1
Currency Risk Exposure, Hedging and Stock Return
An Empirical Investigation of Large US Chemical Companies
from 2008 to 2011
Thesis
Master of Science in Business Administration
Specialization: Finance
University of Groningen
Faculty of Economics and Business
2
Abstract
Scrutinizing the issue of currency risk, we examine the currency risk exposure and the effect of hedging on exposure and stock return for 58 large US chemical companies for the period from February 2008 to January 2012. In contrast to the literature, our analysis suggests significant impacts of exchange rate movements on stock returns. Inspecting the effects of hedging, we find evidence that suggests hedging significantly reduces exposure but only when using a particular model specification. The significant coefficients of hedging are, however, not found in other model specifications, raising questions as to whether hedging really does reduce currency risk exposure as suggested in the literature. Our analysis shows no significant direct effect of hedging on stock returns during this period. This raise questions as to what exactly is the hedging objective during this period.
JEL classification
F31, G12, G32
Keywords
3
I. Introduction
Since the breakdown of the Bretton Woods fixed exchange rate system in 1971, exchange rate changes have become an important risk factor for companies across the globe. In addition, companies and industries become increasingly globalized as their activities expand across borders and oceans. These effects combined leave firms more exposed to currency risk.
A widely held theoretical belief is that exchange rate changes are a major macroeconomic uncertainty that should significantly affect company performance and value, notwithstanding whether the company is domestic or international (see, e.g., Shapiro (1975); Levi (1994); Muller and Verschoor (2004c)). Yet, empirical evidence has found weak to non-existent influence of exchange rate changes on companies’ stock returns (see, e.g., Jorion (1990);Amihud (1993); Bodnar and Gentry (1993)). A possible explanation for this lack of empirical support is that companies are well-informed about their exposures and are taking measures to limit currency risk (Bartov and Bodnar (1994)). However, because the effects of exchange rate movements in the long-term are hard to determine, the effectiveness of these measures for future operations is questionable. Moreover, it is doubtful whether these measures are value enhancing for the firm because managers have their own incentives to hedge currency risk that are different from, and sometimes in direct contrast to, shareholders’ value maximization (Smith and Stulz (1985)).
4 The novel contribution of this thesis is in the construction of a hedging variable that indicates the degree of currency risk hedging of a company. Based on this variable, we examine the relationships between hedging and the firm’s currency risk exposure and stock return. To do this, we employ a dataset of 58 large US chemical companies for the period from February 2008 to January 2012. We choose large chemical companies located in the US because doing this offers a set of companies with likely high levels of foreign involvement, similar exposure characteristics, and good data availability, thanks to high information disclosure requirements in the US. The timeframe of the dataset is selected because it allows us to study the relationships as a consequence of the financial crisis and to obtain more up-to-date results.
Due to the Autoregressive Conditional Heteroskedasticity (ARCH) effects in the stock return time-series, GARCH(1,1) model is used. Estimated Generalized Least Squares (EGLS) period Seemingly Unrelated Regression (SUR) is used for the corresponding panel regressions. Simple Ordinary Least Squares (OLS) is used in the cross-sectional exposure regressions where we regress exposure against hedging and other control variables.
Our analysis shows significant impacts of exchange rate changes on stock returns, which are to some extent overwhelming compared to the effects from the market return factor. We also find evidence that suggests hedging significantly reduces exposure but only when using a particular model specification: the Fama-French three-factor model. The significant coefficients of hedging are, however, not found in other model specifications, raising questions as to whether hedging really does reduce currency risk exposure as suggested in the literature. No evidence is found on the direct positive effect of hedging on stock return during this period. This raise questions as to what exactly is the hedging objectives during this period.
In the next section, we give a literature review. After that, the theoretical model and methodologies are presented, followed by the empirical results and the discussion. The final section concludes this paper.
II. Literature Review A. A theoretical overview about hedging
5 the major macroeconomic uncertainties facing firms in open economies. Companies’ international transactions and translations are subjected to exchange rate movements that affect companies’ incomes, operating cash flows, and value (Choi and Prasad (1995)). In addition, similar to other macroeconomic factors, exchange rate volatility does not affect different companies in exactly the same way. The currency risk exposure of different companies is contingent on the company’s specific characteristics, such as, their international market profiles (Western or Asian developing countries), production strategies1 (centralized in one country or decentralized to different countries and regions), and other financial-specific elements (revenues, debt, operational cash flows). Due to these differences, Muller and Verschoor (2006) deem a firm-level study, instead of an aggregate study, to be more appropriate.
Companies can employ various hedging strategies to reduce exposure to exchange rate fluctuations. To varying degrees, companies can pass the changes in costs caused by exchange rate changes to customers. In addition, companies can also choose the location for the production facilities as well as the currency they borrow. Moreover, they have at their disposal a wide range of hedging instruments such as currency options, swaps and futures that can be used to lower currency risk. Scrutinizing in detail each of these factors is, however, beyond the scope of this study.
Modern portfolio theory, conditional on frictionless market economies without financial distress and bankruptcy costs, stresses that only non-diversifiable (or systematic) risks are charged a premium over market returns. Currency risk, however, is considered to be diversifiable in the sense that investors can effectively diversify by holding various stocks of companies from different countries. Hence, investors with a well-diversified portfolio are less exposed to exchange rate variations (Hull (2007)). Hedging activities carried out by companies, therefore, add no value.
A1. Effects of hedging on the market value of equity
There are companies that employ hedging schemes and those that do not. Whether hedging adds value is unclear in theory where there are broadly three different lines of thoughts. The first asserts that hedging activities have a net present value of zero. This is conditional on
1
6 frictionless and efficient market economies with no transaction costs and financial distress. Modigliani and Miller (1958) point out that without market deficiencies, a firm’s financial strategy has no impact on its value. It simply influences the way value is distributed between the firm’s claimants. The second states that hedging increase value as it reduces the expected financial distress costs (Smith and Stulz (1985)). It also helps firms maintain high levels of competitiveness, improves internal forecasting ability and facilitates investment (Brown (2001)). The last school of thought argues that hedging activities have a negative net present value because hedging activities are costly. Despite the costs, the company’s management still has an incentive to hedge since their wealth and human capital are insufficiently diversified. Hedging lowers their risk at no extra personal cost. This presents an agency problem in which managers hedge and benefit themselves at the expenses of the shareholders (Smith and Stulz (1985)). Besides this, a second reason as to why hedging can be a negative net present value activity is that managers may use derivatives and other financial instruments to speculate, thereby taking on an excessive amount of risk that can be value-damaging for the company (see, e.g., Geczy, Minton and Schrand (2007); Allayannis, Lei and Miller (2012)).
7
A2. Competitive impacts and operational benefits of hedging
From the treasury perspective, currency risk hedging can help firms maintain a stable operating profit margin. Currency risk management increases the accuracy of planning and pricing decisions and mitigates the negative effects of exchange rate changes on the firm’s competitiveness (Brown (2001)). It allows companies to engage in competitive pricing strategies in different foreign and domestic markets without being subjected to a high probability of earning a negative margin in one of these markets.
Effective currency risk hedging contributes to lower levels of uncertainty within the firm’s working-environment and improves its internal operations. Lower uncertainty contributes to more accurate domestic and foreign business forecasts, which allows the management to operate more effectively (Brown (2001)). In addition, it results in lower levels of cash flow volatility, which strengthens internal contracting and stimulate investment (Minton and Schrand (1999)). Hedging facilitates internal contracting by reducing the uncertainty surrounding the contracting process and leaves managers with the variable factors that are most relevant to their function within the company. Hence, it keeps well-performing managers from being subjected to punishment for fluctuations outside their control. Risk-averse managers are, therefore, in a better position to make optimized business decisions.
