高频数据下基于组合预测思想的统计套利策略创新设计与实证研究

采用两只股票的日数据和5种高频数据,借鉴组合预测思想,综合利用协整模型和新卡尔曼滤波模型,与统计套利策略具体目标相结合,设计出新统计套利组合策略,实证分析数据频率、策略选择对统计套利效果的影响。结果表明：运用高频数据及引入卡尔曼滤波模型均有效,但卡尔曼滤波模型与协整模型不存在明显优劣之分,选择组合策略是必要的;组合策略收益性显著优于采取单一模型的套利策略;组合策略下的套利组合随数据频率提高,收益率波动性更小、更稳定;组合策略接近市场中性,能很好地免疫市场风险。     

Jai Alai arbitrage strategies

This paper presents arbitrage and risk arbitrage betting strategies for Team Jai Alai. This game is the setting for the analysis and most results generalize to other sports betting situations and some financial market applications. The arbitrage conditions are utility free while the risk arbitrage wagers are constructed according to the Kelly criterion/capital growth theory that maximizes asymptotically long-run wealth almost surely.     

Long-Run Diversification Potential in Emerging Stock Markets

In this paper we use cointegration tests to examine the long-run diversification potential of 13 emerging capital markets. The Johansen [18] and Johansen and Juselius [19] cointegration procedures are applied to the U.S. and 13 emerging capital markets in three geographical regions of the world. None of the three regions examined possesses cointegrated markets. The lack of cointegration indicates that the correlation between returns from each market is independent of the investment horizon Return correlations using weekly data correspond to the long-run investment horizon correlation. Correlations among the returns from these countries are low on average and occasionally negative. The apparent independence of markets within these three emerging regions suggests that diversification across these countries is effective.     

Regime switches between dividend and bond yields

The relation between bond and equity returns serves as a proxy for estimating the premia investors' demand on their equity portfolio holdings and assessing the substitution effects between the two markets. With this in mind, we examine empirically the co-movements and the underlying information between equities and bonds. Our approach relies on the comparison between bond and dividend yields — a relation better known as the gilt-equity yield ratio–GEYR — by examining the characteristics of the cointegration relation between the bond and equity yields. In this context, this paper's contribution is that it lifts the restrictions of linearity both in the long-run cointegration relations and in the underlying short-run relations presented in the VECM. Specifically, we apply the regime-switching framework of Gregory and Hansen (Gregory, A. W. & Hansen, B. E. (1996). Residual-based tests for cointegration in models with regime shifts. Journal of Econometrics 70, 99–126) for the long-run equilibriums and the Markov Switching VECM, established by Krolzig (Krolzig, H.M., 1997. Markov switching vector autoregressions. Modelling statistical inference and application to business cycle analysis. Springer, Verlag) for the short run ones. Our aim is to examine the allocation of capital among the UK bond (or else, gilt) and stock markets for the period of 01-1987 to 01-2007, in a fashion that better reflects the structural breaks and regime shifts of the underlying market conditions. Our findings confirm the substitution effects among stocks and bonds in the long run and highlight the importance of market conditions for the allocation of capital among stocks and bonds.     

Price Linkages Between the US,Japan and UK Stock Markets

This study presents an empirical analysis of the short- and long-term relationships among stock prices in the US, Japan and the UK. We re-examine the evidence of market linkages and cointegration between S&P 500, Nikkei 225 and FTSE-100 stock indices. The results suggest that mature markets are cointegrated, indicating a stationary long-run relationship. Furthermore, Granger causality tests show a bi-directional causality between Nikkei 225–FTSE-100, and unidirectional causalities between S&P 500–FTSE-100 and S&P 500–Nikkei 225. These findings suggest that the potential for diversifying risk by investing in mature markets is limited.     

Cointegration and causality between macroeconomic variables and stock market returns

The purpose of this study is to investigate whether current economic activities in Korea can explain stock market returns by using a cointegration test and a Granger causality test from a vector error correction model. This study finds that the Korean stock market reflects macroeconomic variables on stock price indices. The cointegration test and the vector error correction model illustrate that stock price indices are cointegrated with a set of macroeconomic variables—that is, the production index, exchange rate, trade balance, and money supply—which provides a direct long-run equilibrium relation with each stock price index. However, the stock price indices are not a leading indicator for economic variables, which is inconsistent with the previous findings that the stock market rationally signals changes in real activities.     

On the stability of long-run relationships between emerging and US stock markets

While most studies have found no cointegration between emerging and US stock markets, some recent studies do find a long-run relationship exists between these markets. In view of these mixed findings, this study examines the stability of long-run relationships between a number of emerging stock markets and the US stock market using recursive cointegration analysis. The results show that no long-run relationship exists between emerging markets and the US market over most of the sample period throughout 1997. However, we do find clear evidence of cointegration in response to the recent global emerging market crisis in 1997–1998. We conclude that significant crisis events can change the degree of cointegration between international stock markets and, therefore, need to be taken into account in studies of long-run relationships between international stock markets.     

