آیا بازده بازار اوراق قرضه شرکتی قابل پیش بینی است؟
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|13176||2012||17 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 15660 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Banking & Finance, Volume 36, Issue 8, August 2012, Pages 2216–2232
This paper examines the predictability of corporate bond returns using the transaction-based index data for the period from October 1, 2002 to December 31, 2010. We find evidence of significant serial and cross-serial dependence in daily investment-grade and high-yield bond returns. The serial dependence exhibits a complex nonlinear structure. Both investment-grade and high-yield bond returns can be predicted by past stock market returns in-sample and out-of-sample, and the predictive relation is much stronger between stocks and high-yield bonds. By contrast, there is little evidence that stock returns can be predicted by past bond returns. These findings are robust to various model specifications and test methods, and provide important implications for modeling the term structure of defaultable bonds.
One of the most enduring issues in finance and economics is the question of whether returns on risky assets are predictable. This important issue has been the focus of an extensive literature on asset prices dating back more than a century. Despite an enormous amount of past efforts, whether future asset price changes can be meaningfully predicted is still a subject of ongoing debates and intensive empirical research (see, for example, Ang and Bekaert, 2007, Campbell and Thompson, 2008, Welch and Goyal, 2008, Rapach et al., 2010 and Sekkel, 2011).1 The literature of asset return predictability has focused on the stock market. There is substantial evidence that stock returns are predictable, either by past price changes or economic variables (see Campbell et al., 1997, Ang and Bekaert, 2007, Campbell and Thompson, 2008 and Rapach et al., 2010). Recent efforts have been directed to identifying the predictive components of asset returns at different return horizons, evaluating the predictive power of predictors using more robust tests, and determining how much predictability is compatible with efficiency consistent with risk-based asset pricing models. Notwithstanding extensive research on equity return predictability, there are only a few studies on corporate bond return predictability (see Keim and Stambaugh, 1986, Kwan, 1996, Hotchkiss and Ronen, 2002 and Downing et al., 2009) and empirical evidence is inconclusive. Kwan (1996) shows that significant negative contemporaneous correlation exists between returns of individual stocks and yield changes of bonds issued by the same firm, and that stock returns predict future bond yield changes. Unlike Kwan, 1996 and Hotchkiss and Ronen, 2002 find that corporate bond returns cannot be predicted by past stock returns based on a sample of 20 high-yield bonds from the National Association of Securities Dealers (NASD). By contrast, Downing et al. (2009) show that stock returns predict convertible bond returns in all rating categories but predict returns of only BBB- and junk-rated nonconvertible bonds. In this paper, we examine the predictability of corporate bond returns in a narrow sense by focusing on serial dependence and causality tests. Similar to mainstream equity premium studies, we examine return predictability at the aggregate level. We employ bond market index data constructed from transaction prices, instead of dealer quotes used in a number of studies (see, for example, Kwan, 1996 and Gebhardt et al., 2005). Our empirical analysis draws heavily on the rich literature in random walk and causality tests (Granger, 1969 and Campbell et al., 1997). Similar to Chen and Maringer (2011), we account for nonlinearity in corporate bond index returns. Standard methods of return predictability tests are not robust to nonlinear dependence. To overcome this problem, we employ an advanced generalized spectral method (Hong and Lee, 2005) to detect nonlinear dependencies in returns and to perform robust tests. Furthermore, we conduct causality tests on bond and stock returns by taking into account heteroskedasticity in the error term and potential nonlinearity in the causal relationship. Knowledge of bond price dynamics is important for formulating optimal strategies for asset allocation and hedging. Corporate bonds account for a significant portion of investors’ wealth, with a market size near 6 trillion dollars (see Abhyankar and Gonzales, 2009), so understanding corporate bond price dynamics is essential for academics and practitioners. This paper, to the best of our knowledge, is the first that provides comprehensive time-series analysis on serial and cross-serial dependencies in transaction-based corporate bond index returns. We find strong evidence of serial and cross-serial dependence in corporate bond returns. Empirical analysis reveals a complicated nonlinear structure of serial dependence in corporate bond returns. Investment-grade and high-yield bond returns can be predicted by past stock returns both in-sample and out-of-sample, and the predictive relation is much stronger between stocks and high-yield bonds. By contrast, there is little evidence that stock returns can be predicted by past bond returns. These findings persist even after controlling effects of conditional heteroskedasticity, volatility-induced mean return changes, and time-varying interest rates. The remainder of the paper is organized as follows. In Section 2, we describe the hypotheses and methodology for testing linear and nonlinear serial dependence in returns. In Section 3, we propose vector autoregressive regression models (VAR) and Granger causality tests with homoskedastic and heteroskedastic returns. In Section 4, we present test results for serial and cross-serial dependence in stock and bond market returns and examine the robustness of results to different model specifications and return measures. In Section 5, we examine the sensitivity of corporate bond returns to concurrent and lagged stock and government bond returns. In Section 6, we conduct out-of-sample tests on return predictability. Finally, we summarize our findings and conclude the paper in Section 7.
نتیجه گیری انگلیسی
An issue central to financial research is the predictability of asset returns. There is substantial evidence that stock returns are predictable. An important question is whether returns are also predictable for other asset classes. This paper examines this issue for the corporate bond market and employs empirical methodologies that are robust to nonlinearity in serial return dependence and conditional heteroskedasticity. Empirical evidence strongly suggests that corporate bond market returns are predictable. There is evidence of return predictability for both investment-grade and high-yield bonds. These results are robust to alternative model specifications, return measures, and exclusion of extreme observations. Results show that corporate bond market returns exhibit higher autocorrelation and a more complicated structure of serial dependence than stock market returns. Stock market returns lead both high-yield and investment-grade bond returns, whereas there is little evidence that corporate bond market returns lead stock market returns. This lead–lag relation is stronger between high-yield bond returns and stock returns. The lead–lag relation remains strong when we control the effects of interest rates and serial and cross-serial dependence in bond returns. Moreover, out-of-sample tests show results consistent with in-sample tests. Results show that the past stock return is useful information for predicting both speculative- and investment-grade bond returns out of sample. Our findings provide important implications for corporate bond modeling and asset pricing tests. Results suggest that tests of the risk-return tradeoff in corporate bonds should take into account the cross-serial dependence between bond and stock returns. Our findings also impose restrictions on the specification of the term structure model. In particular, our results suggest that the term structure model of defaultable bonds should account for the serial and cross-serial dependence in corporate bond price changes in order to provide a more satisfactory explanation for corporate bond price behavior. The cause of the predictability of corporate bond returns is not immediately clear. Return predictability could be due to bond illiquidity, transaction cost, market structure, or other frictions. An exploration of the cause for our results is an important extension of this paper, and we leave this for a future study.