Chapter 1 examines alternative specifications of futures-based forecasting models to improve upon existing approaches constrained by restrictive assumptions and limited information sets. We replace historical averages with rolling regressions and incorporate current market information through the deviation of the current basis from its historical average. To address potential non-stationarity and structural changes in the cash-futures price relationship, we employ a five-year rolling estimation window. Our findings indicate that the rolling regression approach yields significant improvements in both accuracy and information content of cotton season-average price forecasts, primarily at short forecast horizons. Chapter 2 addresses challenges in vulnerability assessment for semi-arid regions dependent on rainfed agriculture, where extreme weather events pose significant risks to household livelihoods. Despite advancements in remotely sensed technology, accurately estimating weather variability's impact on household livelihoods remains challenging. This study evaluates the effects of weather anomaly measures, spatial resolutions (i.e., geographic level at which the weather anomaly measures are evaluated), and household characteristics on household likelihood of falling into poverty (i.e., vulnerability) estimates. Combining household consumption data for Niger with remotely sensed agro-environmental measures, we find significant variations in vulnerability estimates based on the use of various weather condition measures (3 percentage points, equivalent to 600,000 households), spatial resolutions (8 percentage points, totalling 1.6 million households), and household characteristics (10 percentage points, equivalent to approximately 2 million households). Chapter 3 evaluates student learning outcomes from student involvement in hands-on learning settings, specifically focusing on student-managed investment funds. To assess the changes in the obtained technical and practical skills, we combine knowledge tests with grading rubrics. As part of practical skills, we consider commodity market analysis, critical thinking, informed decision-making, and insightful interpretation of market analysis results. We evaluate our students' understanding of commodity markets and their practical trading skills before and after joining the student-managed investment fund program. We find significant improvements in student learning outcomes, with students showing an average increase of 28% in disciplinary or technical knowledge and 38% in practical skills. Our findings highlight the importance of hands-on learning experiences to bridge the gap between theoretical knowledge and real-world application and in developing the well-rounded skill set demanded by the job market. / Doctor of Philosophy / Chapter 1 explores several alternative specifications of futures-based forecasting models to improve existing approaches constrained by restrictive assumptions and limited information sets. Accurate prediction of cotton prices is vital for the agricultural sector, significantly impacting decisions made by farmers, traders, and policymakers. Reliable forecasts enable farmers to optimize their planting and harvesting strategies, allow traders to manage risk more effectively, and guide policymakers in developing informed agricultural policies. However, the inherent volatility of commodity markets, particularly cotton, presents substantial challenges to price forecasting. Traditional forecasting methods often struggle to capture rapid market changes, resulting in less reliable predictions. Our proposed more responsive forecasting approaches lead to a significant gain in accuracy and information content of cotton price projection and provide valuable insights that can enhance decision-making processes throughout the cotton industry. Chapter 2 explores how extreme weather events, like droughts, affect households in semi-arid regions where people's livelihood largely depends on rain-fed farming. While satellite technology helps monitor environmental changes, it is still challenging to accurately measure how weather changes impact people's lives. Our study focuses on Niger and uses household survey data to assess how various factors influence our understanding of the risk of falling into poverty (i.e., household vulnerability) due to adverse weather events. We found that the methods we use to measure weather conditions, the geographic scale at which we measure them, and the household information we include can all significantly alter our estimates of how many households are at risk of becoming poor. For example, different methods for measuring weather impacts can change estimates of household vulnerability by about 3 percentage points, affecting around 600,000 households. The geographic level (administrative unit level or within a 20 km buffer around an enumeration area) at which we assess weather conditions can shift our estimates by 8 percentage points, which is equivalent to 1.6 million households. Additionally, considering different household characteristics can change our estimates by 10 percentage points, impacting around 2 million households. Our findings are crucial for policymakers who aim to better understand and address the effects of weather on vulnerable communities. Chapter 3 evaluates student learning outcomes from participation in the Commodity Investing by Students program, a student-managed investment fund within the Department of Agricultural and Applied Economics at Virginia Tech. Our study focuses on students from the 2022/23 and 2023/24 academic years, assessing both their technical knowledge and practical skills gained during a year-long involvement in the program. To measure changes in technical skills, we administered knowledge-testing quizzes before and after the training class. Practical skills, such as commodity market analysis, critical thinking, informed decision-making, and insightful interpretation of market analysis results, we evaluated through trading projects submitted during and at the end of the training class. We grade these student submissions using a specific practical skill evaluation rubric. We find significant improvements in student learning outcomes. On average, students demonstrated a 28% increase in disciplinary knowledge and a 38% improvement in practical skills. Our findings highlight the effectiveness of hands-on learning in improving both technical knowledge and practical skills that are highly valued in today's job market.

