Cash flow instability and the demand for liquid assets by firms in Brazilian manufacturing.
Models developed to analyze the demand for liquid assets by firms typically emphasize the need for immediate and reliable purchasing power in the face of an uncertain future. Any factor reducing the stability of the cash flow may encourage firms to hold a greater quantity of liquid assets as "insurance" against risk (Miller and Orr, 1966, 1968).
One such factor that has generally been neglected by previous studies is a firm's involvement in international trade. Exchange rate variability, the existence of more complex negotiations and contractual obligations than domestic trade, and the need to make business decisions with less complete information will probably cause cash flow to be less certain for firms engaged in international trade. Very little empirical work has focused on the impact of this factor on liquid asset demand, Katz, Rosenberg and Zilberfarb (1985) being an exception. They found that firms involved in exporting have a greater demand for liquid assets than firms which sell only in domestic markets.
The major goal of this paper is to empirically test, using a large sample of manufacturing firms in Brazil, the impact of a firm's involvement in international trade on its demand for liquid assets. We start with a standard approach to the firm demand for liquid assets (a scale variable and industrial control variables as independent predictors) and incorporate foreign involvement into the model in the form of exchange rate variability and the extent of foreign trade. We also incorporate the variability of both the domestic inflation rate and interest rate since these variables may also influence the stability of a firm's cash flow. In the process of this investigation we also examine whether subsidiaries of multinational corporations react differently than domestically owned firms to these sources of instability.
II. THE APPROACH
The model of firm liquid asset holding has generally emphasized two independent variables: a measure of the firm's size of operation (a scale variable) and a rate of return available on some non-money asset. While past studies have argued for and utilized many different scale variables (e.g., sales, assets, total value of output, and business receipts) the most common is sales.
There is even less consensus on the rate of return variable. Some studies use the rate on a short-term financial asset while others prefer longer term assets whether they are financial or physical. Many studies do not even use a rate of return variable in the model because of data limitation or the belief that all rates of return tend to quickly move in unison thereby eliminating any profit incentive to switch assets. Another justification for excluding a rate of return variable is provided by Meltzer (1963). He argues that holdings of liquid assets are essentially a function of wealth and the rate of interest. Sales are related to wealth by the internal rate of return and a variable which measures changes in the capital-labor ratio and changes in business activity (how intensively productive capacity is used). This complex decision process is simplified for annual data, Meltzer argues, because for any particular industry (in our study any particular firm) in any given year the interest rate, capital-labor ratio and internal rate of return are constant. Thus liquid asset demand can be explained by sales in cross-section regression equations. Numerous studies of firm demand for money or liquid assets omit the interest rate from the model: DeAlessi (1966), Falls and Natke (1988), Katz et al. (1985), Meltzer (1963), Vogel and Maddala (1967) and Whalen (1965).
This study chooses a basic model which uses total firm output as the scale variable and excludes any measure of the rate of return on alternative assets. Since a firm's behavior is influenced by the nature of its industrial environment (e.g., the manufacturing process, market structure, firm rivalry, inventory policies, the nature of and attitudes towards risk) industrial dummy variables are also included in the basic model. This is a procedure generally followed since Whalen (1965). The focus of this paper, however, is the introduction of variables that measure the impact of a firm's foreign involvement on its demand for liquid assets.
One variable expected to affect the stability of a firm's cash flow is exchange rate variability. This volatility makes a firm's revenue on foreign sales or its cost of production (via imported components or materials) less certain. Any increase in cash flow instability creates a desire to hold more liquid assets as a hedge against an uncertain future. This suggests that an increased variability in exchange rates would increase the demand for liquid assets.
Exchange rate variability should also cause substitution among assets within a firm. Domestic and foreign firms will probably use different methods of substitution and, therefore, react differently to exchange rate variability. Management of domestic firms are expected to substitute assets they view as safer than the riskier cruzeiro denominated assets so that holdings of the latter assets decrease and holdings of foreign currency denominated assets rise. This substitution has no impact on the accounting category called liquid assets since the measure includes both domestic and foreign currency denominated holdings. In contrast, managers of multinational corporations view the Brazilian subsidiary as only one of a global network of firms. These managers are likely to shift funds to other subsidiaries (i.e., minimize balances held in the Brazilian subsidiary) when the exchange value of the Brazilian currency becomes more volatile in order to avoid a reduction in the asset value of the Brazilian subsidiary. This shifting of funds to other subsidiaries in the face of exchange rate variability would decrease the subsidiary's liquid asset holdings. McGibany and Nourzad (1995) found that increases in exchange rate variability resulted in reduced money holdings using aggregate U.S. data. Thus the impact of exchange rate variability on liquid asset holdings is not clear and may differ between multinational and domestic firms.
