Hedge or Speculation?

Evidence of the use of derivatives by Brazilian firms during the financial crisis

Jos Luiz Rossi Jnior1 Insper Institute of Education and Research

Abstract This paper analyzes the use of foreign exchange derivatives by non-financial publicly traded Brazilian companies from 2007 to 2009. Using balance-sheet data on firms positions in derivatives and their foreign exchange exposure, this study finds that a significant number of companies speculated in the derivatives market. Two types of speculators are identified: companies that significantly increased the volume of derivatives used during this period but used them in line with their currency exposure and companies that adopted positions that would have been inadvisable had the aim been to hedge their currency exposure. Despite the differences between the two types, there is one similarity: both tried to obtain gains through the continuous process of domestic currency appreciation. The study shows that companies that allegedly have an informational advantage on the foreign exchange market - exporters and companies with foreign-currency-denominated debt - are more likely to speculate. No other theory about the reasons why companies speculate can explain the behavior of these companies.

Keywords: Speculation; hedge; derivatives; exposure; risk management.

JEL Classification: G32; G30.

E-mail: joselrj1@isp.edu.br. Address: Rua Quat 300 sala 414, Vila Olmpia, So Paulo, Brazil, 04546042. Phone:+55 11 45042437. Fax: + 55 11 45042350.

1. Introduction Beginning in the 1980s, a vast body of theoretical literature on corporate finance, intended to analyze why companies use derivative instruments, began to develop. Some of the reasons cited include systems of progressive taxation, bankruptcy costs, and agency costs. Based on these models, the empirical literature has attempted to determine why companies use derivatives. The conclusions vary by sector, period, country, and econometric techniques used in the analysis, thereby making it difficult to generalize about the real motivations of companies using these financial instruments. Most of these models assume that the use of derivatives is intended to minimize firm-level cash flow volatility and that the interaction between imperfections in the financial markets and this reduction in companies cash flow volatility can generate gains to the firms that compensated the costs, leading to the use of derivatives to add value to firms. Thus, the use of derivatives can be considered as hedging and a part of risk management activities of the firm. However, the recent global financial crisis shows that it is possible for companies to use derivatives for reasons other than minimizing the volatility of their cash flow. In many countries, especially emerging countries such as Brazil, Poland, and Mexico, several companies reported severe financial losses directly after the devaluation of local currencies. Such losses were attributed to the use of derivatives for speculative purposes. Evidence about companies speculation has been presented by Geczy et al. (2007), Adam and Fernando (2006), Bodnar, Hayt and Marston (1998), and Dolde (1993), among others. The literature discusses different factors that may motivate companies to speculate. Stulz (1996) argues that speculation can result when firm executives believe that they have a comparative informational advantage compared to the market. Such an advantage, it is presumed, will allow the firm to make better decisions and obtain higher gains. Stulz (1996) also discusses the possible relationship between speculation and financial distress. Companies with less potential for financial distress can take more risks because they are less vulnerable to the effects of possible bad outcomes. In contrast, companies in distress might also be able to speculate more because gains will benefit shareholders and losses will affect bondholders. Stulz (1996) argues that by reducing risk, these companies may reduce the possibility of good outcomes that would enable them to improve their status. Finally, the author argues that firm compensation schemes can lead managers to speculate when their compensation packages reward

them for such behavior. In Campbell and Kracaw (1999) and Adam, Dasgupta, and Titman (2007), speculation is the optimal outcome in equilibrium for firms with financial constraints and a profit function that is convex in investment. These authors argue that factors such as a firms scale of operations, growth opportunities, liquidity, and financial distress will influence whether it speculates. Some studies involve attempts to empirically determine firm motivations for speculation. Beber and Fabbri (2011) used data from a panel of North American nonfinancial companies from 1996 to 2001 to show that companies adjust their derivative positions according to past movements in the exchange rate. This idea is consistent with the argument that firms select their positions according to their perspective on the exchange rate trajectory. The authors show that companies led by CEOs who have MBA degrees, are younger, and have less prior experience tend to speculate more. They ultimately conclude that their results are consistent with the theory that confident managers take more risks. Aabo et al. (2010), using a survey targeted to Danish companies, found that the involvement of non-financial departments in risk management increases the probability of speculation by companies. In their paper, the authors suggest that firm size and international involvement also exert a positive influence on the probability that these companies speculate. Geczy et al. (2007), using a survey targeted toward North American companies, have observed that risk management is the main reason why companies use derivatives. However, they also suggest that once the fixed costs of hedging are paid, firms will extend their positions based on the assumption that they can generate profits (but not increase risk) using derivatives due to a comparative informational advantage in the market in question. For example, the authors show that companies with revenue in a foreign currency or with subsidiaries overseas have a greater probability of speculating in the exchange rate market because they believe themselves to have an informational advantage with regard to the exchange rate; they have specific expertise in this market and believe that they can profit from this increased exposure. Furthermore, the authors indicate that a relationship exists between corporate governance and the use of derivatives of speculative purposes. The authors suggest that a firms compensation system can persuade a manager to speculate but that internal control systems can limit potential abuses.

The objective of this study was to analyze the factors that lead companies to speculate in the currency derivatives market using data from publicly traded Brazilian companies from 2007 to 2009. This study is innovative in several ways. First, this study focuses on a different economic environment than previous studies focusing on developed countries. Firms in emerging countries face a larger foreign exchange exposure and more volatile exchange rates than do developed countries, which make risk management activities more important for these companies. Unlike other studies, this study has identified companies as speculators using data from firms yearly balance sheets indicating their use of derivatives and foreign exchange exposure. After the global financial crisis, the Brazilian legislature demanded greater transparency. Not only must the value and position of all derivatives used by a company be declared, but the degree of exposure must also be accurately reported. The use of such data gives this study an advantage over previous ones, which identified companies as speculators either through surveys (Geczy et al. (2007), Aabo et al. (2010)) or through indirect proxies for speculation (Beber and Fabbri (2011)).2 The results show that a significant number of companies speculated in the derivatives market and that this number varied in the established period. In 2008, out of the 98 companies that used derivatives, 38 (38.7%) speculated in the exchange rate market. In 2009, out of the 76 companies that used derivatives, 16 (21.0%) were classified as speculators. Two types of speculators are identified: companies that significantly increased the volume of derivatives used in the period but that used them in accordance with their exchange rate exposure and companies that adopted positions contrary to their exchange rate exposure. The first group consists of companies with positive exchange rate exposure (net exporters) that increased their short position in the exchange rate market, whereas the second is composed of companies with negative exchange rate exposure (net importers or net debtors) that also hold a short position in the exchange rate derivative market. Both groups have tried to obtain gains through the continuous process of domestic currency appreciation, as evidenced by their short position in the foreign exchange market. These results are consistent with those found by Beber and Fabbri (2011), which indicate that companies use past movements in the exchange rate to
2

Geczy et al. (2007) argue that in the USA, information on the use of derivatives presented in firms yearly balance sheets is not sufficient to identify speculation.

