Estimating Return Parameters with Short Historical Data: The Case of U.S. Treasury Inflation-Protected Securities

Abstract

Academic research on consumption and portfolio choices over an individual’s life cycle has been flourishing for almost four decades. Since the seminal work of Merton (1969) and Samuelson (1969), economists’ models have supplied many asset allocations that would achieve optimal intertemporal consumption. Additional considerations on one’s labor decisions or unforeseen rare events have substantiated these models’ predictions and encouraged their implementation by industry practitioners. However, applications of even fairly complex models have quickly exposed the limitations and shortcomings of simplistic assumptions. For example, most models restrict the investment opportunity set to stocks and a riskless asset. Bonds, derivatives and housing are seldom included while other categories are dismissed altogether. Although some model specifications could easily allow for more financial alternatives, it is sometimes data availability (or lack thereof) that determines what variables are included. In this respect, Treasury Inflation-Protected Securities (TIPS) are challenging. The U.S. Government’s first issue was in 1997, long after TIPS were introduced in the U.K., Canada, Israel, Turkey and New Zealand. Naturally, data that span over only 15 years are of little use for the study of an individual’s optimal lifetime behavior: oftentimes, the planning horizon is set at age 90 or 100, with retirement starting at 65. Ultimately, the too recent history of TIPS returns has deterred fund managers from offering them as long-term/retirement assets and, in turn, has curbed their market reach. This project develops a rigorous methodology to estimate the real term structure of TIPS between 1970 and 1997. The resulting series is then merged with the actual data to produce a continuous series of real interest rates from 1970 until now. A summary description of our techniques and preliminary results is attached. The goodness of the newly obtained rates is tested by computing historical correlations with: Moody’s Seasoned AAA Corporate Bond Yield, the 30-Year Conventional Mortgage Rate, the OFHEO’s House Pricing Index, the Dow Jones Industrial Average Index and the S&P500 Index. We perform the exercise for all maturities (5, 10, 20 and 30 years). The outcome of this analysis is twofold. First, it provides practitioners with a reliable tool to better measure the risk-return tradeoff of portfolio strategies with TIPS. Second, it helps establishing guidelines that policy makers can use to choose qualified default investment alternatives (QDIA). Both dimensions offer fertile ground to financial experts who have long cheered TIPS as valuable retirement vehicles.