B. Empirical evidence on hedging, currency risk exposure and its characteristics
8 A possible explanation for the insignificant effects of exchange rate changes on stock returns is the autocorrelation in the exchange rate change series (Muller and Verschoor (2006)). Expected exchange rate movements must already be included in the price of stocks. This means that only the unexpected variations affect price at the moment they occur. Amihud (1994) postulates taking into account the past changes in exchange rate. Gao (2000) goes further and suggests taking into account macroeconomic indicators instead of lagged exchange rate movements.
Another possible explanation for the low significance of the currency risk exposure is the fact that companies have effectively hedged currency risk. In the absence of hedging, stock returns are likely to be more strongly affected by exchange rate movements compared to when firms hedge. To the extent that exposure is completely covered, we should expect to find no statistically and/or economically significant effect of exchange rate changes on stock returns. It is apparent that exchange rate hedging schemes, if effectively executed, can directly influence the level of companies’ exposures. Albeit numerous suggestions about the influence of hedging on exposure, few have incorporated the effects of hedging on the exposure coefficients into their researches. This might be the result of lacking data on hedging.
Evidence shows that exposures are to a large extent not actively managed (Marshall (2000)), even though most nonfinancial companies are reported to employ currency risk hedging strategies for their foreign transactions (Bartram, Brown and Fehle (2006)). Hedging is found to marginally reduce exposures in some empirical studies (see, e.g., De Jong, Ligterink and Macrae (2002); Muller and Verschoor (2004d)). Companies usually practice selective hedging, in which self-produced evaluations and predictions are utilized (see, e.g., Di Iorio and Faff (2000); Glaum (2002)). This, however, makes economic sense only if the firm possesses higher quality information and is able to make better estimations and predictions than the market. If this is not the case, the potential for ineffective hedging is to be expected
9 rate movements (Doidge, Griffin and Williamson (2006)). In addition, the firm information disclosures are often uninformative and obscured by unconvincing assumptions and decorative statements (Roulstone (1999)). Investors, therefore, can only thoroughly and accurately evaluate of the effects of exchange rate changes on the firm once they gain access to information on the past performance. A complete reflection of the effects of exchange rate changes on stock returns is therefore delayed until the information on past performance is released to the public. Studies, thus, suggest that lagged exposures need to be taken into account.
Williamson (2001) examines the possibilities of non-linear currency risk exposures in his study of the car manufacturing industry. He suggests a magnitude effect of the exchange rate changes on the company value. A quadratic term of the exchange rate changes, as a result, is added into the model. However, this method assumes an identical response to an appreciation as well as a depreciation, which is considered to be unrealistic. Convex and concave functional forms, proposed by Bartram (2004), allow stock return’s reactions to an appreciation to differ from those due to a depreciation. Even though the method increases the level of significance of currency risk exposure, it lacks a strong theoretical reasoning that provides a rationale for selecting a particular model specification. In addition, robustness tests carried out by Bartram (2004) give little evidence for non-linear exposure.
With regards to data frequency, theory assuming market efficiency asserts that currency risk exposure estimation should not depend on the frequency of the observation used. Empirical researches, on the other hand, point out that because of market inefficiencies and of the high complexity of the relation between firm value and exchange rate changes, the observation frequency matters for the estimation of the exposure coefficient (Muller and Verschoor (2006)).
10 and disclose the more fundamental long-term relationship between equity returns and exchange rate movements. Estimations using lower frequency data (e.g. monthly) are therefore preferred to those using higher frequency data (e.g. daily).
To sum up, due to the significant impact of exchange rate changes on firms’ international transactions and translations and on the increase in firms’ international operations in recent years, we should expect significant currency risk exposures. In addition, the effect of hedging on the market value of equity can be positive if hedging decisions are made to maximize shareholders’ value and are not significantly influenced by the management’s distorted incentives. If this is not the case, hedging may not add value or may even be value-reducing. Finally, evidence for a significant effect of hedging on exposure in the literature is mixed. If hedging schemes are carried out properly, we should expect it to be able to lower firm’s exposure. Nevertheless, if firms are unable to ensure this, hedging might be ineffective in reducing exposure.
In the following section, we present a framework to (1) examine the currency risk exposures of companies, (2) to investigate whether hedging can reduce exposure and (3) to see whether hedging can influence stock returns.
III. Theoretical Model A. Currency risk exposure
Below we present a framework that we use to measure the currency risk exposure of each individual firm. Individual firms’ exposures are then used as the dependent variable in the second stage regression (described in the next sub-section), where we examine how exposure is affected by hedging, firm size and the degree of foreign involvement of the firm. Based on this framework, we run panel data regressions with pooled stock returns of all companies to obtain the currency risk exposure of the sample as a whole and to examine the potential effects of hedging on the stock returns of companies.
11 creates a number of statistical issues, Adler, Domas and Simon (1986) propose the use of exchange rate changes and stock returns as a mean to obtain stationary series and overcome the estimation issues associated with the non-stationary series:
(1) In this equation, is the total return of company i in time t; represents the exchange rate
change in time t; , which is company i’s currency risk exposure, indicates the level of sensitivity of company i’s stock returns to movements in the exchange rates; and denotes respectively the constant and the error term.
Exchange rate is defined as the number of units of foreign currency per unit of domestic currency. An increase in the exchange rate, which is an appreciation of the domestic currency, causes domestic exporting goods to be more expensive for the foreign consumers, which may results in a fall in foreign demand and revenues. In contrast, firms that import goods from abroad will benefit from such an appreciation since the imported products become cheaper. Export-oriented companies therefore tend to exhibit negative foreign exchange exposure with the exchange rate defined in this way.
Adler et al. (1986)’s exposure gauges the part of stock return variance of company i that exhibits correlation with the exchange rate changes. It describes the total currency risk exposure of company i. As there are a number of macroeconomic factors that covary with both the stock returns and the exchange rate, it is important that they are incorporated into the model to avoid overestimation/underestimation of the currency risk exposure. We start with the Jorion (1990) model that includes the market return factor:
(2) Beside the variables described above, is the total market return in time t. measures the sensitivity of company i’s stock returns to market risk and measures firm i's exposure to contemporaneous exchange rate changes adjusted for the market exposure .
12 exchange rate changes is also included into the regression to capture the stock return exposure to lagged exchange rate changes:
(3)
where and represent respectively the firm's stock return exposure to contemporaneous
and lagged exchange rate changes in time t. is the market return factor; designates the change in the percentage growth rate of industrial production; denotes the change in expected future inflation; is the unexpected inflation; is the market risk premium and is the term structure of interest rates. Among the five macroeconomic factors, market risk premium and term structure of interest rates affect the discount rate. Higher risk premium and/or expected inflation are associated with higher discounted rates and, thus, a lower present value. It is therefore expected that the two variables are negatively associated with stock returns of companies. A higher expected future inflation means a higher expected real money demand, which is associated with higher levels of expected real activity (Kaul (1987)). This implies a positive relationship between stock return and expected inflation. The relationship between stock return and unexpected inflation, on the other hand, is hypothesized to be negative due to the fact that the central bank usually follows a counter-cyclical monetary policy supposedly to maintain a predetermined level of inflation (Kaul (1987)). Higher industrial production indicates higher levels of productivity and future cash flows that imply a positive relationship between stock return and industrial production.
Finally to see how hedging affect stock returns we add a variable H, which indicates the degree of hedging:
(4)
13 significant β8i suggests value-destroying effects. Otherwise, there is no evidence for the effects of
hedging on stock return.
