Journal of Financial Risk Management
Vol.3 No.3(2014), Article ID:49546,17 pages DOI:10.4236/jfrm.2014.33010

Foreign Currency Derivatives and Firm Value: Evidence from New Zealand

Hao Li1, Nuttawat Visaltanachoti1, Robin H. Luo2*

1School of Economics and Finance, Massey University, Auckland, New Zealand

2Faculty of Business, ALHOSN University, Abu Dhabi, United Arab Emirates

Email: *robin.h.luo@gmail.com

Copyright © 2014 by authors and Scientific Research Publishing Inc.

This work is licensed under the Creative Commons Attribution International License (CC BY).

http://creativecommons.org/licenses/by/4.0/

Received 12 July 2014; revised 15 August 2014; accepted 28 August 2014

ABSTRACT

The benefit of corporate hedging remains controversial. While hedging could reduce the likelihood of adverse outcome, it will incur additional costs that may offset such benefit. This study provides some evidences to resolve the debate. We examine the benefits of foreign currency derivatives usage in 134 non-financial firms listed on the New Zealand Stock Exchange. New Zealand dollar experiences relatively high volatility so it is an ideal setting to examine whether the currency derivative usage could add value to the firm. Using Tobin-Q and other of its variants as a proxy of firm value, we find no evidence supporting the notion that the use of foreign currency derivatives can enhance a firm value.

Keywords:Currency, Derivatives, Risk Management

1. Introduction

Corporate hedging, commonly referred to risk management, is actively practiced in both large and small economies (Prevost, Rose, & Miller, 2000), however, the question “Why do firms hedge?” is still asked daily. There are two schools of theory that demonstrate why firms use corporate hedging. One is based on owner’s diversification motives or corporate managers’ personal utility maximisation. It suggests that the main purpose of corporate hedging is to reduce the likelihood that managers will suffer from adverse outcomes from uncertainties. The other is based on shareholder value maximisation. It implies that corporate hedging can reduce the likelihood that firms will encounter financial difficulties. By hedging those possibilities, firms could improve their ability to finance all of their profitable investment opportunities, hence maximising shareholders value.

In contrast, there are also reasons why firms do not hedge. Transaction costs of hedging, such as commissions paid to dealers, bid-ask spread and transaction fees charged by OTC or stock exchange, are the first concern. Secondly, in order to accomplish the purpose of hedging, firms have to evaluate the trade-off between costs and benefits of any particular hedging strategy. Such evaluation requires expensive financial professionals to participate and is time-consuming. Moreover, hedging strategies, especially hedging with derivatives, involve complex transactions; firms must frequently monitor those transactions and implement internal control of those transactions. Because derivatives are risky, any inappropriate transaction that violates hedging objectives could bring about serious financial problems. Finally, it may complicate firms’ financial reports. Preparations and adjustments for tax and accounting consequence of derivatives hedging have to be implemented and they bring extra cost. lso significant at the 10% level. But the ratio of capital expense to sales, debt to equity ratio and ratio of foreign sales to total sales are not significant at all three levels.

In column 2, three variables are very prominent, namely, ROA, R & D over total assets and advertising over total assets. Only R & D over total assets has the expected sign, while the other two have negative signs. Earlier in the investment growth section of control variable description, the cautiousness of using advertising expenses was discussed. Since a number of listed firms did not disclose their advertising expense separately, the advertising expenses for those particular firms were recorded as zero. This could create a significant bias and is shown in this OLS regression. A possible explanation for the negative ROA sign may be the fact that firm market value is largely dependent on its ability to generate future cash flow, not its current period profit. The other two variables, size and the dividend dummy, have the expected sign for coefficient like stated by Lang and Stulz (1994).

Another point of attention is that there are a number of extreme Q values in the original dataset. These extreme values could affect the test, therefore certain adjustments are needed. Like in the earlier univariate test section, firms with Q values that are greater than 4 are eliminated and left blank cells for regression.

