CAPÍTULO IV: MARCO PROPOSITORIO
4.2 PROPUESTA DE AUDITORÍA
4.2.3 FASE III: Ejecución
4.2.3.1 Hallazgos
It should be noted that in some circles the concept known
as target efficiency would be called effectiveness.
However, as target efficiency is a recognised term in
social security, it will be retained here.
Roter (1975) discusses the need to tackle the problem of
overall benefit efficiency, defined as the extent to
which the actual implementation of a programme meets 'the
original targets conceptualized by policy-makers'. For
this purpose two elements of benefit efficiency are
identified. Target efficiency - 'the degree of
discrepancy between eligibility requirements as formally
specified in programme regulations and the original
targets or objectives which prompted the setting up of a
programme' - and operational efficiency, the outcomes of
actual programme implementation compared with the said
eligibility requirements. Roter explains the trade-off
namely that increased target efficiency generally
requires more complex operational rules and regulations
leading to an increase in errors and a decrease in
operational efficency and ultimately a breakdown in
implementation with ad hoc rules being substituted at the
local level and a greater mismatch between policy as
analysed and policy as implemented. In essence an attempt
to improve the equity of the benefit has a trade-off in
operational efficiency which may have inequitable
consequences.
For determining target efficiency Roter suggests a data
set needs to be based either on a population census or a
representative sample of the general population, and for
operational efficiency 'a broad data base in which both
eligible persons and non-recipients are included with a
known probability so that both groups can be identified'.
It may help to understand the populations which need to
be identified by considering the following diagram
Figure 5.1 The Various Populations of Analysis in Benefit Evaluation Studies Recipient Population Eligible Population f b d Target Population
Clearly for a perfectly efficient benefit these three populations would coincide so that one policy objective would be to maximise (a) as a proportion of the union of the above population s e t s . The other subsets in the diagram can be explained as follows.
b - manifestation of non-take-up and operational errors c - 'two wrongs make a right', rules do not define as
originally intended but are incorrectly applied giving desired outcome
d - rules fail to embrace target e - rules fail to exclude non-target
f - rules fail to exclude non-target but benefits fail to reach their eligible population either through
operational error or non-take-up
g - administrative error and fraudulent claims.
(a) + (c)
(a) + (b) + (c) + (d)
that is the ratio of the target population who receive
benefits to the total target population, and a measure of
inefficiency would be
(b) + (d)_______ ,
(a) + (b) + (c) + (d)
that is the proportion of those who do not receive a
benefit which was intended for them.
If these were considered to be acceptable definitions of
efficiency (a point taken up later in this sub-section),
assuming that benefit receipt can be regarded as a
success irrespective of whether it is a by-product of
misapplication of the rules and not considering the
actual level of benefit received, then this requires the
identification of two populations for any given benefit -
namely the recipients and the target.
For an existing benefit the identification of the
recipient population is fundamental to any comparison
with alternatives. To identify the target population of a
benefit requires an explicit definition of the objectives
of that benefit and this may well be difficult to
establish - the longer the scheme has been in existence
the harder it is likely to be. To estimate the recipient
population of a proposed benefit requires an
take-up and administrative errors vary between benefits.
If such a model could be developed then the basic output
from a policy evaluation exercise, which would normally
be estimated on the basis of an approximation of the
eligible population, could be used to transform the
expected distribution of benefits amongst the eligible
population into that for the recipient population.
Roter lists the factors which are generally acknowledged
as leading to increased take-up as being the promotion of
a more appealing service through advertising, the
reduction of policing methods which can cause
embarrassment to potential claimants and stigmatisation
of the service, and the reduction of the cost of the
service to potential claimants in terms of expense and
effort relative to its value.
The Supplementary Benefits Commission (1978) found strong
evidence that the proportion of sick and unemployed
receiving their title to Supplementary Benefit rose as
the length of their PIE extended. Moreover about
two-thirds of those with an unclaimed title to benefit
were living in households which had combined incomes
above the Supplementary Benefit level.
Holdaway and Partridge (1981) report that a Delphi study
involving eight bocal Office managers revealed benefit
complexity being defined as ' the number of different
distinguishable operations through which a claim must
pass'. Other factors frequently proposed as having an
influence on error rates are the number of claimants
relative to the number of staff, the quality of Local
Office staff and the amount of training which they
receive and the level of staff turnover.
If it was felt unacceptable to assume the receipt of
benefit to be a success irrespective of whether the
amount of benefit received was correct then subsets (a)
and (c) could have the additional condition imposed upon
them of ' received within x% of the correct amount of
benefit'. Those people excluded as a result would be
added into subsets (b) and (d) respectively.
There is, then, a need for quantitative research into the
interaction between benefit complexity and operational
efficiency. This would have two main objectives. Firstly
to establish a better understanding of how proposed
policies will work in practice rather than theory. If it
is the case that schemes with simpler rules are more
faithfully implemented than schemes with more complex
rules then a direct comparison of results of evaluation
studies based on eligible populations rather than
potential recipient populations will be deficient.
population.
If this model could then be extended in both directions
it may also be possible to relate increased benefit
complexity to attempts to improve equity in the
eligibility rules, and to relate operational efficiency
to administrative costs. This would then enable analyses
of benefit policy alternatives to include a consideration
of their implications for the cost of administering them.
This has not been feasible in the past and yet whilst
National Insurance benefits cost around 4% of their value
in benefit payments to administer the corresponding
figure for Supplementary Benefits is around 17%. Nobody
would argue that to decrease the cost of administering
the benefit system is undesirable - whether such savings
should be used to enable an improvement in the quality of
service or to save on running costs may be a different
matter.
This is an area where the OR modelling approach could be
particularly valuable and could be a first important step
towards relating operational problems and policy making
more closely. To this end this subject will be addressed
again in Chapter 9.
The suitability of Roter's definition of benefit
efficiency could be questioned by policy makers.
instance, focus on the sub-populations e, f, and g. The
important point to recognise for the moment, however, is
that all these various sub-populations ought to be
considered in benefit evaluation studies. If such groups
are to be enumerated then clearly the data base for the
information system needs to cover the general population
or else not even existing benefits can be properly
evaluated.