Recessions are often attributed in popular opinion and by policymakers to large, identifiable, negative shocks, which hit specific important sectors. For example, the slow recovery from the last U.S. recession was attributed to the attacks of September 11, 2001, which hit the travel sector. Terrorism also hurt tourist destinations like Israel, while natural disasters wrecked the tourism and fishing sectors in Louisiana (Hurricane Katrina) and South Asia (the 2004 tsunami).
It is not obvious how such a temporary sectoral shock gets propagated across sectors and through time, thus affecting the entire economy negatively even after it is gone. This paper discusses one propagation mechanism and demonstrates it in an environment that is based on the monetary search model of Kiyotaki and Wright (1989, 1993). Since that model is supposedly too narrow for analyzing broad macroeconomic issues, I explain the intuition up front. In the model there are n types of agents and goods, and agents specialize in production and consumption: Type i produces good i and consumes good i+1 (mod n). If the production of good i is temporarily impossible the income of type i agents is reduced. They will buy less of their consumption good, good i+1, next period. Their low demand for good i+1 will, in turn, reduce the money holdings of that good’s producers type i+1 agents so those agents will later buy less of good i+2, and so on. The propagation of the shock continues in a chain reaction from one sector to another even if the shock to good i lasts only one period.
Of course, well-functioning credit markets could help agents in mitigating these effects, but the use of money in an economy proves that at least some of its credit markets are not well functioning. The propagation mechanism is therefore closely related to the essentiality of money.
The propagation of changes in purchasing power is implicit in the static Keynesian multiplier, but here it is dynamic and creates a long, auto-stabilizing business cycle in a sectoral context. This paper complements the literature on sectoral linkages through intermediate goods (Long and Plosser 1983, Scheinkman and Woodford 1994), in which a sectoral shock deprives other sectors of a production input. Here there are no intermediate goods: Sectors are connected by liquidity from the demand side instead of the supply side. Kiyotaki and Moore’s (1997a) “credit chains” model also uses intermediate goods, in order to justify limited credit relationships: A default by one debtor causes a liquidity contagion because each agent owes money to its supplier.
They derive an immediate contagion similar to the Keynesian multiplier. I show that a liquidity contagion is possible even in an economy in which any type of credit is completely impossible, and I describe a complete business cycle. The business cycle described in Kiyotaki and Moore’s (1997b) “credit cycles” model also has contagion but not due to illiquidity: It comes from changes in asset prices of collateral used by both sectors. Wallace (1997) already uses a monetary search model to analyze propagation of a shock, but he does so for a monetary shock whereas here the shock is real.
Money is not new to business cycle analysis, but its role here is fundamentally different from the roles previously given to it. The literature has analyzed the effects of exogenous or endogenous changes in money supply. In contrast, money belongs here simply because when it is needed in its most fundamental role a medium of exchange it propagates a real, isolated shock into a persistent, economy-wide business cycle.
The paper is organized as follows. Section 2 presents the basic model and the propagation of a one-period shock to one sector. Sections 3 and 4 show robustness to properties of the shock and to consumption preferences, respectively, while generating additional business cycle regularities. Section 5 concludes. An appendix shows robustness to the indivisibilities assumed in the text.
Download
PDF Ebook The Propagation of Sectoral Shocks in a Monetary Search Model
