The paper by Jegadeesh and Pennacchi extends a sizable literature on continuoustime models of interest rate behavior, some of it theoretical, some empirical. In some cases, the focus is on statistical modeling, frequently with the goal of a rich parameterization that captures known stylized phenomena (for example, current yield curve, nonconstant volatility structure), but that is nevertheless solvable in closed form. Examples of this type of approach are Ho and Lee (1986) and Hull and White (1990). In other cases, the focus is structural, making a connection between interest rates and other fundamental economic variables, such as inflation or production. Pennacchi (1991), for example, has adopted this approach in an earlier paper. I will argue in this comment that the present paper is a useful addition to this literature. It extends the literature in a logical direction that was clearly worthwhile investigating, and demonstrates a high level of technical competence. However, I will also argue that the interpretation of the results is overstated in terms of both statistical fit and of implications about fundamental economic relationships. Consider first the possible objectives of modeling interest rates. One such objective is simply to capture the statistical propvrties of interest rate time series. If those properties are modeled accurately, the results can be applied to the pricing of financial instruments or to the management of price risk on such instruments. A second alternative objective is to explain the relationship between interest rates and other economic variables. This strategy could be useful in the construction of economic models and perhaps in performing policy analysis. The ultimate objective of the modeling exercise determines the appropriateness of any specific interest rate model. In the Jegadeesh-Pennacchi paper, both statistical modeling and structural interpretation are explicit or implicit goals. Modeling accuracy is a stated objective of the paper, and there are suggestions that the model may be used for pricing and hedging purposes. Pricing and hedging, however, are not directly attempted in the paper. Structural interpretation is not explicitly a goal of the paper, but it is a de facto goal as evidenced by the extensive discussions of Fed credibility and of the interpretation of the estimated market prices of risk.