Table 7 has a similar format to Table6 It presents the summary statistics of OLS regression that are generated by Eview for each control variable. White Heteroskedasticity is controlled during processing the regression. Again each control variable is listed in column 1. This time, only the ratio of capital expenses to total sales and R & D expenses over total assets are significant at all three levels. ROA, advertising over total assets and debt to equity ratio are significant at the 10% level. Size is significant at both the 5% and 10% levels. In column 2, all variables have the expected sign as Lang and Stulz (1994) stated, except for ROA and the ratio of advertising expenses to total assets. After eliminating the effect of extreme Q values, the magnitude of the ROA coefficient is smaller than before but still has a negative sign. Four control variables, namely, diversification dummy, dividend dummy, ratio of foreign sales to total sales and debt to equity ratio, are not significant at all three levels. This implies that all four of these control variables may not have any direct relationship with firm market value

Table 6 . Simple Q. 

This Table shows the following univariate regression: Q = β0 + β1*X with 134 observations. T-statistics are computed from White heteroskedasticity-consistent standard errors.

Table 7. Adjusted Q. 

This Table shows the following univariate regression: Adjusted-Q = β0 + β1*X with 134 observations. T-statistics are computed from White heteroskedasticity-consistent standard errors.

for New Zealand firms. Further, the ratio of foreign sales to total sales, which represents the geographic diversification, does not have statistical significance both before and after adjusting Q. The ratio of advertising expenses to total assets still has a negative sign. Due to its uncertainty status, this ratio should not be included in the multivariate test. Therefore, five of the following control variables will not be included in the multivariate test: 1) diversification dummy; 2) dividend dummy; 3) ratio of foreign sales to total sales; 4) debt to equity ratio; and 5) ratio of adverting expenses to total assets.

Table 8 presents the summary statistics of OLS regression for the dependent variable FCD dummy. White Heteroskedasticity is controlled while processing the regression. The explanatory variable represents the dependent variable FCD dummy with respect to different Tobin’s Q. One interesting factor in row (a) is that the FCD dummy has a negative sign for its coefficient. This implies that hedging with FCDs is negatively related to firm market value. This is quite contrary to the risk management theories. However, one thing should be considered, and this is that there are a number of extreme Q values in the original dataset. These extreme Q values could affect the multivariate test, therefore certain adjustments are needed. Like in earlier sections, firms with Q values greater than 4 are eliminated and blank cells are left for regression. In row (b), the coefficient of major dependent variable FCD dummy becomes positive this time. This is consistent with the corporate risk management theory although without statistical significance. Therefore, adjusted Q is used in the following multivariate test with its respect dependent and control variables.

Table 8. Main dependent variable. 

This Table shows the following univariate regression: Tobin-Q = β0 + β1*X with 134 observations. T-statistics are computed from White heteroskedasticity-consistent standard errors.

4.2.2. Multivariate Test

Table 9 presents the summary statistics of OLS regression that are generated by Eview for the multivariate test. The dependent variable is the adjusted Q; FCD represents FCD dummy; ROA represents return on total assets ratio; Growth represents the ratio of capital expense to total sales, and R & D/assets represents the ratio of R & D expenses to total assets. White Heteroskedasticity effect is controlled. Every control variable is statistically significant except for ROA. The ratio of R & D expenses to total assets has a very strong coefficient among all variables. The coefficient of FCD dummy is greater than single OLS regression, but it still does not have any statistical significance. This is consistent with the univariate test, which shows that statistically it is not certain whether there is any association between using FCDs and firm market value.

4.3. Robustness Checks

The methodology used in this study defines Tobin’s Q as the ratio of market value of debt and equity to a firm's total assets, which is also known as simple Q. Various other methodologies are used by prior researchers to state Tobin’s Q. Thereby, investigation of how sensitive our test results are to this Q measurement is necessary. In this section, two alternative measurements are constructed for robustness checks. The first alternative Q measurement follows the methodology from Lewellen and Badrinath (1997), which is denoted by QL in this section. According to Lewellen and Badrinath (1997), the replacement cost of assets is calculated by adding inventories to the sum of the replacement cost of fixed assets. The replacement cost of fixed assets is estimated by the book value of non-current fixed assets after current year depreciation adjustments. The replacement cost of inventories is the closing balance of the current year’s book value of inventories. Since LIFO is not allowed in New Zealand, the LIFO reserve effect does not need to be considered. The second alternative Q measurement is simply calculated by market value of debt and equity divided by total sales, which is denoted by QS in this section. Again, the book value of a firm’s total liability is used to represent debt value since the market value of each firm’s debt is quite hard to get.