Foreign transactions are also generally more complex than those related to domestic trade. The firm must deal with tariff and non-tariff barriers to trade (e.g., product standards and entry procedures) and variable administration of customs requirements. Acquiring information about foreign markets is expected to yield benefits by raising revenue or reducing cost. But information is also costly to acquire. The economically optimal quantity of information (where the expected marginal cost is equated to the expected marginal benefits) therefore reduces uncertainty over cash flow but does not eliminate it. The extent of foreign trade acts as a single proxy for the complexity and information influences.
An increase in foreign involvement, holding exchange rate variability constant, may increase instability in a firm's cash flow and thus lead to increases in their holdings of liquid assets as a hedge.(1) Alternatively, one could argue that as a firm's foreign involvement increases, it acquires valuable market information and experience which, over time, reduces the marginal cost of acquiring further information. This additional information reduces the cash flow instability problem and the desire to hold liquid assets. In summary, the effect of foreign involvement on a firm's liquid asset holdings is not clear.
One domestic source of instability for a firm's cash flow is the variability of inflation. Variable and unpredictable inflation rates can make liquid assets less attractive as an asset choice for a firm's portfolio than physical assets. Investment in plant and equipment produces goods and services which can serve as a good hedge against inflation and yield a higher expected rate of return than investment in liquid assets. Silveira (1973) has suggested that there were limited substitution possibilities for money in Brazil because of underdeveloped bond and equity markets and the repression of interest rates caused by the imposition of government price controls. He asserted that goods may assume a predominant role among substitutes for money or liquid assets.
This hypothesized negative relationship between the demand for money and inflation rate variability was found by Blejer (1979) in his study of several Latin American nations. Silveira (1973) reached a similar conclusion using Brazilian data while Khan (1982) has suggested that this negative relationship holds for any nation with rapid and sometimes erratic inflation.
Another source of domestic instability is variability in the rate of interest. If increased interest rate volatility creates greater uncertainty about expected revenues and payments then a risk-averse firm will hold more liquid assets as a hedge against that uncertainty. Friedman (1982) suggested that this relationship existed. McGibany and Nourzad (1986) found some empirical support for this hypothesis using U.S. data and Silveira (1973) reached similar conclusions using aggregate data from Brazil.
Some of these independent variables, however, are interrelated. The impacts of inflation on cash flow could be captured in three places: (1) the explicit measure of inflation variability; (2) the inflation volatility implicit in the variability of nominal interest rates; and (3) inflation variability implicit in the variability of nominal exchange rates. This may cause difficulty in isolating what separate impacts inflation, interest rate and exchange rate variability have on liquid asset holdings.
We also seek to determine if ownership influences firm behavior. Numerous empirical studies of Brazilian industries have concluded that multinational subsidiaries behave differently than domestically owned firms in choice of production technique, trade, technology transfer, response to government policy and impacts on income distribution.(2) Differences in firm behavior could take several forms in our study. Multinational subsidiaries may hold less liquid assets, other things being equal, because they have superior financial management and greater access to foreign financial markets through the parent firm or the subsidiary network (a parallel shift in the demand for liquid assets function). However, multinational subsidiaries may also react differently to specific stimuli than domestic firms. For example, the liquid asset holdings of subsidiaries might be less sensitive to fluctuations in exchange rates because it uses the forward market more often.(3) Also if multinationals have more information about foreign markets, they are less likely to "hedge" against the increased cash flow instability created by foreign trade. These potential differences in behavior dictate that interaction terms between ownership and the measures of international instability be included in the model.
III. DATA AND EMPIRICAL MODEL
The data for the empirical analysis consists of annual observations on 256 manufacturing firms which engaged in foreign trade during the 1974 through 1976 period. These firms are in fifteen two-digit industries as classified by the Brazilian Geographic and Statistical Institute (see Appendix for classification system). The term "multinational" refers to subsidiaries of any multinational corporation based outside Brazil, while "Brazilian" is used to describe privately owned domestic firms. All monetary measures are expressed in real Brazilian cruzeiros using 1972 as the base year.
