ABSTRACT
The above chapter was based on the idea that information from the project analysis should enable all decision-makers, regardless of their normative views, to make a factually well-founded welfare evaluation. In practice, that may be too strong a requirement. To make a full evaluation, different decision-makers may need quite different information; at the same time, their capacity for handling additional information may be limited. If the information that would have been required for every decision-maker to make a well-founded ranking of projects is so extensive that they are, in practice, unable to understand and relate to it anyway, normative objectivity will not be very helpful. In such cases, indicators can be useful. By the term indicator, I mean a piece of information that gives a hint or indication of something, but without providing final or exact answers to the question of interest. Indicators can do no more than indicating what the answer itself may be. A number providing the precise answer to the exact question at hand is a measure rather than an indicator. Since indicators do not provide all relevant parts of the full picture, they must always be interpreted with care. Nevertheless, they can be useful as warning signals for when to make the effort of digging deeper. If you take your infant to the physician for a routine check, the physician will probably measure the baby’s weight, his body temperature, the circumference of his head, and a few other measures. If all these values are as expected, and you report no other alarming observations, the physician will not proceed to ordering blood tests or ultrasound images just to be absolutely sure that nothing is wrong. The weight, temperature, and circumference numbers are measures of weight, temperature, and circumference; viewed as a whole, they are indicators of the baby’s health. If the set of indicators is not as expected, we must look deeper. But if we do indeed fear that the baby is not well, the physician will need more specific information to find out what exactly is the matter, and how to heal it. The indicators are merely warning signals; they are neither final answers
to the question of the baby’s health, nor sufficient information to establish a diagnosis and recommend treatment. Assume now that bureaucrats in a ministry have conducted a project analysis comparing two project alternatives. They have summarized their report by recommending one project, while also pointing out what they find to be the main pro and contra arguments of the case. The politicians who are going to make the decision do not have the capacity to look into all the details, so they read only the summary, not the entire report. If the bureaucrats are thought to be competent and honest, with a good understanding of what causes political controversy, a politician may assume that everything she ought to know is indeed mentioned in the summary. Unless the summary causes her any alarm, she may assume that the decision an uncontroversial one and simply trust the bureaucrats’ recommendation. If, however, she gets some signal from the report summary or elsewhere that the case is not so uncontroversial after all, she might want to spend more effort in looking into it, for example by reading the entire report and seeking advice elsewhere. Since an indicator does not provide the entire answer, it must be interpreted with care. But what does it mean to interpret something with care? To do so, one should have a reasonably good understanding of (1) what the indicator measures, and (2) the relationship of this measure to the question one is really interested in. If one wants to indicate an infant’s health, his body temperature is interesting because anything substantially above or below 37°C would be abnormal; without this knowledge, the temperature indicator would be of no help. It is thus a substantial advantage if an indicator can be given a simple, descriptive interpretation, which can be easily be related to the primary question in an intuitively understandable way.