B. Effect of hedging on exposure
To examine the potential effects of hedging on exposure, we employ the following cross-sectional regression:
̂ (5) In this equation, ̂ is the sum-exposure to contemporaneous and lagged exchange rate changes, i.e., . These exposures are estimated from equation (4) for firm i. denotes the degree of hedging as described above. designates the size of company and (foreign sales ratio) represents the share of foreign sales in the firm’s total sales. With regards to firm size, larger firms have advantages in terms of economies of scale in hedging and, thus, are able to hedge more effectively (see, e.g., He and Ng (1998); Batram (2002)), while smaller firms have a higher incentive to hedge because of their higher cash flow volatility (Fok, Carroll and Chiou (1997)). The effect that dominates, therefore, determines the sign of the variable . is an important determinant of the currency risk exposure because it serves as a proxy for the firm’s level of international operations (Gao (2000)). However, as the level of foreign operation increases, firms tend to relocate their production facilities to the foreign country and obtain local financing that in effect localizes their operation to the foreign location. This reduces their currency risk exposure without using financial hedging instruments such as currency options, futures or swaps (Allayannis and Ofek (2001)). Among these two contradicting effects, the one that dominates, therefore, determines the sign of .
14
C. Alternative models
It is pointed out that the measurement of currency risk exposure is sensitive to model specifications (see, e.g., Dominguez and Tesar (2006); Hsin et al. (2007)). Thus, while complying with the well-established Jorion’s model, we also employ an alternative specification based on Fama and French (1993) three-factor model, where the five additional macroeconomic variables in our specified model (equation 4) are replaced the by size and value factors.
(6)
In this equation, (small minus big), which is the small firm factor, is intended to mimic the risk factor in returns related to size, (high minus low), which is the value factor, is intended to mimic the risk factor returns related to book-to-market equity. In addition to this, we run a full model controlling for all five macroeconomic variables as well as the Fama-French size and value factors.
(7)
Using firms’ exposures from these two additional model specifications we repeat the cross-sectional exposure regression (equation 5) in order to examine whether the results differ among the three model specifications.
D. Regression methods
15 information about the volatility in the previous period and the fitted variance from the model during the previous period (Brook (2008)).
In the panel regression, the redundant fixed effects test and the Hausman correlated random effects test reject the necessity to assume fixed or random effects. Since the residuals with a given cross-section exhibit conditional heteroskedasticity (due to the ARCH effect), estimation using the SUR is more efficient compared to simple Pooled-OLS. For confirmation, we use a LaGrange-Multiplier test suggested by Breusch and Pagan (1980) to examine whether the residuals are indeed correlated. The test rejects the null hypothesis of no correlation between the residuals, thus, justifying the preference of the Seemingly Unrelated Regression (SUR) model over OLS. We use the Period SUR Estimated Generalized Least Squares for estimation of the pooled regression. This method allows for arbitrary heteroskedasticity and general correlation between the residuals for a given cross-section.
We use simple OLS regression to estimate the cross sectional exposure regression (equation 5) since there is no need for a more advanced method.
IV. Data and measures A. Sample selection
Given the data availability, 58 large US chemical companies are chosen. We select companies from only one industry because it is shown that the exchange rate sensitivity of industries significantly differs from each other (Choi and Prasad (1995)). Some industries respond positively to an appreciation of the home currency, while others respond negatively. Selecting companies from only one industry avoids this issue. We pick the chemical industry for our study because inputs and outputs of this industry are considered to be highly standardized. Standardized products can enter different countries’ market without (many) additional adjustments. Therefore, the exchange rate sensitivity of the chemical industry is likely to be high.
16 when expanding to foreign markets and are likely to have high levels of foreign operations. Finally, information disclosure requirements in the US are among the highest in the world that benefits us in terms of data availability.
For the first stage regression, monthly data is collected from February 2008 to January 2012 for the 58 companies. As discussed earlier, the use of monthly data is generally supported in the literature (see, e.g., Chow et al. (1997a); Chow and Chen (1998); Griffin and Stulz, 2001; Dominguez and Tesar (2001a) and Muller and Verschoor (2006)). Studies suggest that that using monthly data allow models to capture the long swings in value that currencies experience and reveals the more fundamental long-term relationship between equity returns and exchange rate movements.
B. Measures
B.1. Dependent Variables
The firm’s total return is measured by the total stock return index, collected from Datastream. The index is calculated as log returns on a monthly basis, including dividend. The dependent variable in the second stage regression (equation 5) is the foreign exchange exposure obtained from the individual firms’ stock return regressions (equations 4, 6 and 7).
B.2. Independent Variables
The exchange rate movement is proxied for by change of the US Dollar Index. This index measures the value of the dollar relative to other world currencies. The weight given to each currency in the calculation is based on the volume of trade between the dollar and the corresponding currency. Data for the index is collected from Datastream.
The market return factor is proxied for by the value-weighted return on all NYSE, AMEX, and NASDAQ stocks. The Fama-French size and value factors, HML and SMB, are collected from Professor Kenneth R. French’s Data Library.2 SMB, calculated as the average return from three small caps portfolios minus the average return on three big caps portfolios, is intended to mimic the risk factor in returns related to size (Fama and French (1993)). HML, calculated as the average return on two value portfolios minus the average return on two growth
2
17 portfolios, is intended to mimic the risk factor returns related to book-to-market equity (Fama and French (1993)).
Chen et al. (1986)’s five macroeconomic factors are the change in the industrial production growth rate (IP), change in the expected future inflation (ExInf), unexpected inflation (UnInf), risk premium (RP), and term structure of interest rates (TS). Data for the change in the percentage growth rate of the US industrial production (IP) is collected from the US Federal
Reserve. Data for the US actual and expected inflation is collected from the Bank of St. Louis Federal Reserve. For the change in expected future inflation, we use the University of Michigan
Inflation Expectation series calculated as the median expected price change over the next 12 months, based on the MICH Survey of Consumers. The unexpected inflation is calculated by subtracting the expected inflation in the previous period from the actual inflation of the current period. Following Chen et al. (1986) and Jorion (1991), we calculate the risk premium by taking the difference between government bond yield and lower medium grade bond yield. The series is constructed based on the US 10-year Treasury Constant Maturity Rate and Moody’s Seasonally Adjusted BAA Corporate Bond Yield. Data for both are collected from the Board of Governors
of the Federal Reserve System. In accordance with Chen et al. (1986), we calculate the term
structure of interest rates by subtracting the 3-month T-bill (DataStream) rate at time t-1 from the 10-year treasury rate at time t. This variable can be considered as a measure for the unanticipated return on long-term bonds.
18 disclosed. Section 7A provides details on the degree of exposure after hedging. Besides companies that either hedge extensively or do not hedge, there are companies, which are midway between these two extremes (i.e., do hedge but only to a certain extent). The hedging variable for the companies in this grey area takes the value of 0.5. Form 10-K reports of the 58 companies in our sample are retrieved from the SEC’s EDGAR Database.
Finally, data on firm size (number of employees measured in thousands) and foreign sales ratio (proportion of foreign sales relative to the total sales) is retrieved respectively from Bureau
van Dijk’s Orbis database and from companies’ 2010 10-K reports. The use of the number of
19 Table 1: This table presents the correlation matrixes of the independent variables. Panel A and B show correlation matrixes of the independent variables used in the stock return regressions and the exposure regressions, respectively. Abbreviations of the variables are presented in panel C.
