OpenUCT :: Browsing by Author "Rajaratnam, Kanshukan"

Browsing by Author "Rajaratnam, Kanshukan"

Now showing 1 - 20 of 27

Open Access
A simulation-based approach to assessing the relationship between mutual fund size and performance
(2020) Molyneux, Matthew; Rajaratnam, Kanshukan
This study examines how mutual fund size affects performance. Academic literature on this topic is extensive but has yielded conflicting results. Some studies find a distinct relationship between fund size and risk-adjusted returns while others do not; some studies also posit that an optimal fund size exists where risk-adjusted returns are maximised. The size of equity mutual funds in South Africa and the market dynamics of the Johannesburg Stock Exchange provide an interesting context within which to analyse the relationship between size and performance. In this study, hypothetical portfolios are created, and an allocation procedure is used to distribute capital to these hypothetical portfolios. The allocation procedure distributes capital to the portfolio stocks by controlling for each stock's yearly volume traded. This method works to distribute capital up until a certain fund size; beyond that size, the hypothetical portfolio might no longer be fully invested in the random portfolio. To control for this, the simulation model engages in a routine to discard the stock with the lowest volume-traded level from the portfolio and reselect another stock from the investable universe with a higher volume-traded level. This process is repeated until the portfolio is fully invested. Stock selection and investment dates are randomised and variance reduction techniques are used to improve the efficiency of the simulation, and 10 000 simulation runs are performed. The results of the simulation found a non-monotonic relationship between mutual fund size and performance over a one-year holding period, consistent with some research internationally and in South Africa. Over a two- and three-year holding period, mutual fund size and returns, however, seem to be negatively correlated. Over the three holding periods, the study suggests that the optimal equity mutual fund size in South Africa is approximately ZAR 2bn. Portfolios with assets under management greater than ZAR 2bn see their returns decrease noticeably as fund size continues to increase. These findings are supported by comparing simulated returns to actual benchmark returns over the same random periods. The results of this study suggest that mutual funds should be aware that consistent increases in assets under management could negatively affect performance and that all funds should ensure that total assets under management do not exceed ZAR 2bn.
Open Access
Affordable rental housing delivery in Kenya
(2021) Olonde, Victor Otieno; Mooya, Manya Mainza; Rajaratnam, Kanshukan
Rental housing sector remains a significant housing option and an essential component of a vibrant housing market and construction industry as a whole. However, rental housing markets in most developing countries have been characterized by market failure because of the inability to provide adequate rental units commensurate with the urban population's demand. One of the major reasons for the inefficiency is attributed to the little attention by the private developers/landlords towards the rental housing development, yet they are perceived to be the major players on the supply side. The main objective of this research is to critically examine the rental housing market in Kenya and determine why despite the high demand for affordable rental housing, there has not been adequate corresponding supply of good quality housing units, a phenomenon which denotes market failure. This research aims at exploring the institutional environment to find out what has hindered delivery of adequate affordable rental housing despite high demand. This study has been guided by critical realism philosophical perspective and combines the conceptualisation in the Institutional Analysis and Development (IAD) framework and the theoretical richness in the New Institutional Economics (NIE) to develop the conceptual framework for analysing market failure in the rental housing market. The study utilizes mixed methods research design where both qualitative and quantitative research approaches have been employed, comprising a combination of cross-section survey of lower-middle income tenants and interviews of existing landlords, developers, key informants from the relevant public and private sector stakeholders. The main findings of this study underscore the significance of institutional environment in influencing the outcomes of the housing market, and note that as currently constituted, the institutional framework is not practically well-matched to support delivery of affordable rental housing units and as such discourages developers from the rental sector. The frameworks ranging from policy, regulatory and financial systems coupled with inferior performance of rental sector compared to development for sale systematically, but inadvertently lead to market failure in the rental housing market. This study recommends formulation of a distinct Rental Housing Policy and consequently Rental Housing Act complete with implementation framework to deal with issues intrinsic to the rental sector. It has made various and diverse contributions to the existing body of knowledge which comprise theoretical, contextual, empirical and policy perspective.
