Limitations of Capital Asset Pricing Model (CAPM) continue to present inconsistent empirical results despite its rm mathematical foundations provided in recent studies. In this thesis, we examine how estimation errors of the CAPM could be minimized using the cross-validation technique, a concept that is widely applied in machine learning (CV-CAPM). We apply our approach to test the assumption of CAPM as a well-diversified portfolio model with data from S&P500 and Dow Jones Industrial Average (DJIA). Our results from the CV-CAPM validate that both S&P500 and DJIA are well-diversified market indices with statistically insignificant variation in unsystematic risks during and after the 2007 financial crisis. Furthermore, the CV-CAPM provides the smallest root mean square errors and mean absolute deviations compared to the traditional CAPM.

Historic market volatility has made risk management decisions by firms in the agricultural supply chain more challenging. Market risk measurement methods, such as Value-at-Risk, were developed in the financial industry to objectively measure, and thus better comprehend, market risk's effect on positions. This thesis gives a thorough background of the issues involved with risk measurement. Different scenarios were then used to demonstrate how the risk measurement method can be applied to the agricultural processing margin. In this thesis, the flour milling margin was used to demonstrate how a firm can incorporate sophisticated risk analytics into its risk management decision making process. Multiple scenarios were developed to account for different situations faced by flour millers. Ocean freight, exchange rate risk, futures price risk, basis risk and flour price risk are all included to provide examples of how market risk measurement can be beneficial to industry participants.

This thesis explores and evaluates the forecasting application of the machine learning method Gradient Boosting Decision Trees. This method is used to forecast the demand of the online grocery market with a 7-day time horizon. The thesis was conducted in collaboration with the online grocery company Mathem. The model is applied and evaluated on three different periods representing the spring, summer and fall. The main evaluation metric is the mean absolute percentage error (MAPE), and clear differences were found depending on the predictability of the period. Apart from the model and its application to demand forecasting, the related risk was investigated. This was done by studying the Value-at-Risk and Expected Shortfall associated with discrepancies between the forecasted and actual values over the three periods. The most important conclusion of the case study at Mathem is that overestimation in the forecast is more costly in terms of monetary value than underestimating. It is also found that this is highly dependent on the cost structure of the company's operation and could therefore vary between companies. Thus, the study has contributed to understanding the applications of machine learning models in forecasting processes as well as the risks related to over/underestimating the demand of the online grocery market. / I denna uppsats utforskas och utvärderas maskininlärningsmetoden Gradient Boosting Decision Trees och dess tillämpningsområde inom prognostisering. Metoden används för att prognostisera efterfrågan på onlinehandel av dagligvaror med en 7 dagars tidshorisont. Uppsatsen gjordes i samarbete med företaget Mathem som är aktiva inom denna sektor. Modellen appliceras och utvärderas på tre olika tidsperioder som representerar våren, sommaren och hösten. Modellen själv utvärderas med avseende på måttet mean absolute percentage error (MAPE) och tydliga skillnader mellan tidsperioderna observerades relaterat till variansen i datan. Förutom modellen och dess applikation på prognostisering av efterfrågan utforskades även de relaterade riskerna. Detta genomfördes genom att studera riskmåtten Value-at-Risk och Expected Shortfall baserade på differensen mellan de prognostiserade och de faktiska värdena under dessa tre tidsperioder. Den viktiga slutsatsen som kunde dras från uppsatsen i samband med Mathem var att det, i monetära termer, är dyrare att prognostisera för högt än för lågt. En ytterligare slutsats som kan dras är att den monetära risken är mycket beroende av kostnadsstrukturen hos företaget och skulle därför kunna variera mellan olika företag. Således har denna uppsats bidragit till förståelse för användandet av maskininlärning för prognostisering samt riskerna kopplade till över- och underskattning av efterfrågan i marknaden för dagligvaruhandel online.