The firm's balance sheet and income statements were taken from reports filed with the Ministry of Finance. Data on firm exports and imports were collected from CACEX which was Brazil's import/export authority. Many of the largest firms in Brazilian manufacturing are in the sample including most of the prominent firms in each industry. Some small privately owned firms are also included to provide a more representative sample.(4)
The use of annual data is subject to the usual caveats: (1) all monetary measures are aggregated over one fiscal year and this may be too long to capture significant variations; (2) the year end figures may not be representative due to various "window dressing" practices; and (3) firms end their fiscal years at different times during the calendar year.
The regression equation is estimated in conventional log-linear form:
LLA = [b.sub.o] + [b.sub.1]LOUTP + [b.sub.2]LEXRD + [b.sub.3]LFI + [b.sub.4]LINFD + [b.sub.5]LTBRD + [b.sub.6]OWN + [b.sub.7]EXROWN + [b.sub.8]FIOWN + [b.sub.9]INDXX + e (1)
where e is a residual with the usual properties.
The dependent variable is the natural log of real liquid assets (LLA). Liquid assets corresponds to the usual accounting definition and is approximately the sum of cash, bank deposits, and short term securities. Given the accounting nature of the data it was impossible to break down liquid assets into its subcategories of short-term securities, bank deposits and cash. This is not a severe limitation. The Internal Revenue Service data for the United States, a primary source of data for previous work (Hunter, 1978; Katsimbris and Miller, 1980; Meltzer, 1963; Vogel and Maddala, 1967; Whalen, 1965), includes some savings deposits in its "cash" category. Another major source of data (Securities Exchange Commission) defines "cash" to include time deposits and other financial assets. Our use of liquid assets roughly corresponds to many of the previous empirical studies and is less subject than "money" to variation since firms may substitute among assets within this measure with no impact on its level. Frazer (1964) and Wilbraite (1975) have concluded that a division between money and money-substitutes is not meaningful when examining firm financial behavior. Therefore, a broader accounting category like liquid assets may exhibit greater stability over time than a narrower category such as cash particularly in Brazil which experienced a high inflation rate over the period.
The scale variable is the natural log of the firm's total output (LOUTP) which is the sum of the firm's sales and change in inventories. Industrial classification is also controlled for in the analysis using a set of dummy variables (INDXX). Metals manufacturing (IND11) is the default category since it contains the largest number of firms. LFI is the natural log of the ratio of exports plus imports to total sales and measures the extent of a firm's activity in foreign markets. The variables LINFD and LTBRD, are the natural logs of the standard deviation, based on monthly observations, of the wholesale Brazilian inflation rate and the thirteen-week Brazilian Treasury bill rate respectively.(5) The dummy variable OWN is one if the firm is at least twenty-five percent foreign-owned and zero otherwise, a practice common in the literature on multinational corporations. The slope interaction terms designed to capture differences in multinationals' behavior toward foreign sources of cash flow instability are FLOWN and EXROWN. These two variables are the product of OWN with LFI and LEXRD, respectively.
For multinational firms the exchange rate variability measure LEXRD is the standard deviation, based on monthly observations, of the nominal exchange rate between the Brazilian cruzeiro and the currency of the country in which the firm's parent company is located.(6) For domestically owned firms a measure of the cruzeiro's exchange value is obtained by averaging its bilateral exchange rates in terms of the currencies of the G-10 countries plus Argentina, Mexico and Panama. These thirteen nations are Brazil's largest trading partners. Natke (1982) found that 94% of exports by domestically owned manufacturing firms in Brazil were to developed countries and 75% of their imports were from these nations. Since the exact pattern of trade for each firm in each year is not known a simple average of these exchange rates is used.
This differential measurement of exchange rate variability is based on two factors. One is the belief that the management of multinational firms focus on global performance denominated in the currency unit of the parent firm. For example in the 1970s, common stock ownership of U.S. multinationals predominantly remained in the U.S. and the performance of subsidiaries, individually and collectively, was evaluated in terms of the U.S. dollar. Therefore, it is the variability of the U.S. dollar-cruzeiro exchange rate which best measures, for U.S. firms, the risk of holding assets in Brazil. A second justification of the use of the parent nation exchange rate for multinational firms is provided by trade flows. A study of multinational subsidiaries in Brazil (Natke, 1982) found that a large portion of trade flows were with the parent nation (49% for U.S. based multinationals and 46% for multinationals based in Europe, Canada and Japan). Newfarmer and Mueller (1975) also reported that a high percentage of Brazilian subsidiaries' trade was with the parent nation. Therefore variability of the parent nation currency versus the cruzeiro best measures the variability of cash flow associated with a firm's trade flows.
