Household Composition and Preferences: A Collective Approach to Household Consumption

• Published date: 01 September 2021
• Author: Bart Leeuwen,Rob Alessie,Jochem Bresser
• Date: 01 September 2021
• DOI: http://doi.org/10.1111/roiw.12483

© 2020 The Authors. Review of Income and Wealth published by John Wiley & Sons Ltd on behalf of

International Association for Research in Income and Wealth

591

HOUSEHOLD COMPOSITION AND PREFERENCES: A COLLECTIVE

APPROACH TO HOUSEHOLD CONSUMPTION

By Bart van Leeuwen*

University of Groningen

AND

roB aLessie and Jochem de Bresser

Tilburg University, Netspar

This paper tests whether preferences over bundles of market goods are different for single persons and

members of couples. We use a collective model which incorporates economies of scale in consumption.

Detailed individual consumption data enable us to estimate a model that allows individual preferences

for some goods to depend on household composition. The hypothesis that singles and couple members

of the same gender have the same preferences is rejected. This suggests that preferences may change

when household composition changes. We produce indifference scales for members of couples and a

refined poverty line measure for couples. Indifference scales for women and men are respectively 81

and 59 percent of their household’s expenditure. These measures are highly sensitive to the preference

equality assumption.

JEL Codes: D13, I31, I32

Keywords: collective model, economies of scale, equivalence scales, poverty analysis, welfare comparisons

1. introduction

Households consisting of two (or more) persons make expenditure decisions

in a fundamentally different way than single-person households. When people start

living together, they are faced with the concept of sharing in two distinct ways.

First, the members of the newly formed household will no longer have complete

discretion over their household’s expenditure. Instead household members share

control over the budget. Second, the household members share in the use of some

purchased goods. The multiple-person household thereby realizes a more efficient

use of goods compared to a single-person household. Since the publication of the

studies by Chiappori (1988, 1992) and Apps and Rees (1988) such ideas have been

Notes: This paper makes use of data from the LISS panel. The LISS panel data were collected by

CentERdata (Tilburg University, The Netherlands) through its MESS project funded by the Netherlands

Organization for Scientific Research. We gratefully acknowledge comments on earlier versions of this

paper from two anonymous referees, Raun van Ooijen, Marcos Rangel, Adriaan Soetevent, Arthur van

Soest, and Frederic Vermeulen, and conference participants at the 2016 SOM conference, 2016 ESPE

conference, and 2017 SEHO conference.

*Correspondence to: Bart van Leeuwen, Faculty of Economics and Business, University of

Groningen, Nettelbosje 2, Groningen, 9747 AE, The Netherlands (bart.van.leeuwen@rug.nl).

Review of Income and Wealth

Series 67, Number 3, September 2021

DOI: 10.1111/roiw.12483

This is an open access article under the terms of the Creative Commons Attribution License, which

permits use, distribution and reproduction in any medium, provided the original work is properly

cited.

bs_bs_banner

Review of Income and Wealth, Series 67, Number 3, September 2021

592

© 2020 The Authors. Review of Income and Wealth published by John Wiley & Sons Ltd on behalf of

International Association for Research in Income and Wealth

formalized in collective household decision-making models. These models can par-

tially explain why single-person and couple households allocate resources differ-

ently. An issue that has received little attention in the collective modeling literature

is whether living with others makes certain goods intrinsically more attractive. This

paper examines what role preference differences play in explaining differences in

expenditure patterns of single persons and members of childless couples. Using a

data set with individual consumption data, we test whether preferences of single

persons and couple members are equal, controlling for gender.

We use a model based on Browning etal. (2013) (BCL hereafter) that allows

preferences over the consumption of a subset of individually assignable goods and

services to depend on the composition of the household. Decisions are made in

cooperation as in Chiappori (1988). The consumption allocation is Pareto efficient.

Efficiency gains from shared consumption are modeled as economies of scale in a

household production process as in BCL.

Data used in this study come from the Longitudinal Internet Studies for the

Social Sciences (LISS). The LISS data record expenditure on an exhaustive set

of goods. We group this expenditure data into five non-assignable categories and

three individually assignable categories per household member. In our sample 36%

of total expenditure on goods other than housing or utilities is recorded at the

individual level.

The LISS data reveal two key facts about the expenditure patterns of singles

and couples that we would like to explain. First, couples allocate a smaller budget

share to non-assignable goods (e.g. housing and utilities) than singles. Prior studies

have shown that such differences can be rationalized using a model that includes a

form of economies of scale in consumption. We use the model of BCL and modify

it to account for the second fact: Couples allocate a larger budget share to leisure

expenditure than singles, but couples and singles allocate roughly the same budget

shares to other assignable goods. The basic model proposed in BCL cannot ratio-

nalize this pattern, as it imposes the assumption that singles and couple members

have the same preferences.

Our innovation is to weaken this assumption by allowing preference parame-

ters appearing only in assignable goods demand functions to differ between singles

and members of couples. For a given budget and prices, singles and couple mem-

bers allocate the budget differently among assignable goods but similarly among

non-assignable goods. We refer to the original assumption as full household com-

position independence of preferences (HCIP) and the weakened assumption as

partial HCIP. We use the model identified under partial HCIP to test the nested

full HCIP assumption, and reject it. Partial HCIP better explains allocation pat-

terns overall, and in particular the difference between singles and couples in terms

of the allocation to assignable goods.

We find that imposing full HCIP substantially biases estimates of the col-

lective model. Economies of scale are overestimated, and the bargaining power

of women in couples is underestimated. These estimates can be used to produce

such policy relevant measures as indifference scales and poverty rates. We calculate

indifference scales for couple members in the spirit of BCL. Indifference scales tell

us how much income a person would need living alone to be as well-off as he or

she would be when living as part of a couple. Based on estimates from the partial