Master of Science / Department of Statistics / James W. Neill

Paper [I] models a game, where two temporary work agencies (TWAs) compete to fill a vacancy at a client firm (CF). They simultaneously choose how much effort to expend, based on their expectation of how good their opponent’s best candidate will be. I then show that this will make the TWAs overconfident, as the rational way of judging your own probability of winning is not looking at the opponents expected best, but comparing how much effort your opponent will expend. Paper [II] examines the misaligned incentives in the temporary work agency sector, where we first look at pure recruiting contracts, that either require payment on delivery, or payment on some specified point in time. We then look at the incentives of recruit-and-rent contracts, where the worker is leased to the client firm. We assume that the better the worker, the higher the probability that the client firm is going to want to hire him/her. If that happens then the TWA will no longer get revenues from said worker, incentivizing the TWA to not always deliver the first match it finds, if it is too good. Lastly we look at how competition can dampen this perverse incentive. Paper [III] models the waiting behavior that can occur if a TWA is contracted to find a worker for a specific time far in the future; the TWA will postpone effort. This behavior is modeled for two types of TWAs; one that is rational and plans ahead, and another that does not plan ahead at all, but instead only looks at the immediate future. I find that the one that only looks at the immediate future starts exerting effort earlier than the planner. After looking at optimal contracts under perfect monitoring and hidden action I provide two extensions. I first show that for the principal to want to delegate search to a rational TWA, the agent has to be better than the CF, by some factor, as it has to make up in efficiency what the principal loses in moral hazard, when the agent waits longer than the principal would like it to. Lastly I prove that it is profit maximizing for the principal to contract one agent and give it a deadline earlier than when the principal would need the worker, and then replace that agent with a competitor if the first one has not succeeded by that earlier deadline. Paper [IV] estimates at the effect of family experience on relative transition probability into the temporary work agency sector. Using register data for all of Sweden we run a bias-reduced logistic regression, where we include various factors that affect the probability of young adults (aged 18-34) entering the sector. This paper ties in to the literature on occupational inheritance, as well as the literature on changing social norms. We find that having had a parent, sibling or partner in the TWA sector increases your probability of entering.

No description available.

This research consists of two essays. The first essay entitled” Stock Return Forecasting with Sum-of-the-Parts Methodology: Evidence from Around the World”, examines forecasting ability of stock returns by employing the sum-of-the-parts (SOP) modeling technique introduced by Ferreira and Santa-Clara (2011).This approach decomposes return into three components of growth in price-earnings ratio, earnings growth, and dividend-price ratio. Each component is forecasted separately and fitted values are used in forecast model to predict stock return. We conduct a series of one-step ahead recursive forecasts for a wide range of developed and emerging markets over the period February 1995 through November 2014. Decomposed return components are forecasted separately using a list of financial variables and the fitted values from the best estimators are used according to out-of-sample performance. Our findings show that the SOP method with financial variables outperforms the historical sample mean for the majority of countries. Second essay entitled,” Equity Premium Predictability under Regime Shifts: International Evidence”, utilizes the modified version of the dividend-price ratio that alleviates some econometric concerns in the literature regarding the non-stationary and persistent predictor when forecasting international equity premium across different regimes. We employ Markov switching technique to address the issue of non-linearity between the equity premium and the predictor. The results show different patterns of equity premium predictability over the regimes across countries by the modified ratio as predictor. In addition, transition probability analysis show the adverse effect of financial crisis on regime transition probabilities by increasing the probability of switching between regimes post-crisis 2007 implying higher risk perceived by investors as a result of uncertainty inherent in regime transitions.

Timely identification of undesirable pavement crack conditions has been a major task in pavement management. Up to date, myriads of pavement performance models have been developed for forecasting pavement crack condition with the traditional preferred techniques being the use of regression relationships developed from laboratory and/or field statistical data. However, it becomes difficult for regression techniques to predict the crack performance accurately and robustly in the presence of a variety of tributary factors, high nonlinearity, and uncertainty. With the advancement of modeling techniques, two innovative breeds of models, Artificial Neural Networks and Markov Chains, have drawn increasing attention from researchers for modeling complex phenomena like the pavement crack performance. In this study, two distinct models, a recurrent Markov chain, and an Artificial Neural Network (ANN), were developed for modeling the performance of pavement crack condition with time. A logistic model was used to establish a dynamic relationship between transition probabilities associated with the pavement crack condition and the applicable tributary variables. The logistic model was then used conveniently to construct a recurrent Markov chain for use in predicting the crack performance of asphalt pavements in Florida. Florida pavement condition survey database were utilized to perform a case study of the proposed methodologies. For comparison purpose, a currently popular static Markov chain was also developed based on a homogeneous transition probability matrix that was derived from the crack index statistics of Florida pavement survey database. To evaluate the model performance, two comparisons were made; (1) between the recurrent Markov chain and the static Markov chain; and (2) between the recurrent Markov chain and the ANN. It is shown that the recurrent Markov chain outperforms both the static Markov chain and the ANN in terms of one-year forecasting accuracy. Therefore, with high uncertainty typically experienced in the pavement condition deterioration process, the probabilistic dynamic modeling approach as embodied in the recurrent Markov chain provides a more appropriate and applicable methodology for modeling the pavement deterioration process with respect to cracks.

