate banks on the basis of their trustworthiness for any intermediary. Furthermore, since the incen-

and fairness when dealing with their customers and tive to behave in the same way as the others do natu-

when managing their portfolios and providing finan- rally implies that the financial industry can either set-

cial advice. This “bank-fairness index” may be tle on a bad equilibrium in which all cheat or instead

reported on a scale between 0 and 10, with higher val- in a good equilibrium where all play honestly, one

ues meaning a more reliable intermediary – a metric can think of a role for regulation/supervision that

that any investor can understand. The “bank-fairness encourages intermediaries to coordinate on a differ-

index” is similar to the rating system adopted for ent, no-cheating equilibrium.

issuers of specific securities and its role would be

analogous to that of standard rating: making avail-

able to the investors synthetic information that aggre- A trust-based compensation scheme

gates the judgment of an expert observer (and based A second, more direct mechanism to raise trust is to

on a multitude of data) on the quality of the issuers, provide incentives to build it. If the compensation of

subject to periodic revisions. In contrast to standard the investor’s manager depends on the level of trust

rating, the “bank-fairness index” is aimed at measur- investors have in their asset manager, the latter have

ing a bank’s ability and reliability in its role as dele- strong incentives to behave in a trustworthy manner

gated portfolio manager and in general as provider of and this, perhaps slowly, will raise the investors’ trust

financial advice that un-experienced investors use in and his willingness to invest. As trust increases, the

their financial decisions. Banks with an internal investor will also tend to concentrate more assets

organisation that discourages the exploitation of with a single manager, thus avoiding costly duplica-

conflicts of interest or that distributes easily under- tions of relationships. A mechanism of this sort

standable information to its customers, that allocate could be implemented for instance by relying on the

qualified personnel to financial consulting services, information that intermediaries have to collect from

etc. would obtain a high rating, attracting more cus- the investors to comply with the EU’s Markets in

tomers and this would provide enough incentives for Financial Instruments Directive (MiFID). The infor-

them to adopt actions that discourage the exploita- mation in this directive is presently essentially per-

tion of conflicts of interests. These banks would be ceived as a burden and unutilised. One could insert

compensated for the extra costs they incur with specific questions that the investors can report

increased trust from their customers. Reliance on a anonymously on how much they trust the intermedi-

rating system – which is a voluntary choice of a bank ary, the portfolio manager and in general the person

– is credible precisely because it entails some costs to 68

EEAG Report 2010 Chapter 2

they deal with when making financial decisions. Promoting investors’ financial education

Manager pay could then respond to the level of trust A third type of strategy is to take actions that pro-

(or its change) of the pool of customers he is respon- mote the financial education of the investors – for

sible for. One benefit of the trust-based compensa- instance transparently lobbying with the government

tion scheme is that it naturally leads the bank man- for having financial education taught at schools,

ager to adopt a long horizon. Since building trust making financial education material certified by

takes time and is accumulated only slowly, if only third parties available to investors etc., since people

because those with low levels of trust do not experi- with lower levels of financial education and financial

ment (or experiment less) and thus do not learn (or experience are more likely to be victims of financial

learn slowly), they cannot learn immediately the deception by intermediaries. The main reason is that

increased trustworthiness of the bank manager. unsophisticated investors are more vulnerable to

Furthermore, since trust is slow to accumulate but deception because they are more dependent on the

fast to vanish, once a reputation of trustworthiness intermediary advice for their financial choices.

is obtained it becomes costly to dispel it, strengthen- Second, they are also more subject to interpretation

ing the incentives to behave in a trustworthy manner. problems when investments result in negative returns

Obviously, this too, like all incentive schemes, can be and are thus more likely to think that they have been

distorting. In particular, if one encourages building cheated. Consistent with this view, Butler et al.

trust one provides incentives not only to create but (2009) find that the probability a person being de-

also to extort trust especially if this is a less costly ceived by a bank or insurance company is much

activity than creating trust by behaving in a trust- higher for people with low levels of education.

worthy manner. One way to limit this possibility is to Furthermore, this probability is higher also for peo-

integrate an investor’s opinions with those of some ple that – holding constant their level of education –

internal auditing committee at certain frequencies. live with parents with low education. This feature

Another is to rely on the legitimate interest of the has an important implication: since the family is an

other managers for having their colleagues behave important channel through which reliable financial

honestly, particularly those that are located in close education is obtained, raising the level of financial

proximity. The reason for this is that if manager A education has important spillovers through the fam-

cheats his investors, also the trust of the investors of ily and informal (but reliable) network channels. An

manager B will be affected, as the Quint Tatro tale in intermediary that promotes financial education sig-

the introduction illustrates. Thus, one could rely on nals its intention to be willing to deal with experi-

an internal reporting system that allows and actual- enced and sophisticated investors, with enough abil-

ly encourages managers to reports cases of abuses ity not to fall victims to financial abuses and distort-

and manipulation of investors’ trust. ed advice. Because of this the investors’ trust should

To strengthen the scheme even further, also the com- increase. Needless to say investment in financial edu-

pensation of the top management of the bank, in par- cation pays off in the very long run; however the

ticular its CEO (and maybe also the board of direc- return to the intermediary in terms of increased

tors) could be linked to the trust index of the bank trustworthiness may be more immediate if the inter-

customers. mediary’s commitment to transfer power to the

investor through this channel is credible. Credibility

To sum up, the adoption of a “trust-based compen- would be enhanced if the sponsoring of financial

sation scheme” is a practical way to induce a financial education programmes is part of a broader policy

organisation and its workers to limit the incentives to aimed at limiting intermediaries’ incentives to

deceive poorly informed investors and to treat them deceive investors, such as the trust-based compensa-

fairly by always acting in their best interest. Since this tion scheme and the bank fairness index.

commitment is translated into a compensation

scheme, it should be credible and thus able to modify

investors’ beliefs. In other words, trust is the in- Conclusions

vestors’ belief that those who manage their savings The dramatic drop in trust following the revelation

and provide them with financial advice are trustwor- of information of pervasive cheating in financial

thy. For intermediaries hoping to increase investors’ markets is likely to have a very strong negative

trust, the only way is to invest in increasing their impact on investors’ willingness to bear risk and

trustworthiness. 69 EEAG Report 2010

Chapter 2 thus on the cost of risk capital. Insofar as trust lev- Guiso, Luigi, Paola Sapienza and Luigi Zingales (2009), “Trust and

the Fragility of Financial Markets”, mimeo.

els were exceedingly optimistic, their downward Guiso, Luigi and Tullio Jappelli (2008), “Financial Literacy and

revision should be partially welcome as it may help Portfolio Diversification”, EIEF working paper 8/12.

punish dishonest financiers and help restore market Jappelli, Tullio (2009), “Economic Literacy: An International

discipline. Comparison”, CSEF working paper 209.

Lusardi, Annamaria and Olivia Mitchell (2009), “Financial Literacy.

Evidence and Implications for Financial Education”, TIAA-CREF

However, since trust has fallen across the board, its Institute Trends and Issues, May.

decline also affects the honest intermediaries, limit- Pinotti, Paolo (2008), “Trust, Honesty and Regulations”, MPRA

ing the flow of capital to industry in general. We paper 7740.

have proposed a number of measures to rebuild Sapienza, Paola and Luigi Zingales (2008), “The Financial Trust

Index. The results: Wave I”,

trust. The measures proposed all try, from different http://www.financialtrustindex.org/resultswave1.htm

angles, to limit the scope for intermediaries’ oppor- Carlin, B., F. Dorobantu and S. Viswanathan (2009), “Public Trust,

tunist behaviour – that is to raise their trustworthi- the Law, and Financial Investment”, Journal of Financial Economics,

forthcoming.

ness – and because of this, increase trust. In each Knell, Markus and Helmut Stix (2009), “Trust in Banks? Evidence

case, the policy is not imposed; adhering to it is from normal times and from times of crises”, Oesterreichische

Nationalbank, mimeo .

instead to the discretion of the intermediary.

However, as we have argued, there is no automatic

mechanism that guarantees that intermediaries will

all agree to voluntarily adopt these policies. Rather,

if dishonest behaviour is dominant among interme-

diaries, even the honest ones may on their own be

unwilling to adopt these measures and help the

economy move to a better outcome where competi-

tion drives out dishonest behaviour. We have also

argued that regulation by itself, without the involve-

ment of the intermediaries, may fail to restore trust;

however regulatory agencies may play a very impor-

tant role in coordinating the selection of the honest

equilibrium. For instance, using “moral suasion” to

persuade even a small but important number of

intermediaries to “play the honest game” may be

enough to trigger a response of the same type by the

dishonest ones and influence the whole industry

outcome.

References

Aghion, Philippe, Yann Algan, Pierre Cahuc and Andrei Shleifer

(2010), “Regulation and Distrust”, Quarterly Journal of Economics,

forthcoming.

Arrow, Kenneth (1972), ”Gifts and Exchanges”, Philosophy and

1, 343–62.

Public Affairs,

Gambetta, Diego (2000) “Can We Trust Trust?” in Gambetta, Diego

(ed.) University of

Trust: Making and Breaking Cooperative Relations,

Oxford, 213–37.

Cole, Shawn, Xavier Giné, Jeremy Tobacman, Petia Topalova, Robert

Townsend and James Vickery (2009) “Barriers to Household Risk

Management: Evidence from India”, mimeo Harvard University.

DeBruine, Lisa M. (2002), “Facial Resemblance Enhances Trust”,

Proceedings of the Royal Society of London B, 269 (1948): 1307–12.

Guiso, Sapienza, and Zingales (2004), “The Role of Social Capital in

Financial Development”, 94(3), 526–56.

American Economic Review,

Guiso, Sapienza, and Zingales (2008), “Trusting the Stock Market”,

63(6), 2557–600).

Journal of Finance,

Guiso, Sapienza, and Zingales (2009), “Cultural Bias in Economic

Exchange?” forthcoming.

Quarterly Journal of Economics, 70

EEAG Report 2010 Chapter 3

F with the overall changes to fiscal positions? What

ROM FISCAL RESCUE TO are the existing and projected levels of public debt

GLOBAL DEBT relative to GDP?

• Was the consensus correct? Did we need a fiscal

stimulus? Can we identify the effects?

1. Introduction • There is now significant concern about debt-to-

GDP ratios. Are they too high? How and when

A broad consensus seemed to have been reached since should they be reduced?

the onset of the financial and economic crisis that

governments needed to undertake collective action to

provide a fiscal stimulus to prevent a deep and long- 2. What has happened to fiscal deficits during the

lasting recession. crisis?

For example, a much-cited note by the IMF at the In 2009 every EU member-state government had a

end of 2008 argued that the “optimal fiscal package budget deficit. In almost all cases, these deficits are

should be timely, large, lasting, diversified, contin- expected to rise in 2010. These deficits varied consid-

1

gent, collective and sustainable”. erably between countries, and the reasons for the size

of the deficit also varied. Most countries introduced

The European Council of the EU agreed a “European some discretionary fiscal stimulus in response to the

Economic Recovery Programme” (EERP) in De- financial and economic crisis, by cutting taxes or

cember 2008, which called for a discretionary fiscal increasing spending. These discretionary measures

stimulus of at least 1.5 percent of GDP. This was re- were small relative to the size of the deficits.

garded as “a crucial contribution to tackling the glob-

al economic crisis in which all countries with suffi- In this section we present some evidence on the pat-

cient fiscal space need to play a role in filling short- tern of the deficits both over time (since 2004, and up

2

term demand gaps”. to 2010 using European Commission forecasts), and

across countries. We also describe the extent to which

At its meeting in April 2009, the G20 stated: “We are these deficits were generated by discretionary mea-

undertaking an unprecedented and concerted fiscal

expansion, which will save or create millions of jobs sures, and the extent to which they were due to reduc-

which would otherwise have been destroyed, and that tions in tax revenues or rises in expenditure.

will, by the end of next year, amount to $5 trillion,

raise output by 4 percent, and accelerate the transition A starting point is the measurement of government

to a green economy. We are committed to deliver the debt. Measuring government indebtedness is diffi-

scale of sustained fiscal effort necessary to restore cult, since in principle it should include the extent

growth.” As recently as September 2009, the G20 stat- of future liabilities due to pension provisions and

ed: “We will continue to implement decisively our nec- other factors. There are also difficult issues with

essary financial support measures and expansionary respect to interventions in the banking sector. For

monetary and fiscal policies, consistent with price sta- example, if a government guarantees a loan, then

bility and long-term fiscal sustainability, until recov- typically that is not recorded as an increase in gov-

ery is secured.” ernment debt, even though the government has a

contingent liability. Box 3.1 describes how such

The issues addressed in this chapter are: financial sector interventions are typically recorded

• What has happened so far? What discretionary in national accounts. The figures shown in this

stimulus has taken place? How does this compare chapter are taken from Eurostat and the European

Commission, which are based on a consistent

1 IMF (2008). approach across the EU.

2 European Commission (2009a). 71 EEAG Report 2010

Chapter 3 Bo x 3. 1 Measurin g th e im pactt on

n governm entt debtt off financi all sectorr inter ventionss

EU governments have made significant interventions into the financial sector since the beginning of the financial crisis.

The classification of the costs of these interventions, and their effect on various measures of government debt, are

1

European System of Accounts 1995

generally estimated in accordance to the .

Several accounting issues arise with respect to financial sector interventions. One is whether the intervention represents an

institution becoming part of the public sector, and hence its debt becoming a public sector liability.

A second issue is which aspects of the financial accounts of an institution are relevant for measuring public sector debt.

public sector net debt

The most commonly-used measure of public sector debt is known as . This includes the financial

current

liabilities of financial companies which have moved into public ownership. However, only financial assets are

netted out against these liabilities. Because other financial assets are not included, the measure does not give a realistic

general government

indication of the increase in the overall net indebtedness of the public sector. An alternative measure,

gross debt , does net off all financial assets, and can therefore give very different indications of debt.

public sector net debt

For example, it is estimated that, for the UK, the total increase in as a result of financial sector

interventions is approximately £1.1 trillion to £1.6 trillion, which could raise the debt-to-GDP ratio in the UK by more

than 100 percent. (Only a small part of this increase is included in the figures shown here; by far the largest part of this

general government

reflects the public ownership of Royal Bank of Scotland, Lloyds TSB and HBOS). But the increase in

gross debt is estimated at only around £77 billion. While this is still, of course, a large amount, it gives a very different

picture of the extent of the financial sector interventions.

Finally, government debt guarantees – including those in place before the crisis, and those introduced during the crisis –

are typically not included in the figures for debt, even though they represent a contingent liability on the government.

Table 3.6 on page 76 gives an indication of the extent of the public sector interventions in the banking sector during the

crisis.

1 See Eurostat (1995). The European Committee on Monetary, Financial and Balance of Payments Statistics (CMFB) reviewed financial

interventions and reported its opinion on their appropriate accounting treatment in March 2009. These were reviewed in detail by Kellaway (2009).

Table 3.1 shows the public sector

balances of each member state Tabl e 3. 1

since 2004; 2009 and 2010 are Budg ett bal an cess off EU

U m e mb err stat es,, 2 00 4–2 01 0

projections made by the Euro- perc e ntt GDP

pean Commission. It is clear that 200 4 200 5 200 6 200 7 200 8 200 9 201 0

deficits rose sharply in 2009. In Austria – 4.5 – 1.7 – 1.7 – 0.7 – 0.5 – 4.2 – 5.3

Belgium – 0.4 – 2.8 0.2 – 0.3 – 1.2 – 4.5 – 6.1

2007, the EU as a whole had a Bulgaria 1.6 1.9 3.0 0.1 1.5 – 0.5 – 0.3

deficit of only 0.8 percent of Cyprus – 4.1 – 2.4 – 1.2 3.4 0.9 – 1.9 – 2.6

GDP. That rose to 2.3 percent in Czech Rep. – 2.9 – 3.6 – 2.6 – 0.6 – 1.4 – 4.3 – 4.9

Denmark 1.9 5.0 5.0 4.5 3.6 – 1.5 – 3.9

2008, and then jumped to 6 per- Estonia 1.7 1.5 2.9 2.7 – 3.0 – 3.0 – 3.9

cent in 2009, and to 7.3 percent Finland 2.2 2.6 3.9 5.2 4.1 – 0.8 – 2.9

in 2010. France – 3.6 – 3.0 – 2.3 – 2.7 – 3.4 – 6.6 – 7.0

Germany – 3.8 – 3.3 – 1.5 – 0.2 – 0.1 – 3.9 – 5.9

Greece – 7.4 – 5.2 – 3.1 – 3.9 – 5.0 – 5.1 – 5.7

Romania is the only country Hungary – 6.4 – 7.8 – 9.3 – 4.9 – 3.4 – 3.4 – 3.9

that reduced its deficit between Ireland 1.4 1.7 3.0 0.2 – 7.1 – 12.0 – 15.6

Italy – 3.6 – 4.4 – 3.3 – 1.5 – 2.7 – 4.5 – 4.8

2008 and 2009, but then it had a Latvia – 1.0 – 0.4 – 0.5 – 0.4 – 4.0 – 11.1 – 13.6

relatively high deficit of 5.4 per- Lithuania – 1.5 – 0.5 – 0.4 – 1.0 – 3.2 – 5.4 – 8.0

cent of GDP even in 2008. Luxembourg – 1.1 0.1 1.4 3.6 2.6 – 1.5 – 2.8

Some countries have seen a Malta – 4.7 – 2.9 – 2.6 – 2.2 – 4.7 – 3.6 – 3.2

Netherlands – 1.8 – 0.3 0.6 0.3 1.0 – 3.4 – 6.1

notable worsening of the fiscal Poland – 5.7 – 4.3 – 3.9 – 1.9 – 3.9 – 6.6 – 7.3

position. Ireland jumped from a Portugal – 3.4 – 6.1 – 3.9 – 2.6 – 2.7 – 6.5 – 6.7

small surplus in 2007 to a de- Romania – 1.2 – 1.2 – 2.2 – 2.5 – 5.4 – 5.1 – 5.6

Slovakia – 2.4 – 2.8 – 3.5 – 1.9 – 2.2 – 4.7 – 5.4

ficit of 12 percent of GDP in Slovenia – 2.2 – 1.4 – 1.3 0.5 – 0.9 – 5.5 – 6.5

2009. Likewise, Latvia went Spain – 0.4 1.0 2.0 2.2 – 3.8 – 8.6 – 9.8

from a small deficit in 2007 to a Sweden 0.6 2.0 2.4 3.8 2.5 – 2.6 – 3.9

UK – 3.3 – 3.3 – 2.6 – 2.6 – 5.4 – 11.5 – 13.8

deficit of 11 percent of GDP in EU27 – 2.9 – 2.5 – 1.4 – 0.8 – 2.3 – 6.0 – 7.3

2009. The UK also moved in a Source: 2004–2008, Eurostat; Forecasts 2009–2010 European Commission (2009a).

similar way. 72

EEAG Report 2010 Chapter 3

A small number of countries Tabl e 3. 2

have had substantial deficits for a Fiscall sti mul uss m e asuress i n 2 00 9/10

0

number of years: notably Greece, perc e ntt GDP

P

Hungary, Italy, Malta, Poland, 200 9 201 0

Portugal and to a lesser extent, Total Expenditure Revenue Total

the UK. There are significant dif- Austria 1.8 0.4 1.4 1.8

Belgium 0.4 0.2 0.2 0.4

ferences across counties in 2009, Bulgaria 0.0 0.0 0.0 0.0

ranging from Bulgaria with a Cyprus 0.1 0.1 0.0 0.0

deficit of only 0.5 percent of Czech Republic 1.0 0.5 0.5 0.5

Denmark 0.4 0.3 0.1 0.8

GDP, to Ireland with a deficit of Estonia 0.2 0.2 0.0 0.3

12 percent of GDP. Finland 1.7 0.6 1.1 1.7

France 1.0 0.7 0.3 0.1

Germany 1.4 0.6 0.8 1.9

These deficits were only partly Greece 0.0 0.0 0.0 0.0

due to discretionary responses Hungary 0.0 0.0 0.0 0.0

to the economic and financial Ireland 0.5 0.3 0.2 0.5

Italy 0.0 0.2 – 0.2 0.0

crisis. This is shown in Table 3.2 Latvia 0.0 0.0 0.0 0.0

which indicates the size of the Lithuania 0.0 0.0 0.0 0.0

discretionary fiscal stimulus in Luxembourg 1.2 0.1 1.2 1.4

Malta 1.6 1.3 0.3 1.6

each country in 2009 and in Netherlands 0.9 0.4 0.5 1.0

2010 (taking into account those Poland 1.0 0.3 0.7 1.5

measures already announced). Portugal 0.9 0.9 0.0 0.1

These are measured relative to Romania 0.0 0.0 0.0 0.0

Slovakia 0.1 0.1 0.0 0.0

the position in 2008, recording Slovenia 0.6 0.5 0.1 0.5

all discretionary changes in Spain 2.3 1.0 1.3 0.6

these two years. Sweden 1.4 0.6 0.8 1.6

UK 1.4 0.4 1.0 0.0

EU27 1.1 0.5 0.6 0.7

Almost all EU governments in- Figures for 2010 represent changes with respect to 2008, i.e. include permanent

troduced a fiscal stimulus in measures taking effect in 2009 plus the net effect of measures taking effect in

2010.

2009, though some maintained Source: European Commission (2009a).

a neutral position. The largest

discretionary changes were in

Spain, with a stimulus of nia and Slovakia both have revenues of 32.2 percent

2.3 percent of GDP, made up of an increase in of GDP, and expenditures of 38.5 percent and

spending of 1 percent and a reduction in taxes of 38.3 percent respectively.

