Our empirical result, firstly, points out DLI reveals an extremely significant positive relation to percentage bid-ask spread for each month in Table 4, proving DLI is one of the primary determinants of liquidity cost. This is to say, when financial condition is exacerbating, the percentage bid-ask spread increases, indicating financially troubled firms suffer the cost of reduced liquidity. The kind of relation between financial performance and equity liquid situation is consistence with the argument of Agrawal, et al., (2004) which demonstrates the higher spreads for firms experiencing financial dilemma are the result of an increase in informed trading in the stocks of these firms. It is more interesting that especially during the period of expected high financial distress costs, the increasing credit risk of poorly financial condition firms will contribute to the percentage bid-ask spread to raise more than usual period (see Table 6). Even though DLI is not huge distinction between period 1 and period 2, it does make higher liquidity cost in period 2. On an interesting basis we consider it is indeed in the period of Enron crisis that default risk obviously increases more liquidity cost. In addition, the credit risk of firms falling into very high default risk regime strongly increate more liquidity costs than those in low default risk regime, and this phenomenon is very obvious as the presence of high financial distress costs is severe.
From what we have mentioned above, the result of this paper asserts some important meaning for investors. In the beginning, as the presence of high financial distress costs is expected, default risk obviously increases the cost of liquidity more than time of lower financial distress costs. Because the relation between default risk and equity liquidity becomes stronger as the financial distress costs are higher, the happenings of firm’s bankrupt and enormous dump of prices are generally clustered,
especially firms with deteriorating financial condition. This is the reason why the worse equity liquidity is able to precipitate default of poor performance firms and corporate scandal and listed company bankruptcy events always appear in a clustered chain reaction. Moreover, dissimilar degrees of credit risk are able to manipulate the liquidity cost in enormously diverse weights.
Security liquidity measures the cost of taking ownership positions in firm’s equity. A stock with lower liquidity cost is always close to having excellent marketability that benefits investors greatly by allocating their ownership positions in lower cost. By and large, higher spread usually accompanies higher trading cost and bringing down the stock price. Vassalou and Xing (2004) assert stocks with both of higher default risk and smaller size or higher book-to-market ratio possess greater equity return. Moreover, according to our findings, it is suggested that higher default risk stocks in small size or high BM are particularly noteworthy when the mental situation of expected high financial distress generally exists in stock market. And auspiciously default risk level of all stocks could be classified into two or three classes according to the thresholds of DLI estimated in this paper. The stock price of those “very high default risk” category will go to down rapidly because of higher liquidity costs, so it will be correct and elegant strategy to sell or short those stocks lying in the “very high default risk” category.
REFERENCES
Agrawal, V., Kothare, M., Rao, R. K. S., and Wadhwa, P., 2004, “Bid-ask spreads, informed investors, and firm’s financial condition,” The Quarterly Review of Economics and Finance 44, 58–76.
Altman, E. I., 1968, “Financial ratios, discriminant analysis and the prediction of corporate bankruptcy,” Journal of Finance 23, 589–609.
Altman, E. I. 1984, “A further empirical investigation of the bankruptcy cost question,” Journal of Finance, 39, 1067–1089.
Amihud, Y., & Mendelson, H. 1980, “Dealership market: Market-making with inventory,” Journal of Financial Economics 8, 31–53.
Bagehot, Walter(pseud.), 1971, “The only game in town,” Financial Analysis Journal, 27, 12-14.
Benston, G., and Hagerman, R. (1974), “Determinants of bid-asked spreads in the over-the-counter market,” Journal of Financial Economics, 1, 353–364.
Black, Fischer, and Myron Scholes, 1973, “The pricing of options and corporate liabilities,” Journal of Political Economy, 81, 637–659.
Black, F. , 1986, “Noise,” Journal of Finance, 41, 529–544.
Brook, Y., & Rao, R. K. S. 1994, “Shareholder wealth effects of directors’ liability limitation provisions,” Journal of Financial and Quantitative Analysis, 29, 481–497.
Chan, K.S., 1993, “Consistency and limiting distribution of the least squares estimator of a threshold autoregressive model,” The Annals of Statistics ,21, 520-533.
Copeland, T. E., and Galai, D., 1983, “Information effects of the bid-ask spread,
“Journal of Finance, 38, 1457-1469.