Panel A: Correlation matrix - stock returns regresion
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) θt 1 (2) θt-1 -0.08 1 (3) Rmt -0.14 -0.10 1 (4) ExInft -0.46 -0.11 0.16 1 (5) UnInflt 0.20 0.09 -0.28 -0.01 1 (6) IPt -0.07 -0.27 0.24 0.21 0.01 1 (7) RPt -0.09 -0.03 0.07 -0.03 -0.72 -0.43 1 (8) TSt 0.07 0.24 -0.03 -0.24 -0.59 -0.37 0.73 1 (9) SMBt -0.13 -0.04 0.51 0.18 -0.10 -0.03 0.20 0.02 1 (10) HMLt -0.06 0.04 0.61 0.18 -0.28 0.02 0.15 0.12 0.50 1
Panel B: Correlation matrix - cross-sectional exposure regression
(1) (2) (3)
(1) Hi 1
(2) Sizei -0.11 1
(3) FSRi -0.13 0.03 1
Panel C: Abbreviation
θt Exchange rate change
θt-1 1-period lageed exchange rate change
Rmt Market return
ExInft Expected inflation
UnInflt Unexpected inflation
IPt Industrial Production
RPt Risk Premia
TSt Term structure
SMBt Difference between returns of small caps and large caps
HMLt Difference between returns of portfolios of value firms and portfolios of growth firms
Hi Degree of Hedging
Sizei Size of the company
20 Table 2: Descriptive statistics are reported in this table. Panel A describe the time series independent variables used in the stock return regressions for the period from February 2008 to January 2012. and are respectively the contemporaneous and lagged exchange rate change, is the market return factor, is the industrial production, and denote respectively the expected future inflation and unexpected inflation, is the market risk premium and is the term structure of interest rates. (small minus big), which is the small firm factor, is intended to mimic the risk factor in returns related to size, (high minus low), which is the value factor, is intended to mimic the risk factor returns related to book-to-market equity. designates the size of the company and (foreign sales ratio) represents the share of foreign sales in the firm’s total sales. Panel C describe firms’ contemporaneous and lagged exchange rate exposures using respectively the adjusted Jorion model, the Fama-French three-factor model and the full model.
Panel A: Time series variables for stock return regressions
Mean Median Maximum Minimum Std. Dev. Skewness Ex. Kurtosis Normality N
θt 0.00 0.00 0.07 -0.07 0.03 0.32 0.80 0.60 48 θt-1 0.00 0.00 0.07 -0.07 0.03 0.27 0.76 0.58 48 Rmt 0.33 1.20 11.34 -17.15 6.07 0.92 0.53 0.25 48 ExInft 0.00 0.00 1.20 -1.20 0.44 0.50 0.06 0.35 48 UnInflt -1.13 -0.53 2.14 -7.09 2.52 0.98 0.58 0.10 48 IPt -0.13 0.15 1.30 -4.10 1.05 0.95 0.07 0.15 48 RPt 3.37 3.02 6.01 2.40 0.99 0.00 0.30 0.01 48 TSt 3.02 2.81 5.98 0.96 1.14 0.01 0.34 0.05 48 HMLt 0.58 0.37 10.64 -4.27 2.77 0.14 0.15 0.51 48 SMBt 0.02 -0.45 19.72 -8.75 4.64 0.14 0.08 0.47 48
Panel B: Firms' characteristics
Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis Shapiro–Wilk N
Sizei 10.43 4.10 129.00 0.35 19.41 4.45 26.02 0.00 57
FSRi 0.48 0.50 0.86 0.04 0.21 -0.20 2.28 0.34 57
Panel C: Firms' stock return exposure to contemporaneous and lagged exchange rate changes
Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis Shapiro–Wilk N
γi (Jorion) -1.45 -1.16 0.46 -8.10 1.33 -2.45 12.27 0.00 58
γ'i (Jorion) -0.07 0.01 2.38 -2.19 0.83 -0.15 3.84 0.06 58
γi + γ'I (Jorion) -1.52 -1.12 1.48 -5.72 1.50 -0.70 3.27 0.03 58
γi (Fama-French) -1.87 -1.82 0.89 -5.33 1.09 -0.55 3.93 0.27 58
γ'i (Fama-French) -0.45 -0.41 2.75 -1.98 0.84 0.98 6.62 0.00 58
γi + γ'i (Fama-French) -2.32 -2.21 -0.10 -7.08 1.45 -0.77 3.66 0.03 58
γi (Full) -1.54 -1.42 0.28 -4.47 1.01 -0.83 3.85 0.02 57
γ'i (Full) -0.29 -0.19 1.32 -2.62 0.77 -0.58 3.90 0.05 57
21
V. Empirical Results
Table 1 and 2 respectively show the correlation matrixes and the summary statistics. The correlation matrix shows high correlation between the term structure of interest rates and the risk premium (correlation of 0.73) and between the risk premium and the unexpected inflation (correlation of -0.72). The Variance Inflation Factor (VIF) test for multicollinearity (Appendix table IV) indeed indicates the multicollinearity problem around the term structure of interest rates and the risk premium. After omitting these variables there is no more evidence for multicollinearity as shown in appendix table IV. Tests for skewness, kurtosis and normality, described in Bai and Ng (2005), for time series data are presented in panel A of table 2. The variables are shown to be slightly right-skewed ( for a normal distribution) and slightly leptokurtic ( for a normal distribution), indicating that data distributions have higher peaks and thicker tails compared to a normal distribution. The normality tests reported in the final column show that data for time series variables are indeed marginally deviated from a normal distribution ( for a normal distribution).
Panel B and C of table 2 present descriptive statistics of the variables used in the second stage regressions. The Shapiro-Wilk test is applied to test for the equality of the distribution.3 The results indicate that the distribution of Size significantly differ from normal (p-value = 0). The normality of FSR is, however, not rejected. The currency risk exposures are left-skewed, leptokurtic and different from normally distributed (p-value of the total exposure from the adjusted Jorion model, the Fama-French three-factor model and the full model are respectively 0.03, 0.03 and 0.06).
Individual stock return regressions are presented in appendix tables I, II and III using respectively the adjusted Jorion model (eq. 4), the Fama French three-factors model (eq. 6) and the full model (eq. 7). Interestingly, out of the total 58 companies, only 13, 15 and 17 report a significant market return’s coefficient using the three model specifications respectively. Most companies have a low market beta. In contrast, there are more than double the numbers of companies that report a significant exchange rate change coefficient (36, 53 and 41 companies respectively for the three model specifications). Significant effects of the lagged exchange rate changes are found in 10, 21 and 15 companies using the three model specifications. The
3
22 contemporaneous currency risk exposure and the lagged exposure are negative for most of the companies.
Cross-sectional individual firm regression (eq. 5) results are reported in table 6. Panel A, B and C present respectively the results from different model specifications using all of the exposure coefficients from the first stage (panel A), using only the significant exposure coefficients (panel B) and using only the negative and significant exposure coefficients (panel C). Columns (1)-(4), (5)-(8) and (9)-(12) in each panel show the regression results using respectively the exposure coefficients from the adjusted Jorion model, the Fama-French three-factor model and the full model as the dependent variable. The results from the three panels are in general in line with each other. Hedging coefficients are negative is all regressions. Using currency risk exposures from the Fama-French three-factor model as the dependent variable (Column (5)-(8)) gives negative and significant coefficients of hedging, while using exposures from the adjusted Jorion model or the full model gives insignificant coefficients. The relationship between FSR and exposure is negative and significant in all regression results. Size, on the other hand, is positive and insignificant in most of the results.
The stock returns panel-regression results are shown in table 7. Column (1)-(4) show the regression results without any control variable. Column (6)-(8), (9)-(10) and (11)-(14) present respectively the regression results using the adjusted Jorion model, the Fama-French three-factor model and the full model. Due to the fact that the risk premium is highly correlated with the term structure as well as with the unexpected inflation, in column (6) and (11), we exclude risk premium out of the regression to see if other results are affected. In column (13) we leave out both risk premium and term structure, which are the variables associated with the multicollinearity problem according to the VIF tests, to completely eliminate the multicollinearity problem. In all cases, signs and significance levels of all variables are not affected.
23 of , and respectively in the adjusted Jorion model, the Fama-French model and the full model (column (8), (10) and (14)). The coefficients of hedging are positive but insignificant in all regressions.