1.3 percent. On average, though, the EU as a whole

introduced a discretionary stimulus of only 1.1 per- Across the whole of the EU, revenues have been very

cent of GDP. Note though that evidence presented consistent as a proportion of GDP, at just over 44 per-

in Chapter 1 suggests that changes in structural cent in each of the 7 years shown. Revenues in 2009

deficits – that part of the deficit that is not auto- and 2010 are lower than in the preceding years, but

matic – were larger than implied by the discre- only fractionally. There is more variation over time for

tionary responses listed in this chapter. individual countries, although in most countries rev-

enues typically only changed in 2009 by less than one

Tables 3.3 and 3.4 split up the deficits in each country percent of GDP.

by considering the size of tax revenues (Table 3.3) and

public spending (Table 3.4) as a proportion of GDP. The substantial rises in deficits therefore appear to be

Of course, there is considerable variation between mainly driven by increases in spending as a propor-

countries. Not surprisingly, the Scandinavian coun- tion of GDP, rather than reductions in taxation as a

tries have the highest revenues: in 2009, Sweden has proportion of GDP. Some countries – typically those

revenues of 53 percent of GDP, Denmark 52.8 per- with large increases in their deficits – have seen sub-

cent, and Finland 52 percent. Their expenditures are stantial rises in spending as a proportion of GDP. But

similarly high: 56.6 percent for Sweden, 55 percent for note that GDP fell in many countries in 2009. The rise

Denmark and 52.8 percent for Finland. Some of the in the spending ratio may therefore not represent only

newer members states are at the other extreme: Roma- a real increase in spending but also a reduction in

73 EEAG Report 2010

Chapter 3 GDP. By contrast, falling nation-

Tabl e 3. 3 al income tends to reduce tax rev-

Rev en uess off EU

U m em b err states,, 20 04 –2 010

0

perc e ntt GDP

P enues automatically: so it is likely

200 4 200 5 200 6 200 7 200 8 200 9 201 0 that revenues as a proportion of

Austria 49.5 48.2 47.7 48.0 48.2 47.0 47.3 GDP would remain relatively

Belgium 49.1 49.4 48.7 48.1 48.4 48.5 48.2 constant in a downturn.

Bulgaria 41.3 41.2 39.5 41.5 39.0 40.8 40.9

Cyprus 38.7 41.2 42.2 46.4 44.9 44.1 44.1

Czech Rep. 42.2 41.4 41.2 42.0 40.9 40.7 41.1 This analysis of revenues and

Denmark 56.4 57.8 56.6 55.4 55.4 52.8 53.4 expenditures helps to identify

Estonia 35.7 35.5 37.1 38.2 37.9 38.2 38.4 the automatic stabilisers of the

Finland 52.3 52.9 52.6 52.5 52.5 52.0 51.3

France 49.6 50.4 50.4 49.6 49.3 49.4 49.9 economic downturn. However,

Germany 43.3 43.5 43.8 44.0 43.8 43.5 42.3 other factors may also be rele-

Greece 38.0 38.1 39.1 40.1 39.9 40.8 40.0 vant. In some cases, such as the

Hungary 42.6 42.3 42.7 44.8 46.5 46.1 46.4

Ireland 35.1 35.4 37.0 35.9 33.8 33.7 33.9 UK, 2007 spending plans in-

Italy 44.2 43.8 45.4 46.4 46.0 46.5 46.5 tended spending to rise sharply,

Latvia 34.7 35.2 37.7 35.5 35.5 34.1 34.7 financed by higher revenues.

Lithuania 31.8 32.8 33.1 33.9 34.0 34.8 36.0

Luxembourg 41.4 41.6 39.9 40.8 43.3 44.0 42.9 Moving into the recession,

Malta 40.8 41.8 41.2 40.4 40.6 41.1 41.2 spending plans were not re-

Netherlands 44.3 44.5 46.2 45.6 46.4 46.1 45.6 duced, but revenues were much

Poland 36.9 39.1 39.9 40.2 39.2 40.2 40.3

Portugal 43.1 41.6 42.3 43.1 43.2 42.6 42.4 lower than expected, leading to

Romania 32.3 32.3 33.1 34.0 33.1 32.2 32.5 the very high deficit.

Slovakia 35.3 35.4 33.5 32.5 32.7 32.2 32.1

Slovenia 43.6 43.8 43.3 42.9 42.7 41.7 41.6 Where do these deficits leave the

Spain 38.5 39.4 40.5 41.0 36.6 36.4 36.9 level of outstanding debt as a

Sweden 56.1 57.2 56.5 56.3 55.7 53.0 52.7

UK 39.6 40.8 41.6 41.4 42.3 41.4 41.6 proportion of GDP for EU coun-

EU27 44.0 44.4 44.9 44.9 44.5 44.3 44.1 tries? For the EU as a whole, the

Source: 2004–2008, Eurostat; Forecasts 2009–2010 European Commission (2009a). measured debt-to-GDP ratio has

increased from 58.7 percent in

2007 to 72.6 percent in 2009, and

Tabl e 3. 4

Gover nm entt Ex pe ndit uress off EU

U m em berr st ates,, 20 04 –2 010

0 it is projected to rise again to

perc e ntt GDP

P 79.4 percent in 2010. This figure

200 4 200 5 200 6 200 7 200 8 200 9 201 0 is likely to continue to rise even

Austria 54.0 49.9 49.4 48.7 48.7 51.6 52.1 after 2010.

Belgium 49.5 52.2 48.5 48.3 49.9 48.5 48.2

Bulgaria 39.7 39.3 36.5 41.5 37.4 39.5 39.3 Of course, there is again consid-

Cyprus 42.8 43.6 43.4 42.9 44.0 44.0 45.0

Czech Rep. 45.1 45.0 43.8 42.6 42.4 45.9 47.6 erable variation across coun-

Denmark 54.6 52.8 51.6 51.0 51.7 55.0 57.0 tries: from Estonia with debt of

Estonia 34.1 34.0 34.2 35.5 40.9 45.0 47.3 under 7 percent of GDP to Italy

Finland 50.1 50.3 48.7 47.3 48.4 52.8 54.3

France 53.2 53.4 52.7 52.3 52.7 55.6 56.4 with a ratio of 113 percent.

Germany 47.1 46.8 45.3 44.2 43.9 48.2 49.0 There is some evidence that

Greece 45.4 43.3 42.2 44.0 44.9 45.3 45.2 countries with a lower debt ratio

Hungary 48.9 50.1 51.9 49.7 49.8 50.8 52.0 before the crisis have responded

Ireland 33.7 33.7 34.0 35.7 41.0 45.8 49.1

Italy 47.7 48.2 48.7 47.9 48.7 51.2 51.1 with a greater overall fiscal stim-

Latvia 35.8 35.6 38.2 35.9 39.5 46.8 49.8 ulus. For example, Ireland’s

Lithuania 33.3 33.3 33.6 34.9 37.2 39.5 42.7 ratio shot up from 25 percent in

Luxembourg 42.5 41.6 38.6 37.2 40.7 44.2 45.7

Malta 45.5 44.7 43.7 42.6 45.3 44.4 44.8 2007 to 61 percent in 2009, and

Netherlands 46.1 44.8 45.6 45.3 45.5 48.3 50.2 the UK from 44 percent to

Poland 42.6 43.4 43.8 42.1 43.1 46.1 46.8 68 percent. But there is little evi-

Portugal 46.5 47.6 46.3 45.8 45.9 48.9 48.7

Romania 33.5 33.5 35.3 36.6 38.5 38.5 38.9 dence that this was a discre-

Slovakia 37.6 38.2 36.9 34.4 34.9 38.3 39.4 tionary response, whereby coun-

Slovenia 45.8 45.3 44.6 42.4 43.6 47.7 48.6 tries that were more able to pro-

Spain 38.9 38.4 38.5 38.8 40.5 45.2 47.1

Sweden 55.6 55.2 54.1 52.5 53.1 56.6 57.3 vide a fiscal stimulus did so.

UK 42.9 44.1 44.2 44.0 47.7 50.5 52.4 Instead, the underlying reasons

EU27 46.9 46.9 46.3 45.7 46.8 50.1 51.1 appear more to do with the

Source: 2004–2008, Eurostat; Forecasts 2009–2010 European Commission (2009a)

74

EEAG Report 2010 Chapter 3

3.1 Macroeconomic models

Tabl e 3. 5

Debt-to-GDP

P r atioss off EU

U m e mb err stat es,, 2 004 –2 01 0 3.1.1 Theory

perc e ntt

200 4 200 5 200 6 200 7 200 8 200 9 201 0 Theoretical macroeconomic mod-

Austria 64.8 63.7 62.0 59.4 62.5 70.4 75.2 els have explored a variety of

Belgium 94.4 92.2 87.9 84.0 89.6 95.7 100.9

Bulgaria 37.9 29.2 22.7 18.2 14.1 16.0 17.3 channels through which a fiscal

Cyprus 70.2 69.1 64.6 59.4 49.1 47.5 47.9 stimulus can affect the economy.

Czech Rep. 30.4 29.8 29.6 28.9 29.8 33.7 37.9 There are of course fundamental

Denmark 43.8 37.1 31.3 26.8 33.3 32.5 33.7

Estonia 5.0 4.5 4.3 3.5 4.8 6.8 7.8 differences across paradigms as

Finland 44.2 41.4 39.2 35.1 33.4 39.7 45.7 regards the effectiveness of fiscal

France 64.9 66.4 63.7 63.8 68.0 79.7 86.0 policy. Neoclassical models em-

Germany 65.6 67.8 67.6 65.1 65.9 73.4 78.7 phasize that fiscal measures are

Greece 98.6 98.8 95.9 94.8 97.6 103.4 108.0

Hungary 59.4 61.7 65.6 65.8 73.0 80.8 82.3 either irrelevant (Ricardian equiv-

Ireland 29.7 27.5 24.9 25.0 43.2 61.2 79.7 alence prevents tax cuts from

Italy 103.8 105.8 106.5 103.5 105.8 113.0 116.1 boosting private demand) or

Latvia 14.9 12.4 10.7 9.0 19.5 34.1 50.1

Lithuania 19.4 18.4 18.0 17.0 15.6 22.6 31.9 counterproductive (public spend-

Luxembourg 6.3 6.1 6.7 6.9 14.7 16.0 16.4 ing crowds out private spending).

Malta 72.2 69.8 63.7 62.1 64.1 67.0 68.9 Keynesian models emphasise that

Netherlands 52.4 51.8 47.4 45.6 58.2 57.0 63.1

Poland 45.7 47.1 47.7 44.9 47.1 53.6 59.7 fiscal policy can actually crowd-in

Portugal 58.3 63.6 64.7 63.5 66.4 75.4 81.5 private expenditure, especially

Romania 18.7 15.8 12.4 12.7 13.6 18.2 22.7 when economic resources are

Slovakia 41.4 34.2 30.4 29.4 27.6 32.2 36.3

Slovenia 27.8 27.0 26.7 23.4 22.8 29.3 34.9 underutilised in a recession. An

Spain 46.2 43.0 39.6 36.2 39.5 50.8 62.3 important lesson from these con-

Sweden 51.2 51.0 45.9 40.5 38.0 44.0 47.2 trasting theoretical analyses, how-

UK 40.6 42.3 43.4 44.2 52.0 68.4 81.7 ever, is that the macroeconomic

EU27 62.2 62.7 61.3 58.7 61.5 72.6 79.4

Source: 2004–2008, Eurostat; Forecasts 2009–2010 European Commission (2009a). response to fiscal expansion can

vary widely, depending on the

degree of slack in the economy,

the monetary policy response, as

planned spending prior to the crisis and the degree well as the relevance of market distortions, ranging

to which the economies were affected by the finan- from credit constraints and other financial imperfec-

cial crisis. tions to nominal rigidities.

We discuss the implications of these deficits and their It is useful to start our analysis with a brief reconsid-

effects on the debt ratio in Section 4 below. eration of the standard neoclassical model with flexi-

ble prices and well-functioning labour and goods

markets, see e.g. Baxter and King (1993). A specific

3. Is fiscal stimulus effective? Evidence from the reason to do so is that this model clarifies the impor-

literature tant difference between wealth and substitution

effects from fiscal measures, which are often blurred

Most economists and policymakers have agreed that together in the popular account of the way fiscal pol-

the adverse economic effects of the current crisis icy works. In the classical exercise proposed by the lit-

could not have been contained without a strong fis- erature, a temporary increase in government spending

cal stimulus. Nonetheless, there are also sceptics who is eventually matched by an increase in lump-sum tax-

denounced the large fiscal expansions from 2008 as a ation which has the same present value (the timing of

waste of resources that could actually jeopardise the taxes does not matter, as Ricardian equivalence holds

recovery because of their lasting negative impact on in this case). The increase in spending raises output

government finances. Not surprisingly, the long- somewhat, but unambiguously lowers consumption.

standing debate on the fiscal transmission mecha- The fall in consumption occurs for two reasons. First,

nism, and especially on the size of the fiscal multi- as agents anticipate rationally the time path of future

plier, is raging once again. In this section, we briefly spending, they also feel that their net-of-tax wealth

reconsider the theoretical and empirical arguments has fallen by the full increase in the tax burden. Under

in this debate. 75 EEAG Report 2010

Chapter 3 Tabl e 3. 6 Publi c in terv ent ionss i n th e ba nkin g se ctorr

perc e ntt GDP

Guarantees on bank Relief on impaired asset and Guarantees

Capital injections liabilities liquidity and bank support on deposits

Total Effective Total Total ( 000, or

approved capital approved Guarantees approved Effective percent of

measures injections measures granted measures interventions GDP)

Austria 5.5 1.7 25.7 6.8 7.1 2 100 percent

Belgium 5.3 6.1 70.8 16.3 8.1 8.1 100

Bulgaria 0 0 0 0 0 0 50

Cyprus 0 0 0 0 0 0 100

Czech Republic 0 0 0 0 0 0 50

Denmark 6.1 2.4 253 2.5 0.3 0.3 100 percent

Estonia 0 0 0 0 0 0 50

Finland 0 0 27.7 0 0 0 50

France 1.2 1.2 16.6 5.5 0.2 0.2 70

Germany 4.4 2 18.6 7.2 1.4 1.4 100 percent

Greece 2 1.5 6.1 1.2 3.3 1.8 100

Hungary 1.1 0.1 5.9 0 0 2.6 100 percent

Ireland 6.6 6.5 164.7 164.7 0 0 100

Italy 1.3 0.1 NA 0 0 0 c. 103

Latvia 1.4 0.9 25.7 2.8 10.9 4.7 50

Lithuania 0 0 0 0 0 0 100

Luxembourg 6.9 7.9 12.4 NA 0.9 0.9 10

Malta 0 0 0 0 0 0 100

Netherlands 6.4 6.8 34.3 7.7 11.4 5.5 100

Poland 0 0 0 0 0 0 50

Portugal 2.4 0 10 3.3 0 0 100

Romania 0 0 0 0 0 0 50

Slovakia 0 0 0 0 0 0 100 percent

Slovenia 0 0.4 32.8 6.3 0 0 100 percent

Spain 0 0 18.6 2.1 2.8 1.8 100

Sweden 1.6 0.2 48.5 11 12.6 0 50

UK 3.5 2.6 21.7 11.3 16.4 14.7 50

EU27 2.7 1.7 20.5 7.8 2.1 1.4

Source: European Commission (2009b) spending. If the increase in public expenditure is per-

standard assumptions, agents react to the negative manent and immediately implemented, it is the wealth

wealth shocks by reducing consumption and leisure. shock associated with the higher tax burden that con-

The reduction in leisure in turn implies an increase in stitutes the lion’s share. Otherwise, most of the adjust-

labour supply, which increases output and lowers the ment in consumption and leisure is driven by

real wage. The second effect works through intertem- intertemporal substitution. To clarify this point, sup-

poral substitution of future for present consumption. pose that, over the long run, the growth rate is zero

If the increase in spending is temporary, interest rates and the real interest rate is 3 percent. All else equal, a

(long and short) rise on impact, reflecting the relative temporary increase in spending as high as, say, 10 per-

scarcity of current output due to the additional de- centage points of GDP for one year would generate

mand by the government. In response to real interest tax liabilities reducing households’ permanent income

rate movements, households postpone their spending 4

by a mere 0.3 percentage points of GDP

plans. Similarly, real wages may temporarily rise in the – quite a

short run, creating an incentive to work more on small amount, relative to the size of the upfront

3

impact. spending expansion.

The distinction between wealth and substitution

The relative weight of these two effects, wealth and effects is a key element in assessing the effectiveness

intertemporal substitution of consumption, depends of fiscal stabilisation policy. By its very nature, fiscal

crucially on the evolution over time of the change in

3 4

While consumption is typically crowded out by government spend- This result is obtained by calculating the constant flow of real

ing, investment can respond in different ways, depending on the spec- taxes, which is equal, in present discounted term, to the increase in

ification of the model and especially on the persistence of the shock net debt financing the spending expansion. In the example in the

to public spending. There are also a number of extensions of the text, the increase in debt is 10 percentage points of GDP.

t

neoclassical model which could also accommodate a positive effect Hence 10=Σ(1+r) x=x(1+r)/h implies that the additional tax pay-

of the rise in government spending on consumption. ment (x=.3/1.03) must be approximately equal to 0.3 per period.

76

EEAG Report 2010 Chapter 3

stimulus is temporary. The wealth shock associated higher if a short-run expansion in spending is even-

tually offset, at least in part, by a decline in spend-

with changes in the tax burden will affect house- ing below trend, rather than exclusively by a rise in

holds’ consumption decisions in a limited way. The taxes. This is because the reversal in government

focus should instead be placed on “intertemporal spending generates expectations of a decline in

substitution”. short-term interest rates in the future, which has an

immediate effect on long-term rates. With sticky

As stressed by the Keynesian literature, an important prices (and a relatively accommodative monetary

argument in favour of fiscal stabilisation is provided by stance), it is possible that this effect may dominate

models allowing for financial frictions. A fiscal stimu- the upward pressure on long-term rates resulting

lus is likely to be effective, for instance, when some from the additional government spending. It may

households are credit-constrained, so that their spend- well be possible that consumption would be crowd-

ing decisions become sensitive to disposable income, as ed in, rather than crowded out, on impact (see also

opposed to permanent income. If current income Chapter 1 in the 2009 EEAG Report).

increases due to either a government expansion which

raises economic activity and therefore wage payments, Note that monetary and fiscal interactions in both the

or a cut in taxes, these households are likely to spend short and the long run work mostly through the

more. Depending on the proportion of credit-con- intertemporal substitution channel already discussed

strained households in the economy, the positive early on in this chapter, in particular through their

response of their demand may drive up overall con- influence on the path of the long-term rates relevant

sumption. Similar results may be predicted by models for private demand decisions.

where firms (entrepreneurs) are credit constrained,

although the transmission mechanism is different, see A new generation of models building on Eggertsson

e.g. Villaverde (2010) for a recent discussion. and Woodford (2004) suggest that the fiscal expan-

sions may be extremely valuable in deep recessions

Models with nominal rigidities call attention to an in which monetary policy is constrained in setting

additional important element, that is, the interactions interest rates by the zero lower bound. In this case,

among monetary and fiscal policy. In both the tradi- absent fiscal policy, deflationary pressures from

tional and the new Keynesian models, the effect of a large recessionary shock may give rise to a defla-

fiscal stimulus is largely determined by the stance of tionary spiral: with the interest rate at zero, insuffi-

the central bank. The fiscal multiplier is indeed deter- cient demand causes firms to cut prices; to the

mined by the targeting rule (interest or exchange rate) extent that pricing decisions are staggered, falling

pursued by monetary authorities. For example, in the prices generate expectations of lasting deflation; for

classical Mundell-Fleming model, fiscal policy is a given nominal interest rate, these translate into

more effective if the country adopts a fixed exchange higher real rates; and higher real rates further weak-

rate regime, so that the domestic policy rate is en demand, reinforcing the fall in output. A fiscal

anchored to the foreign interest rate by the uncovered expansion can however stop this adverse mecha-

interest parity condition. Similarly, in the new nism, by raising demand and therefore contrasting

Keynesian model, if the central bank could (and the pressure towards lowering prices – a case dis-

would be willing to) target a constant interest rate in cussed by Christiano et al. (2009), Corsetti et al.

real terms, under standard assumption this monetary (2010), Erceg and Lindé (2010) and Eggertsson

stance would completely determine the evolution of (2009) among others. These contributions are of

consumption: any variation in government spending particular interest in the current situation, not only

would exclusively be reflected in changes in output because they explicitly address issues in fiscal stabil-

(Woodford 2010). The general point here is that some isation when interest rates are already effectively at

degree of monetary accommodation in the short run zero, but also because they provide theoretical

raises the macroeconomic impact of an increase in instances of very large multipliers for government

government spending. 5

spending (although not necessarily for tax cuts).

However, as consumers and firms are forward look- 5 Eggetsson (2009) emphasises that, when monetary policy is stuck at

ing, the impact of fiscal stimulus also depends on the zero lower bound, fiscal policies should aim directly at stimulat-

ing aggregate demand. These policies include temporary increases in

(private expectations about) how fiscal consolida- government spending and tax cuts, such as an investment tax credit

or a cut in sales taxes (by virtue of their direct effect on aggregate

tion will take place in the future. Corsetti et al. demand rather than aggregate supply). Tax cuts that lower marginal

(2009a), for instance, show that fiscal multipliers are costs may instead exacerbate the risk of a deflationary spiral.

77 EEAG Report 2010

Chapter 3 Yet, they mostly rely on restrictive assumptions The different approaches briefly summarized in the

Box have been used on aggregate data in a number of

regarding the origin of the shocks underlying the countries to identify the sign and size of the multipli-

global slowdown. er effect, and the effect on other variables, such as

consumption, employment, interest rates and ex-

As discussed in last year’s EEAG Report, a leading change rates.

explanation of the unexpected and strong drop in

demand during the last months of 2008 and through- Tables 3.7 and 3.9 summarise some of the estimates in

out 2009 attributes the recession to a rise in perceived 6 The estimates

the literature of the fiscal multiplier.

uncertainty (see Chapter 2 of the 2009 EEAG Re- shown in these tables are – where it is possible to iden-

port). Such an interpretation raises the issue of how tify – the cumulative peak effect on GDP of an exoge-

fiscal policy could stabilise inefficient and large fluc- nous shock to government spending (Table 3.7) or

tuations in economic activity in the face of rising taxation (Table 3.9). The cumulative peak effect can

uncertainty. This issue defines an important chapter occur immediately, or several quarters after the initial

in the economics of fiscal stabilisation, largely yet to shock.

be written. We will not discuss all of these papers. However, it is

It is indeed plausible that during a financial turmoil, worth exploring some of the results in a little more

when expectations are down, the role of fiscal policy detail. For example, in one well-known paper,

is to inject “optimism” in private markets, helping Blanchard and Perotti (2002) present results from two

people to re-gain confidence. Concretely, the govern- models, which vary according to whether a determin-

ment could commit to insure people against some istic or stochastic trend is added to the model. The

very bad outcomes: to the extent that the crisis is dri- results shown in Tables 3.7 and 3.9 are for the latter

ven by self-validating expectations, such a commit- case. Looking at point estimates, output rises by 0.9

ment can in principle coordinate expectations away following a unit increase government spending. This

from those outcomes. An advantage of this approach effect occurs in the first quarter, and thereafter

to fiscal policy is that the premise of the stabilisation declines. By contrast, with a deterministic trend, the

strategy would be fully consistent with the leading peak effect reaches 1.29, but occurs only after

diagnosis of what causes the crisis. 15 quarters. Since government spending is itself a

component of GDP, these two estimates generate dif-

The design of fiscal stabilisation coherent with this ferent predictions for the sign of effect on the other

view is however quite complex. Some of the trade-offs elements of GDP: negative under the stochastic trend,

are already debated heavily, by and large contrasting and positive for the deterministic trend, although

the interests of Wall Street with the interest of Main both are close to zero.

