Elsevier

Economics Letters

Volume 76, Issue 3, August 2002, Pages 303-309
Economics Letters

Reconsidering the labeling effect for child benefits: evidence from a transition economy

https://doi.org/10.1016/S0165-1765(02)00086-1Get rights and content

Abstract

I consider expenditure on food, alcohol and tobacco, and clothing in households with children in Slovenia at the start of its transition to a market economy. I exploit the unique structure of its child benefit program in order to look for evidence that labeling a cash transfer a child benefit influences how that cash transfer is spent. Contrary to much of the recent literature on the non-fungibility of income, I do not find evidence supporting a labeling effect.

Introduction

Child benefits compensate households for some of the cost of raising a child.1 If recipient households treat child benefits as income for the child, it may benefit a child more than other types of income received by the household. This labeling effect is of considerable policy importance in the transition economies of Eastern and Central Europe where child benefits are used to mitigate the impact on children of the economic insecurity associated with the transition (Barr, 1994). If labeling income a child benefit helps children in ways that other types of income do not, then a case can be made for a universal child benefit. However, if income labeled a child benefit is treated like any other income, then a universal child benefit transfers income from poor households without children to rich households with children. A universal benefit may also imply a transfer from poor households with children to rich households with children if budget constraints force lower benefit levels. This study considers evidence of a labeling effect of a child benefit on household expenditures using data from early in the transition in Slovenia.

The Slovene case is the best available for examining evidence of a labeling effect in child benefits in a transition economy. Almost every Central and Eastern European country provides a universal child benefit that is fixed given the number and ages of children in a household. Thus, it is impossible to disentangle the effect of a child benefit on household resource allocation from demographic effects. However, in the case under consideration in this paper (Edmonds, 2001, describes the Slovene child benefit program, Otroski Dodatek, in greater detail), child benefits are means-tested based on the household’s self-reported income in the previous calendar year. Thus, it is possible to examine the relationship between the child benefit and household expenditures conditional on both household demographics and the current year’s income. I consider expenditures on food, alcohol and tobacco, and clothing. I am never able to reject the hypothesis that the child benefit is treated the same as any other source of income.

Section snippets

Existing evidence on labeling effects

Under standard microeconomic demand theory a child benefit should be spent like any other source of income.2 However, a large body of empirical evidence reports relationships between income sources and the resulting

New evidence on labeling child benefits in Slovenia

I consider evidence of a labeling effect in child benefit income using the 1993 Family Budget Survey (FBS) collected by the Statistical Office of the Republic of Slovenia (1993). The FBS contacts households in February of a calendar year asking participants to collect receipts and earnings statements throughout the year and collects the data in December. From the FBS, I observe income and expenditure in 1993, and the child benefit income I observe is means-tested based on 1992 earnings. My

Conclusion

I do not find evidence in support of a labeling effect for child benefits in Slovenia. I consider whether households spend child benefits as they do other income on food, alcohol and tobacco, and clothing. For each expenditure category, I never reject the hypothesis that child benefits are treated like non-benefit income. The results of this paper depend on last year’s income not influencing consumption this year given this year’s income and various regression controls. This assumption is far

Acknowledgements

I am grateful to Darja Basin, Nada Stropnik, and Martina Trbanc for providing me with information about the child benefit program in Slovenia. For data, I appreciate the help of Tomaz Banovec, Milan Vodopivec, Erika Znidarsic, and the Statistical Office of the Republic of Slovenia. I thank Anne Case, Angus Deaton, Bo Honore, Doug Miller, Nina Pavcnik, Jon Skinner, and Nada Stropnik for helpful comments and discussions.

References (13)

There are more references available in the full text version of this article.

Cited by (20)

  • Do high-income households ‘label’ family cash benefits? Evidence on family expenditures from Australia

    2023, Social Science Research
    Citation Excerpt :

    Previous studies have investigated labelling effects in two ways, either by comparing households' marginal propensity to consume (MPC) one dollar from family cash benefits against one dollar from any other source or by examining interventions that may have encouraged labelling in a ‘treated’ vis-à-vis a comparison group. MPC studies have typically focused on children's clothing as a child-assignable good, providing mixed evidence for labelling of family cash benefits across countries (cf. Kooreman, 2000; Edmonds, 2002; Blow et al., 2012). Only one study in the MPC strand has explicitly compared lower and higher-income families, although only in terms of food expenditures, and found a larger MPC from family cash benefits for lower-income families, in the context of Germany's universal child benefit (Raschke, 2016).

  • Cash by any other name? Evidence on labeling from the UK Winter Fuel Payment

    2014, Journal of Public Economics
    Citation Excerpt :

    Kooreman finds some evidence of a labeling effect (i.e. the child benefit is spent on child-related goods); in contrast, Blow, Walker and Zhu's results suggest that child benefit is spent disproportionately on adult-related goods.2 Finally, Edmonds (2002) also looks at child benefit payments (in this case among families in Slovakia) but finds no evidence of a labeling effect. These studies use quasi-experimental difference-in-difference designs, exploiting differential changes over time in the real value of benefits for different types.

View all citing articles on Scopus
View full text