24 Table 3:Estimation is based a cross-sectional sample with 58 US chemical companies. Currency risk exposures of companies are used as the dependent variable. In panel B, only the significant exposures are used, while all exposures are used in panel A. In panel C, only the negative and significant exposures are used. Regressions include the variable H representing level of hedging by firms. The Size is proxied for by the number of employees (measured in thousands). FSR is the Foreign Sales Ratio capturing the degree of foreign involvement of the firm. Columns (1)-(4), (5)-(8) and (9)-(12) in each panel show respectively the regression results using the exposure coefficients from the adjusted Jorion model, the Fama-French three-factor model and the full model as the dependent variable.
Symbol '***', '**', and '*' indicate statistical significant at 1, 5 and 10%, respectively.
Adjusted Jorion Model Fama-French Model Full Model
Panel A: All Exposures
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Constant -1.47 -1.61 -0.03 -0.13 -2.03 -2.12 -0.72 -0.80 -1.85 -1.67 -1.02 -0.78 Hi -0.12 -0.07 -0.30 -0.26 -0.75* -0.71 -0.92** -0.89** 0.04 -0.04 -0.08 -0.18 Sizei 0.01 0.01 0.01 0.01 -0.02 -0.02 FSRi -2.87*** -2.97*** -2.63*** -2.69*** -1.63* -1.77** R-square 0.00 0.02 0.16 0.18 0.05 0.06 0.19 0.21 0.00 0.05 0.06 0.11 Ads. R-sqr -0.02 -0.01 0.13 0.14 0.03 0.03 0.16 0.16 -0.02 0.01 0.02 0.06 F-statistics 0.07 0.60 5.05*** 3.91*** 2.9* 1.73 6.52*** 4.57*** 0.01 1.26 1.64 2.11 Num. Obs. 58 57 57 56 58 57 57 56 57 56 56 55
Panel B: Significant Exposures
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Constant -2.03 -2.29 -0.30 -0.48 -1.84 -1.928 -0.542 -0.637 -1.95 -2.148 -0.242 -0.42 Hi 0.10 0.22 -0.10 0.008 -0.74 -0.69 -0.9** -0.852* -0.04 0.058 -0.291 -0.209 Sizei 0.016* 0.017 0.007 0.008 0.013 0.01 FSRi -3.25*** -3.46*** -2.6*** -2.64*** -3.18*** -3.23*** R-square 0.00 0.09 0.256 0.37 0.05 0.06 0.20 0.21 0.00 0.05 0.24 0.29 Ads. R-sqr -0.03 0.04 0.21 0.31 0.03 0.02 0.16 0.16 -0.02 0.003 0.20 0.24 F-statistics 0.05 1.74 6.04*** 6.44*** 2.73 1.57 5.99*** 4.07** 0.01 1.07 6.58*** 5.4*** Num. Obs. 39 38 38 37 53 52 52 51 46 45 45 44
Panel C: Negative Significant Exposure
25 Table 4: The stock returns panel-regression results are shown in this table. Estimation is based on a 2784 firm-month observations for the period from February 2008 to January 2012. Column (1)-(4) show the regression results without any control variable. Column (6)-(8), (9)-(10) and (11)-(14) present respectively the regression results from the adjusted Jorion model, the Fama-French three-factor model and the full model. Due to the fact that the risk premium highly correlated with the term structure as well as with the unexpected inflation, in column (6) and (11) we exclude risk premium out of the regression to see if other results are affected. In column (13) we leave out both risk premium and term structure, which are the variables associated with the multicollinearity problem according to the VIF test, to completely eliminate the multicollinearity problem. In all cases, signs and significance levels of all variables are not affected. and are
respectively the contemporaneous and lagged exchange rate change in time t, is the market return factor, is the industrial production, and
denote respectively the expected future inflation and unexpected inflation, is the market risk premium and is the term structure of interest rates. (small minus big), which is the small firm factor, is intended to mimic the risk factor in returns related to size, (high minus low), which is the value factor, is intended to mimic the risk factor returns related to book-to-market equity.
Symbol '***', '**', and '*' indicate statistical significant at 1, 5 and 10%, respectively.
26
VI. Discussion
Our analysis indicates significant effects of exchange rate changes on stock return for the majority of companies in the dataset during the period from February 2008 to January 2012. These are in contrast with the past empirical findings where no significant results are recorded (see, e.g., Jorion (1990); Jorion (1991); Bodnar and Gentry (1993); Amihud (1994); Choi and Prasad (1995); Miller and Reuer (1998); Chow, Lee and Solt (1997b)).The 58 large US chemical companies mostly report a negative currency risk exposure during this period. This indicates a high amount of revenue being repatriated to headquarters from foreign subsidiaries and/or profits gained from exporting activities.
Our analysis of firms’ exposures using cross-sectional regressions suggest significant and negative effects of hedging on exposure when using exposures from the Fama-French three factor model in the first stage. These findings are in line with Choi and Jiang (2009) who assert strong capabilities of multinational companies in reducing currency risk through financial hedging. Utilizing only one set of control variables (the Fama-French size and value factors), Choi and Jiang (2009) obtain currency risk exposures of individual firms and regress them against a financial hedging dummy. However, as shown in our analysis, once the five macroeconomic factors are controlled for, the significant effects of hedging disappear. As pointed out by Dominguez and Tesar (2006) and Hsin et al. (2007), the measurement of currency risk exposure is sensitive to model specifications. This raises questions on the reliability of Choi and Jiang (2009)’s findings and on the true effects of hedging on currency risk exposure.
It is important to point out that both a highly negative exposure coefficient and a highly positive exposure coefficient indicate a high currency risk exposure. Therefore, in the exposure regression (eq. 5), the negative coefficient of hedging can mean either more negative exposure (i.e., even more negative ) as a result of more hedging or less positive exposure (i.e., lower positive ) as a result of more hedging. The true effects of hedging on exposure, thus, depend on the pre-hedging value of the exposure coefficients. The pre-hedging exposures, however, cannot be observed. Concluding that hedging reduces exposure is therefore correct only if the pre-hedging exposures of firms are positive.
27 stock returns less sensitive to changes in either the market return or the exchange rate. The combined effect of a negative market beta, a negative currency risk exposure and a negative correlation between market return and exchange rate change, therefore, makes stock returns more stable.
Panel regression results (table 7) show insignificant coefficients for hedging in all three model specifications. This is at odds with Allayannis and Weston (2001) who, using a dataset of 378 non-financial companies in 1993, found financial hedging to be a value-enhancing strategy, as well as with Nelson et al. (2005) who show significant positive effects of the use of currency derivatives on the market value of equity over the period 1995-1999. Lacking evidence for strong direct benefits of hedging on firm value as there were in the past might be an indication of altered hedging incentives, distorted by managers’ personal objectives, which is made clearer as a result of high market fluctuations during the financial crisis.
VII. Conclusion
Even though theory predicts a sizeable currency risk exposure for firms, supportive empirical findings are yet to be found. Recent studies suggest a reason for this lack of empirical support arises from the fact that companies have taken measures to reduce their exposures to a level so low that it cannot be detected. In addition, albeit numerous papers about the factors that drive hedging decisions, the issue of whether hedging benefits companies and its shareholders has not been thoroughly investigated. It is pointed out that hedging practices and purposes are subjected to changes over time adhering to changing management incentives and market conditions, specifically as a result of the 2008 financial crisis. Whether hedging maintains its level of effectiveness and the objective of maximizing shareholders’ value in the recent periods is questionable.
28 effectiveness of hedging activities in recent years. Finally, we found no evidence for a significant effect of hedging on stock return. This indicates that hedging adds no value to companies during this period. This creates concerns about the management’s hedging decisions. Whether it maintains the objective of maximizing shareholders’ value, as suggest in past empirical findings, is doubtful. We left this question to future scrutiny.
The variable hedging remains a limitation of this thesis. The information on the hedging activities of companies is obtained through analyzing companies’ 10-K reports. In such a report, albeit information disclosing requirements, a company can still neglect their disclosure of internal information in order to avoid damaging their competitive position. As is often the case, information is masked with questionable assumptions and vague statements. Moreover, despite the efforts by the SEC to standardize the reported information, reporting practices of firms still vary making comparison difficult to be executed.
