This study illustrates the use of ACNielsen Homescan Panel (HSD) in three separate demand analyses of dairy products: (1) the effect of using cross-sectional data in a New Empirical Industrial Organization (NEIO) study of ice cream firm mergers in San Antonio; (2) the estimation of hedonic price models for fluid milk by quart, halfgallon and gallon container sizes; (3) the estimation of a demand system including white milk, flavored milk, carbonated soft drinks, bottled water, and fruit juice by various container sizes. In the NEIO study a standard LA/AIDS demand system was used to estimate elasticities evaluating seven simulated mergers of ice cream manufactures in San Antonio in 1999. Unlike previously published NEIO work, it is the first to use crosssectional data to address the issue associated with inventory effects. Using the method developed by Capps, Church and Love, none of the simulated price effects associated with the mergers was statistically different from zero at the 5% confidence level. In 1995 Nerlove proposed a quantity-dependent hedonic model as a viable alternative to the conventional price-dependent hedonic model as a means to ascertain consumer willingness to pay for the characteristics of a given good. We revisited Nerlove’s work validating his model using transactional data indigenous to the HSD. Hedonic models, both price-dependent and quantity-dependent, were estimated for the characteristics of fat content, container type, and brand designation for the container sizes of gallon, half- gallon, and quart. A rigorous explanation of the interpretation between the estimates derived from the two hedonic models was discussed. Using the Almost Ideal Demand System (AIDS), a matrix of own-price, crossprice, and expenditure elasticities was estimated involving various container sizes of white milk, flavored milk, carbonated soft drinks, bottled water, and fruit juices, using a cross-section of the 1999 HSD. We described price imputations and the handling of censored observations to develop the respective elasticities. These elasticities provided information about intra-product relationships (same product but different sizes), intrasize relationships (different products same container size), and inter-product relationships (different products and different sizes). This container size issue is unique in the extant literature associated with non-alcoholic beverage industry.

There are many models used to estimate demand elasticities. We present a complex review of these studies in our thesis. Our empirical goal is to compare LES, Translog and QUAIDS demand systems according to their performance. In parallel, we estimate the elasticities of meat demand in the Czech Republic for the period 2010 - 2015 using the data of the household budget survey. Comparing the systems by the Akaike and Schwarz criterion, LES demonstrates the best fit for this kind of data. The average of price elasticity for different kinds of meat in the examined period is -0.99, income elasticity then equals to 1.12. These results can have important implications for tax policy, or for commercial use. JEL Classification F12, F21, F23, H25, H71, H87 Keywords Demand, comparison, LES, Translog, QUAIDS, meat Author's e-mail 55606678@fsv.cuni.cz Supervisor's e-mail milan.scasny@czp.cuni.cz

This dissertation has four interrelated studies, namely (1) the characterization of milk purchase choices which included the purchase of organic milk, both organic and conventional milk and conventional milk only; (2) the estimation of a single-equation household demand function for organic and conventional milk; (3) the assessment of binary choice models for organic milk using the Brier Probability score and Yates partition, and (4) the estimation of demand systems that addresses the censoring issue through the use of econometric techniques. In the first paper, the study utilized the estimation of both multinomial logit and probit models in examining a set of causal socio-demographic variables in explaining the purchase of three outcome milk choices namely organic milk, organic and conventional milk and conventional milk only. These crucial variables include income, household size, education level and employment of household head, race, ethnicity and region. Using the 2004 Nielsen Homescan Panel, the second study used the Heckman two-step procedure in calculating the own-price, cross-price, and income elasticities by estimating the demand relationships for both organic and conventional milk. Results indicated that organic and conventional milk are substitutes. Also, an asymmetric pattern existed with regard to the substitution patterns of the respective milk types. Likewise, the third study showed that predictive outcomes from binary choice models associated with organic milk can be enhanced with the use of the Brier score method. In this case, specifications omitting important socio-demographic variables reduced the variability of predicted probabilities and therefore limited its sorting ability. The last study estimated both censored Almost Ideal Demand Systems (AIDS) and Quadratic Almost Ideal Demand System (QUAIDS) specifications in modeling nonalcoholic beverages. In this research, five estimation techniques were used which included the usage of Iterated Seemingly Unrelated Regression (ITSUR), two stage methods such as the Heien and Wessells (1990) and the Shonkwiler and Yen (1999) approaches, Generalized Maximum Entropy and the Dong, Gould and Kaiser (2004a) methods. The findings of the study showed that at various censoring techniques, price elasticity estimates were observed to have greater variability in highly censored nonalcoholic beverage items such as tea, coffee and bottled water.

