MSc.MF Theses and Dissertations (2019)
Permanent URI for this collection
Browse
Recent Submissions
Now showing 1 - 4 of 4
- ItemAn Inclusive pension model for Kenya’s informal sector with late entries and early exit rates(Strathmore University, 2019) Lagat, Cherono AsumptaThe purpose of this study is three-fold: first, we develop 1'1 pension model that uses preretirement mobile phone airtime expenditures to accumulate the pension fund. Secondly, we · calculate the exit and entry rates into the comprehensive pension scheme. Finally, we determine the expenditure patterns experienced post-retirement and use these patterns to advise on the daily amount required to be charged per minute above the current rate in order to facilitate a comfortable post-retirement life. The data utilized in this study was retrieved from various secondary sources. Inflation and interest rates data -were retrieved from Kenya's Central Bank database. The entry and exit data into informal pension schemes was retrieved from Eagle Africa the administrators of Mbao Pension scheme the largest informal pension scheme in Kenya. The mortality rates were retrieved from the World Health Organization and the life expectancy from ·world Atlas, Lancet and World Life Expectancy. Pre-retirement data was retrieved from November 2013 from an integrated survey on land ownership and tenure, provision, access and control of basic services, asset ownership, financial resources, evictions and demolition of houses, as well as thirty-two key informant interviews with informal small-scale service provider’s facilitated by Strathmore University. The inflation and interest rates were forecasted using ARIMA (1,9,5)-GARCH (0,1) model while the backward entry and exit data points were simulated in R. Our results show that an unemployed Kenyan spends approximately KShs. 2, 000.37 a month considering inflation this amount will translate KShs. 4, 025.45 to maintain the same life style post-retirement assuming the person joins the scheme at 18 and exits at the age of 55. Given the expenditure pre-retirement of this group of people, it will require them to be charged KShs. 3.41 per minute above the current rate in order to raise an amount sufficient to sustain their lifestyle post-retirement.
- ItemComparison of survival analysis approaches to modelling credit risks(Strathmore University, 2019) Mungasi, Sammy MonyonchoCredit risk is a critical area in finance and has drawn considerable research attention. As such, survival analysis has widely been used in credit risk, in particular, to model debt's time to default mechanisms. In this study, we revisit different survival analysis approaches as applied in credit risk defaulters' data and assess their performance in light of the Kenyan context. In practice, inconsistency in the validity of credit risk models used by many companies when predicting and analysis of loan default is a common phenomenon that occurs unexpectedly. Loan defaults often cause major loses to creditors' and can be of great benefit if quantified correctly in advance by using correct models. Here, we address the unbiasedness, analysis, and comparison of survival analysis approaches, particularly, the models of credit risk. We carry out data analysis using the Cox proportional hazard model and its extensions as well as the mixture cure and non-cure model. We then compare the results systematically by investigating the most efficient awl preferable model that produces best estimates in the Kenyan real data, sets. Results show the Cox Proportional Hazard (Cox PH) model is more efficient in the analysis of Kenyan real data set compared to the frailty, the mixture cure, and non-cure model.
- ItemApplication of long-short term memory Deep Neural Network in financial forecasting(Strathmore University, 2019) Wanyonyi, Watua PeterThe goal of this research was to apply Long-Short Term Memory Deep Neural Networks in financial forecasting. In order to predict the financial data, we used long-short term model and we compared its performance to ARIMA-GARCH hybrid model. In the study we used the ARIMA-GARCH time series model and studied its limitations in time series forecasting. We then introduced Deep Neural Network model (DNN) so as to improve accuracy which was tested on different financial datasets. Lastly we compared the results of the models employed using the root mean square error (RMSE) and p-value; LSTM had RMSE of 0.09989178 while the ARIMA-GARCH had RMSE of 0.0178. It was then concluded that the long-short term memory (LSTM) model, which is one of the DNN models, had significantly better than the ARIMA-GARCH Hybrid model in prediction/forecasting financial returns on FTSEIOO and S&PSOO indices.
- ItemConsumer credit risk modelling using machine learning algorithms: a comparative approach(Strathmore University, 2019) Nyangena, Brian OkemwaConsumer credit risk scoring involves the assessment of the risk that is associated with a customer that applies for credit. The ability to confidently identify customers that will repay the credit and those that will not is therefore, an important aspect for any institution. The purpose of this study is to model consumer credit risk using machine learning models and compare the results to the traditional logistic model. The aim is to identify whether there is improved performance in the classification of default among customers when machine learning algorithms are used. Additionally, the study aims to identify how different customer characteristics affects their default experience. The data used was obtained from Kenya Metropol between 2014-2017 and had customer details such as age, loan amount, marital status and sex among others, during this period. 5 models are used to model the default experience namely: Logistic regression, Random Forest, Support Vector Machine, Gradient Boosting and Multi-layer Perceptron Neural Network. The efficiency of the models was assessed using the following metrics; Accuracy, Precision, Recall, F1-score and Precision-Recall curve. Due to the imbalanced nature of credit data set, the F1-score, which is a weighted average of the Precision and Recall, was eventually used as the metric to determine the best model for credit scoring. The findings showed that Random Forest performed the best, having an F1-score of 0.307. The machine learning algorithms outperformed the logistic model and showed an improved performance in the classification of default, especially in identifying false positives. It was also established that male customers had a higher default probability, younger customers were more likely to default and single customers defaulted more than married customers