Arthur Warga

Bond Returns, Liquidity, and Missing Data

This paper examines holding period returns to constant duration portfolios of U.S. Government notes and bonds, and measures the return premium generated by liquidity differences in bonds. The approach compares constant duration portfolios constructed in two distinct ways: one using only bonds issued in the most recent Treasury auction, one using all other bonds. Comparisons of the resulting series are made meaningful by choosing narrow duration ranges in the portfolio formation procedure. This, in turn, induces a missing data problem in the created time series. Parameters of return distributions are therefore estimated employing a maximum likelihood framework that explicitly accounts for the missing data. It is estimated that recently issued bonds are priced to return a premium of around 55 basis points per annum over otherwise equivalent instruments.

BOND PRICE DATA AND BOND MARKET LIQUIDITY

This paper attempts to characterize liquidity-driven noise in the CRSP Government Bond price data set by comparing these price records to the independently collected Shearson Lehman Brothers (SLB) Bond Data Base. We argue that discrepancies between the data sets are due largely to liquidity-driven price errors, and we show that they are systematically related to certain bond characteristics. On the other hand, these discrepancies are small in magnitude and are approximately mean zero. We examine data filters based on observable bond characteristics and show that these filters can reduce the noise in price records while preserving their mean zero nature. The effects of these errors on performance evaluation are investigated by comparing results using filtered and unfiltered data.

Risk and the Performance of Real Estate Investment Trusts: A Multiple Index Approach

This paper analyzes the returns of a sample of Real Estate Investment Trusts and examines their risk-adjusted performance using both single index (i.e., CAPM) and multiple index (i.e., APT) models. It is shown that while the performance rankings of the investment trusts are not very sensitive to the risk-adjustment model, the actual performance measures do sometimes differ substantially. Unfortunately, because of the high volatility of these real estate investments, the differences in investment performance across trusts generally are not statistically significant. Copyright American Real Estate and Urban Economics Association.

Stock Returns as Predictors of Interest Rates and Inflation

This study examines whether stock returns provide forecasts of changes in interest rates and inflation. In contrast to earlier work that indicated that changes in expected inflation negatively affect stock returns, we find a statistically significant positive relation between stock returns and future inflation rate changes as well as a significant positive relation between stock returns and future interest rate changes. Real estate investment trusts, which are particularly interest- and inflation-sensitive securities, provide better forecasts than a broad market index. Finally, we find that most of the evidence supporting the forecasting ability of stock returns occurs in the October 1979 to October 1982 period when the Federal Reserve Board chose not to counteract interest rate changes.

The Market Value of Debt, Market Versus Book Value of Debt, and Returns to Assets

Empirical studies usually measure the value of debt based on book rather than market value, even though the underlying theory is almost always based on market values. This paper documents how using book value to measure debt can distoret debt-equity rations and cost of capital calculations.

EXPERIMENTAL DESIGN IN TESTS OF LINEAR FACTOR MODELS

For various reasons, cross-sectional tests of linear factor models have employed the returns on portfolios of assets (as opposed to individual securities). There is a good deal of flexibility in how one goes about choosing which portfolios to examine (the experimental design). This article points out that some choices can lead to collinearity among factor loadings with potentially harmful consequences for power. A diagnostic procedure for detecting collinearity is discussed, and several illustrations of the procedure on existing tests of linear factor models are used to illustrate their value to the researcher. Finally, a series of experiments is carried out to validate the diagnostics and to confirm their assessment of the experimental design in the models examined.