The U.S. sheep industry has been declining in size for many years. Many factors have contributed to the decline of the sheep industry including declining consumption of lamb and mutton, the growth in manmade fiber use, scarcity of labor, and predator losses. In an effort to slow the rate of decline in the U.S. sheep industry, the U.S. Congress passed the Wool Act of 1954. In 1993, Congress passed a three-year phase out of the Wool Act incentive payments with the last payments occurring in 1996. The 2002 Farm Bill included a marketing loan program for wool. The loan rates are set to $0.40 per pound for un-graded wool, $1.00 per pound for graded wool. In recent years exchange rate changes have had a large impact on the industry affecting lamb and wool trade. The U.S. is the second largest producer of mohair and Texas accounts for over 85 percent of the U.S. mohair production. Mohair also received incentive payments through the Wool Act. Mohair payments were also phased out along with the wool incentive payments. Moreover, the 2002 Farm Bill reinstated support for the industry by implementing a loan program with loan rates of $4.20 per pound of mohair. This analysis uses capital stock inventory accounting methodology to model the supply side of the sheep industry. Demand is incorporated using traditional single equations and complete demand system estimation methods. OLS, 2SLS, and 3SLS models are developed and tested for the single equations estimation methods. The OLS model is used to model the impacts of three different levels of loan rates for wool. Also, an OLS mohair model is developed and used to examine the impacts of three different levels of loan rates for mohair. Results indicate that the sheep industry will continue to decline even with the marketing loan program for wool in the 2002 Farm Bill. However, a higher loan rate for wool would reduce the decline rate of the industry. The Angora goat industry will continue to decline in size, but with a higher loan rate for mohair, the number of goats clipped would increase.

This study analyzes the economic and demographic drivers of household demand for at-home consumption of nonalcoholic beverages in 1999. Drivers of available intake of calories, calcium, vitamin C, and caffeine associated with the purchase of nonalcoholic beverages also are analyzed. The 1999 ACNielsen HomeScan Panel, purchased by the U. S. Department of Agriculture, Economic Research Service, is the source of the data for this project. Many different classifications of beverages were analyzed including milk(whole, reduced fat, flavored, and non-flavored), regular and low-calorie carbonated soft drinks, powdered soft drinks, isotonics(sports drinks), juices(orange, apple, vegetable, and other juices), fruit drinks, bottled water, coffee(regular and decaffeinated), and tea(regular and decaffeinated). Probit models were used to find demographic drivers that affect the choice to purchase a nonalcoholic beverage. Heckman sample selection models and cross tabulations were used to find demographic patterns pertaining to the amount of purchase of the nonalcoholic beverages. The nutrient analysis indicated that individuals receive 211 calories, 217 mg of calcium, 45 mg of vitamin C, and 95 mg of caffeine per day from all nonalcoholic beverages. A critical finding for the nutrient analysis was that persons within households below 130% of poverty were receiving more calories and caffeine from nonalcoholic beverages compared to persons within households above 130% of poverty. Likewise, persons in households below 130% of poverty were receiving less calcium and vitamin C from nonalcoholic beverages compared to persons in households above 130% of poverty. Price and cross-price elasticities were examined using the LA/AIDS model. Methodological concerns of data frequency, beverage aggregations, and censoring techniques were explored and discussed. Own-price and cross-price elasticities for the beverages were uncovered. Price elasticities by selected demographic groups also were investigated. Results indicated that price elasticities varied by demographics, specifically for race, region, and presence of children within the household. The information uncovered in this dissertation helps to update consumer demand knowledge and nutritional intake understanding in relation to nonalcoholic beverages. The information can be used as a guide for marketing strategists for targeting and promotion as well as for policy makers looking to improve nutritional intake received from nonalcoholic beverages.

