Behavioral life cycle theory and consumption response to income changes: an empirical analysis of survey data

The standard theory about consumption behavior, the Permanent Income Hypothesis theory predicts that individuals smooth optimally consumption over time and consumption is not determined by the current level of income but rather by the overall lifetime expected income, that is the permanent income. However, empirical evidence doesn’t fully support the theory and shows excess sensitivity of consumption to current income. While different explanations were proposed, I look for evidence to explain the excess sensitivity under the framework of behavioral life cycle theory, analyzing the consumption and income data. In the first part of this dissertation I briefly review the literature about life cycle theory/permanent income hypothesis, its discrepancies and the main behavioral economics findings that can explain deviations from the theory; the behavioral life cycle model proposed by Thaler (1988) explains how these deviations might be caused by behavioral biases that affect how household consumption react to income changes; these bias include mental accounting, loss aversion and the self-control problem. In the second part, I present the survey data used, both regarding US households: one is the Consumer Expenditure Survey (CES) dataset and the other is the Survey on Consumer Expectations (SCE) dataset; in addition, I will present, the computational methods of data analysis and OLS method used to analyze these data. In the third part, I use the behavioral framework and computational techniques discussed to test the behavioral life-cycle hypothesis and to empirically analyze the US household expenditure and expectation datasets in order to study consumption response to predictable and unpredictable income changes. In conclusion, consumption data suggests that there is some evidence for excess sensitivity; This evidence is weak for income increases so that we cannot fully reject the permanent income hypothesis. However, when income decreases consumption response seems to be more sensitive; this result is in line with the loss aversion prediction such that, under uncertainty, “gain” and “losses” are treated differently. Also, behavioral life cycle theory’s mental accounting explanation of excess sensitivity is supported by data, showing that households frame current income and wealth differently while making consumption decisions.