A driving cycle is a velocity profile over time. Driving cycles can be used for environmental classification of cars and to evaluate vehicle performance. The benefit by using stochastic driving cycles instead of predefined driving cycles, i.e. the New European Driving Cycle, is for instance that the risk of cycle beating is reduced. Different methods to generate stochastic driving cycles based on real-world data have been used around the world, but the representativeness of the generated driving cycles has been difficult to ensure. The possibility to generate stochastic driving cycles that captures specific features from a set of real-world driving cycles is studied. Data from more than 500 real-world trips has been processed and categorized. The driving cycles are merged into several transition probability matrices (TPMs), where each element corresponds to a specific state defined by its velocity and acceleration. The TPMs are used with Markov chain theory to generate stochastic driving cycles. The driving cycles are validated using percentile limits on a set of characteristic variables, that are obtained from statistical analysis of real-world driving cycles. The distribution of the generated driving cycles is investigated and compared to real-world driving cycles distribution. The generated driving cycles proves to represent the original set of real-world driving cycles in terms of key variables determined through statistical analysis. Four different methods are used to determine which statistical variables that describes the features of the provided driving cycles. Two of the methods uses regression analysis. Hierarchical clustering of statistical variables is proposed as a third alternative, and the last method combines the cluster analysis with the regression analysis. The entire process is automated and a graphical user interface is developed in Matlab to facilitate the use of the software. / En körcykel är en beskriving av hur hastigheten för ett fordon ändras under en körning. Körcykler används bland annat till att miljöklassa bilar och för att utvärdera fordonsprestanda. Olika metoder för att generera stokastiska körcykler baserade på verklig data har använts runt om i världen, men det har varit svårt att efterlikna naturliga körcykler. Möjligheten att generera stokastiska körcykler som representerar en uppsättning naturliga körcykler studeras. Data från över 500 körcykler bearbetas och kategoriseras. Dessa används för att skapa överergångsmatriser där varje element motsvarar ett visst tillstånd, med hastighet och acceleration som tillståndsvariabler. Matrisen tillsammans med teorin om Markovkedjor används för att generera stokastiska körcykler. De genererade körcyklerna valideras med hjälp percentilgränser för ett antal karaktäristiska variabler som beräknats för de naturliga körcyklerna. Hastighets- och accelerationsfördelningen hos de genererade körcyklerna studeras och jämförs med de naturliga körcyklerna för att säkerställa att de är representativa. Statistiska egenskaper jämfördes och de genererade körcyklerna visade sig likna den ursprungliga uppsättningen körcykler. Fyra olika metoder används för att bestämma vilka statistiska variabler som beskriver de naturliga körcyklerna. Två av metoderna använder regressionsanalys. Hierarkisk klustring av statistiska variabler föreslås som ett tredje alternativ. Den sista metoden kombinerar klusteranalysen med regressionsanalysen. Hela processen är automatiserad och ett grafiskt användargränssnitt har utvecklats i Matlab för att underlätta användningen av programmet.

Performance-based maintenance contracts differ significantly from material and method-based contracts that have been traditionally used to maintain roads. Road agencies around the world have moved towards a performance-based contract approach because it offers several advantages like cost saving, better budgeting certainty, better customer satisfaction with better road services and conditions. Payments for the maintenance of road are explicitly linked to the contractor successfully meeting certain clearly defined minimum performance indicators in these contracts. Quantitative evaluation of the cost of performance-based contracts has several difficulties due to the complexity of the pavement deterioration process. Based on a probabilistic analysis of failures of achieving multiple performance criteria over the length of the contract period, an effort has been made to develop a model that is capable of estimating the cost of these performance-based contracts. One of the essential functions of such model is to predict performance of the pavement as accurately as possible. Prediction of future degradation of pavement is done using Markov Chain Process, which requires estimating transition probabilities from previous deterioration rate for similar pavements. Transition probabilities were derived using historical pavement condition rating data, both for predicting pavement deterioration when there is no maintenance, and for predicting pavement improvement when maintenance activities are performed. A methodological framework has been developed to estimate the cost of maintaining road based on multiple performance criteria such as crack, rut and, roughness. The application of the developed model has been demonstrated via a real case study of Miami Dade Expressways (MDX) using pavement condition rating data from Florida Department of Transportation (FDOT) for a typical performance-based asphalt pavement maintenance contract. Results indicated that the pavement performance model developed could predict the pavement deterioration quite accurately. Sensitivity analysis performed shows that the model is very responsive to even slight changes in pavement deterioration rate and performance constraints. It is expected that the use of this model will assist the highway agencies and contractors in arriving at a fair contract value for executing long term performance-based pavement maintenance works.

No description available.

No description available.

In this study, we consider high frequency transaction data of NYSE, and apply statistical methods to characterize each trade into two classes, influential and ordinary liquidity trades. First, a median based approach is used to establish a high R-square price-volume model for high frequency data. Next, transactions are classified into four states based on the trade price, trade volume, quotes, and quoted depth. Volume weighted transition probability of the four states are investigated and shown to be distinct for informed trades and ordinary liquidity trades. Furthermore, four market reaction factors are introduced and studied. Logistic regression models of the influential trades are established based on the four factors and odds ratios are used to select the cutoff points.