The estimation of demand equations provides the earliest example of the use of statistical and econometric techniques on economic data. It is possible to identify two distinct approaches to the estimation of demand equations. The first and original approach concentrated on the demand for particular goods by paying attention to any special characteristics of the single market involved. The second approach involved simultaneous estimation of complete systems containing the demand equations for every commodity group purchased by consumers. The estimation of a complete system of demand equations in principle enables us to obtain better estimates of each equation in the system than the first approach because of interaction in the demand behavior of different commodities. This study is directed towards the estimation of demand systems for agricultural products in the OECD countries. Three representatives demand systems with their extensions, namely the Rotterdam Model, An Almost Ideal Demand System (AIDS), and CBS model are used. These models are estimated by Seemingly Unrelated Regression (SUR) method. The procedures to estimate demand systems suggest significant empirical regularities for agricultural products in the OECD countries. The main contribution of this study is its procedure for model selection. This procedure implies the superiority of AIDS and CBS models over the Rotterdam model.

Cette thèse étudie les conditions de taxation indirecte optimale qui permettent d'améliorer l'adéquation de la ration alimentaire des ménages français aux recommandations nutritionnelles. L'approche adoptée exige d'estimer un système complet de demande alimentaire afin de mettre à jour les élasticités-prix. Les comportements de consommation alimentaire des ménages sont décrits par une forme fonctionnelle AI (Deaton et Muellbauer, 1980). Deux spécifications quadratique et dynamique autorisant, respectivement, de possibles non-linéarités des courbes d'Engel et la persistance d'habitudes de consommation sont testées. L'estimation se fait par les moindres carrés linéaires itérés (Blundell et Robin, 1999) sur données de pseudo-panel. Une base de données s'étalant sur 156 périodes de 4 semaines, de 1996 à 2007, a été constituée par l'aggrégation en 8 cohortes de ménages et le regroupement de coupes transversales issues de TNS Worldpanel. Dans la lignée des mécanismes théoriques de taxation optimale initiée par Ramsey (1927), un application à un objectif de politique nutritionnelle permet de dériver les variations de prix, ou taux de taxe, optimales. Chacune apparait comme une fonction des élasticités-prix, directes et croisées, de la demande et d'indicateurs d'adéquation aux recommandations. L'impact des variations de prix sur l'adéquation aux recommandations nutritionnelles du panier alimentaire, ainsi les taux de taxation de produits optimaux sont estimés. L'incidence de ces scénarios de politique sur le bien-être des ménages et les inégalités nutritionnelles est abordée dans un dernier chapitre. / This thesis aims at simulating optimal prices satisfying public health recommendations in terms of nutrient adequacy. This implies to estimate a complete food demand system in order to compute price elasticities. Food consumption behaviors are described by an AI functional form (Deaton and Muellbauer, 1980). The demand system is estimated using the Iterated Least Square Estimator developed by Blundell and Robin (1999). Augmented specifications to control for, respectively, non-linearities in expenditure patterns and habit persistence are tested. We use French household expenditure data drawn from TNS Worldpanel covering 156 periods of 4 weeks from 1996 to 2007. Given the nature of our data, households are split into 8 cohorts. Issues afferent to pseudo-panel data are presented. Nutrient adequacy is defined using three nutrient only-based indicators: the MAR (Mean Adequacy Ratio), the LIM (Score des Composés à Limiter) and the SAIN (Score d'Adéquation Individuel aux recommandations Nutritionnelles). Optimal prices are derived following Ramsey's approach to optimal taxation; Maximizing social welfare under nutritional constraints results in optimal price variations or tax rates, each defined as a function of all direct and cross price elasticities and the above mentioned indicator for all food groups. Further developments allow to estimate the impacts of price variations on nutrient adequacy. Incidence on social welfare and nutritional inequalities is assessed in a last chapter.

Income and price affect chickpea trade expenditure and consumption expenditure share respectively. An empirical model was estimated to examine the trade effect through the expenditure channel using Almost Ideal Demand System and thus considering non-homotheticity in preferences. The results of the analysis indicated that global chickpea trade has increased from 100000 metric tons in 1988 to about 2.5 million metric tons in 2015. Between the same period consumption and production of chickpea had an increasing trend. USA and Canada had become part of the top 10 chickpea producers by 2015 signifying the increasing demand of chickpea in western countries. Factors that affected relative chickpea trade to importers income were relative market size of the exporter, bilateral distance and contiguous borders. Also, a percentage increase in the adjusted mean income of chickpea consuming country will lead to 94% decrease in the consumption of chickpea when country pair effects are considered.

