How to overcome consumption measurement errors
In this paper, Buda et al., to remedy the issues raised by Meyer and Sullivan above, propose a new dataset using naturally occurring transaction data, organized systematically according to national accounting principles, which can reproduce current official statistics on aggregate consumption with a high degree of precision. The data consists of 3 billion individual transaction-level data by 1.8 million customers of Banco Bilbao Vizcaya Argentaria (BBVA). These transaction data enable the authors to construct a large-scale, high-quality, and highly detailed consumption panel with the following advantages:
- The consumption panel aggregates, on average, to the same level of final household consumption as reported in official national accounts, thereby overcoming the severe downward bias in aggregate consumption implied by traditional consumption surveys.
- This consumption panel allows for the construction of macro-consistent distributional national accounts for consumption and reveals that consumption inequality is substantially larger than that implied by traditional consumption surveys, particularly at the top.
National Accounts in a World of Naturally Occurring Data: A Proof of Concept for Consumption
Authors: Gergely Buda, Stephen Hansen, Tomasa Rodrigo, Vasco M. Carvalho, Alvaro Ortiz, José V. Rodríguez Mora
From: Barcelona School of Economics, University College London, BBVA Research, University of Cambridge, University of Edinburgh
