µL

With globalization and knowledge economy, although distance between spaces and obstacles to social and economic activities are reduced, does variation of space not exist in regional space anymore? Has effect of regional space on economic development changed? Many economic activities, especially research and development, still rely on face-to-face communication, but geographical link is still beneficial to technology sharing and decrease in cost. It¡¦s observed that research and development that knowledge economy relies on still emphasizes geographical link, localization of learning networks, and face-to-face communication, so clustering of high-tech industry is still popular in the world. In order to solve the abovementioned questions, this research reviews and discusses the theory and research of clustering economy, and analyzes the relationship between clustering of high-tech industry and innovation on the basis of Taiwan¡¦s cities or counties with last twenty¡¦s survey of industrial, business, and service industries. This research discusses effect on clustering of technology industry and its innovative variants. In addition, it¡¦s hypothesized that clustering level of technology industry and innovation of industry are interfacing with each other, in other words, it¡¦s hypothesized that variants as Taiwan¡¦s technology industry and innovation are interfacing with each other instead of being independent. The relationship between clustering of Taiwan¡¦s technology industry and innovation is investigated by 2SLS. The result shows that clustering of technology industry and innovative production have positive correlation, that is to say, clustering of technology industry is beneficial, and it¡¦s also beneficial to innovation of industry, so clustering of industry speeds up. Results of this research show that technology industry on this island has notable MAR professional dynamic external economics, implying that the more professional clustered one region¡¦s technology industry is, the more beneficial to one region¡¦s professional growth of technology industry is. The more noticeable Porter regional competitive dynamic external economic effect on technology industry is, the more competitive a technology industry is, and the more prevalent one region¡¦s technology is. Technology industry has noticeable human resource dynamic external economics, showing that the better educational background of one region¡¦s population is, the more growth of one region¡¦s technology industry will be. Clustering industry of agglomeration economy is the primary reason of formation of cities or space centering of economic activities, and it motivates innovation of an industry and development of a city. This research suggests that government should take advantage of clustering knowledge capital of space to multiply effect of clustering and knowledge capital to do space management in terms of future urban space management.

On the one hand, time-honored practices of open outcry, bazaar and other flea markets, as well as Sotheby's and Christie's, the traditional auction houses. On the other hand, the Internet technology quickly and cheaply connects people anywhere in the world at anytime. A true love story then began and gave rise to online auctions, which are undoubtedly one of the greatest successes of e-commerce. Thanks to the Internet, individual consumers, average people, generally used to posted prices, discover an efficient mechanism of price setting. The weak popularity of auctions until the dawn of the third millennium (auctions were intended to B2B transactions, or to B2C transactions but in specific markets, such as the art market) certainly justifies the lack of interest of researchers in marketing for this mechanism. On the other hand, economists made auctions one of their favorite topics, as this mechanism constitutes a wonderful application of game theory, which has been playing a preponderant role in economics for twenty-five years. Entitled “Essays on Online Auction Design and Bidding Behavior”, this doctoral dissertation helps enrich the traditional economic approach of auctions by a behavioral and dynamic approach. The original part of the research is structured in three main stages, which go deeper and deeper in the study of bidders' behavior and auction dynamics. The whole research is based on online auctions organized by the French airline company Air France and by the auction site eBay. In the first stage of the research, on the basis of hypotheses coming from the auction theory, the impact of auction rules on bidders’ participation and seller's revenue is econometrically studied, using simultaneous equations models and the 2SLS method. Does a high opening bid have a negative effect on the number of bidders and a positive one on the auction revenue, as predicted by the auction theory? Do sequential auctions of similar items lead to similar auction price? These are examples of questions that are investigated. While the auction theory literature typically takes the number of bidders as exogenously given, we chose to consider it as endogenous. Even though an increasing number of researchers acknowledge the endogenous feature of auction participation, no empirical study really takes this feature into account. Hypotheses from the auction theory turn out to be confirmed, as far as the minimum bid and the number of bidders are concerned. Interesting results are found regarding the performance of sequential auctions of similar items, since auctions in a sequence turn out to yield increasing revenues. Furthermore, the impact of the starting bid turns out to be moderated by the effect of these sequential auctions. In the second stage of this research, a disaggregated perspective (at the bidder’s level) is adopted, since a typology of bidding behaviors based on bidders' choices is built. This part is thus aimed at investigating whether heterogeneity exists in bidder’s behavior in online auctions, contrary to the longstanding assumption of homogenous, rational and strategic bidders made by the auction theory. A clustering analysis is conducted, based on decisions that bidders have to make during an auction, whose keywords are certainly: when? How? How much? For example: do bidders submit a small or a large number of bids? Do they react quickly when they lose auction leadership? When do they make their bids: at the beginning of the auction, at the end of it, or during the whole auction? What increment do they use? Results, based on two different samples related to Air France and eBay auctions provide a description of different bidding behaviors. The analysis focusing on Air France auctions highlights five types of bidding behavior: jump bidders, rational bidders, active bidders, bottom fishers and pioneers. Slightly different results are obtained for eBay auctions, since six types of bidding behavior are highlighted: snipers, evaluators, unmasking bidders, bottom fishers, pioneers and late pioneers. A closer look is given to the group of bidders who turned out to win an auction, in order to determine whether these bidders are characterized by specific bidding decisions, and thus, by specific bidding behaviors. The third stage focuses on the dynamic bidding process of an English auction, by specifically studying the impact of signals – namely, promotional messages – sent during an online auction on the final auction price. It proposes to test a model of the genesis and the impact of these messages aimed at informing current bidders and potential bidders about the item or urging them to submit a bid. This impact is modeled through a disaggregated and dynamic model. It exploits the recent behavioral view that each bid submitted by a bidder in an English auction is a particular decision that may be influenced by signals sent by the auctioneer during the auction, that is, between the decisions that the bidder makes. This model simultaneously takes into account the following three factors: (i) the direct impact of marketing messages on the auction price when messages affect bidders’ valuations, (ii) the indirect impact of messages on this price when messages attract a new bidder to the auction, and (iii) the possibility that the auctioneer’s strategy for sending messages depends on past events in the auction and on the timing of past messages. In this model, we thus propose that messages influence final auction prices through a dual-path system. The model also reflects an important feature of auction messages, which is the real time interactivity between the auctioneer and auction participants. The results, obtained through Bayesian inference, support the proposed model. This doctoral dissertation globally helps better understand bidders’ behavior in view of the rules set by the seller, and estimate how this seller can maximize his/her revenues. The main original aspects of this research consist in considering the English auction as a dynamic process and in focusing on bidders’ behavior, these two elements being ignored by economists. From a managerial point of view, this research can help online auction designers to better design their auctions in order to maximize their revenue.