Customer Relationship Management has been an integral part of the enterprise since two decades. Today, enterprises that focus on customer satisfaction need to manage their relationships with their customers effectively. This demand has allowed software vendors to create CRM solutions. The technology and broadband advancement allowed the CRM vendors to enhance their product portfolio by developing web-based CRM systems, in addition to their CRM on-premise solutions. These vendors adopted the business model in which CRM on-demand systems are provided via monthly-subscription fees, decreasing the total cost of ownership massively for enterprises in need of these systems. This business model is especially attractive for Small-To-Medium Enterprises who are searching for cost-efficient CRM systems. Currently, CRM on-demand market is quite saturated with more than 40 vendors providing similar solutions. Furthermore, CRM on-demand is delivered via Software-as-a-service method, which is a relatively new technology with unique benefits along with drawbacks. Thus, it’s of vital importance for managers in SMEs to make the right decision while evaluating the CRM on-demand option and systems. This research is meant to address this issue by building a proposed framework for CRM on-demand system evaluation. The inductive research uses qualitative and quantitative approaches for data collection and analysis. The evaluation criteria for CRM on-demand systems at a functional and general level were proposed. The general criteria were refined via collecting data from CRM on-demand experts and users in SMEs by structured questionnaires. Combining these criteria created the proposed framework which was applied to evaluate two major CRM on-demand systems in the market. The results indicate that CRM on-demand systems cover the basic functionalities of CRM including sales, marketing and service modules and offer enhanced functionality such as mobile CRM, social CRM and customizations. The research also revealed drawbacks of CRM on-demand systems such as disintegration with legacy applications, limited language support, limited country availability and technology maturity which needs to be addressed in the future. This research provides valuable insight for managers in SMEs when selecting CRM on-demand systems for their companies. Furthermore, the academicians interested in CRM and cloud computing could improve this initial proposed framework and adapt it further to different cases.

Die Anwendung eines Mikrosimulationsmodells als Instrument der wirtschaftswissenschaftlichen Forschung erfreut sich großer Beliebtheit. Die Verwendung dieses Modelltyps hat vor allem zwei Gründe. Zum einen sind die in den letzten Jahren extrem gesunkenen Kosten der Informationsverarbeitung ursächlich für die mittlerweile häufige Verwendung. Zum anderen besitzt diese Analyseform Eigenschaften, die kein anderes Instrument aufweist. Insbesondere für die Abschätzung der Verteilungswirkungen von Steuer- und Transferreformen ist die Mikrosimulation eine nahezu ideale Methode. Durch die Einbindung eines Nachfragesystems kann sowohl das Anwendungsgebiet eines Mikrosimulationsmodells erweitert als auch dessen potentielle Leistungsfähigkeit gesteigert werden. Die Arbeit hat im Wesentlichen die Beantwortung der Fragestellungen aus den folgenden zwei Fragenkomplexen zum Ziel: - Wann ist die Einbindung eines Nachfragesystems in Mikrosimulationsmodelle sinnvoll und wie kann ein solches System in das eigentliche Mikrosimulationsmodell integriert werden? - Welche Nachfragesysteme können aus den in Deutschland in Querschnittsform vorhandenen Mikrodatensätzen geschätzt werden und welche theoretischen Eigenschaften besitzen diese Systeme bzw. deren in einem zweistufigen Budgetierungsprozess einsetzbare Kombinationen? / The application of a microsimulation model as an instrument for economic research enjoys great popularity. The use of this type of model primarily has two reasons. In the first place the decreasing costs of data processing led to the frequent usage. Furthermore this type of analysis has unique characteristics compared to other instruments. Particularly microsimulation is an almost ideal method for the estimation of the impact of tax and allocative transfer reforms. With the integration of a demand system the field of application of a microsimulation model can be enlarged as well as its potential capacity increased. This paper has the aim to answer basically two batteries of questions: - In what case the integration of a demand system in a microsimulation model is reasonable and how can such a system be integrated in the model? - Which demand systems can be estimated from the microdata available in Germany and which characteristic properties do such systems offer, especially the combinations applicable in a two-stage budgeting environment?