Anthony D. Hall

A cointegration analysis of treasury bill yields

This paper shows that yields to maturity of U.S. Treasury bills are cointegrated, and that during periods when the Federal Reserve specifically targeted short-term interest rates, the spreads between yields of different maturity define the cointegrating vectors. This cointegrating relationship implies that a single non-stationary common factor underlies the time series behavior of each yield to maturity and that risk premia are stationary. An error correction model which uses spreads as the error correction terms is unstable over the Federal Reserves policy regime changes, but a model using post 1982 data is stable and is shown to be useful for forecasting changes in yields.

A Cointegration Analysis of Treasury Bill Yields

This paper shows that yields to maturity of U.S. Treasury bills are cointegrated and that, during periods when the Federal Reserve specifically targeted short-term interest rates, the spreads between yields of different maturity define the cointegrating vectors. This cointegrating relationship implies that a single nonstationary common factor underlies the time-series behavior of each yield to maturity and that risk premia are stationary. An error-correction model that uses spreads as the error-correction terms is unstable over the Federal Reserves policy regime changes, but a model using post 1982 data is stable and is shown to be useful for forecasting changes in yields. Copyright 1992 by MIT Press.

A nonlinear time series model of El Niño

A smooth transition autoregressive model is estimated for the Southern Oscillation Index, an index commonly used as a measure of El Nino events. Using standard measures there is no indication of nonstationarity in the index. A logistic smooth transition autoregressive model describes the most turbulent periods in the data (these correspond to El Nino events) better than a linear autoregressive model. The estimated nonlinear model passes a battery of diagnostic tests. A generalised impulse response function indicates local instability, but as deterministic extrapolation from the estimated model converges, the nonlinear model may still be useful for forecasting the El Nino Southern Oscillation a few months ahead.

Greisenisation in the granite of Cligga Head, Cornwall

Abstract The central part of the Cligga Head granite is cut by numerous greisenbordered veins, between which the granite shows incipient greisenisation. The greisenised rocks were found to be enriched in Fe, Ca, F, B, Li, Mn, Rb, Sn, W and Zn, and depleted of Na, Al, Ba, Cu and Sr. The feldspars, micas and tourmalines also show changes in composition related to greisenisation. The chemical changes brought about by postmagmatic alteration at Cligga Head are broadly similar to those observed in other examples of greisenisation, but each occurrence has its own particular characteristics.

Using Bayesian variable selection methods to choose style factors in global stock return models

This paper applies Bayesian variable selection methods from the statistics literature to give guidance in the decision to include/omit factors in a global (linear factor) stock return model. Once one has accounted for country and sector, it is possible to see which style or styles best explains current asset returns. This study does not find compelling evidence for global styles, once country and sector have been accounted for.

A Survival Analysis of Australian Equity Mutual Funds

Determining which types of mutual (or managed) investment funds are good financial investments is complicated by potential survivorship biases. This project adds to a small recent international literature on the patterns and determinants of mutual fund survivorship. We use statistical techniques for survival data that are rarely applied in finance. Of specific interest is the hazard rate of fund closure, which gives the variation over time in the conditional probability of fund closure given fund survival to date. For a sample of 251 retail investment funds in Australia from 1980 to 1999 we identify a hump-shaped hazard function that reaches its maximum after about five or six years, a pattern similar to the UK findings of Lunde, Timmermann and Blake (1999). We also consider the impact of monthly and annual fund performance (gross and relative to a market benchmark). Returns relative to the benchmark are much more important than gross returns, with higher relative returns associated with lower hazard of fund closure. There appears to be an asymmetric response to performance, with positive shocks having a larger impact on the hazard rate than negative shocks.

Bayesian variable selection in logistic regression: predicting company earnings direction

This paper presents a Bayesian technique for the estimation of a logistic regression model including variable selection. As in Ou & Penman (1989), the model is used to predict the direction of company earnings, one year ahead, from a large set of accounting variables from financial statements. To estimate the model, the paper presents a Markov chain Monte Carlo sampling scheme that includes the variable selection technique of Smith & Kohn (1996) and the non-Gaussian estimation method of Mira & Tierney (2001). The technique is applied to data for companies in the United States and Australia. The results obtained compare favourably to the technique used by Ou & Penman (1989) for both regions.

4 Modeling the term structure

Publisher Summary This chapter describes the methods of modeling the term structure that are to be found in the econometrics and finance literatures. By utilizing a factor representation, the chapter shows that there are many similarities in the two approaches. However, there are also some differences. Within the econometrics literature, it is common to assume that yields are integrated processes and that spreads constitute the co-integrating relations. Although the finance literature takes the stance that yields are near integrated, but stationary, it emerges that the models used in that literature would not predict that the spreads are co-integrating errors if one actually replaced the stationarity assumption by one of a unit root. The reason for this outcome is found to lie in the assumption that the conditional volatility of yields is a function of the level of the yields. Empirical work tends to support such a hypothesis and suggests that the consequences of such a relationship can be profound for testing propositions about the term structure. The chapter also presents a number of stylized facts about a set of data on yields that prove useful in assessing the likely adequacy of many of the models that are used in finance for capturing the term structure.

Regulatory Tools and Price Changes in Futures Markets

There is an extant literature investigating the relation between futures price limits and the volatility of futures price changes. An equally impressive number of papers investigates margin levels and their relation with price volatility. Very few papers explicitly model the indirect relation, through volatility, between margins and limits. Brennans (1986) model is an exception. In his model, price limits help control contract default risk, thereby reducing required margins and ultimately lead to lower transaction costs. The crucial assumption in Brennans model is the absence of accurate price signals when prices are locked at the limit. The paper extends Brennans model with more realistic price change distributions that capture the typical characteristics of futures prices such as fat tails and time-varying volatility. It also discusses how learning can occur and how this may affect cost minimising optimality of regulation. Copyright 2001 by Blackwell Publishers Ltd/University of Adelaide and Flinders University of South Australia

Evaluating the Impact of Inequality Constraints and Parameter Uncertainty on Optimal Portfolio Choice

We present new analytical results for the impact of portfolio weight constraints on an investor’s optimal portfolio when parameter uncertainty is taken into account. While it is well known that parameter uncertainty and imposing weight constraints results in reduced certainty equivalent returns, in the general case, there are no analytical results. In a special case, commonly used in the funds management literature, we derive analytical expression for the certainty equivalent loss that does not depend on the risk aversion parameter. We illustrate our theoretical results using hedge fund data, from the perspective of a fund-of-fund manager. Our contribution is to formalize the framework to investigate this problem, as well as providing tractable analytical solutions that can be implemented using either simulated or asset manager returns.