determine their expectations regarding future movements and then they take positions in the foreign exchange market based on their view of the exchange rate trajectory. The study shows that companies that allegedly have some type of informational advantage in the foreign exchange market exporters and companies with foreign currency denominated debt are more likely to speculate. No other theory on why companies speculate is able to explain their behavior. This study is divided as follows. Section 2 presents the data used and the methodology proposed for classifying the companies as speculators or hedgers. Section 3 presents the econometric analysis. The last section presents the conclusion. 2. Data and Methodology for Classification Two data sources were used in this research: Bloomberg and yearly firm balance sheets. Bloomberg provides stock market data and accounting data for all Brazilian publicly traded non-financial companies. The data on firms use of derivatives and accounting exchange rate exposure were collected directly from their yearly balance sheets. Information on derivatives and foreign-currency-denominated debt is available in the explanatory notes on firms yearly balance sheets. The total amount of foreigncurrency-denominated debt is located under loans and financing, and the information on derivatives is located under financial instruments. The study includes data from a sample of 200 non-financial companies in 2007-2009. For companies with subsidiaries also publicly traded, consolidated balance sheets were used.3 This study is different from those previously mentioned partially because Brazilian firm data on the use of derivative became available in 1996.4 However, after the global financial crisis of 2007-2008 and the problems that arose in relation to the use of derivatives (with substantial losses accrued by some companies), the CVM (Comisso de Valores Mobilirios), a Brazilian regulatory authority, released statement 475/08 in combination with resolution 566/08, which both made the dissemination of data regarding derivatives more transparent and facilitated the analysis of firms accounting exchange rate exposure. CVM demands companies to use specific explanatory notes to disclose qualitative and quantitative information on all of their financial instruments, whether recognized as asset or liability.

It is thus assumed implicitly that companies belonging to the same conglomerate should maintain a common financial policy. 4 Securities and Exchange Commission (CVM), Normative Statement 235/1995.

According to the legislation, the explanatory notes on the use of financial instruments must enable users to evaluate their relevance for the firms financial position and overall results. This statement is especially true for derivative financial instruments. Businesses are required to provide tables that display the quantitative information on the use of derivatives and the risks associated with each instrument. Also based on statement 475/08, companies must now provide a sensitivity analysis chart for each type of market risk deemed relevant by the administration, with data for the financial instruments to which the entity is exposed, including all derivative financial instruments. This chart must identify the types of risk that can result in material losses for the company, including operations with derivative financial instruments; identify the methods and assumptions used in preparing the sensitivity analysis; define the most likely scenario in assessing the management (in addition to two scenarios that, if they occurred, would yield in adverse results for the company); and estimate the impact of those scenarios on the fair value of the financial instruments operated by the company. For the operations with derivative financial instruments, the company must disclose the object (the element being protected) and the derivative financial instrument used to indicate the net exposure of the company in each of the scenarios isolated. Thus, it is possible to use firms explanatory notes to obtain the total volume of the derivatives separated by type of instrument and firm position (short or long). Furthermore, it is possible to determine the foreign exchange rate exposure of the firm and, consequently, to calculate the firms net accounting exposure. Table 1 presents year-by-year summary statistics indicating the number of users of derivatives, their distribution by derivative instrument, the total notional amount of the derivative instruments used by the firms and the net position of the firms. The data shown in table 1 indicate temporal variation in the use of derivatives when both the number of firms that use derivatives and the total notional amount are analyzed. In 2007, 67 firms (38.5%) used some type of derivative; this number increased to 98 firms (49.0%) in 2008 and fell to 68 firms (34.0%) in 2009. A similar pattern can be observed in table 1 in relation to the total volume of derivatives (notional) regardless of the normalization used. Between 2007 and 2008, there was an increase in the volume of derivatives used by the firm. The data for 2009 (after the financial crisis of 2008) show

a reduction in both the number of firms that use currency derivatives and the total (notional) amount used by the firms. The data presented in table 1 indicate a difference between international trends and the Brazilian case. Among Brazilian firms, swaps are the most widely used derivative because the devaluation of the domestic currency is seen as a risk for firms that therefore hedge themselves through swaps.5 (Swaps are the most appropriate derivative when firm exposure takes place over a longer, predetermined period, as when companies have foreign-currency-denominated debt.6) An econometric analysis, to be presented later, confirmed this motivation for the use of derivatives by Brazilian companies. Table 1 also shows the net position of firms in the derivatives market. This value is calculated using the difference between firm long and short positions in US$ in the various types of derivatives. Negative values indicate a short position in US$, whereas positive values indicate the opposite. Again, a pattern has emerged from the analysis. Between 2007 and 2008, there was an increase in the net short positions of Brazilian firms in US$. The ratio between the net position of firms and their total assets rose from 2.34% in 2007 to 3.20% in 2008. In addition, the number of firms with net short positions increased from 23 in 2007 to 44 in 2008, whereas the number of companies with net long positions in US$ stagnated at 54 during 2007-2008. In 2009, after the onset of the global financial crisis and the bankruptcy of the Lehman Brothers Bank in the U.S., there was a change in the average net position of companies with a fall in the number of companies with net short positions; this number fell from 44 in 2008 to 26 in 2009. In short, the data presented in table 1 indicate, as discussed by Beber and Fabbri (2011) that firms adjust their derivatives positions in tandem with the market and based on past changes in the exchange rate. Figure 1 shows the evolution of the R$/US$ exchange rate during this period. Beginning in 2004, the exchange rate exhibited continuous appreciation, varying from 3.12 R$/US$ in May 2004 to 1.56 R$/US$ in July 2008. With the financial crisis, the exchange rate depreciated over 50% in one

Rossi (2007) has already presented these results for the period from 1996 to 2007. For a discussion of the foreign exchange derivatives market in Brazil, see Oliveira and Novaes (2007). 6 Allayannis et al. (2003) show that the use of foreign-currency-denominated debt in emerging countries is a result of capital structure decisions and not of risk management.

semester, reaching a value of 2.37 R$/US$ in February 2009. As the crisis cooled, it started to increase in value, reaching its pre-crisis rate only in 2011. In an environment marked by the continuous appreciation of the domestic currency (as shown in figure 1) and the strong performance of the Central Bank of Brazil in buying international reserves7, companies would have been able to take positions in the derivatives market that would have allowed them to profit from the appreciation process by predicting that it would continue. Stulz (1996) and Geczy et al. (2007) make a similar suggestion. The onset of the financial crisis in 2008, which was an exogenous external shock, corresponded to a reversal in the trajectory of exchange rates that led companies to reassess their positions in the derivatives market. 2.1 Identification of speculation Distinguishing between firms that use derivatives for speculative purposes and those that use them exclusively for hedging is the first step in this analysis of the motivations of companies to speculate. Ideally, the speculation by firms should be identified using an optimal hedging model. Deviations from the optimal hedge ratio may be seen as indication that companies would be speculating in the derivatives market. Beber and Fabbri (2011) thus construct a proxy for speculation based on a regression of the derivative holdings used by the company on the variables that the authors see as key reasons for hedging. The authors consider companies with more volatile deviations to be those who use derivatives differently than is recommended according to the fundamentals of hedging. These companies, in turn, are also seen as more likely to be speculating in the derivatives market. Unfortunately, there is no consensus regarding the optimal model for evaluating hedging. Furthermore, reasons for hedging should also be correlated with reasons for speculation, which makes it difficult to identify firms that speculate. The paper will use a model similar than the model used by Beber and Fabbri (2011) to verify the robustness of the results. To resolve the problems mentioned above, the majority of studies identify speculative firms using surveys. Both Geczy et al. (2007) and Aabo et al. (2010) base
7