Street. Moreover, there is a budget constraint on the

stimulus: in light of the uncertainty surrounding their Blanchard and Perotti develop their model further, in

effects, actions must be such that they do not put fis- an attempt to identify the effects on the different com-

cal sustainability in peril. We will return to these ponents of GDP. Their results are shown in Table 3.8.

issues below. For both models, they find a positive effect on private

consumption, although the size of this effect is quite

3.1.2 Time series and panel analysis different between the two models. This is inconsistent

with the basic neoclassical model, which would pre-

Empirical work has generated a wide variety of esti- dict a reduction in consumption. They also find nega-

mates of fiscal multipliers – that is, of the effect on tive impacts on investment, exports and imports. Note

output of a fiscal stimulus. Certainly, we would expect that the peak effect on GDP is not equal to the sum of

the multiplier to depend on the type of fiscal stimulus. the peak individual components of GDP. This is

But the range of estimates generated in the literature because the peak effects occur at different times. For

probably owes more to the difficulty of identifying the example, the peak effect on GDP overall with the

effects, the variety of techniques used and the possi- deterministic trend is in quarter 15. But in that quar-

bility that the multiplier may vary over time and ter, the cumulative effect on each of the components

across countries. The Box gives an indication of the is not at its maximum.

different approaches used to identify the effects of an

exogenous discretionary shock to government spend-

ing or taxation. 6 These tables draw on the survey by Hebous (2009).

78

EEAG Report 2010 Chapter 3

Box

x 3. 2 Estimatin g thee fiscall m ultiplierr

A variety of techniques have been used to estimate the effects of discretionary fiscal policy on the economy, including the

effects on output, consumption, employment and other factors.

The starting point for most models is a Vector Autoregression (VAR) model. This is of the form

X = A(L)X + U

t t1 t

X

where represents a vector of variables (typically, output, government spending and taxation, although more recent

t A(L)

approaches include variables such as the stock of debt, exchange rates and interest rates), is a distributed lag

U

function, and is a vector of error terms. The key issue in using a VAR model is to identify the effects of an exogenous

t

change to either government spending or to tax revenue. Four approaches have been used.

A recursive approach to estimating the effect of a shock to one variable (Sims, 1980) is to assume an ordering of the

variables such that one variable – typically government spending – does not react contemporaneously to other variables.

The second – typically tax revenue – responds contemporaneously only to the first, government spending. The third –

typically output – responds contemporaneously to both, and so on for more variables.

A structural approach (SVAR in Tables 3.7 and 3.9) is used by Blanchard and Perotti (2002), based on a particular

structure of the residuals from the estimated relationship. External is used to identify the contemporaneous effects of

output on taxation and government spending. These generate instruments which can be used to estimate the

contemporaneous effect of taxation and government spending on output. Essentially, the idea is that there are decision and

implementation lags which prevent new government spending decisions in a quarter (year) from responding to

contemporaneous economic circumstances. Hence, innovations to spending not systematically explained by the evolution

of the business cycle (contemporaneous and lagged output gaps), the own dynamics of spending (lagged spending), and

the state of the public finance (debt) can be treated as unexpected (structural) shocks to fiscal policy, whose effect on the

economy gives information about fiscal transmission. A potential problem in this approach is that the variation in public

spending defining these fiscal shocks may actually be the subject of a political debate prior to implementation. Hence they

are to some extent anticipated by the private sector, and thus they cannot necessarily be treated as unexpected. As a partial

solution to this problem Beetsma et al. (2006) use annual instead of quarterly observations.

A “narrative” approach aims to exploit some clearly exogenous shocks to one of the variables in the system. For example,

Ramey and Shapiro (1998) identify changes to government expenditure in several episodes of military build-up in the

USA. These episodes are exploited by introducing dummy variables into the VAR. The response of the system to these

dummy variables provides a direct estimate of the multiplier.

A clear advantage of a narrative approach is that anticipation effects can be accommodated in the estimation by tracing the

timing in which the political discussion about policies with clear fiscal implications (such as going to a war) begins. An

important open issue however is that the approach is more effective, the larger the variation in spending or tax changes to

be proxied by the dummies. Recent papers have encompassed both of these latter approaches: see Perotti (2007), Ramey

(2008). Perotti (2007) provides a useful comparison of these techniques.

A fourth approach is a sign restriction approach, proposed by Uhlig (2005), and used by Mountford and Uhlig (2009) and

Pappa (2009a). This involves imposing sign restrictions on the impulse responses of some variables.

A useful critical discussion of identification has recently been provided by Barro and Redlick (2009), who emphasize the

problem of reverse causation (output growth explaining more spending) in the macro literature. In their approach, the best

identification strategy consists of focusing on episodes of large variations in defence spending (for the US: World War II

and the Korean War).

How large is the government spending multiplier? tion crucially hinges on financial development, trade

Returning to Table 3.7, there is clearly considerable openness, the state of public finances, the exchange

variation in estimates of the impact of a shock to gov- rate regime and, last but not least, the health of the

ernment spending on overall GDP. Most, though not financial sector, see e.g. Perotti (1999), Giavazzi and

all, of the estimates are positive (with a few in excess Pagano (1990), Giavazzi, Jappelli, and Pagano (2000),

of 1). However, it is worth stressing that confidence Ilzetzki, Mendoza, and Vegh (2009), and Corsetti,

intervals are quite large: in most cases the point esti- Meier and Mueller (2009b). Linear estimations aver-

mates are not significantly different from zero. aging out multipliers across economic conditions

(which can vary over time) may hide large and signif-

From a theoretical perspective, this overall conclusion icant differences. For this reason, the fiscal transmis-

should not come as a surprise. Theory has long sion mechanism should be systematically analysed

emphasized that the effectiveness of fiscal stabiliza- conditional on different economic environments.

79 EEAG Report 2010

Chapter 3 Tabl e 3. 7 Estimat ess off t hee effe ctss off a g ov ern me ntt sp en din g in cre asee sh o ck

y Dataa Perio d Tec hni qu e Multipl ierr forr

Stud outp utt

Ramey and Shapiro (1998) USA 1947–96 Narrative approx 1

Fatas and Mihov (2001) USA 1960–96 Recursive 0.3

Blanchard and Perotti (2002) USA 1960–97 SVAR 0.9

Australia 1960–79 SVAR – 0.1

Australia 1980–01 SVAR 0.21

Canada 1960–79 SVAR 0.59

Canada 1980–01 SVAR – 0.28

Germany 1960–74 SVAR 0.41

Perotti (2005) Germany 1975–89 SVAR 0.4

UK 1960–79 SVAR 0.48

UK 1980–01 SVAR – 0.20

USA 1960–79 SVAR 1.13

USA 1980–01 SVAR 0.31

Heppe-Falk et al. (2006) Germany 1974–04 SVAR 0.62

Ravn (2007) Australia, Canada, SVAR 0.52

UK, USA

Giordano et al. (2007) Italy 1960–79 SVAR 0.06

Favero and Giavazzi (2007) USA 1980–06 SVAR 0.13

USA 1980–06 SVAR 0.02

Gali et al. (2007) USA 1954–03 SVAR 0.78

USA 1955–06 Recursive 1

Caldara and Kamps (2008) USA 1955–06 SVAR 1

USA 1955–06 Sign restriction approx 0.5

USA 1955–06 Narrative 0

Beetsma et al. (2006) EU14 1970–04 Recursive 1.2

De Castro and De Cos (2008) Spain SVAR 1.31

Ramey (2008) USA 1947–03 Narrative approx 1

Canada 1970–07 Sign restriction 0.18

EU 1991–07 Sign restriction 0.16

Pappa (2009b) Japan 1970–07 Sign restriction 0.13

UK 1970–07 Sign restriction 0.13

USA 1970–07 Sign restriction 0.74

Bilbiie et al. (2008) USA 1957–79 Recursive 1.71

USA 1983–04 Recursive 0.94

Mountford and Uhlig (2009) USA 1955–00 Sign restriction 0.44

fiscal transmission: output and consumption multipli-

Specifically, focusing on OECD countries, Corsetti et ers are small and positive; trade balance turns into a

al. (2009b) contrast average linear estimates of the deficit; the exchange rate experiences a short-lived real

government spending multiplier, with estimates appreciation, followed by a weakening. But these

explicitly allowing for “non-linearities”. In their average linear responses are not necessarily confirmed

analysis, the estimated linear effect of a government when the estimation is conditional on specific eco-

spending shock is in line with the VAR literature on nomic features/environments. In

accord with standard theory, the

study finds that spending policies

Tabl e 3. 8 are more effective in relatively

Effectss off go ver nm entt spe ndi ng

g sh oc k o n co mp on entss off GD P,,

from

m Bl an ch ard

d an d P erottii ( 200 2)) closed economies (“openness

Deterministic trend Stochastic trend matters”) and under a peg (“the

estimates estimates exchange rate regime matters”);

Peak effect Quarter Peak effect Quarter and less effective or even counter-

Government spending 1.14 4 1.00 1

Consumption 1.26 14 0.46 2 productive in economies with

Investment – 1.00 5 – 0.98 9 high public debt (“the state of

Exports – 0.80 9 – 0.37 13 public finances matters”).

Imports – 0.49 9 – 0.08 9

GDP

P 1.39

9 15

5 0.95

5 1 Most strikingly, multipliers are

significantly larger during years

80

EEAG Report 2010 Chapter 3

The extent to which multipliers change over time is

with financial and banking crises, identified using the also examined in other papers. For example, Perotti

information in Reinhart and Rogoff (2008). Indeed, (2005) finds that multipliers are not constant over

during such episodes the point estimate for the output time or across countries (see Tables 3.7 and 3.9). In

multiplier is a multiple of the other estimates. While, particular, this paper presents some evidence that the

given the reduced number of observations featuring a size of the multiplier has declined over time. A num-

crisis (Spain, Japan, Finland and Norway), confi- ber of possibilities for differences over time are dis-

dence intervals are large, these results appear to cor- cussed by Perotti, including countries becoming more

roborate the argument that fiscal support to econom- open and introducing flexible exchange rate regimes.

ic activity has been key to stabilising output during These explanations are consistent with the results in

the current crisis. Corsetti et al. (2009b). However, it should be kept in

mind that the export/GDP ratio is small in many

Table 3.9 summarises estimates of the multiplier countries, and also that the evidence of crowding out

arising from a change in taxation. Beginning with of net exports by fiscal shock is controversial.

Blanchard and Perotti (2002) again, they find that

GDP falls by 0.7 in response to unit increase in taxes. Another possibility is the gradual relaxation of credit

However, there is a large variation in estimates from constraints over time. Since credit-constrained indi-

other papers. The sign of the effect on output is not viduals are more likely to change their consumption

agreed, and neither is the size. One well-known in response to a change in the real income, relaxing

recent approach is that of Romer and Romer (2010), these constraints is likely to reduce any positive effect

who identify various exogenous shocks to taxes in on consumption of a positive fiscal shock.

the US, and trace out their effects in a single equa-

tion model of output. They find a very high effect of However, while credit constraints may have been pro-

taxation. gressively relaxed with the process of deregulation

and market liberalisation, their incidence can still be

These results are challenged by Favero and expected to fluctuate along business cycle move-

Giavazzi (2009). They point out various restric- ments. This observation raises a fundamental prob-

tions in the Romer and Romer approach – for lem of using the estimates in Tables 3.7 and 3.9 to

example, that only tax shocks are incorporated into identify the effects of a fiscal stimulus during the

the model. Relaxing these restrictions, they find recession, as the extent of credit constraints is itself

much lower estimates: before 1980 the estimate affected by the recession. As the financial crisis in

never exceeds 1, and after 1980 it is not significant- 2008–09 generally reduced the supply of credit, more

ly different from zero.

Tabl e 3. 9 Estimat ess off t hee effe ctss off a t ax

x i ncr eas e sho ck

y Dataa Perio d Tec hni qu e Multipl ierr forr

Stud outp utt

Blanchard and Perotti (2002) USA 1960–97 SVAR – 0.69

Australia 1960–79 SVAR 0.46

Australia 1980–01 SVAR 0.36

Canada 1960–79 SVAR 0.03

Canada 1980–01 SVAR – 0.30

Germany 1960–74 SVAR 0.22

Perotti (2005) Germany 1975–89 SVAR – 0.02

UK 1960–79 SVAR – 0.10

UK 1980–01 SVAR 0.23

USA 1960–79 SVAR – 0.69

USA 1980–01 SVAR 0.43

Favero and Giavazzi (2007) USA 1980–06 SVAR 0.02

USA 1955–06 Recursive 0

Caldara and Kamps (2008) USA 1955–06 SVAR 0

USA 1955–06 Sign restriction – 0.8

Romer and Romer (2009) USA Narrative – 3.0

Mertens and Ravn (2009) USA 1947–06 Narrative – 2.17

Mountford and Uhlig (2009) USA 1955–06 Sign restriction – 0.2

81 EEAG Report 2010

Chapter 3 individuals are likely to have moved into a position of mentioned above in regards to spending policy: the

being denied credit. That in turn would make a tax estimated parameters may vary according to econom-

cut, for example, more effective in expanding con- ic conditions, and so be an unreliable guide to the

7

sumption and hence GDP. Once again, the estimates multiplier in any other period or country.

by Corsetti et al. (2009b) regarding the effect of

spending expansions in crisis periods appear to sup- 3.2 Microeconomic factors

port this notion. The discussion of the macroeconomic evidence

implies that the strength of the fiscal multiplier, and

This is an example of a more general problem with its effects on economic welfare, depends on the partic-

empirical analysis based on VAR models, already ular measures used and the underlying state of the

7 Auerbach and Feenberg (2000) discuss the automatic stabilising economy. Table 3.10 summarises the fiscal stimulus

properties of the US income tax with reference to the proportion of

consumers who are credit-constrained. measures adopted by EU member states in 2009

Tabl e 3. 10

0 Fiscall sti mul uss m e asuress i n th e EU,, 200 9

Nett stim uluss Nett c ontr acti on

n

Country Percent of GDP Country Percent of GDP

Rev en uee 0.3

Estonia

0.3

Personal income taxes, including Belgium 0.8

Ireland

0.2

social contributions, capital gains Bulgaria 1.0

Latvia

0.5

tax and dividends taxes Czech Republic 0.9

Luxembourg

0.3

Denmark 0.8

Romania

0.6

Germany 0.2

Slovakia

0.5

Spain 0.5

Cyprus 0.6

Lithuania 0.2

Malta 0.3

Netherlands 0.6

Poland 0.2

Portugal 0.6

Slovenia 0.9

Finland 0.7

Sweden 0.3

UK 0.4

Ireland

0.2

Corporate income tax and other Czech Republic 0.2

Greece

0.1

business taxes Denmark 0.4

Italy

0.1

Germany 0.4

Lithuania

0.2

France 0.3

Luxembourg

0.2

Cyprus 0.1

Hungary

0.2

Netherlands 0.2

Sweden

0.2

Poland 0.1

Portugal 0.4

Slovenia 0.1

UK 0.3

VAT and other indirect taxes Belgium 0.1 Bulgaria 0.4

Spain 0.6 Ireland 0.2

Portugal 0.2 Greece 0.2

UK 0.5 Cyprus 0.7

Latvia 0.7

Lithuania 0.3

Malta 0.1

Netherlands 0.2

Poland 0.1

Romania 0.9

Slovenia 0.2

Slovakia 0.1

Finland 0.8

Bulgaria

0.4

Other taxes Spain 1.1

Estonia

0.1

France 0.7

Ireland

0.1

Cyprus 0.5

Greece

0.1

Portugal 0.1

Italy 0.3

Luxembourg 0.2

Poland 0.1

Romania

82

EEAG Report 2010 Chapter 3

continued: Table 3.10

Exp en dituree Ireland 1.2

0.1

Public investment, support for Bulgaria France 0.2

0.4

business, infrastructure and research Czech Republic Netherlands 0.1

0.3

Denmark 0.4

Germany 1.0

Spain 1.3

Cyprus 0.7

Luxembourg 0.1

Hungary 0.9

Malta 0.9

Poland 0.6

Portugal 1.0

Romania 0.3

Slovenia 0.1

Slovakia 0.3

Finland 0.3

Sweden 0.4

UK Ireland 0.7

0.1

Social expenditure Belgium Hungary 0.6

0.3

Bulgaria Poland 0.2

0.2

Czech Republic 0.2

Greece 0.1

France 0.2

Italy 2.1

Latvia 1.4

Lithuania 0.6

Luxembourg 0.2

Portugal 0.1

Romania 0.5

Slovakia 0.2

UK Lithuania 0.7

0.1

Housing, labour market, education Belgium Netherlands 0.4

0.1

expenditure Germany 0.8

Estonia 0.3

Greece 0.1

France 0.1

Malta 0.8

Slovenia 0.1

Finland 0.1

Sweden 0.6

Czech Republic

0.1

Other spending Belgium 0.1

Ireland

0.4

Germany 0.3

Italy

0.2

Greece 1.4

Lithuania

0.2

Cyprus 0.5

Hungary

0.1

Slovenia 1.2

Malta

0.1

Finland 0.7

Poland

Source: European Commission (2009a)

(more details are shown in the Appendix), which is increasing its lifetime wealth. So, such a measure

would tend to generate some additional spending,

taken from European Commission (2009a). but also additional saving. In the context of the

increased uncertainty generated in a recession, it

It is easy to see how different measures with the same would be plausible to believe that much of the

fiscal cost could have different impacts on aggregate 8

increased wealth would initially be saved.

demand. For example, a credit-constrained house-

hold would generally like to increase its consump- To see how various measures may have different

tion, but is unable to do so because of the lack of effects, compare three forms of fiscal stimulus, say: (a)

opportunity for borrowing. A tax cut aimed at such an increase in social security benefits for the less well-

households would be immediately translated into an off, who are more likely to be credit-constrained; (b) a

increase in consumption, boosting aggregate de-

mand. The effect of the same tax cut on a household 8 Some evidence on households’ response to tax rebates after the

start of the crisis is provided by Parker, Souleles, Johnson and

that could already borrow as much as it wanted to McClelland (2009), who study the impact of the 90 billion dollar tax

rebates in 2008 on consumer spending in the US. This study finds

would be much smaller. Broadly, we would expect the response to be largest for lower-income households and home-

such a household to hardly regard the tax cut as owners.

83 EEAG Report 2010

Chapter 3 reduction in the general VAT rate; and (c) a reduction The apparent preference for measures targeted at per-

in the income tax rate. Neither (b) nor (c) need to be sonal consumption may not be surprising if it was

targeted towards groups that are more likely to spend believed that firms would be reluctant to undertake

the additional income. It is therefore likely that there significant investment in the midst of a recession:

would be a greater effect on aggregate demand per investment is generally more volatile than consump-

euro of a fiscal stimulus from (a). tion, and is perhaps less likely to respond to any form

of fiscal stimulus. Nevertheless, investment clearly has

As Table 3.10 indicates (and the Appendix sets out longer term benefits in creating conditions for greater

in more detail), many countries did enact increases output in the future.

in social expenditure, which are more likely to lead

to increases in aggregate demand. Unlike most of Public spending measures were more heavily targeted

the other categories in the Table, only three coun- towards such activity. Many countries increased pub-

tries actually reduced social expenditure. By con- lic investment or provided additional support for

trast, few countries sought to stimulate their infrastructure spending and research.

economies by reducing the VAT rate – and many

sought to offset the costs of a stimulus elsewhere by Overall, though, the Table suggests that there has

increasing the rate. Where the VAT rate was cut, it been no firm consensus amongst EU governments on

was sometimes a temporary measure: the UK, for the appropriate fiscal response to the financial and

example, introduced a lower rate for one year only. economic crisis. In each of the categories in the Table,

The fact that the VAT rate was due to rise again some countries created a positive discretionary stimu-

may have provided a greater stimulus to higher lus while others created a negative one. This to some

spending, as we discuss below. However, many extent may reflect differing conditions between coun-

countries also reduced income taxes and other taxes tries. But it also suggests that there has been consider-

on individuals. able uncertainty about the appropriate types of fiscal

policy required to best combat the crisis.

Similar considerations apply to business investment.

Here measures designed to create additional incen-

tives to invest, such as a more generous definition of 4. When and how should deficits be reduced?

taxable profit through increasing depreciation In some parts of the EU at least, the political debate

allowances, may have had some effect for firms that has moved swiftly on from the need for a fiscal stimu-

are not credit-constrained. Of course, even these lus to a recognition that fiscal deficits have grown sub-

firms may be unwilling to respond to such incentives, stantially and need to be reduced. This raises two

given the uncertainty surrounding the returns to related questions. First, how quickly should deficits

investment. be reduced? Second, how can they be reduced?

Such measures are unlikely to have any impact on the

investment of firms which cannot raise finance. By 4.1 The costs of maintaining high fiscal deficits

contrast, for such firms, measures which encourage A first point to note is related to the analysis above.

lending by banks would have a greater impact on The larger part of the fiscal deficits currently facing

investment. Alternatively, a simple cut in the tax rate EU governments does not result from discretionary

may achieve this, by allowing firms to retain a higher policy in response to the financial and economic cri-

proportion of pre-tax profit. However, this would sis. It is due to partly to an automatic response, and

only be true for firms that were profitable entering the partly to structural factors which would have created

recession. For firms making a taxable loss, a reduction larger deficits in any event. The automatic responses

in the tax rate may represent a cost, as the value of result from lower growth, which is translated into

any tax rebate would be lower. lower than expected tax receipts and higher than

expected costs of social transfers.

