استهلاک های ارز، نقل و انتقالات مالی، و عدم تجانس شرکت

The present paper investigates five episodes of currency collapse from the perspective of non-financial firms operating in Argentina, Brazil and Mexico. We focus on two aspects: wealth and income transfers from borrowing firms to lenders and firm heterogeneity. At the firm level, we find that the currency collapses are preceded and associated with sharply rising financial transfers from firms to lenders. The debt and income structure is central in explaining the asymmetric firm dynamics. Most affected are firms with high levels of unhedged foreign-currency debt. At the country level, Argentina, Brazil, and Mexico display three contrasting examples. Argentina has a large currency mismatch, Brazil balances the currency denomination of debt and income (natural hedge), and Mexico occupies an intermediate position.

This paper studies income and wealth transfers associated with currency crises. It provides evidence based on micro data for three countries (Argentina, Brazil and Mexico) that the currency composition of firm liabilities, as well as the contribution of exports to revenues, determines the extent to which firms are affected by large currency depreciations. The distributive dimension of currency crises has been largely understudied with the exception of Halac and Schmukler (2004). This question is, however, central to the understanding of currency crises to the extent that these transfers have macroeconomic effects and determine the length and protractedness of currency crises. In a world with perfect markets distributive issues are irrelevant. The efficient market hypothesis argues that the liability structure of firms does not matter for investment decisions. Firms are indifferent in financing investment through equity issues, domestic debt or foreign-currency denominated debt. Similarly, profits are zero in a world with frictionless markets but are non zero in the presence of rigidities, for instance exchange rate pass-through, nominal price stickiness, search and matching frictions, moral hazard, and adverse selection on financial markets. In particular, the literature on the financial accelerator puts forward that there is a relationship between unanticipated exchange rate fluctuations and firm profitability (and/or net wealth). In the presence of imperfect information commercial banks use firm profitability (and/or net wealth) as explanatory variables for the probability of default. Unanticipated exchange rate movements affect the liability structure of firms indebted in foreign currency and generate transfers between lenders and borrowers. Banks cut credit supply to these firms, which are forced to scale back investment. The probability of default can therefore be inferred by firms’ profitability and their debt-to-asset ratio. In Stiglitz and Weiss (1981), profitability accounts for the ability of firms to meet interest payments. In Bernanke et al. (1999), the debt-to-asset ratio is a proxy used by creditors to measure the potential losses in case of firms going bankrupt. Currency collapses (large nominal depreciations or devaluations) in emerging markets are associated with two types of transfers between lenders and borrowers: income transfers (flows) and wealth transfers (stocks). Debt accumulation in foreign currency without hedging the currency exposure raises the income transfers from borrowers to lenders, when the exchange rate depreciates. The depreciation raises the domestic-currency value of the stock of foreign-currency liabilities and translates into a transfer of wealth from borrowers to lenders. In addition, interest payments in domestic currency increase due to the mentioned stock effect and an associated increase in the risk premium on external finance caused by the reduction in wealth (Bernanke et al., 1999). A firms’ ability to meet interest payments is greatly determined by the export structure of their revenues. The exchange rate depreciation also improves firms’ competitiveness and foreign sales. The degree of export orientation of firms, which works as a natural hedge against exchange rate risks, is a second distributive issue considered here alongside the composition of debt. In fact, it is shown that, with the exception of Brazil, firms indebted in foreign denominated currency were not necessarily exporting firms. The present paper investigates and compares five episodes of currency collapses from the perspective of non-financial firms operating in Argentina, Brazil and Mexico. The impact of currency collapses on the income distribution between borrowers and lenders is a complex issue that needs to be settled empirically, ideally at the micro-level. To this purpose we have gathered a detailed firm-level data set on the financial statements of non-financial firms over the period 1994 to 2004. First, the income statements of firms are decomposed into net financial payments and earnings from real activity over time to quantify the impact of the currency collapses on these two components of income. Compared to investment, which is the usual variable considered, firms’ income has a distributive aspect since it allows one to separate the generated surplus from financial payments and gives a clear picture of the income distribution. The profit decomposition is done at the average country level and across firms with varying degrees of foreign-currency debt and income. We find that the currency collapses are preceded and associated with sharply rising financial transfers to lenders and that the debt and income structure is central in explaining the asymmetric firm dynamics. Second, using panel regressions, the impact of the currency collapses on firms’ income is quantified after controlling for differences in the currency composition of debt, the degree of export orientation, and macroeconomic conditions. To account for the differences across countries and currency collapses, the regressions are estimated for each country individually. In particular we stress the importance of being indebted in foreign currency as well as being an exporter. This part shares similarities with existing works by Aguiar, 2005, Bleakley and Cowan, 2008, Bonomo et al., 2003, Gilchrist and Sim, 2007 and Pratap et al., 2003. The country studies of Gilchrist and Sim (2007) on Korea, and Aguiar, 2005 and Pratap et al., 2003 on Mexico find that investments of firms that are indebted in foreign currency are constrained by adverse balance sheet effect that dominates the competitiveness effects of devaluations. The opposite is found by Bleakley and Cowan (2008) who analyze currency mismatches and depreciations for a sample of firms from Argentina, Brazil, Chile, Colombia and Mexico during the 90s. Focusing on investment, they find evidence that firms indebted in foreign currency matched the currency composition of their liabilities with the exchange rate sensitivity of their profits, suggesting that liability dollarization has not been the driving force of the recessionary impact of the currency depreciations. Finally Bonomo et al. (2003) find that the currency composition of debt had no significant impact on investment in Brazil. As they argue, the results of Bleakley and Cowan (2008) might be driven by the fact that Brazilian firms represent more than 50% of the observations in their sample. Apart from a different perspective of our work, our value added with regard to the aforementioned studies is that we make use of quarterly data for firms from Argentina, Brazil, and Mexico from 1994 to 2004, whereas the other studies relied on annual data. This allows us to track the direct impact of currency collapses. Contrary to Aguiar, 2005, Bonomo et al., 2003, Gilchrist and Sim, 2007 and Pratap et al., 2003, we do not focus on a single economy, and, in contrast to Bleakley and Cowan (2008), we do not pool firms from different countries together in the estimations. Similarly, we analyze each episode of currency collapses separately. Our empirical study has therefore a comparative dimension and takes into account differences in the initial conditions of firms across countries, and differences in the circumstances of each currency collapse. The estimations show that the currency collapses had indeed a different impact across firms and countries. In particular, we find evidence in favor of Aguiar, 2005, Gilchrist and Sim, 2007 and Pratap et al., 2003 for the cases of Mexico and Argentina, i.e. the adverse balance sheet effects tend to dominate the positive competitiveness effects on income. On the contrary, the competitiveness effects have been dominant in Brazil, which is in line with the findings of Bleakley and Cowan, 2008 and Bonomo et al., 2003. The remainder of the paper is organized as follows. In Section 2, average income is decomposed into its real and financial part and examined separately for the three economies over time. The income decomposition is done as well for high and low exporters, and firms with high and low levels of foreign-currency debt. In Section 3, we present the econometric results. The final section summarizes the main results of this paper and draws conclusions.

