تاثیر بانک فدرال رژیمهای سیاست پولی در گسترش مبادله نرخ بهره آمریکا

This paper analyzes the asymmetric impacts of various economic shocks on swap spreads under distinct Fed monetary policy regimes. The results indicate that (a) during periods of aggressive interest rate reductions, slope of the Treasury term structure accounts for a sizeable share of the swap spread variance although default shock is also a major player. (b) On the other hand, liquidity premium is the only contributor to the 2-year swap spread variance in monetary tightening cycles. (c) The impact of default risk varies across both monetary cycles and swap maturities. (d) The effect of interest rate volatility is generally more evident in loosening monetary regimes.

This study documents the asymmetric impacts of financial market shocks on the US interest rate swap spreads over different monetary policy regimes. Analyzing swap spreads is an interesting and important research agenda as swap spreads measure the price of interest rate swaps, which have experienced exponential growth over the past decade due to the widespread use by corporations and financial institutions for risk management. Increasingly municipal governments are embracing swaps too. For example, city council members in Durham, North Carolina voted on April 4, 2005 on whether to proceed with an interest rate swap arranged by Rice Financial Products. A similar swap deal was adopted by the city of Houston last year.1Fig. 1 shows the growth of interest rate derivatives from 1988 to 2004. Interest rate options increased from $316 billion to $27.1 trillion, while interest rate swaps soared from slightly over one trillion to $150 trillion.2 The explosive growth in interest rate swaps speaks for the importance of understanding the key drivers in the swap pricing mechanism. In addition to the unprecedented popularity enjoyed by interest rate swaps for risk management, the financial market has increasingly used the swap curve for bond and derivative securities pricing due to the dwindling liquidity in the Treasury market.3A plain vanilla interest rate swap is a contractual agreement between two counterparties to exchange periodic interest payments based upon a certain amount of notion principal. Typically, the swap buyer makes a fixed interest payment in exchange for a variable cash flow based upon a floating London Interbank Offered Rate (LIBOR). The interest rate that determines the fixed payment is the swap rate. The swap spread is thus measured by the difference between the swap rate and the Treasury rate of similar maturity. In a complete financial market, arbitrage will ensure a zero swap spread. However, when the financial market is less than complete, and counterparty default risk exists, we observe a positive swap spread. Since the swap spread is the effective price of an interest rate swap and it changes over time, it is essential that we understand what sways its stochastic nature. There is a growing body of scholarship that addresses the changes in swap spreads. Treating a swap as an option to receive (pay) fixed and long an option to pay (receive) floating cash flows, Sorensen and Bollier (1994) argue that the value of a swap depends on the value of the option to default, which in turn is determined by the term structure of interest rates and interest rate volatility. Therefore, swap spreads are determined, partially, by the default risk which can be proxied by the interest rate volatility and the shape of the yield curve. On the other hand, Grinblatt (1995) contends that generic swaps are default-free. In particular, he attributes the swap spread to the liquidity differences between government securities and Eurodollar borrowings. Swap spreads contain a convenience yield (liquidity premium) which is absent in the more liquid government securities. Changes in liquidity premium, therefore, result in a corresponding change in the swap spread. The same argument that swaps are virtually default-free is echoed by Collin-Dufresne and Solnik (2001). Similarly, He (2000) opines that in practice interest rate swaps are close to default-free because the net interest payment streams involved are much smaller in a swap agreement than, say, in a bond. Empirical evidence to date examining the impact of liquidity and default risk premiums on swap spreads are somewhat mixed. Minton (1997) finds that corporate quality spread, a measurement of bilateral counterparty default risk, is not statistically related to the swap rates. However, aggregate default spread, a measurement of unilateral default risk is significantly related to the changes in swap rates. Huang, Neftci, and Jersey (2003) show that liquidity consideration is the primary driver of swap spreads. Liu, Longstaff, and Mandell (2004) report that although both default and liquidity premiums are salient components of swap spreads, the former is almost entirely a compensation for the variation in the liquidity component of the spreads, and the risk of changes in the default probability is virtually not priced by the market. Duffie and Singleton (1997) conclude that both credit and liquidity risks have impacts on swap zero spreads although the shock to swap spreads accounts for the majority of their own variations. Other studies confirming more or less similar views include Brown, Harlow, and Smith (1994), Lang, Litzenberger, and Liu (1998), Sun, Sundaresan, and Wang (1993), In, Brown, and Fang (2003), Lekkos and Milas, 2001 and Lekkos and Milas, 2004, Fehle (2003). Nevertheless, there are at least three issues being central to the swap spread analysis: some have been studied, yet others remain unaddressed. The first issue is what factors might have contributed to the dynamics of swap spreads. As discussed above, this question is perhaps one of the most researched swap spread topics so far, yet empirical findings are inconclusive. The second contention