This study investigates the efficacy of various machine learning (ML) and deep learning (DL) models for short-term load forecasting (STLF) in the Electric Reliability Council of Texas (ERCOT) grid. A dual comparative approach is employed, evaluating models based on temporal features alone as well as in combination with actual and forecasted weather variables. The research emphasizes region-specific forecasting by capturing heterogeneous load patterns for ERCOT's individual weather zones and aggregating them to predict total load. Model evaluation is conducted using accuracy and bias metrics, with particular attention to high-demand months and peak load hours. The findings reveal that Generalized Additive Models (GAM) consistently outperform other models, most importantly during summer months and peak load hours. / Master of Science / This study investigates the efficacy of various machine learning (ML) and deep learning (DL) models for short-term load forecasting (STLF) in the Electric Reliability Council of Texas (ERCOT) grid. A dual comparative approach is employed, evaluating models based on temporal features alone as well as in combination with actual and forecasted weather variables. The research emphasizes region-specific forecasting by capturing heterogeneous load patterns for ERCOT's individual weather zones and aggregating them to predict total load. Model evaluation is conducted using accuracy and bias metrics, with particular attention to high-demand months and peak load hours. The findings reveal that Generalized Additive Models (GAM) consistently outperform other models, most importantly during summer months and peak load hours.

<p>This doctoral thesis is comprised of four papers that all relate to the subject of Time Series Analysis.</p><p>The first paper of the thesis considers point estimation in a nonnegative, hence non-Gaussian, AR(1) model. The parameter estimation is carried out using a type of extreme value estimators (EVEs). A novel estimation method based on the EVEs is presented. The theoretical analysis is complemented with Monte Carlo simulation results and the paper is concluded by an empirical example.</p><p>The second paper extends the model of the first paper of the thesis and considers semiparametric, robust point estimation in a nonlinear nonnegative autoregression. The nonnegative AR(1) model of the first paper is extended in three important ways: First, we allow the errors to be serially correlated. Second, we allow for heteroskedasticity of unknown form. Third, we allow for a multi-variable mapping of previous observations. Once more, the EVEs used for parameter estimation are shown to be strongly consistent under very general conditions. The theoretical analysis is complemented with extensive Monte Carlo simulation studies that illustrate the asymptotic theory and indicate reasonable small sample properties of the proposed estimators.</p><p>In the third paper we construct a simple nonnegative time series model for realized volatility, use the results of the second paper to estimate the proposed model on S&P 500 monthly realized volatilities, and then use the estimated model to make one-month-ahead forecasts. The out-of-sample performance of the proposed model is evaluated against a number of standard models. Various tests and accuracy measures are utilized to evaluate the forecast performances. It is found that forecasts from the nonnegative model perform exceptionally well under the mean absolute error and the mean absolute percentage error forecast accuracy measures.</p><p>In the fourth and last paper of the thesis we construct a multivariate extension of the popular Diebold-Mariano test. Under the null hypothesis of equal predictive accuracy of three or more forecasting models, the proposed test statistic has an asymptotic Chi-squared distribution. To explore whether the behavior of the test in moderate-sized samples can be improved, we also provide a finite-sample correction. A small-scale Monte Carlo study indicates that the proposed test has reasonable size properties in large samples and that it benefits noticeably from the finite-sample correction, even in quite large samples. The paper is concluded by an empirical example that illustrates the practical use of the two tests.</p>

