Income or Consumption: Which Better Predicts Subjective Well‐Being?

• Published date: 01 November 2019
• Author: Thomas Carver,Arthur Grimes
• Date: 01 November 2019
• DOI: http://doi.org/10.1111/roiw.12414

© 2019 Internation al Association for Re search in Inco me and Wealth

S256

INCOME OR CONSUMPTION: WHICH BETTER PREDICTS

SUBJECTI VE WELL-BEING?

by Thomas Carver

Motu Economi c and Public Policy Res earch

AND

arThur Grimes*

Motu Economi c and Public Policy Res earch and Victoria Uni versity of Wellington

The positive relationship between income and subjective well-being has been well documented.

However, work assessing the relationship of alternative material well-being metrics to subjective well-

being (SWB) is limited. Consistent with the permanent income hypothesis, we find that a consumption-

based measure out-performs (surveyed) income in predicting subjective well-being. When objective

measures of consumption are combined with self-assessments of a household’s standard of living,

income becomes insignificant altogether. We obtain our result utilizing household-level data from

Statistics New Zealand’s New Zealand General Social Survey which contains measures of income, SWB

and a measure of material well-being called the Economic Living Standard Index that combines meas-

ures of consumption flows and self-assessments of material well-being.

JEL Codes: D63, I31

Keywords: consumption, income, life satisfaction, subjective well-being

1. inTroduCTion

We examine whether surveyed income or a consumption-based measure bet-

ter predicts subjective well-being (SWB) where SWB is measured by an officially

surveyed question on how people feel about their life. In this context, SWB may

be considered as a proxy for utility. Since its inception, economics has attempted

to understand the relationship between utility, consumption and income. Unlike

Adam Smith (1776), who focused on consumption as a goal or John Stuart Mill

(1863) who focused on happiness (utility) as a goal, twentieth century policy-mak-

ers tended to focus more on increasing measures of national production, such

as Gross Domestic Product (GDP) or national income measures such as Gross

National Income (Jaszi, 1986) in setting economic policy.

Note: This work was funded through Marsden Fund Grant MEP1201 from the Royal Society of

New Zealand. We thank RSNZ for this funding assistance and Statistics New Zealand for access to the

data. We also thank two referees of this journal and participants at the IARIW-Bank of Korea confer-

ence “Beyond GDP: Past Experiences and Future Challenges on the Measurement of Economic Well-

being” (Seoul, April 2017) for helpful comments on an earlier version.

*Correspondence to: Arthur Grimes, Motu Economic and Public Policy Research; and Victoria

University of Wellington, New Zealand (arthur.grimes@motu.org.nz).

Review of Inc ome and Wealth

Series 65, Numb er S1, November 2019

DOI : 10.1111 /roi w.1241 4

bs_bs_banner

Review of Income and Wealth, Series 65, Number S1, November 2019

S257

© 2019 Internation al Association for Re search in Inco me and Wealth

Economists commonly model utility inter alia as a function of consumption

which is maximized subject to a budget constraint. The budget constraint incor-

porates current and future income, wealth and prices. One lens through which to

understand our study is that it is an evaluation of whether a direct measure of

consumption (a component of the utility function) out-performs current income

(a component of the budget constraint) in proxying utility.

The limitations of measures such as GDP as indicators of well-being have

been systemically documented (Stiglitz et al ., 2009; hereafter SSF) leading to

increased interest in holistic well-being measures. SWB is one such conceptual

framework for measuring quality of life. One set of SWB metrics is collected by

directly asking individuals to evaluate their happiness / life satisfaction, either as a

whole or in particular domains (e.g. health, work) (Boarini et al., 2006).

Developments in psychology and behavioral economics have increased confidence

in the use of such metrics (Kahneman et al., 1997). Many studies have presented

supporting evidence demonstrating that surveyed life satisfaction is a reliable and

valid measure of well-being (Layard, 2011; Helliwell et al., 2013). Other measures

of “happiness” such as positive and negative affect concentrate more on shorter

term emotions, whilst eudaimonic measures of life purpose are more relevant to

personal life choices (OECD, 2013b).1 While these well-being concepts differ from

one another, several studies show that they are positively correlated (e.g. Delhey

and Kroll, 2013). We focus on evaluative well-being (life satisfaction) as our mea-

sure of SWB.

Our approach has been shaped by the recommendations of SSF for the mea-

surement of well-being. Three of their key recommendations are: 1) to concentrate

on consumption and wealth over production; 2) to emphasize the household per-

spective rather than the individual; and 3) to utilize subjective measures of well-be-

ing. The data contained within the New Zealand General Social Survey (hereafter

NZGSS) enable us to pursue certain implications of these recommendations.

Specifically, NZGSS contains a measure called the “Economic Living Standard

Index” (hereafter ELSI), a consumption-based measure of living standards. ELSI

assesses a household’s level of consumption and, to a lesser extent, wealth via a

combination of objective and self-rated questions.

Our central research question is to ascertain which of two measures of mate-

rial well-being—household income or ELSI—better predicts subjective well-being.

The correlation between life satisfaction and income is well established (Kahneman

and Deaton, 2010). The novelty of our approach is to assess if a proxy for con-

sumption (ELSI) is more informative. There are four factors that suggest that ELSI

may prove superior: First, Friedman’s permanent income hypothesis postulates

that current consumption is determined by lifetime resources, and thus current

consumption should be a better indicator than current income of lifetime living

standards (Friedman, 1957). Second, Deaton (amongst others) has demonstrated

the veracity of self-rated measures of material well-being (Deaton, 2010 and 2016).

The ELSI indicator that we use includes subjective responses relating to income

adequacy in line with Deaton’s findings. (We test our relationships using both the

1To illustrate the concepts, a surgeon may feel unhappy because a patient dies but still believe that

their life has purpose. They may be satisfied with their life as a whole possibly because happy moments

outweigh unhappy moments across their life and also because of their high degree of life purpose.