From January 2004 to December 2009, the Brazilian international reserves rose from US$ 53.2 billion to US$ 239.0 billion through the purchase of foreign currency by the Central Bank. This accumulation of reserves served as a implicit government guarantee to firms that sudden fluctuations in exchange rates would be avoided. Thus, companies would have been willing to increase their exposure. For the details of the theory of implicit guarantees in a fixed exchange rate regime, see Burnside, Eichembaum and Rebelo (2001), Schneider and Tornell (2003).

their classification of speculation on surveys8, posing the following questions to companies: How often does your view on exchange rates lead you to 1) change the timing of hedging; 2) change the volume of hedging; 3) actively take a position in the derivatives market. Stulz (1996) uses the term "selective hedging" to refer to the strategies of managers who incorporate their market views into their hedging policy. The first two questions thus consider this form of speculation. Geczy et al. (2007) employ a more restricted view of speculation. The authors consider firms as speculators if they respond positively to the last question, arguing that this type of speculation is the type that should be of most concern to regulators. In contrast, Aabo et al. (2010) consider all questions as proxies for speculation. The use of surveys for this purpose has potential drawbacks, such as selection bias and low response rates, among others. Furthermore, these studies cannot clearly indicate the position of companies in the foreign exchange market because they do not consider the volume of derivatives; they analyze the intentions but not the real position of the companies in the derivatives market. Therefore, this study tries to identify a company as speculator directly based on data from its annual balance sheets instead of employing the methods already in use. In this way, this study is similar to that of Oliveira and Novaes (2007), which was the first to try to identify the speculation by firms using only the data available from annual balance sheets. Initially, we calculate firms accounting exchange rate exposure of based on the difference between their total revenue in a foreign currency and the sum of their expenses and debt in a foreign currency. Foreign currency revenue includes sales, cash assets, income from overseas subsidiaries and other foreign currency revenues. Foreign exchange expenses include those associated with imports and payments to suppliers and others. The net position in derivatives can be calculated using the difference between the long and short positions expressed in US$. Finally, we compare companies exchange rate exposure and their net derivative position. Two types of speculators are identified. A firm is classified as using derivatives for speculative purpose if, in the same year, it had a net position in the derivatives market that was the opposite of what it would have needed to hedge its

Geczy et al. (2007) use data collected by the Wharton survey of financial risk management.

exchange rate exposure. Essentially, net exporters (with positive exchange rate exposure) that hold long positions in dollars and net importers (with negative exchange rate exposure) that hold short positions in dollars. Furthermore, companies without foreign exchange exposure that take positions in the derivatives market are also considered speculators (SPEC1). A second group of firms that speculate is identified if these companies take positions in line with their foreign exchange exposure during one year but significantly increase their exposure in the derivatives market relative to that of the previous year without a proportional increase in their foreign exchange exposure in the same period. This definition is more consistent with the definition of selective hedging provided by Stulz (1996). (SPEC2) Analyzing firms this way requires researchers to determine an ad-hoc value for changes considered speculative. The study initially considered a change over 30% in the total net notional amount to be indicative of speculation.9 Companies that initiated the use of derivatives but did not continuously use them in subsequent years are also considered speculators. We performed several exercises to verify the robustness of this definition. 2.2 Reasons for speculation and characteristics of the firm The literature shows different reasons why companies speculate with derivatives. Several factors help to create a relationship between firm size and speculation. If transaction costs are important and there are economies of scale for speculation, large companies will be more likely to speculate in the derivatives market. However, if the size of a company is a proxy for financial constraints faced by the firm, as Campbell and Kracaw (1999) and Adam, Dasgupta and Titman (2007) indicate, then more constrained (smaller) companies should be expected to speculate more than large companies do. Adam, Fernando and Salas (2007), for example, analyzed data from companies in the gold-mining sector and found a negative relationship between speculation and firm size. The variable size (SIZE), calculated using the log of total assets of the companies, is included so that this relationship can be studied.

Beber and Fabbri (2011) found that 96% of companies in the sample vary the total amount of derivatives used at least 5% every year and two-thirds vary at least 30%. In the study, it was observed that between 2007 and 2008, the companies increased their position in derivatives by an average of 20.8%. Between 2008 and 2009, there was an average decrease of 2.27%.

Informational advantages in the foreign exchange market can be important in getting firms to speculate. Stulz (1996) argues that speculation can occur when a company believes that it enjoys a comparative informational advantage over the rest of the market and thus is in a position to obtain gains. Empirically, Geczy et al. (2007), Aabo et al. (2010), and Beber and Fabbri (2011) confirmed this hypothesis, establishing a relationship between the international orientation of a company and its probability of speculating on the foreign exchange market. We then use the ratio between foreign sales and total sales (Foreign Sales), the ratio between foreign-currency-denominated debt and total debt (Foreign Debt) and a dummy variable if the company had overseas subsidiaries (Foreign Operations) to test this hypothesis. Stulz (1996) discusses the possibility that there is a relationship between speculation and financial distress. According to the author, companies with low bankruptcy risk should be more likely to speculate because it will be easier for them to absorb the negative outcomes that may result from speculation. The author also suggests that the principal-agent relationship may lead managers of companies in distress to be more likely to speculate because this action will benefit shareholders, whereas risk management will only reduce the likelihood of good outcomes that improve the companys situation. The ratio between the book value of long-term debt to total assets (Debt Ratio) can also be used to analyze this relationship. Campbell and Kracaw (1999) and Adam, Dasgupta, and Titman (2004) suggest that speculation may be linked to the existence of a convex investment function. Thus, companies with good opportunities for growth but low short-term liquidity and high external financing costs should be more likely to speculate with derivatives. The market-to-book ratio (Growth) and the ratio between current assets and current liabilities (Quick) also played a role in these analyses as proxies of growth opportunities and liquidity. Companies may see speculation simply as increasing firm risk. According to Geczy et al. (2007), riskier companies will be more prone to speculate. We add the volatility of daily returns of companies as a measure of risk-taking behavior.10 Geczy et al. (2007) show the existence of a relationship between corporate governance and the use of derivatives for speculation. The authors observe that companies with lower levels of governance are more likely to be speculate but that

10

In the study, the volatility of daily returns is normalized by the volatility of market returns (Ibovespa).

