کانال های بین المللی از سیاست پولی غیر متعارف بانک فدرال

Previous research has established that the Federal Reserve's large scale asset purchases (LSAPs) significantly influenced international bond yields. We use dynamic term structure models to uncover to what extent signaling and portfolio balance channels caused these declines. For the U.S. and Canada, the evidence supports the view that LSAPs had substantial signaling effects. For Australian and German yields, signaling effects were present but likely more moderate, and portfolio balance effects appear to have played a relatively larger role than in the U.S. and Canada. Portfolio balance effects were small for Japanese yields and signaling effects basically nonexistent. These findings about LSAP channels are consistent with predictions based on interest rate dynamics during normal times: Signaling effects tend to be large for countries with strong yield responses to conventional U.S. monetary policy surprises, and portfolio balance effects are consistent with the degree of substitutability across international bonds, as measured by the covariance between foreign and U.S. bond returns.

In response to the extreme credit market disturbances in the fall of 2008, the Federal Reserve lowered the Federal funds rate target to near-zero, announced unprecedented bond purchases, and offered forward guidance to markets to reduce expectations of future short rates. Eventually, the Fed would announce three rounds of asset purchases that would total over $3 trillion from November 2008 through 2013. Federal Open Market Committee (FOMC) statements and speeches described the motives for these asset purchases in several ways but repeatedly returned to the themes of directly supporting credit markets—especially for housing—to reduce medium- and long-term U.S. interest rates in order to ultimately stimulate real activity. Other central banks, that is, the Bank of Japan, the Bank of England and the European Central Bank, would later initiate or expand similar programs. A growing literature studies the empirical effects of these unconventional policies. For the United States, the event study estimates of Gagnon et al. (2011) and Krishnamurthy et al. (2011) establish that the Fed's asset purchases strongly affected domestic bond yields. Neely (2013) finds that the purchases had substantial international effects on bond and foreign exchange markets. Announcements of large scale asset purchases (LSAPs) can affect government bond yields through both signaling and portfolio balance channels. The signaling channel implies that investors interpret asset purchase announcements as implying a lower path for future short-term interest rates, which reduces the expectations component of long-term interest rates.1 On the other hand, asset purchases can also affect prices of imperfectly substitutable assets through the portfolio balance channel. A purchase of U.S. bonds can reduce the term premia in both U.S. long-term yields and in international substitutes. A crucial question is how important signaling and portfolio balance channels are empirically for the effects of these asset purchases on government bond yields. For the U.S., the term structure estimates of Gagnon et al. (2011) appear to indicate that portfolio balance effects dominate, and these authors conclude that the signaling effects are negligible. On the other hand, Bauer and Rudebusch (2013b) and Christensen and Rudebusch (2012) find a substantially larger role for the signaling channel of asset purchase announcements.2Neely (2013) argues that the large impact of the Fed's LSAP announcements on international yields are consistent with a portfolio balance effect but he does not directly evaluate the relative importance of signaling/portfolio balance effects. There has been no serious analysis of the channels through which the Fed's LSAP announcements affected international bond yields. This paper aims to fill that gap by using term structure models to evaluate the relative importance of LSAP channels in mediating the impact of the Fed's asset purchases on international bond yields. In addition to U.S. yields, we study the effects on interest rates in Canada, Germany, Australia, and Japan.3 We consider announcements associated with the three LSAP programs during the period from 2008 to 2012: QE1, QE2, and QE3. Before presenting our results on the relative importance of LSAP channels of unconventional policy, we investigate what past data would lead us to expect for each country. We predict the impact of U.S. LSAPs on expectations of foreign short-term interest rates by analyzing how conventional U.S. monetary policy surprises affect foreign yields. For example, the strong reaction of Canadian yields to conventional U.S. monetary shocks implies a significant signaling effect for that country's markets. Analysis of the covariances between real foreign and U.S. bond returns predict that Australia and Germany would show the strongest portfolio balance channel effects. Using dynamic term structure models (DTSMs) we estimate changes in short-rate expectations and term premia around key LSAP announcements. Their respective contribution to the observed decreases in long-term yields is a measure of the importance of the signaling and portfolio balance channels. Importantly, changes in short-rate expectations should be viewed as conservative estimates of the importance of the signaling channel for two reasons: First, a successful monetary policy action aimed at easing financial conditions stimulates future growth and would raise short-rate expectations for the more distant future, counteracting the decreases in expectations due to signaling effects. Second, signaling near-zero policy rates would tend to lower interest rate risk and the term premium, even without any portfolio balance effects.4 The resulting inference can be quite sensitive to model choice. To guard against model-dependent conclusions and to obtain robust evidence, we estimate six alternative models for each country. In addition to a conventional maximally-flexible model, we correct for small-sample bias and restrict model parameters to obtain more reliable results. We evaluate the term structure models using criteria that include out-of-sample forecast accuracy. Models that impose greater peristence on the expected short rate or that restrict the dynamic evolution of the risk factors (as in Duffee, 2011) have the best forecasting performance. Based on the forecast accuracy, we can weight well-performing models more heavily, and can partly address the issue of model uncertainty. We find that the unconventional policy announcements had the most substantial signaling effects for the U.S. and Canada. In both countries, changes in expected future policy rates contributed significantly to lower long-term yields in those two countries. This finding holds for all three LSAP programs. The strong signaling effects on Canadian rates is consistent with the sensitivity of Canadian interest rates to signals from conventional U.S. monetary policy surprises. Overall, signaling effects likely accounted for a very substantial part of the sizable effects of the Fed's LSAP announcements on U.S. and Canadian long rates. For Australia and Germany, we also find strong signaling effects. They appear slightly more moderate than for the U.S. and Canada, however, and these estimates entail model uncertainty. Again, the results are quite consistent across the three LSAP programs.5 In these countries, portfolio balance effects probably played a relatively larger role than they did for the U.S. and Canada, which is consistent with the predictions based on interest rate dynamics during normal times. For Japan, signaling effects are negligibly small and portfolio balance effects can entirely explain the modest LSAP announcement effects.6 A very weak signaling channel parallels the weak reaction of Japanese yields to conventional U.S. monetary policy and the small portfolio balance effects are consistent with the relatively weak covariance between Japanese and U.S. long bond returns. This paper is part of a quickly growing literature on the effects of unconventional monetary policies on financial markets. In addition to the event studies of the LSAPs cited above, other papers include Joyce et al., 2011, D’Amico and King, 2013, D'Amico et al., 2012 and Hamilton and Wu, 2012, and Li and Wei (2013). Our paper is also related to the literature on the effects of (conventional) U.S. monetary policy surprises on international asset prices (Andersen et al., 2003, Faust et al., 2007, Ehrmann and Fratzscher, 2009 and Ammer et al., 2010). Two papers in particular study spillovers of conventional monetary policy surprises on international yields: Craine and Martin (2008) find that U.S. monetary policy surprises affect Australian interest rate and equity markets but Australian surprises do not measurably affect U.S. financial markets. Hausman and Wongswan (2011) establish that conventional U.S. policy surprises affect interest rates, equity prices, and exchange rates in 49 different countries. The strong effect on foreign interest rates that they document suggests a high potential for international signaling effects for U.S. monetary surprises, which is consistent with our findings. We extend this literature to study the channels through which unconventional policies affect international yields. The paper is structured as follows: Section 2 discusses details of the Fed's first LSAP program and the extent to which an event study approach can assess its effects on financial markets. Section 3 reviews the signaling and portfolio balance channels and predicts their relative importance with independent empirical evidence. Section 4 presents and discusses the DTSMs that we use for our empirical analysis. Section 5 contains the empirical results on the importance of signaling and portfolio balance channels for international LSAP effects. Section 6 concludes.

