branches of a central ministry on the basis of relatively clear and
unambiguous rules enforced through a single bureaucracy, suffer
from a 35 per cent error rate. In the case of the Child Support Agency,
its first Annual Report referred to a study by its Chief Child Support
Officer who found, of 1,380 assessments checked, only 25 per cent
were judged correct, 39 per cent were found to be incorrect, whilst
in 35 per cent of cases insufficient information was recorded to tell
if the assessment was right or wrong. When policies are imple-
mented through a series of agencies, each of which expects to have
some influence on the nature and interpretation of the policy, then
clearly ‘perfect implementation’ becomes still less likely. Inter-
organisational bargaining will doubtless affect the outcomes of
policies, and with different agencies in different parts of the country
considerably different outcomes may result (see Figure 8.1).
Pressman and Wildavsky (1973) in an American study graphically
entitled Implementation: How Great Expectations in Washington
Are Dashed in Oakland or Why It’s Amazing that Federal Programs
Work at Alldemonstrates that if a series of administrative agree-
ments or clearance stages are necessary for implementation, even
with 99 per cent of agreement at each ‘clearance’, overall probability
of perfect implementation falls below 50 per cent after 68 clearances.
Is ‘perfect implementation’ always desirable? Local conditions may
differ radically from those central policy makers had in mind in
formulating their response to ‘the’ problem. Barrett and Fudge (1981)
attack the traditional British ‘top/down’ approach to public problem
solving, arguing that local communities can deploy scarce resources
much more effectively to meet their real need rather than the
centrally perceived ‘problem’. Lindblom (1959), as we have seen,
defended incrementalism as a policy-making procedure in cases
where it is difficult to define a clear consensus on policy goals and
circumstances are rapidly changing – as is the case with much public
policy. If the central policy is a radical one then the analysis of
Bachrach and Baratz (1970) referred to earlier may well help to
explain its non-implementation. Equally any central government
may find local areas (particularly those controlled by a different
political party) will stonewall on the implementation of economic and
fiscal policies with a severe local economic impact.
In some cases it may even be the case that policies are not even
intended to be implemented. Edelman’s (1977) study of political226 POLICIES