Improving Consumer Finance Measures: Evidence from Linked Survey and Administrative Data

Abstract

Consumer finances are a central construct in marketing research. Yet, its measurement is fragmented. It is unclear how measurement decisions affect conclusions on the role of consumer finances in consumer behavior. To address this issue, we construct a unique data set that links self-reported survey data on consumer finances to observed household income and wealth from high-quality administrative data based on tax registries. We compare subjective and objective financial measures and assess their reliability and concurrent validity across a range of marketing-relevant outcomes. First, we show that self-reported personal income measures are generally reliable, particularly when extreme values are winsorized or when categorical response formats with sufficient granularity are used. Second, we find that reports of household income are more error-prone than personal income but can be improved by correcting for the number of income earners in the household. Third, subjective financial well-being measures, although less frequently used in marketing research, outperform objective measures in explaining variance in consumption, financial problems, and happiness. Combining measures of income and subjective financial well-being yields the highest explanatory power for consumption behavior. We develop a decision framework that guides researchers in choosing the most suitable financial measures based on data availability and research objectives.

Gepresenteerd op