This doctoral thesis is comprised of four papers that all relate to the subject of Time Series Analysis. The first paper of the thesis considers point estimation in a nonnegative, hence non-Gaussian, AR(1) model. The parameter estimation is carried out using a type of extreme value estimators (EVEs). A novel estimation method based on the EVEs is presented. The theoretical analysis is complemented with Monte Carlo simulation results and the paper is concluded by an empirical example. The second paper extends the model of the first paper of the thesis and considers semiparametric, robust point estimation in a nonlinear nonnegative autoregression. The nonnegative AR(1) model of the first paper is extended in three important ways: First, we allow the errors to be serially correlated. Second, we allow for heteroskedasticity of unknown form. Third, we allow for a multi-variable mapping of previous observations. Once more, the EVEs used for parameter estimation are shown to be strongly consistent under very general conditions. The theoretical analysis is complemented with extensive Monte Carlo simulation studies that illustrate the asymptotic theory and indicate reasonable small sample properties of the proposed estimators. In the third paper we construct a simple nonnegative time series model for realized volatility, use the results of the second paper to estimate the proposed model on S&amp;P 500 monthly realized volatilities, and then use the estimated model to make one-month-ahead forecasts. The out-of-sample performance of the proposed model is evaluated against a number of standard models. Various tests and accuracy measures are utilized to evaluate the forecast performances. It is found that forecasts from the nonnegative model perform exceptionally well under the mean absolute error and the mean absolute percentage error forecast accuracy measures. In the fourth and last paper of the thesis we construct a multivariate extension of the popular Diebold-Mariano test. Under the null hypothesis of equal predictive accuracy of three or more forecasting models, the proposed test statistic has an asymptotic Chi-squared distribution. To explore whether the behavior of the test in moderate-sized samples can be improved, we also provide a finite-sample correction. A small-scale Monte Carlo study indicates that the proposed test has reasonable size properties in large samples and that it benefits noticeably from the finite-sample correction, even in quite large samples. The paper is concluded by an empirical example that illustrates the practical use of the two tests.

In this thesis, we focus on a forecasting tool, Martingale Model of Forecast Evolution (MMFE), to model the evolution of forecasts in a production-distribution system. Additive form is performed to represent the evolution process. Variance-Covariance (VCV) matrix is defined to express the forecast updates. The selected demand pattern is stationary and it is normally distributed. It follows an Autoregressive Order-1 (AR(1)) model. Two forecasting procedures are selected to compare the MMFE with. These are MA (Moving average) and ES (Exponential smoothing) methods. A production-distribution model is constructed to represent a two-stage supply chain environment. The performance measures considered in the analyses are the total costs, fill rates and forecast accuracy observed in the operation of the production-distribution system. The goal is to demonstrate the importance of good forecasting in supply chain environments.

Rail transit investments require highest amount of investment costs of all modes and considering the high cost involved, it is particularly important that their performance justifies this high cost and that expectations from these investments are met. Therefore, in the world, it has become an important field of research to study the performances of rail systems in order to assess whether these expectations are met. In Turkey, there is a growing interest in constructing rail transit systems in the cities. However, there has been limited number of studies on the performance of these investments. There are researches on individual systems / yet, there has not been a comprehensive, systematic and comparative evaluation of the rail transit experience of Turkish cities. It is not clear with what expectations these systems are built or whether these expectations are met. There seems to be an urgent need to study these rail investments, with a particular focus on their planning, investment objectives and outcomes. This thesis analyzes the expectations from the rail transit systems in Turkey and answers the question whether these expectations are met. In order to understand the objectives under the planning and decision making processes in the implementation of Turkish rapid rail transport investments, a sample group was selected among the cities currently operating rail transit systems: &amp / #272 / stanbul, Ankara, &amp / #272 / zmir and Bursa. The study sets the objectives in planning and implementing rail transit systems drawn by the answers in the semi-structured interviews. It compares the expectations with the actual outcomes. As the primary indicators of performance, cost and ridership forecast and outcome data are also collected and considered in the comparison. It is found that the main success in all case study cities was the increase in public transport usage after the opening of the rail transit systems. On the other hand, systems performed rather poor in terms of other expectations, such as attaining ridership forecasts, being built within budget, creating an integrated public transport system, traffic reduction, air pollution reduction, improvement of city image, etc. Hence there is a gap between expectations and outcomes.