IV. EMPIRICAL RESULTS
The results of estimating the model by ordinary least squares using pooled data are presented in the second column of Table 1. A Chow test of the basic model concluded that the model was homogeneous over time thus allowing all three years of data to be pooled over time. The influence of LEXRD alone is insignificant at traditional levels. However, the t-statistic for the coefficient EXROWN (2.04) suggests that, as expected, the impact of exchange rate fluctuations on liquid asset holdings depends upon whether a firm is Brazilian or foreign-owned. For multinationals the total influence of exchange rate fluctuations is given by the sum of the coefficients of LEXRD and EXROWN. This sum is -.058 with a standard deviation, not shown in Table 1, of .021(t = -2.74). Thus it appears that while exchange rate variations are not important in determining liquid asset holdings for domestically owned firms they are to multinational corporations. These results suggest that multinational firms reduce liquid assets when exchange rates are more volatile.
The coefficients of both LFI and FLOWN are statistically insignificant as is the sum of the two coefficients (i.e., the total impact of foreign trade activity on multinational firms' holdings of liquid assets). This suggests that the degree of involvement in international trade is apparently not empirically important to either type of firm in determining liquid asset holdings unlike the conclusions of Katz (1985).
As expected an increase in the volatility of the wholesale inflation rate reduces the amount of liquid assets held. These results affirm the conclusions of studies using aggregate data in developing nations (Silveira, 1973; Blejer, 1979; Khan, 1982). Increased volatility of interest rates apparently increase liquid asset holdings affirming the conclusions of Silveira (1973) for aggregate data from Brazil and McGibany and Nourzad (1995) for aggregate U.S. data.
The coefficients associated with OWN and LOUTP have the anticipated signs and are statistically significant at the five percent level or better. Foreign owned firms hold less liquid assets than their domestic counterparts. On the other hand, increases in output lead to a less than proportionate rise in these holdings. The existence of economies of scale for firms is consistent with other studies of aggregate money demand in Brazil (Barbosa, 1978; Cardoso, 1983; Contador, 1974; Pastore, 1969; Rossi, 1988; da Silva, 1973; Silveira, 1973). Only firms in the pharmaceutical industry behave differently from those in metals manufacturing by holding fewer liquid assets.
[TABULAR DATA FOR TABLE 1 OMITTED]
The model was also estimated with only one interaction term and with none since it is recognized that the presence of such variables may cause multicollinearity problems. These outcomes are presented in columns three through five of Table 1. Both the magnitude and statistical significance of the LEXRD coefficient are sensitive to the specification. This parameter is insignificant whenever EXROWN is present (Equations 1 and 4) and significant (negative) whenever EXROWN is absent from the model (Equations 2 and 3).
The total influence of exchange rate fluctuations on the liquid asset holdings of multinationals, given by the sum of LEXRD and EXROWN, is significantly negative. The magnitude of this impact is approximately the same whether both interaction terms are included (Equation 1) or if only EXROWN is included (Equation 4). Thus it appears that the impact of exchange rate fluctuations on liquid asset holdings of multinationals is not sensitive to the specification.
The coefficients of LINFD, LFI, and OWN are also apparently dependent upon the model employed in the analysis. With both interaction terms absent, the magnitudes of the coefficients of these three variables are much smaller compared to when EXROWN and FIOWN are both present. The qualitative results for the other variables are basically unchanged.
The divergent responses by multinational and Brazilian firms to exchange rate fluctuations leads one to question whether members of the two groups also react differently to other stimuli. To examine this broader question, separate regression equations were estimated. This allows multinationals and domestically owned firms to respond differently to each independent variable in the model. These results are presented in Table 2.
One contrast between the groups is in their reaction to exchange rate variations. The response of multinationals to an increase in these fluctuations is a small but statistically significant, reduction in liquid asset holdings. On the other hand, domestic firms significantly increase their holdings when exchanges rate become more volatile.
Another major difference between ownership types is readily apparent in the LFI parameter. The degree of foreign involvement does not significantly affect liquid asset holdings of the foreign subsidiaries. As expected, however, Brazilian companies increase such holdings as the extent of their involvement in foreign markets increases. Multinationals' constant international involvement might reduce the marginal cost of acquiring further information and reduce the uncertainty over cash flow which arises from foreign transactions. Subsidiaries can count on the experience of the parent firm and centralized international operations. Brazilian firms having less continuous experience in the international economy, face greater uncertainty over cash flow for a given level of foreign involvement. Their liquid asset holdings, therefore, are more sensitive to variations in foreign involvement.