DeAngelo, H. D., & DeAngelo, L. 1990, “Dividend policy and financial distress: An
empirical investigation of troubled NYSE firms,” Journal of Finance, 45, 1415–1432.
Demsetz, H., 1968, “The cost of transacting,” Quarterly Journal of Economics, 82, 33-53.
Easley, D., Kieffer, N. M., O’Hara, M., & Paperman, J. B. 1996, “Liquidity,
information, and infrequently traded stocks,” Journal of Finance, 51, 1405–1436.
Gilson, S. C. 1989, “Management turnover and financial distress,” Journal of Financial Economics, 25, 241–262.
Glosten, L. R., and Milgrom, P. R., 1985, “Bid, ask and transaction prices in a specialist market with heterogeneously informed traders,” Journal of Financial Economics, 14, 71-100.
Hansen, Bruce E. 1999, “Threshold effects in non-dynamic panels: Estimation, testing, and inference,” Journal of Econometrics, 93, 345-368.
Huang, R. D., and Stoll, H. R., 1994, “Market microstructure and stock return predictions,” Review of Financial Studies, 7, 179-213.
Huang, R. D., and Stoll, H. R., 1996, “Dealer versus auction markets: A paired comparison of execution costs on NASDAQ and the NYSE,” Journal of Financial Economics, 41, 13-357.
Kyle, A. S., 1985, “Continuous auctions and insider trading,” Econometrica, 53, 1315-1335.
Lang, L., Stulz, R. M., and Walking, R. A., 1989, “Managerial performance, Tobin’s Q and the gains from successful tender offers,” Journal of Financial Economics, 24, 137–154.
Lindenberg, E., and Ross, S., 1981, “Tobin’s Q ratio and industrial organization, ” Journal of Business, 54, 1–32.
Lin, J.-C., Sanger, G.C., and Booth, G., 1995, “Trade size and components of the
bid-ask spread,” Review of Financial Studies 8, 1153-1183.
Merton, Robert C., 1974, “On the pricing of corporate debt: The risk structure of interest rates,” Journal of Finance, 29, 449–470.
Odders-White, E. R. and Ready, Mark J., 2004, “Credit Ratings and stock liquidity,”
Review of Financial Studies, forthcoming.
Stoll, H. R., 1978a, “The supply of dealer services in securities markets,” Journal of Finance, 33,1122-1151.
Stoll, H. R., 1978b, “The pricing of security dealer services: An empirical study of NASDAQ stocks,” Journal of Finance, 33, 1153-1172.
Stoll, H. R., 2000, “Friction,” Journal of Finance, 55, 1479-1514.
Tinic, S., and West, R. 1972, “Competition and the pricing of dealer services in the over the counter stock market,” Journal of Financial and Quantitative Analysis, 7, 1707–1727.
Titman, S. 1984, “The effect of capital structure of a firm’s liquidation decision,”
Journal of Financial Economics, 13, 137–151.
Vassalou, M., and Xing, Y. 2004, “Default Risk in Equity Returns,” Journal of Finance, 59, 831-868.
Warner, J. B. 1977, “Bankruptcy costs: some evidence,” Journal of Finance, 32, 337–347.
Weiss, L. A. 1990, “Bankruptcy resolution: direct costs and violation of priority of claims,” Journal of Financial Economics, 27, 285–314.
White, H. 1980, “Aheteroskedasticity-consistent covariance matrix estimator and a direct test for heteroscedasticity,” Econometrica, 48, 817–838.
Wruck, K. H. 1990, “Financial distress, reorganization, and organizational efficiency,” Journal of Financial Economics, 27, 419–444.
