url
27
Apr
2014

ON THE DETERMINANTS OF DERIVATIVE HEDGING BY INSURANCE COMPANIES: Multivariate Results

ON THE DETERMINANTS OF DERIVATIVE HEDGING BY INSURANCE COMPANIES: Multivariate ResultsTables IV present the cross-sectional and panel data results of the participation decision analysis for the life and non-life insurance sectors over the 2001-2003 period. Note that the non-life results for the year 2001 are not reported. This is because none of the non-life insurers employed derivative instruments in 2001. Nearly all the independent variables in the cross-sectional analysis are highly significant, while in the panel data analysis some of the independent variables are not significant. There could be two reasons for this. First, each period has its own features which may feed through to the results in the former analysis. Second, the sample size is relatively small in the cross-sectional analysis so the results may be influenced by extreme values. Note that the sign and significance of the coefficients of some of the variables change from one period to another. Relative to the panel data analysis, the cross-sectional analysis may yield less efficient estimates due to less degrees of freedom. We will concentrate on the panel data results to simplify the exposition.
The coefficient for the firm size variable is positive and significant for the life sector, which supports the notion that informational economies and economies of scale are more important than the costs associated with financial distress. This result is consistent with prior studies such as Hoyt, Colquitte and Hoyt and Cummins, Phillips, and Smith which find that life insurers with greater size tend to be involved in hedging activities.
As to the impact of durations of assets and liabilities on derivative use, as expected we find that the larger the duration gap, the more likely it is that life companies use derivatives. This suggests that life companies which are more exposed to interest rate risk employ derivatives to manage this risk. This has been documented in prior studies, such as Colquitt and Hoyt and Cummins, Phillips, and Smith. It is not surprising that the duration gap between assets and liabilities is not significant in the non-life sector as they write short-term insurance business and therefore duration gap is less relevant than for their life insurance counterparts.
Contrary to the underinvestment hypothesis that firms with greater growth opportunities tend to engage in more derivative transactions, the coefficient for growth opportunity is found to be negative and significant for the life sector. This evidence contradicts the findings of Nance, Smith, and Smithson and Froot, Scharfstein and Stein and may be due in part to the long-term nature of life insurance.
The sign of the tax loss carried forward coefficient is positive for non-life companies, as expected, but negative for life companies. The negative sign for life companies is inconsistent with prior studies (Colquitt and Hoyt, 1997; Cummins, Phillips, and Smith, 1997; Geczy, Minton, and Schrand, 1997; Gay and Nam, 1998) which argue that firms with higher tax preference items such tax loss carried forward and other tax credits are more likely to use derivatives. A possible reason for our result is that life insurers with greater tax losses carried forward have suffered losses in the recent past and may tend not to participate in derivative activities for fear of a further loss.
The coefficients for the managerial ownership variable are negative and significant. How can these unexpected results be explained? Since most insurers in Taiwan are small unlisted companies whose stocks are owned by a limited number of shareholders, the size effect discussed above may be the dominant factor.
The foreign exposure coefficients are positive and significant, as expected. Up to 95 percent of derivative instruments used by insurance firms are currency forward contracts designed to hedge against the volatility of foreign exchange. This finding is consistent with prior studies, such as Davies, Eckberg and Marshall, which argue that firms with higher levels of foreign sales are more likely to hedge currency exposure.
The coefficients for interest rate risk exposure in the non-life sector and investment-linked insurance products sold by life insurance companies are both positive and significant. This suggests that risks associated with changes in interest rates and sale of investment-linked products are managed by derivative instruments. this

Table IV Participation Decision Model for Life Insurance Sector

Variable ExpKtedSign 2001-2003 2001 2002 2003
Coefficient г-statistic Coefficient z-statistic Coefficient ^-statistic Coefficient r-stattstic
Constant -54.3 88T -6.2711 -89.1173 -111.7610 -100.3456 -129.8170″ -79.1886 -109.8440
SIZE 2.9610 6.7791 4.3887 88.9300 -02459 -4.4730 5.9948 1219290
FIN 0.0000 13176 0.0000 -25.4860 0.0000 -40.3360 0.0006 154.0110
MISAST -0.0559 -02447 -0.0349 -0.4040 03942 27.4100 0.3286 259410
MBUA 0.0510 25690″ 0.0347 7.4750 02759 43.2020 -0.1602 -18.8570
GROWTH -0.0870 -35503 -0.1228 -24.8960 -0.7200 -94.6700 -0.2857 -125.8390
REINS 0.0386 02677 0.2301 32.6930 0.7243 93.5180 -0.2752 -475090
TLCF -0.2209 -15396 -0.1666 -10.6660 -1.1400 -94.3760 1.Q271 21.6830
CH 2.0838 3.0546 -13.1629 -60.2730 15.8909 93.5840 13.2944 535090
h£H -248.9013 -4.1696 -117.4013 -7.3470 126.0124 7.4290 -221.7411 -119750
FX 37.1100 5.1002 83.3244 52.1130 253.0621 166.2440 48.6293 46.6830

Tables IV present the cross-sectional and panel data results of the participation decision analysis for the life and non-life insurance sectors over the 2001-2003 period. Note that the non-life results for the year 2001 are not reported. This is because none of the non-life insurers employed derivative instruments in 2001. Nearly all the independent variables in the cross-sectional analysis are highly significant, while in the panel data analysis some of the independent variables are not significant. There could be two reasons for this. First, each period has its own features which may feed through to the results in the former analysis. Second, the sample size is relatively small in the cross-sectional analysis so the results may be influenced by extreme

About The Author

Kevin J. Brandon

Home | Site Map | Contacts

Copyright © 2013 - 2019 Investment And Finance Online. All rights reserved