is the differential impacts of these factors on swap spreads across a spectrum of maturities. For example, Lekkos and Milas (2001) find that the impact from changes in the term structure on swap spreads is not uniform across swap maturities. They identify that an increasing yield curve slope is positively related to short-term swap spreads, but negatively related to the long-term swap spreads. The last issue deals with the underlying effect of economic/market conditions on swap spreads. Lang et al. (1998) report that swap spreads are procyclical. However, these studies investigate the direct impact of economic conditions on the swap spreads. Yet, the impact could be complicated. For example, while fixed rate payer gains value when the yield curve becomes steeper, the increasing default risk may offset this value enhancement. Therefore, a more interesting question that begs to be answered, then, is whether expectations of economic or financial market conditions may alter the structural model that determines swap spreads. Despite numerous debates over whether it is liquidity or credit risk that drives swap spreads (e.g., Huang et al., 2003), no attempts have been made to examine whether the relative importance of these factors and consequently the swap spreads generating process vary in different stages of market conditions. It is conceivable that one model fitting the data well under an expansionary monetary policy environment may work poorly under a restrictive policy environment. For instance, default risk could be important during one phase of the monetary cycle, but not in others. Similarly, liquidity premium may be substantial during high interest rate periods, but inconsequential during others. Failure to take any structural change into consideration will most likely produce data aggregation bias and result in wrongful statistical inferences. Addressing this issue is our primary objective and a distinguishing feature of this study. In a recent paper, Lekkos and Milas (2004) models swap spreads based upon a non-linear smooth transition VAR framework, which allows for different effects on swap spreads depending on whether the slope of the term structure is steep or flat. However, yield curve changes are often the consequence not the cause of changing monetary policies. Furthermore, the paper provides no error variance decomposition analysis and has a shorter sampling period excluding the most recent era of the Internet bubble burst. There are theoretical and statistical considerations on why the structural model for swap spreads may behave differently during disparate market conditions. One of the predominant factors that affect investor behaviors is their risk aversion. The argument of changing investor behaviors associated with varying risk appetite is supported by Jackwerth (2000), which discovers that the risk aversion function dramatically alters forms due to the 1987 stock market crash. Specifically, it's found that risk aversion function is positive and decreasing in wealth precrash and it becomes negative and increasing in wealth postcrash. The literature on swap spreads implicitly assumes a constant risk aversion coefficient over different market cycles. Since risk aversion is unobservable, it poses as an omitted variable problem in the swap spread model. Statistically speaking, the impact of an omitted variable should not be constrained to the intercept of the structural model; it also affects the regression parameters. Therefore, the marginal contributions of liquidity and credit risks on swap spread variations may differ during various market cycles due to changes in the investor's risk aversion level. Though not well understood, implied volatility smile in the options market also has changed its shape to volatility smirk after the Black Monday of October 1987. Asymmetric response of equity market performance to economic shocks during different economic cycles is reported in McQueen and Roley (1993), which argues that the stock market's response to macroeconomic news depends on the state of the economy. In this paper, we employ a vector autoregressive model to test the factors that compel the movement of swap spreads during various monetary policy regimes. Potential factors considered in the model include swap spreads, the Treasury term structure, corporate credit spreads, Eurodollar rates, and interest rate volatility. Impulse response functions and variance decompositions help enrich our analyses. Aggregating all time-series data, it is revealed that economic shocks other than swap spreads' own innovation account little for the spread variance. On the other hand, examining swap spread determinants under different Fed monetary policy regimes, however, we find that all relevant factors exhibit significant, albeit asymmetrical, impacts on swap spreads. Generally speaking, default premium, volatility, and the term structure slope are more important during phases of monetary loosening. During these periods, swap spreads' own innovation gives way to other economic shocks. The contribution of the term structure slope to the spread variance is particularly strong during the most recent aggressive interest rate cuts. Moreover, using impulse response and variance decomposition analyses, it is recognized that the effects of economic shocks on swap spreads display more dissimilarities across time periods than across swap maturities. We, therefore, observe an unstable structural model for swap spreads. Hence, failure to disaggregate the time series data will produce spurious conclusions. The rest of the paper is organized as follows. In Section 2, we offer some discussions on swap pricing and focus on the economic determinants of swap spreads. Data source and variable definitions are provided in Section 3. Section 4 presents the statistical methodologies and empirical results, and Section 5 draws the conclusions.