internal control systems specific to the use of derivatives can limit potential abuses. Lel (2009) analyzed the impact of internal and external corporate governance structures on the use of derivatives in companies from several countries exposed to foreign exchange risk. The results of the study suggest that companies with strong governance use derivatives to mitigate their exposure to changes in exchange rates, whereas companies with weak governance tend to use derivatives in a manner that risks harming the company. In a related study, Allayannis et al. (2009) used a sample composed of companies from several countries to demonstrate that risk management policies add value to firms with strong governance regimes. Thus, in this study, a proxy variable was used for firm governance levels. The So Paulo stock exchange groups firms according to different governance levels, from lowest to highest: new market, level 1, level 2, traditional, and organized counter. The study then uses this classification system to analyze the relationship between governance and speculation. Finally, Beber and Fabbri (2011) show that past movement in the foreign exchange rate is an important factor in companies decisions to speculate. Thus, this study also used time dummies to analyze the likelihood that macroeconomic factors common to all companies would affect their decision to speculate. 2.3 Derivatives User, Hedger and Speculator Profile Table 3 illustrates the profiles of companies that use derivatives and those classified as hedgers and speculators. The results found in table 3 indicate the extent of speculation among Brazilian companies. In 2008, out of 98 companies that used derivatives, 38 (38.7% of derivatives users) speculated in the foreign exchange market. In 2009, there was a decrease in the numbers, 16 (21.0%) out of 76 derivative users were classified as speculators. Of the 38 speculors in 2008, 16 took positions that were unexpected whether it uses derivatives for hedging reasons given their foreign exchange exposure, and 22 markedly increased the volume of their derivatives without a proportional increase in their foreign exchange exposure. Of the 16 speculators in 2009, 11 fell into the first category and 5 into the second. The number of hedgers (derivative users not classified as speculators) was stable both in 2008 and in 2009; 60 companies were classified as hedgers. One of the advantages that this study has over previous studies is that it is possible to use these study data to examine the net position of the derivatives of these firms and their foreign exchange exposure over time. The data presented in table 3 show that hedgers exhibit negative foreign exchange exposure and are, on average, long in US

dollars. This evidence is consistent with the fact that swaps are the derivative most widely used by Brazilian companies, which have used it as a form of hedging their foreign exchange-denominated liabilities. On average for the period, the speculators exhibit positive foreign exchange exposure and net short positions in US$. That is, they are net exporters that bet on the appreciation of the domestic currency, hoping to obtain gains via their positions. Interestingly, this market view is common to the two types of speculators. The companies classified as speculators because they have taken positions that are inconsistent with their exposure (SPEC1) show negative figures for their average foreign exchange exposure (net importers) and hold short positions in dollars, with a net position in derivatives similar to that of companies classified according to the second classification (SPEC2) but with positive foreign exchange rate exposure (net exporters). Similarly, in an analysis resembling this study but that only considered the first type of speculators, Oliveira and Novaes (2007) discovered that companies with positive foreign exchange rate exposure (net exporters) were long in dollars in 2002: they were betting on the depreciation of the domestic currency. The authors argued that this phenomenon happened because of the extraordinary volatility of the domestic currency during that year, which in turn was caused by uncertainty resulting from the election of a new government. Table 3 also makes three types of comparisons among the companies in the sample. Initially, the data compare companies that use derivatives with all of the companies in the sample. Next, a comparison is made between the companies classified as hedgers and speculators. Finally, the two types of speculators are compared. The asterisks in the table illustrate the results of a test of means. The null hypothesis is that the difference between the mean of the two groups is null. The asterisks show where the null hypothesis is rejected. The results presented in table 3 show that derivative users are larger than the non-users. This finding confirms the existence of transaction costs in the use of derivatives. The data also confirm that greater international integration, as represented by a larger fraction of revenue and debt held in a foreign currency and by the operation of overseas subsidiaries, led these firms to use derivatives. The data also indicate that a positive relationship exists between the possibility of financial distress and the use of derivatives given that companies that use derivatives have higher debt ratios than those that do not.

The data in table 3 show that companies that use derivatives are less risky than those that do not because they have less volatile returns than do non-users and users present better governance than non-user. The data in table 3 show the differences between the companies that use derivatives for speculative purposes and those that use it for hedging. Considering the whole period, the results indicate that speculators are smaller than hedgers. This finding may indicate that financial restrictions may play a role in firm decision to speculate. As discussed by Stulz (1996), companies that are believed to have an informational advantage in the exchange market should tend to speculate more. Table 3 confirms this theory; speculators exhibit a greater fraction of foreign-currency-denominated revenue and debt and finally speculators present more volatile returns, indicating these firms are riskier than hedgers but this result is not robust for 2009. In 2008, the results also indicate that companies classified as speculators are more liquid than those that use derivatives for hedging, but this fact is not robust for the whole period. The data presented in table 3 confirm the differences between the two types of companies classified as speculators. The companies that took an active position in the derivatives market (SPEC1) have a higher proportion of their debt denominated in foreign currency, a higher debt ratio and growth opportunities. In addition, the companies that increased their exposure (SPEC2) have a greater proportion of their revenue in a foreign currency. 3 Empirical Analysis This section presents the empirical results of the analysis of derivative use by Brazilian companies. In all regressions, a logit model was estimated; the dependent variable varied according to the regression performed. In addition to the characteristics of the companies, sectoral and time dummies are included in the regressions. The results shown in table 4 confirm that transaction costs play an important role in the use of derivatives by companies. The results indicate that the likelihood that a company will use derivatives increases with the size of the company. The results also show a positive relationship between the use of derivatives and foreign currency revenue and debt, confirming that the greater the firms exchange exposure, the greater its likelihood of using derivatives is. The results show that companies with greater growth opportunities are more likely to use derivatives, as suggested by Froot et al. (1997). No other theory seems to explain the Brazilian firms decision to use derivatives.

Interestingly, the sign of the dummy for 2008 is positive and significant, implying that macroeconomic factors played a role in the use of derivatives by companies. Some sectoral dummies (not shown) also had a statistically significant impact in the use of derivatives by the companies. The results presented in table 4 show that in the period under analysis, only the proxy variables for company size and the ratio of foreign-currency-denominated debt to total debt were robust to the likelihood that a company would be classified as a hedger. This finding indicates that when Brazilian companies use derivatives as a hedge instrument, they use them to protect their liabilities from exchange rate fluctuations. In turn, these results confirm why swaps are the most widely used derivative, as also found by Rossi (2007). Although they are only significant for 2008, the results shown in table 4 show a negative relationship between a companys use of derivatives for protection and its revenue in a foreign currency, indicating that the exporters paid close attention to market timing in the derivatives market in 2008. There is also(weak) evidence that growth opportunities are important in the likelihood of a firm being a hedger. Table 5 clearly indicates why companies speculated in the derivatives market. Stulz (1996) argues that companies can speculate in the derivatives market if they believe they have some type of informational advantage that allows them to gain from speculating in the market. The results shown in table 5 indicate that companies with revenue in foreign currency and with foreign-currency-denominated debt show a higher probability of speculating in the foreign exchange market. In addition to indicating company characteristics, the results show that the common macroeconomic factors represented by the dummy variable for 2008 also affected the decision to speculate. The results shown in table 5 indicate that the time dummy for 2008 is positive and statistically significant. The results for the two types of speculators in table 5 show that companies with a higher ratio of foreign currency debt to total debt were more likely to take positions that were inconsistent with their exchange exposure (SPEC1) and that companies with a higher ratio of foreign currency revenue to total revenue were more likely to increase their positions in line with their exchange exposure but beyond its past values (SPEC2). Macroeconomic effects (as represented by the dummy for 2008) were significant for this group, which means that the possible effect of appreciation was stronger in this group.