This dissertation examines the challenges associated with disruptions in supply chain financing and the railway transportation network. The study is divided into six chapters: In Chapter 1, we introduce the core problems under investigation. Chapter 2 investigates supply chain financing, emphasizing trade credit and bank credit—two predominant external financing mechanisms. Given the inherent uncertainties in demand, interest rates, and supplier credit ratings, this chapter introduces a stochastic programming model accounting for demand uncertainty. Subsequently, a robust optimization program is applied, whose complexity demands a specialized solution methodology. By analyzing a case study centered around a prominent U.S. retailer, the research reveals key insights into decision-making processes related to financing, the effects of bargaining power on portfolio mix and profits, and the relative importance of interest rate uncertainties over supplier credit ratings. Chapter 3 introduces a game-theoretical model designed to hedge financing risks in supply chains, with a focus on the application of insurance for both trade and bank credits. To support the design of effective supply chain finance contracts, three distinct contracts are developed, aiming to synchronize both financial and material flows within the supply chain. A significant feature of this chapter is the data-driven approach employed to address the potential bankruptcy risks that can arise from borrowing loans. Alongside this, a novel solution algorithm is introduced to solve the proposed non-convex models. A case study involving Ford Motor Company and a Chicago-based retailer enriches the research with real-world context. The findings offer several managerial insights: the strategic advantages of different insurance services vary based on the risk attitudes and profit margins of participants. For example, when a retailer operates with a lower profit margin, the use of Trade Credit Insurance (TCI) is recommended in conjunction with a risk-seeking retailer, while a risk-averse retailer might diminish the benefits of TCI. Conversely, with high profit margin retailers, the adoption of Payment Protection Insurance (PPI) is advised under all conditions. In Chapter 4, a game-theoretical model for risk mitigation within railway transportation is introduced. This model addresses random disruptions by employing strategies like repair, re-routing, third-party services, and leasing capacity from competing rail companies. Through a U.S. case study, the efficacy of these strategies is examined, with renting railcars emerging as a particularly potent approach to enhance resilience and reduce third-party expenses. The research further suggests that negotiations extending delivery dates can significantly diminish post-disruption costs. Finally, Chapter 5 summarizes the primary contributions of this research, laying the groundwork for prospective studies in this domain. / Thesis / Doctor of Philosophy (PhD)

No description available.

Credit risk mitigation that can be applied by commercial banks in assessing the lending decision /credit risk when advances and equity investments are considered for BEE classified companies. / A research report presented to the Graduate School of Business Leadership, University of South Africa / The previous political dispensation limited black people’s participation in the South African economy. Poor credit records, lack of training, resulting in skills and capacity gaps further limited entry into the lending market. These aspects are considered the main limitations in obtaining finance for the Small, Medium and Micro Enterprises (SMMEs). This research report focuses on how credit risk can be mitigated by commercial banks in lending to Black Economic Empowerment (BEE) companies in the medium to large market. Exploratory research was conducted using various methods to achieve methodological triangulation. These methods consisted of a literature review, interviewing experts in the field and case studies. A qualitative research approach was followed. It was found that the lack of own contribution and security were still prevalent in the medium to large market, but the quality of management (little training and skills) was deemed not to be a limitation as suitable credit risk mitigants were identified. No credit risk mitigants were identified to mitigate poor credit records. It is postulated that by adopting and applying the identified credit risk mitigants, commercial banks can increase their success rate in lending to BEE companies. It will further assist in the transformation of black people and compliance with the Financial Services Charter. It is recommended that a similar study be conducted in the agriculture, hunting, forestry and fishing industry. The reasons why BEE companies applications are declined could also be investigated. Further studies could also explore other external factors such as economical, legal and social that could have an influence on the funding of BEE companies.