Open Access
An investigation into higher and partial moment portfolio selection frameworks
(2019) Polden, Stuart John; Rajaratnam, Kanshukan
This dissertation highlights the importance of considering higher moments and partial moments of the distribution when conducting portfolio optimisation and selection. This is due partly to the weaknesses of mean-variance optimisation, as discussed throughout the dissertation, and the appropriateness of considering higher moments to better meet the investors utility functions. This dissertation investigates the usage of two bi-objective optimisation frameworks, a Skewness/Semivariance framework previously suggested by Brito et al (2016), and a proposed upside and downside semivariance framework (referred to as Semivariance/Semivariance), developed from Cumova and Nawrocki’s (2014) general upper partial and lower partial moment framework. It solves the endogeneity issue present in the co-semivariance matrices, through the usage of a direct multi-search algorithm. The two frameworks were tested across multiple datasets, including one of pure stocks and one of asset classes, to test the ability to both allocate assets and select stocks. The performance was measured through nominal returns, statistical tests, Sharpe ratios, Sortino ratios, and Skewness/Semivariance ratios. The results reveal the Semivariance/Semivariance optimisation process to outperform the Skewness/Semivariance optimisation in the majority of the cases investigated. This suggests it may be a superior selection optimisation process. Furthermore, the Semivariance/Semivariance portfolios remain competitive with the benchmark portfolios selected in this dissertation, often outperforming them on an absolute return and ratio basis; however, this outperformance has not consistently proven to be statistically significant.
Open Access
An investigation into reference-day risk-free metrics in the context of modern portfolio theory on the JSE
(2019) Feinstein, Samuel G; Rajaratnam, Kanshukan
Modern portfolio theory (MPT), asset pricing models and broader financial modelling are dependent upon the accuracy of input parameters. For example, the accuracy of expected returns, standard deviations and correlations as an input into MPT will result in a more efficient selection of the optimal portfolio. These metrics are exposed to reference-day risk which is the variation in input estimation due to the selection of initial reference-day in calculations. This paper examines whether a change in reference-day, the day on which a metric is calculated, significantly affects estimates of risk-return metrics on the Johannesburg Stock Exchange (JSE). Thereafter, it applies these findings to the asset allocation problem of constructing a maximum Sharpe portfolio. The objective of this paper is to further prior research through the evaluation of an alternative simulation method and an extension of the range of tested metrics. The advancement of this prior research is achieved through the use of the Cholesky decomposition and a nonparametric bootstrapping procedure to generate reference-day risk-free estimates for average returns, standard deviations, correlations and betas. Furthermore, this paper applies the reference-day risk-free metrics to the construction of optimal multi-asset portfolios in the mean-variance framework. The findings suggest that through the use of a five-year period of monthly returns, the selection of a reference-day materially affects risk-return metrics and the subsequent portfolio characteristics that are based upon these metrics. The performance of portfolios, optimised on each reference day, ranged between 10% during the out-of-sample period. Additionally, using traditional end of month data resulted in underperformance of out-of-sample, overstated average returns, understated standard deviations and lower correlations between asset classes. Based on these findings we propose an alternative bootstrapping method for calculating reference-day risk-free metrics which reduces the effect of reference-day risk. The purpose of this methodology is to use these estimates for portfolio construction, risk management and asset pricing. The results of this paper indicate that reference-day risk makes a material difference in portfolio construction.
Open Access
An investigation into the use of multiple cryptocurrencies in a diversified portfolio
(2018) Kibble, Alexander; Rajaratnam, Kanshukan
This study investigates the possible diversification benefits of multiple cryptocurrencies (Bitcoin, Ethereum and Litecoin) in a diversified portfolio from the perspective of a South African investor over the period 30 July 2015 to 20 December 2017. Cryptocurrencies are mostly still in their infancy, and reliable information regarding their usefulness as an asset class in a diversified portfolio is scarce to come by. This study adopts a quantitative research methodology which incorporates the following statistical methods: i) mean-semivariance optimisation; ii) Kendall Tau-b correlations; and, iii) autocorrelation function for serial correlations. The JSE All Bond Index is used as bond investment proxy, a combination of the JSE Top 40, Resources Index and Financial-Industrials Index is used as an equity investment proxy, and the LBMA Gold PM is used as a gold investment proxy. The study found that all three cryptocurrencies under investigation yielded risk-return benefits for a diversified portfolio. The alternative cryptocurrencies (Ethereum and Litecoin) exhibited higher levels of downside risk (semideviation) than Bitcoin, but proportionately greater returns. Hence, the addition of these two cryptocurrencies to a portfolio that includes Bitcoin and traditional assets resulted in an expansion of the efficient frontier. Ethereum exhibited slightly lower correlations to Bitcoin than Litecoin, which is most likely attributed to its greater technological differences, but performed worse as a diversifier. All three cryptocurrencies yielded similar low to very low correlations to all traditional assets, including gold - representative of the potential diversification benefits. The autocorrelation function resulted in high positive serial correlations for all three cryptocurrencies, indicative of strong trending behaviour and high volatility.