TABLE 1 Descriptive statistics of selected variable
N Mean Median S.D Minimum Maximum
PSP(%) 4416 0.1285 0.1054 0.0837 0.0195 0.8140
CLP 4416 40.0159 38.1913 19.7966 2.8175 159.9965
TRADE 4416 1,068 859 681 162 5,005
SIGMA 4416 2.0782 1.8841 0.8603 0.4489 6.0690
MV(million’s) 4416 19,612 7,855 42,885 506 486,720
DLI(%) 4416 4.3084 0 14.8112 0 100
PSP = the monthly average percentage bid-ask spreads for company i CLP= the monthly average close price for company i
TRADE= the monthly average of the number of trade for company i
SIGMA= the standard deviation of daily stock returns for a month for company i MV= market value of company i
DLI= Default Likelihood Indicators of company i N= the number of observation
TABLE 2 Pearson correlations coefficients of selected variable
PSP CLP TRADE SIGMA MV DLI
PSP 1.00000 -0.5991 (<.0001)
- 0.0665 (<.0001)
0.4142 (<.0001)
-0.0991 (<.0001)
0.3041 (<.0001)
CLP 1.00000 0.1838
(<.0001)
-0.2895 (<.0001)
0.2110 (<.0001)
-0.0156 (0.2793)
TRADE 1.00000 0.2064
(<.0001)
0.6439 (<.0001)
0.0125 (0.3836)
SIGMA 1.00000 -0.0601
(<.0001)
0.2338 (<.0001)
MV 1.00000 -0.0616
(<.0001)
DLI 1.00000
TABLE 3 Cross-sectional regression results
The monthly regression of PSP against four control variables (logCLP, logTRADE, SIGMA, logMV) is estimated by ordinary least squares estimation.
Coefficients on independent variables
INTER logCP logNT SIG logMV
For each month the first row in the “coefficients on independent variables” column gives the estimated coefficient, and the second row is the t-values OLS S.E.
* Level of significance is 5%
** Level of significance is 1%
TABLE 4 Cross-sectional regression results (key regressor:DLI)
The monthly regression of PSP against four control variables (logCLP, logTRADE, SIGMA, logMV) and one key variable of interest (DLI) is estimated by generalized least squares (heteroskedastic).
5
Coefficients on independent variables
INTER logCP logNT SIG logMV DLI
For each month the first row in the “coefficients on independent variables” column gives the estimated coefficient, and the second row is the t-values OLS S.E.
* Level of significance is 5%
** Level of significance is 1%
TABLE 5 Results of panel data regression model (key regressor: PDLI)
Panel data regression of PSP against four control variables (logCLP, logTRADE, SIGMA, logMV) and one key variable of interest (logPDLI) is estimated by fixed effect approach. N=276
5 Prediction Independent variables Whole Period
T=16
N= the number of the sample firm T= the number of the sample month
Period 1: February 2001-September 2001, number of observations=2208 Period 2: October 2001-May 2002, number of observations=2208
* Level of significance is 5%
** Level of significance is 1%
For each month the first row in the “coefficients on independent variables” column gives the estimated coefficient, and the second row is the t-values OLS S.E.
TABLE 6 Results of single threshold model (key regressor: PDLI)
Panel data threshold regression of PSP against four control variables (logCLP, logTRADE, SIGMA, logMV) and one key variable of interest (PDLI) is estimated by fixed effect approach. And DLI is used to be threshold variable. N=276
it
γ 0.1880 0.1627 0.1073
F 50.8996** 31.5678* 63.9163 **
PDLI=100 × DLI
N= the number of the sample firm T= the number of the sample month
Period 1: February 2001-September 2001, number of observations=2208 Period 2: October 2001-May 2002, number of observations=2208
* Level of significance is 5%
** Level of significance is 1%
For each period the first row in the “coefficients on independent variables” column gives the estimated coefficient. In the next two rows, () and [] are the t-values respectively calculated by OLS S.E and White S.E.
γˆ is the estimator of threshold and if p-value for F is less than the desired critical value, the null hypothesis of no threshold effect is rejected.
TABLE 7 Results of double threshold model only for Period 2 (key regressor:
Prediction Independent variables Coefficient
- Log CP -0.1233
N= the number of the sample firm.
T= the number of the sample month
Period 2: October 2001-May 2002, number of observations=2208
* Level of significance is5%
** Level of significance is 1%
For each period the first row in the “coefficients on independent variables” column gives the estimated coefficient. In the next two rows, () and [] are the t-values respectively calculated by OLS S.E and White S.E.
γˆ is the estimator of threshold and if p-value for F is less than the desired critical value, the null hypothesis of no threshold effect is rejected.
Appendix
FIGURE A.1
PSP V.S. DLI(Mean)
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18
200103 200104 200105 200106 200107 200108 200109 200110 200111 200112 200201 200202 200203 200204 200205 Month
PSP%
0 1 2 3 4 5 6 7 DLI%
PSP%
DLI%
PSP V.S. DLI(Median)
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
200103 200104 200105 200106 200107 200108 200109 200110 200111 200112 200201 200202 200203 200204 200205 Month
PSP%
-0.001 0 0.001 0.002 DLI%
PSP%
DLI%
TABLE A.1
Descriptive statistics and t-test results of DLI
Panel B of Table A.1 shows the average DLI of all firms in period 1 and period 2 are 4.7616% and 3.9428% respectively. In Panel B of Table A.1, the t-value is -1.34 and the p-value is 0.1827 without less than 0.05. The t-test result means that DLI for period 2 fails to significantly differ from that for period 1.