The rapid growth of gasoline consumption in the USA for the last decades brings much concern to scientists and politicians. Therefore many researchers investigated the influence of the main factors that have an impact on gasoline demand. In our study we tried to estimate gasoline demand in the USA, using national time series data for the period 1984-2010. Gasoline demand function considered in this paper includes price, income, fuel efficiency and gasoline consumption in previous year, as the main explanatory variables. The model is estimated using simultaneous equations and cointegration and error correction model (ECM). The results of both methods show a significant price and income effect on gasoline demand. The price is found inelastic and its impact on gasoline demand is very small, however when we correct for endogeneity of price variable, we obtain higher price elasticity. The results on income elasticities obtained from two methods are dubious, since the two methods gave us the different results. In whole, an income raise will lead to an increase of consumption, gasoline demand is inelastic with respect to income in the short-run, while in the long-run it is found to be elastic according to 2SLS method, while the results of cointegration method indicate that gasoline response to income changes is higher in the short-run than in the long-run. Lag of error term suggests that around 57% of adjustment between short-run and long-run occurs during the first year.

The primary aim of this study is to explore the connection between rural - urban migration and unemployment in Turkey and examine whether this internal migration has an effect on increasing the unemployment rates. By using Two Stage Least Squares (2SLS) panel data techniques and fully identifying these very concepts: migration from rural areas to urban areas, unemployment and internal migrations effects on the unemployment, an attention will be taken to the (negative) impact of internal migration on unemployment in Turkey.

nÃo hà / As vendas de castanha de caju do Cearà para o exterior representam cerca de 70% das exportaÃÃes do mercado de amÃndoas do Brasil. As amÃndoas sÃo exportadas principalmente para o mercado dos Estados Unidos da AmÃrica. A despeito disso, nota-se que o total da parcela das vendas do produto no Cearà que atendem o mercado externo representa cerca de 20% do que à consumido no resto do mundo. Dada a grande representatividade do mercado local nos cenÃrios nacional e internacional e a recente perda de participaÃÃo frente ao surgimento de fortes concorrentes, este trabalho busca estimar como os beneficiadores de castanha local reagem a determinados conjuntos de variÃveis. Assim, estimamos uma equaÃÃo de oferta para o setor beneficiador de castanhas de caju do CearÃ. AlÃm de preÃo e quantidade, foi coletado um conjunto de variÃveis de controle que quando inseridos na equaÃÃo estimada podem fazer com que persista alguma simultaneidade com as vendas cujo destino à o mercado exportador. Dessa forma, o mÃtodo dos MÃnimos Quadrados em Dois EstÃgios (MQ2E) foi escolhido para eliminar problema de identificaÃÃo e obter estimativas consistentes para a oferta, por corrigir possÃveis vieses de simultaneidade. Este trabalho evidencia que o preÃo nÃo apresentou significÃncia a 5%, possivelmente por ser determinado no mercado internacional e, portanto, foge ao domÃnio dos beneficiadores cearenses, os quais se comportam como tomadores de preÃo.Demonstra ainda que o cÃmbio, a energia, a crise econÃmica dos Estados Unidos da AmÃrica iniciada em 2008 e a renda disponÃvel americana apresentaram significÃncia a 1% e sinais esperados na curva de oferta. / The sales of cashew nut from Cearà to the world represents around 70% of the market exports of almonds from Brazil and around 20% of what is consumed in the world. They are mainly exported to the US market. Given the great representability of the local market in national and international scenarios and the recent loss of market share due to the emergence of strong competitors, this work seek to estimate how the local producers of cashew nut react to determined variables group. We estimated one offer equation for the producing sector using the two stages least square (2SLS). Besides price and quantity, a group of control variables were collected, and when inserted in the estimated equation they can make some simultaneity with the sales destined to the foreign market. Knowing this, the two stages least square (2SLS) was the method chosen to eliminate identification problems and get estimations to correct possible remaining bias of simultaneity. This work evidences that the price did not show significance of 5%, possibly to be determined by the international market, and so, it is beyond the domain of the CearÃ`s producers as this one acts as a price taker. It shows also that the coin rates, energy , the US crysis started in 2008 and the available US earnings represents 1% significance and signs awaited in the offer curve.

