|
Article Excerpt
"This seeming inabilility to find a significant (and consistent!) impact of demand is surprising." (Nordhaus 1972, p. 35)
IN CLASSICAL MICROECONOMIC models, product prices move in such a way as to stabilize production. If demand increases, firms raise prices, and this dampens the demand for their products. But the link between demand and prices, which follows directly from textbook theory, has been hard to find in the data. Estimates of price equations typically show that prices respond strongly to factor prices, but they are much less responsive to demand (for references, see, e.g., Nordhaus 1972, Gottfries 1991, Bils and Chang 2000). Bils and Chang (2000) confirmed this result in a recent study. Shea (1993) found that prices typically do rise with demand, but with a considerable lag. (l) Menu costs can explain slow price adjustment, but not an asymmetric response to cost and demand shocks.
On the macroeconomic level, researchers find similarly weak effects of demand on prices. Impulse-response functions from structural VAR models show very little movement of wages and prices in the first year after a monetary shock (Blanchard 1989, Christiano, Eichenbaum, and Evans 1999). To fit this fact, modern macro-econometric models with sticky prices often incorporate an implausible degree of nominal price stickiness. Smets and Wouters (2005), for example, estimated the average duration of prices to be about 2 1/2 years in the United States and the euro area. This is far longer than one finds in micro data (Bils and Klenow 2004; see also Altig et al. 2005). Backward-looking indexation schemes are commonly introduced in empirical macro models so as to make inflation more persistent (Christiano, Eichenbaum, and Evans 2005).
The puzzling behavior of prices suggests that some important elements are missing in the textbook treatment of price determination. Nominal rigidities cannot be the whole story but there must also be some real rigidity, which slows down price adjustment. (2) In this paper, we argue that long-term customer relations, financial constraints, and interaction between prices and investment may generate real price rigidity and slow down price adjustment in response to a demand shock.
In a customer market, buyers repeatedly purchase a good. Customers who are attracted by low prices tend to remain loyal and customers lost because of high prices are hard to win back. The seminal paper by Phelps and Winter (1970) formalized this idea. They analyzed a firm's choice between a high price, which increases profits today, and a low price, which attracts customers and increases profits in the future. Gottfries (1991) and Chevalier and Scharfstein (1996) showed that if firms in a customer market are financially constrained, markups may be countercyclical. In a recession, financially constrained companies abstain from price cuts in order to maintain cash flows and pay their debts; during booms, they can afford to pursue a more aggressive price policy. Empirical evidence consistent with this hypothesis is reported by Bhaskar, Machin, and Reid (1993), Chevalier and Scharfstein (1996), Gottfries (2002), and Asplund, Eriksson, and Strand (2005).
If firms are financially constrained and sell their products in a customer market, one would expect to see important interactions between investment in physical capital and investment in market share. High demand implies high cash flow, but also a need for additional capacity. High predetermined investment expenditure should make it more likely that a firm becomes financially constrained. The purpose of this paper is to explore this interaction between investment and price decisions theoretically and empirically.
We develop a dynamic model of a firm, which sells its output in a customer market. The firm has two assets: physical capital and the customer stock (market share). It can invest either in physical capital or in the customer stock, by charging a low price to attract new customers. The firm is financially constrained: it does not issue new shares, managers (or owners) dislike fluctuations in dividends, and only a fraction of its investments can be financed by borrowing. (3) We also allow for the fact that the completion of an investment project is a prolonged process. According to Nickell (1978), the whole completion process takes about 23 months, whereas Hall (1977) found that investments are completed in 21 months. To capture this in a simple way, we assume a 1-year implementation lag.
We solve the model numerically and find optimal decision rules for price and investment. Without financial constraints, we get conventional pricing behavior; prices respond immediately and positively to cost and demand shocks. But if financial constraints are important, we get sluggish price adjustment after a demand shock. To understand this, consider an unexpected permanent decrease in demand. With investment predetermined and demand falling, the firm finds itself in a financial squeeze. In order to finance ongoing investments and avoid drastic cuts in dividends, the financially constrained firm abstains from price cuts. In subsequent periods, investment is reduced; the firm becomes less financially constrained and cuts price in order to increase its market share. Hence, there is a form of lagged price adjustment after a demand shock.
A wage increase has an immediate effect on the price because higher wage costs raise marginal cost and also make firms more financially constrained. Thus, our model can explain an asymmetric response to cost and demand shocks. Furthermore, the model predicts a positive relation between investment and prices because, ceteris paribus, high predetermined investment tends to make firms more financially constrained. This is a new and testable prediction, which differentiates our theory from other explanations of countercyclical markups, such as those presented by Rotemberg and Saloner (1986), Bils (1989), Rotemberg and Woodford (1992), Ireland (1998), and Bils and Chang (2000).