Table 10 presents the comparison of alternative Q measures. QS has the same extreme value problems like simple Q encountered in the previous section. For example, one QS is 2638 and the other QS is 281, which are very different. After eliminating these two extreme values, the mean value of QS is reduced to 4.6648. Compared with the Q used in previous tests, both QL and QS have a higher mean value and a higher standard deviation. The mean QL is nearly double the mean QS, and the mean QS is nearly double the mean Q. There is definitely a steadily increasing spread among the three alternative Q measurements. The spread of QL is much wider than the spread of QS. For instance, the 10th percentile of QL is 1.4521 but its 90th percentile rose to 25.9908.

To reduce the wide spread and make alternative Qs more centralised, the natural log of QL and QS is used. After taking the natural log of two alternative Qs, their data spread is reduced and their mean value came closer to the practical value. Table 11 presents the results of OLS regression that uses the natural log of QL as the dependent variable. The total number of observations is still 134, while White Heteroskedasticity effect is controlled. Similar to the earlier multivariate test, the dependent variable FCD dummy still does not have any statistical significance but has the expected coefficient sign. Most control variables are statistically significant except for the ratio of R&D to total assets. Return on total assets still has a negative coefficient but with higher significance.

Table 12 presents the results of OLS regression that uses natural log of QS as the dependent variable. The total number of observations is still 134, while White Heteroskedasticity effect is controlled. Although most the control variables and dependent variable FCD dummy have opposite coefficients compared with the earlier multivariate test, nearly all of these variables including constant are not statistically significant at any of the three levels. Only the growth control variable has the expected coefficient sign and has statistical significance.

Table 9. Multivariate OLS regression.

Table 10. Alternative Q comparison.

This Table shows the following multivariate regression: Adjusted-Q = β0 + β1*X1 + ... + βk*Xk with 134 observations. T-statistics are computed from White heteroskedasticity-consistent standard errors.

Table 11. Multivariate regression for QL.

Table 12. Multivariate regression for QS.

This Table shows the following multivariate regression: Log(QL) = β0 + β1*X1 + ... + βk*Xk with 134 observations. T-statistics are computed from White heteroskedasticity-consistent standard errors.

This Table shows the following multivariate regression: Log(QS) = β0 + β1*X1 + ... + βk*Xk with 134 observations. T-statistics are computed from White heteroskedasticity-consistent standard errors.

The results of the robustness checks show that both alternative Q measurements have similar test results as simple Q, which is used in the earlier multivariate test. In the two above OLS regressions, the two coefficients of FCD dummy are nearly zero, which have less magnitude than the earlier multivariate test. Although the FCD dummy in all three Q measures does not have statistical significance, the P-value of FCD dummy in the earlier multivariate test is closer to the 10% statistical significance bench mark. Statistically, no matter which kind of Q measurement methodology is used, it is not certain whether there is any association between hedging with FCDs and firm market value.

5. Conclusion

This study examines whether the use of FCDs can cause higher finn market value within New Zealand. A new dataset is built based on 2007 annual reports of 134 non-financial finns listed in the New Zealand Stock Exchange. Tobin’s Q is used as an approximation for finns’ market value.

The original dataset is separated into two groups, one with foreign sales and the other one without. The main hypothesis is divided into two sub-hypotheses based on the finns’ foreign sales status, which allows the univariate tests to have a clear outline. Both null sub-hypotheses, for different reasons, cannot be rejected. The first sub-hypothesis cannot be rejected due to a lack of statistical significance. The second sub-hypothesis cannot be rejected due to the fact that T-statistics do not satisfy T-test requirements. Therefore, statistically, there is no direct relationship between the use of FCDs and firm market value for firms without foreign sales. For firms with foreign sales, statistically, it is not certain if there is any causality between the use of FCDs or firm market value either. Due to the possibility that extreme Q values in the target dataset may affect the tests results materially, adjusted Q is used to further test two-hypotheses. After eliminating Qs that are greater than 4, both null hypotheses still cannot be rejected due to a lack of statistical significance. There is a possible reason that other control variables may affect the univariate tests, so multivariate tests are utilised to further test the main hypothesis. After testing the relationship of Q with each control variable, five proposed control variables are eliminated from the final OLS regression. However, whether using simple Q or adjusted Q, the multivariate tests still show that there is no causality between the use of FCDs and firm market value.

