Conflict, terrorism and consumer confidence

Consumer confidence and psychology, shown to matter for the economy in developed countries, may be crucial even in the context of developing countries where markets are relatively unstable and erratic. This warrants a deeper understanding of the factors that drive developing economies, including measures of consumer expectation and sentiment. This paper uses novel and unique data from a recent endeavor by the central bank of Pakistan to periodically collect survey data on consumer sentiment from a large sample of households across the nation.

Specifically, we examine how consumer sentiment responds to the conflict environment. A host of factors crucial for development are sensitive to existence of violent conflict – it can lower growth (Rodrik 1999) and human capital investment (Ichino and Walter-Ebmer (2004), Leon 2012) and affect individuals’ social, risk, and time preferences (Voors et al 2012). Using information on location, we match consumer sentiment information to events of violent conflict from the Global Terrorism Database and from Armed Conflict Location & Event Data. The consumer sentiment phone interviews, conducted with households every other month, collect households’ sentiment about prices, general economic position, and private economic and financial condition. Households are interviewed during the first 10 days of the month; we identify the short-term effect of exposure to terror events by comparing responses of individuals in the same location and month if they happen to get interviewed just after a terror event in their locality relative to households who got interviewed before the event.

We find that in general highly conflict prone areas demonstrate lower rate of consumer satisfaction. Comparing individuals surveyed just after terror events to other individuals in their region, we find that recent local terror events lower consumers’ sentiment about future and current prices. However there are considerable variations across demographic groups. Particularly, the sentiment of more vulnerable sections of the society, including females, low income and low education households, is most negatively impacted by events of violent conflict. These are also the groups who demonstrate high biased-ness in price expectations (in line with Souleles 2004 and Madeira and Zafar 2012). This heterogeneity in reaction to violent events may be that these groups are more affected by them; or that they make more inefficient forecasts, thus exaggerate their responses more. Violent events, like bombings and armed assault, are most significant. The consumers in large cities are also the ones who are more likely to react to terror events.