Open Access
Cluster management synergy valuation: Synthesis and illustration of a discounted cash flow synergy valuation model for cluster management organisations
(2016) De Kock, Neil; Rajaratnam, Kanshukan; Kruger, Ryan
The practice of cluster management has become an integral component to the modern cluster business environment. This research develops a framework for the valuation of synergies generated by a cluster management organisation (CMO) to be used as either a method of (ex-post) management evaluation or (ex-ante) for capital budgeting purposes. The theoretical framework is synthesised from clustering and business alliance (predominantly Mergers and Acquisitions (M&A) and Joint Ventures (JV)), literature. The case of the South African Furniture Initiative (SAFI) was used to inform model development and to illustrate practical application of the theoretical synergy valuation model. The case study found that the synergy valuation model faces problems with practical application due to the wide variety of activities commonly associated with CMO goals and objectives. It concludes that even though a synergy framework would provide a useful tool for evaluation and capital budgeting, further research is required to develop a more accurate method of impact estimation.
Open Access
Comparison of Logistic Regression and Classification Trees to Forecast Short Term Defaults on Repeat Consumer Loans
(2021) Naicker, Keeland; Rajaratnam, Kanshukan
This dissertation highlights the performance comparison between two popular contemporary consumer loan credit scoring techniques, namely logistic regression and classification trees. Literature has shown logistic regression to perform better than classification trees in terms of predictiveness and robustness when forecasting consumer loan default events over standard twelve-month outcome periods. One of the major shortcomings with classification trees is its tendency to overfit data eroding its robustness, making it vulnerable to underlying population characteristic shifts. Classification trees remains a popular technique due to its ease of application (algorithm machine learning basis) and model interpretation. Past research has found classification trees to perform marginally better than logistic regression with respect to predictiveness and robustness when modelling short term consumer credit default outcomes related to previously unseen new customer credit loan applications. This dissertation independently tested this finding on reloan consumer loan data, repeat customers who renewed loan facilities at a significant South African micro lender. This dissertation tests the finding if the classification tree technique would outperform logistic regression when modelling this new type of loan data. Credit scoring models were built and tested for each respective technique across identical data sets with the intent to eliminate bias. Robustness tests were constructed via careful iterative data splits. Performance tests measuring predictiveness and robustness were conducted via the weighted sums of squared error evaluation approach. Results reveal logistic regression to outperform classification trees on predictiveness and robustness across the designed uniform iterative data splits, which suggests that logistic regression remains the superior technique when modelling short term credit default outcomes on reloan consumer loan data.
Open Access
A cost benefit analysis of operational risk quantification methods for regulatory capital
(2016) Nyathi, Mandla; Rajaratnam, Kanshukan; Toerien, Francois
Operational risk has attracted a sizeable amount of attention in recent years as a result of massive operational losses that headlined financial markets across the world. The operational risk losses have been on the back of litigation cases and regulatory fines, some of which originated from the 2008 global financial crisis. As a result it is compulsory for financial institutions to reserve capital for the operational risk exposures inherent in their business activities. Local financial institutions are free to use any of the following operational risk capital estimation methods: Advanced Measurement Approach (AMA), the Standardized (TSA) and/ the Basic Indicator Approach (BIA). The BIA and TSA are predetermined by the Reserve Bank, whilst AMA relies on internally generated methodologies. Estimation approaches employed in this study were initially introduced by the BCBS, largely premised on an increasingly sophisticated technique to incentivise banks to continually advance their management and measurement methods while benefiting from a lower capital charge through gradating from the least to the most sophisticated measurement tool. However, in contrast to BCBS's premise, Sundmacher (2007), whilst using a hypothetical example, finds that depending on a financial institution's distribution of its Gross Income, the incentive to move from BIA to TSA is nonexistent or marginal at best. In this thesis I extend Sundmacher (2007)'s work, and I test one instance of AMA regulatory capital (RegCap) against that of TSA in a bid to crystalise the rand benefit that financial institutions stand to attain (if at all) should they move from TSA to AMA. A Loss Distribution Approach (LDA), coupled with a Monte Carlo simulation, were used in modelling AMA. In modelling the loss severities, the Lognormal, Weibull, Burr, Generalized Pareto, Pareto and Gamma distributions were considered, whilst the Poisson distribution was used for modelling operational loss frequency. The Kolmogorov-Smirnov and Akaike information criterion tests were respectively used for assessing the level of distribution fit and for model selection. The robustness and stability of the model were gauged using stress testing and bootstrap. The TSA modelling design involved using predetermined beta values for different business lines specified by the BCBS. The findings show that the Lognormal and Burr distributions best describes the empirical data. Additionally, there is a substantial incentive in terms of the rand benefit of migrating from TSA to AMA in estimating operational risk capital. The initial benefit could be directed towards changes in information technology systems in order to effect the change from TSA to AMA. Notwithstanding that the data set used in this thesis is restricted to just one of the "big four banks" (owing to proprietary restrictions), the methodology is representable (or generalisable) to the other big banks within South Africa. The scope of this study can further be extended to cover Extreme Value Theory, Non-Parametric Empirical Sampling, Markov Chain Monte Carlo, and Bayesian Approaches in estimating operational risk capital.