Another drawback in this thesis is the use of a trade-weighted exchange rate index. Such an index ignores the issue of low or negative correlations among exchange rates. This may result in underestimation of the currency risk exposure because the index cannot capture the diverging movements of different currencies’ values. Hence, selecting the exchange rates between domestic currency and the most relevant foreign currencies for the company can give more precise results.
29
References
Adler, M., Dumas, B., 1984, Exposure to Currency Risk: Definition and Measurement, Financial Management 13, 2, 41-50.
Adler, M., Dumas, B., Simon, D., 1986, Exchange risk surprises in international portfolios, Journal of Portfolio Management, 44–53.
Allayannis, G., Weston, J., 2001, The use of foreign currency derivatives and firm market value, Review of Financial Studies 14, 1, 243-276.
Allayannis, G., Lel, U., Miller, D., 2012, The use of foreign currency derivatives, corporate governance, and firm value around the world, Journal of International Economics 87, 65-79.
Allayannis, G., Ofek, E., 2001, Exchange rate exposure, hedging and the use of foreign currency derivatives, Journal of International Money and Finance 20, 273–296.
Amihud, Y., 1994. Exchange rates and the valuation of equity shares. In: Amihud, Y., Levich, R.M. (Eds.), Exchange rates and corporate performance. Irwin, New York, pp. 49–59.
Anderson, B. P., Makar, S. D., Huffman, S. H., 2004, Exchange rate exposure and foreign exchange derivatives: do ineffective hedgers modify future derivatives use?, Research in International Business and Finance 18 (2) , 205--216 .
Arnott, R., Hsu, J., Moore, P., 2005, Fundamental Indexation, Financial Analysts Journal 61.
Bai, J., Ng., S., 2005, Tests for skewness, kurtosis, and normality for time series data, Jornal of Business & Economics Statistics 23, 49-60.
Bartov, E., Bodnar, G.M., 1994, Firm valuation, earnings expectations, and the exchange rate exposure effect, Journal of Finance 49, 1755–1785.
Bartram, S., 2004, Linear and nonlinear foreign exchange rate exposure of German non-financial corporations, Journal of International Money and Finance 23, 673–699.
Bartram, S., Brown, G.W., Fehle, F.R., 2006, International evidence on financial derivative usage, Financial Management 38, 185-206.
Berkman, H., and M.E. Bradbury, 1996, Empirical Evidence on the Corporate Use of Derivatives, Financial Management 25, 5-13.
Bessembinder, H., 1991. Forward contracts and firm value: investment incentive and contracting effects. Journal of Financial and Quantitative Analysis 26, 519–532.
Bodnar, G.M., Gentry, W.M., 1993. Exchange rate exposure and industry characteristics: evidence from Canada, Japan and the USA, Journal of International Money and Finance 12, 29–45.
30
Chamberlain, S., Howe, J.S., Popper, H., 1997, The exchange rate exposure of US and Japanese banking institutions, Journal of Banking and Finance 21,871–889.
Chen, N.F., Roll, R., Ross, S.A., 1986, Economic forces and the stock market, Journal of Business 59, 383–403.
Choi, J., and Prasad, A., 1995, Exchange Rate Sensitivity and Its Determinants: A Firm and Industry Analysis of US Multinationals, Financial Management 24, 3, 77-88.
Choi, J., Ijang, C., 2009, Does Multinationality matter? Implications of operational hedging for the exchange risk exposure, Journal of Banking & Finance 33, 1973-1982.
Chow, E.H., Chen,H.L., 1998, The determinants of foreign exchange rate exposure: evidence on Japanese firms, Pacific-Basin Finance Journal 6,153–174.
Chow, E.H., Lee, W.Y., Solt, M.E., 1997a, The economic exposure of US multinational firms, The Journal of Financial Research 20, 191–210.
Chow, E.H., Lee, W.Y., Solt, M.E., 1997b, The exchange-rate risk exposure of asset returns, Journal of Business 70, 105–123.
De Jong, A., Ligterink, J., Macrae, V., 2002, A firm-specific analysis of the exchange-rate exposure of Dutch firms. Working Paper, Erasmus University Rotterdam.
DeMarzo, P.M., Duffie, D., 1995, Corporate incentives for hedging and hedge accounting, The Review of Financial Studies 8, 743–771.
Di Iorio, A., Faff, R., 2000. An analysis of asymmetry in foreign currency exposure of the Australian equities market. Journal of Multinational Financial Management 10, 133–159.
Doidge, C., Griffin, J., Williamson, R., 2006, Measuring the economic importance of exchange rate exposure, Journal of Empirical Finance 13, 550–576.
Dominguez, K.M.E., Tesar, L.L., 2001a, A re-examination of exchange rate exposure, American Economic Review 91, 396–399.
Dominguez, K.M.E., Tesar, L.L., 2006, Exchange rate exposure, Journal of International Economics 68, 188–218.
Fama, E.F., French, K.R., 1993, Common risk factors in the returns on stocks and bonds, Journal of Financial Economics 33, 3-56.
Fok, R.C., Carroll, C., Chiou, M.C., 1997, Determinants of Corporate Hedging and Derivatives: A Revisit’, Journal of Economics and Business 49, 569-585.
Gao, T., 2000. Exchange rate movements and the profitability of US multinationals, Journal of International Money and Finance 19, 117–134.
Geczy, Christopher C., Minton, Bernadette A., Schrand, Catherine M., 2007, Taking a view: corporate speculation, governance, and compensation, Journal of Finance 62, 2405–2443.
31
Griffin, J.M., Stulz, R.M., 2001, International competition and exchange rate shocks:a cross-country industry analysis of stock returns, Review of Financial Studies 14, 215–241.
He, J., Ng, L., 1998, The foreign exchange exposure of Japanese multinational corporations, Journal of Finance 53, 733–753.
Hsin, C., Hou, S., Chang, F., 2007, Stock Return Exposure to Exchange Rate Risk: a perspective from delayed reactions and hedging effects, Journal of Multinational Financial Management 17, 384-400. Hull, 2007, Options, Futures and Other Derivatives (Prentice Hall, New Jersey). Jensen, Michael C., and William H. Meckling, 1976, Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, Journal of Financial Economics 3, 4, 305-360.
Jorion, P., 1990, The Exchange-Rate Exposure of US Multinationals, The Journal of Business 63, 3, 331-345.
Jorion, P., 1991, The Pricing of Exchange rate exposure in the Stock Market, The Journal of Financial and Quantitative Analysis 26, 3, 363-376.
Kaul, G., 1987, Stock returns and inflation: The Role of the Monetary Sector, Journal of Financial Economics 18, 253-275.
Levi, M.D., 1994. Exchange rates and the valuation of firms. In: Amihud, Y., Levich, R.M. (Eds.), Exchange Rates and Corporate Performance. Irwin, New York, pp. 37–48.
Marshall, A.P., 2000, Foreign exchange risk management in UK, USA and Asia Pacific multinational companies, Journal of Multinational Financial Management 10, 185–211.
Minton, B.A., Schrand, C., 1999, The impact of cash flow volatility on discretionary investment and the costs of debt and equity financing, Journal of Financial Economics 54, 432–460.
Modigliani, F., Miller, M., 1958, The cost of capital, corporate finance and the theory of investment. American Economic Review 48, 261–297.
Muller, A., Verschoor, W., 2006, Foreign exchange risk exposure: Survey and suggestiongs, Journal of International Financial Management 16, 385-410.
Muller, A., Verschoor, W.F.C., 2004c, Asymmetric foreign currency risk exposure: evidence from US multinationals, LIFE working paper, Maastricht University.
