THE PURPOSE OF THIS CHAPTER

In this chapter we look in detail at the Kim–Wright (2005) model, and at its natural successor, the D'Amico, Kim and Wei (2010) model. Both are very good examples of latent-variable models, and they attempt to give a coherent account of both nominal and real rates (and, to some extent, even of liquidity).

One of the tasks for which this class of models can be used is the prediction of future yields and future inflation. Hence the first reason for the oracular reference in the quote that opens the chapter. As predicting yields is only a quarter of a tiny step away from predicting excess returns, these models also produce predictions for risk premia and yield expectations. These we analyze in the last part of the chapter. The main results are that “a large portion of the decline in long-term yields and distant-horizon forward rates since the middle of 2004 [is due] to a fall in term premiums”, and that “about two-thirds of the decline in nominal term premium owes to a fall in real term premiums, but estimated compensation for inflation risk has diminished as well”.

This is interesting in itself. Apart from the specific findings, the Kim–Wright model provides a blueprint for the latent-variable approaches discussed further in Chapter 34. It is important to become familiar with this approach, because it is very common in the modern term-structure literature. See, in this respect, the references quoted in Section 31.10.

Finally, the Kim–Wright model and its cousins are frequently referred to by central banks, and are used by policymakers as one of the tools to assess the magnitude and sign of the risk premium, the extent to which inflation expectations are ‘well-anchored’, etc. So, whatever the merits of the model, in itself this is a powerful reason for going to the oracle others oracles go to. The last section of the chapter mentions a few of the currently most popular (first-order) oracles.