- ItemCorrelating factors of U.S. presidential speeches with stock market movements - a machine learning approach(Stellenbosch : Stellenbosch University, 2023-03) Rees, Pablo; van Lill, Dawie; Stellenbosch University. Faculty of Economic and Management Sciences. Dept. of Economics.ENGLISH SUMMARY: The literature relating textual to stock market data is deep, but the relationship between speeches given by political figures and stock markets is relatively undefined. This research begins to rectify this by exploring the relationship between U.S. presidential speeches and daily price movements in the S&P 500 index. It was possible to explore this relationship by using natural language processing techniques, econometric time-series analysis, and machine learning models. It was found that models including presidential speech data can achieve prediction accuracy of about 60% over an S&P 500 index price movement proxy. This is an increase of about 0.3% (0.599 vs 0.601) over the models that did not include the presidential speech data (without losing ground in either recall or precision). Notably, this result was drawn from 71 years of data at a daily resolution. Thus, it is concluded that presidential speeches hold predictive power over stock market movements and that this relationship can be used to improve the power of predictive models.
- ItemA comparison between existing mortality risk algorithms and machine learning techniques(Stellenbosch : Stellenbosch University, 2022-12) Scholtz, Jenny; Burger, Rulof; Retief, Riani; Stellenbosch University. Faculty of Economic and Management Sciences. Dept. of Economics.ENGLISH SUMMARY: This thesis assesses the feasibility and benefits of using the patient data of a large private South African hospital group to estimate a model of mortality risk using flexible machine learning techniques. Specifically, I investigate whether such a model would have been able to outperform a commonly used medical scoring system, SAPS 3, in predicting mortality during the second half of the Covid-19 pandemic. A LightGBM machine learning model is shown to be much more accurate in predicting mortality (76.15% accuracy, compared to 56.58% for SAPS 3) for the Covid-19 positive sample. Roughly half of this gain in predictive accuracy is obtained from using the most recent and relevant data to train the model, while the remaining lift is attributable to allowing the model to find patient symptoms and attributes that are measured but ignored by SAPS 3. Interestingly, the flexible functional form of the machine learning models, which allow the predictors to affect mortality through non-linearities and interactions, has a negligible effect on predictive accuracy. The same method is also found to produce more accurate forecasts for patients who tested negative for Covid-19, but this improvement is smaller than for Covid-19 positive sample. The results of this thesis illustrate that machine learning methods are valuable tools to predict patient outcomes, particularly when there are unexpected shifts in the relationship between patient features and patient outcomes. Large hospital groups can obtain more accurate forecasts from a dynamic scoring system which is frequently frequently retrained on their own patient data.
- ItemFiscal instruments and the elderly population in South Africa : a distributional analysis(Stellenbosch : Stellenbosch University, 2022-04) Huffer, Calvin; Jansen, Ada; Stellenbosch University. Faculty of Economic and Management Sciences. Dept. of Economics.ENGLISH SUMMARY: Saving an adequate amount of money to finance retirement spending is a growing issue that affects the elderly. In response to the incomplete retirement savings markets that the elderly face, national governments employ fiscal instruments in an attempt to provide social security to the economically vulnerable elderly population. In South Africa, the government utilizes four primary fiscal instruments to address the lack of retirement savings among the elderly: the old age grant, a higher tax rebate for the elderly, an additional medical expenses tax credit for out-of-pocket expenses and medical aid fees of those 65 years and older, and preferential tax treatment of pension contributions. Combined, these four fiscal instruments form the social security net for the elderly in South Africa, although concerns of some individuals still slipping through the net remain. Therefore, an analysis into the effectiveness of these fiscal instruments in addressing incomplete markets and issues related to intended beneficiaries is undertaken to evaluate social security provision for all segments of the South African elderly population. The fiscal instruments are assessed both individually and in combination by means of economic theory and microsimulation using the South African tax-benefit Microsimulation Model (SAMOD). Microsimulation using the SAMOD allowed for a comparison of current fiscal instruments against alternative scenarios to assess the relative costs and benefits of policy reforms in terms of government revenue, poverty, and inequality. The underpinning household survey data of the SAMOD allows for the static economic implications of potential reforms to be evaluated and inform the effectiveness of the current social security system. The South African context and existing empirical evidence are used to determine whether policy reforms would be socially beneficial to the elderly. This investigation finds that while the fiscal instruments employed by the South African government are reasonably established for those in low- and higher-income groups among the elderly population, those in the middle-income groups may fall through the gaps of the social security system. Reform of the existing fiscal instruments to include these individuals would create more comprehensive social security for the elderly in South Africa and it can be accomplished with a relatively small impact on government expenditure. More extensive reform could be accomplished with greater fiscal expenditure on the elderly, but even within the confines of the budgetary constraints, modest reform of the fiscal instruments affecting the elderly can have positive societal benefits.
- ItemThe law and economics of potential competition in digital markets(Stellenbosch : Stellenbosch University, 2022-04) Friday, Megan; Boshoff, Willem Hendrik; Stellenbosch University. Faculty of Economic and Management Sciences. Dept. of Economics.ENGLISH SUMMARY: This research aims to explore how authorities in different jurisdictions respond to potential competition in the online automotive classifieds platform markets. This is done by means of a review of cases in South Africa, the United Kingdom and Australia. This thesis considers how actual competition should be measured within markets as well as how potential competitors are identified and ranked against one another according to the firm that imposes the largest competitive constraint on its rivals. Previous research and literature have identified the method to be used when defining a relevant potential competitor in a market. This thesis evaluates how potential competition considerations should affect the evaluation of a merger or acquisition, specifically in the online automotive classifieds platform markets. Potential competition issues arising in intermediation platform markets exists as more firms are developing their own intermediation platforms and competition authorities are seeking to regulate these firms that bring about novel competition issues compared to those experienced by a traditional firm. This research seeks to provide an understanding to its readers and authorities on how potential competition concerns should be addressed and handled in platform markets, particularly those including a digital element. This research evaluates the literature surrounding this topic by taking a case studies approach.
- ItemSimulated horizontal mergers in vertically related markets(Stellenbosch : Stellenbosch University, 2022-04) Marais, Wihan; Boshoff, Willem; Stellenbosch University. Faculty of Economic and Management Sciences. Dept. of Economics.ENGLISH SUMMARY: A global trend of increasing retail concentration and a heightened concern about buyer power demand more sophisticated and flexible merger screening tools. Parametrisable merger simulation models that suitably account for the effects of bargaining competition may pose a solution. Thus, this dissertation examines the predicted effects o f retail consolidation in a vertically related market using the merger simulation tool developed by Tschantz and Froeb (2019). With the flat logit nested demand function for differentiated products of Boshoff e t a l. (2020), the predicted retail merger effects of five different bargaining and non-bargaining models for a 1 × 2 industry are evaluated. The primary contribution of this study is its unique application of Nashin-Shapley (NiS) bargains in the study of retail consolidation. It is novel in so far it compares the inherent implications of Nash-in-Nash (NiN) and NiS bargains for the appraisal of retail mergers. Pre-merger competitive outcomes with NiN predetermine a finding o f an anticompetitive merger. Contrarily, predictions with NiS as its starting point do not indicate that a merger would cause consumer harm, and therefore, would not be prohibited. In addition, the results suggest that retail consolidation consistently poses a viable option through which retailers can improve their profitability and bargaining positions. However, merger incentives are moderated by the competitiveness of an outside market.