Thesis (MBA)--Stellenbosch University, 2008. / ENGLISH ABSTRACT: The budgeting process is an integral part of the annual business cycle of most organisations. The budget consists of numerous uncertain inputs, which are frequently used to produce a single EBIT figure. This implies that there is a risk of not achieving the budget that is not quantified and apparent from the prepared budget. In this report, the differences between the budgets of two business units of Exxaro Resources were analysed to gain a better understanding of the information hidden beyond the figures quoted on the surface. The budgets of Exxaro KZN Sands, a heavy minerals producer, and Zincor, a zinc refinery, were analysed to compare the respective risk approach of each. Simplified deterministic models were first constructed that contained the most important budget risk drivers. These were validated with comparisons to the official budgets. Historical actual data from 2006 and 2007 was then obtained from the business archives for the risk drivers. Probability distributions were then generated that fit the distributions of the historical data. These risk distributions were then used as input variables in a Monte Carlo simulation, performed in Crystal Ball. The EBIT for each business was thus simulated as a probability distribution. The simulation showed that the two business units applied very different approaches to budget risk. The actual budgeted EBIT of Exxaro KZN Sands of a loss of R167 579 945 had a more than 99% chance of being exceeded, showing a very conservative, worst case approach to budgeting. Zincor had only a 29% probability of exceeding their budgeted EBIT of R202 783 091, and incorporated a much larger risk of not achieving EBIT into the budget. The budgets of both business units were not suitable for the most important functions of budgeting, namely target setting, strategic planning and valuation of the business. It is recommended that Exxaro implements a procedure to standardise the risk approach to budgeting in the organisation. The budget process must firstly have guidelines to indicate how risk drivers’ values should be chosen for the official budget. Recommendations regarding average values, best three months or any other methodology will ensure that different business units follow a comparable approach. Secondly, Monte Carlo simulation must be performed on simplified business models. The KPI trees currently being used for continuous improvement provide a base model for this purpose. The Monte Carlo simulation will provide a more sophisticated and quantified analysis of risk, and give a further indication of the inherent variability of a specific business unit. Lastly, scrutiny of the Monte Carlo can indicate the biggest drivers of risk. Measures can then be implemented to better understand, or reduce, the variability of the main risk drivers. This will lead to more accurate budgeting, and a better understanding of the inherent budget risk. / AFRIKAANSE OPSOMMING: Die begrotingsproses is ‘n integrale deel van die jaarlikse besigheidsiklus van meeste organisasies. Die begroting bestaan uit etlike onseker insette, maar word meestal gebruik om ‘n enkele syfer vir inkomste te bereken. Dit beteken dat daar ‘n risiko is dat die begroting nie behaal gaan word nie, wat nie duidelik na vore tree in die begroting nie. In hierdie verslag word die verskille tussen die begrotings van twee besigheidseenhede van Exxaro Resources geannaliseer om insig te verkry rakende die inligting versteek agter die ooglopende getalle. Die begrotings van Exxaro KZN Sands, ‘n swaar minerale produsent, en Zincor, ‘n zink rafinadery, is geannaliseer om die onderskye risikobenaderings te vergelyk. Die eerste stap was om vereenvoudigde deterministiese modelle te bou wat die belangrikste begrotingsrisikodrywers bevat het. Die modelle is gevalideer deur die winste te vergelyk met die amptelike besigheidsbegrotings. Historiese data van 2006 en 2007 is versamel van die risikodrywers. Verdelings van waarskynlikheid is toe gekies wat die historiese data beskryf het. Die verdelings is gebruik as inset veranderlikes in ‘n Monte Carlo simulasie, gedoen in Crystal Ball. Die wins van elke besigheid is dan as ‘n waarskynlikheidsverdeling gegenereer. Die simulasie het aangetoon dat die twee besighede uiteenlopende benaderings tot begrotingsrisiko het. Die begrote verlies van R167 579 945 van Exxaro KZN Sands het ‘n hoër as 99% kans gehad om behaal te word. Dit dui op ‘n uiters konserwatiewe benadering, met die mees pessimistiese waardes vir risiko drywers in die begroting. Zincor het sleg ‘n 29% waarskynlikheid gewys om die begrote wins van R202 783 091 te behaal, en het aansienlik meer risiko in die begroting ingebou. Beide die benaderings was nie geskik vir meeste van die funksies waarvoor begrotings gebruik word nie, naamlik doelwitstelling, strategiese beplanning en waardasie van die besigheid. Dit word aanbeveel dat Exxaro ‘n prosedure implementeer om die risikobenadering te standariseer. Die begrotingsproses moet eerstens riglyne hê rakende die benadering tot risikodrywers. Daar moet aanbeveel word of gemiddelde waardes, beste drie maande of ‘n ander benadering gevolg moet word, om seker te maak dat verskillende besigheidseenhede dit vergelykbaar uitvoer. Tweedens moet Monte Carlo simulasie gedoen word op vereenvoudigde besigheids modelle. Die KPI bome wat tans vir deurlopende verbetering gebruik word is ‘n ideale basis vir die proses. Die Monte Carlo simulasie bied ‘n meer kwantifiseerbare benadering tot risiko analise, en dui ook aan wat die verwagte afwyking in ‘n besigheid se inkomste is. Laastens gee die Monte Carlo simulasie ‘n aanduiding oor wat die groot risikodrywers in die besigheid is. Stappe kan dan geimplimenteer word om die risikos te bestuur. Die resultaat sal meer akurate begrotings wees, asook meer insig in die inherente risiko in die begroting.