In addition, by considering the possibility that different methodology of calculating Q may alter the tests results, the robustness checks are carried out. Two alternative methodologies of Q measurement are introduced. One, denoted by QL, follows Lewellen and Badrinath’s (1997) methodology. The other one, denoted by QS, is simply calculated by the ratio of market value of debt and equity to total sales. After modifying extreme Qs for both measures, OLS regression is used to test whether there is any association between the use of FCDs and firm market value. Both OLS regressions have similar results to the earlier multivariate tests. These regressions show that there is no clear association between the use of FCDs and firm market value.

There is no evidence to support the value added benefits of hedging with FCDs in this study. This shows that, statistically, there is no evidence that the use of FCDs can cause higher firm market value for New Zealand firms. This finding is consistent with findings of Guay and Kothari (2003), which shows that the use of derivatives is of minor economic significance. One possible reason for this result is that the benefit of using FCDs may be offset by their high initial set up price and transaction costs. Another reason could be that of New Zealand’s unique status. Only a few multinational firms list their stocks crossly in both their home country and New Zealand. Finally, since creating a new dataset takes a lot of time and effort, a single cross sectional dataset may not serve this study well. Panel datasets with more than 5 years time series data could be considered in future research to see if there is any improvement. Therefore, this study encourages further risk management empirical study with respect to this small and deregulated economy of New Zealand.

