Commercial Paper Rates and Stock Market Excess Returns

Transcription

1 Journal of Finance and Investment Analysis, vol. 2, no, 213, ISSN: (rint version), (online) Scienress Ltd, 213 Commercial Paer Rates and Stock Market Excess Vichet Sum 1 Abstract This study investigates how commercial aer rates resond to the innovations in stock market risk remiums. The unrestricted vector autoregression (VAR) analysis of monthly data from 1997:1 to 212:M6 shows that the changes in the one-, two-, and three-month non-financial and financial commercial aer rates ositively resond to the innovations in the excess returns on the CRSP value weighted index. The resonse is esecially strong during the first few months following shocks to stock market risk remiums. The Granger-causality test results show the changes in commercial aer rates can be redicted by the excess returns on the CRSP value weighted index. The findings from this study rovide evidence that there is a link between equity market and commercial aer market. JEL classification numbers: G2, G12, G14 Keywords: stock market excess returns, commercial aer rates, VAR 1 Introduction There has been a growing interest to investigate the cross-market linkage and transmission because returns on equity and other debt instruments can be exlained by many common factors. Fama and French (1993) rovide emirical evidence of the linkage between equity and bond markets by showing that the market factor, size factor (SMB), growth factor (HML), and maturity, default risks and dividend yields are rime candidates in exlaining excess return variability of government bond returns. In addition, both stock and bond rices are affected by discount rate and exected inflation (Cambell & Ammer, 1993). Evidence of the linkage between stock and bond markets is also 1 University of Maryland, Eastern Shore. vsum@umes.edu Article Info: Received : December 2, 212. Revised : December 3, 212. Published online : February 1, 213

2 78 Vichet Sum emirically documented in studies conducted by Sum (212a and 212b) which show that stock and bond returns can be redicted by business and consumer confidence. Another study by Wilson and Jones (199) shows that January effect is observed in the returns on commercial aer and bond. Kairys (1993) finds a ersistent relationshi between stock returns and commercial aer rates. Although there are many factors exlaining rice and return behavior in the equity and debt markets, the overall financial market factor is one of the most common factors included in various asset ricing models including the single-index model (Share, 1966), three-factor model (Fama & French, 1992) and four-factor model (Carhart, 1997). Consequently, it is the roblem of this study to investigate how commercial aer rates resond to the innovations in stock market risk remiums. This study is unique because it alies the vector autoregression analysis to examine the imulse resonse functions of commercial aer rates to excess returns on the stock market. The vector autoregression analysis also allows the causality link between stock market excess returns and commercial aer rates to be established. This study is necessary because it furthers the understanding the common factor exlaining returns on the equity and commercial aer market. In addition, the results of this study rovide useful information for asset and risk management. 2 Method and Data Monthly data of stock market excess returns and changes in the one-, two-, and three-month non-financial and financial commercial aer rates from 1997:1 to 212:M6 are obtained from the CRSP database and Federal Reserve Bank rovided by the University of Pennsylvania s Wharton Research Data Services (WRDS). This study emloys the unrestricted vector autoregression (VAR) analysis to investigate if changes in the one-, two-, and three-month non-financial and financial commercial aer rates resond to the innovation in the stock market risk remiums. This study is also set u to test if stock market excess returns cause the changes in the one-, two-, and three-month non-financial and financial commercial aer rates. R mt R ft = α + i=1 λ i (R mt i R ft i ) + i=1 φ i CP t i + ε t (1) CP t = α + i=1 λ i (R mt i R ft i ) + i=1 φ i CP t i + ε t (2) where: R mt = Return on CRSP value-weighted index (%) in time t R mt i = Return on the CRSP value-weighted index (%) in time t-i R ft = One-Month Treasury bill rate (%) in time t R ft i = One-Month Treasury bill rate (%) in time t-i CP t = change in commercial aer rates in month t CP t i = change in commercial aer rates in month t-i

3 Commercial Paer Rates and Stock Market Excess 79 3 Results Before erforming the vector autoregression analysis, Schwarz's Bayesian information criterion (SBIC), the Akaike's information criterion (AIC), and the Hannan and Quinn information criterion (HQIC) tests are erformed to determine the aroriate length of lags to be included in the model. Two lags are recommended by the tests for the one-, two-, and three-month non-financial commercial aers and one-month financial commercial aer rates. Three lags are suggested for the two-, and three-month financial commercial aer rates. The results show that the changes in the one-, two-, and three-month non-financial and financial commercial aer rates ositively resond to the innovations in the excess returns on the CRSP value weighted index. The resonse is esecially strong during the first few months following shocks to stock market risk remiums. The Granger-causality test results show the changes in commercial aer rates can be redicted by the excess returns on the CRSP value weighted index. varbasic, r, cnf Figure 1: The Orthogonal Imulse Resonse Functions (OIRF) of the Changes in the One-Month Non-Financial Commercial Paer Rates to CSRP Value-Weighted Excess Table 1: Granger Causality Walt Tests R m R f ΔCP R m R f All ΔCP R m R f ΔCP All in one-month non-financial commercial aer rates.

