Abstract

This thesis adds to the body of literature seeking to improve the estimation of the optimal hedge ratio used in hedging money market and fixed income securities in Canada. A more accurate or improved depiction of the hedge ratio is of considerable importance and is the primary goal of this thesis. The specific futures contract analyzed in this thesis is the Canadian Bankers' Acceptance Futures contract, the BAX, which was introduced on the Montreal Exchange in the early eighties as an instrument to manage interest-rate risk. Institutional features of the BAX market and the growth of this market, particularly in the second part of the 1990s, are described in Chapter 2. The efficiency of the BAX market is also addressed in this chapter where cointegration analysis is used to investigate the unbiasedness hypothesis. Chapter 3 presents a univariate analysis of the BAX, and its underlying instrument the Canadian Bankers' Acceptance or BA, in a general framework that permits the statistical evaluation of myriad dynamic volatility models which have been used in such contexts. The successful estimation of these general models requires a considerable amount of data and necessitates the daily sampling frequency which is used in this thesis. Chapter 4 presents a brief background and motivation for the hedge ratio, surveys early attempts at characterizing the hedge ratio and presents the bivariate models needed to estimate the hedge ratio. Most importantly, the general univariate framework used extensively in the previous chapter is extended to the bivariate case. The time-varying BA/BAX Hedge Ratio is then estimated. The hedging performance of these models is then evaluated. We also discuss whether there is any practical value in using daily data, versus weekly data, in the determination of the hedge ratio.