Syftet med uppsatsen är att presentera en översikt av studier där instrumentvariabler används för att estimera den kausala effekten av att ta ut pension på hälsa. Studier där två olika typer av instrument används inkluderas i forskningsöversikten: (1) studier där åldersgränser används som instrument, och (2) studier där reformer används för att skapa instrument. I samtliga studier används ”Two stage least squares” för att estimera effekten. Villkoret om att instrument behöver vara exkluderade ur den strukturella modellen för att den kausala effekten av att ta ut pension på hälsa ska kunna identifieras diskuteras, och tillvägagångssätt och vissa resultat från tester som används i studierna för att uppfylla villkoret presenteras. Även villkoret om att det krävs tillräckligt starka instrument för att identifiera den kausala effekten diskuteras, och tillvägagångssätt och tester från studierna presenteras. Huvudresultaten från studierna om effekten av att ta ut pension på hälsa presenteras och diskuteras. Resultaten varierar mellan studierna, vilket antas bero på att måtten som används för att mäta hälsa varierar, att populationerna som studeras skiljer sig från varandra, och att metoderna skiljer sig åt mellan studierna.

Although a number of studies have been conducted to understand the impact of ICT on a country’s economic output; a complete account for ICT’s influence on real GDP per capita has yet to be visualized. This study aims to investigate this topic, by exploring the hypothesis whether ICT has an impact (causal relationship) on real GDP per worker in 41 European countries from 1996-2010. Confidence in the judicial system is taken as an instrument to solve for endogeneity problems. ICT is defined as mobile, broadband and internet penetration per 100 people. We find that confidence in the justice system is a strong instrument for ICT. Therefore, using IV 2sls regression, it is observed that ICT does have a strong positive correlation and causal relationship with GDP per worker, with an elasticity of 0.4 between the variables. In light of the results, the author recommends that economic policies must be tailored to support ICT development in countries.

This thesis examines the role of grandchildren's existence in the retirement timing decision- making process of grandparents. Previous literature has focused mostly on other aspects of retirement and potential causes that can affect its timing. Using the Two-Stage least squares estimation on the SHARE dataset, representing 17 European countries and Israel, we estimate the desired effect with respect to various data limitations (age groups, gender, child existence). Residential proximity is used as the instrument for estimation. Having at least one grandchild yields a statistically significant result that increases on average the likelihood of retirement by 19% when compared to a non-grandparent while holding other factors constant. As a secondary outcome, the estimated effect of an additional child on retirement likelihood is negative. JEL Classification C36, C51, J26 Keywords grandchild, retirement, Instrumental variable, SHARE, IV, wide-ranging data, 2SLS Title The impact of grandchildren on retirement timing: evidence from SHARE data

Economic inequality and its effects on income growth is a topic that has been researched extensively. Previous studies with the aim to determine the impact of income inequality on economic growth have functioned more as pointers than solid facts. This study contributes to the continued discussion within the growth economy by exploring the relationship between the growth rate of average income and income inequality using panel data of Swedish municipalities from 2000-2020. Produced independently, this study uses the Gini coefficient as a tool to measure the income inequality among all 290 municipalities in Sweden. The result from the panel estimation shows no clear effect on the impact of income inequalities on income growth. However, when endogeneity is considered a 2SLS regression with the age structure as an instrument variable is used, the estimation shows that income inequality at local level has a positive association to growth of average income. Coherent with earlier studies of income inequality among Swedish municipalities, the result implies that as income inequality increases between income groups, average municipality experiences increase in growth of average income. Furthermore, this disposition compares the estimated results to a system GMM estimation.