Common and country-specific components in national stock prices

Previous studies on cointegration among national stock markets show conflicting results. This paper designs a general state space model to investigate the importance of a common world-wide component and the existence of country-specific components in national stock market indices. For G7 countries, there exist country-specific permanent and transitory components. Based on a variance decomposition analysis, Germany, Italy, and Japan's country-specific permanent shocks account for about half of their total permanent shocks. Most of France, Germany, and the United Kingdom's transitory shocks are country-specific. G7 countries do not seem to cointegrate around one common stochastic trend, and potential long-run international diversification benefits still exist.     

The Asymmetric Wealth Effect in the US Housing and Stock Markets: Evidence from the Threshold Cointegration Model

Previous studies commonly use a linear framework to investigate the long-run equilibrium relationship between the housing and stock markets. The linear approaches may not be appropriate if adjustments from disequilibrium are asymmetric in both markets. Nonlinear adjustments are likely to be observed since the two markets respond rather differently to negative shocks where the stock market is more volatile but price rigidity is found in the housing market. In this paper, we firstly propose two hypotheses on the long-run equilibrium relationship of the US housing and stock markets, and then employ the threshold cointegration model to investigate the potential asymmetric relationships between the two markets. Our empirical results reveal that cointegration exists among the markets, but adjustments toward its long-run equilibrium are asymmetric. Further evidence points out that a rapid mean reversion occurs in one regime where the stock price outperforms the housing price, and no significant reversion is found in the other regime, supporting the hypothesis of the existence of an asymmetric wealth effect among the two markets in the US. Furthermore, evidence from the asymmetric vector error correction model shows that significant error corrections toward the equilibrium exist in the short run only when the stock price exceeds the real estate price by the estimated threshold level, reassuring the finding of the asymmetric wealth effect.     

Long-run dynamics of official and black-market exchange rates in Latin America

This paper attempts to investigate the long-run dynamic relationship between official and black-market exchange rates for four Latin America markets namely, Argentina, Brazil, Chile, and Mexico. We follow (Moore, M. J., & Phylaktis, K. (2000). Black and official exchange rates in the Pacific Basin: Some tests of dynamic behaviour. Applied Financial Economics, 10, 361–369.) and we distinguish between long-run informational efficiency and short-run predictability in a sense that these notions are compatible with cointegration and error-correction mechanisms (ECM). Our findings indicate a constant black-market premium for each country, which is taken as strong support for long-run informational efficiency between the official and black markets for foreign currency. In addition, the evidence of short-run predictability is not considered as a violation of market efficiency, but it is the outcome of optimal arbitrage by rational economic agents.     

Long-run dynamics of official and black-market exchange rates in Latin America

This paper attempts to investigate the long-run dynamic relationship between official and black-market exchange rates for four Latin America markets namely, Argentina, Brazil, Chile, and Mexico. We follow (Moore, M. J., & Phylaktis, K. (2000). Black and official exchange rates in the Pacific Basin: Some tests of dynamic behaviour. Applied Financial Economics, 10, 361–369.) and we distinguish between long-run informational efficiency and short-run predictability in a sense that these notions are compatible with cointegration and error-correction mechanisms (ECM). Our findings indicate a constant black-market premium for each country, which is taken as strong support for long-run informational efficiency between the official and black markets for foreign currency. In addition, the evidence of short-run predictability is not considered as a violation of market efficiency, but it is the outcome of optimal arbitrage by rational economic agents.     

Integration Between the Baltic and International Stock Markets

Using cointegration tests, this paper analyzes the existence of long-run relationships among Baltic stock markets and major international stock markets, including the United States, Japan, Germany, the United Kingdom, and France. Bivariate and multivariate cointegration tests indicate a common trend linking Latvia to European markets. Evidence indicates that the German market dominates this long-run relationship. In general, short-term Granger causality indicates causality running from the European markets to the Baltic markets, as well as among the Baltic states, excepting Latvian and Lithuanian short-term effects on the Estonian market. Overall, the results suggest that international investors can obtain diversification benefits given a long-term investment horizon because of the low degree of integration between the Baltic and international capital markets.     

Can earnings fixated investors survive in a competitive securities market? Implications for sustained price anomalies and mark-to-market accounting

We model the dynamic survival of earnings fixated investors in a competitive securities market that allows for learning and arbitrage and that is populated by heterogeneous investors. Our model is distinct from those based on aggressive trading by overconfident investors. We prove that in the absence of noise traders, rational investors will drive out earnings fixated investors from the market in the long-run. More interestingly, we show that in a market with noise traders, some proportion of earnings fixated investors survive in long-run equilibrium for all feasible model parameter values. Furthermore, under no circumstances can the earnings fixated investors be driven out of the market completely. On the contrary, for some parameter values, the earnings fixated investors drive out the rational investors entirely. These results rationalize the long-run sustainability of common pricing anomalies. They also highlight potential benefits to society of mark-to-market accounting.     

