The paper uses a reduced-form vector autoregressive framework to study the effects of quantitative easing and operation “twist”, as well as a conventional monetary expansion, on corporate bond yields and spreads. We construct rating- and maturity-based weekly bond portfolios using TRACE and simulate monetary policies as shocks to the Treasury yield curve. We find that none of the policies can persistently lower corporate spreads, and that operation twist is the only policy capable of lowering corporate yields. This latter finding can be accounted for by the operation twist’s ability to keep the monetary base constant and, therefore, to flatten the riskless yield curve without generating inflationary expectations.
As the US economy continued to falter during the financial crisis despite the near-zero federal funds rates, the Federal Reserve resorted to unconventional monetary policies. The Large-Scale Asset Purchases program, commonly known as quantitative easing (QE), was introduced in November 2008 and the Maturity Extension Program, known as operation “twist” (OT), was launched in March 2009. The objective of these policies was to reduce the yields on Treasuries as well as on corporate bonds, thereby lowering costs of borrowing. In light of this objective and the significance of debt financing for US firms, it is important to understand the expected effects of QE and OT on the yields and spreads of corporate bonds with different ratings and maturities. This is the goal of our paper.
Unlike much of the existing literature (e.g.,Gagnon et al., 2011, Krishnamurthy and Vissing-Jorgensen, 2011, Hamilton and Wu, 2012, Wright, 2012 and D’Amico and King, 2013), we do not conduct an ex-post analysis and instead opt for an a priori investigation of the effects that monetary policies could be expected to generate at the time of their implementation. Researchers have lamented that confounding effects, such as the simultaneity of macroeconomic shocks, complicate the evaluation of monetary policies (e.g., Gilchrist and Zakrajšek, 2013), and our a priori, full-sample approach allows us to circumvent these difficulties. We study the data from 2004 to 2012, which includes the financial crisis but it is not limited to it (e.g., differently from Wright, 2012). Finally, we construct corporate bond portfolios using raw transactions data, which allows us to obtain more accurate results vs. studies that have relied on indices (e.g., Longstaff, 2010).
The vast majority of monetary policy operations are implemented via purchases and sales of Treasuries, and previous research has shown that Treasury transactions affect Treasury yields (e.g., Gagnon et al., 2011, Krishnamurthy and Vissing-Jorgensen, 2011 and Hamilton and Wu, 2012). Accordingly, we simulate QE and OT, as well as a conventional monetary policy, as shocks to the Treasury yield curve. We then use a reduced-form vector autoregressive (VAR) framework to compute responses of each of the series in the system. Section 2 presents the methodology and describes our portfolio construction approach, Section 3 reports the results, and Section 4 concludes.
Our results suggest that the monetary policy’s overall ability to lower corporate bond yields and spreads is modest at best. This can be accounted for by the higher inflationary expectations associated with an expansion, as well as by signalling effects: any monetary expansion can be perceived as signalling a future economic slowdown, thereby increasing bond risk premia and thus offsetting any direct impact. All in all, this paper supports four concepts. First, an a priori approach to evaluating policies can help obtain accurate results in the presence of confounding effects that plague ex-post analyses. Second, the accuracy of results can be improved by avoiding the use of bond indexes and instead constructing bond portfolios from transactions data. Third, it is essential to base the investigation on a broad data sample that includes crisis as well as tranquil periods. Finally, signalling effects of monetary policy can dampen—and even reverse—the desired impact of the policy.
Our paper has identified unconventional policies with shifts in the Treasury yield curve, but it is well-known that such policies have been stated and implemented in terms of Fed balance sheet’s size and composition. Our goal in this paper has been to document how developments in the Treasury term structure may affect different segments of the corporate bond market. Even though we believe this remains an aspect of the transmission mechanism of monetary policy of key relevance, the issue of the real effectiveness of unconventional policies remains. However, a body of research has exactly addressed the issue of their transmission to the riskless yield curve, through event studies: see, e.g., D’Amico and King, 2013 and Gagnon et al., 2011. This literature has concluded that QE has affected the Treasury term structure. Moreover, Swanson (2011) has investigated the implications of OT and found non-negligible effects, with long-term Treasury yields lowered and short-term bill rates simultaneously increased by amounts comparable to the experiments we have performed in this paper.
The precise impact of monetary policies is likely to depend on the state of the economy at the date of the implementation of a policy shift. Thus, the results presented in this paper should be viewed simply as measuring the average effects of monetary policies on various segments of the corporate bond market. Future research should further advance our understanding of these effects by explicitly modelling multiple regimes.