To explore whether the dynamics of prices and investment are qualitatively consistent with our model, we estimate structural price and investment equations on a large data set for manufacturing plants 1990-98. The data source is unique; it provides a wide coverage of plants in Swedish industry and is not limited to data for a certain branch or a small number of products as is usual in micro data studies of pricing behavior. Plant specific price indices have been constructed by Statistics Sweden using a mixture of plant specific unit values and disaggregate producer price indices.
To disentangle how prices respond to cost and demand shocks, we exploit the openness of the Swedish economy. Using industry data for export and import shares, foreign and domestic production and prices, and exchange rates, we construct firm-specific measures of demand and competitors' prices. Industries differ in their dependence on foreign markets and in their exposure to foreign competition and this is a source of considerable cross industry variation in demand and competitiveness, which should help us to disentangle the effects of costs, demand, and competitors' prices. To deal with simultaneity, we use foreign demand and prices as instruments. These variables are reasonably exogenous for a small open economy and can be seen as a small open economy alternative to the demand instruments used by Hall (1988), Shea (1993), and Ghosal (2000).
As predicted by our theory, wage costs and competitors' prices affect the price, but demand variables have small and mostly insignificant effects. Investment has a positive effect on the price, which is both statistically and quantitatively significant. In fact, the estimated effect is larger than predicted by our model. Combined with adjustment lags in investment, this implies slow price adjustment after a demand shock. This real rigidity may play an important role in the propagation of business cycle shocks.
The theory is set up in Section 1 and the numerical solution is presented in Section 2. Data and variable definitions are presented in Section 3 and estimation issues are discussed in Section 4. Baseline results are presented in Section 5, Section 6 contains some robustness checks, and Section 7 concludes.
1. A MODEL OF PRICE AND INVESTMENT DYNAMICS
The firm's customer stock is [X.sub.t] and each customer buys [Y.sup.[sigma].sub.t] units, where [Y.sub.t] is a demand shock, so production is
[Q.sub.t] = [X.sub.t][Y.sup.[sigma].sub.t]; [sigma] > 0. (1)
The customer stock changes slowly, increasing or decreasing over time depending on the price charged by the firm, [P.sub.t], relative to the average market price, [P.sup.0.sub.t]:
[X.sub.t] - [X.sub.t-1]/[X.sub.t-1] = -[epsilon]([P.sub.t] - [P.sup.0.sub.t]/[P.sup.0.sub.t]); [epsilon] > 0. (2)
This relation can be motivated in alternative ways. Phelps and Winter (1970) provided theoretical foundations for such an equation based on imperfect information, Gottfries (1986) considered a mix of imperfect information and switching costs, while Klemperer (1987, 1995) emphasized switching costs. Ravn, Schmitt-Grohr, and Uribe (2006) derive a similar equation assuming that consumers form habits over individual varieties of goods.
The functional form of the demand function determines the importance of competitors' prices in the optimal price policy. The more convex the demand curve, the more important are competitors' prices for the firm's optimal price. In macroeconomic models with static monopolistic competition, demand is often assumed to be constant-elastic, i.e., concave, so the markup is independent of competitors' prices. We have chosen the demand curve which is linear in the relative price because it leads to pricing behavior that is roughly consistent with what one finds empirically; most studies find that the price depends on costs as well as on competitors' prices. (4)
The production function takes the CES form
[Q.sub.t] = [(1 - [alpha])[([A.sub.t][F.sub.t]).sup.[rho] + [alpha][K.sup.[rho].sub.t-1]].sup.1/[rho];0 < [alpha] < 1, (3)
where [K.sub.t-1] is capital at the end of period t - 1, [F.sub.t] is a flexible production factor, and At is an exogenous technology factor. There is an adjustment cost associated with investment:
C([DELTA][K.sub.t], [K.sub.t-1]) = c/2[([DELTA][K.sub.t]/[K.sub.t-1]).sup.2] [P.sup.k.sub.t][K.sub.t-1], (4)
where [P.sup.k.sub.t] is the price of capital goods. We do not allow for fixed adjustment costs or irreversibility. Both are likely to be important, but they would make the model much more complicated (Carlsson and Laseen 2005).
As is well known, dividends are much more stable than investment and borrowing, and new share issues play a modest role in the financing of investments; most investment is financed by retained earnings and borrowing. To introduce financial constraints in a way that is broadly consistent with these observations, we make four assumptions:
(i) The firm does not issue shares. This may be because of adverse selection problems or because owners or managers fear loss of control. (5)
(ii) Owners or managers dislike fluctuations in dividends. In a small entrepreneurial firm, where the owner has all his capital invested in the firm and lives on the dividends, the owner's preference for smooth consumption translates into a preference for smooth dividends. (6) More generally, it seems clear that managers or owners dislike fluctuations...
|