Muller, A., Verschoor, W.F.C., 2004d, Foreign exchange risk exposure and the use of foreign currency derivatives. Working Paper, Radboud University Nijmegen.
Nelson, J., Moffitt, J., Graves, J., 2005, The impact of hedging on the market value of equity, Journal of Corporate Finance 11, 851-881.
Nguyen, H., Faff, R., 2003, Can the use of foreign currency derivatives explain variations in foreign exchange exposure? Evidence from Australian companies, Journal of Multinational Financial Management 13, 193–215.
Oosterhof, C.M., 2001, Corporate Risk Management: An Overview, SOM-theme E: markets and institutions, Working Paper, University of Groningen.
32
Roulstone, D., 1999. Effect of the SEC financial reporting release on derivative and market risk disclosures. Accounting Horizons 13, 343–363.
Shapiro, A.C., 1975, Exchange rate changes, inflation and the value of multinational corporations, Journal of Finance 30, 485–502.
Smith, C.W., and R.M. Stulz, 1995, The Determinant of Hedging Policies, Journal of Financial and Quantitative Finance 20, 4, 391-405.
The American Chemistry Council, 2012, Global Business of Chemistry.
Williamson, R.G., 2001, Exchange rate exposure and competition: evidence from the automotive industry, Journal of Financial Economics 59, 441–475.
33
Appendices
Table I:This table shows stock return regression results of 58 large US chemical companies for the period from February 2008 to January 2012 (monthly), using the adjusted Jorion model. For simplicity and an easy comparison with the two following tables, only results of the main explanatory variables are reported. is the market return factor, and are respectively the contemporaneous and lagged exchange rate change. Not reported in the table are the five macroeconomics factors, used as control variables, which are: industrial production, expected future and unexpected inflation, market risk premium and term structure of interest rates.
Symbol '***', '**', and '*' indicate statistical significant at 1, 5 and 10%, respectively.
Company name Constant Rmt θt θt-1 R-squared Adjt R-sq Durbin-Watson
AIR PRODUCTS & CHEMICALS INC -0.112 0.001 -1.157*** -0.178 0.36 0.23 2.48
ALBEMARLE CORP -0.02 0.000 -1.52** 0.057 0.32 0.18 2.28
AMERICAN VANGUARD CORP 0.0276 -0.005 0.407 -1.744** 0.18 0.02 2.17
ARCH CHEMICALS INC 0.194** 0.000 -1.142** -0.211 0.17 0.00 2.21
ASHLAND INC -0.144 0.005* -0.925 0.23 0.23 0.07 1.84
AVON PRODUCTS INC -0.007 0.005* -2.103*** -0.486 0.46 0.35 2.00
CABOT CORP 0.218 -0.004 -0.932 0.038 0.18 0.01 1.80
CALGON CARBON CORPORATION 0.19* 0.002 -0.226 0.758 0.04 -0.15 2.69
CAMBREX CORP 0.006 -0.006 -3.266*** -0.398 0.27 0.12 2.05
CARLISLE COMPANIES INC 0.013 0.000 -0.466 0.025 0.11 -0.07 2.04
CELANESE CORPORATION -0.114 0.007*** -1.936*** -1.147*** 0.33 0.19 2.16
CF INDUSTRIES HOLDINGS, INC. -0.102 0.009*** -2.433*** -0.105 0.17 0.00 3.06
CHEMTURA CORPORATION -0.391 0.005 -4.107*** 1.256 0.10 -0.08 1.81
CHURCH & DWIGHT CO INC 0.09 -0.000 -0.101 -0.153 0.18 0.01 2.37
CLOROX CO 0.079** 0.001 -0.189 -0.277 0.08 -0.11 1.41
CYTEC INDUSTRIES INC 0.131 0.001 -1.985** -0.17 0.41 0.28 2.16
DOW CHEMICAL COMPANY (THE) 0.123 0.004* -0.918* 0.002 0.21 0.05 2.09
EASTMAN CHEMICAL CO 0.165*** 0.005** -0.913** -0.013 0.28 0.13 2.27
ECOLAB INC 0.058 -0.001 -0.613 0.137 0.19 0.02 1.97
ELIZABETH ARDEN INC 0.202* 0.005* -0.89* 0.151 0.20 0.04 2.15
FERRO CORP 0.251* -0.002 -0.73 0.688 0.24 0.08 2.08
FMC CORP -0.066 0.002 -1.257** 0.322 0.16 -0.01 2.98
H.B. FULLER COMPANY 0.016 0.002 -0.747 0.045 0.17 0.00 2.42
HUNTSMAN CORPORATION 0.323 0.006 -0.327 -0.165 0.19 0.02 1.91
INNOSPEC INC. 0.124 -0.000 -1.949* -1.764* 0.38 0.25 2.01
INTERNATIONAL FLAVORS & FRAGRANCES INC0.054 0.002*** -1.114*** 0.355 0.32 0.18 2.08
KOPPERS HOLDINGS INC. 0.136 0.004 -3.025*** -0.054 0.49 0.38 1.97
KRONOS WORLDWIDE, INC. 0.369** -0.006 -2.146** -1.431* 0.23 0.07 2.24
LSB Industries, Inc. 0.011 0.003 -2.813*** -0.446 0.27 0.13 1.98
LUBRIZOL CORP 0.072 0.002 -0.653 -1.158*** 0.20 0.04 2.55
METHANEX CORPORATION 0.048 0.007** -1.886*** 0.173 0.45 0.34 2.15
MINERALS TECHNOLOGIES INC 0.071 0.000 -0.636 0.017 0.23 0.07 2.63
MONSANTO CO -0.144** 0.003 -1.162*** 0.166 0.10 -0.08 2.68
NALCO HOLDING COMPANY -0.073 0.002 -1.745*** -0.655** 0.19 0.03 1.81
NEWMARKET CORPORATION 0.079 0.003 -2.059** -0.427 0.32 0.18 2.29
OLIN CORP -0.165** 0.002 -2.005*** 0.765 0.12 -0.06 2.53
OM GROUP INC 0.132 0.007* -1.735** -2.185*** 0.19 0.03 2.31
OMNOVA SOLUTIONS INC -0.001 0.001 -0.477 1.589** 0.30 0.15 2.22
POLYONE CORPORATION 0.128 0.002 -1.428** 0.729 0.30 0.16 2.45
POLYPORE INTERNATIONAL, INC. 0.111 0.009*** -3.043*** -1.15*** 0.08 -0.11 2.03
PPG INDUSTRIES INC 0.126 0.003 -0.736* 0.799 0.21 0.05 2.36
PRAXAIR INC 0.007 0.002 -0.98*** 0.163 0.42 0.30 2.46
PROCTER & GAMBLE CO 0.126*** 0.000 -0.719*** -0.078 0.32 0.18 1.55
REVLON, INC. 0.258*** -0.001 -1.217 -1.311 0.06 -0.14 1.56
ROCKWOOD HOLDINGS, INC. 0.091 0.002 -4.496*** -0.966 0.54 0.44 2.13
RPM INTERNATIONAL INC. 0.083 0.001 -1.533** 0.205 0.41 0.29 2.21
SCHULMAN A INC -0.045 -0.001 -0.421 0.312 0.16 -0.01 2.07
SCOTTS MIRACLE-GRO COMPANY (THE) 0.025 0.003* -0.532 0.412 0.04 -0.15 2.58
SIGMA ALDRICH CORP 0.052 0.002 -1.252*** -0.159 0.44 0.32 2.22
SOLUTIA INC 0.24 -0.000 -0.886 0.724 0.06 -0.13 1.17
SPARTECH CORP -0.061 -0.003 0.464 1.011 0.18 0.01 2.20
STEPAN COMPANY 0.117 -0.001 -1.523** -0.078 0.26 0.11 1.66
TERRA INDUSTRIES INC -0.117 -0.004 -2.083*** -0.621 0.17 0.00 2.56
TRONOX INCORPORATED -0.311 0.003 -8.099 2.377 0.85 0.56 2.14
USEC INC -0.086 -0.006 -0.779 0.657 0.15 -0.02 2.21
VALHI INC -0.105* -0.005** -1.618*** -1.612*** 0.15 -0.03 2.44
VALSPAR CORP 0.029 0.003 0.0346 0.348 0.10 -0.08 2.33
ZOLTEK COMPANIES INC -0.122 -0.005 -1.432* 0.652 0.43 0.32 2.29
34
Table II: This table shows stock return regression results of 58 large US chemical companies for the period from February 2008 to January 2012, using Fama French three-factor model. The main explanatory variables are the market return factor and the contemporaneous and lagged exchange rate changes, and respectively. The control variables include: (small minus big), which is the small firm factor, is intended to mimic the risk factor in returns related to size and (high minus low) , which is the value factor, is intended to mimic the risk factor returns related to book-to-market equity.