In general, in setting fiscal stimulus packages,

European governments appear to have targeted per- These automatic factors work in reverse as economies

sonal consumption more than business investment. move out of recession: economic growth will raise rev-

While 17 countries did make reforms to business tax- enues (and typically, coming out of a recession, rev-

ation, only 10 of these represented a cut, while 7 actu- enues rise more quickly than the underlying growth of

ally increased their tax take from business. 84

EEAG Report 2010 Chapter 3

Box

x 3. 3 Creditt Def aultt S w apss

A credit default swap (CDS) is a financial instrument in which the buyer makes a series of payments to the seller in

exchange for a payoff if a credit instrument defaults. 100

This can be seen as a form of insurance. For example, suppose A lends to B. Then A could purchase a CDS on that

100

debt which would pay in the event that B defaults on the repayment. The CDS spread is the annual amount that A

pays to the insurer over the length of the contract, expressed as a percentage of the insured amount.

An important difference from a normal insurance contract, though, is that the purchaser of this CDS need not be A. That

is, other investors can purchase the CDS even though they do not bear the underlying risk. That implies that the CDS can

be used as a speculative instrument rather than as a means of insurance. Such opportunities for speculative investment

raise regulatory issues about such contracts.

The underlying credit instrument can be a government-issued security. For such assets, the spread can be used as a

measure of the price of the risk associated with that security by the market. However, there are caveats to using the spread

for such a purpose. The main caveat is that the CDS only has value if the seller of the security is able to make the

insurance payment in the event the government in question defaults. If, for example, the US government were to default,

then it is highly likely that many financial companies would also default, and the CDS insurance payment would not be

made. The spread should therefore be seen as reflecting the joint probability that the government defaults but that the

seller of the CDS does not default.

Leaving that aside, in a well-functioning capital market, the spread on the CDS should be equal to the annual risk

premium which the government would have to pay on issuing debt. The yield on such debt will also reflect other factors,

1

such as expectations of future interest rates. So the difference in yields will not exactly match CDS spreads. This is why

the CDS spread itself is potentially a useful measure of the risk premium.

1 The differences in yields across countries are shown in Figure 1.26 in Chapter 1.

the economy). And as unemployed and others find would not be reached until 2019, and the debt/GDP

work, social transfers are reduced and replaced by ratio would peak at 106 percent. On the other hand,

higher tax revenues. if expenditure were kept constant in real terms, and

growth were 2.5 percent per year, a fiscal surplus

It is possible to do some basic calculations to esti- would be reached in 2016, with debt peaking at

mate how long it would take for economic growth to 95 percent of GDP.

lift the EU back to a position of budget balance.

Take as a starting point the projections made by the While there is of course, considerable uncertainty

European Commission for revenue and public about future growth rates, and about the ability of EU

expenditure as a proportion of GDP in 2010; these governments to hold down public expenditures rela-

are 44.1 percent and 51.1 percent respectively, and tive to the rate of economic growth, these projections

would result in an overall stock of public debt of suggest that it will be some years – and possibly a

79.4 percent of GDP. decade or more – before the EU can reach a fiscal sur-

plus and begin to cut the aggregate stock of debt. Of

Now suppose that the EU returns to a steady state course, this will happen more quickly in some coun-

2 percent growth from 2011 onwards. Assume also tries than in others.

that revenues rise slightly faster than economic

growth (so that the elasticity of revenues with respect What are the costs of maintaining such high levels of

to GDP is 1.1) and that public expenditure is held debt? The most obvious cost is that of servicing the

constant in real terms. Under this scenario, the EU as debt through interest payments. At the end of 2009,

a whole would return to fiscal surplus in 2017, reach- the yields on 10 year bonds issued by EU govern-

ing a peak debt/GDP ratio in 2016 of approximately ments lie mostly in a range between 3.2 percent and

100 percent. 3.8 percent, although some countries lie outside this

range (for example, Ireland and Greece). Yields on

Obviously, the return to a fiscal surplus would be shorter-dated bonds tend to be lower than this. But

faster if economic growth were higher, and slower if very roughly, the nominal cost of servicing debt at

public expenditure rose in real terms. For example, the 2010 level of around 80 percent of GDP is

if instead, public expenditure grew at just 0.5 per- approximately 3 percent of GDP. For example, the

cent per year in real terms, then a fiscal surplus European Commission’s current projections of the

85 EEAG Report 2010

Chapter 3 Figure 3.1 shows the development of CDS spreads on

interest liability in 2010 are: 3 percent of GDP in 10 year bonds issued by a number of governments

Germany, 3.1 percent in France, 3.1 percent in the since the beginning of 2008. Prior to the development

UK, 1.9 percent in Spain, and 4.8 percent in Italy. As of the financial crisis, these spreads were typically less

the stock of debt inevitably rises, these costs will than 0.1 percent. They increased dramatically over the

increase further. course of the crisis, reaching much higher levels – and

in the case of Ireland, around 3.5 percent. Since then,

The cost will rise as interest rates rise above their cur- they have declined again, though they remain well

rent low levels. And the interest rate for the debt of above their pre-crisis levels.

any country will depend on how risky that debt is per-

ceived by financial markets. The risk of the debt The spreads at the end of 2009 differ considerably

depends on a number of factors. Clearly, it depends across countries. Spreads in Germany have fallen

on the size of the outstanding debt as a proportion of back considerably to around 0.3 percent. The

GDP. But it also depends on the rate at which that spreads in France and the USA are slightly higher at

debt is increasing, and the state and prospects of the around 0.4 percent. However, the UK’s spread is

economy. only a little under 1 percent, Ireland is at 1.6 percent

and Greece is at 2.6 percent. The higher spread in the

In addition, some countries have provided guarantees UK may reflect the high level of the current deficit in

to private sector bank debt which could result in large the UK, even though its stock of debt is not out of

liabilities but which are not reflected in the measures line with that in France or Germany. It may also

of the current stock of public debt. Some evidence on reflect the potential liabilities that the UK has in

the extent of the contingent liabilities taken on by providing guarantees to the UK financial system.

governments through their financial sector interven- The high spreads for Ireland and Greece reflect their

tions is shown in Table 3.10, which shows the size of a much higher risk.

number of different measures undertaken by EU gov-

ernments to stabilise their banking sectors. The cost to governments in terms of higher interest

payments due to the risk reflected in these CDS

Three forms of intervention are shown: capital injec- spreads is, however, modest. Most government debt is

tions, guarantees on bank liabilities, specific relief on issued at a fixed rate of interest. The selling price of a

impaired assets and other direct bank support. The new government bond will reflect the risk which the

Table also shows the extent of guarantees on market attaches to that bond, and this implicitly

deposits. As would be expected, these forms of inter- defines the premium which the government must pay.

vention varied considerably across countries. To take Also, as risk rises, the market price of existing debt

one notable example, the UK has injected capital into falls, reflecting the higher rate of return required by

banks worth around 2.6 percent of GDP; it has pro- the market. But it is the owners of existing bonds that

vided guarantees for bank liabilities of over 11 per- bear this cost through the reduction in the value of

cent of GDP, and has injected a further nearly 15 per-

cent of GDP in supporting

banks through relief for im- Figure 3.1

paired assets and other mea-

sures. As such, the UK clearly

has contingent liabilities that are

not reflected in the figures for

public debt presented in Section

2 of this chapter. Other countries

also have huge contingent liabili-

ties, notably Ireland, and to a

lesser extent, Belgium and the

Netherlands.

One way of assessing the per-

ceived risk is to look at the

spreads of credit default swaps

on sovereign debt. 86

EEAG Report 2010 Chapter 3

their asset: the cash paid by the government on exist- ing a rise in income tax. If implemented at an early

stage, when the economy is still at the beginning of its

ing debt does not change. recovery path, the rise would be expected to generate

some reduction in spending, which would be a nega-

The cost to governments of the risk premia associat- tive shock to the economy. Now compare that option

ed with these CDS spreads therefore apply only to with an announcement that the rise in income tax will

new debt and not to the stock of debt. New debt take effect with sufficient delay – say in one or two

includes both new borrowing and the replacement of years’ time. In this case, the government would be

debt which matures, and so exceeds the fiscal deficit. delaying the reduction in the deficit, presumably in

(For example, a government with a new debt of say the hope of allowing the economy to recover further

10 percent of GDP and a risk premium of 0.5 percent before implementing the change. The problem with

would need to pay an additional 0.05 percent – i.e. one this is that individuals who face a future rise in their

twentieth of one percent of GDP – a year to service income tax perceive a reduction in their lifetime

this debt.) While these amounts may still be signifi- wealth immediately. They therefore consider them-

cant – especially for governments with high current selves to be worse off now, and consequently would

deficits – they are small relative to the overall costs of be likely to reduce their spending now. There is of

servicing the stock of debt. course the possibility that individuals would seek to

bring income forward from the following year in

Of course, the longer that these risk premia are main- order to benefit for the lower tax rate while it lasts.

tained, though, the more their cost will build up as This could provide a stimulus to the economy. But

more and more of the stock of debt has been issued at shifting income across years is generally much harder

relatively high risk premia. To reduce these risk pre- than shifting consumption. So it seems implausible

mia in the short, medium and long term, it is neces- that this effect could outweigh the effects on the

sary for governments to demonstrate that they have economy of the reduced spending.

credible plans to reduce the deficits in the medium

term, thereby reducing the possibility of eventual The main lesson here is that, from the vantage point

default. of the effect of changes in taxes on permanent

income, the announcement of a future income tax rise

One option here would be simply to point to the type may not have specific advantages over the announce-

of calculations set out above: that with economic ment of an immediate income tax rise. The timing of

growth and holding down expenditure rises, then the fiscal adjustment, however, can and does make a

deficits will eventually be closed. But other policies large difference through channels other than perma-

may also be required. We now discuss options for such nent income.

policies. A clear instance is provided by taxes on consumption,

4.2 Options for reducing deficits such as VAT and excise duties. Suppose the govern-

ment announced that the rate of VAT would rise in

The most obvious problem facing governments that one year’s time. This would also reduce the lifetime

wish to reduce their fiscal deficits is that doing so may wealth of individuals in the sense that, for a given

generate a negative fiscal stimulus, reducing or even income, they would be able to afford to buy less goods

overturning any economic recovery. The evidence for and services. The higher tax burden would tend to

this is summarised above in Section 3. Given the depress the economy, as would be the case with other

depth of the recession which faced the EU in 2008 tax rises.

and 2009, governments should be cautious in raising

taxes or cutting expenditure to reduce their deficits. However, the announcement could provide an impor-

The costs associated with a delay in such policies are tant fiscal stimulus, since there would be a clear incen-

relatively small compared with the possible costs of tive to bring forward spending to take advantage of

restricting economic growth. the lower VAT rate before it was increased. This would

provide an immediate stimulus to current private

But there is an important timing issue. As discussed demand, despite raising additional revenue in the

in Section 3, timing works mostly through intertem- medium term.

poral substitution effects. Consider, for example, the

possibility that a government may try to develop a In some ways such a policy mirrors the fiscal stimulus

measure announced by the UK government in

credible strategy for reducing its deficit by announc- 87 EEAG Report 2010

Chapter 3 December 2008: to reduce the VAT rate from 17.5 per- ness of current fiscal stimulus, as their deflationary

cent to 15 percent for a fixed period of about one year. impact materialises when the economy is still strug-

Arguably the most effective element of this stimulus gling with the aftermath of the recessionary shock. A

was its fixed time period. A permanent reduction in credible plan gradually phasing in spending cuts over

the VAT rate may not have had a large effect at the a two year horizon not only can reduce this risk: it can

point at which the country entered a recession. But also enhance the expansionary impact of the ongoing

the fact that the rate increased again a year later is fiscal stimulus.

likely to have had a more significant impact on con-

sumption in 2009. A final point to note here concerns the need for coor-

dination across countries. As noted in the introduc-

A second important instance (already discussed in tion, in 2008 there was general agreement that enact-

Section 3) is the possibility of designing consolida- ing a fiscal stimulus would be more effective if all (or

tion packages including cuts of government spending at least many) countries followed a similar policy,

below trend. Anticipation of lower public demand in increasing demand everywhere. By contrast, a single

the future tends to contain long-term interest rates country enacting a stimulus on its own would see

(as future short term rates will be lower). Lower real much of the stimulus flowing abroad through the pur-

rates in turn stimulate current demand. The effect of chase of imports.

anticipated cuts is expansionary because, to the ex-

tent that firms set prices subject to nominal rigidities, But in light of the need to consolidate debt, mea-

today’s prices will already optimally incorporate ex- sures to reduce public deficits across the world sum

pectations of the path of future demand and infla- up to a global recessionary impulse. In this case,

tion. With sticky prices, prospective spending cuts (all international policy coordination may still be benefi-

else equal) lower prices, containing the dynamic of cial insofar as it would be a way to internalise the

inflation, and thus allowing the central bank to be negative demand spillovers on foreign output created

more expansionary. by fiscal adjustment in a country. To wit: the same

way in which coordination leads to stronger global

For the above mechanism to work, however, the cen- stimulus at the start of a recession, coordination

tral bank must be able to control policy rates, i.e. the would lead to gradualism in fiscal consolidation

economy cannot be in a situation in which the central once the initial stimulus is withdrawn. If all coun-

bank would like to lower policy rates, but it cannot, tries simultaneously reduced their deficits by increas-

because these are already at zero. In these circum- ing taxes and reducing spending ignoring spillovers,

stances, as shown by Corsetti et al. (2010), the timing aggregate demand would fall everywhere too much,

of fiscal adjustment is crucial. and adjustment would create a much greater reces-

sionary impulse, possibly harming the nascent world

With a near-zero nominal interest rate, implementing recovery.

spending cuts too soon would add to the deflationary

pressure of the ongoing recession. These pressures However, coordination is not necessarily desirable.

may end up raising inefficiently the interest rates in The risk is that gradualism in the name of coordi-

real terms, and may possibly exacerbate the zero- nation could provide an excuse to delay the adop-

imple-

lower-bound problem. In contrast, a tion of the necessary measures to preserve stability.

delayed

mentation of spending cuts can be quite beneficial, as Appealing to the need for a coordinated fiscal con-

it would help the central bank maintain an expan- solidation, for instance, incumbent governments

the economy exits from

sionary monetary stance may leave unpopular decisions for future govern-

after

the zero-lower-bound constraint (and it may shorten ments to make.

the period of the zero lower bound episode). Conversely, in the current circumstances it makes

These considerations are important in light of the fact sense that the worst hit countries or the countries with

that the large rise in public debt requires fiscal consol- the most fragile public finances should adjust upfront

idation to be substantial. Households reasonably and most deeply so as to prevent the spreading of

expect adjustment not to take place exclusively via concerns about fiscal sustainability. The benefits from

increases in taxes but also via some cut in spending. coordination, which may be small initially, can quick-

With interest rates still close to zero, anticipation of ly turn largely negative if this ends up interfering with

early spending cuts may actually harm the effective- the most efficient path of debt consolidation.

88

EEAG Report 2010 Chapter 3

• There are costs of maintaining high levels of debt,

5. Conclusions though these should not be exaggerated. Especially

In this Chapter we have discussed a number of issues at low interest rates, the cost of servicing debt is of

surrounding the large rises in fiscal deficits in Europe. the order of 3 percent of GDP, though again there

The key points raised are as follows. is considerable variation across member states.

Two factors could increase this cost in the short to

• There have been large increases in budget deficits medium term. First, interest rates are likely to rise.

throughout the EU, leading to considerable rises in Second, public debt appears increasingly risky to

the stock of public debt as a percentage of GDP. In the market, which implies that higher risk premia

2009, the total deficit in the EU was around 6 per- could be charged.

cent of GDP, and it is expected to rise further in • Although these risk premia are currently not large,

2010. There has been a corresponding increase in they could be lowered – or at least prevented from

outstanding debt, rising to 72 percent of GDP in growing – if governments announced credible

2009 and to nearly 80 percent of GDP in 2010. strategies to reduce deficits over the medium term.

• There are also wide variations across countries. A downside of such a strategy is that announce-

The UK, Ireland and Latvia have particularly high ments of future tax rises may hamper the economy

deficits, though in all three cases their outstanding immediately as individuals perceive their lifetime

debt is moderate. Italy, Greece and Belgium have income to be lower.

much higher outstanding debt, though all three • One way of reconciling the need for a credible

have had high debt for some years. deficit-reduction strategy with the need to avoid

• These high deficits have generally not reflected dis- harming a fragile economy is to announce rises in

cretionary changes by EU governments. While taxes on spending – such as VAT – to take effect

most governments introduced a discretionary fiscal from some future period, say in one year’s time.

stimulus in 2008 and 2009, these were small relative This would induce individuals to bring spending

to the overall deficits. The form of these discre- forward, which would provide a temporary stimu-

tionary changes (and even their sign) has varied lus to the economy.

considerably between countries. • Another way consists of announcing well-designed

• There is considerable empirical evidence that a fis- measures bringing government spending on goods

cal stimulus has a positive effect on output, and services below trend, to be implemented suffi-

although there are many problems in measuring ciently far in the future as to avoid the risk of

the effect, so that the size of the fiscal multiplier is exposing the economy to additional deflationary

not known with any certainty. In any case, there is pressures when policy interest rates are still close to

little reason to suppose that effects estimated on zero. Provided that they are not implemented too

historic data are likely to be valid in the midst of a early, future spending cuts are beneficial to the

recession. This is particularly the case when inter- recovery, as they contain the rise in long-term

est rates are effectively at zero and the economy is interest rates (as well as attenuating concerns about

shaken by an ongoing financial and economic cri- debt sustainability).

sis, when there may be very large multipliers for • A final point concerns co-ordination. In attempt-

government spending. There is also little reason to ing to stimulate the economy there were gains from

suppose that different forms of fiscal intervention co-ordination. For an individual country, a stimu-

have similar effects. lus to spending might be largely reflected in

• The scope for reducing deficits depends crucially increased imports, creating demand for goods and

on the rate of economic growth achieved over the services produced elsewhere. A coordinated policy

next few years, and the degree to which real public reduces this risk. In principle, the same argument

spending can be curtailed. For example, a simple also applies (with a different sign) to fiscal adjust-

calculation suggests that if spending is kept con- ment. If all countries implemented a contrac-

stant in real terms throughout the EU, then eco- tionary fiscal adjustment simultaneously and inde-

nomic growth of around 2 percent would see the pendently, without internalizing negative output

aggregate EU deficit reduced to zero by 2017, with spillovers abroad, then this would be likely to ham-

outstanding debt reaching a peak of around 100 per the economic recovery. This adverse effect

percent of GDP. Of course, some countries would would be reduced if such policies were introduced

need a higher growth rate to achieve fiscal balance in a coordinated way, possibly leading to more

within this period. gradualism.

89 EEAG Report 2010

Chapter 3 • However, coordinated gradualism should not Favero, C. and F. Giavazzi (2007), “Debt and the effect of fiscal pol-

icy”, NBER Working Paper 12822.

interfere with the adoption of measures necessary Galí, J., J.D. López-Salido and J. Vallés (2007), “Understanding the

to preserve stability. The worst hit countries or the Effects of Government Spending on Consumption”, Journal of the

5, 227–70.

European Economic Association

countries with the most fragile public finances Giavazzi, F.and M. Pagano (1990), “Can Severe Fiscal Contractions

should adjust upfront and most deeply, to prevent be Expansionary? Tales of Two Small European Countries”, NBER

5, 75–111.

Macroeconomics Annual

the spreading of concerns about fiscal sustainabil- Giavazzi, F., T. Jappelli and M. Pagano (2000), “Searching for Non-

ity. If gradualism in the name of coordination Linear Effects of Fiscal Policy: Evidence from Industrial and De-

veloping Countries”, 44, 1259–89.

European Economic Review

feeds doubts about debt consolidation, then no Giordano, R., S. Momigliano, S. Neri and R. Perotti (2007), “The

coordination is a much better option. effects of fiscal policy in Italy: Evidence from a VAR model”, Euro-

22, 707–733.

pean Journal of Political Economy

Hebous, S. (2009), “The Effects of Discretionary Fiscal Policy on

Macroeconomic Aggregates: A Reappraisal”, mimeo, University of

Frankfurt.

References Heppke-Falk, K., J. Tenhofen and G. Wolff (2006), “The Macro-

economic Effects of Exogenous Fiscal Policy Shocks in Germany: a

Auerbach, A. and D. Feenberg (2000), “The Significance of Federal Disaggregated Analysis”, Deutsche Bundesbank Discussion Paper 41.

14,

Taxes as Automatic Stabilizers”, Journal of Economic Perspectives Ilzetzki E., E. Mendoza and C.A. Vegh (2009), “How Big are Fiscal

37–56. Multipliers?”, CEPR Policy Insight 39.

IMF Staff Position Note (2008), “Fiscal Policy for the Crisis”,

Barro, R. J. and C. J. Redlick (2009), “Macroeconomic Effects from SPN/08/01.

Government Purchases and Taxes”, NBER Working Paper 25369. Kellaway, M. (2009), “Public Sector Interventions in the Financial

Baxter, M. and R. King (1993), “Fiscal Policy in General Equilibri- Crisis”, London, Office for National Statistics.

83, 315–34.

um”, American Economic Review Mertens, K. and M. Ravn (2009), “Understanding the Aggregate

Effects of Anticipated and Unanticipated Tax policy Shocks”,

Beetsma R., M. Giuliodori and F. Klaasen (2006), “Trade Spillovers mimeo, EUI.

of Fiscal Policy in the European Union: a Panel Analysis”, Economic

48, 640–87.

Policy, Mountford, A. and H. Uhlig (2009), “What are the Effects of Fiscal

Policy Shocks?, 24. 960–92

Journal of Applied Econometrics

Bilbiie, F., A. Meier and G. Müller (2008), “What Accounts for the Pappa, E. (2009a), “The Effects of Fiscal Shocks on Employment

Changes in U.S. Fiscal Policy Transmission?”, Journal of Money and Real Wages”, 50, 217–44.

International Economic Review

40, 1439–70.

Credit and Banking, Pappa, E. (2009b), “The Effects of Fiscal Expansions: an Interna-

Blanchard, O. and R. Perotti (2002), “An Empirical Characterization tional Comparison”, mimeo, Universitat Autònoma de Barcelona.

of the Dynamic Effects of Changes in Government Spending and Parker J. A., N. S. Souleles, D. S. Johnson and R. McClelland (2009),

117, 1329–68.

