Working Paper: CEPR ID: DP16883
Authors: Calogero Brancatelli; Roman Inderst
Abstract: This paper revisits the effects of income changes on consumption of private households by focusing on a commonly disregarded and yet sizeable component of household expenditures: consumption of food and non-food consumer packaged goods. We exploit a new data source from the Netherlands that combines on the level of individual households administrative data from tax records with household scanner data, thus minimizing measurement error for both expenditures and the key explanatory variable, household disposable income. Even after controlling for differences in needs and for consumption volume, we document significant variation in expenditures and thereby reveal substantial scope for potential savings. Still, even though the Netherlands experienced a recession and a subsequent recovery in the analysed period from 2011 to 2018, we find only an economically small relationship with income, which is also not higher for households with low income or low liquidity. Despite remaining small in magnitude, we document inter alia a much higher coefficient for single households. We can exclude various potentially confounding effects as we show that retailers practice national pricing and as we control for sample composition and potential substitution between in-house and out-of-house consumption.
Keywords: income effects; consumer-packaged goods; administrative data
JEL Codes: D12; E21; E32; M30
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Disposable household income (D19) | CPG expenditures (H51) |
| 10% increase in disposable household income (D19) | 0.17 to 0.18% increase in CPG consumption (D19) |
| Household composition (D10) | CPG expenditures (H51) |
| Financial wealth (G51) | CPG expenditures (H51) |
| Nutritional intake (I12) | CPG expenditures (H51) |
| Local economic conditions (R11) | CPG expenditures (H51) |