The third contrast between these two types of firms concerns the effect of inflation rate fluctuations. While the coefficient of LINFD is significantly negative for both groups, it is much larger for domestically controlled companies.
Table 2. SEPARATE REGRESSION RESULTS Independent Domestic Multinational Variables Firms Firms Constant .2603 -2.1486 (.33) (-3.75)(***) LOUTP .8062 .8644 (17.87)(***) (21.62)(***) LEXRD .5235 - .0627 (2.29)(**) (-2.58)(***) LFI .0694 -.0480 (2.45)(**) (-1.30) LINFD -.8528 -.2464 (-2.33)(**) (-1.66)(*) LTBRD .1851 .2196 (1.72)(*) (1.98)(**) IND10 .3367 -1.3632 (1.67) (-3.40)(***) IND12 -.0202 .3139 (-.10) (1.74)(*) INDI3 -.1011 .0812 (-.72) (.54) IND14 -.1169 .2593 (-.71) (1.41) IND15 -.0040 (-.01) IND16 -.0033 .3549 (-.08) (.88) IND17 -.2994 .1219 (-1.89)(*) (.45) IND18 -.0492 (-.19) IND20 -.4830 .1285 (-2.35)(**) (.73) lND21 -.6181 -.1045 (-2.69)(***) (-.58) IND23 .2075 -.3205 (.47) (-1.20) IND24 -.0543 .4495 (-.32) (1.93)(*) IND26 -.3056 .1138 (-2.01)(**) (.42) IND30 -.1550 (-.52) Summary Statistics [R.sup.2] .6167 .5684 F 30.72(***) 34.07(***) Note: t statistics in parentheses. *, **, *** indicates significance at the 10%, 5%, and 1% levels respectively. In this sample there are no multinational firms in industries 15 and 18 and no domestic firms in industry 30.
It is interesting to note that there are also major differences between multinationals and domestic firms in the industry control variables. For example, multinationals in the cement/ceramics industry (IND10) hold significantly fewer liquid assets than such firms in metals manufacturing (the default category) whereas there are no differences across the two industries for Brazilian firms. Differences also arise in capital goods (IND12), paper products (IND17), petroleum/basic chemicals (IND20), pharmaceutical (IND21), fibers and fabrics (IND24), and food products (IND26). These results and others presented in Table Two suggest that there are differences between multinationals and domestically controlled firms with respect to liquid asset management.
V. CONCLUSIONS AND INFERENCES
Multinational subsidiaries respond differently to exchange rate variability than do Brazilian firms. Across all model specifications, multinationals respond to increased exchange rate variability by significantly decreasing their holdings of liquid assets. The size of the appropriate coefficients varies little across specifications: they range from -0.052 to -0.062. When multinationals and Brazilian firms are allowed to react differently to each of the model's independent variables (Table 2), domestic firms increase their holdings of liquid assets in the face of increased exchange rate variability.
Different propensities to use the forward exchange rate market might explain the difference in behavior toward exchange rate variability. Multinationals reduce liquid asset holdings by purchasing contracts in forward exchange markets in response to increased volatility of exchange rates. A survey by Soenen and Aggarwal (1987) supports this line of argument. In a survey of 259 corporations from the United Kingdom, the Netherlands and Belgium a substantial majority (roughly 72%) engage in forward exchange rate contracts.
Altering transfer prices, making foreign payments only on net rather than gross transactions and leading and lagging payments are tactics that can reduce the quantity of a subsidiary's funds exposed to exchange rate risk. As exchange rate variability rises, multinationals use these tactics more intensively which reduces the Brazilian subsidiaries' holdings of liquid assets. For example, over the sample period the cruzeiro experienced substantial declines versus the major international currencies. This encouraged multinationals to lead payments from Brazil to other subsidiaries, delay payments to the Brazilian subsidiary, raise prices of goods imported to Brazil and lower export prices. On the other hand, small Brazilian firms with relatively low levels of foreign involvement and inconsistent trading experience may find limited opportunities to use forward exchange markets since the marginal cost of employing this technique exceeds any perceived marginal benefits in risk reduction. These firms are also unable to engage in the same types of financial management techniques (e.g., leads and lags of payment) as multinationals can or use as extensively those techniques that are available to them.