Open Access
Determinants of the cost of credit for project finance debt in Africa
(2016) Hatzilambros, Constantin; Rajaratnam, Kanshukan
This study investigates the characteristics of project finance transactions and establishes the cost determinants for non-recourse project finance in Africa within the energy, oil and gas, mining and infrastructure sectors. Essentially, this thesis will be investigating what the main cost determinants are which lenders use to price the risk in project finance transactions. Project finance risks such as market, operational, sponsor, political / regulatory and environmental risks are investigated. A loan transaction database is used to fit these risks to determine the relevant loan parameters available in the database, employing a regression model is used to obtain which loan parameters, and, in turn, risks, lenders price into the cost of the loans. The database represents non-recourse project finance transactions throughout Africa from 1995 to 2015 and was filtered down 89 loan entries that contained the most important loan parameters. Empirical results suggest that secured loans are priced in a different category to unsecured loans, increasing the All-In credit-spread by 196.94 bps (P-value < 0.1%) if the loan parameter is moved from an unsecured to a secured loan. Political / regulatory risk, which had a 27.697 bps increase in the All-in Credit-spread (P-value < 2.3%). This can be attributed to being a result of a country's risk ranking, which was found to be the most significant pricing determinant for non-recourse loans on the African continent.
Open Access
The effect of FED's quantitative easing policy on listed companies and sectors in South Africa
(2017) Chacha, Terry; Rajaratnam, Kanshukan
This paper examines the effect of Quantitative Easing (QE) on listed companies and sectors in South Africa. The unconventional monetary policy carried out by the developed markets had spill over effects in emerging market economies. We focus on the policies performed by the United States. Our interest is to find out whether the QE announcements had any impact on the returns of listed companies and sectors in South Africa. An exploratory analysis is done on the macroeconomic and financial indicators in SA to provide grounds for doing the analysis on the listed companies. This analysis shows that the exchange rate and portfolio inflows were impacted by QE. However, other local factors were in play in affecting the exchange rate. The shrinkage in the global economic activity affected the Gross Domestic Product (GDP) growth rate. The changes in inflation cannot be attributed to QE. Most of the portfolio inflows were in the bond market and since some were directed to the equity market we proceed to check whether stocks and sectors had abnormal returns as a result. Our empirical analysis shows that only three companies had significant Cumulative Abnormal Returns (CARs) in the three phases of QE. On the sector front, nine out of the 34 sectors had significant CARs every time QE was announced. A broader classification of these sectors into industries shows that the industries represented are industrials, consumer goods, consumer services and financials. In QE1, the industrials industry and the consumer services industry had negative CARs but in QE2 and QE3, they had positive CARs. The consumer goods industry had positive CARs during the three phases of QE. This research concludes that QE1 had the greatest impact on the Johannesburg Stock Exchange (JSE) and its impact was negative. QE2 had a positive impact on the JSE since most companies and sectors had significant positive CARs. The impact of QE3 on sector abnormal returns was almost neutral. We also provide an investment strategy on the JSE using various indices for the periods following QE2 and QE3. Out of the 14 indices used, the small caps index is given a higher weighting in both portfolios due to its low risk.
Open Access
Enhancing students' learning through practical knowledge taught by industry professionals
(Clute Institute, 2013) Rajaratnam, Kanshukan; Campbell, Anita
A topic of interest in teaching business courses is incorporating the practical aspect of the subject matter into teaching as this helps to bridge theory and real-world practice. Research indicates that students gain a deeper understanding of material when theory is contextualized through real-life practical examples. However, given the traditional career-path of academics in finance in countries such as South Africa, a significant proportion of finance lecturers have little or no relevant practical experience in the subject matter. In this paper, we discuss a strategy implemented in finance courses at sophomore and senior levels in order to link theory and practice. Guest speakers were invited from industry to contextualize the topics for the students. Students' perceptions on the benefit they derived from the speakers were deduced from statistical analyses of student evaluations. The results indicate that the experience was positive and aided in their understanding of the subject.