Taxes on Output”, Quarterly Journal of Economics “Consumer Spending and the Economic Stimulus Payments of

2008”, mimeo, The Wharton School, University of Pennsylvania.

Caldara, D. and C. Kamps (2008), “What are the Effects of Fiscal

Policy Shocks? A VAR-based Comparative Analysis”, European Perotti. R. (1999), “Fiscal Policy in Good Times and Bad”, Quarterly

Central Bank Working Paper 877. 114(4), 1399–1436.

Journal of Economics

Perotti, R. (2005), “Estimating the Effect of Fiscal Policy in OECD

Christiano, L. J., M. Eichenbaum and S. Rebelo (2009), “When is the Countries”, Proceedings, Federal Reserve Bank of San Francisco,

Government Expenditure Multiplier Large?”, NBER Working Paper http://www.frbsf.org/economics/conferences/0503/fpoecd.pdf.

15394. Perotti, R. (2007), “In Search of the Transmission Mechanism of

Corsetti, G., A. Meier and G. Müller (2009a), “Fiscal Stimulus with Fiscal Policy”, 2007 22, 169–226.

NBER Macroeconomic Manual

Spending Reversals”, International Monetary Fund Working Paper Ramey, A. (2008), “Identifying Government Spending Shocks: It’s

09/106. All in the Timing”, NBER Working Paper 15464.

Corsetti, G., A. Meier and G. Müller, G. (2009b), “When Where and Ramey, A. and M. D. Shapiro (1998), “Costly Capital Reallocation

How Does Fiscal Stimulus Work”, mimeo, EUI. and the Effects of Government Spending”, Carnegie-Rochester Con-

48, 145–94.

ference Series on Public Policy

Corsetti, G., K. Kuester, A. Meier and G. Müller (2010), “Debt Con- Ravn, M., S. Schmitt-Grohé and M. Uribe (2007), “Examining the

solidation and Fiscal Stabilization of Deep Recessions”, American Effects of Government Spending Shocks on Consumption and the

forthcoming.

Economic Review Papers and Proceedings, Real Exchange Rate”, mimeo, Duke University.

De Castro, F. and P. de Cos (2008), “The Economic Effects of Fiscal Reinhart, C. M. and K. Rogoff (2008), “Banking Crises: An Equal

Policy: the Case of Spain”, 30, 1005–28.

Journal of Macroeconomics Opportunity Menace”, NBER Working Paper 14587.

Eggertsson, G. (2009), “What Fiscal Policy is Effective at Zero Romer, C. and D. Romer (2010), “The Macroeconomic Effects of Tax

Changes: Estimates Based on a New Measure of Fiscal Shocks”,

Interest Rate”, Federal Reserve Bank of New York Staff Report 402, forthcoming.

American Economic Review,

November. Sims, C.A. (1980), “Macroeconomics and reality”, 48, 1–48.

Econometrica

Eggertsson G. and M. Woodford (2004), “Optimal Monetary and

Fiscal Policy in a Liquidity Trap”, NBER International Seminar on Tenhofen, J. and G. Wolff (2007), Does Anticipation of Government

2004.

Macroeconomics Spending Matter? Evidence from an Expectation Augmented VAR,

Deutsche Bundesbank Discussion Paper 14.

Erceg, C.J. and J. Lindé (2010), “Is there a Fiscal Free Lunch in a Uhlig, H. (2005) What are the effects of monetary policy? Results

Liquidity Trap?” CEPR Discussion Paper 7624. from an agnostic identification procedure. Journal of Monetary Eco-

52, 381–419.

nomics

European Commission (2009a), “Public Finances in EMU”, Euro-

pean Commission Directorate-General for Economic and Financial Van den Noord, P. (2000), The Size and the Role of Automatic Fiscal

Affairs. Stabilizers in the 1990s and Beyond, OECD Economics Department

Working Paper 230.

European Commission (2009b), “Sustainability Report 2009”, Euro-

pean Commission Directorate-General for Economic and Financial Villaverde J. F. (2010), “Fiscal Policy in a Model With Financial

Frictions”, forth-

Affairs. American Economic Review Papers and Proceedings,

coming.

Eurostat (1995), “European System of Accounts”, Woodford M. (2010), Simple Analytics of the Government

http://circa.europa.eu/irc/dsis/nfaccount/info/data/ESA95/en/titelen.htm Expenditure Multiplier, presented 2010 AEA meetings.

Fatás, A. and I. Mihov (2001), “The Effects of Fiscal Policy on

Consumption and Employment: Theory and Evidence”, CEPR

Discussion Paper 2760. 90

EEAG Report 2010 Chapter 3

tax

various

on

cuts income

nd percent); percent)

percent)

tax

Finla deductibles

tax Increasing

Lowering pension

Income 0.1

0.7 0.1 (–

(– (– insurance rate

Estonia un- percent)

employment contribution

in

Increase 0.3

(+

cuts

mark percent)

percent) for

09 tax companies

credit

Den

20 Income 0.1

0.3

inn Tax (–

(–

e

op GDP)

Eur c

ubli

n contri-

social of

percent) percent)

i off goods

ep

asuress Write-down

entt

R Reduced security

ch

h butions

erc capital

0.5 0.2

e Cze

m (p (– (–

uss ess

mul easur semi-

the rate

rate gov-

sti the accommodation percent);

percent)

m of percent)

Cyprus governmental

VAT

organisations organisations

Fiscall tax

of from

dividend

semi Application

uee Reduction ernmental

corporate

en hotel

reduced income

the

ev 0.2 0.5

(<0.1

r No

for on

(– (–

Tax

x thresh-

in- percent)

contribution

pension

percent)

minimum

man-

aria income

Bulg in 0.9

the

(0.7

Increase datory (–

in social

sured olds rate

Cut

percent);

tax percent)

tax for

income income reduction

um percent)

construction

Belgi 0.2 0.1 residential

Flemish Federal

(– (– 0.1

VAT

cuts cuts (–

allowances

base tax

per-

tax contributions taxes

on

income

tax

from

income measures taxes

taxes

income gains

x:: Deductions Investment

Corporate

ndi Dividend

Personal business Capital

Social

Appe Other

sonal VAT

91 EEAG Report 2010

Chapter 3 alcohol excises percent)

of tobacco

Increases (0.05

and enterprises

contribution construction

percent);

tax the

state funded

social finance housing

scheme

to percent)

of percent)

contributions percent)

0.5 Suspension Supporting

in mandatory

minimum Boosting

Boosting

to rental

Increase (+ pension access

0.6 (0.05

basis (0.2

(+ of

and percent)

invest- transport

repair percent);

maintenance (0.1

Municipal

Building Green

ments

GDP) (0.1

off

entt

erc percent) Infrastructure

of

(p Indexation pensions

uress (+0.2

eas

m rate

ituree

of fees

land- airports the public

tourism tourism land

cancellation of

percent) VAT

local

the and of

on percent)

stay

percent) on end percent

of encouraging Application building

duty of

of authorities promotion

Reduction at minimum

overnight xp

by domestic

Boosting

hoteliers Increase

Increase fees 0.12 e

excise levied

petrol GDP) (<0.1

(0.13

(0.15 entt

and

ing on

(– nm

Gover in

needs sector

pensions

property

electricity local salaries

kerosene,

excise and allo-

the

taxes

administrative capital

for

cigarettes budgetary

economic percent) percent) percent) percent)

percent) contribution percent);

of in

in in

in for

valuations healthcare spending

and property Increase

Increase Increase Increase

Increase on cations

Higher

coal,

rates (0.3 (0.5 (0.1 (0.3

(0.3

and the

for (1

social percent)

in

Increase benefits (0.1 spending

expenditure

taxes business

to expenditure

investment

related measures

Environmental Infrastructure

duties industry

for

Measures Housing

housing Support

Excise Public Social

Other and

92

EEAG Report 2010 Chapter 3

contribution

minimum tax- pension

rate

tax;

per- of the

PIT the

increase percent).

a the

income into

Latvi of funded

Maintaining

of and

of exemption increasing

Reduction threshold accruing

social

increase scheme

wage;

sonal (0.63 state

of

percent)

invest- municipal

percent) (0.05 Italy

infrastructure (0.1 Funding mergers

ment stand- percent)

tax

band for

percent); of

infrastructure related

percent) d percent); rules

of Introduction

Irelan levy

tax

Widening (0.1

rate income interest

Stricter

0.1

(0.1 relief

(0.7

ard (–

GDP)

and

percent) percent) off

ary

consumption

Government

investment entt

Hung erc

wages

0.4 0.6 (p

(+ (– ess

easur

offsetting ex- income

petrol of

annual stock pertain-

the con- introduction

tax-

the supple-

m

Compensating percent) percent); line

infrastructure percent)

of above

percent) e

on

investments on

uee

Greec

on for euro

tax wage

with

impact in

measures duty rules

en tax

increase tribution

mentary options,

One-off income

ev 60.000

payers

(0.15 new to

(>0.1 (<0.1

(1.2 cise with

r

the The ing

Tax

x a

incl. allowance social

of rates

tax

support,

any Re-introduction )

percent) percent)

percent

income contributions

in

Germ Reduction

commuter

Income 0.1

0.2 0.3

in

cut (–

(– (–

invoices

tax subsidies

labour

the of ce

in percent) percent)

Acceleration of Fran

Reduction on payment

through

wedge (0.1 (0.1

education

research issues tax tax contributions

labour public employment expenditures

measures from

income income

environmental

on

on on

on

on Deductions

Spending

Spending Spending

Spending

Spending Personal personal

market Social

sector Other base

93 EEAG Report 2010

Chapter 3 non-alcoholic

standard and

excise

VAT alcohol,

percent). percent). percent)

wage petrol,

reduced of

of beverages

minimum Increases tobacco,

Increase on

0.34 1.92 0.74

certain

taxes

rates

and

(+ (+ (+

energy/banking/insurance

company percent)

percent);

tax percent)

income on of 0.3

0.2

0.1 tax revaluation

Corporate (+

on

(+

One-off

(– sectors

Taxes

assets

relief stamp

percent)

rate

capital excise

tax

cor- percent)

dates VAT

rate of

percent) percent)

percent) rate

and gains

Advancing in in

in (0.3 Reduction

tax

payment standard

Increase Increase

Increase

poration top

capital duties 0.1

gains duty

(0.1 (0.1

(0.3 (–

energy percent)

tax on

(surcharge)

Temporary companies

of

profits 0.1

(+ of

enterprises from divi-

payment introduction same of

the the

percent) percent)

percent) for

percent duties

gains

on and

in in

percent the holds

rate

Increase Increase

advance tobacco alcohol

for capital

dends; excise

(<0.1

(0.15 (0.15

80 tax

The

rate rate

65

to a

private rules

incl. favourable percent)

depreciation

investment,

to

Support 0.1

more (– capital percent)

tax of taxation

pay- on

corpo- complementary

of percent); turnover pharmaceutical

the percent)

the

and percent)

Acceleration government on

of companies in

(0.1

insurance dividend

to Increase tax Change

rations the

ments 0.3 0.1

gains

New

(0.1

on

(– (–

allowances

tax taxes

to

taxes

on

income related

measures taxes Environmental

taxes duties

gains

Investment

Corporate Measures

Dividend

business housing

Capital Excise

Other VAT

94

EEAG Report 2010 Chapter 3

social percent)

in payments

Increase 2.1

(+

tax

of

fighting household

evasion/avoidance support percent)

percent)

Intensified One-off income 0.2

0.1 (+

(+ wages public

pension

change of on

investment social

insurance Reprioritisation percent);

of levy”

percent) percent)

welfare

related percent);

percent); sector in

percent)

Introduction levy, of assets in Reduction

“Pension

transfers

Transfer package Savings

GDP) public public

Social

pay

health social 1.2 0.3 0.4

fund (0.3 (0.3

(0.5 (– (– (–

in off

entt (+

district

public nominal and

per-

the

payment social schemes

percent)

erc public

percent); Modernisation

of 0.25

monthly monthly pro-

the

the

(p percent)

for

Suspension in

and (–

of

uress transfers

in

Capping gramme

Savings

pension subsidy heating

0.15

wages

freeze

sector

salary

0.2 cent);

13th 13th

eas 0.1

(– (–

m

ituree public

for Cohesion

settlement end Fund percent)

percent) Restraining

xp National

e Social

entt (0.2

(0.5

Tax nm

Gover incl. support market

percent) percent) percent)

infrastructure

Investment, Industry support

Labour

0.3 0.1 0.1

(+ (+ (+

and industry favour pay

public housing

package percent)

percent) income work

percent) percent) percent)

households

automobile in

investment

Additional Measure measure

Making

for low

Fiscal 0.3

0.1 (0.1 (0.1 (0.1

Aid of

(–

(– spending education

research

expenditure

business labour public

expenditure

investment measures

measures Infrastructure on on

on

on

industry

for Spending Spending

Spending

Spending

Housing

Support market

Public Social

Other and 95 EEAG Report 2010

Chapter 3 investment investment

percent);

payroll com-

percent) rate

a

eni percent)

of for

tax

of 0.1

Elimination

Slov Additional Additional

Reduction allowance

corporate

0.6 (–

panies

0.1

(–

tax (–

resources social rate

a

of ani percent)

Rationalisation percent) contribution

Rom of

government Increase 0.8

0.3 (+

(–

increase reduc- con-

develop- groups

of gal percent)

percent);

payroll percent)

percent) for

social

Postponement in Portu Temporary tributions

pay Reduction selected

overseas aid of 0.2

service agreed 0.1

0.2 0.1 tion

ment (–

(–

(– (– GDP) percent)

bud- programmes

institutions

opera- income

percent)

govern- off

nd

percent)

of entt

Pola

costs

the 0.6

Personal

erc

in

in 0.25 (–

getary 0.2

tional ment (p

Cuts Cuts tax

(– (– ess

easur

remunera-

employment percent)

high-level

public income

2009

ex-

in items; percent)

m

cut a of

percent); wages

public percent) Malt uee

for Widening

the bands

percent en

(<0.1 personal

penditure

officials' freezing

sector's ev

growth in 0.2

elastic Public

sector Cuts r tax

(0.2

(0.3 tion (–

10 Taxx percent)

percent) tax

expenditure urg

health-care income

bo

Environmental percent) of (0.9

0.2 em Indexation personal

premium brackets

(+ Lux

Higher the sector

0.2 on

(+ tax system

so- pillar)

some income

pension

income percent) into

a percent); percent)

ani of dividends percent)

security

of (2nd

professions of tax

Lithu Reduction corporate

Inclusion Ongoing

personal Increase and

0.45 reform 0.3

(0.17 (0.4

cial tax on

(– (– allowances

tax

issues tax tax contributions

employment expenditures income

from

income income

environmental

on Deductions Investment

Corporate

Spending Personal personal Social

sector Other base

96

EEAG Report 2010 Chapter 3

and

percent) companies

sole SMEs

excise

percent)

for percent)

for

proprietors in (0.9

allowance Increase Support start-up

duties

0.2 (0.1

(– base

percent) percent) investment

excise property

tax

the percent)

0.1 0.1

of Updating

(+ local

Increase (+ Public

duties taxes 1

for (+ to

housing

burden per-

the

firms firms

exports

percent) support

percent) standard percent);

GDP)

of percent)

0.1

to to

Reduction SMEs

tax to (–

liquidity activity,

Support Support

Special

0.15 of

related

Lower GDP)

assets

0.1 VAT cent (0.1

(0.1

and

(– (– off

entt

business percent);

percent) percent)

percent) erc

duties Investment Investment

(p

on uress

Excise

Taxes 0.2 0.3 0.6

0.2 eas

(+ (+ (+

(– m projects tourism

ituree industry

percent)

excise incentives

reform

Environmental percent)

Vehicle investment

percent); percent);

end

percent) percent)

for

Investment

in measures to

xp

(0.3 Licences

Increase Support

Higher

related

e

Motor 0.1

duty entt

(0.1 (0.1

(0.1 (0.2

for

(– nm

Gover

on

of existing

tax the GDP

percent) govern-

capital or the investment:

percent) tax

for

the of

company of

in reduction

Replacement a

an

of by

increase percent

the in

(0.3

Abolition of Increase

children

(0.3

on capital bonus

new ment

paid one 0.7

tax

an

a alcohol

excise fuel, percent)

and

in on

Increase tobacco

duties (0.65 taxes business

to

taxes

on investment

related

measures measures

taxes Environmental

taxes duties

gains industry

for

Measures

Dividend

business housing Support

Capital Excise Public

Other Other

VAT and

97 EEAG Report 2010

Chapter 3 bill

worked

techno-

invest- specific

wage

for

R&D kind

subsidy hours

for sector percent)

percent) percent) percent)

new in in

and

Subsidies Increases transfers

in shorter Public

logies Wage

ment (0.1

(0.2 (0.6 (0.2 on

expenditure services

personnel percent) percent)

expenditure and

in Lower

0.9 goods 1.3

Cuts (– (–

schools ener- telecommu- infra-

in

percent); percent)

of Investment nications

premises

Renewal structure

and (0.2

(0.2 gy percent)

consumption

Intermediate 0.7

(–

institutions percent);

edu- other

local subsidies

maritime

– de- Environmental percent)

in

Infrastructure percent);

percent); at in

in percent)

percent)

percent) Sustainable

Investment velopment Reduction

Reduction

Education (0.1 measures subsidies

facilities cational energy

roads, 0.4

level 1

(0.1

(0.3 (0.1

(0.2 (–

(– expenditure

sector gov-

trans- percent); current

percent) percent)

public governments

in local

Reduction ernment

in

in wages to

0.7 0.5 0.9

Cuts

Cuts fers

(– (– (–

spending education

research issues

expenditure labour public employment expenditures

measures environmental

Infrastructure on

on on

on

on Spending

Spending Spending

Spending

Spending

Housing market sector Other

98

EEAG Report 2010 Chapter 3

m reduction

increase

do busi- percent) percent)

percent)

income

King duties

and percent)

of rate

rate Tobacco

Deferral

taxation

United

d alcohol

Lower 0.1 0.6

0.3 VAT

ness (0.1

(– (–

(–

nds percent)

social care

excise

for

erla percent);

percent) percent)

contributions health

in depreciation investments

Accelerated in

eth (0.1

Reduction premiums

Increase

N Lower

duties

0.3 0.2 0.1

Thee (– (– (–

companies

of improvement

GDP) companies

interest

deductibility corporate deduc- for

under-

on on

percent); percent) percent)

percent) percent) percent)

percent)

income

en contributions

social rules

taxes taxes

Swed off tax of

Changed Changed

pensions for

services

entt income pricing certain

earned tibility

Lower

Lower Lower Lower 0.2 0.2

0.1 0.3 0.2

0.5 0.1 home costs

erc Tax (+ (+

(– (– (–

(– (–

(p

ess Personal system of

easur to

reduction withholdings

percent)

percent) returns with

n the

m in

Spai Tax

uee of

Reduction taxpayers

Specific

en VAT

Change

Income 0.56

0.47

ev tax

r of (–

(–

Tax

x social and on

second

transfers

percent) pillar duties

contributions percent) percent)

tax in

a

aki Changes the tobacco

pension

Income capital Excise

0.2

Slov from (0.4 (0.2

(– allowances

tax taxes

tax tax contributions to

taxes

on

income related

from

income income measures measures

taxes Environmental

taxes duties

gains

Deductions Investment

Corporate Measures

Dividend

Personal personal business housing

Capital Excise

Social Other Other

VAT

base 99 EEAG Report 2010

Chapter 3 housing

industry

busi-

spending

Front-loading percent) percent) percent)

expenditure

for and

and

Support

capital Social

ness

(0.2 (0.2 (0.2 expenditure

deductible projects

infra- edu-

ex- market

percent) percent) percent)

percent)

Exceptional in

in measures

structure

Increase Increase Labour

penses cation

0.1 0.3 0.1

(0.1 GDP) (– (– (–

education coaching,

off main-

invest- infra-

entt percent)

percent) percent) and

research unemployed

expenditure

and

erc of

of activation

Increased

Increased

Increased structure

(p training

tenance

in 0.1

0.2 0.1

ment

uress and (+

(+ (+

eas Public

m Government

ituree improve

percent); percent) percent)

percent)

the strategic

Local

end

of Investment

mortgages tax

Abolition for

xp to

Central 0.72 0.27

0.15 0.21 sectors

certain

wealth e Fund Fund

nmentt (+ (+

(– (– Gover welfare 9a).

percent)

new 00

percent)

of

for in (2

purchases measures

Changes

(0.1

Subsidy n

mmissio

cars (0.5 o

spending education C

research issues

expenditure

business labour public an

n

employment expenditures

expenditure

investment measures pe

environmental

Infrastructure Euro

on

on on

on

on

industry

for Spending

Spending Spending

Spending

Spending

Housing

Support ce::

market

Public Social sector Other

and Sour

100

EEAG Report 2010 Chapter 4

I tive, (ii) be efficient, for example, not involving

MPLICATIONS OF THE CRISIS strongly increasing or variable tax rates, and (iii) that

US

FOR ADJUSTMENT NEEDS they should be politically feasible. The latter is cer-

tainly an important issue in the US, where tax paths

that would be easy to implement in Europe may be

1. Introduction politically impossible in the US.

The US economy is arguably following an unsus- Despite these caveats, an important policy question

tainable trajectory. The main indicators of this are a is whether current fiscal policy, including current

large current account deficit, a large federal budget spending programmes and tax law, is sustainable

deficit and trend-wise increasing costs of Social given forecasts of economic growth and demo-

Security and Medicare. In this chapter, we will dis- graphic change. If this is not the case, creditors will

cuss these observations and to what extent the in some way or another force the government to

financial and economic crisis may have changed the change its policy. Pre-crisis estimates of the long-

outlook. Before this, we need to define what we run sustainability of US fiscal policy suggests that

mean by sustainability. An often used definition of government income would have to be raised and/or

sustainability is that the inter-temporal budget expenditure cut. Gokhale and Smetters (2003) cal-

restriction is satisfied. In the context of fiscal sus- culated that the required change is dramatic – to be

tainability, this means that the discounted sum of sustainable, the government’s fiscal balance would

current and future government revenues (at least) have to be permanently improved by 6.5 percent of

covers the discounted sum of current and future GDP over pre-crisis long run forecasts (see

outlays plus the value of current outstanding debt. Box 4.2). To the extent that the crisis will have long-

Similarly, a sustainable path of current accounts run negative effects on economic growth and activ-

must imply that the discounted value of income gen- ity, the required adjustments are even larger. The

erated by exports of goods and services covers the Congressional Budget Office (CBO, 2009a), has

cost of imports and other transfers to abroad plus calculated that the required adjustment has

the initial value of foreign debt. Sustainability increased and is now 8.1 percent of GDP. They also

implies that that the ratios of government debt to show that by waiting, the required adjustment con-

GDP and foreign debt to GDP do not follow an

1

explosive path. tinues to grow; to 9.7, 12.1 and 15.5 percent of

GDP if adjustments wait for another one, two or

Defining a sustainable fiscal policy as one that satis- three decades, respectively. These figures illustrate

fies the inter-temporal budget constraint is not free of the claim by Kotlikoff (2006) that the US is going

problems. First, it can be argued that it is trivial in the bankrupt.

sense that if extraordinary income, like inflation

taxes or the implicit income generated by defaults is Recognising that fiscal sustainability requires a long-

included, all government spending will be financed in run perspective on future spending and revenues, our

one way or another. Second, the definition requires aim in this chapter will, however, be narrower. We will

that the future is predictable, which is difficult, but focus on two questions. We will first describe how fis-

the definition also allows different future scenarios. cal deficits have evolved over recent years and in par-

The multitude of different future possibilities calls ticular study how the forecasts for the coming decade

for other criteria to be included in the definition of have changed during the crisis. This allows us to dis-

sustainability. Specifically, one may require that the cuss whether the crisis has increased the urgency of

future paths of government revenues and outlay fiscal consolidation. The second question is to what

should (i) be fair from an inter-generational perspec- extent the current account deficit of the US is an indi-

cator of a non-sustainable consumption pattern that

needs to be corrected.

