Browsing by Author "MacDonald, lain"

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Restricted
    An Investigation into the suitability of using GARCH process for pricing options on the SAFEX all share index futures contracts
    (1999) Miller, Saul; Dorrington, Robert; MacDonald, lain
    This thesis primarily sets out to investigate the possibility of incorporating autoregressive conditional heteroskedasticity (ARCH) assumptions in an option-valuation model for All Share Index option contracts, as an alternative to the constant variance assumption required by the Black-Scholes option-pricing model. This involves an assessment of whether the log-returns of the ALSI futures (the instruments underling the ALSl option) follow an ARCH process. A secondary objective is to assess the potential for using an ARCH process to model the ALST spot log returns. This could have the following uses: • Pricing over-the-counter ALSI spot options. • Using the forecast spot return ARCH volatility as a proxy for the forecast ALSI future log return volatility if they have similarly. This is desirable for pricing options on new futures contracts when there is insufficient historical futures data available to fit an ARCH model. Evidence of ARCH presence is determined by examining autocorrelation in the square error terms of the log returns. Although some statistically significant autocorrelations were found, the lags which exhibited these significant autocorrelations showed no pattern. Furthermore, lags which exhibited these significant autocorrelations changed over time.
  • Thumbnail Image
    Item
    Open Access
    A comparison of methods for modelling rates of withdrawal from insurance contracts
    (2009) Smith, Bradley; MacDonald, lain
    Withdrawal from insurance contracts can be a significant risk for insurers. Withdrawal rates can be difficult to predict because withdrawal is influenced by a number of inter-related factors related to, inter alia, the sales process, characteristics of the insurance contract, characteristics of the contract holder, and economic variables. Existing methods used to model and predict withdrawal rates are initially reviewed. Two additional methods which have been proposed in the literature as means for modelling insurance risks are neural networks and Bayesian networks. These two methods are utilised in order to build models to compare their predictive ability with a commonly used method for modelling withdrawal rates, namely logistic regression.