During the early 1990s Brazil experienced a significant increase in exchange rate variability. To the extent government policies contributed to this rise they have according to our results led multinational firms to reduce liquid asset holdings of their Brazilian subsidiaries. However these same policies may have resulted in domestic firms increasing their holdings of such assets. Thus foreign subsidiaries may be illiquid relative to their domestic counterparts.
Table 2 also shows that Brazilian firms increase their liquid asset holdings as foreign involvement rises while multinationals' behavior is not altered. Multinationals' experience in foreign markets is a cumulative learning process that in many cases has a long history implying low marginal cost for incremental trading activity of a single subsidiary. We believe that the results of Table 2 (when all variables are allowed to change across ownership groups) more closely reflects actual firm behavior.
As indicated by the coefficient of OWN, domestic firms hold a greater amount of liquid assets than multinationals even when controlling for other influences. This difference may simply reflect better liquid asset management practices, economies of scale in liquid asset management available to the larger global firms, or better access to foreign financial markets should the need for short-term borrowing arise.
Another difference in firm behavior is in the reaction to a variable inflation rate in Brazil. Domestic firms reduce their holdings of liquid assets to a greater extent than the multinationals when faced with increased inflation variability. The explanation may lie in factor substitution policies. During periods of high and variable inflation rates when the return on real capital is increasing relative to the return on liquid assets Brazilian firms may purchase more capital or possibly increase their product inventories by reducing their holdings of financial assets particularly cash. We have no hard evidence that either of these substitutions were pursued by Brazilian firms during the time of the study although Silveira (1973) has suggested that problems in Brazilian financial markets created conditions where "goods may assume a predominant role among substitutes for money" (p. 120).
However, Cardoso's (1983) empirical evidence suggests that goods inventories are not an important alternative to holding money in Brazil. One problem with firms increasing inventory stocks is the possibility that they could be caught holding relatively high priced inventory if inflation were to suddenly and unexpectedly decline.(7) Multinationals are less likely to purchase more capital to reduce liquid asset holdings because variable inflation rates create uncertainty about the future value of the subsidiary's assets in terms of the currency of the parent firm.
In the first three years of the current decade there was a dramatic decline in the volatility of the Brazilian inflation rate. This trend has been reversed in more recent years. To the exent government policies contributed to these changes they have probably caused both domestic and foreign owned firms to more frequently adjust their liquid asset holdings. According to our results these changes would have been larger for domestically owned firms.
One similarity between the two ownership groups is in their response to interest rate volatility. In this case both significantly increase their holdings of liquid assets as these rates become more uncertain, a conclusion that studies using aggregate data in Brazil (Silveira, 1973) and the U.S. (McGibany and Nourzad, 1995) have also reached.
One may wonder how the current model performs with respect to the substantial body of literature on firm liquid asset demand. First, consider the scale elasticity. These elasticities range from 0.3076 in Katsimbris and Miller's (1982) study to 1.249 for that of Katz et al. (1985). The elasticities reported in this study compare favorably with the literature: between 0.853 and 0.855 in Table 1; 0.806 for Brazilian firms and 0.864 for multinationals in Table 2. Our results suggest that the basic model specification used in previous studies can be modified to include measures of international and domestic instability without substantially altering their results. This opens up a field of inquiry into firm financial behavior that is potentially productive without laying waste to past efforts.
BRAZILIAN GEOGRAPHIC AND STATISTICAL INSTITUTE Industrial Classification Description of Industry 10 cement, ceramics 11 ferrous, non-ferrous metals and products 12 capital equipment 13 electrical equipment 14 automotive and transportation 15 wood products 16 furniture 17 paper products 18 rubber products 20 petroleum and basic chemicals 21 pharmaceutical 23 plastics 24 fibers and fabrics 26 food products 30 miscellaneous equipment and products
* Direct all correspondence to: Gregory A. Falls, Central Michigan University, Department of Economics, Mt. Pleasant, MI 48859.
1. A secondary effect of increased foreign involvement could be a reduction in exchange rate variability for the firm if there is a diversification of trading partners (nations). Since an increased volume of trade does not necessarily imply a greater diversification of trade and trade direction data was not available for all firms and years this potential impact has been ignored.
2. The following selected studies of Brazilian industries are sufficient to justify that behavioral differences based on ownership did exist during the time period of this study: Connor and Mueller (1982), Falls and Natke (1988), Lall and Streeten (1977), Natke (1986), Newfarmer and Marsh (1981), Newfarmer and Mueller (1975), and United Nations (1985).
