The Effect of Hurricane Matthew on Consumer Spending

Abstract

Severe weather events, such as blizzards and hurricanes, can temporarily disrupt economic activity. The imprints of these events on aggregate statistics can make it challenging for macroeconomists to analyze and forecast economic conditions. As one illustration, the minutes from March FOMC meetings in six of the past seven years cited unseasonable temperatures or winter storms as influencing economic activity. Weather events present an opportunity to observe how consumers adjust their spending in the face of unanticipated shocks. Thus far, however, there has been little analysis of the spending effects from weather events, partly due to a lack of data at both a sufficiently high frequency and a geographically detailed level. For example, official statistics, such as retail sales from the Census Bureau, are only estimated nationally at a monthly frequency. In this note, we take a step forward in this regard using a new dataset of transaction volumes to examine how consumers reacted to Hurricane Matthew, which struck the East Coast in October 2016. Our results reveal that the hurricane significantly reduced consumer spending in the affected states (Florida, Georgia, South Carolina, and North Carolina) in early October. Although the level of spending after the storm quickly returned to normal, very little of the preceding shortfall appears to have been made up in the subsequent weeks, suggesting that, on net over the span of a few weeks, the hurricane had a negative effect on spending. The drop in activity was most apparent in the discretionary components of spending, such as restaurants, as opposed to necessities, such as grocery stores.2 The net negative effect on spending implies that inter-temporal smoothing through this temporary shock is either incomplete or slow to occur.