In short, in the case of Brazil, continuous domestic currency appreciation made companies take speculative positions in relation to the future trajectory of the exchange rate, probably because they believed that they had superior knowledge of the evolution of the exchange rate due to their international integration. As table 5 illustrates, the results for all other possible reasons for the companies to speculate are statistically insignificant and are not robust.11 3.1 Robustness The estimations shown in tables 4 and 5 were performed using all of the companies present in the sample. Usually, the analyses are performed only with companies that have some type of exchange rate exposure. In the Brazilian context, this separation is more difficult because even if the company does not have some form of foreign currency revenue or debt, it may have foreign competitors or contracts adjusted by the dollar or using type of price index highly correlated with the dollar. 12 Therefore, it was necessary to perform the analysis with all publicly traded companies. The results presented in the first part of table 6 consider only companies with assets and liabilities that were in some way exposed to exchange rate movements, therefore companies without foreign revenues, debt or subsidiaries were excluded. The results shown in table 6 are similar to those found previously; however, liquidity seems to affect the probability that companies will use derivatives for hedging. The results show that liquidity is complementary to the use of derivatives for hedging, as shown by the positive and significant relationship between the two variables. The second robustness exercise with results shown in table 6 breaks firm decision-making down into two stages. In the first stage, the companies decided whether to use derivatives; next, assuming that they have decided to use them, they would determine whether to use them for speculative purposes. Thus, in the second stage of this study, the regressions were conducted only for the companies that had used derivatives. The main results are unchanged, showing that companies with higher ratios of foreign currency revenue to total revenue and foreign-currency-denominated debt to total debt are more likely to speculate in the market with derivatives. Contrary to previous results, the results derived from this exercise indicate the existence of a relationship between the firm size and speculation. Of those companies that decided to
11

Adam, Dasgupta and Salas (2007) estimate a non-linear relationship between speculation and financial distress. A quadratic term for the debt ratio was also included in the regression, but it was not found to be significant in any specification. 12 This occurs with companies in the electric sector, for example.

use derivatives in the period under study, smaller companies were more likely to speculate. These results are consistent with those of Adam, Fernando, and Salas (2007), who considered such findings an indication of a relationship between financial restrictions and speculation. Finally, in this study, it was necessary to analyze the robustness of the definition of speculation used here. A change in the volume of derivatives was considered the definition of the second type of speculator. The results (not shown) of an analysis using several levels of such change indicate no significant changes in the main results given variation in derivatives volume between 20 and 50%. Therefore, it was concluded that the results were robust to the variation involved in the study. Following Beber and Fabbri (2011), a two steps procedure can be used to identify the speculation by the firms. Initially, the ratio of the total notional amount of derivatives to total assets should be regressed against the variables that can explain why companies use derivatives for hedging. Deviations are considered evidence that companies change their position in derivatives based on factors other than hedging. Companies with more volatile deviations are more likely to be speculating in the foreign exchange market. The last two columns of table 6 use this methodology. The results indicate that larger companies with a greater proportion of their revenue and debt in a foreign currency, which also enjoy greater growth opportunities, less liquidity and better governance, have a higher ratio between the total notional amount of derivatives and total assets. In accordance with previous results, the second step indicates the positive relationship between speculation, external sales and foreigncurrency-denominated debt and indicates that none of the other reasons can explain speculation by these companies. 4. Conclusion This study has analyzed the use of derivatives for Brazilian companies from 2007 to 2009, which includes the period in which the global financial crisis occurred. It is essential to analyze this period because large financial losses afflicted non-financial companies in several emerging countries due to the speculative use of currency derivatives. The study contributed to the literature in several ways. First, it focuses on one country, Brazil, where the question of risk management is extremely important for companies because most of them have some form of exposure to exchange rates and exchange movements tend to be more pronounced than in developed countries. In

addition, the disclosure rules established by regulatory agents enable to obtain not only the total (notional) amount of derivatives used by a particular company but also the net position of the company in the foreign exchange derivatives market and the accounting exposure of the company to exchange rate fluctuations. Therefore, unlike other studies, this study distinguishes between companies that use derivatives for speculative purposes and hedging purposes, using data from their annual balance sheets to evaluate time-based changes in their position in the derivatives market and their exchange exposure. The results show that a considerable number of companies speculated in the derivatives market in the period under analysis. In 2008, approximately 38.7% of companies that used derivatives were classified as speculators. Furthermore, the study confirms that in 2009, the proportion of speculative companies among those that used derivatives fell to 21.0%. The analysis of the net position of companies in currency derivatives and their foreign exchange exposure over time made it possible to identify two types of speculators: companies that significantly increased the volume of derivatives that they used in this period but that used derivatives proportional with their exchange exposure and companies that adopted positions incompatible with the aim of limiting their foreign exchange exposure. The first group includes companies with positive exchange rate exposure (net exporters) that increased their short position in the foreign exchange derivative market, whereas the second is composed of companies with negative exchange rate exposure (net importers or net debtors) that also chose a short position in the foreign exchange derivative market. The both groups tried to obtain gains through the continuous process of domestic currency appreciation, as evidenced by their short position in the exchange market. These findings corroborate the results found by Beber and Fabbri (2011) indicating that companies use past exchange rate movements to form their expectations about future movements and that they take positions in the exchange market based on their vision of the exchange rate trajectory. The study shows that companies that allegedly have some type of informational advantage in the exchange market (exporters and companies with foreign-currencydenominated debt) are more likely to speculate and that no other theory about the motives that lead companies to speculate is capable of explaining this behavior.

Why companies believe that they can obtain gains by speculating in the derivatives market is still an open question. Even the most internationalized companies do not generally have sufficient expertise regarding the market to achieve significant gains. This generalization is even truer in the foreign exchange rate market; even the literature that tries to predict its movements indicates that no model is good enough in all periods to predict the exchange rate. Whether for behavioral reasons on the part of managers as suggested by Beber and Fabbri (2011) or because of implicit government or Central Bank guarantees to firms, some reason might encourage managers to think that they can obtain real gains by speculating. None of the existent theoretical models show to be useful to explain companies behavior and it seems to be a fruitful way for future research in the field. References Aabo, T., Hansen, M., and Pantzalis, T. Who pulls the trigger? Evidence on Corporate Currency Speculation. Aarhus School of Business, 2010. Adam, T., Dasgupta, S. and Titman, S. Financial Constraints, Competition, and hedging in Industry Equilibrium. Journal of Finance 62, 2445-2473, 2007. Adam, T. and Fernando, C. Hedging, Speculation and Shareholder Value. Journal of Financial Economics 81, 283-309, 2006. Adam, T., Fernando, C. and Salas, J. Why do firms hedge selectively? Evidence from the gold industry. Working paper, University of Oklahoma and National University of Singapore, 2007. Allayannis, G., Lel, U., and Miller, D. Corporate governance and the hedging premium around the world. Darden Business School Working Paper No. 03-10, 2009. Allayannis, G., Brown, G. and Klapper, L. Capital Structure, Foreign Debt and Financial Risk: Evidence from East Asia. Journal of Finance 58, 2667- 2690, 2003. Bodnar, G., Hayt, G., and Marston, R. 1998 Wharton Survey of Financial Risk Management by U.S. Non-Financial Firms. Financial Management 27, 70-91, 1998. Brown, G., Crabb, P., and Haushalter, D. Are Firms Successful at Selective Hedging?. Journal of Business 79, 2925-2949, 2006. Beber, A. and Fabbri, D. Who times the Foreign Exchange Market? Corporate Speculation and CEO Characteristics. University of Lausanne and FAME Working Paper, 2011. Bodnar, G.M., G.S. Hayt and R.C. Marston, 1998 Wharton Survey of Derivatives Usage by U.S. Non-Financial Firms, Financial Management 27, 70-91, 1998.