3. The extent of trading on the forward exchange market may be substantial when one considers optimal hedging strategies suggested by previous studies. Swanson and Caples (1987), for example, find that the optimal hedging ratios are 75 and 94% for the German mark and British pound respectively.
4. The firm level data were originally compiled during a research project under the direction of Richard Newfarmer and Lawrence Marsh at the University of Notre Dame. The project was funded by the U.S. Department of Labor.
5. These two variables change across years but are constant for all firms in a given year. The relationships between inflation, the interest rate and the exchange rate imply collinearity of independent variables and larger standard errors of the estimated parameters. Although simple correlations are not a conclusive measure of the degree of multicollinearity in the model, they do indicate the problem is not likely to be severe. The simple correlation between LTBRD and LINFD is -0.051; between LEXRD and LINFD is 0.236; between LEXRD and LTBRD is 0.097. We should also note that the nature of volatility discussed in deriving the model is ex ante while the variability measured in the empirical section is ex post. This, of course, is a problem for most empirical work involving uncertainty unless surveys of expectations are undertaken at regular intervals.
6. Nominal exchange rates were used in the study. Pauls (1987) has pointed out that four issues are potentially involved in the selection of a weighted average exchange rate index: the selection of weights, the selection of currencies, the use of real versus nominal measures, and the selection of an appropriate price index for calculation of real exchange rates. These issues are more difficult to resolve because of two economic events in Brazil over the sample period: inflation and price controls. High rates of inflation in 1974-1976 perhaps suggests that a real exchange rate be used but several troubling questions remain. Are the decisions of firm management influenced by real or nominal exchange rates? Does the pattern of real rates diverge enough from nominal rates that their respective variances are substantially different? If real exchange rates are chosen, which price index is appropriate? GDP deflators and consumer price indices include many goods and services not traded in international markets (e.g., housing, medical services) while the commodity composition of wholesale price indices may vary substantially across nations. Furthermore, the presence of price controls in Brazil biases any conversion of nominal to real exchange rates using the wholesale price index.
On an empirical level, McGibany and Nourzad (1995) found that their conclusions were not dependent on the choice of a real or a nominal exchange rate to measure volatility.
7. We would like to thank an anonymous referee for bringing this point to our attention.
Barbosa, Fernando de Holanda. 1978. "A Demanda de Moeda no Brasil: Una Resenha de Evidencia Empirica," Pesquisa e Planejamento Economico, 8: 33-82.
Blejer, Mario I. 1979. "The Demand for Money and the Variability of the Rate of Inflation: Some Empirical Results," International Economic Review, 20: 545-549.
Cardoso, Eliana A. 1983. "A Money Demand Equation for Brazil," Journal of Development Economics, 12: 183-193.
-----. 1991. "Da Iercia a megainflacao: O Brasil nos anos 80," Pesquisa e Planejamento Economico, 21: 129-159.
Conner, John and Willard F. Mueller. 1982. "Market Structure and Performance of U.S. Multinationals in Brazil," Journal of Development Studies, 18: 329-353.
Contador, R. 1974. "Desenvolvimento Financeiro, Liquidez e Substituciao Entre Ativos no Brasil: A Experencia Recente," Pesquisa e Planejamento Economico, 4: 245-284.
DeAlessi, Louis. 1966. "The Demand for Money: A Cross-Section Study of British Business Firms," Economica, 33: 288-302.
Falls, Gregory A. and Paul A. Natke. 1988. "The Demand for Liquid Assets: A Firm Level Analysis," Southern Economic Journal, 53: 630-642.
Frazer, William J. 1964. "The Financial Structure of Manufacturing Corporations and the Demand for Money: Some Empirical Findings," Journal of Political Economy, 72: 176-183.
Friedman, Benjamin. 1982. "Federal Reserve Policy, Interest Rate Volatility, and the U.S. Capital Raising Mechanism," Journal of Money, Credit and Banking, 14: 721-745.
Hunter, Helen Manning. 1978. "Corporate Demand for Cash: The Influence of Corporate Population Growth and Structure," Review of Economics and Statistics, 60: 467-471.
Katsimbris, George M. and Stephen M. Miller. 1980. "Distribution Effects and the Business Demand for Money," Journal of Macroeconomics, 2: 287-305.
-----. 1982. "The Nonfinancial Corporate Demand for Cash: A Disaggregated Analysis," Quarterly Review of Business and Economics, 22: 96-103.