Open Access
Evaluation of Asset Pricing Models in the South African Equities Market
(2020) Moyo, Nigel A P; Rajaratnam, Kanshukan
Asset pricing models have been of interest since their origin in modern finance. The Capital Asset Pricing Model is a widely used tool and is one of the early developed asset pricing models in modern finance. There are continual improvements of this model with the evident multifactor models of Fama and French (2015), Carhart (1997) and the South African two – factor arbitrage pricing models of Van Rensburg (2002) and Laird-Smith et al. (2016). This research empirically investigates the performance of eight-different multi-factor asset pricing models in describing average portfolio returns in the South African Johannesburg Stock Exchange. We find that the Carhart (1997) four factor model comprising of the market factor, size factor, value factor and the momentum factor is the most parsimonious model and thus better explains the average portfolio returns in the South African JSE. This model is an improvement of the Fama and French (1992) three factor model. Additionally, we investigate the performance of the two factor Asset Pricing Theory (APT) model of Laird-Smith et al. (2016) and Van Rensburg (2002) that consists of the South African Financial Index (SAFI) and the South African Resources Index (SARI). We observe that the model performs better than the traditional CAPM that is widely used in industry. Adding the SAFI and the SARI to the six-factor model results in an eight-factor model that has a significant improvement in explaining average returns. The results indicate that the market factor, the South African Financial Index and the South African Resources Index (SARI) poorly explain each other but their linear combination improves the eight-factor asset pricing model in explaining average portfolio returns in the South African market. The eight – factor model comprises of the market, size, value, investment, profitability, momentum factors and the two South African indices namely, the South African Financials Index (SAFI) and the South African Resources Index (SARI).
Open Access
An investigation of empirical properties of South African bonds
(2017) Mate, Janet; Rajaratnam, Kanshukan; Majoni, Akios
This study investigates empirical properties of South African bonds over the period 2000 to 2016. In particular, it investigates i) mean reversion in bond returns; ii) the correlation between bond returns and the inflation rate; and, iii) the correlation between bond returns and equity returns. An understanding of bond return dynamics would allow bond investors to assess which bond properties work in their favour. Thus this study seeks to guide bond investors, and to add to the knowledge of the bond market concerning bond return dynamics in an emerging market economy. The study employs a quantitative research methodology, using a nonexperimental research design. The investigation is carried out at the macroeconomic level using the JSE All Bond Indices as the bond investment proxy, the FTSE/JSE All Share Total Return Index as the equity investment proxy, and the Consumer Price Index as the proxy used to measure the inflation rate. The sample autocorrelation function is used to test for mean reversion and the Kendall Tau-b correlation test is used for the correlation investigations. This study does not find statistically significant evidence of long term mean reversion but finds statistically significant evidence of short-term mean reverting behaviour in the period 2013-2016. Furthermore, this study reveals that short-term serial correlations vary and are sensitive to political developments in the economy. The correlation analysis between bond returns and the inflation rate and bond returns and stock returns did not return statistically significant correlation values. However, further analysis provided evidence against the use of bonds as an inflation hedge and of diversification benefits to be reaped from combining bonds and stocks together in a portfolio.
Open Access
Leading the Economy: Do both the banking sector and stock market development independently lead economic activity in South Africa? Evidence from South Africa using Granger-causality tests
(2018) Mellon, Richard; Rajaratnam, Kanshukan
This study investigates and concludes on thus the predictability of economic activity in South Africa through the use of share price indices and banking sector asset levels as leading indicators. This study investigates share price and banking sector information using both nominal and real quarterly and monthly time-series data for the period March 1998 to October 2017. For share prices analysis, it takes market segmentation on the JSE into account, and examines causality between the All-Share index, the Industrial index, the Resources index and the Financial index against GDP and the index of Industrial Production (IIP) using the test proposed by Granger (1969). For the banking sector development analysis, it takes the South African Reserve Bank’s disclosure of all banking institution assets and examines if changes in those asset levels Granger-cause the changes to GDP and IIP. This study builds on the work performed by previous studies, specifically on Sayed, Auret and Page (2017) and Har, Ee, and Tan (2009), which test the leading relationship of the stock market for economic activity. This study adopts a similar approach to these studies, while also making certain adjustments and additions to their methodology, aiming to produce more robust findings. This study not only tests for a new relationship between banking development and economic activity, but it also conducts several additional stationarity tests to provide more conclusive evidence of the data’s stationarity before Granger-causality testing is performed. The additional stationarity tests in this study establish that some time series data, which Sayed et al. (2017) concluded to be stationary, is in fact not stationary, and this contrary finding directly impacts the subsequent Granger-causality testing and results. This study also notes and corrects Sayed et al.'s(2017) methodology which fails to perform subsequent stationarity testing on its differenced time series data, and thus fails to prove that the transformed data is satisfactorily stationary and acceptable for Granger-causality testing. Another adjustment we make to the methodology is the interpretation of the Index for Industrial Production (IIP), which we view as a volume based index rather than a price based index that can be adjusted for inflation, which was the position of Sayed et al. (2017). The empirical investigation of this study reveals some positive evidence in favour of the JSE as a leading indicator of economic activity, where unidirectional causality is established between the four market segmentation proxies and the macroeconomic variables. This is however less conclusive than the findings of previous South African studies, which is explained. For the banking sector’s development analysis, the empirical tests produce inconsistent findings across monthly and quarterly data, leaving one unable to confirm a causal relationship existing between the banking sector’s development and economic activity.
