The SRAs have assisted the government to deal with one of the most difficult requirements of the reforms - the specification of outcomes. Although many of the SRAs are too broadly drawn to qualify as outcome statements, substantial progress has been made, but the quality is uneven. Moreover, the Estimates contain numerous outcome measures and link them to both the Votes and to output classes. I have not closely reviewed the hundreds of such measures, but a general conclusion is that while there has been substantial improvement, more work needs to be done in this area.
My main concern, however, is not about measurement but about the conceptual distinction between outcomes, which are the responsibility of Ministers, and outputs, which are the responsibility of chief executives. This distinction leads to a further one: the political accountability of Ministers versus the managerial accountability of chief executives. The latter can be formalised in agreements and other contract-like arrangements; the former depends on elections and public opinion. Both, however, demand transparency with respect to what has been committed and what has been accomplished.
The Public Finance Act 1989 defines the relation of outputs to outcomes in causal terms. Outputs are "the goods or services that are produced by a department, Crown agency, Office of Parliament, or other body"; outcomes are "the impacts on, or the consequences for, the community of the outputs or activities of the Government." In other words, outputs produce outcomes. These definitions are meant to establish accountability for government performance. In my view, however, the definitions mistake the relationship of outcomes and outputs. There is no inherent causal link between the two types of measures. Some outcomes may derive from specified governmental outputs, many do not. Alternatively, producing the right outputs does not ensure that the desired outcomes will materialise.
I prefer an alternative definition of outcomes along the following lines: outcomes are measures that indicate progress, or the lack thereof, in achieving public objectives. Outcomes can be stated in absolute terms, such as the rate of low-weight births or the incidence of infant mortality, or they can be stated in terms of change, the difference in low-weight birth or infant mortality rates between one time period and another. Changes in the rate of either indicator may be due to government action or exogenous circumstances or (most likely) a combination of both.
The issue is not one of definition but of cause and effect and, derivatively, of credit and blame. In my view, a system that would hold politicians accountable for conditions they only partly control invites loose definitions and evasions of responsibility. If outcomes were reported as the causal results of the outputs purchased by Ministers, politicians would be positioning themselves to be blamed for matters that are not truly their doing. In this situation, they are likely to devise expedient escape routes; one of the most popular is to define outcomes vaguely so that progress cannot be measured.
Outcomes should be seen not as measures of impact but as indicators of direction. They should be employed more for formulating policy than for maintaining accountability. They are powerful directional signals: are things getting better or worse? Have the initiatives taken by the government moved it closer or further from achieving stated objectives? Is the government progressing toward the conditions it favours? Particular outcomes may or may not be the product of outputs, but even when they are not, the government should take notice of them, analyse their significance, seek to explain what has (or has not) happened, and develop appropriate policy responses. Even if it is not accountable, the government should take outcomes into account.
Suppose, for example, that birth weights were declining or infant mortality rising? Even if government action did not cause these developments, it should respond to the problem. One response might be to change the mix of outputs; another might be to fundamentally reexamine its objectives and policies; a third might be to thoroughly evaluate its programmes and the assumptions on which they are based. Sometimes there are no solutions, but government still cannot walk away from the problem.
The output-outcome nexus has another defect: it bifurcates government into two compartments: management and politics. It is with all the capacity and intelligence it can muster that government defines and pursues outcomes. Politicians contribute to the mix, as do managers, analysts, and others. When managers tender advice, they do not desist from commenting on outcomes, nor should they. Public policy is enriched by the continuing iteration and feedback between policy and operations, outputs and outcomes, politicians and managers.