Symbol '***', '**', and '*' indicate statistical significant at 1, 5 and 10%, respectively.
Company name Constant Rmt θt θt-1 R-squared Adjt R-sq Durbin-Watson
AIR PRODUCTS & CHEMICALS INC 0.004 -0.001 -1.965*** -0.491 0.53 0.48 2.61
ALBEMARLE CORP 0.023 -0.004 -1.959*** -0.395 0.36 0.28 2.23
AMERICAN VANGUARD CORP -0.008 -0.008* 0.886*** -1.627*** 0.03 -0.09 2.25
ARCH CHEMICALS INC 0.023 -0.007** -1.58*** -0.745 0.22 0.13 2.43
ASHLAND INC 0.025 -0.008** -1.92*** -1.254*** 0.50 0.44 2.02
AVON PRODUCTS INC 0.003 -0.000 -2.489*** -0.474 0.56 0.51 2.26
CABOT CORP 0.024 -0.006 -1.85*** -0.348 0.22 0.13 1.87
CALGON CARBON CORPORATION 0.011 0.002 -0.479 0.378 0.03 -0.08 2.79
CAMBREX CORP 0.016 -0.008 -3.155*** -0.697 0.29 0.21 2.08
CARLISLE COMPANIES INC 0.018 -0.002 -1.043* -0.031 0.21 0.12 2.14
CELANESE CORPORATION 0.028 -0.005 -3.355*** -1.455*** 0.62 0.58 2.21
CF INDUSTRIES HOLDINGS, INC. 0.014 0.006 -2.188*** -0.534 0.18 0.08 2.94
CHEMTURA CORPORATION -0.1*** 0.011** -4.198*** 2.164*** 0.06 -0.05 1.62
CHURCH & DWIGHT CO INC 0.012 -0.004** -0.452* -0.13 0.23 0.14 1.86
CLOROX CO 0.007*** -0.001 -0.33** -0.121 0.16 0.06 1.56
CYTEC INDUSTRIES INC 0.032* -0.002 -2.629*** -0.699* 0.36 0.28 2.01
DOW CHEMICAL COMPANY (THE) 0.032* -0.002 -2.629*** -0.699* 0.36 0.28 2.01
EASTMAN CHEMICAL CO 0.024*** 0.002 -3.225*** -0.138 0.36 0.29 2.32
ECOLAB INC 0.01 -0.004* -0.831*** -0.121 0.32 0.24 2.07
ELIZABETH ARDEN INC 0.032 -0.001 -1.331* -0.316 0.32 0.24 2.30
FERRO CORP 0.028 -0.012 -3.58*** -1.98** 0.37 0.29 2.02
FMC CORP 0.000 0.003 -1.402*** -0.016 0.17 0.07 3.02
H.B. FULLER COMPANY 0.014 -0.008** -1.524*** -0.694 0.41 0.34 2.55
HUNTSMAN CORPORATION 0.015 0.003 -1.8* -0.596 0.21 0.12 2.06
INNOSPEC INC. 0.053*** 0.008** -2.451*** -1.828*** 0.28 0.20 2.05
INTERNATIONAL FLAVORS & FRAGRANCES INC0.004 -0.002 -1.096*** 0.219 0.43 0.36 2.12
KOPPERS HOLDINGS INC. -0.003 -0.005 -3.243*** 0.072 0.51 0.45 1.93
KRONOS WORLDWIDE, INC. 0.08*** -0.004 -2.81*** -1.843* 0.16 0.06 2.08
LSB Industries, Inc. 0.017 -0.01** -2.486*** -1.077*** 0.41 0.33 2.41
LUBRIZOL CORP 0.038*** -0.003 -1.54*** -1.032*** 0.34 0.26 2.80
METHANEX CORPORATION 0.006 0.001*** -1.813*** -0.989** 0.50 0.44 2.51
MINERALS TECHNOLOGIES INC 0.008 -0.001 -1.157** -0.483 0.20 0.11 2.54
MONSANTO CO -0.014 -0.003 -1.117** -1.085** 0.12 0.01 2.34
NALCO HOLDING COMPANY 0.042*** -0.002 -1.863*** -0.223 0.34 0.26 1.96
NEWMARKET CORPORATION 0.052*** -0.007 -2.691*** -0.947* 0.39 0.32 2.42
OLIN CORP 0.016 -0.004 -2.247*** 0.008 0.43 0.36 2.57
OM GROUP INC -0.001 -0.004 -1.928** -0.744 0.30 0.22 2.47
OMNOVA SOLUTIONS INC 0.036 -0.007 -1.943** -0.349 0.39 0.32 2.36
POLYONE CORPORATION 0.035 -0.01 -2.391** -0.544 0.47 0.41 2.33
POLYPORE INTERNATIONAL, INC. 0.056*** 0.002 -0.938* -0.203 0.15 0.05 2.21
PPG INDUSTRIES INC 0.007 -0.005 -1.174*** 0.033 0.41 0.34 2.38
PRAXAIR INC 0.016*** -0.001 -1.233*** -0.237 0.48 0.42 2.29
PROCTER & GAMBLE CO 0.009 -0.001 -0.626*** -0.102 0.27 0.18 1.72
REVLON, INC. 0.079* -0.003 -2.39 -1.936 0.12 0.02 1.65
ROCKWOOD HOLDINGS, INC. 0.02 -0.013*** -5.33*** -1.745*** 0.64 0.59 2.50
RPM INTERNATIONAL INC. 0.019 -0.002 -1.819*** 0.07 0.45 0.38 2.34
SCHULMAN A INC 0.003 -0.008** -1.059* -0.455 0.14 0.03 2.07
SCOTTS MIRACLE-GRO COMPANY (THE) 0.008* -0.003** -0.813*** 0.637*** 0.09 -0.02 2.53
SIGMA ALDRICH CORP 0.011 -0.002 -1.771*** -0.413*** 0.47 0.41 2.33
SOLUTIA INC 0.034 0.001 -0.223 -0.311 0.05 -0.07 1.70
SPARTECH CORP -0.016 -0.009 -1.481 -0.189*** 0.16 0.06 1.80
STEPAN COMPANY 0.026 -0.00** -1.598** -0.475 0.38 0.31 1.67
TERRA INDUSTRIES INC 0.013 -0.006 -1.184** 0.242 0.14 0.04 2.24
TRONOX INCORPORATED 0.026 -0.001 -3.099*** 2.747*** 0.79 0.63 1.91
USEC INC -0.002 -0.009 -0.803 0.041 0.13 0.03 2.31
VALHI INC 0.068** 0.009 -2.056*** -1.924*** 0.00 -0.12 2.24
VALSPAR CORP 0.025*** -0.005*** -0.687*** 0.22* 0.24 0.15 2.53
ZOLTEK COMPANIES INC 0.006 -0.017*** -4.094*** -0.401 0.26 0.17 2.11