<p>Volvo in Arvika produces wheel loaders, and the production is based on forecasts. When a machine is ordered, the customer can choose what type of equipment he or she wants, and these equipments are also made forecasts on. This is made by giving each equipment an estimated procentual usage that shows how many of the machines that will use this option. Today two people are working with the forecasts, planer A in Eskilstuna and planer B in Arvika. Planer A makes a forecast based on the historical outcome and planer B then makes adjustments of this based on how many options that are ordered. Volvo in Arvika is having problems with the accuracy of the forecasts and because of this they have got too much in stock. But how big are the forecast deviations, what is the cause of it, in what or which places does the process lack? What can be made to make more accurate forecasts, how can you get a more affective process with less work made? To answer these questions we surveyed the process and analyzed it to find strong and week spots. We found that Planer A has a lack of information about how the forecast influence the stock in Arvika, that she gets pour feedback from production, that Planer B is the only one with knowledge about the forecast work in Arvika. We also made a benchmarking with the factory in Braås to see how they differ. Just like in Arvika there are two people working with the forecasts, but in Braås both of them are located close to the production and they share the options equal. They are also able to fill in for each other if someone would be sick.</p><p>To see how much the forecast differ from market demand, we have analyzed forecast data from nine different options for eight months. It turned out that the automatic calculated forecasts are a bit high and that planner B lower them.</p><p>Our conclusion is that the forecasts should be made only in Arvika, and not as it is today when half of it is made in Eskilstuna. There should also be documents and routines on how the work shall be done. This is to make it easier for people that will do the same job in the future.</p> / <p>Volvo Wheel Loaders (WLO) i Arvika tillverkar hjullastare och gör detta mot prognos. Till hjullastarna finns olika utrustningar som kunden kan välja mellan och även dessa gör Volvo prognoser på. Detta görs genom att de uppskattar hur många procent av maskinerna som kommer använda varje utrustning och lägger in det i ett program. Idag arbetar två personer med prognoserna, planerare A på huvudkontoret i Eskilstuna och planerare B på plats i Arvika. Planerare A gör först prognosen med avseende på historiskt utfall, sedan justerar planerare B dessa gentemot bl.a. orderingång. WLO har problem med träffsäkerheten i sina utrustningsprognoser och detta har medfört höga lagernivåer och därmed bundet kapital. Hur stora är prognosavvikelserna, vad är det som gör att prognoserna blir fel, på vilket eller vilka ställen i processen är det som bristerna uppstår? Vad kan de göra för att få bättre prognoser, hur kan man effektivisera processen så att det blir mindre arbete? För att svara på dessa frågor började vi med att kartlägga prognostiseringsprocessen och sedan analysera den för att få fram svagheter och styrkor. Här fann vi t ex att planerare A inte har någon kunskap om hur prognoserna påverkar lagret i Arvika och att hon får för dålig feedback från produktion, att planerare B är ensam kunnig om prognosarbetet vilket leder till problem när han är sjuk eller borta av andra skäl.</p><p>Sedan gjorde vi även en processjämförelse med Volvo Braås för att se hur de skiljer sig åt. I Braås är det två personer som tar fram prognoserna och de arbetar med hälften av utrustningsnumren var. Båda sitter nära produktion och är även väl insatta i varandras arbete om någon av dem skulle vara borta.</p><p>För att få fram hur prognoserna avviker från utfallet har vi gått igenom prognoshistorik för nio olika typer av utrustningar och sedan gjort beräkningar på det materialet. Det visade sig att prognoserna som automatiskt beräknas ofta ligger för högt och att planerare B sänker dessa.</p><p>Vi har kommit fram till att allt arbetet med prognoserna borde ske på plats i Arvika och inte som i nuläget när hälften görs i Eskilstuna. Man bör även införa rutiner på hur arbetet med prognoserna ska gå till och göra dokument på detta så att det är lättare för personer som ska ta över eller måste sätta sig in hur det fungerar.</p>