The empirical literature on the impact of currency collapses on the macroeconomy is inconclusive. Since the transmission mechanisms can operate in opposite directions, it is important to investigate empirically which channel dominates, taking into account the heterogeneity at the micro-level. In this paper we focused on the distributive aspects associated with currency collapses from the perspective of non-financial firms from Latin America. Two aspects that seemed the most important in our view have been investigated: the income and wealth transfers across borrowers and lenders, as well as the heterogeneity of borrowers. These two aspects are addressed by decomposing firms’ income into its real and financial components over time. In addition panel regressions are estimated to quantify the effect of currency collapses on firms’ profits and we identify the driving forces of the underlying heterogeneity. There are two main results. First we identified three channels through which currency collapses increase transfers from the borrowing firms to their lenders. In the case of unhedged positions in foreign-currency debt, there is a balance sheet channel and the currency depreciation reduces the borrowers’ wealth. The equity-to-asset ratio decreases between 5 and 15 percentage points, depending on the country taken into account. There is as well a flow channel as higher liabilities raise financial payments to lenders. Financial payments peaked at 15% of total assets in Argentina. In addition, high dollar debt firms experienced a 3 to 12% drop in profitability in the three affected economies. Eventually, there is an interest rate effect as the country risk premium reached 20 to 70%. Second, we identified two levels of heterogeneity. At the firm level, those firms that benefited from the devaluation were specialized in the export sector, while having rather low levels of foreign-currency debt. On the other hand, firms losing from the devaluation were specialized in the domestic sector, while having a high level of foreign-currency liabilities. In Argentina, winning firms experienced a 4% increase in profitability, while losing firms experienced a 10% decrease in profitability. At the country level, Argentina, Brazil, and Mexico display three contrasting examples. Argentina has been an economy with a large currency mismatch, Brazil had some sort of natural hedge, and Mexico occupied an intermediate position. Consequently, the wealth and income transfers have been three times higher in Argentina than in Brazil.