ترکیب بدهی و اثر ترازنامه نرخ ارز در برزیل : تجزیه و تحلیل در سطح بنگاه

In this paper, we study the interaction between macroeconomic environment and firms’ balance sheet effects in Brazil during the 1990's. We start by assessing the influence of macroeconomic conditions on firms’ debt composition in Brazil. We found that larger firms tend to change debt currency composition more in response to a change in the exchange rate risk. We then proceed to investigate if and how exchange rate balance sheet effects affected the firms’ investment decisions. We test directly the exchange rate balance sheet effect on investment, but the results were not statistically significant. We then pursue an alternative investigation strategy, inspired by the credit channel literature. According to this perspective, Tobin's q can provide an adequate control for the competitiveness effect on investment. Our results provide supporting evidence for imperfect capital markets, but not for a balance sheet effect in Brazil. The main effect we found is that firms in industries with higher proportion of imported inputs tend to invest less when the exchange rate is depreciated.

The macroeconomic environment interacts with the firms’ balance sheet structure in a two-way relationship. On one hand, macroeconomic environment is central in shaping the capital markets, determining what kind of contracts is feasible and enforceable. Moreover, it also affects the incentives faced by firms when selecting their financial contracts. Conversely, the firms’ balance sheet structure affects crucially the result of macroeconomic policies, influencing policymakers’ choices of regimes and policy rules. In this paper, we study the balance sheet effects of exchange rates and interest rates in Brazil since 1990, using a panel data set with firm level variables. For this endeavor, we also consider how the macroeconomic environment affected the balance sheet structure and interacted with firms’ balance sheet effects. Balance sheet effects on investment and production rely on capital market imperfections. According to the credit channel literature (see Bernanke and Gertler, 1995), imperfect information creates a wedge between internal and external finance. An adverse shock to the net worth of a financially constrained firm increases its cost of external financing and decreases the ability or incentive to invest, and to implement production plans. It should impact firms differently, being stronger for firms that face higher premium of external finance costs relative to internal finance (see Hubbard, 1998). There is substantial empirical evidence that proxies for firms’ net worth affect investment more for low net worth than for high net worth firms (Hubbard, 1998). Therefore, to the extent that exchange rate and interest rate variations affect firms’ net worth, their balance sheet effect should matter for determining investment. Firms will see their financial condition deteriorate whenever they have substantial debt at floating interest rates, and the relevant real interest rate increases. This can happen if they have foreign denominated debt and the real exchange rate depreciates, entailing an exchange rate balance sheet effect. An interest rate effect takes place when firms have substantial short-term domestic debt or long-term debt contracted at floating rates, since their loans will be rolled over at higher rates. In the case of interest rates, it engenders a financial accelerator, which magnifies the traditional interest rate channel (Bernanke et al., 1999). In the case of exchange rate, it should counteract the expansionary effect of the competitive channel (Aghion et al., 2001).1 Thus, while the question in the interest rate literature is about the magnitude of the recessive impact of interest rate rises the debate in the recent exchange rate literature is about whether exchange rate devaluations are expansionary or contractionary. Harvey and Hoper (1999) argued, in an investigation which used firm level indicators that the exchange rate balance sheet effect greatly exacerbated the Asian Crisis. Bleakley and Cowan, 2002 and Forbes, 2002 tested the empirical relevance of exchange rate balance sheet effects using multinational panel regressions with firm level data. The former work used a panel data for over 500 non-financial firms in five Latin American countries, dominated by Brazilian firms (52.5% of the observations). They found that holding foreign-currency denominated debt was associated with more investment during exchange rate devaluations, contrary to the predicted sign. However, Forbes (2002) found that more indebted firms had lower net income growth after a large depreciation. Although she used a larger sample of countries, she only examined large depreciations. One advantage in focusing one specific country is that we can take into consideration the relevant specifics of the macroeconomic environment in this economy. The macroeconomic conditions changed drastically in the last 12 years. First there was an important trade liberalization, which occurred in the early 1990s. Simultaneously, the control of international financial flows was softened, increasing the access of Brazilian firms to foreign liabilities. The Real plan in 1994 ended the high inflation period in Brazil, unveiling the new incentives, and at the same time creating additional ones. The sudden reduction of inflation rate and its volatility contributed for the strengthening of credit relations and for the lengthening of debt maturities. This happened at first in an environment characterized by low volatility of the real exchange rate. In the beginning of 1999, the exchange rate was allowed to float. This change of exchange rate regime was complemented by the adoption of an inflation targeting monetary regime. As a result, exchange rate became very volatile while interest rate policy became focused on bringing inflation to target. At firm level, investment is the candidate variable potentially more influenced by balance sheet effects. Balance sheet effects might additionally affect production. Since firms cash flow could be an important channel through which balance sheet deterioration affect investment, we also investigate how cash flows affect investment (also measure of capital market imperfection), and how they are influenced by our balance sheet effects. We start by studying the relation between the macroeconomic environment and the balance sheet structure. After briefly providing some background information about the macroeconomic reforms in Brazil, we proceed by analyzing the determinants of debt composition. Our main finding is that larger firms react more to an increase in exchange rate risk than smaller firms by reducing the proportion of foreign currency debt in their liabilities. We then study the balance sheet effects. We perform several tests, starting with some basic equation where the balance sheet effect of exchange rate is tested directly. We do not find any significant effect of firm's dollar indebtedness on investment when exchange rate is devalued. The only robust effect we found is that firms in industries with higher proportion of imported inputs invest less when the real exchange rate is more depreciated. We also investigate if by allowing exchange rate balance sheet effects to vary over periods we could find different results. Our findings suggest that if exist a negative exchange rate balance sheet effect, it is due to the floating exchange rate period. We then explore the link between balance sheet effects and investment through a common test of capital market imperfection. Our claim is that Tobin's q could provide a better control for the competitiveness effect. Our results provide evidence for imperfect capital markets, but not for the existence of balance sheet effects of exchange rate. We also follow the capital market imperfection tradition by testing if the balance sheet effects are stronger for firms with characteristics that make them more likely to be financial constrained, but the results we found go in the opposite direction: large firms tend to have a negative balance sheet effect. We proceed as follows. In the next section, we provide the general macroeconomic background for the period of analysis. In Section 3 we describe our database. The investigation methodology is outlined in Section 4. Section 5 presents the results and the last section concludes.