Assessing the Distributional Impact of 'Imputed Rent' and 'Non-Cash Employee Income' in Microdata: Case studies based on EU-SILC (2004) and SOEP (2002)

Abstract

There is general agreement about the importance of integrating non-monetary income components into cash-based income measures in order to improve the comparability of distribution results across time and space. Based on micro-data from the first round of EU-SILC (EU Statistics on Income and Living Conditions) collected in 12 EU countries in 2004 and placing strong emphasis on methodological aspects of determining non-cash incomes, this paper investigates the incidence, scope and distributional impact of imputed rent and private use of company cars. Imputed rent can be calculated using a variety of approaches. Thus in the present paper EU-SILC-based results for Denmark, France and Finland - with each of these countries applying a different method - are contrasted to findings derived from the 2002 wave of the German Socio-Economic Panel Study (SOEP). This makes it possible to operationalize three different methods for the very same population, thus supporting sensitivity analyses keeping all other influential factors constant. In general, there is a tendency toward decreasing inequality and poverty once including imputed rent in the income measure. Inequality decomposition by age shows a rising income advantage from owner-occupied housing (including tenants with reduced rent payments) for the elderly, confirming the importance of owner-occupied (and reduced rent) housing as a means of old age provision. However, while we can identify some cross-national differences, we cannot clearly determine the degree to which the variation between Denmark, Finland, and France actually depends on the choice of the methodology applied to measure IR. A further complication results from the fact that for Denmark and Finland, only a gross measure of IR is given in the currently available EU-SILC data, while for France, a net measure is included. Apparently a gross measure is not appropriate for the type of welfare-oriented analysis presented here. Furthermore, we find evidence that the substantive results on inequality are sensitive to the choice of the approach in the German data, thus supporting the principal demand for an extensively harmonized measure across countries. Results for company cars show a great deal of variety in the incidence of this noncash employee income component (as low as 2 to 3% in Ireland and Norway and as much as 25% in Finland and Sweden). While this income source accounts for not more than 2% of overall compensation in any of the countries considered, there is a general finding that the additional consideration of company cars in the employee income measure yields higher degrees of wage dispersion. Concluding, the empirical assessment of EU-SILC data demonstrates the relevance of including non-cash components when comparing income and wage distribution results across Europe. At the same time, it becomes apparent that cross-national comparability is very much a matter of data availability (e.g., for a measure of net IR, this pertains to information about the relevant costs to be deducted at the micro-level) as well as the national framework (e.g., size of the private rental market as basis of the opportunity cost approach to generating IR). As such, any deviation from a generally proposed approach to capture such non-cash income effects will have to be well justified so as not to jeopardize cross-national comparability. However, given explicit cross-national variation, e.g., in the tax and transfer regimes, functional equivalents for capturing non-cash income components are being sought, and not necessarily national applications of pre-defined algorithms.