محدودیت های نقدی متفاوت، اهرم مالی و تقاضا برای پول: شواهدی از یک پانل کامل از شرکت های تایوانی

This paper studies firms’ demand for money by developing a differential-cash-constraint framework with firms’ entire wage bills requiring cash in advance and a fraction of investment purchases being financed by credits. In addition to conventional scale and opportunity-cost factors, firms’ financial status and profitability are crucial determinants for their money demand behavior. Employing a new data set consisting of a panel of Taiwanese firms over 1990–97, our econometric analysis lends empirical support to our theory. The estimates suggest that economies of scale in firms’ cash management are present and that lower financial leverage or higher profitability raises money demand significantly.

Theoretical and empirical studies on money demand have been at the center of the stage of monetary economics. Such issues are not only important for understanding the relationships between money, inflation and growth, but also crucial for conducting monetary policy and evaluating its welfare implications. While there is a large body of research on household demand for money, the literature examining money demand by firms is rather thin. Our paper contributes to the latter literature both theoretically and empirically. Specifically, we develop a generalized differential cash constraint model by taking an explicit account of firms’ financial status. We then perform a systematic empirical test of the validity of our theory using a newly available panel data set of reliable quality. Four decades ago, Miller and Orr (1966) extend the Alais–Baumol–Tobin inventory model to study firms’ demand for money under uncertainty, suggesting the presence of economies of scale as money serves as a medium of exchange for firm transactions. Although the property of scale economies is empirically supported by Selden, 1961 and Frazer, 1964, Meltzer (1963) uses cross-section data and finds that the sales elasticity of money demand by firms is very close to unity, thereby implying the absence of economies of scale.1 Since then, Whalen, 1965, Vogel and Maddala, 1967 and Falls and Natke, 1988 have revisited this issue. By considering the portfolio choice aspect, their cross-sectional analyses reconfirm Selden’s finding. Whalen and Vogel and Maddala suggest a tendency for cash economization when they incorporate assets as well as transactions into the demand function, obtaining estimates of the sales elasticities ranging from 0.86 to 1.08 (with only four industries below unity). Falls and Natke find that the sales elasticity of liquid assets of Brazilian manufacturing firms is about 0.9 after controlling for foreign ownership, industry structure, and macroeconomic variables. Thus, a general conclusion drawn from this later literature is that the economies of scale in cash management are essentially negligible. More recently, Mulligan (1997) re-estimates firm’s money demand based on the value of time approach developed by Karni, 1973 and Bomberger, 1993, by constructing a pseudo panel data set from COMPUSTAT, which is clearly superior to the conventionally used cross-sectional data. In particular, Mulligan uses the wage rate to measure the cash manager’s value of time and finds that an increase in the value of time results in higher money demand. His estimates of sales elasticities around 0.83 suggest moderate economies of scale of money in facilitating firm transactions. Bover and Watson (2005) correct measurement errors, accounting for firm unobserved effects that are correlated with sales. They find economies of scale in cash management in the US, but not in the UK or Spain. Our paper contributes to the literature in two important aspects. First, we construct an intertemporal model of money demand by firms, in contrast to the conventional inventory or value of time approach.2 The central feature of the theory developed here is the consideration of differential cash constraints in which firms’ entire wage bills require cash in advance whereas a nontrivial fraction of investment purchases can be financed by credits. This framework is not only simplistic to permit analytic solution for the money demand function, but also realistic to capture different cash requirements for different types of firms’ purchases. An infinitely lived firm owner optimizes over labor, capital investment, asset holdings and real cash balances. In steady-state equilibrium, a firm’s demand for money depends crucially on its wage bills and investment purchases, as well as on the nominal interest rate and an array of profitability and financial variables that may influence labor employment and the endogenous differential cash constraint factor. Second, we perform empirical analyses based on a new data set that contains 345 Taiwanese companies of various sizes and industrial categories over the period of 1990–97. Although the data at the firm level are much smaller in sample size than the COMPUSTAT data, we are able to construct a complete panel from the original source particularly suitable for the purpose of our study. Our empirical work based on the generalized differential cash constraint theory can be thought of as a complement to the conventional inventory and the value of time models. While a firm’s demand for money is increasing in wage bills and investment purchases, the coefficient estimates for wage bills far exceed those for investment purchases, justifying our differential cash constraint hypothesis. Moreover, the results lend support to the presence of economies of scale for firms’ transactions use of money and indicate that lower financial leverage or higher firm profitability tends to induce greater money demand. The overall scale elasticity is in the range of 0.58–0.68, implying significant scale economies in firms’ cash management. The interest rate elasticity is higher than the conventional estimates, falling between 0.70 and 0.98 (in absolute value). While the financial leverage is a consistently powerful explanatory variable for both the fixed-effects model and the instrumental variable method, turnover rates and term and risk premia have statistically significant effects in some specifications. These findings are robust to an alternative two-step estimation and to deletion of extreme sales values. The remainder of the paper is organized as follows. Section 2 develops an intertemporal model of a representative firm owner facing a differential cash constraint that enables us to characterize the property of money demand in steady-state equilibrium. While Section 3 describes the data and the empirical methodology, Section 4 presents the panel data estimation using the fixed-effects model as well as the instrumental variable method. Finally, we conclude the paper and acknowledge the limitations of our study in Section 5.

We have revisited the issue of firm’s demand for money by developing a generalized differential cash constraint model and undertaking empirical tests with a complete panel of Taiwanese firms over the period of 1990–97. We support the hypothesis of differential cash constraints and the property of economies of scale in the transaction role of money for firms where the overall scale elasticity is in the range of 0.56–0.78. The nominal interest rate has a conventional negative effect on money demand with the elasticity measures falling between 0.70 and 0.98. Several firm profitability and financial variables also play significant roles in influencing the money demand behavior, including the profit/sales ratio, the leverage (debt/equity) ratio and the term and risk premia. In the end, we would like to acknowledge the limitations of this study. On the one hand, the data is at annual frequency, which is not suitable for studying within-the-year adjustments in cash. On the other, since our sample includes only firms listed on the Taiwan Stock Exchange, the resultant selection bias problem exists. Thus, our conclusions, especially those regarding scale economies in the transactions use of money, may not be entirely applicable to small enterprises in which cash management is arguably less effective. Nonetheless, our empirical evidence provides an interesting contrast with that in previous work, and our generalized differential cash constraint theory can be applied to performing econometric analysis using available data from other countries.