4 8 Vichet Sum varbasic, r, cnf Figure 2: The Orthogonal Imulse Resonse Functions (OIRF) of the Changes in the Two-Month Non-Financial Commercial Paer Rates to CSRP Value-Weighted Excess Table 2: Granger Causality Walt Tests R m R f ΔCP R m R f All ΔCP R m R f ΔCP All in two-month non-financial commercial aer rates. varbasic, r, cnf Figure 3: The Orthogonal Imulse Resonse Functions (OIRF) of the Changes in the Three-Month Non-Financial Commercial Paer Rates to CSRP Value-Weighted Excess

5 Commercial Paer Rates and Stock Market Excess 81 Table 3: Granger Causality Walt Tests R m R f ΔCP R m R f All ΔCP R m R f ΔCP All in three-month non-financial commercial aer rates. varbasic, r, cf Figure 4: The Orthogonal Imulse Resonse Functions (OIRF) of the Changes in the One-Month Financial Commercial Paer Rates to CSRP Value-Weighted Excess Table 4: Granger Causality Walt Tests R m R f ΔCP R m R f All ΔCP R m R f ΔCP All in one-month financial commercial aer rates.

6 82 Vichet Sum varbasic, r, cf Figure 5: The Orthogonal Imulse Resonse Functions (OIRF) of the Changes in the Two-Month Financial Commercial Paer Rates to CSRP Value-Weighted Excess Table 5: Granger Causality Walt Tests R m R f ΔCP R m R f All ΔCP R m R f ΔCP All in two-month financial commercial aer rates. varbasic, r, cf Figure 6: The Orthogonal Imulse Resonse Functions (OIRF) of the Changes in the Three-Month Financial Commercial Paer Rates to CSRP Value-Weighted Excess

7 Commercial Paer Rates and Stock Market Excess 83 Table 6: Granger Causality Walt Tests R m R f ΔCP R m R f All ΔCP R m R f ΔCP All in three-month financial commercial aer rates. 4 Conclusion This study investigates how commercial aer rates resond to the innovations in stock market risk remiums. The unrestricted vector autoregression (VAR) analysis of monthly data from 1997:1 to 212:M6 shows that the changes in the one-, two-, and three-month non-financial and financial commercial aer rates ositively resond to the innovations in the excess returns on the CRSP value weighted index. The resonse is esecially strong during the first few months following shocks to stock market risk remiums. The Granger-causality test results show the changes in commercial aer rates can be redicted by the excess returns on the CRSP value weighted index. The findings from this study rovide evidence that there is a link between equity market and commercial aer market. Finally, the findings rovide useful information for asset and risk management. References [1] M. Carhart, M., On ersistence in mutual fund erformance, Journal of Finance, 52(1), (1997), [2] J. Y. Cambell and J. Ammer, What moves the stock and bond markets? A variance decomosition for long-term asset returns, Journal of Finance, 48(1), (1993), [3] E. F. Fama and K. R. French, The cross-section of exected stock returns. Journal of Finance, 47(2), (1992), [4] E. F. Fama and K. R. French, Common risk factors in the returns on stocks and bonds, Journal of Financial Economics, 33(1), (1993), [5] J. J. P., Kairys, Predicting sign changes in the equity risk remium using commercial aer rates, The Journal of Portfolio Management, 2(1), (1993), [6] W. F. Share, Mutual fund erformance, Journal of Business, 39(1), (1966), [7] V. Sum, Random effects of business and consumer confidence on stock market returns: Cross-sectional evidence. Working aer (213): htt://aers.ssrn.com/abstract= [8] V. Sum, Determinants of U.S. Government Bond Risk Premia. Working aer (213): htt://aers.ssrn.com/abstract= [9] J. W. Wilson and C. P. Jones, Is there a January effect in cororate bond and aer returns? The Financial Review, 25(1), (199),