Volvo in Arvika produces wheel loaders, and the production is based on forecasts. When a machine is ordered, the customer can choose what type of equipment he or she wants, and these equipments are also made forecasts on. This is made by giving each equipment an estimated procentual usage that shows how many of the machines that will use this option. Today two people are working with the forecasts, planer A in Eskilstuna and planer B in Arvika. Planer A makes a forecast based on the historical outcome and planer B then makes adjustments of this based on how many options that are ordered. Volvo in Arvika is having problems with the accuracy of the forecasts and because of this they have got too much in stock. But how big are the forecast deviations, what is the cause of it, in what or which places does the process lack? What can be made to make more accurate forecasts, how can you get a more affective process with less work made? To answer these questions we surveyed the process and analyzed it to find strong and week spots. We found that Planer A has a lack of information about how the forecast influence the stock in Arvika, that she gets pour feedback from production, that Planer B is the only one with knowledge about the forecast work in Arvika. We also made a benchmarking with the factory in Braås to see how they differ. Just like in Arvika there are two people working with the forecasts, but in Braås both of them are located close to the production and they share the options equal. They are also able to fill in for each other if someone would be sick. To see how much the forecast differ from market demand, we have analyzed forecast data from nine different options for eight months. It turned out that the automatic calculated forecasts are a bit high and that planner B lower them. Our conclusion is that the forecasts should be made only in Arvika, and not as it is today when half of it is made in Eskilstuna. There should also be documents and routines on how the work shall be done. This is to make it easier for people that will do the same job in the future. / Volvo Wheel Loaders (WLO) i Arvika tillverkar hjullastare och gör detta mot prognos. Till hjullastarna finns olika utrustningar som kunden kan välja mellan och även dessa gör Volvo prognoser på. Detta görs genom att de uppskattar hur många procent av maskinerna som kommer använda varje utrustning och lägger in det i ett program. Idag arbetar två personer med prognoserna, planerare A på huvudkontoret i Eskilstuna och planerare B på plats i Arvika. Planerare A gör först prognosen med avseende på historiskt utfall, sedan justerar planerare B dessa gentemot bl.a. orderingång. WLO har problem med träffsäkerheten i sina utrustningsprognoser och detta har medfört höga lagernivåer och därmed bundet kapital. Hur stora är prognosavvikelserna, vad är det som gör att prognoserna blir fel, på vilket eller vilka ställen i processen är det som bristerna uppstår? Vad kan de göra för att få bättre prognoser, hur kan man effektivisera processen så att det blir mindre arbete? För att svara på dessa frågor började vi med att kartlägga prognostiseringsprocessen och sedan analysera den för att få fram svagheter och styrkor. Här fann vi t ex att planerare A inte har någon kunskap om hur prognoserna påverkar lagret i Arvika och att hon får för dålig feedback från produktion, att planerare B är ensam kunnig om prognosarbetet vilket leder till problem när han är sjuk eller borta av andra skäl. Sedan gjorde vi även en processjämförelse med Volvo Braås för att se hur de skiljer sig åt. I Braås är det två personer som tar fram prognoserna och de arbetar med hälften av utrustningsnumren var. Båda sitter nära produktion och är även väl insatta i varandras arbete om någon av dem skulle vara borta. För att få fram hur prognoserna avviker från utfallet har vi gått igenom prognoshistorik för nio olika typer av utrustningar och sedan gjort beräkningar på det materialet. Det visade sig att prognoserna som automatiskt beräknas ofta ligger för högt och att planerare B sänker dessa. Vi har kommit fram till att allt arbetet med prognoserna borde ske på plats i Arvika och inte som i nuläget när hälften görs i Eskilstuna. Man bör även införa rutiner på hur arbetet med prognoserna ska gå till och göra dokument på detta så att det är lättare för personer som ska ta över eller måste sätta sig in hur det fungerar.

