BBSA Research Projects (2015)

Permanent URI for this collection


Recent Submissions

Now showing 1 - 5 of 27
  • Item
    An analysis of longevity risk in a portfolio of life annuitants
    (Strathmore University, 2015) Njeri, Sharon
    Longevity risk has economic significance for governments, individuals and corporations. There is need to analyze the expected future lifetime of a population anticipating to receive lifetime benefits in Kenya. This paper performs such an analysis on the annuitants of a Kenyan life insurance company by making use of the Lee carter model and further describes ways to manage this risk. The results of this paper are directly relevant to annuity providers. It is found that life expectancy is increasing in the future up to some point where it gradually decreases. Conclusions made are that several models should be used to investigate this risk so as to reduce model errors and that impact of the data used is financially material.. Suggestions are that companies should create Value at Risk estimates of capital to cover this risk over one year periods and that more research should be done on managing longevity risks by use of capital markets in the country.
  • Item
    Incorporating dependence into the pricing of joint life annuities.
    (Strathmore University, 2015) Ohuru, Vincent Nyarango
    This study incorporated dependence of lives into the pricing of joint life annuities in Kenya. Currently the Kenyan industry practitioners base this pricing on the traditional assumption of independence oflives which has long been discredited as it ignores the fact that the underlying lives face similar risks . To value benefits under these policies accurately, a statistical model that assesses the impact of survivorship of one life on another was applied in this study. The Markovian model was applied to incorporate dependence into the pricing ofjoint life annuities. In the Markovian model the researcher considered three types of dependence which are : a) the instantaneous dependence that is due to a catastrophic event affecting both lives; b) the short-term dependence that can cause death to a surviving partner after the death of the spouse; c) the long-term dependence that results from the association between lifetimes, to generate the joint probabilities of survival that were then incorporated into the pricing of joint life annuities.
  • Item
    Determining incurred but not reported (IBNR) reserves using a collective risk model framework for a general insurance business line in a Kenyan insurance company
    (Strathmore University, 2015) Mboko, Willis Meshack
    General Insurance companies set up reserves in order to meet future claim liabilities. Reasonable forecasting of these liabilities is therefore an integral part of an insurer's business. Actuarial methods such as the chain ladder method have long been used to estimate insurer liabilities. Stochastic improvements of the chain ladder method mhave also been developed in order to obtain a standard error around a point estimate and a full distribution in some instances. However, such methods depend on heavily aggregated data in run-off triangles. Such an aggregation leads to loss of potentially predictive information. This paper uses specific claim information such as reporting delay and size of individual claims to forecast claim liabilities. It uses the collective risk model to combine expected claim size and frequency of claims. The data-set used in modelling is a realistic liability business from a Kenyan insurer. A final comparison of existing methods (Mack and Over-Dispersed Poisson) and the collective risk model is done for validation purposes. From the case-study findings, the collective risk model was preferred over traditional stochastic methods since it had a lower predictive error and was more realistic in modelling the claim process.
  • Item
    Forecasting the time varying-beta of nse-20 share companies: Bi-variate garch (1, 1) model vs kalman filter method
    (Strathmore University, 2015) Maywa, Nickson K.
    This research paper forecasts the time -varying daily beta of ten stocks listed in the Nairobi Securities Exchange 20- Share Index by use of a Bivariate GARCH (1, 1) model and the Kalman filter method. A comparison of the forecasting ability of the GARCH model and the Kalman filter method is made. Forecast errors based on the retUI11 forecasts are used to evaluate the outof- sample forecasting ability of both the GARCH model and the Kalman method. Two measures of error are used: MAE and MSE. The results are inconclusive, based on MSE the Kalman method is superior while based on MAE, the Bivariate GARCH (1, 1) method appears to provide more accurate forecasts of the time-varying beta .
  • Item
    Testing for the correlation between geographical area of operation and accident risk in PSV insurance industry
    (Strathmore University, 2015) Odunga, Hesbon Busera
    The Kenyan public service vehicle insurance has been marred with a lot of uncertainty in terms of the expected outcome of claim frequency and claim severity. At least eight insurance companies have either collapsed or have been placed under statutory management in the last twenty years. In an attempt to address the problem in the industry, this research paper concentrated on the possible underwriting inadequacies of the insurance companies operating in the industry in terms of their failure to consider all the significant risk rating factors that can improve the premium pricing. The research therefore applied the unused observables test to test for the existence of asymmetric information in the industry. It checked for the significance of geographical area as an additional variable that is unused in the market. The research made use of accident records for fifteen regions within Nairobi to test for the correlation between accident risk and geographical area. Secondary data that was used for this research project was obtained from the Traffic Police Headquarters for Nairobi Area. The null hypothesis was stated as, there being no correlation between geographical area and the number of accidents. A statistical analysis was then undertaken and an econometric model ran using ordinary least squares. The cover type of the policy was controlled for to produce unbiased results. The results of the analysis revealed the existence of significant correlation between the number of accidents and the geographical area leading to a rejection of the null hypothesis. The correlation was positive for some regions and negative for other regions. Therefore the research recommended the inclusion of geographical area as one of the risk rating factors in premium pricing if an actuarially fair premium is to be charged by the PSV insurance companies.