VOL. 74 NO. 3 KING AND PLOSSER: MONEY, CREDIT, AND PRICES households)purchase transaction services, current inputs, future production is higher dy, and allocate labor services to transaction and the current returns to additional units of activities in fixed proportions where the scale factors of production are higher. These variable corresponds to total payments to offsetting effects are analogous to the income actors of production and thus is closely and substitution effects of a real interest rate related to next period,s output (for house- change. Essentially, the small impact of a holds the scale variable is x, +i,). The sec- shift in p, on the amount of output allocated ond and third assumptions eliminate n to capital accumulation (akut/a% =0)re- dhr, du, and kdt from the vector fects the idea that the income and substitu decision variables tion effects on consumption are roughly There is a discounted dynamic program- ffsetting. On the other hand, the substitu- ming problem whose solution corresponds tion effect of such shifts on labor supply is the competitive equilibrium of this simplified presumed to dominate so that anut/ap, >0, model economy. The decision rules for the which generates procyclical work effort problem are stationary functions of the state Once quantity behavior is determined, variables y, and p,. Rather than solve this equilibrium factor prices, interest rates, and problem for an explicit specification of pref- share prices are straightforward to erences and technologies, the essential fea- In particular, competitive prices tures of the interactions between the final to marginal rates of substitution n at optimal goods industry and the financial industry can planned quantities. For example, there is a be analyzed by employing the following re- riskless commodity interest rate r, that we strictions on the decision rules; 0< ak y/ay discuss below. We also can construct the <1, akur/a, =0, anu/dy,>0, and u)u]=(o, -u)/ur. In our setup, this ex- These restrictions follow from assumptions pected return exceeds the riskless rate, bout preferences and production opportuni- E,(ru)>r, since the holders of these shares ties. For example, an increase in the amount must be compensated for bearing production of the initial stock nvolves additional risk wealth so that the consumption of final prod uct and leisure are expected to rise. Agents Real Business Cu e 2d the nsiae however, choose to spread some portion of this wealth increment over time and do so by cated to capital services so that 0< ak w/ay, correlation(comovement)of real produgsitive ncreasing the amount of commodity allo- In our real business cycle model, <1. The other conditions on the decision credit, and transaction services arises from rules require stronger restrictions on prefer- the general equilibrium of production and ences and production possibilities. For ex- consumption decisions by firms and house- ample, an increase in y, raises the marginal holds. The timing patterns among these vari- product of labor if capital and labor are ables, however, depends on the source of the complements in production. If the wealth variation in real output ffect on labor supplied, which arises from Unexpected output events(s,) operate by the increased output of final goods next altering the initial conditions pertinent for period, is outweighed by the increase in the economic agents' plans for consumption, in- real wage(marginal product of labor) then vestment, and hours of work. As discussed hours worked rises. 4 bove,an unexpected wealth increment(5,> Analogously, an increase in involves 1)leads to higher net investment than would both wealth and substitution effects. Given otherwise have been the case. Furthermore, that real output increases and exhibits positive serial correla 4 This result also requires that the amount of time tion. during the course of such an economic old is small relative to total time allocated to market expansion, the volume of credit( shares)is also high as firms finance relatively large
THE AMERICAN ECONOMIC REVIEW JUNE 1984 mounts of goods in process. This positive econometricians(Christopher Sims, 1972 correlation between the total volume of credit 1980), and monetary historians (Milton and real activity is potentially an important Friedman and Anna Schwartz, 1963)view prediction of our framework, especially since monetary variables as"leading"measures of evidence presented by Benjamin Friedman real activity One way of generating a diferent timing between total credit and output than be- pattern is through shifts in the intertemporal tween the individual components of credit opportunities of the economy as a whole and real activity Real events of this type, respresented by e The movements in final goods production alter agents'allocations of leisure and con- duces a higher volume of transaction sumption between the present and the future services demanded by firms and households for a given level of national wealth. A higher Thus, our model generates the positive co- than average shock (, >1), under the as- movement of output with measures of sumptions outlined above, expands hours bank clearings, long noted by empirical re- worked with little accompanying change i ample, Wesley Mitchell, 1930, pp. 116-51). cial services are an intermediate Phuc? searchers in the business cycle area(for ex- consumption or capital. The fact that fina Finally, real rates of return move in a which can be produced more rapidly than countercyclical direction as agents'oppor- the final product-leads tunities to spread wealth over time are sub- the quantity of such services and of bank ject to diminishing returns (i. e, total time is deposits. Consequently, movements in hours in fixed supply) worked, interest rates, and security prices, The predictions of our model focus on the deposits-financial services, and trade credit flow of transaction services. It is important all occur prior to the expansion of to provide a link between these flows and the The subsequent increment to time t+1 stock of deposits that has been the more wealth(stemming from the joint impact of traditional focus of monetary analysis. It is the exogenous shift, P, and agents'responses convenient to assume that the stock of de- to that shift) works much like the above posits IS proportional to the flow of transac. discussion of unexpected output events tion services and can be represented by Yd Typically, we suspect the initial phases of Under this assumption, our model implies business fluctuations incorporate a combina- that the volume of inside money(deposits)is tion of both types of shocks (i. e, shifts in positively correlated with output with a rough current and expected future production pos coincidence in timing. More generally, this sibilities) may reflect the role of deposits as a store of IL. Currency, Deposits, and Prices process At least some cycle episodes, however, are In order to investigate the relation be- commonly viewed as involving a different tween nominal aggregates and the real busi timing pattern. For example, traditional ness cycle it is necessary to augment the business cycle analysts (Arthur Burns and Mitchell, 1946), modern time-series macro- sIms(1980)discusses reverse causation and output working through central bank esent setup is a first step toward the type It is sufficient for our purposes that f small-scale general equilibrium model that is neces- any monotonic in- sary to evaluate the reverse causation argument. creasing function. Although this assumption is a conven- 7 We deliberately employ the idea of a"leading vari- tional assumption with physical capital, it is nevertheless able"in ure the common ad hoc element ling. For example, tran elements of these alternative discussions action services le, require any specifie It is commonly stressed that asset prices and returns checking accounts that incorporate information about predictabl equilibrium model d mation is also incor- in the volume of debits re porated into all quantity decisions such as effort, con the stock of deposits sumption, an