1 See, e.g., Neck and Sturm (2008) on this. 101 EEAG Report 2010

Chapter 4 say 3 percent thus requires a primary surplus of

Before discussing the empirical issues, we need to (0.2 0.03 =) 0.6 percent of GDP. During the tran-

demonstrate some arithmetic facts and make some *

sition to the steady state from a lower debt ratio,

definitions. The definition of sustainability as the crowding out is smaller than in steady state.

derived from the inter-temporal budget constraint Therefore, the transition involves inter-generational

of the government implies that public debt as a transfers from future to current generations.

share of GDP is non-explosive. However, this does

not rule out a permanent deficit. As shown in

Box 4.1, a permanent fiscal deficit will lead to a

d

where is the

stable debt-to-GDP ratio of d/g g 2. US government debt

growth rate of nominal GDP. If, for example the

deficit is 3 percent, a 5 percent growth rate of nom- Let us now focus on the development of the debt of

inal GDP leads to a stable debt-to-GDP ratio of the US government. An immediate problem is that

60 percent. Given that interest rate on government there are a number of different definitions of gov-

debt tends to be larger than the growth rate of ernment debt. The two main dimensions in which

GDP, a positive debt crowds out other types of these measures differ are (i) what branches and levels

spending or requires higher government revenues, of government are included and (ii) whether the

i.e. it must be financed by a primary surplus. In measure is gross or net of financial assets like loans

steady state this crowding out is equal to the rela- to the private sector, shares in publically traded com-

tive difference between the interest rate and the panies and foreign exchange reserves. A particular

growth rate times the deficit. As an example, if problem, making international comparisons diffi-

interest rates are 6 percent and the growth rate of cult, is that assets and liabilities related to the pen-

GDP is 5 percent, the interest rate is 20 percent sion system are treated differently in different coun-

larger than the growth rate. A permanent deficit of tries. For example, the US government has an un-

Box 4.1 The arithmetic of stable deficits

The law-of-motion for public debt can be written

(1)

( ) /(

1 )

b b d g

1 1

t t t t

where is the debt-to-GDP ratio in period is the nominal growth rate of GDP between period and and

b t+1, g t t+1 d

t+1 t+1 t

is the fiscal deficit. Here we note that the term 1+ , i.e., GDP-growth, tends to reduce the debt ratio. All else equal, a

g

t+1

high rate of GDP growth dilutes the debt-to-GDP ratio by spreading a given debt over a larger GDP base.

It now follows immediately that as long as growth is positive, any constant deficit will lead to a non-explosive debt ratio

and is therefore consistent with the general definition of sustainability. Specifically, assuming for simplicity a constant

GDP growth rate, a constant deficit leads to a constant debt ratio of We see this by assuming all variables in

d d/g.

equation (1) to be constant and then solving for b; d

Ÿ

(

1 )

b g b d b .

g

This is, however, generally not a free lunch for the government, since a higher debt requires more interest payments.

Specifically, note that the fiscal deficit can be written

d r b p (2)

t r t t

where is the interest rate on public debt and is defined as government revenues minus net interest

r p the primary surplus,

t t

payments. Using the steady state result in this equation and solving for we find that in a steady state with deficit

b=d/g p,

the primary surplus must satisfy

d,

r g

p d . (3)

g

Thus, as long as the interest rate is higher than the GDP growth rate, a permanent deficit leads to a debt buildup that

requires a permanent primary surplus. Specifically, at any point in time, sustainability requires that outstanding

government debt equals the discounted sum of all future primary surpluses.

The arithmetic of budget deficits can easily be adapted to the issue of sustainability of the foreign affairs of a country. The

equivalent of a budget deficit is a current account deficit, and a permanent current account deficit is therefore in principle

consistent with sustainability. Specifically, a constant current account deficit leads to a stable foreign debt ratio equal

CA b f

to where is the growth rate of nominal GDP. The counterpart of the primary surplus is the trade balance plus other

CA/g, g

net income from abroad not related to the debt position. Sustainability requires that the initial foreign debt is matched by

future surpluses from trade and other income. 102

EEAG Report 2010 Chapter 4

Box

x 4. 2 Lo n g runn fiscall and

d generati onall im bal ancess

To evaluate the long run fiscal sustainability and inter-generational fairness, the development of government debt is an

insufficient indicator. A major reason is that various long-term government payment obligations, notably in Social

Security and Medicare, are underfinanced and the shortfall is not included in the national debt. Gokhale and Smetters

(2003) argue that the government should report two more informative measures, namely the fiscal imbalance (FI) and the

generational imbalance (GI). The first is defined as the difference between the present value of projected government

spending (not counting interest payments) plus current debt and the present discounted value of projected revenues. Both

spending and revenues are calculated using current spending programmes and tax law but take into account trends in

demography and expected growth. If FI is positive, it means that sooner or later government spending must be reduced

and/or revenues (taxes) increased to prevent government bankruptcy – i.e., the current fiscal system is not sustainable.

past and living

The generational imbalance is defined as the share of the fiscal imbalance that is due to generations. A

higher GI measure thus implies that past and living generations are passing on a debt to future generations.

Gokhale and Smetters (2003) estimate that FI is of an order of 4 times GDP, i.e. much larger than the official debt. About

80 percent of this debt is due to the Medicare system being seriously underfinanced in the long run. Around 40 percent of

the deficit in the Medicare system is due to past and living generations. This means that past and living generations are

handing over a debt larger than GDP to future generations via Medicare. The imbalance in social security is smaller but is

mostly due to past and living generations, implying that another debt almost as large as GDP is handed over to future

generations.

The calculations by Gokhale and Smetters imply that US fiscal policy is very far from being sustainable. Changes must

necessarily be made and the later these are undertaken the larger they must be. If measures are taken immediately, they are

still very large. If no spending cuts are undertaken, revenues must permanently be increased by 6.5 percent of GDP. If this

extra revenue is to be achieved by income taxes, these have to be raised by 16.6 percentage points according to Gokhale

and Smetters calculations. Alternatively, Social Security and Medicare spending could be permanently cut in half.

The European Commission regularly publishes sustainability reports (EU 2009). The key index is the S2 indicator that in

principle is similar to the GI measure. However, rather than expressing the deficit in the inter-temporal budget, it is

defined as the necessary improvement in the government budget to reach sustainability. The S2 indicator shows that also

EU needs to change its fiscal policies. In particular, the aging population implies that the fiscal balance of the EU as a

whole needs to be improved by 3 percent permanently. This number should be compared to the 6.5 percent necessary

improvement for the US. In the latest sustainability report, the consequences for sustainability of various post-crisis

scenarios are evaluated. It is shown that if growth remains lower than what was previously expected for the coming

decade, only gradually returning to trend at 2020, an extra percentage point improvement in the fiscal balance is required

to reach sustainability. measure of net debt, where the government’s financial

funded pension liability to its own employees assets are deducted from its financial liabilities. The

amounting to around 10 percent of GDP that is not

2 third curve is the Congressional Budget Office’s

For other coun-

included in the official debt figures. (CBO) measure of US federal government debt held

tries, this implicit debt may be larger or smaller than 4

for the US. by the public. Assets and intra-government assets

and liabilities are here netted out.

In Figure 4.1, we report three

measures of the US government’s Figure 4.1

debt. The upper curve is the

OECD’s measure of general gov-

ernment gross financial liabilities

including forecasts for year

3 Here, all branches

2009–2011.

and levels of the government are

consolidated, but financial assets

of the government are not de-

ducted from the debt. The next

curve from above is the OECD

2 OECD reports that this debt was 10.2 per-

cent of GDP in 2005.

3 OECD Economic Outlook No. 86 Annex

Tables.

4 Historical data from CBO (2008) and

forecasts from CBO (2009). 103 EEAG Report 2010

Chapter 4 In Figure 4.1, we see that the Figure 4.3

curves for OECD net and gross

debt are fairly parallel. A closer

look reveals that net debt has

been around 70 percent of the

gross figure for the whole period.

The higher debt level for the fore-

cast years 2009–2011, shows that

while the net debt remains at

around 70 percent of gross debt,

the difference has increased from

less than 20 to close to 30 per-

centage points. We also note that

while the CBO measure of net

debt differs from OECD mea-

sures, the differences are quite

small, reflecting small levels of

debt in state and local govern- line, where data from 2008 are the forecasts done in

ment. Since the net debt is the one most relevant for January 2008.

our notion of sustainability and since CBO produces

biannual long-run forecasts of the US fiscal balances, The CBO forecasts changed dramatically during the

we will use the CBO measure for the analysis in this macro-financial crisis. In August 2009, the forecast

section. was that the debt of the US government would in-

crease at a high speed. The increasing trend in the

In January 2008, just before the full consequences of debt as a share of GDP during the 1980s seems to be

the financial crisis for the macroeconomic develop- back and perhaps even stronger than before, as indi-

ments became apparent, the CBO forecasted that cated by the red dashed line.

the fiscal deficit would continue to be on a decreas-

ing trend. After three years of declining deficits, CBO’s federal deficit forecasts calculated in January

CBO forecasted a small reduction in deficits to 2009 and August 2009 are shown in Figure 4.3. The

around 1.5 percent of GDP for 2008–2011, a small- largest change in the forecast is for the year 2009. In

er deficit for 2012 and surpluses thereafter. Thus, January 2008, the forecast was a deficit of 1.5 per-

the pre-crises forecast implied substantially falling cent of GDP. Half way into 2009, this forecast had

debt-to-GDP ratios over the coming decade. That changed to an unprecedented deficit of 11.2 per-

meant that the trend towards increasing debt was cent. In fact, this is almost twice as large as the pre-

broken. This is shown in Figure 4.2 as the blue solid vious record deficit of 6.0 percent in 1983. Fur-

thermore, the CBO forecast

implies that the deficit remains

Figure 4.2 above the previous record in

2010 and 2011.

Furthermore, while GDP growth

was according to the 2008 CBO

forecast projected to rebound

substantially in 2012, the budget

deficit is not expected to return to

the previous track. The 2008 real

GDP growth forecast was a mod-

erate 2.9 percent for 2012 and

2.6 percent for 2013 and 2014.

During this period, the 2008 fore-

cast implied a turnaround in

deficits, from negative in 2012 to

104

EEAG Report 2010 Chapter 4

fiscal stimulus act contains tax

Figure 4.4 cuts to individuals and firms as

well as increased spending on

infrastructure, education and

transfers such as more generous

unemployment benefits. The

CBO estimated that the bill in-

creases the deficits by 1.3 per-

cent of GDP for the fiscal year

2009, which ended on 30 Sep-

tember 2009. The budgetary

effects peak during the fiscal

year 2010 at 2.8 percent of GDP

and then fall to below 1 percent

of GDP the following year.

The large debt build-up during

the crisis means increasing inter-

positive from 2013 onwards. The revised CBO growth est payments in the long run. In the 2009 forecast,

forecasts implied a high 5.0 percent real GDP growth interest payments are nearly three times as high as in

rate for 2012 and 4.5 and 3.0 percent growth for 2013 the 2008 forecast, i.e., 3.3 percent rather than 1.2 per-

and 2014, respectively. Despite these optimistic cent of GDP.

growth forecasts, deficit-to-GDP ratios are projected

to be close to 4 percentage points higher during this As we noted in the introductory section, sustainabili-

period than previously forecasted. Subsequently, there ty requires that the outstanding government debt is

is even a tendency for the discrepancy between the two balanced by future primary surpluses. Before the 2009

forecasts to increase. crisis, government debt was in the order of 35 percent

of GDP. For this debt ratio to remain stable, future

The gigantic increase in the deficit for 2009 is due to primary surpluses needed to average a few 10ths of a

a 4.1 percentage point reduction in revenues and a percent of GDP. In Figure 4.6, we depict the primary

5.7 percentage point increase in outlays. As seen in surplus forecast before and after the crisis. Before the

Figure 4.4, the difference between the two revenue crisis the forecasts implied the primary surplus should

forecasts is temporary, reflecting a projected re- stabilise at around 2 percent of GDP. This clearly

bound of tax bases. By 2011, the difference is exceeded what was necessary to keep the debt ratio

1.1 percentage points and falls to close to zero after constant but was, as noted above, far from sufficient

that. The picture for outlays is quite different and to compensate for the projected long-term increases in

depicted in Figure 4.5. Even at the end of the fore-

cast period, i.e., in 2018, total

outlays as percentage of GDP Figure 4.5

are in the 2009 forecast are close

to 4 percentage points higher

than in the 2008 forecast. Most

of the increases in outlays are

projected to fade away over time.

For example, the Troubled Asset

Relief Program (TARP) and

costs associated with the sup-

port to Fannie Mae and Freddie

Mac add a full 3 percent of

GDP to the projected deficit for

2009. The American Recovery

and Reinvestment Act also adds

substantial amounts to the fed-

eral deficit. This comprehensive 105 EEAG Report 2010

Chapter 4 rent account deficit is worri-

Figure 4.6 some. If it is unsustainable, it

must eventually be corrected,

and there is ample empirical evi-

dence that such corrections often

are done in the form of a “sud-

den stop”, where creditors

abruptly halt the financing of a

deficit. If such a sudden stop

were to happen to the US, this

would obviously have dramatic

consequences for the global

economy. Even if the US current

account deficit was not directly

responsible for the crisis, it con-

tinues to be a worry. However, it

is important to note that there

the costs of Medicare and Medicaid after the depict- are substantial discrepancies in different measures of

the US current account. As we will see, these mea-

ed forecast period. sures differ substantially in how alarming a picture

they paint. By definition, a country’s current

The 2009 crisis has changed the picture quite dramat- account deficit is equal to the change in its net for-

ically. The debt build-up due to the crisis is around eign liabilities. Measuring the current account by

50 percent of GDP, which will require additional summing flows of income should yield the same

future primary surpluses in the order of a quarter to result as measuring the change in the values of assets

a third of a percent of GDP. Instead, the latest fore- and liabilities. However, since income generated by

cast implies primary surpluses around 2 percentage revaluations in assets and liabilities are not necessar-

points of GDP lower than the pre-crisis forecast. ily recorded accurately, there are large discrepancies

Clearly, if there were doubts about fiscal sustainabili- 5

between the measures.

ty before the crisis, these doubts have strengthened

substantially. There now seems to be no other way for Figure 4.7 shows two measures of the US current

the US than to immediately reconsider its fiscal poli- account from 1981 to 2008. The solid line with

cies. This must be done sooner or later, but on squares is the OECD measure, calculated from

grounds of intergenerational equity and efficiency, transactions data. The solid line without squares is

sooner is better than later. instead measured as the yearly change in the net

asset position of the US. The dotted lines are linear

regression trends. As we see, there are substantial

3. US current account differences between the two curves. The transaction-

based measure is consistently negative, except for a

For some time, the US current account has been at the small surplus in 1981 and 1991. There is also a

centre of policy discussions about the so called global strong negative trend such that the current account

imbalances. A key feature of these imbalances is the worsens by 1.7 percentage points per decade. The

large current account deficit in the US, financed by valuation based measure is instead dominated by

surpluses in swiftly developing Asian countries, large swings, reflecting large revaluations of the

notably China, and in oil exporting countries. The asset positions of the US. The trend in the latter

sustainability of the current account deficit of the US measure is barely negative at – 0.5 percentage points

has been questioned. Furthermore, the large demand per decade. The average deficit over the whole peri-

for liquid assets from surplus countries has led to low od differs substantially – it is 2.8 percent of GDP

interest rates in a time of high output growth rates for the transaction-based measure and only 1.6 per-

and, in particular, optimistic expectations about cent for the valuation based. The difference is large

future growth.

Although it is clear that a balanced current account 5 See Gourinchas and Rey (2007a,b), Curcuru et al. (2008) and

at all times should not be a policy target, a large cur- Hausman and Sturzenegger (2007).

106

EEAG Report 2010 Chapter 4

foreign debt ratio might contin-

Figure 4.7 ue to increase, but not indefi-

nitely. The reason is that nomi-

nal GDP growth tends to

reduce debt as a fraction of

GDP. When GDP grows, the

debt can grow at the same speed

without increasing the debt-to-

GDP ratio. Since a growing

debt is equivalent to a current

account deficit, the latter is con-

sistent with a constant debt

ratio. Applying the calculations

in Box 4.1, we find that this

constant debt ratio is given by

the current account deficit

divided by the nominal growth

rate of GDP.

also for the last decade, where the valuation-based

measure is at a deficit of 2.2 percent and the trans- Again assuming a 5 percent growth rate of US

action based at 4.8 percent. Also including 2009 will nominal GDP, a current account deficit of say

reduce this difference but only by about 1 percent- 3 percent would lead to the foreign debt stabilising

age point. 3

at / = 60 percent of GDP, i.e. approximately a

5

doubling of the current debt. A debt of 60 percent

Looking at the two curves, it is clear that the dis- of GDP is large but not large enough to be obvi-

crepancies between the two current account mea- ously impossible to sustain. Again in parallel to the

sures have increased quite dramatically over time. case of fiscal deficits, a constant debt ratio needs to

To understand this, we need to consider the fact that be balanced by a positive trade balance in the same

behind the trend-wise deterioration of the US net way government debt must be financed by a posi-

asset position, there are much larger changes in tive primary surplus. In other words, part of the

gross positions. Since 1986, when the net US foreign revenue generated from abroad must be used to pay

debt was approximately zero, net foreign debt has the percentage difference between the interest rate

increased to 24 percent of GDP in 2008. However, and GDP growth times the deficit and cannot be

gross assets have increased from 33 percent in 1986 used for consuming imports. Given that the interest

to 139 percent in 2008 and gross debt from 33 to rate on US foreign debt remains only marginally

164 percent of GDP during the same period. Thus, above the growth rate of GDP, sustainability in

the ratio of debt to assets has only increased from itself requires only quite small increases in the

unity to 1.17. Under-reporting of income generated trade balance to stabilise this higher foreign debt

by the increasing large gross foreign asset position level.

yields an increasing divergence between the two

measures. Since revaluations of foreign assets and So far, we have implicitly assumed that the return on

liabilities are more volatile than export and import, assets and liabilities is equal. However, the US has

the volatility of a valuation-based measure increas- consistently earned a higher return on its foreign

es relative to the OECD measure as gross positions assets than what it pays on its foreign debt.

increase. Gourinchas and Rey (2007a), for instance, report

that over the post-Bretton Woods period, the aver-

As we have seen, both measures of the current age real return on US foreign assets has been as high

account show a downward trend in the current 6.8 percent while the real interest rate paid on liabil-

account. Clearly, the continuation of these trends is ities has been 3.5 percent. While there is some

not consistent with sustainability. However, in exact debate on the magnitude of the US return privilege

parallel to the case of fiscal deficits, sustainability (Curcuru et al. 2008), there is little doubt that net

does not necessarily rule out a stable deficit. inflow of asset income has been historically positive.

