LING HU

Home Address:    420 Temple Street, Apt #429    New Haven, CT 06511    Tel: (203) 668-7870 Office Address:   Department of Economics   Yale University   Box 208268   New Haven, CT 06520-8268   Fax: (203) 432-5779 Birth Date: August 5, 1974 Citizenship: China

Fields of Concentration

Econometrics Financial Economics Empirical Finance

Desired Teaching:

Econometrics Financial Economics Corporate Finance Fixed Income Security Analysis Investment Management

Comprehensive Examinations Completed:

May 1999 (Oral) Econometrics, Finance May 1998 (Written) Microeconomic and Macroeconomic Theory

Dissertation Title:

Dependence Patterns across Financial Markets: Methods and Evidence

Committee:

Professor Peter C. B. Phillips Professor Robert J. Shiller Professor Ray C. Fair Professor Roger G. Ibbotson

Expected Completion Date:

May 2002

Degrees:

Ph.D., Department of Economics, Yale University, expected May 2002 M. Phil., Department of Economics, Yale University, May 2000 M.A., Department of Economics, Yale University, December 1998 B.A., School of Economics, Beijing University, July 1997

Fellowships, Honors and Awards:

Yale University Dissertation Fellowship, 2001-2002 Cowles Foundation Prize, 2001 Yale University Graduate, Fellowship, 1997-2001 Beijing University Outstanding Students Scholarship, 1995-1997

Teaching Experience:

Teaching Assistant, Fixed Income Trading Strategy, Yale School of Management, 2001 Teaching Assistant, Financial Engineering, Yale School of Management, 2001 Teaching Assistant, Fixed Income Security Analysis, Yale School of Management, 2000 Teaching Assistant, Econometrics and Data Analysis, Yale University, 2000 Teaching Assistant, Corporate Finance, Yale School of Management, 1999 Teaching Assistant, Portfolio Theory and Financial Markets, Yale University, 1999 College Tutor in Economics, Math and Science Tutoring Program, Yale University, 1998

Research Experience:

Research Assistant for Professor Arturo Bris, Yale School of Management, Summer 1999 Research Assistant for Professor Koichi Hamada, Economics Department, Summer 1998

Papers:

"Dependence Patterns across Financial Markets: Methods and Evidence," 2001 (job market paper). "Dynamics of Federal Funds Target Rate: A Nonstationary Discrete Choice Approach," with Peter C. B. Phillips, 2001 (job market paper and has been submitted to American Economic Review for publication). "Nonstationary Discrete Choice," with Peter C. B. Phillips, 2001 (submitted to Journal of Econometrics for publication). "Dynamic Behavior of the Volatility of Term Structure and Time-varying Risk Premium," 2000. "Robust Tests of Foreign Exchange Market Efficiency with Evidence from European 92-93 Currency Crisis," 1999.

Professional Activities:

Referee for Econometric Theory

References:

Professor Peter C. B. Phillips Department of Economics Yale University Box 208281 New Haven, CT 06520-8281 Tel: (203) 432-3695 Fax: (203) 432-5429 E-mail: peter.phillips@yale.edu Professor Ray C. Fair Department of Economics Yale University Box 208281 New Haven, CT 06520-8281 Tel: (203) 432-3715 Fax: (203) 432-6167 E-mail: ray.fair@yale.edu Professor Robert J. Shiller Department of Economics Yale University Box 208281 New Haven, CT 06520-8281 Tel: (203) 432-3708 Fax: (203) 432-6167 E-mail: robert.shiller@yale.edu Professor Roger G. Ibbotson Yale School of Management Box 208200 New Haven, CT 06520-8200 Tel: (203) 432-6021 E-mail: roger.ibbotson@yale.edu

Dissertation Abstract:

