2007 Annual International CHRIE Conference & Exposition
439
interest rate changes and foreign exchange exposure. These findings are also consistent with prior research that has
documented the importance of firm size and the foreign sales ratio as important determinants of hedging. Finally,
the quick ratio, as a substitute for hedging, provides weak evidence that firms with higher liquidity face greater
exposure. These firms rely more on internal financing and are better able to withstand unexpected shocks and
consequently, have a lower incentive to hedge.
To test the robustness of my results, I rerun Models 1-4 in Table 5 using alternative variables. First I
measure derivatives in alternative ways, as the total notional value of all derivatives (including speculative
derivatives and foreign currency derivatives), notional value of hedging only derivatives (including only foreign
currency derivatives), and the net notional value of hedging derivatives. In all these cases, my results are similar
except when I use the total notional value. The coefficient for derivatives is insignificant in Model 1 and Model 4 at
the 10% level. It appears that including speculative derivatives in the measure may contribute to noise in the
measure since these derivatives increase the volatility of earnings. Second, I perform the analysis over a shorter
sample period from 2003-2004. The results are similar for all models but stronger for Model 3 where derivatives,
interest coverage, firm size and foreign sales ratio are all significant at the 5% level when the number of
observations is increased. Third, I recomputed the exposure coefficient using the CRSP value-weighted market
index and rerun my regressions. Again the results are similar but weaker. More specifically, the coefficient on
derivatives is significant at the 10% level for Model 2 but insignificant in Model 1 and 3. However, Bodnar and
Wong (2003) show that the CRSP value-weighted index may be biased and therefore they recommend the CRSP
equal-weighted index to correct for potential bias. Finally prior research has identified a number of incentives to
hedge such as managerial risk aversion, information asymmetry, tax incentive, and operational hedging. I measure
these incentives using proxy variables including the number of CEO stock option and percentage of CEO share
ownership, number of analysts monitoring the firm (information asymmetry), tax loss carry-forwards (tax incentive)
and diversification (operational hedging). However, these variables do not add any explanatory power to the
regression and are found to be generally insignificant in all regressions. I also attribute this insignificance to sample
size limitations.
Do'stlaringiz bilan baham: |