要旨：In this paper, we revisit a wide variety of existing economic geography models in a many-region setup, and investigate the spatial scale of agglomeration and dispersion intrinsic to each model. By characterizing spontaneous agglomerations specific to each model in a unified analytical framework, we show that these models reduce to two canonical classes: one with global dispersion force and the other with local dispersion force. Their formal distinction is that the global dispersion force is dependent, while the local one is independent of the distance structure of the model. These classes exhibit two stark differences. First, on the response to transport costs, local dispersion is triggered by lower costs, while global one by higher costs. Second, on the agglomeration pattern, multiple agglomerations emerge and are spread over the regions in the former, while the agglomeration always takes the form of a unimodal regional distribution of mobile agents in the latter. The knowledge of these variations in agglomeration behavior is crucial for empirical analyses of agglomerations. The prevailing reduced-form regressions (e.g. Redding and Turner, 2015, §20.4) typically assume a monotonic relation between interregional transport accessibility and the size of agglomeration in a given region. But, it is not generally the case when agglomerations are endogenous. The well-accepted structural approaches (e.g., Redding and Rossi-Hansberg, 2017, §3) adopt the models of the latter class, which in turn implies a negative correlation between the level of transport costs and the size of agglomeration. But, this correlation is specific to the class of their model, and the estimated impacts would differ (in fact would be opposite) if models of the other class were adopted.

要旨：With the tremendous progress of technology in recent years, the automation of production is proceeding. Moreover, it is said that the development of artificial intelligence will accelerate production automation in the future. Does automation bring growth or decline of the regions? In order to investigate the problem, this paper constructs a measure which shows how much impact of automation each state has. The measure is used to analyze the effect of automation on interstate migration in the period 2004 - 2015. It is shown that people migrate to a state with more impact of automation. Especially, skilled occupations and some unskilled occupations, which is related to skilled occupations, tend to migrate. These results suggest that automation may accelerate agglomeration of skilled occupations.

要旨：This paper builds on the canonical model of city systems with micro-foundations in terms of input sharing. This model is amended so that the production of local intermediates goes along with emissions in addition to labor. This modification affects the balance between agglomeration economies and crowding costs in favor of the latter. Emissions may have a negative impact on the welfare of consumers locally and globally for the entire city system. National and local governments are assumed to be benevolent and set environmental policies which take the negative effect of emissions on consumers into account. The paper characterizes and compares the social optimum with the allocations under local governments and land developers, allowing for environmental policies to be set either nationally or in decentralized fashion.