Thesis (Ph. D.)--Ohio State University, 2006. / Title from first page of PDF file. Includes bibliographical references (p. 72-76).

A guaranteed minimum income benefit (GMIB) is a long-dated option that can be embedded in a deferred variable annuity. The GMIB is attractive because, for policyholders who plan to annuitize, it offers protection against poor market performance during the accumulation phase, and adverse interest rate experience at annuitization. The GMIB also provides an upside equity guarantee that resembles the benefit provided by a lookback option. We price the GMIB, and determine the fair fee rate that should be charged. Due to the long dated nature of the option, conventional hedging methods, such as delta hedging, will only be partially successful. Therefore, we are motivated to find alternative hedging methods which are practicable for long-dated options. First, we measure the effectiveness of static hedging strategies for the GMIB. Static hedging portfolios are constructed based on minimizing the Conditional Tail Expectation of the hedging loss distribution, or minimizing the mean squared hedging loss. Next, we measure the performance of semi-static hedging strategies for the GMIB. We present a practical method for testing semi-static strategies applied to long term options, which employs nested Monte Carlo simulations and standard optimization methods. The semi-static strategies involve periodically rebalancing the hedging portfolio at certain time intervals during the accumulation phase, such that, at the option maturity date, the hedging portfolio payoff is equal to or exceeds the option value, subject to an acceptable level of risk. While we focus on the GMIB as a case study, the methods we utilize are extendable to other types of long-dated options with similar features.

A guaranteed minimum income benefit (GMIB) is a long-dated option that can be embedded in a deferred variable annuity. The GMIB is attractive because, for policyholders who plan to annuitize, it offers protection against poor market performance during the accumulation phase, and adverse interest rate experience at annuitization. The GMIB also provides an upside equity guarantee that resembles the benefit provided by a lookback option. We price the GMIB, and determine the fair fee rate that should be charged. Due to the long dated nature of the option, conventional hedging methods, such as delta hedging, will only be partially successful. Therefore, we are motivated to find alternative hedging methods which are practicable for long-dated options. First, we measure the effectiveness of static hedging strategies for the GMIB. Static hedging portfolios are constructed based on minimizing the Conditional Tail Expectation of the hedging loss distribution, or minimizing the mean squared hedging loss. Next, we measure the performance of semi-static hedging strategies for the GMIB. We present a practical method for testing semi-static strategies applied to long term options, which employs nested Monte Carlo simulations and standard optimization methods. The semi-static strategies involve periodically rebalancing the hedging portfolio at certain time intervals during the accumulation phase, such that, at the option maturity date, the hedging portfolio payoff is equal to or exceeds the option value, subject to an acceptable level of risk. While we focus on the GMIB as a case study, the methods we utilize are extendable to other types of long-dated options with similar features.