This paper aims to investigate the determinants of US interest rate swap spreads during different monetary cycles, which are classified according to the Fed's monetary policy regimes. The motivation is that monetary policy changes tend to exert an immediate impact on the financial market, and evidence of asymmetric impacts of expectations of economic states on financial markets is abundant. Chow test and CUSUM test lend further support in favor of our sample partitioning. Economic determinants of swap spreads considered in this study include default premium, liquidity premium, the term structure slope, and interest rate volatility. A vector autoregressive model combined with impulse response functions and variance decomposition method are employed for the analyses. Reaffirming Duffie and Singleton (1997), we find that swap spread explains the largest proportion of its own forecast error variance. Extending Duffie and Singleton (1997), we highlight the finding that the dominant role of swap spread's own shock varies substantially across time periods under different monetary policy regimes. Economic shocks other than swap spreads' own innovation account for little spread variances when data are aggregated. This bias no longer exists when the sample is disaggregated. Both the impulse response and variance decomposition analyses confirm significant yet differential impacts exerted by economic shocks on swap spreads across various monetary policy cycles. The key findings can be summarized as follows: (1) When the sample is partitioned in accordance with the monetary policy regimes, the Treasury slope, liquidity premium, interest rate volatility, and default premium all contribute to the swap spread variance depending on the time periods analyzed. (2) Although the swap spreads' own shock attains a dominating position in explaining the spread variance, the effect, however, is much smaller during periods of loose monetary policies. (3) In periods of aggressive interest rate reductions (e.g. period IV), Treasury slope accounts for a substantial portion of spread variance while default risk is also a major contributor. On the other hand, during periods of interest rate hike (e.g. period II), Treasury slope appears to play a more limited role. (4) In a tightening cycle of the monetary policy (e.g. period II), liquidity premium is the only contributor to the 2-year swap spread variance other than swap spreads' own shock. For the 10-year swap spread, liquidity premium, however, has more effect in period III. (5) The impact of default risk varies across both monetary cycles and swap maturities. Specifically, default risk plays an important role in 2-year swap spreads during periods of weak economic activities, which correspond to monetary loosening regimes. On the contrary, it shows minute effect during periods of increasing or stable interest rates. With 10-year swap spreads posing a different investment horizon, the relative contribution of default premium becomes more significant under stringent monetary policies. (6) The importance of interest rate volatility in the 2-year swap spreads analysis is more visible during periods of loose monetary policies, but almost non-detectable with the enforcement of tight monetary policies. This paper, therefore, documents some asymmetrical impacts of economic shocks on swap spreads. In this regard, the empirical results share the same spirit of McQueen and Roley (1993), which find that whether the stock market responds to macroeconomic news depends on the state of the economy. Our findings are also in line with Jackwerth (2000), which points out that risk aversion function changes direction after a market crash. It seems justifiable to say that analyzing the impact of economic shocks on swap spreads will produce dubious statistical inferences if time series data are aggregated over monetary cycles.