Appunti del corso di Economic Assessment of Urban Transformation (Prof. Caragliu)

CBA: LIMITATIONS

- It is substantially individualistic, although it takes into account the cost and benefits accruing to all individuals. (It is based on an anthropocentric view of the world).

- Appraisal techniques are imperfect.

- Conclusions are often strongly influenced by some assumptions, such as in the case of the discount rate to be used.

- Evaluating some goods is often awkward if not impossible (e.g. what is the value of human life?).

- It is difficult to balance quantifiable costs and benefits with those costs and benefits that instead cannot be quantified.

- The notion of equity (patrimonio netto) underlying CBAs is clearly limited.

MULTI-CRITERIA ANALYSIS (MCA)

All costs and benefits linked to the following topics/parameters are difficult to define and so difficult to be quantified by means of classical CBA.

- Quality of life (e.g. aesthetically improvement of an urban area);

- Crime reductions, preservation of human rights;

- History, cultural heritage, art and

• Culture
• Health
• Welfare
• Environmental sustainability
• Public hygiene (e.g. safety on the workplace)
• Education

Oftentimes, projects and policies involve both qualitative and quantitative effects.

Let us focus on the example of building a new highway

Multi-criteria Analysis (MCA) techniques include all those evaluation methods that are not mainly based on monetary evaluations.

MCA applications often combine different evaluation criteria translating in monetary value different criteria.

MCA techniques cover a large number of approaches different from one another, contrary to what happens in the case of CBA (which is a much more coherent set of techniques). Some typologies of MCA do not help much in choosing between alternative policies, but often they are extremely useful.

MCDA (Multi-Criteria Decision Analysis) represents a vastly used type of MCA. Its aim is to rank projects from the "best" to the "worst".

Analogously to

Run a sensitivity analysis on the results to test the effect of a change in the scores or weights on the analysis' final results.

Stage 1: SET THE CONTEXT IN WHICH THE CHOICE TAKES PLACE.

1. Establish the objectives of the analysis, identify decision makers and other individuals involved in the project.
   • Key player: anyone who can contribute usefully and directly to the MCDA (rank project from the best to the worst).
   • Stakeholder: anyone affected (either positively or negatively) by the decisions made about the project to be implemented. They can also not directly take part in the analysis, but their interests should be represented by at least one key player.
2. Establish the operational aspects of the analysis.
   • How much and when will stakeholders and key players contribute to the MCDA?
   • Which type of MCDA will be used? How will it be implemented?
3. Consider the context of the appraisal.
   • What is the current situation? Which objectives do we want to achieve by implementing the policy? Which

elements can facilitate/hamper the project?

Stage 2: IDENTIFY ALTERNATIVE OPTIONS.

- Options are often stated on the basis of the acting/non acting principle. However, often alternatives can be built starting from aims defining at the previous stage. In other cases, when the number of options is a priori very high, one can decrease it by selecting those that can be deemed as being more interesting or representative, so that the analysis is easier to be carried out and the costs are minimised. 3 di 11fi fl fl

Stage 3: IDENTIFY AIMS AND CRITERIA.

1. Identify criteria to evaluate the impacts of each alternative project.
2. To evaluate each option one must think about its consequences, or, in other words, on the impacts of each option. We evaluate the impacts, not the options (think of the consequentialism principle)!
3. The effects of each policy can be quite diverse. Each typology of impact represents a criterion. Criteria represent specific and measurable objectives.

Let us assume we want to

that a table for each option is needed.

Example. Let us assume that the project we want to appraise is the construction of a high-speed train (HST). At stage 2 we identified 4 alternatives (rows on the matrix), while at stage 3 we defined criteria for our evaluation (columns on the matrix)2.

Attribute scores to each option.

Problem: how can we combine evaluations expressed in monetary terms with others indifferent metrics and produce an evaluation for each option that can be compared?

We must translate the impacts represented on the performance matrix into a common unit of measurement.

Relative scales can be established using the best result (i.e. the most desired outcome, for instance the smallest number of trees cut, to which a value of 100 is attributed) and the least desirable (for instance, the highest number of trees cut, set equal to 0) for each criterion as benchmarks.

Intermediate levels are assigned a score (clearly between 0 and 100) proportional to the two extreme values.

For instance (cfr. next slide), how can we attribute a score to OptionA? We must translate on a scale ranging from 0 to 100 what on the performance matrix is expressed on 0-700 scale (1200-500=700).- Notice that this formula is valid only for costs; benefits must be replaced their distance from the minimum value here

Stage 5: BESTOW WEIGHTS

Scales of preference for the different criteria cannot (yet) be combined and compared, since a value of 100 for a criterion may not necessarily be exactly equally important as a 100 score for a second criterion.

Bestowing a weight to each criterion shows:

• The size of the differences between the options of each criterion (e.g. whether the number of trees cut is roughly constant across different options the corresponding weight will be low);
• The relevance for each criterion, given preferences of the individuals involved in the project, (e.g. if environmental issues are socially perceived as relevant I will choose a high weight for the corresponding

1. Stage 6: COMBINE WEIGHTS AND SCORES OF EACH OPTION

so that a general evaluation can be obtained, in values that can be mutually compared.

- At this stage the appraiser must multiply each score by the weight attributed to that criterion. This process must be repeated for all criteria and options.

- Summing the values of the different criteria we obtain an overall appraisal for each option.

- More formally, let us assume that the score for option i of criterion j is defined as sij, and that the weight for each criterion is written as wj. With n criteria, the weighted score for option i can be written as:

Pay attention: the simple method represented by this formula, in which the final score is given by the product of each starting score times the weight of the corresponding criterion, can be used iff criteria are mutually independent.

E.g. let us assume in our analysis a new criterion: the number of grey owls killed. Since owls build their nests on three branches,

This criterion will not be independent from the number of trees cut. We therefore run the risk of counting twice the same impact.

Back to the previous example, weighted scores are shown in the table below:

• N. B.: THE SUM OF ALL WEIGHTS MUST BE EQUAL TO 1!
• Fase 7: EXAMINE RESULTS.

Overall scores give an idea of how much an option is preferable with respect to the others, i.e. we can build a ranking of preferences among the different alternatives.

Stage 8: Run a sensitivity analysis.

Does a change in the choice of weights relevantly impact the final results? A sensitivity analysis represents a useful device to understand whether the approximation in the definition of the impacts and the misalignment between individuals or groups of people influence the final result.

6 di 11fiSTRONGHOLDS AND WEAKNESSES OF MCA

Teacher: This year's topic is solved: The nation speed limit should be lowered to 55 miles per hour."

Homer Simpson: 55! That's ridiculous! Sure, they'll save a

few lives. But, millions will be late!

MCA identi es a ranking of preferences between alternative options making reference to• a list of goals made explicit at the onset of the analysis, for which measurement criteria havebeen defined so that a score can be attributed indicating the degree to which each goal hasbeen achieved.

A key issue in MCA is represented by the evaluation made by the appraiser about the• objectives, the criteria and the goals upon which the analysis is built.

This subjectivity represents a potential issue for MCA. However, w.r.t. CBA, MCA can contribute• a great deal of advantages to the evaluation process, whose scope is vastly widened toencompass aspects that would be difficult to capture with CBA.

Another limitation associated to MCA is represented by the fact that it cannot