Open Access
Loss distributions in consumer credit risk : macroeconomic models for expected and unexpected loss
(2016) Malwandla, Musa; Rajaratnam, Kanshukan; Clark, Allan
This thesis focuses on modelling the distributions of loss in consumer credit arrangements, both at an individual level and at a portfolio level, and how these might be influenced by loan-specific factors and economic factors. The thesis primarily aims to examine how these factors can be incorporated into a credit risk model through logistic regression models and threshold regression models. Considering the fact that the specification of a credit risk model is influenced by its purpose, the thesis considers the IFRS 7 and IFRS 9 accounting requirements for impairment disclosure as well as Basel II regulatory prescriptions for capital requirements. The thesis presents a critique of the unexpected loss calculation under Basel II by considering the different ways in which loans can correlate within a portfolio. Two distributions of portfolio losses are derived. The Vašíček distribution, which is the assumed in Basel II requirements, was originally derived for corporate loans and was never adapted for application in consumer credit. This makes it difficult to interpret and validate the correlation parameters prescribed under Basel II. The thesis re-derives the Vašíček distribution under a threshold regression model that is specific to consumer credit risk, thus providing a way to estimate the model parameters from observed experience. The thesis also discusses how, if the probability of default is modelled through logistic regression, the portfolio loss distribution can be modelled as a log-log-normal distribution.
Open Access
Market Betas on the JSE: Factor selection, estimation and empirical evaluation
(2017) Laird-Smith, James; Rajaratnam, Kanshukan
This paper examines the nature and significance of market betas on the Johannesburg Stock Exchange (JSE). The identity of market betas is determined by means of Principal Component Analysis (PCA) performed on the returns of the FTSE/JSE Africa Index Series. A scree test shows two factors necessary for inclusion in the appropriate Arbitrage Pricing Theory (APT) model. Based on the promax rotated factor loadings, it is argued that the Financials (J580) and Basic Materials (J510) indices ought be used as the appropriate observable index proxies for the first and second factors respectively. Regarding the estimation of beta, this paper makes the case for the use of Reduced Major Axis (RMA) regression over the traditional Ordinary Least Squares (OLS) approach. A number of characteristics are assessed when arriving at this conclusion. Importantly, it is shown that the traditional OLS regression method chronically underestimates the magnitude of the beta parameter whereas RMA regression does not. In addition, it is shown that, while OLS beta values are more stable in absolute terms than RMA beta values, the RMA values are more stable when adjusted for their magnitude. This paper does not make use of a thin trading filter to narrow the sample of stocks for empirical evaluation. Instead, an examination is made of the significance of beta values at the point at which they are estimated. This is accomplished by means of a rolling window of regressions. It is shown that, while most stocks do exhibit betas which are consistently significant over their listing period, many stocks do not. Some stock returns result in almost no significant beta values while some others exhibit beta values which are significant for only a portion of their listing period. It is shown that a median beta p-value value of 5% is an appropriate 'significance filter' for limiting the sample of stocks to only those significant for the majority of their listing period. Using only these stocks, an empirical evaluation of beta is conducted using portfolios sorted on both OLS and RMA beta values. It is found that neither beta measure explains the cross-section of returns in the case of resource stocks. However, in the case of non-resource stocks the results show a clear divergence between the methods. In the case of OLS sorted portfolios, the results show a negative relationship between beta and returns. This surprising and counterintuitive result has also been arrived at by other researchers and is the opposite of what the APT would predict. However, in the case of RMA sorted portfolios, this pattern reverses itself, showing a positive relationship between beta and returns. For some holding periods, this is shown to be significant, providing evidence in support of the APT. As a result it is demonstrated that OLS regression not only underestimates the magnitude of beta, but that it distorts the results of empirical tests. On this basis it is argued that RMA regression ought replace OLS regression as the preferred method of beta estimation for the JSE.
