@phdthesis{01JZ0FKKB552K6A14DEKSEET9P,
  abstract     = {{This dissertation advances our understanding of household financial behavior by leveraging high-frequency transaction data from BNP Paribas Fortis, one of Belgium’s largest banks. Through three interconnected chapters, it demonstrates how high-frequency transaction data, combined with robust methodological approaches, can illuminate key questions in household finance and consumption behavior.

Chapter 2 establishes the methodological foundations for working with transaction data in economic research. It develops a comprehensive framework for transforming raw banking transactions into research-ready economic data, addressing challenges from data cleaning to economic classification while ensuring privacy and representativeness. The chapter’s contribution lies in developing systematic approaches to bridge the gap between banking operations and economic concepts—going “from payment to purpose” by establishing crucial components for standardized data definitions and methodologies. This standardization effort represents an important step toward creating common practices in transaction data research, enabling consistent measurement and comparison across different datasets and studies. 

Chapter 3 tackles the “zoo” of conflicting consumption responses to income changes found in the literature by analyzing permanent, transient, and recurrent changes—both positive and negative—within a single empirical framework. Using a large panel of Belgian households constructed from transaction data, it provides the first comprehensive examination of how these different types of income changes affect consumption behavior within a consistent methodological setting. The analysis reveals that while households exhibit the strongest responses to permanent income changes, as predicted by the permanent income hypothesis, they also show significant reactions to transient changes and systematic differences in how they handle recurrent changes. These patterns challenge strict interpretations of consumption smoothing while supporting broader lifecycle frameworks. By examining multiple types of income changes simultaneously, the chapter provides robust empirical evidence for how households actually adjust their consumption. 

Chapter 4 demonstrates how leveraging the full granularity of transaction data—moving beyond monthly aggregation to examine daily payment patterns—can illuminate fundamental puzzles in consumption theory. By focusing on payment frequency, a feature only observable in high-frequency transaction data, the chapter provides strong empirical evidence on how frequently evaluating one’s financial situation shapes consumption decisions. The analysis reveals that career starters who receive wages more frequently, either weekly or biweekly, exhibit consumption responses nearly twice as large as those paid monthly. This amplification effect persists even after accounting for a wide range of variabels, and most notably individual fixed effects, suggesting that payment frequency fundamentally alters how individuals adjust their consumption to income changes. They also react asymmetrically, spending more of a bonus than they reduce their consumption when faced with an equivalent income reduction. These findings help explain several long-standing consumption puzzles that traditional theories like the permanent income hypothesis struggle to reconcile. Just as myopic loss aversion helps explain seemingly excessive risk premiums in financial markets, the frequency of income receipt appears to systematically influence consumption decisions in ways absent from standard theoretical frameworks. This parallel between investment and consumption behavior suggests that incorporating behavioral factors—particularly evaluation frequency and loss aversion—into consumption theories could provide a more parsimonious explanation for observed consumption patterns.

Together, these chapters make significant contributions at a crucial moment when the scientific community is beginning to harness the potential of transaction data. The thesis first establishes a comprehensive methodological framework with clear standards for transforming raw banking data into economically meaningful measures, providing a foundation for both current and future researchers working with transaction datasets. It then demonstrates the power of this framework through two substantive applications in household finance. By exploiting the richness of transaction data, the research reveals how both the type of income change and the frequency of wage payments—features difficult or impossible to study with traditional data sources—fundamentally shape household consumption responses. These findings not only advance our understanding of household financial behavior but also illustrate how transaction data, when properly structured and analyzed, can illuminate economic phenomena that before were difficult to study empirically.}},
  author       = {{Weytjens, Johannes}},
  language     = {{eng}},
  pages        = {{XIV, 119}},
  publisher    = {{Ghent University. Faculty of Economics and Business Administration}},
  school       = {{Ghent University}},
  title        = {{From payment to purpose : using financial transaction data for economic research on consumption dynamics}},
  year         = {{2025}},
}