Financial crisis and persistence: evidence from sticky expectations consumption growth model

ABSTRACT

We estimate the degree of stickiness in aggregate consumption growth for the U.S. considering the effects of the Great Recession. The behavior of stickiness estimate in the crisis is somewhat as the U-shaped pattern. Our findings imply that during the crisis consumers’ attentiveness to aggregate information has slightly increased, thereby reducing the persistence of aggregate consumption growth. However, the reduction in persistence is transitory. Since 1980, the U.S. faced five recessions and in most of them the degree of stickiness declined, albeit temporarily.

KEYWORDS:

• Financial crisis
• consumption growth
• sticky expectations
• persistence

JEL CLASSIFICATION:

• C5
• E2

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Other theoretical explanations of persistence in aggregate consumption growth exist in the literature, for example, see Sims (2003), Reis (2006), Deaton (1991), Carroll (1992), Ludvigson and Alexander (2001) and Pischke (1995), among others. Sims (2003) and Reis (2006) examined the economic agents’ imperfect attentiveness to exogenous macroeconomic shocks. Angus. (1991), Carroll (1992) and Ludvigson and Alexander (2001) considered the buffer-stock models which allow for precautionary saving motives, impatience and restrictions on borrowing. Pischke (1995) examined consumers’ incomplete information on economy-wide variables. Moreover, there is a vast amount of literature on habit persistence and the equity premium puzzle (for example, Christiano, Eichenbaum and Evans, 2005; Smets and Raf 2004) and the business cycle (for example, Sundaresan 1989; Abel 1990; Constantinides 1990).

² They used the GMM technique.

³ We derive price volatility using the CPI. GARCH model is used to attain the series.

⁴ These estimates are fairly consistent with the FIML; we do not report the FIML results to conserve space.

⁵ The Wald test results for other $\beta$s are not reported to conserve space; however, they are available from the authors upon request.

⁶ Analogous results are obtained using the FIML and these are not reported for brevity. The estimates of $\beta$are statistically significant at the conventional levels in all samples.

⁷ These results are not reported to conserve space.