Open Access
Natural Language Financial Forecasting: The South African Context
(2021) Katende, Simon; Er, Sebnem; Nyirenda, Juwa; Rajaratnam, Kanshukan
The stock market plays a fundamental role in any country's economy as it efficiently directs the flow of savings and investments of an economy in ways that advances the accumulation of capital and the production of goods and services. Factors that affect the price movement of stocks include company news and performance, macroeconomic factors, market sentiment as well as unforeseeable events. The conventional prediction approach is based on historical numerical data such as price trends and trading volumes to name a few. This thesis reviews the literature of Natural Language Financial Forecasting (NLFF) and proposes novel implementation techniques with the use of Stock Exchange News Service (SENS) announcements to predict stock price trends with machine learning methods. Deep Learning has recently sparked interest in the data science communities, but the literature on the application of deep learning in stock prediction, especially in emerging markets like South Africa, is still limited. In this thesis, the process of labelling announcements, the use of a more statistically relevent technique called the event study was used. Classical textual preprocessing and representation techniques were replaced with state-of-the-art sentence embeddings. Deep learning models (Deep Neural Network (DNN)) were then compared to Classical Models (Logistic Regression (LR)). These models were trained, optimized and deployed using the Tensorflow Machine Learning (ML) framework on Google Cloud AI Platform. The comparison between the performance results of the models shows that both DNN and LR have potential operational capabilites to use information dissemination as a means to assist market participants with their trading decisions.
Open Access
Portfolio diversification utilising rolling economic drawdown constraints and risk factor analysis
(2018) Mills, Bradley; Rajaratnam, Kanshukan
This study investigates a new asset allocation technique termed Factor Adjusted Rolling Economic Drawdown (FAREDD), whereby resources are allocated to different assets by way of integrating Principle Component Analysis (PCA) with existing Rolling Economic Drawdown Methods (REDD). The primary purpose of this model is to create a portfolio with low drawdown levels, that can withstand turbulent market periods thus protecting portfolio value through providing stronger diversification benefits while still seeking to maximise risk adjusted and overall return. This will have strong implications for investors as it could provide an additional method and tool to be considered during the asset allocation decision stage if they have a strong drawdown aversion. The concept of FAREDD is developed in this study within a South African context and compares this method with several traditional allocation methods including mean-variance optimised models, risk parity as well as traditional rolling economic drawdown models. So far, at the point of writing this study, the author has been unable to find any previous studies documenting this type of application of PCA to REDD. In addition to this, all previous studies that has investigated rolling economic drawdown has been conducted exclusively on the United States of America. The literature finds that REDD provides a viable and superior alternative to traditional asset allocation in the long run. Thus, as part of this study, a second objective is to investigate whether REDD models provide sufficient protection and superior returns in a developing economy with a significantly lower number of available liquid assets and higher volatility due to increased political, economic and business risk, when compared to alternative more traditional allocation techniques. The key findings of this study are that the FAREDD model does outperform the traditional REDD model that it is compared to for the period and it also meets the objective of providing low drawdowns and volatility while achieving strong risk-adjusted returns. However, the model does not provide the strongest drawdown protection of all portfolios tested. The FAREDD model is surpassed by the minimum-variance portfolio in this regard but from a risk adjusted basis and an overall return perspective it far outperforms the minimum-variance portfolio. Therefore, the performance of the FAREDD model is mixed and its optimality would need to be assessed relative to an investor’s risk appetite and risk-return trade-off. In addition to this, the paper finds that the performance of traditional REDD models in the South African context are mixed when compared to traditional asset allocation techniques thereby indicating that REDD models may not be superior in the South African market place at all times. However, they can provide relevant and potential asset allocation alternatives for mangers to consider.
Open Access
Pricing Offshore Services: Evidence from the Paradise Papers
(2021) Gawronsky, Marcus; Gebbie, Timothy; Rajaratnam, Kanshukan
The Paradise Papers represent one of the largest public data leaks comprising 13.4 million con_dential electronic documents. A dominant theory presented by Neal (2014) and Gri_th, Miller and O'Connell (2014) concerns the use of these offshore services in the relocation of intellectual property for the purposes of compliance, privacy and tax avoidance. Building on the work of Fernandez (2011), Billio et al. (2016) and Kou, Peng and Zhong (2018) in Spatial Arbitrage Pricing Theory (s-APT) and work by Kelly, Lustig and Van Nieuwerburgh (2013), Ahern (2013), Herskovic (2018) and Proch_azkov_a (2020) on the impacts of network centrality on _rm pricing, we use market response, discussed in O'Donovan, Wagner and Zeume (2019), to characterise the role of offshore services in securities pricing and the transmission of price risk. Following the spatial modelling selection procedure proposed in Mur and Angulo (2009), we identify Pro_t Margin and Price-to-Research as firm-characteristics describing market response over this event window. Using a social network lag explanatory model, we provide evidence for social exogenous effects, as described in Manski (1993), which may characterise the licensing or exchange of intellectual property between connected firms found in the Paradise Papers. From these findings, we hope to provide insight to policymakers on the role and impact of offshore services on securities pricing.
