The case against performance and results measurement in public management is pretty well-known by now and has been made by many critics, including in previous issues of Optimum itself. (See for example: Carroll, 2000; Fountain, 2001; van Thiel and Leeuw, 2002). Christopher Pollitt, one of the wisest and most elegant scholars of comparative public management, has summarized the three main lines of attack against measurement under the categories of: (1) conceptual, (2) motivational and (3) technical. (Pollitt, 2000) 

The “conceptual” problems of measurement are all related to the difficulty of agreeing on what should count as “good.” Goods in public management are almost as contestable as goods in public policy, which accounts, in part, for why performance indicators can never be complete, or completely “objective,” or perhaps even stable over long periods of time. They are debatable, and so they change regularly – much too regularly – according to fashion and power. As a result, comparisons – even over time, let alone between organizations – can become very difficult, sometimes even meaningless.

The “motivational” problems of measurement have to do with “gaming” and other motivationally-based distortions that can undermine measurement systems, and regularly do. The introduction of any measurement system simultaneously introduces an element of “politics” – either “real” politics (in the public and democratic sphere) or bureaucratic politics. Depending on measurement outcomes, people are liable to lose or gain reputation, prestige, resources, performance pay, career prospects, or power itself. So the “politics of performance” lead to distortions such as: the suppression of measures exposing persistent differences in performance; relabeling or redirecting of activities or outputs to conform to measurement categories; over-emphasis on things that are easily measured and the neglect of other equally or more important activities where measurement is more difficult or even unattractive, and so on.

Measurement also encounters persistent “technical” problems including the difficulty of establishing plausible causal linkage between inputs, outputs and desired or actual outcomes; a tendency to over-emphasize process, output, and economy and efficiency measures instead of measures of outcomes, quality, or user satisfaction; the difficulty of establishing adequate times series of indicators; and a tendency for measurement systems to drive toward an increasing comprehensiveness which is ultimately self-defeating and brings abut the collapse, curtailment or abandonment of the measurement system. Measures tend to “wear out,” and this contributes to overall instability.

Having surveyed the criticisms of results measurement, Pollitt’s “‘message’ however is not that the measurement approach should be abandoned.” On the contrary, he suggests it can be helpful “to examine the limits and common problems of measurement as a way of knowing about the character of public services, but not to deny either its fundamental usefulness or the scope for its further development. It should be readily acknowledged that it is usually much harder – if not impossible – to form a reliable judgment as to the quality of public services without measurement” (Pollitt, 2000: 122, 140).