Brown, G.W., Crabb, P. and Haushalter, D. Are Firms Successful at Selective Hedging? Journal of Business 79, 29252949, 2006. Burnside, C., Eichenbaum, M. and Rebelo, S. Hedging and Financial Fragility in Fixed exchange Rate Regimes. European Economic Review 45, 1151-1193, 2001. Campbell, T. and Kracaw, W. Optimal Speculation in the Presence of Costly External Financing. In Corporate Risk: Strategies and Management, G. Brown and D. Chew, editors, Risk Publications, London, 1999. Dolde, W. The Trajectory of Corporate Financial Risk Management. Journal of Applied Corporate Finance 6, 33-41, 1993. Froot, K., Scharfstein, D., and Stein, J. Risk management, coordinating investment and financing policies. Journal of Finance 48, 1629-1658, 1993. Geczy, C., Minton, B. and Schrand, C. Why firms use currency derivatives. Journal of Finance 52, 1323-1354, 1996. Geczy, C., Minton, B. and Schrand, C. Taking a view: Corporate speculation, governance and compensation. Journal of Finance 62, 2405-2443, 2007. Lel, U. Currency hedging and corporate governance: a cross-country analysis. Indiana University Working Paper No. 858, 2009. Oliveira, F. and Novaes, W. Demand for Foreign Exchange Derivatives in Brazil: Hedge or Speculation. Central Bank of Brazil working paper series 152, 2007. Rossi, J. The use of currency derivatives by Brazilian companies: an empirical investigation. Brazilian Review of Finance 5, 205-233, 2007. Schneider, M. and Tornell, A. Balance sheet effects, bailout guarantees and financial crises. Review of Economic Studies 71, 883-913, 2004. Stulz, R. Rethinking risk management. Journal of Applied Corporate Finance 9, 824, 1996.

Figure 1 Trajectory of Exchange Rate R$ / US$

Figure 1 shows the trajectory of the exchange rate of R$ / US$ from January 2004 to September 2011. Source: Central Bank of Brazil.

Table 1 Summary Statistics of the Use of Derivatives by the Companies

Table 1 shows a summary by year of the use of derivatives by the companies. Notational / total assets is the ratio between the total notational value of derivatives and the book value of assets of the company. Notational / external sales is the ratio between the total notational value of derivatives and the total value of revenue in foreign currency. Notational/ Foreign Currency Debt is the ratio between the total notational value of derivatives and the total value of the debt denominated in foreign currency. Net Position/ Total Assets represents the ratio between the net total value of derivatives and the book value of assets. The net position is calculated from the difference between the total notational amount of long and short positions in US$. Long means the total number of companies with long positions in US$, and Short represents the total number of companies with short positions.

Swap NDF, Forward and Options Both None Notational / Total Assets (all firms) Notational / Total Assets (Only Derivative Users) Notational / External Sales Notational / Foreign Currency Debt Net Position / Total Assets Long Short

2007 57 34 14 123 4.39% 11.4% 53.0% 47.7% -2.34% 54 23

2008 63 55 20 102 6.19% 12.6% 61.4% 92.5% -3.20% 54 44

2009 51 43 16 122 3.44% 8.73% 60.9% 78.5% 0.24% 50 26

Table 2 Reasons for Speculation and Company Characteristics

Table 2 shows summary statistics for the main variables used in the study and the correlation between the different variables. Size is the log of the book value of the total assets of the company (R$ millions). Foreign Sales represents the ratio of foreign currency sales over total sales. Foreign Debt represents the ratio of debt denominated in foreign currency over total debt. Growth is the ratio of market value of assets over the book value of assets. Debt Ratio represents the ratio between the book value of longterm debt and the total assets. Quick Ratio is the ratio between the current assets and the current liabilities. Foreign Operations is a dummy that takes the value of 1 if the company has overseas subsidiaries, and Governance is a proxy variable for corporate governance, constructed according to the levels of governance adopted by the So Paulo Stock Exchange (BOVESPA). Volatility Stock Returns represents the ratio between the daily volatility of returns of the company and the volatility of the market index (Ibovespa). The statistics are calculated for all companies (N=200) from 2007 and 2009.

Size Mean Standard Deviation Size Foreign Sales Foreign Debt Foreign Operations Growth Quick Debt Ratio Governance Volatility Stock Returns 13.92 1.98 1.00 0.18 0.30 0.18 -0.13 -0.01 0.42 0.36 -0.37

Foreign Sales 10.7% 19.7% 1.00 0.40 0.22 0.06 0.07 0.13 0.10 -0.08

Foreign Debt 11.6% 16.3%

Foreign Operations 0.148 0.365

Growth 1.19 1.04

Quick 1.63 1.48

Debt Ratio 25.7% 20.6%

Governance 1.95 1.28

Volatility Stock Returns 2.33 1.99

1.00 0.11 -0.03 0.03 0.36 0.16 -0.10

1.00 0.00 0.01 0.07 0.19 -0.13

1.00 -0.03 0.02 0.09 0.05

1.00 0.02 0.12 -0.04

1.00 0.27 -0.20

1.00 -0.31

1.00

Table 3 - Company Profiles

Table 3 shows the different profiles of the companies. The column User is a comparison between companies that use derivatives and those that do not. Furthermore, table 3 compares the companies classified as hedgers and speculators and the two types of speculative companies (SPEC1 and SPEC2). Net Exposure represents the net exchange rate exposure of companies; its calculation is in the text. Net Derivative Position is the difference between the notational value of long and short positions in US$ of derivatives of the company. Size is the log of the book value of the total assets of the companies (R$ millions). Foreign Sales represents the ratio of external sales over total sales. Foreign Debt represents the ratio of debt in foreign currency over total debt. Growth is the ratio of market value of assets over the book value of assets. Debt Ratio represents the ratio between the book value of long-term debt and the total assets. Quick Ratio is the ratio between the current assets and the current liabilities. Foreign Operations is a dummy that takes the value of 1 if the company has overseas subsidiaries and Governance is a proxy variable for corporate governance, constructed according to the levels of governance adopted by the So Paulo Stock Exchange (BOVESPA). Volatility Stock Returns represents the ratio between the daily volatility of returns of the company and the volatility of the market index (Ibovespa). *, ** indicates statistically significant differences at 5% and 10%, respectively.