Open Access
Quantitative models for prudential credit risk management
(2021) Malwandla, Musa; Rajaratnam, Kanshukan; Clark, Allan
The thesis investigates the exogenous maturity vintage model (EMV) as a framework for achieving unification in consumer credit risk analysis. We explore how the EMV model can be used in origination modelling, impairment analysis, capital analysis, stress-testing and in the assessment of economic value. The thesis is segmented into five themes. The first theme addresses some of the theoretical challenges of the standard EMV model – namely, the identifiability problem and the forecasting of the components of the model in predictive applications. We extend the model beyond the three time dimensions by introducing a behavioural dimension. This allows the model to produce loan-specific estimates of default risk. By replacing the vintage component with either an application risk or a behavioural risk dimension, the model resolves the identifiability problem inherent in the standard model. We show that the same model can be used interchangeably to produce a point-in-time probability forecast, by fitting a time series regression for the exogenous component, and a through-the-cycle probability forecast, by omitting the exogenous component. We investigate the use of the model for regulatory capital and stress-testing under Basel III, as well as impairment provisioning under IFRS 9. We show that when a Gaussian link function is used the portfolio loss follows a Vašíček distribution. Furthermore, the asset correlation coefficient (as defined under Basel III) is shown to be a function of the level of systemic risk (which is measured by the variance of the exogenous component) and the extent to which the systemic risk can be modelled (which is measured by the coefficient of determination of the regression model for the exogenous component). The second theme addresses the problem of deriving a portfolio loss distribution from a loan-level model for loss. In most models (including the Basel-Vašíček regimes), this is done by assuming that the portfolio is infinitely large – resulting in a loss distribution that ignores diversifiable risk. We thus show that, holding all risk parameters constant, this assumption leads to an understatement of the level of risk within a portfolio – particularly for small portfolios. To overcome this weakness, we derive formulae that can be used to partition the portfolio risk into risk that is diversifiable and risk that is systemic. Using these formulae, we derive a loss distribution that better-represents losses under portfolios of all sizes. The third theme is concerned with two separate issues: (a) the problem of model selection in credit risk and (b) the problem of how to accurately measure probability of insolvency in a credit portfolio. To address the first problem, we use the EMV model to study the theoretical properties of the Gini statistic for default risk in a portfolio of loans and derive a formula that estimates the Gini statistic directly from the model parameters. We then show that the formulae derived to estimate the Gini statistic can be used to study the probability of insolvency. To do this, we first show that when capital requirements are determined to target a specific probability of solvency on a through-the-cycle basis, the point-in-time probability of insolvency can be considerably different from the through-the-cycle probability of insolvency – thus posing a challenge from a risk management perspective. We show that the extent of this challenge will be greater for more cyclical loan portfolios. We then show that the formula derived for the Gini statistic can be used to measure the extent of the point-in-time insolvency risk posed by using a through-the-cycle capital regime. The fourth theme considers the problem of survival modelling with time varying covariates. We propose an extension to the Cox regression model, allowing the inclusion of time-varying macroeconomic variables as covariates. The model is specifically applied to estimate the probability of default in a loan portfolio, where the experience is decomposed the experience into three dimensions: (a) a survival time dimension; (b) a behavioural risk dimension; and (c) calendar time dimension. In this regard, the model can also be viewed as an extension of the EMV model – adding a survival time dimension. A model is built for each dimension: (a) the survival time dimension is modelled by a baseline hazard curve; (b) the behavioural risk dimension is modelled by a behavioural risk index; and (c) the calendar time dimension is modelled by a macroeconomic risk index. The model lends itself to application in modelling probability of default under the IFRS 9 regime, where it can produce estimates of probability of default over variable time horizons, while accounting for time-varying macroeconomic variables. However, the model also has a broader scope of application beyond the domains of credit risk and banking. In the fifth and final theme, we introduce the concept of embedded value to a banking context. In longterm insurance, embedded value relates to the expected economic value (to shareholders) of a book of insurance contracts and is used for appraising insurance companies and measuring management's performance. We derive formulae for estimating the embedded value of a portfolio of loans, which we show to be a function of: (a) the spread between the rate charged to the borrower and the cost of funding; (b) the tenure of the loan; and (c) the level of credit risk inherent in the loan. We also show how economic value can be attributed between profits from maturity transformation and profits from credit and liquidity margin. We derive formulae that can be used to analyse the change in embedded value throughout the life of a loan. By modelling the credit loss component of embedded value, we derive a distribution for the economic value of a book of business. The literary contributions made by the thesis are of practical significance. The thesis offers a way for banks and regulators to accurately estimate the value of the asset correlation coefficient in a manner that controls for portfolio size and intertemporal heterogeneity. This will lead to improved precision in determining capital adequacy – particularly for institutions operating in uncertain environments and those operating small credit portfolios – ultimately enhancing the integrity of the financial system. The thesis also offers tools to help bank management appraise the financial performance of their businesses and measure the value created for shareholders.