Using the concept of a copula, this dissertation shows how to estimate association across financial markets, with a focus on the structure of dependence, rather than the degree of dependence. Much economic analysis and multivariate econometric modeling of financial markets focuses on the degree of dependence. Perhaps, in consequence, statements relating dependence among markets are now fairly common in financial market discussion, one example being the observation that U.S.-European markets are more correlated than Japan-European markets. However, another important aspect of association is the structure of dependence across markets and this aspect is often omitted from analysis and discussion. As demonstrated in the dissertation, at the same correlation level, very different patterns of dependence structures are possible. Using a continuous function that we call a mixed copula model, this dissertation offers an alternative approach to capturing various dependence patterns between component random variables. The thesis develops an inferential apparatus for this approach and the methodology is applied to several major financial markets. The empirical findings are shown to have implications that seem important for a wide range of studies including risk management, portfolio choice, asset pricing, and tests for contagion in financial markets. In the first chapter, a mixed copula model is constructed so that it can capture three basic patterns of positive dependence between two markets. In summarizing these three patterns of dependence, we are looking for answers to questions such as following: are markets more likely to crash together, boom together, or do they tend to have symmetric comovements in both directions? Plotting the mixed copula function against the comovements between two markets, the curve can display three shapes corresponding to those three patterns, what we call the L shape, the J shape, and the U shape. A mixed copula model also facilitates the separation of the concepts of degree of dependence and pattern of dependence, and these concepts are embodied in two different groups of parameters -- association parameters and weight parameters. This is useful in the context of testing for stability of dependence over time, when it would be possible to observe a switch in the dependence pattern while the degree of dependence is largely unchanged. The opposite is also possible. To estimate a mixed copula across financial markets, we recommend using semi-parametric method, which means that the marginal distributions are estimated non-parametrically and the empirical distributions are then used to estimate the parameters in the copula. One advantage of this approach is that we do not need to specify the marginals, so there are no detrimental effects on copula estimation from a specific choice of marginals. Estimation and inference procedures for this approach are provided. The second chapter investigates the empirical dependence patterns in several major financial markets, using estimated mixed copula functions. In the four empirical applications considered in this dissertation, examples of all the three shapes have been found. The evidence suggests that either a U shape or a combination of U and L shapes satisfactorily summarizes the dependence in foreign exchange markets; that the L shape is very significant in international stock markets; that the dependence between U.S. domestic stock market and bond market also displays L shape; and, finally, that cross-country dependence in economic growth patterns displays a combination of U and J shapes. In particular, the findings from international stock markets in this paper corroborate the proposition in Longin and Solnik (2001): it is a bear market, rather than volatility, that is the driving force in increasing dependence across international equity markets. Based on the empirical findings, we propose some implications that seem important in financial studies. In the first place, use of multivariate normality and correlation coefficients to measure dependence may significantly under-estimate certain risk. A numerical example is given in computing value at risk (VaR) for a portfolio composed of equities in different countries. The VaR computed using the estimated mixed copula is very close to the actual VaR, which is much lower than that computed under the Gaussian assumption. Second, in hedging risk and in asset diversification dependence patterns as well as dependence degree should enter into market valuation. Finally, other things being equal, assets should be priced differently if they have different dependence patterns with the aggregate market. For example, assets having stronger dependence with the aggregate market in crashes (L-shape dependence) should have higher expected return due to the ‘shape risk’ compared to assets having stronger dependence with the aggregate market in booms (J-shape dependence), all other things being equal.

Other Papers:

"Dynamics of the Federal Funds Target Rate: A Nonstationary Discrete Choice Approach," is a paper jointly written with Peter. C.B. Phillips. We apply a discrete choice approach to model the empirical behavior of the Federal Reserve in changing the Federal funds target rate, which is the benchmark for short-term market interest rates. In particular, our methods allow the explanatory variables to be nonstationary as well as stationary. The approach is useful in this problem since many economic fundamentals that contribute to the determination of optimal interest rate display nonstationarity over time, such as measures of inflation and employment. We use the PIC criterion (Phillips and Ploberger (1996)) for model selection and the model successfully predicts the majority of the target rate changes during the time period we consider (1985-2001). In particular, it captures the six consecutive rate cuts in the first part of 2001. Based on the model implied optimal interest rate, we find that there is a lag in the Fed's reaction to economic shocks and that the Fed is more conservative in raising the interest rate than lowering it. "Nonstationary Discrete Choice," jointly written with Peter. C.B. Phillips, develops an asymptotic theory for time series discrete choice models with explanatory variables generated as integrated processes and with multiple choices and threshold parameters determining the choices. The theory extends recent work by Park and Phillips (2000) on binary choice models, allowing for multiple choice and threshold parameter estimation. As in this earlier work, the maximum likelihood (ML) estimator is consistent and has a limit theory with multiple rates of convergence (n3/4 and n1/4) and mixture normal distributions where the mixing variates depend on Brownian local time as well as Brownian motion. An extended arc sine limit law is given for the sample proportions of the various choices. The new limit law exhibits a wider range of potential behavior that depends on the values taken by the threshold parameters and is flexible enough to accommodate many different potential applications. References: Longin, F. and B. Sonik (2001). "Extreme Correlation of International Equity Markets," Journal of Finance, 56, 649-676. Park, J.Y. and P.C.B. Phillips (2000). "Nonstationary Binary Choice," Econometrica, 68, 1249-1280.   Phillips, P.C.B. (1996). "Econometric Model Determination," Econometrica, 64, 763-812. Phillips, P.C.B. and W. Ploberger (1996). "An Asymptotic Theory of Bayesian Inference for Time Series," Econometrica, 64, 381-413.