2008 -2009 User Number of Firms Net Exposure (US$ million) Net Derivative Position (US$ million) Size Foreign Sales Foreign Debt Growth Debt Ratio Quick Ratio Foreign Operations Governance Volatility Stock Returns 174 -189.6 +137.6 15.0* 0.166* 0.207* 1.02 0.332* 1.65 0.235* 2.33* 1.73* Hedgers Specul. 120 -361.4 +257.7 15.26* 0.098 0.162 1.06 0.329 1.61 0.250 2.34 1.61 54 +36.8 -131.1 14.52 0.316* 0.305* 0.926 0.339 1.74 0.203 2.30 2.03** SPEC1 27 -400.5 -208.5 14.48 0.252 0.400* 1.08** 0.381** 1.61 0.074 2.51 2.14 SPEC2 27 +326.7 -103.7 14.56 0.380* 0.210 0.76 0.297 1.88 0.333* 2.07 1.92 User 98 -189.1 +82.6 14.9* 0.172* 0.228* 0.846 0.337* 1.61 0.234* 2.34* 1.62*

2008 Hedgers Specul. SPEC1 SPEC2 60 -351.7 +227.9 15.32* 0.082 0.180 0.871 0.339 1.48 0.250 2.42 1.40 38 +67.69 -146.8 14.26 0.315* 0.304* 0.808 0.333 1.81** 0.211 2.21 1.97* 16 -373.2 -197.9 14.22 0.213 0.422* 0.775 0.411* 1.62 0.125 2.43 2.11 22 +388.3 -109.7 14.32 0.389* 0.218 0.852 0.276 1.94 0.272 2.04 1.87 User 76 -353.0 +207.2 15.2* 0.157* 0.179* 1.25 0.326* 1.71 0.236* 2.32* 1.88*

2009 Hedgers Specul. SPEC1 SPEC2 60 -371.3 +287.5 15.20 0.115 0.144 1.26 0.319 1.74 0.250 2.26 1.81 16 -85.3 -93.7 15.13 0.317* 0.309* 1.20 0.353 1.59 0.187 2.50 2.15 11 -440.2 -223.8 14.97 0.308 0.369* 1.42* 0.388 1.59 0.00 2.63 2.18 5 +55.6 -192.7 15.48 0.337 0.178 0.708 0.307 1.60 0.60* 2.20 2.08

Table 4 - Determinants of the Use of Derivatives and of Use of Derivatives for Protection
Table 4 shows the results of the regressions for the determination of reasons that lead companies to use derivatives and use them for hedging. All regressions are performed using a Logit model. For the regressions for derivatives users (User), the dependent variable takes a value of 1 if the company has used derivatives and 0 if it has not. For the hedger regressions, the dependent variable takes on a value of 1 is the company is classified as a hedger and 0 if it is not. Size is the log of the book value of total assets of the companies (R$ millions). Foreign Sales represents the ratio of external sales over total sales. Foreign Debt represents the ratio of debt in foreign currency over total debt. Growth is the ratio of market value of assets over the book value of assets. Debt Ratio represents the ratio between the book value of long-term debt and the total assets. Quick Ratio is the ratio between the current assets and the current liabilities. Foreign Operations is a dummy that takes the value of 1 if the company has overseas subsidiaries, and Governance is a proxy variable for corporate governance, constructed according to the levels of governance adopted by the So Paulo Stock Exchange (BOVESPA). Time and Sectoral dummies are added to some regressions as indicated. Robust standard deviations are indicated between parentheses. *, ** indicates statistically significant differences at 5% and 10%, respectively. Panel is a Logit panel regression with random effects. Columns 2008 and 2009 only use data from these respective years.

User Size Foreign Sales Foreign Debt Growth Debt Ratio Quick Ratio Foreign Operations Governance Stock Returns Volatility 2008 2009 Sectoral Dummies N R2 Logit 0.761 (0.130)* 1.89 (1.02)** 6.56 (1.04)* 0.250 (0.104)* -0.053 (0.636) -0.098 (0.097) 0.440 (0.360) 0.129 (0.103) -0.045 (0.060) 0.733 (0.279)* 0.112 (0.276) Yes 600 0.388 Panel 2.21 (0.49)* 3.30 (1.16)* 12.4 (2.82)* 0.589 (0.302)** -0.136 (1.93) -0.441 (0.370) 0.976 (1.38) 0.383 (0.372) -0.262 (0.222) 2.04 (0.268)* 0.179 (0.507) Yes 600 2008 0.679 (0.245)* 3.65 (1.53)* 6.83 (1.92)* 0.250 (0.278) 0.537 (1.09) -0.065 (0.132) 0.443 (0.689) 0.252 (0.197) -0.075 (0.101) 2009 0.772 (0.186)* 1.61 (0.84)* 7.37 (1.94)* 0.266 (0.154)** -0.734 (1.24) -0.029 (0.138) 0.356 (0.642) -0.035 (0.171) 0.029 (0.116) Logit 0.644 (0.140)* -1.38 (1.20) 2.65 (1.06)* 0.299 (0.134)* -0.425 (0.775) 0.042 (0.067) 0.813 (0.476)** 0.061 (0.115) -0.091 (0.101) 0.056 (0.275) Panel 1.35 (0.47)* -3.88 (2.61) 5.33 (2.51)* 0.491 (0.389) -0.712 (1.78) 0.052 (0.242) 1.58 (1.11) 0.107 (0.303) -0.126 (2.31) 0.038 (0.450)

Hedger 2008 0.798 (0.229)* -3.26 (1.86)** 1.67 (0.47)* 0.523 (0.282)** -0.526 (1.21) 0.098 (1.32) 0.883 (0.694) 0.189 (0.179) -0.243 (0.145)** 2009 0.596 (0.178)* -0.251 (1.64) 3.81 (1.62)* 0.237 (0.150) -0.346 (1.08) 0.051 (0.081) 0.686 (0.625) -0.041 (0.163) -0.025 (0.128)

Yes 200 0.405

Yes 200 0.359

Yes 400 0.2768

Yes 400

Yes 200 0.338

Yes 200

Table 5 - Determinants of Speculation

Table 5 shows the result of regressions for the determination of the reasons that lead companies to speculate with derivatives. All regressions are performed using a Logit model. For the regressions for all of the companies classified as speculators (Speculation), the dependent variable takes on the value of 1 if the company has been thus classified and 0 if it has not. For the regressions with the types of speculation (SPEC1 and SPEC2), the dependent variable takes on the value of 1 if the company was classified as their respective classification and 0 if not. Size is the log of the book value of total assets of the companies (R$ millions). Foreign Sales represents the ratio of external sales over total sales. Foreign Debt represents the ratio of debt in foreign currency over total debt. Growth is the ratio of market value of assets over the book value of assets. Debt Ratio represents the ratio between the book value of long-term debt and the total assets. Quick Ratio is the ratio between the current assets and the current liabilities. Foreign Operations is a dummy that takes the value of 1 if the company has overseas subsidiaries, and Governance is a proxy variable for corporate governance, constructed according to the levels of governance adopted by the So Paulo Stock Exchange (BOVESPA). Volatility Stock Returns represents the ratio between the daily volatility of returns of the company and the volatility of the market index (Ibovespa). Robust standard deviations are indicated between parentheses. *, ** indicates statistically significant differences at 5% and 10%, respectively. Panel is a Logit panel regression with random effects. Columns 2008 and 2009 only use data from these respective years.