Suppose that the US current account deficit sta- To the extent that this return privilege remains, a

bilised at some level. In such a scenario, the US net

107 EEAG Report 2010

Chapter 4 more highly leveraged portfolio generates more ed return on assets is higher than the cost of bor-

income to the US and even a portfolio where liabil- rowing. Quite surprisingly, however, Gourinchas

ities are larger than assets can generate a positive and Rey (2007a) find that the return privileges are

income that can be used to finance a negative trade mostly accounted for by the return effect. Of the

balance. three plus percentage point return privilege in the

post-Bretton Woods period, only about a quarter

US net nominal capital income from abroad is can be accounted for by the composition effect.

where and are the nominal rates of Doubt is cast on this result, however, by Curcuru et

r A–r L, r r

a l a l

return on assets and liabilities, respectively. As al. (2008), who argue that the US earns little excess

noted by Obstfeld and Taylor (2005), there is a tip- return within asset classes. As yet, this issue is unre-

, such that only if the ratio

ping point, given by solved.

r /r

a l

is larger than the tipping point, net capital

L/A

income is negative. Given the historical real capital It seems reasonable that the more developed financial

returns and adding an inflation rate of 2 percent, markets in the US allow its citizens to take on more

the tipping point is 1.6, i.e., substantially above the risk. Thus, the US is in a better position to take ad-

current value, indicating that if historical rates are vantage of the equity premium. To the extent the

on its foreign

maintained, the US still return premium is due to the equity premium, it seems

makes money

asset portfolio. reasonably safe to assume that it will remain. Less,

however, is known about return differences within

Suppose that the real rate of return paid on US for- similar assets. In a recent paper, Hassan (2008), shows

eign debt remains at 3.5 percent and that the infla- that returns on similar assets denominated in different

tion rate is 2 percent. Suppose also that the current currencies do vary systematically. Assets denominated

account stabilizes at 3 percent. Then, if the US had in currencies of large OECD countries have paid sub-

no foreign assets or had no return privilege, the stantially lower returns over the period 1980–2007.

trade balance must be positive at a value given by The estimates are stark – increasing the size of a coun-

one third of a percent of GDP if the nominal GDP try from zero to 10 percent of world GDP reduces the

growth rate is 5 percent. With current assets of rate of return by over 2 percent. The result appears

139 percent of GDP and a return privilege of quite robust and holds for bonds of different maturi-

3.3 percent, this generates excess revenues of 4.5 per- ties as well as for stocks. Hassan also provides a theo-

cent of GDP, i.e., much larger than the trade balance retical explanation for his findings. A shock to the

required without the return privilege. If the return productivity in a country affects its real exchange rate.

privileges were to disappear, the US would therefore Since a large country has a larger impact on the

need to make very substantial adjustments, in par- world, its exchange rate is more negatively correlated

ticular reduce consumption very dramatically. Such with world consumption of tradables. Therefore,

a change would most likely require large dollar assets denominated in the currencies of large coun-

depreciation. tries provide a better insurance against fluctuations in

tradables consumption and yields a lower expected

Recently, several authors have analyzed the reasons return.

for the return privilege. Gourinchas and Rey (2007a)

decompose the return difference into a return effect There is a systematic currency bias in the US for-

and a composition effect. The latter comes from bor- eign portfolio – assets are more often denominated

rowing in one type of assets and investing in another, in foreign currency than liabilities. Therefore,

for example borrowing in fixed income short maturity Hassan’s results point to a fundamental factor indi-

instruments and investing in stocks. The return effect cating that the return privilege of the US may

instead arises to the extent that the US manages to get remain, also if it is not due to maturity conversion.

higher returns on assets than on liabilities within each But clearly, much caution is warranted regarding

asset class. this prediction. The surprisingly strong apprecia-

tion of the dollar in autumn 2008, when the crisis

It is well known that maturity conversion is an took a sharp negative turn, and the dollar behav-

important part of US foreign investment strategy – iour afterwards, have also focused attention on the

Gourinchas and Rey call the US the “venture capi- hedging properties of different currencies and

talist” of the world. By investing in foreign equity assets against “disaster” risk. According to a lead-

and borrowing in short maturity bonds, the expect- ing opinion among market participants, also men-

108

EEAG Report 2010 Chapter 4

tioned by Fed chairman Ben Bernanke in a public are willing to pay has widened during the crisis.

Although increased taxes and spending cuts should

speech in November 2009, the US currency not be undertaken before the economy is well on its

strengthens whenever markets’ assessment of glob- way out of the crisis, we urge the US government as

al systemic risk increases. Specifically, this opinion well as fellow economists to help prepare US citizens

seemingly contrasts different regimes driving global to accept substantial changes. A serious discussion

risk factors, whereas fear of disasters would moti- about the introduction of a federal value added tax

vate a flight to quality into the largest economic and increased personal income taxes must begin

and military power of the world. This observation immediately.

points in the same direction as the argument of

Hassan (2008) – the US enjoys a return privilege We also believe that the US government should sys-

because of the currency composition of its assets tematically produce indicators of its long-run fiscal

and liabilities – although, again, the quantitative viability and the consequences for intergenerational

relevance is unclear. redistribution of current policies. Such indicators

should be based on forecasts of spending and rev-

enues given current laws and long-term projections of

4. Concluding discussion growth and demographic change. Sustainability indi-

cators are already produced for the EU by the Euro-

The medium-run forecasts of government debt pean Commission but would seem to be also useful

before the crisis were consistent with sustainability, for the US An important purpose of a systematic

showing a primary surplus for the coming decade. In reporting of sustainability indicators is to affect the

the longer run, however, there is no doubt that very public debate. Hopefully this can lead to an increased

large adjustments are necessary and will be under- awareness that spending programmes must be

taken, voluntarily or by the force of creditors. In financed. Then, the government’s inter-temporal bud-

particular the cost of Medicare, Medicaid and Social get constraint may be satisfied in a process with a

Security is forecasted to grow substantially, mostly deliberate consideration of the consequences for effi-

due to a larger share of elderly. This is shown in ciency and intergenerational equity of different poli-

Figure 4.8. cies. The alternative is that creditors will eventually

force a change in policy in which these considerations

The worrisome factors behind the longer run sus- are likely to be absent.

tainability have certainly not disappeared during the

current economic crisis. The difference between Regarding the current account deficit, we are less

before and after the crisis is instead that now also the worried by the actual deficit and its long run trends

medium-run forecasts point towards non-sustain- than by the US’s vulnerability to its foreign return

ability. We are also worried that the discrepancy privileges. Given that these remain, they yield a suf-

between what US citizens’ expectations of what the ficient income to finance a large trade deficit. Were

government should provide and how much tax they they, however, to disappear,

quite dramatic adjustments

would be required, likely includ-

Figure 4.8 ing a large depreciation of the

dollar. It is difficult to make

forecasts for the return privilege.

On the one hand, it appears

unlikely that financial markets in

China and other emerging mar-

ket economies will develop

quickly. On the other hand, the

restructuring of the financial

sector in the US made necessary

by the current crisis may reduce

its productivity advantage and

thus its ability to generate excess

returns on the world capital

markets.

109 EEAG Report 2010

Chapter 4 References

CBO (2008), The Budget and Economic Outlook: Fiscal Years 2008

to 2018, January 2008.

CBO (2009a), The Long-Term Budget Outlook, June 2009.

CBO (2009b), The Budget and Economic Outlook: An Update,

August 2009.

Curcuru, S., T. Dvorak and F. Warnock (2008), “Cross-Border

Returns Differentials”, Quarterly Journal of Economics 123(4),

1495–1530.

European Union (2009), “Sustainability Report 2009”, European

Commission, Directorate-General for Economic and Financial

Affairs,

http://ec.europa.eu/economy_finance/publications/publica-

tion15998_en.pdf

Gokhale, J. and K. Smetters (2003), “Fiscal and Generational

Imbalances: New Budget Measures for New Budget Priorities”, The

AEI Press.

Gourinchas, P.-O. and H. Rey (2007a), “From World Banker to

World Venture Capitalist: The US External Adjustment and the

Exorbitant Privilege”, in R. Clarida (ed.), G7 Current Account Im-

balances: Sustainability and Adjustment, Chicago: University of

Chicago Press, 11–55.

Gourinchas, P. -O. and H. Rey (2007b), “International Financial

Adjustment”, Journal of Political Economy 115, 665–703.

Hassan, T. (2008), “Country Size, Currency Unions, and

International Asset Returns“, The University of Chicago Booth

School of Business, mimeo.

Hausmann, R. and F. Sturzenegger (2007), “The Missing Dark

Matter in the Wealth of Nations, and Its Implications for Global

Imbalances”, Economic Policy 22, 469–518.

Kotlikoff, L.J. (2006), “Is the United States Bankrupt?”, Federal

Reserve Bank of St. Louis Review 88(4), July/August.

Neck, R. and J.-E. Sturm (editors) (2008), Sustainability of Public

Debt, MIT Press.

Obstfeld and Taylor (2005), “Sources of America’s ’Exorbitant

Privilege”, mimeo, University of California. 110

EEAG Report 2010 Chapter 5

T F C : R debt more problematic, which increases the likeli-

HE INANCIAL RISIS ISKS hood of an exit from the euro area or of a default

AND CHALLENGES FOR THE on public debt. The rise of the spreads during the

crisis suggests that over a ten-year horizon and for a

EURO AREA peripheral country, markets do not consider those

possibilities as rare events.

1. Introduction One case in point is Greece. In December 2009, its

sovereign debt was downgraded to BBB. The spreads

The current crisis has led many analysts to re-assess shot up again as debt is quickly growing well beyond

the role of the euro. At face value, the euro area has 100 percent of GDP, while low competitiveness due to

done relatively well at avoiding the massive finan- past cumulated inflation differentials makes it difficult

cial crisis of the Anglo-Saxon countries. Does the to exit the recession. Possible scenarios include out-

crisis prove the virtues of the euro, or can it be a right default, exiting the euro area, or a bail-out from

source of tensions that stress the viability of the core euro countries. None of these scenarios is

monetary union? In this chapter we discuss these favourable for the euro. A bail-out can be especially

issues. We acknowledge that membership in the problematic if it fails to prevent contagion to other,

euro area has helped to eliminate the possibility of much larger economies with a public debt overhang,

a “twin crisis”, i.e. a joint banking and balance of such as Belgium or Italy, for which a bail-out would

payment crisis in the member countries. To the be too costly.

extent that such crises are self-fulfilling rather than

driven by fundamentals, this is unambiguously ben-

eficial. On the other hand, the crisis brings about 2. The international transmission of the crisis

some scenarios that may be problematic for the

euro area. One such scenario is a rapid, excess Historically, macroeconomic shocks that originate

appreciation of the euro reflecting a flight out of in the United States eventually spread to Europe,

US assets. Another is a balance-of-payments crisis but this happens with a substantial lag. Typically,

in Central and Eastern European countries. Despite the transmission is thought to take place through

the fact that these countries are not members of the 1

international trade. Essentially, a recession in the

monetary union, they are slated to join some day, US is associated with a fall in import demand by US

and financial and macroeconomic fragility there consumers, which reduces the demand for foreign-

affects the euro area. produced goods and thus depresses aggregate

demand in the rest of the world. The effect is small

Finally, we document a number of asymmetries and because the share of imports in consumption expen-

imbalances between the core members of the mone- ditures is not very large; and it is associated to a lag

tary union, in particular with respect to inflation because it takes some time for consumers to rebal-

differentials and net foreign asset positions. It is ance their expenditure and for exporters to realise

unclear whether the crisis has exacerbated or damp- that demand has fallen and to adjust their employ-

ened these asymmetries. But the evolution of ment and production decisions. Thus, Krugman

spreads in government yields during the crisis sug- (2008) has argued that for aggregate demand in the

gests that the credibility of the euro area is not rest of the world to be reduced by 1 percent, the US

absolute. It is plausible that these asymmetries, would need to be in a recession where output has

while not accentuated by the crisis, undermine the fallen by 8 percent.

credibility of the area, which itself becomes more of

an issue in times of crisis. That is, a shrinking eco- 1 The academic literature on the international transmission of busi-

nomic activity may make imbalances such as low ness cycles is large. The reader may refer to Clark and van Wincoop

competitiveness, high trade deficits or high public (2001), Canova and Dellas (1993), or Calderon (2008).

111 EEAG Report 2010

Chapter 5 equity. This is illustrated by the following example.

In that respect, the recent crisis seems unique in that Suppose an investor has 10 shares, worth 10 each,

despite having originated in the United States, its financed with equity of 40 and debt of 60. The total

transmission to the euro area has been instantaneous value of the portfolio is equal to 100. If stock prices

and the magnitude of the recession has been of the fall by 20 percent, equity falls to 10 8 – 60 = 20, i.e.

same order as in the US. The reason for this unusual *

it falls by 50 percent. The ratio of equity over total

pattern is that the transmission mechanism is differ- portfolio value falls from 40 percent to 20/80 = 25 per-

ent; due to financial globalisation, there now exists an cent. Assuming the firm wants to restore that ratio,

international financial transmission mechanism of short of getting new capital, for example by issuing

macroeconomic disturbances, and this mechanism is new shares, it must sell assets to reduce its debt – this

more rapid than the traditional one. Thus, the world is the essence of the deleveraging process. Assume it

economy is now in a regime where economies are sells n shares. Then its debt falls to D’ = 60 – 8 n, and

more interdependent and react more quickly to *

its equity is unchanged at 8 (10 – n) – (60 – 8 n)

shocks in other countries. * *

= 20. To restore a ratio of 20/(20 + D’) = 0.4, we need

D’ = 30, so that the firm needs to sell 3.75 shares. To

The increased financial interdependence is illustrated fix ideas, let us assume that 4 shares are sold.

in Figure 5.1 (taken from Krugman), which shows

that in three decades the level of foreign assets in the Therefore, a fall in asset prices induces investors to

balance sheet of financial institution has been multi- reduce their portfolio holdings. Note that this in turn

plied by 5 relative to world GDP. Similarly, a 2007 increases the supply of the asset on the market, which

study finds a portfolio exposure of French banks to may further exacerbate the initial fall in the price.

US assets equal to 22 percent, to which one may want

to add a 15 percent exposure to UK assets. Assume now that the investor is internationally diver-

sified, and owns 50 percent of his portfolio in US

That the transmission of the crisis is now synchro- shares and 50 percent in euro shares. Assume the price

nised is evidenced by the synchronisation of the of those shares is initially equal to 10, and that the

responses of stock markets and real economic vari- investor owns 5 of each share. The initial value of the

ables across economic blocks in the current crisis, as is portfolio is 10 5 + 10 5 = 100, which again we

depicted in Figures 5.2 and 5.3. * *

assume is split between 60 of debt and 40 of equity.

Next, assume that the price of US shares falls to 6. The

The mechanism underlying the financial transmis- new value of the firm’s assets is 6 5 + 10 5 = 80

sion of the crisis lies in the balance sheet of interna- * *

again. The investor must again deleverage, and let us

tional investors and its effect on asset prices. Those in assume that his preferences are such that he wants to

turn affect the “financial accelerator”, which is the keep an equal proportion of each asset. If he sells n

transmission mechanism from the financial to the assets, that will be n/2 of each kind, and the resulting

real sector. value of the portfolio is 6 (5 – n/2) + 10 (5 – n/2)

* *

Financial institutions must hold

a fraction of their liabilities in the Figure 5.1

form of equity rather than debt,

generally for regulatory reasons.

Because their portfolios are val-

ued at market prices, when mar-

ket prices fall, they have trouble

matching their regulatory ratios

if they are leveraged. This is

essentially because their equity,

which is equal to the value of

their assets minus their debt, is

more sensitive to stock prices

than their total assets, because

debt, which does not fall with

stock prices, is subtracted from

total assets when computing 112

EEAG Report 2010 Chapter 5

These developments have diverg-

Figure 5.2 ing effects on the exchange rate.

At any point in time, the ex-

change rate clears the market for

foreign exchange. The demand

and supply for foreign exchange

comes from two motives. First,

exporters and importers need to

acquire foreign currency to

finance their purchases, or con-

versely get rid of the foreign cur-

rency they got in international

transactions. Second, portfolio

investors also generate a demand

and a supply for foreign currency

depending on the denominations

of the assets they want to hold in

= 8 (10 – n). Debt falls by 6 n/2 + 10 n/2 = 8 n, their portfolio. In particular, their demand for, say,

* * * *

and equity is unchanged. These are the same compu- dollar denominated assets will be greater, the greater

tations as before, and thus n = 4. The investor dumps the rate of return on those assets compared to the rest

2 US shares and 2 euro shares on the international of the world is. That rate of return is in turn more

market. We now have a fall in euro stock prices, which favourable when either the rate of return of US assets,

deteriorates the balance sheet of investors who hold expressed in dollars, goes up, or the dollar is expected

those assets. This triggers another wave of deleverag- to appreciate. Nowadays, the second motive for for-

ing, which alters both euro and US markets if those eign exchange transactions plays a far greater role

investors also hold US assets. The spiral continues than the first, because the volume of FOREX trade

until a new equilibrium is found. induced by international capital movements dwarfs

the one associated with international trade in goods

and services.

3. The impact of the crisis on the euro exchange rate Let us now tackle the presumed impact of each aspect

We now discuss how the economic crisis may affect the of the US crisis on the euro/dollar exchange rate.

likely evolution of the exchange rate of the euro vis-à- Consider first the fall in aggregate demand. Let us dis-

vis the dollar. Potentially, the crisis can have a large cuss its impact on the exchange rate by first assuming

effect on the euro area through massive movements in that inflation in the US relative to the rest of the

nominal exchange rates. We start by discussing the world, as well as rates of returns on assets, are

mechanisms by which the recession and the response unchanged. A fall in aggregate demand implies a per-

of policy makers may affect the

exchange rate of the euro vis-à-vis Figure 5.3

other currencies.

In the United States, the crisis is

characterised by

• A severe contraction in aggre-

gate demand

• A massive policy response, in

the form of

– Large scale stimulus pack-

ages that may lift the budget

deficit to some 13 percent of

GDP in year 2009.

– Aggressive cuts in interest

rates by the Fed to a level

close to zero. 113 EEAG Report 2010

Chapter 5 manent improvement in the US’s Figure 5.4

net foreign asset position, as the

US imports less goods from the

rest of the world. Furthermore, it

should be matched by a once-

and-for-all adjustment in the

exchange rate, because any future

movements in the exchange rate

beyond the impact effect of the

shift would be arbitraged away by

financial markets. If the real

exchange rate were to depreciate,

the US trade deficit would im-

prove by even more. If markets

were expecting the US foreign

asset position to be balanced in

the long run before the fall in Thus we see that in the current crisis there are forces

aggregate demand, they would now expect it to be for appreciation along with forces for depreciation.

ever-improving. This is clearly not an equilibrium What has actually happened? Figure 5.4 depicts the

since the US consumer would eventually want to con- evolution of the euro/dollar exchange rate since the

sume part of that added wealth. Therefore, the fall in beginning of 2007. We observe three phases:

aggregate demand has to be matched by an apprecia-

tion of the real exchange rate, which reflects the asso- Initially, the dollar gradually depreciates to end up

ciated lower demand for US goods. below 1.5 dollars per euro. This coincides with the

“pre-crisis” period, during which investors started to

However, this argument holds everything else equal, be increasingly worried about global imbalances and

i.e. assuming that there is no reaction by monetary subprime mortgages.

authorities to the slump in demand and more general-

ly that the return to US dollar denominated assets • Then we have another period of appreciation,

does not fall. In practice we rather expect the latter to which ends in November 2008.

fall, for example because monetary authorities will • Finally, a new depreciation period started in

reduce interest rates to counteract the recession. This January 2009.

would then trigger a shift out of US assets and a

depreciation of the dollar – this effect is likely to Such evolutions are notoriously difficult to interpret,

dwarf the effect of an expected improvement in the in light of the complex forces outlined above. The

net foreign asset position. expectations of market participants play a key role in

shaping them. One important question for the euro

Let us now consider the effect of a fiscal expansion. area is: Can the current trend of depreciation contin-

A fiscal expansion, everything else equal, needs to be ue? If so, this would be a mixed blessing, as it would

financed; the rate of return on dollar denominated trigger a substantial appreciation of the euro and a

debt increases, which attracts foreign capital and trig- loss in competitiveness, and therefore make the reces-

gers an appreciation of the exchange rate. This is sion more persistent in Europe. As documented

what was observed during the 1980s with the so- below, it is likely that competitiveness problems are

called “Reagan” deficits. Again, this is everything else building up in some euro countries, such as Spain,

equal. If markets expect that the additional debt will France and potentially Italy.

be financed by inflation, thus expecting a low return

on US assets, deficits may well trigger capital flight There are some arguments against such a scenario. In

and a depreciation of the dollar. Finally, a monetary particular, euro-denominated assets are not overall

expansion, by lowering nominal interest rates, makes more attractive than US-denominated ones. His-

it more profitable to exit dollar-denominated assets torically, monetary conditions in the euro area typi-

in order to invest one’s money elsewhere where cally have been more restrictive than in the United

returns are higher, and this leads to a depreciation of States. As we have seen in Chapter 1, though, the pre-

the dollar. 114

EEAG Report 2010 Chapter 5

sent difference in interest rates is smaller than ever. Consequently, their net debt falls. This tends to

improve its net foreign asset position which, as we

Thus compared to the recent past there is no particu- have discussed above, is a force for appreciation; thus,

lar reason for a portfolio shift in favour of euro assets. we have an additional mechanism for correcting an

Nor is there any clear evidence that growth prospects appreciation of the dollar. In particular, this rules an

are better in Europe than in the United States: the cri- insolvency/depreciation spiral out by which as the

sis is at least as severe as in the US, the aging problem dollar depreciates US residents would increasingly be

is worse, and, despite the rhetoric of the Lisbon unable to meet their (foreign-denominated) debt,

Agenda, there are no expectations of broad reforms which would trigger a run away from US assets and

that might unleash some unexploited growth poten- 2 Such a mecha-

further depreciation of the currency.

tial – if anything, the crisis has postponed such nism has been important in previous episodes in

reforms. Finally, while budget deficits in the US are emerging economies, for example during the Asian

substantially higher than in the euro area (See crisis or the Argentinean crisis of the 1990s. As we

Chapter 1 and Chapter 4), the initial situation in the discuss below, it is actually more likely to come into

United States is more favourable because its initial play at the periphery of the euro area than in the

level of public debt is lower. Thus even though the cri- United States.

sis has made the US less attractive than before, it does

not seem to justify a massive portfolio shift in favour Thus there are compelling reasons to rule out both, a

of euro assets. This is further compounded by two continuation of the appreciation of the euro beyond

stabilising forces. First, at some point markets seem to 1.5–1.6 dollars/euro and a sudden portfolio shift

internalise the effect of exchange rate misalignments away from US dollar-denominated assets and in

on competitiveness and future trade deficits. For favour of euro-denominated assets. One scenario that

example, in previous EEAG reports we have docu- cannot be ruled out, though, is a sharp rise in expec-

mented that the US dollar/euro exchange rate seems tations of inflation in the US, if say markets antici-

to remain between two boundaries: an upper bound- pate persistently high budget deficits and it appears

ary where a German basket of goods is as expensive that inflation will be the most likely form of taxation

in the US as in Germany (and further euro apprecia- that will be used to reduce the burden of debt. Such

tion would make it cheaper in the US), and a lower a realisation by markets could trigger a sharp drop in

boundary where the converse is true, i.e. a US basket the dollar. In the long run this would not be associat-

of goods is as expensive in Germany as in the US. ed with competitiveness problems in the euro area:

Between these two boundaries, a sort of “no-envy” On average, the rate of depreciation merely offsets

situation holds, with the German basket being cheap- the inflation differential between the two zones.

er in Germany than in the US, while the US basket of However, upon impact the drop may indeed cause

goods is cheaper in the US. While we lack a firm the- competitiveness problems, as the fall in the dollar

ory that would account for such an empirical regular- reflects expected increases in the US price level that

ity, it is possible that these two critical points capture have not yet materialised. Through imports, such a

somewhat the level of bilateral rates beyond which fall may exert deflationary pressure in the euro area

massive arbitrage in goods markets would take place, which would have contractionary effects through

i.e. beyond which trade imbalances would clearly be higher real interest rates, while making it more likely

unsustainable. If so, then intertemporal arbitrage by that a liquidity trap arises. In the longer run, the ECB

speculators would prevent the boundaries from being will be faced with the dilemma between aligning itself

trespassed. Such an interpretation is consistent with to US monetary policy, which amounts to importing

the halt of the preceding phase of appreciation, when US inflation, and fighting an endemic appreciation of

the euro started falling again after hitting 1.55 dollars the euro.

per euro – which is around the level where in the US

the German basket becomes as cheap as the US one. This inflationary scenario is plausible given the mas-

We should then expect the current phase of apprecia- sive liquidity that has been injected in the economy by

tion to stop at around a rate of 1.5. Another stabilis- the Fed and the poor quality of many of the assets

ing mechanism is the well-known valuation effect, that it has acquired in exchange for that. However, at

which was already discussed in our 2008 report. present markets do not anticipate that it will prevail.