References

  1. Allayannis, G., & Weston, J. (2001). The Use of Foreign Currency Derivatives and Firm Market Value. Review of Financial Studies, 14, 243-276. http://dx.doi.org/10.1093/rfs/14.1.243
  2. Allayannis, G., & Mozumdar, A. (2004). The Impact of Negative Cash Flow and Influential Observations on Investment-Cash Flow Sensitivity Estimates. Journal of Banking & Finance, 28, 901-930.http://dx.doi.org/10.1016/S0378-4266(03)00114-6
  3. Bartov, E., Bodnar, G. M., & Kaul, A. (1996). Exchange Rate Variability and the Riskiness of U.S. Multinational Firms: Evidence from the Breakdown. Journal of Financial Economics, 42, 105-132. http://dx.doi.org/10.1016/0304-405X(95)00873-D
  4. Berger, P., & Ofek, E. (1995). Diversification’s Effect on Firm Value. Journal of Financial Economics, 37, 39-65.http://dx.doi.org/10.1016/0304-405X(94)00798-6
  5. Berkman, H., & Bradbury, M. E. (1996). Empirical Evidence on the Corporate Use of Derivatives. Financial Management, 25, 5-13.
  6. Blose, L. E., & Shieh, J. C. (1997). Tobin’s q-Ratio and Market Reaction to Capital Investment Announcements. Financial Review, 32, 449-476. http://dx.doi.org/10.1111/j.1540-6288.1997.tb00434.x
  7. Brown, G. W. (2001). Managing Foreign Exchange Risk with Derivatives. Journal of Financial Economics, 60, 401-448.http://dx.doi.org/10.1016/S0304-405X(01)00049-6
  8. Brown, G. W., Crabb, P. R., & Haushalter, D. (2006). Are Firms Successful at Selective Hedging? The Journal of Business, 79, 2925-2949. http://dx.doi.org/10.1086/508004
  9. DeMarzo, P., & Duffie, D. (1995). Corporate Incentives for Hedging and Hedge Accounting. Review of Financial Studies, 95, 743-771. http://dx.doi.org/10.1093/rfs/8.3.743
  10. Dolde, W. (1995). Hedging, Leverage, and Primitive Risk. Journal of Financial Engineering, 4, 187-216.
  11. Exporters Ready for Dollar Deals (2007). New Zealand Herald. p. B1.
  12. Froot, K., Scharfstein, D., & Stein, J. (1993). Risk Management: Coordinating Corporate Investment and Financing Policies. Journal of Finance, 18, 1624-1658.
  13. Geczy, C., Minton, B., & Schrand, C. (1997). Why Firms Use Currency Derivative? Journal of Finance, 52, 1324-1354.http://dx.doi.org/10.1111/j.1540-6261.1997.tb01112.x
  14. Graham, J., & Rogers, D. (2002). Do Firms Hedge in Response to Tax Incentives? Journal of Finance, 57, 815-839.http://dx.doi.org/10.1111/1540-6261.00443
  15. Guay, W., & Kothari, S. P. (2003). How Much Do Firms Hedge with Derivatives? Journal of Financial Economics, 80, 423-461. http://dx.doi.org/10.1016/S0304-405X(03)00179-X
  16. Haushalter, G. D. (2000). Financing Policy, Basis Risk, and Corporate Hedging: Evidence from Oil and Gas Producers. The Journal of Finance, 55, 107-152. http://dx.doi.org/10.1111/0022-1082.00202
  17. Leland, H. (1998). Agency Cost, Risk Management, and Capital Structure. Journal of Finance, 53, 1213-1243.http://dx.doi.org/10.1111/0022-1082.00051
  18. Mayers, D., & Smith, C. (1982). On the Corporate Demand for Insurance. Journal of Business, 55, 281-296.http://dx.doi.org/10.1086/296165
  19. Mian, S. (1996). Evidence on Corporate Hedging Policy. Journal of Financial and Quantitative Analysis, 31, 419-439.http://dx.doi.org/10.2307/2331399
  20. Nance, D., Smith, C., & Smithson, C. (1993). On the Determinants of Corporate Hedging. Journal of Finance, 18, 267-284.http://dx.doi.org/10.1111/j.1540-6261.1993.tb04709.x
  21. Prevost, A. K., Rose, L. C., & Miller, G. (2000). Derivatives Usage and Financial Risk Management in Large and Small Economies: A Comparative Analysis. Journal of Business Finance & Accounting, 27, 733-759.http://dx.doi.org/10.1111/1468-5957.00332
  22. Servaes, H. (1996). The Value of Diversification during the Conglomerate Merger Wave. Journal of Finance, 51, 1201-1225. http://dx.doi.org/10.1111/j.1540-6261.1996.tb04067.x
  23. Smith, C., & Stulz, R. (1985). The Determinants of Firms’ Hedging Policies. Journal of Finance and Quantitative Analysis, 20, 391-405. http://dx.doi.org/10.2307/2330757
  24. Stulz, R. (1984). Optimal Hedging Policies. Journal of Financial and Quantitative Analysis, 19, 127-140.http://dx.doi.org/10.2307/2330894
  25. Tufano, P. (1996). Who Manages Risk? An Empirical Examination of Risk Management Practices in the Gold Mining Industry. Journal of Finance, 51, 1097-1137. http://dx.doi.org/10.1111/j.1540-6261.1996.tb04064.x
  26. Visvanathan, G. (1998). Who Uses Interest Rate Swaps? A Cross-Sectional Analysis. Journal of Accounting, Auditing and Finance, 13, 173-200.

Appendix 1

NOTES

*Corresponding author.

1Allayannis and Weston (2001) and Guay and Kothari (2003).

2Sources: Statistics New Zealand: http://www.stats.govt.nz New Zealand; Foreign Trade Statistics, US census Bureau: http://www.census.gov/foreign-trade/www/index.html

3Source: New Zealand Herald, Page B1, “Exporters Ready for Dollar Deals” (26th August 2007).

4http://www.irg.co.nz/

5See 2007 Annual Report of Air New Zealand. page 35. Segmental Information.

6https://www.globalfinancialdata.com/index.html

7See 2007 Annual Financial Report for Fisher & Paykel Appliances, page 44.

8See example from 2007 Annual Financial Report for Fisher & Paykel Appliance, page 60, note 25

9According to Allayannis and Weston (2003), 63% of U.S. non-financial firms arc diversified across different industrial segment.