Speculation Logit 0.141 ( 0.137) 1.88 (0.89)* 4.90 (1.12)* -0.147 (0.343) 0.475 (1.14) -0.149 (0.145) -0.676 (0.579) 0.114 (0.194) 0.037 (0.091) 1.05 (0.405)* Yes 400 0.3872 Panel 0.019 (0.176) 2.78 (1.43)** 6.10 (1.85)* -0.317 (0.318) 0.502 (1.37) -0.225 (0.220) -0.829 (0.736) 0.215 (0.219) -0.0109 (0.149) 1.43 (0.55)* Yes 400 2008 -0.043 (0.221) 2.73 (1.27)* 4.53 (1.41)* -0.495 (0.465) 1.06 (1.40) -0.0808 (0.168) -0.706 (0.837) 0.153 (0.277) 0.045 (0.114) 2009 0.299 (0.145)* 0.992 (0.49)* 6.27 (1.89)* 0.016 (0.480) -0.933 (2.58) -0.376 (0.280) -0.470 (0.808) 0.172 (0.329) 0.072 (0.198) Logit 0.154 (0.215) -1.72 (1.34) 6.49 (1.41)* 0.244 (0.314) 0.111 (1.77) -0.231 (0.257) -0.686 (1.02) 0.205 (0.284) 0.095 (0.129) 0.116 (0.478) Yes 400 0.395

Type of Speculation SPEC1 Panel 2008 0.263 -0.088 (0.444) (0.304) -4.55 -2.91 (3.90) (2.03) 12.95 5.51 (5.45)* (1.53)* 0.548 -0.578 (0.539) (0.803) -0.512 2.92 (3.01) (2.15) -0.365 -0.332 (0.587) (0.471) -0.845 0.110 (1.96) (0.915) 0.555 0.158 (0.549) (0.430) 0.319 0.031 (0.356) (0.218) 0.565 (0.826) Yes Yes 400 200 0.395 2009 0.846 (0.316)* -1.57 (2.40) 11.57 (2.82)* 0.688 (0.492) -6.17 (6.37) -0.542 (0.382) 0.340 (0.358) 0.371 (0.604) 0.149 (0.312) Logit 0.158 (0.131) 4.21 (1.08)* 0.285 (1.14) -0.374 (0.271) 0.671 (0.990) 0.021 (0.102) -0.190 (0.642) -0.208 (0.199) -0.022 (0.112) 1.27 (0.488)* Yes 400 0.272 SPEC2 Panel 2008 0.158 0.021 (0.145) (0.203) 4.21 4.97 (1.19)* (1.46)* 0.284 1.14 (1.39) (1.18) -0.374 -0.496 (0.337) (0.441) 0.670 -0.496 (1.23) (1.24) 0.021 0.103 (0.141) (0.150) -0.190 -0.357 (0.584) (0.892) -0.208 -0.124 (0.198) (0.242) -0.022 -0.038 (0.134) (0.134) 1.26 (0.469)* Yes Yes 400 200 0.295 2009 0.388 (0.278) 5.26 (2.81)** 2.41 (4.26) -0.486 (0.481) 0.455 (0.451) -0.560 (0.432) -0.387 (1.08) -0.378 (0.521) 0.026 (0.174)

Size Foreign Sales Foreign Debt Growth Debt Ratio Quick Ratio Foreign Operations Governance Stock Returns Volatility 2008 Sectoral Dummies N Pseudo-R2

Yes 200 0.439

Yes 200 0.326

Yes 200 0.351

Yes 2000 0.336

Table 6 - Robustness Exercises

Table 6 shows the results of some robustness exercises. In (1), similar regressions to those presented in tables 4 and 5 are conducted, but only using the companies with exchange rate exposure present in the sample. In (2), the regressions are performed only in companies present in the sample that use derivatives. In (3), a similar method to Beber and Fabbri (2011) is used to identify firms that speculate. In the first step, the determinants of derivatives usage are analyzed. The independent variable is the ratio of total notational amount of derivatives divided by the total book value of assets. The volatility of regression residues is used as proxy for speculation. Size is the log of the book value of total assets of the companies (R$ millions). Foreign Sales represents the ratio of external sales over total sales. Foreign Debt represents the ratio of debt in foreign currency over total debt. Growth is the ratio of market value of assets over the book value of assets. Debt Ratio represents the ratio between the book value of long-term debt and the total assets. Quick Ratio is the ratio between the current assets and the current liabilities. Foreign Operations is a dummy that takes the value of 1 if the company has overseas subsidiaries, and Governance is a proxy variable for corporate governance, constructed according to the levels of governance adopted by the So Paulo Stock Exchange (BOVESPA). Volatility Stock Returns represents the ratio between the daily volatility of returns of the company and the volatility of the market index (Ibovespa). Time and Sectoral dummies are added to some regressions as indicated. Robust standard deviations are indicated between parentheses. *, ** indicates statistically significant differences at 5% and 10%, respectively.

Size Foreign Sales Foreign Debt Growth Debt Ratio Quick Ratio Foreign Operations Governance Stock Returns Volatility 2008 2009 Sectoral Dummies N R2

User 0.517 (0.138)* 1.40 (0.80)** 4.44 (0.97)* 0.236 (0.141)** 0.208 (0.780) 0.029 (0.121) 0.445 (0.382) 0.018 (0.120) -0.072 (0.066) 0.820 (0.308)* 0.151 (0.291) Yes 425 0.285

Hedger 0.482 (0.141)* -1.44 (0.90)** 0.675 (0.273)* 0.754 (0.287)* -0.311 (0.945) 0.329 (0.179)** 1.10 (0.54)* -0.051 (0.144) -0.082 (0.098) -0.154 (0.255)

(1) Speculation 0.139 (0.160) 1.59 (0.94)** 4.63 (1.15)* -0.444 (0.321) 0.408 (1.08) -0.147 (0.181) -0.680 (0.615) 0.015 (0.212) 0.016 (0.095) 1.08 (0.430)*

Spec 1 0.144 (0.228) -2.23 (1.48) 6.89 (1.53)* 0.030 (0.410) 0.272 (1.63) -0.257 (0.324) -0.419 (1.00) 0.037 (0.260) 0.043 (0.144) 0.041 (0.530)

Spec 2 0.145 (0.140) 3.91 (1.05)* -0.018 (1.12) -0.375 (0.291) 0.515 (1.06) 0.091 (0.152) -0.104 (0.693) -0.233 (0.210) 0.017 (0.092) 1.23 (0.480)*

Speculation -1.07 (0.335)* 4.98 (2.13)* 4.67 (1.74)* 0.084 (0.483) -1.13 (2.42) 0.300 (0.259) 0.344 (0.933) -0.186 (0.275) 0.505 (0.354) 0.847 (0.598)

(2) Spec1 -0.196 (0.256) -3.11 (1.76)** 5.53 (1.36)* 0.627 (0.429) 0.719 (1.97) -0.212 (0.379) -0.047 (1.18) 0.029 (0.306) 0.271 (0.134)* -0.312 (0.647)

(3) Spec2 -0.239 (0.165) 5.20 (1.18)* -3.61 (1.78)* -0.449 (0.417) -0.200 (1.38) 0.203 (0.194) 0.119 (0.643) -0.315 (0.189)** 0.042 (0.137) 1.02 (0.60)** First 0.030 (0.008)* 0.120 (0.063)** 0.497 (0.064)* 0.033 (0.012)* -0.055 (0.050) -0.015 (0.0084)** -0.037 (0.027) 0.025 (0.0080)* -0.0011 (0.0055) 0.029 (0.011)* 0.0013 (0.012) Yes 600 0.770 Speculation 0.0010 (0.0024) 0.0066 (0.0023)* 0.084 (0.025)* 0.011 (0.0083) -0.035 (0.026) -0.0007 (0.0017) -0.016 (0.011) 0.0036 (0.0045) 0.0023 (0.0018) 0.0065 (0.0084) 0.0044 (0.0095) Yes 600 0.085

Yes 282 0.288

Yes 282 0.365

Yes 282 0.395

Yes 282 0.234

Yes 174 0.479

Yes 174 0.384

Yes 174 0.252