Because the US tends to borrow in its own currency, If it were to prevail, the nominal yield on 5 year US

while it is holding assets (such as equities) that are

real, a depreciation of the dollar reduces the value of 2 Such a run would quite often be associated with a bank run, and

the debt of US citizens relative to their assets. therefore one would have a “twin crisis”, as discussed below.

115 EEAG Report 2010

Chapter 5 treasury bonds would be substantially higher than for stantially increase the probability of an attack on

short term maturities, and as we have seen in Chapter the currency. Thus, Ireland, which had a large expo-

1 this is not observed. This may mean either that mar- sure to toxic US assets, was spared the problems

kets do not see an end to the recovery (which would be experienced by Iceland.

necessary to ignite inflation), or that they are confi-

dent in the Fed’s ability to fine tune the rate of infla- Does that mean that the euro is an unambiguous

tion when the recovery comes, by gradually reducing blessing? The answer is “no”, and we have three main

the monetary base. reasons for concern:

• The crisis in some non-members has a severe

impact on some members through the depreciation

4. Is the euro area a safe haven? of those non-members’ currencies.

There has been much debate regarding whether the • The crisis in accession countries generates impor-

euro area has acted as a successful shelter against tant policy dilemmas that may weaken the euro.

the financial crisis. This argument is motivated by • The fiscal and macroeconomic position of at least

the experience of Iceland, where the failure of large one peripheral member country is straining the

banks has led to government insolvency (along with monetary union.

a collapse of the value of the currency). The role of

a single currency in preventing those outcomes has We discuss these three issues successively.

to be clarified. Clearly, participating in a currency

union does not reduce the likelihood of a bank run, 4.1 Depreciation of contiguous currencies

insofar as such a run is motivated by the fact that Essentially, while the euro is overall a blessing in that

the bank is not able to pay back all depositors it protects some countries against a financial crisis,

should it occur, given the illiquidity of the asset side as discussed above, its drawback that the European

of its balance sheet. Thus, a priori, financial crises Monetary Union (EMU) is not an optimal currency

bear little relation to the exchange rate regime. On area may be particularly salient under a crisis. This is

the other hand, macroeconomists have identified because non member countries that trade heavily

“twin crises”, i.e. episodes where a banking crisis with some member countries may experience a large

occurs simultaneously with a balance of payments depreciation of their exchange rate, which will

crisis (see Kaminsky and Reinhart (1999), induce a strong economic contraction in the member

Dornbusch et al. (1995), Sachs et al. (1996)). While country. In the case of the EMU, two member coun-

this literature is still burgeoning, there are reasons tries are in such a situation: Ireland, which trades

to believe that there are complementarities between heavily with the United Kingdom, and Finland,

the two types of crises (which explains that they which trades heavily with Sweden. Figure 5.5 docu-

both happen at the same time in many cases). More ments the very large depreciation of the British

specifically, if the liabilities of financial institutions

are denominated in foreign cur-

rency, expectations of a sudden Figure 5.5

drop in the exchange rate reduce

the solvency of those institu-

tions, which makes it more like-

ly that a run may take place. In

that respect, it is reasonable to

believe that for a number of

small countries in the euro area,

the euro has made a run on the

financial sector less likely, since

it is very unlikely that a massive

fall in the euro would have taken

place, contrary to the case of a

small country with its own cur-

rency where domestic macro-

economic problems can sub- 116

EEAG Report 2010 Chapter 5

the boom was associated with a current account

pound during the crisis as well as the milder depreci- deficit of 25 percent of GDP and the foreign debt

ation of the Swedish krona. reached 140 percent of GDP.

These factors certainly play some role in the fact that To foster early accession to the EMU, Latvia has

Finland and Ireland are the two euro area countries adopted a narrow peg to the euro. This leads to the

where the recession has been most severe, with an problem that markets may force a devaluation

estimated contraction of 8.8 percent and 7.5 percent

3 because they face a high foreign debt, a poor per-

Thus, while membership of

respectively for 2007. formance of the economy and a probable overvalu-

the euro area is favourable for financial stability by ation of the currency. In 2009 the economy con-

shutting down channels for twin crises, it may make tracted at an annual rate of 18 percent. Among the

the actual contractionary impact of the crisis more results is a sharp rise in budget deficits, estimated at

severe by preventing a quick adjustment of the real 13 percent of GDP, which along with the recession

exchange rate. In contrast, non euro countries can create expectations of a devaluation. The govern-

rebalance their economies quite quickly by having a ment has received support from the IMF, but due to

sharp depreciation. The UK, for example, suffered the magnitude of the contraction it cannot meet the

from substantial trade deficits and arguably from an conditionality attached to it in terms of fiscal disci-

overvalued exchange rate; the quick depreciation of pline. All these issues are making a self-fulfilling

the pound has gone a long way toward restoring balance of payments crisis more likely, along with

equilibrium. the rapid appreciation of the real exchange rate that

was accumulated during the period of pegging to

4.2 Critical macroeconomic developments in candidate the euro.

countries Hungary has experienced similar developments, on a

Another critical issue is the effect of the crisis on milder scale: a commitment to a euro peg, strong

Eastern European countries and the timetable of growth and large external imbalances, with a preva-

those countries’ adoption of the euro. In principle, lence of foreign-currency borrowing and again the

these countries have a claim to join the euro after a risk of a twin crisis. Inflation has been less strong than

period of two years of moderate exchange rate fluc- in Latvia (Table 5.3) though, and is probably compat-

tuations, and no devaluation (the so-called ERM-II ible with the exchange rate peg, given the necessary

arrangement). However, a number of these emerg- appreciation of non-traded goods over time vis-à-vis

ing countries are particularly vulnerable to the cri- those prevailing in the euro area. By contrast, infla-

sis. This is not so much due to their exposure to tion in Latvia has been incompatible with the ex-

toxic assets as to the sharp rise in their foreign-cur- change rate peg and is now having a sharp contrac-

rency denominated liabilities during the period of tionary effect through the loss of competitiveness. As

strong growth and imbalances that preceded the in other countries, this tends to correct the trade

crisis. This generates the risk of self-fulfilling bal- deficit because imports massively fall; nevertheless,

ance of payment crises, as investors anticipate that such rebalancing of external trade only comes at the

a collapse in the currency would make a lot of cost of an internal recession and does not eliminate

debtors insolvent, and get rid of their domestic the need for a real depreciation.

assets. The world recession clearly does not help as

these small countries rely more on exports and are

therefore more vulnerable to a slump in world Tablee 5.1

1

aggregate demand. Tradee bal ance/G DP

P Hun gar y an d L atvi a

Hungary Latvia

To illustrate this, let us take two examples: Latvia 2006Q04 – 7.4 – 29.0

2007Q01 – 6.5 – 22.4

and Hungary (see Tables 5.1, 5.2 and 5.3). The rapid 2007Q02 – 7.6 – 23.0

boom in Latvia prior to the crisis was fuelled by 2007Q03 – 6.4 – 25.2

strong capital inflows and international investor 2007Q04 – 6.3 – 19.1

euphoria. As a result, a large stock of foreign debt 2008Q01 – 5.3 – 15.2

2008Q02 – 5.3 – 15.3

was accumulated and up to 90 percent of debt was 2008Q03 – 8.5 – 13.0

denominated in foreign currency (See Stokes (2009)); 2008Q04 – 9.4 – 8.3

2009Q01 – 1.3 1.2

Source: Eurostat.

3 See Table 5. 117 EEAG Report 2010

Chapter 5 In Latvia, monetary authorities have been able so far

Tablee 5.2

2 to defend the currency peg at a remarkably low cost in

Reall G DP

P gro wth,, L atvi a an d H un gary

y terms of interest rates. This is especially surprising

Latvia Hungary given that, as shown on Figure 5.7, money market

2000 6.9 4.9 rates experience large swings that reflect the sensitivi-

2001 8 4.1

2002 6.5 4.4 ty of market expectations to news regarding the pos-

2003 7.2 4.3 sibility of a devaluation or a balance of payment cri-

2004 8.7 4.9 sis (such a disconnect between bank rates and market

2005 10.6 3.5 rates is not observed in Hungary). It is possible that

2006 12.2 4

2007 10 1 covert intervention by the ECB to defend the Latvian

2008 – 4.6 0.6 currency (lats) explains such a pattern.

2009 – 18 – 6.5

Source: Eurostat. How do the macroeconomic problems of peripheral

accession countries affect the euro area? First, they

create pressure for early entry in the euro area. The

T ab lee 5 . 3 point, again, is that the risk of a self-fulfilling attack

Inflatio n , L at v iaa an d H u n g ary

y (%)) would have been nil if those countries had been

Latvia Hungary members of the euro area. A “surprise” adoption of

2007 17 5.5

2008 11 4.5 the euro by the CEECs (as advocated by some com-

Source: Eurostat. 4

mentators ) would kill any prospect of a balance

payment crisis in these countries. The problem is

that, presumably, with a critical mass of vulnerable

Hungary, has maintained high interest rates in order countries in the euro area, the euro itself would

to defend its currency. For example, throughout 2008 eventually become vulnerable. We have seen in the

interest rates in Hungary soared from 7 percent to case of Ireland that euro membership did not pre-

more than 11 percent and they remained above clude a sharp contraction of GDP, and such a con-

9.5 percent throughout most of 2009 (Figure 5.6). The traction is typically associated with large budget

policy dilemma is clear: either the central bank lowers deficits. Having the troubled CEECs join the euro

interest rates and runs the risk of a depreciation and would further weaken the overall budget outlook of

a crisis induced by the insolvency of borrowers in for- the euro area, thus raising pressures for loose mon-

eign currency, or it maintains high nominal and real etary policy while fixing another nail in the coffin of

interest rates and fuels a recession driven by weak the EU’s Growth and Stability Pact. This point is

aggregate demand. So far it has chosen the latter especially relevant in light of the issues faced by

course and the result is a sharp contraction in eco- some peripheral member countries, as discussed in

nomic activity. Since the peak of the crisis, though, the next subsection.

tensions seem to have eased and the central bank has

managed to reduce its interest rate to 5.5 percent. More fundamentally, given the

constraints associated with euro

Figure 5.6 membership, it is unwise that a

country joins the euro area at a

time of crisis, because prices are

more likely to be incorrect. In the

case of Latvia, for example, we

may assume that entry in the euro

area at current exchange rates

will lead to overvaluation and

therefore be associated with a

prolonged slump in that country.

On the other hand, entry in the

euro area immediately after a

4 Marcin Piatkowski and Krzysztof

Rybinski, “ Let us roll out the euro to the

whole Union”, June 11,

Financial Times,

2009.

118

EEAG Report 2010 Chapter 5

crisis, the safe haven hypothesis is

Figure 5.7 currently being tested by markets

for the most highly indebted

countries – especially Greece.

This is apparent when one looks

at the yield on government

bonds of the euro area countries.

Given that these are denominat-

ed in euros, the euro value of a

sovereign bond’s coupon is unaf-

fected by the domestic inflation

rate. Therefore, a higher yield on

such a bond can only reflect the

market’s expectation of outright

default or perhaps an exit from

the euro area and a conversion

of the bonds in the (reintro-

devaluation may lead to under-valuation, especially if duced) domestic currency. At present such a move is

such devaluation is the by-product of a balance-of- 5

not on the political agenda of any member country,

payments crisis. and, in Europe, outright default is only observed in

the context of war or revolution. If in addition to

The second issue is that the problems in Eastern that one expects that the devaluation of domestic

europe may lead to a bailout from Western Europe. debt cannot be forced by a balance of payment cri-

This may happen both because Western banks are sis, due to the protective effect of euro membership,

exposed to substantial credit risk in the East, and we would think that the yields on euro sovereign

because the West may want to inject money in those bonds should be very similar across member coun-

economies in order to stabilise them, in particular so tries. Yet, not only are the spreads substantial, but

as to avoid a postponement of their joining the single they widened considerably during the crisis. Going

currency. Indeed, rescue packages were implemented back to Figure 3.1, which is reproduced in Figure 5.8

during the first half of 2009 under the auspices of the for convenience, we see that for the most ex-

IMF. Such a bailout will make the overall fiscal situa- posed countries, Ireland and Greece, they exceeded

tion of euro area countries more fragile. Again, there 250 basis points at the peak. To put this in perspec-

is a limit to the extent to which the problems of small

countries can be solved by mutualising their liabilities

and diluting them in a larger, more stable area. 5 And the consensus view among economists is that it cannot hap-

Beyond that limit, the stability of the whole area may pen. See Eichengreen (2007).

be in danger. If one compounds

the scenario of an Eastern bail- Figure 5.8

out with the poor situation of a

number of peripheral member

states and the rapidly rising pub-

lic debt in core countries such as

Spain, Germany and France, it is

not far-fetched to argue that such

a limit may be surpassed.

4.3 Fiscal imbalances in

peripheral member states

The third challenge faced by the

euro area is that while it is true

that member countries have

avoided a balance-of-payment 119 EEAG Report 2010

Chapter 5 The other issue regarding Greece is that given the

tive, consider that this can be interpreted as a yearly

probability of total default on the debt. Over a ten- political climate, it is unclear whether a policy of fis-

year period, and assuming the baseline country cal consolidation or wage moderation will be politi-

Germany never defaults, this means that for Ireland cally feasible. Reforms are often met with violent

or Greece the market evaluates such an event as hav- protests and populist electoral platforms tend to gain

10

ing a probability of 1 – (1 – 0.025) = 22.4 percent. the upper hand, as in the 2009 election when the

This is huge. While the tensions have eased some- Socialist party won with a program of wage increases

what, the spreads remain considerable. If neither and greater public spending. It is possible that a radi-

default nor devaluation are possible options, a spec- calisation of Greek politics might lead to new options

ulator could make infinite profits by arbitraging such as exiting the euro being considered, and that

those spreads away. Therefore, there must be some such a possibility is already reflected in the behaviour

reason why default or exiting the euro are more like- of markets.

ly outcomes than we thought. The lesson to be drawn from this discussion is that

To see this, let us take the example of Greece. It while euro membership provides an insurance against

entered the crisis with a ratio of public debt over currency and financial crises, its real effects on periph-

GDP equal to 100 percent, after more than a decade eral countries may lead to such large imbalances that

of very large trade deficits – this latter feature prob- they may end up in a crisis despite the safe-haven

ably reflecting an entry into the euro area at an over- effect.

valued exchange rate. In the absence of euro mem-

bership Greece would probably have experienced a One may interpret recent proposals to issue so-called

balance-of-payments crisis and massive currency “euro bonds” backed by future tax receipts of the

depreciation, as both exchange rate overvaluation European Union as a step toward mutualising claims

and high public debt would have created expecta- between member countries. Given the size of the EU

tions of loose monetary policy in the future. But we budget, additional resources to pay for such bonds

can see from the evolution of spreads and the more must inevitably be the outcome of a strategic game

recent downgrading of Greece’s sovereign debt by between countries in which each member tries to shift

rating agencies that the safe haven mechanism the burden of taxation to the others. Typically, we

works at best imperfectly. Public debt is forecast to expect such a game to benefit the more highly indebt-

hit the 135 percent mark in 2011 (recent revisions of ed countries. Thus, the euro bonds would create an

the deficit put it at some 12 percent of GDP for implicit commitment of the more virtuous govern-

2009). Furthermore, the economy is harmed by its ments to bail-out the least virtuous ones in the future,

poor export competitiveness and the ability of the and at the same time generate perverse incentives for

government to effectively increase tax receipts all countries to increase their debt so as to benefit

remains to be proved. As a result, a default triggered from such a bail-out. This mutualisation indeed par-

by markets’ expectations of the government being tially helps the most indebted countries, but only by

unable to repay its obligations in the future cannot diluting their fiscal insolvency in a wider geographical

be ruled out as a scenario. In such a case, though, area, while it weakens fiscal discipline in the monetary

many analysts would typically expect a bail-out to union. The end result would be an overall weakening

occur by major euro area countries, perhaps with of the euro and an increase in the risk premium over

6

the help of the IMF. But contagion may well spread euro-denominated assets.

to bigger economies with a debt overhang, such as

Belgium, Italy, or even France (as the latter is rapid-

ly headed toward the 100 percent debt/GDP ratio 5. How have member economies reacted to the crisis?

mark). In such a case, bail-out would clearly be

impossible and some form of default would have to We now discuss how the crisis has affected the various

occur. It must be the case that markets do not rule countries participating in the EMU. In dealing with

out an incomplete bail-out and/or a contagion sce- the crisis, the euro area faces a number of specific

nario that would make a complete bailout impossi- challenges due to its heterogeneity and the decen-

ble; otherwise we would not observe such high tralised character of budget decisions. The more the

spreads on Greek public debt. euro area countries are similar in terms of shocks and

policies, the lower are the costs of having the single

6 See, for example,“Greece: A New Deal?“, BNP Paribas note, currency. Thus it is important to understand the

15 Dec 2009, http://www.roubini.com/citation/401470/4/0/11109. 120

EEAG Report 2010 Chapter 5

sources of heterogeneity within the euro area and how Tablee 5.4

4

they affect the response to the crisis of each member Sh aree off U S equit y h eldd

country as well as the scope for a coordinated policy byy eur o ar eaa i nv estorss

response. We now turn to these issues. euro area 45.8

Austria 48.2

Belgium 44.8

5.1 Differences in openness France 42.5

Germany 45.5

As discussed above, one important transmission chan- Italy 44.8

nel is international trade. It is known that different Luxembourg 42.7

Netherlands 54.6

countries in the euro area have different trade intensi- Finland 31.8

ties and therefore different sensitivities to a fall in Greece 40.0

world aggregate demand. Figure 5.9 illustrates this by Ireland 46.8

plotting the fall in the share of exports over GDP dur- Portugal 41.3

Spain 32.4

ing the crisis (i.e. between 2007 and 2009) versus the Source: Lane and Milesi-Ferretti (2005).

initial level of openness (measured as imports plus

exports over GDP): bigger exporters have experienced

a larger external shock. 5.2 Differences in financial exposure

These differences imply differences in the preferred Second, countries may differ in their sensitivity to

policy response to the crisis. Everything else equal, the financial transmission channel. As the above

argument has shown, that channel is stronger, the

• a stronger external shock generates a greater larger the fraction of an investor’s portfolio that is

demand for stimulus coming from the policy invested in US assets. That fraction clearly differs

authorities, but across countries, but a look at the data suggests this

• greater openness means that a larger fraction of is not a big source of heterogeneity. Table 5.4, taken

the stimulus is going to “leak” through imports, so from Lane and Milesi-Ferretti (2005), gives us the

that the net effect of the stimulus is smaller. equity share of euro area countries in the US as of

2005. We see that the exposure rate of the larger

Since the more open economies had the bigger shock, countries is around 45 percent, with the exception of

these two effects go in opposite directions and it is Spain which seems more financially insulated from

therefore not clear what their net response should be. the crisis, with only 32 percent of its equity portfolio

On the other hand, the more open economies are the invested in US assets.

ones that are likely to benefit most from a global coor-

dinated stimulus, whereby the leak-out of activity Therefore, with the exception of Spain, the rate of

associated with imports is compensated by a leak-in exposure to US assets is not a big source of hetero-

associated with exports. geneity. 5.3 Different initial conditions

Figure 5.9 Euro area countries are subjected

to different initial conditions at

the time they enter the crisis.

These initial conditions will in

turn have an effect on the eco-

nomic consequences of the crisis

in a given country, on its margin

of manoeuvre for counter-cycli-

cal policy measures and on the

nature of the policy response

that it prefers. Two important

aspects, in particular, are the evo-

lution of the country’s competi-

tiveness and its trade balance,

121 EEAG Report 2010


PAGINE

140

PESO

16.65 MB

AUTORE

Atreyu

PUBBLICATO

+1 anno fa


DESCRIZIONE DISPENSA

Dispensa al corso di Economia dell'integrazione europea della Prof.ssa Lilia Cavallari. Trattasi del rapporto 2010 dello European Economic Advisory Group, all'interno del quale sono affrontati i seguenti argomenti: situazione economica europea nel 2010, conseguenze della crisi finanziaria sul mercato della finanza, crescita del debito globale, conseguenze della crisi statunitense dei debiti, rischi per l'area Euro.


DETTAGLI
Corso di laurea: Corso di laurea in consulente esperto per i processi di pace, cooperazione sviluppo
SSD:
A.A.: 2011-2012

I contenuti di questa pagina costituiscono rielaborazioni personali del Publisher Atreyu di informazioni apprese con la frequenza delle lezioni di Economia aziendale e studio autonomo di eventuali libri di riferimento in preparazione dell'esame finale o della tesi. Non devono intendersi come materiale ufficiale dell'università Roma Tre - Uniroma3 o del prof Cavallari Lilia.

Acquista con carta o conto PayPal

Scarica il file tutte le volte che vuoi

Paga con un conto PayPal per usufruire della garanzia Soddisfatto o rimborsato

Recensioni
Ti è piaciuto questo appunto? Valutalo!

Altri appunti di Economia aziendale

Turning the page? Eu fiscal consolidation
Dispensa
EEAG report 2011
Dispensa
BCE - Rapporto annuale 2009
Dispensa
Role of independent fiscal policy institutions
Dispensa