Katz, E., J. Rosenberg and B.Z. Zilberfarb. 1985. "The Demand for Liquid Assets and Involvement in Exports: Some Empirical Results," Empirical Economics, 10: 125-129.
Khan, Ashfaque H. 1982. "The Demand for Money and the Variability of the Rate of Inflation: An Empirical Note," Economic Letters, 10: 257-261.
Kuh, Edwin. 1965. Capital Stock Growth: A Micro-Econometric Approach, 2nd ed. London: North-Holland.
Lall, Sanjaya, and Paul Streeten. 1977. Foreign Investment, Transnationals and Developing Countries. Boulder, CO: Westview Press.
McGibany, James M. and Farrokh Nourzad. 1986. "Interest Rate Volatility and the Demand for Money," Quarterly Review of Economics and Business, 26: 73-83.
-----. 1995. "Exchange Rate Volatility and the Demand for Money," International Review of Economics and Finance, forthcoming.
Meltzer, Allan H. 1963. "The Demand for Money: A Cross-Section Study of Business Firms," Quarterly Journal of Economics, 77: 405-422.
Miller, Merton and Daniel Orr. 1966. "A Model of the Demand for Money by Firms," Quarterly Journal of Economics, 80: 413-435.
-----. 1968. "A Model of Money Demand by Firms: Extensions of Analytical Results," Journal of Finance, 23: 735-759.
Natke, Paul A. 1982. Intrafirm Trade and Transfer Pricing: A Study of Manufacturing Firms in Brazil. Ph.D. dissertation, University of Notre Dame.
-----. 1986. "Market Imperfections and Import Pricing Behavior by Multinational Enterprises," Journal of Economic Development, 11: 105-118.
Newfarmer, Richard S. and Lawrence C. Marsh. 1981. "Foreign Ownership, Market Structure and Industrial Performance: Brazil's Electrical Industry," Journal of Development Economics, 8: 47-76.
Newfarmer, Richard S. and Willard F. Mueller. 1975. MNCs in Brazil and Mexico: Structural Sources of Economic and Non-Economic Power. U.S. Congress, Senate Committee on Foreign Relations, Sub-Committee on Multinational Corporations. Washington: Government Printing Office.
Pastore, A.C. 1969. "Inflacao e Politica Monetaria no Brasil," Revista Brasileira de Economia, 23: 92-123.
Pauls, B. Diane. 1987. "Measuring the Foreign Exchange Value of the Dollar," Federal Reserve Bulletin, 73: 411-422.
Rossi, Jose W. 1988. "A Demanda por Moeda no Brasil: O Que Ocorreu e Partir de 1980?" Pesquisa e Planejamento Economico, 18: 37-53.
da Silva, Adroaldo M. 1973. "Demanda de Moeda e Taxa Esperada de Inflacao: Un Estudo Empirico de Argentina, Brasil, Chile, e U.S.A." Estudos Economicos, 3: 59-101.
Silveira, Antonio M. 1973. "The Demand for Money: Evidence from the Brazilian Economy," Journal of Money, Credit and Banking, 5: 113-140.
Soenen. L.A. and R. Agarwal. 1987. "Corporate Foreign Exchange and Cash Management Practices," Journal of Cash Management, 7: 62-64.
Swanson, Peggy E. and Stephen C. Caples. 1987. "Hedging Foreign Exchange Risk Using Forward Foreign Exchange Markets: An Extension," Journal of International Business Studies, 18: 91-98.
United Nations Centre on Transnational Corporations. 1985. Transnational Corporations and International Trade: Selected Issues. New York: United Nations.
Vogel, Robert C. and G. S. Maddala. 1967. "Cross-Section Estimates of Liquid Asset Demand by Manufacturing Corporations," Journal of Finance, 22: 557-575.
Whalen, Edward L. 1965. "A Cross-Section Study of Business Demand for Cash," Journal of Finance, 20: 423-442.
Wilbraite, Barry J. 1975. "Some Essential Differences in the Demand for Money By Households and By Firms," Journal of Finance, 30: 1091-1099.
|Printer friendly Cite/link Email Feedback|
|Author:||Falls, Gregory A.; Natke, Paul A.|
|Publication:||Quarterly Review of Economics and Finance|
|Date:||Jun 22, 1996|
|Previous Article:||Risk and return in trading closed-end country funds: can trading beat holding foreign stocks?|
|Next Article:||The social costs of monopoly and regulation: Posner reconsidered again.|