Panel A: Descriptive statistics of DLI
DLI N Mean Median S.D. Minimum Maximum
Period 1 276 4.7616% 0.0829% 13.5339% 0.0000% 87.5001%
Period 2 276 3.9428% 0.0006% 12.4946% 0.0000% 87.5000%
Diff 276 -0.8187% -0.0006% 10.3646% -73.2764% 79.3695%
Panel B: T-test result of DLI
H0: the difference of DLI for company i between Period 1 and Period 2 equal zero
Variable N t value Pr > |t|
DLI(Diff) 276 -1.34 0.1827
N=the number of the sample firm Period 1: February 2001-September 2001 Period 2: October 2001-May 2002
Diff: the difference of average DLI between Period 2 and Period 1 for each firm
TABLE A.2
Tests for threshold effects
This table is to show whether the threshold effect has observable influence on coefficient estimator.If p-value for F is less than the desired critical value, the null hypothesis of no threshold effect is rejected. And γˆ is the estimator of threshold. The F statistics for threshold effect test is strongly significant with a bootstrap p-value under 0.05, telling us the threshold effect is functional.
Test for February 2001-May 2002(Whole Period) Single threshold model
γˆ1 0.1880
F 50.8996
P-value 0.0000
(10%, 5%, 1% critical values) (20.5702, 23.9565, 33.1793)
Test for February 2001-September 2001(Period 1) Single threshold model
γˆ1 0.1627
F 31.5678
P-value 0.01667
(10%, 5%, 1% critical values) (21.56139, 25.9425, 36.1869)
Test for October 2001-May 2002(Period 2) Single threshold model
γˆ1 0.1073
F 63.9163
P-value 0.0000
(10%, 5%, 1% critical values) (22.2123, 25.4175, 33.4855) Double threshold model#
γˆ1
γˆ2 0.1073 0.6341
F 26.0494
P-value 0.02
(10%, 5%, 1% critical values) (19.2688, 23.5110 , 27.9175)
# There is significant double threshold effect only in period 2 (October 2001-May 2002)
TABLE A.3
Percentage of firms in each regime by month: double threshold model
We find that the percentage of companies in the “very high default risk” category ranges from 0.36%
to 6.88% of the sample in each month. The “low default risk” firms range from 76.81% to 93.12% of the sample over the months. The average of percentage bid-ask spreads for each regime are 0.1197, 0.2087 and 0.2461 in that order. The “very high default risk” class has the highest average of percentage bid-ask spreads, 0.2461. This proves that the firms with higher credit risk will suffer higher liquidity cost.
Firms class DLI<=0.1073 ASPS=0.1197
0.1073<DLI<=0.6341 ASPS=0.2087
DLI>0.6341 ASPS=0.2461
Month No. ％ No. ％ No. ％
2001Feb 212 76.81% 45 16.30% 19 6.88%
2001Mar 257 93.12% 18 6.52% 1 0.36%
2001Apr 246 89.13% 24 8.70% 6 2.17%
2001May 252 91.30% 16 5.80% 8 2.90%
2001June 256 92.75% 13 4.71% 7 2.54%
2001July 255 92.39% 14 5.07% 7 2.54%
2001Aug 253 91.67% 17 6.16% 6 2.17%
2001Sep 250 90.58% 19 6.88% 7 2.54%
2001Oct 238 86.23% 31 11.23% 7 2.54%
2001Nov 243 88.04% 27 9.78% 6 2.17%
2001Dec 250 90.58% 20 7.25% 6 2.17%
2002Jan 250 90.58% 21 7.61% 5 1.81%
2002Feb 250 90.58% 21 7.61% 5 1.81%
2002Mar 253 91.67% 18 6.52% 5 1.81%
2002Apr 261 94.57% 13 4.71% 2 0.72%
2002May 257 93.12% 16 5.80% 3 1.09%
APSP= the average percentage bid-ask spreads for each regime No.= the number of firm