Previous research has found that the Federal Reserve's LSAP program strongly influenced international bond yields (see Neely, 2013). This paper investigates the relative importance of signaling and portfolio balance channels for the international bond yield effects of Fed unconventional policy announcements in 2008–2009. We draw conclusions both about methodology and the importance of signaling and portfolio balance channels. Our methodology demonstrates that the estimated changes in short-rate expectations and term premia are quite sensitive to model specification. We recommend that other researchers studying the term structure of interest rates also consider this model uncertainty, lest they have too much confidence in their (possibly misleading) results. In particular, the conventional (OLS) DTSM estimates would have led to very different conclusions than the ones we draw from a broader set of models. Models that impose high persistence in short rates, which naturally imply larger changes in short-rate expectations and stronger signaling effects, tend to forecast well and appear to be useful, although these are not widely used as of yet. Their good out-of-sample forecasting performance leads us to place more weight on inference from these models. Studying methods to increase statistical confidence—by imposing more parameter restrictions, using more data, and taking advantage of structural dependencies that are well-understood—should clearly be a high priority for the yield curve literature. Our empirical results show that both the signaling channel and the portfolio balance channel likely made substantial contributions to the decline in yields in most countries. For the U.S. and Canada, the evidence for pronounced signaling effects is strongest, and the results are consistent across all three LSAP programs considered. For Germany and Australia, there is also evidence for signaling effects, however with slightly more uncertainty surrounding our estimates. For Japan, the signaling effects are negligible, in line with Japanese rates already being very depressed at short and medium maturities. Our evidence indicates that portfolio balance effects were likely relatively more important for Australia and Germany than for the U.S. and Canada and that portfolio balance effects were modest for Japan. Overall, we find that the evidence on the relative importance of the international effects of the Fed's LSAP programs on foreign yields are largely consistent with past sensitivity to conventional U.S. monetary policy surprises and with the covariance of foreign and U.S. bond returns.