Economic interdependence and common stochastic trends: A comparative analysis between EMU and non-EMU stock markets

Whether economic interdependence among countries is a contributing factor to cointegration and common stochastic trends in international stock markets is indiscernible due to contradictory results from prior empirical work. This study aims to add clarity to this issue through a more distinct grouping of countries and methodological enhancements. A comparative analysis of cointegration is conducted between stock market price indices of major Economic and Monetary Union (EMU) and non-EMU countries. The conventional Johansen methodology is augmented with several diagnostic techniques (that have not been all inclusive in previous studies) to ensure the robustness of test results. Major findings pertinent to investors and policymakers are that economic interdependence appears to be the important contributing factor and that the U.S. stock market does not exert influences on long-run performances of other included stock markets. Furthermore, while the UK is not an EMU member, it may be viewed as a quasi EMU participant due to its stock market being cointegrated with and yet one of the common stochastic trends (besides those of Germany, Italy, and the Netherlands) within the EMU stock markets under investigation.     

Do the stock and CDS markets price credit risk equally in the long-run?

In this paper, we examine the existence and stability of the long-run equilibrium relation between the price of credit risk in the stock and CDS markets for a sample of non-financial iTraxx Europe companies during the 2004–2017 period. We show that standard cointegration tests with no breaks frequently fail to detect cointegration. Once we formally account for the breaks in the cointegrating vector, we are able to detect cointegration over the entire sample period for the vast majority of the companies considered. An application of these results to CDS-equity trading shows that the profitability of traditional trading strategies crucially depends on the presence of cointegration and on the stability of the cointegrating vector. Finally, we find that CDS illiquidity factors decrease the likelihood of the stock and CDS market cointegration.     

Long-Term Co-Memories and Short-Run Adjustment: Securitized Real Estate and Stock Markets

This study considers whether securitized real estate and stock markets have long-term co-memories and implications for short-term adjustment. Our results offer reasonable support for fractional cointegration (characteristic of a long memory process) between securitized real estate price, stock market price and key macroeconomic factors in some economies. The implication is that where fractional cointegration prevails, securitized real estate and common stocks are substitutable assets over the long run and these assets may not be held together in a portfolio for diversification purpose. Furthermore, short-run analysis indicates that the speed of adjustment towards the long-run equilibrium is faster for fractional integrated vector error correction model (FIVECM) than VECM as the former incorporates a long history of past cointegration residuals. Additional comparisons of the two models’ forecasting accuracy show that incorporating fractional cointegration in a VECM model improves the forecasting performance over conventional VECM models. Our results reinforce the notion that cointegration, fractional cointegration and short-run adjustment dynamics are important in understanding market integration/segmentation.     

WEBS,SPDRs, and country funds: an analysis of international cointegration

Prior empirical studies analyzing linkages between international equity markets have suffered because suitable real-world financial instruments representing national equity markets were not available for trading. In March 1996, World Equity Benchmark Shares (WEBS) began trading on the American Stock Exchange. WEBS are open-end index funds that trade like closed-end index funds; they are designed to closely track the international indices developed by Morgan Stanley Capital International. This study utilizes WEBS along with Standard & Poor’s Depository Receipts (SPDRs) to avoid the previously encountered problems associated with nonsynchronous trading, fluctuating foreign exchange rates, non-liquidity, trading restrictions, and index replication. Results indicate that substantial pairwise cointegration exists among the 18 market indices as well as between individual closed-end country funds and their own-country WEBS. In addition, Granger causality tests indicate the existence of short-term causal relationships, suggesting market inefficiencies and the possibility of short-run arbitrage opportunities.     

价格传导关系变化的实证检验

利用变结构协整分析价格指数间从1996年10月～2010年7月的长期均衡关系,并对变结构前后价格传导路径的统计显著性做Granger因果检验结果显示,近年来上下游价格传导的统计显著性有较大提高,需求因素在价格传导中的作用突出,但是生产资料价格向下游传导存在较长的滞后期。最后利用变参数模型的状态空间表示,动态考察价格指数序列间的相互作用,发现各传导路径作用强度的变化并不一致。     

An improved test for statistical arbitrage

We improve upon the power of the statistical arbitrage test in Hogan, Jarrow, Teo, and Warachka (2004). Our methodology also allows for the evaluation of return anomalies under weaker assumptions. We then compare strategies based on their convergence rates to arbitrage and identify strategies whose probability of a loss declines to zero most rapidly. These strategies are preferred by investors with finite horizons or limited capital. After controlling for market frictions and examining convergence rates to arbitrage, we find that momentum and value strategies offer the most desirable trading opportunities.     

Factors affecting the yields on noninvestment grade bond indices: a cointegration analysis

This study examines the long- and short-run dynamics of the yields on noninvestment grade indices. Utilizing cointegration techniques, the traditional yield spread model is found to be inadequate. A revised model finds a long-run relationship between noninvestment grade yields, Treasury securities, and default rates. Error correction models are formulated to model the short-run dynamics of different segments of the market. These models include a long-run equilibrium (between yields, default rates, and Treasuries), mutual fund flows, minor bond ratings, debt subordination measures, a stock index, and a January effect. Segmentation in the noninvestment grade market is also demonstrated.