Théo Mitrakos and Platon Tinios

Statistical investigations of the extent of non-take up, and its potential gravity for the efficacy of social policy, have not been undertaken to this date, with the exception of a few passing remarks. Thus our investigation breaks new ground by attempting to study the phenomenon in Greece quantitatively.

The absence of expressed policy concern for non-take up is counterbalanced by the interest aroused by its mirror image - the collection of social benefits by those not entitled, system abuse. The latter is often mentioned as a manifestation of the grey economy, and many recent institutional and administrative developments were explicitly motivated by a desire to limit it. The first step in our approach is thus to clarify the two concepts. This is done by borrowing a simple framework from statistical sampling theory.

The Greek social welfare system, being particularistic by nature and with very limited use of means testing, limits the kind of data that can be used to test for non-take up. Nevertheless, we identify three cases where testing based on survey data can be consistently employed, employing known statistical methodologies.

The investigation proceeds by employing data from the Household Expenditure Survey and the European Community Household Panel. Use of the latter (or its successor, the Survey of Income and Living Conditions), raises the prospect of producing comparable estimates of take-up in all members of the EU-15 and (soon) for the EU-25. The ability to produce estimates for a ‘difficult’ case such as Greece increases the probability of being able to define indicators on a European-wide basis, which can be useful in the open method of coordination. European databases, apart from comparability, give rise to the possibility of following changes over time. The analysis proceeds by examining non-take up for the specific cases of family benefits and low pensions. Point estimates are derived for both by comparing predicted and actual collection of benefits. However, a note of caution is sounded, by identifying the existence of potentially sizeable system abuse. A key problem encountered is the small number of observations for which inference is possible. The analysis of confidence intervals concludes that though estimates of non-take up are not very stable, they are in all cases significantly greater than zero. The analysis suggests three key suppositions or working hypotheses.

First, a problem of non-take up exists, though its exact magnitude is uncertain. The estimates derived in this paper, even if necessarily couched in very cautious terms, must be taken to imply that the issue is sufficiently important to (a) warrant further study and (b) to be an explicit concern in benefit design and policy formulation.

Second, the widespread parallel problem of under-declaration cannot easily be disentangled from non-take up and could be disguising the latter’s true magnitude. Under-declaration itself is a manifestation of pathology in the administration of social benefits; addressing non-take up cannot happen independently of parallel measures to get to grips with the causes of under-declaration. Combating both should be in the priorities of well-run social protection systems.

Third, dealing with take-up issues should be a priority in any social protection system, which aims to achieve a particular outcome. Progress in tackling non-take up must go hand-in-hand with adapting social policy to be more open, more results oriented and ultimately more efficient in feeding into societal goals.


National Report    Back to top