An Improved Linear Projection Approach to Estimate Daily Real Yields and Expected Inflations in a Latent Multifactor Interest Model

Abstract

The study improves upon the linear projection approach to estimate daily real yields and expected inflations in a latent multifactor interest rate model. It estimates the projection coefficients for inflation factor exclusively from monthly inflation data, rather than from both inflation and nominal yield data, in order to lessen biasedness. Because these coefficients are the same as those in the daily model, the study uses them with daily nominal yield data to estimate the remaining parameters. Using Thailand’s data from March 1, 2001 to August 30, 2013, the study finds that the improved model can fit the nominal yields well. The term structure estimate of real yields has a normal shape, while that of expected inflations is flat. The inflation premiums are significant statistically and economically. They are ten times the ones reported in the past. Inflation premiums cannot be ignored in economic analyses for Thailand.