Le contrat est un acte de prévision et, à ce titre, les prévisions des parties doivent être prises en compte et respectées. A côté des prévisions contractuelles, une notion existe mais est depuis longtemps méconnue : celle de prévisibilité contractuelle. Fondement de nombreux mécanismes bien connus en droit des contrats, comme le dommage prévisible, la force majeure ou encore l'obligation de modérer le dommage, la prévisibilité contractuelle donne naissance à un devoir contractuel : le devoir de prévision, dont la violation constitue une faute de prévision. Outil de moralisation de la relation contractuelle mais également de respect des attentes des parties et de la force obligatoire du contrat, le devoir de prévision a deux conséquences principales : il donne naissance à une obligation d'informer de ce qu'on a pu prévoir, tant lors de la formation que de l'exécution du contrat, et il impose aux parties de mettre en œuvre les mesures raisonnables nécessaires à la bonne exécution du contrat. / The contract is an act of forecasting and, as such, the forecasts of the parties must be taken into account and respected. Besides that, a notion exists but has long been ignored: the notion of contractual predictability. The foundations of many well-known mechanisms in contract law, such as foreseeable damage, force majeure or the mitigation of damages, contractual foreseeability gives rise to a real contractual duty : the duty of forecasting, sanctioned by a fault of forecast. Tools for moralizing the contractual relationship but also for respecting the expectations of the parties and the force of the contract, the duty of forecasting has two main consequences: it gives rise to an obligation to inform what has been foreseen, during training and the performance of the contract, and it requires the parties to implement the necessary reasonable measures for the proper performance of the contract.

nÃo hà / Neste ensaio investiga-se a perda de conteÃdo dos modelos de previsÃo autoregressivos, na medida em que se alarga o horizonte temporal no qual a variÃvel à estimada. O conteÃdo à medido pela reduÃÃo relativa do erro quadrado mÃdio que o modelo proporciona em comparaÃÃo ao processo simplificado de utilizar a mÃdia incondicional da sÃrie temporal. A variÃvel estudada à a arrecadaÃÃo mensal do Imposto sobre CirculaÃÃo de Mercadorias e ServiÃos (ICMS) proveniente do segmento de energia elÃtrica, no Estado do CearÃ, no perÃodo de janeiro de 1999 a setembro de 2010. Utiliza-se o mÃtodo e o modelo computacional formulados por Galbraith (2003), analisando-se a forecast content function, na qual o conteÃdo depende do nÃmero de perÃodos estimados. Os resultados confirmam que, para a sÃrie temporal explorada, quando se eleva o alcance da previsÃo o conteÃdo decai rapidamente, podendo atingir valor inferior a 10% quando o horizonte da previsÃo chega a 5 meses. Verificou-se, ademais, que o uso de sub-amostras via descarte de perÃodos mais antigos agrava a perda de conteÃdo. / In this essay we investigate the loss of content in autoregressive forecast models, as it is increased the horizon of time in which the variable is estimated. The content is measured as the proportionate reduction in medium squared error (MSE) that the model gives, comparing to the simple process by using the unconditional mean of time series. The variable is the monthly collection of ICMS from electric power sector, in Cearà state, in the period from January 1999 to September 2010. We use the method and computational model formulated by Galbraith (2003), analyzing the forecast content function, in which the content depends on the number of estimated periods. The results confirm that, when it increases the range of forecast the content decays quickly, reaching less than 10% when the forecast horizons reaches 5 months. It was found further that the use of subsamples by discarding oldest periods increases the loss of content.