The foodgrain market in India has been under multiplex government intervention. This paper discusses three kinds of intervention: procurement, public distribution system (PDS), and inter-state moving restrictions. Frequently, some types of intervention have been evaluated positively. These evaluations are: (1) PDS gives poor people easier access to food; and (2) procurement has a positive effect on production. Especially, the second effect is know as the 'Dantwala-Mellor conjecture.' This conjecture means that procuring a part of production causes an increase in open market price, and may result in improvement of the producers' welfare.
　Under 'The New Economic Policy' since 1991, agricultural liberalization has been a hot issue. It includes the reduction of agricultural subsidies and the relief of various regulations.
　The essential liberalization of moving restrictions is requested under globalization. In contrast, because of the need to care for poor people, the package of procurement-PDS prograrm will be maintained. However, procurement is imposed mainly upon agricultural developing regions. As a result, procurement policy is connected with the intervention of moving restrictions. In the present paper, the effects of the liberalization of moving restrictions on welfare are analyzed based on the Dantwala-Mellor conjeture.The main conclusions are:
　(1) 　The Dantwala-Mellor effect on production is less under the liberalization of moving restrictions.
　(2) 　The liberalization of moving restrictions increases food subsidies.
　Under the present situation, because the reduction of subsides is focused, this partial liberalization has some difficulties. However, liberalization with procurement and PDS will harm the welfare of poor people. India's agriculture face a dilemma under economic liberalization.
　The effects of export liberalization are also analyzed under the same framework.

　A household-managed land system has gradually replaced the cooperative-managed land system in the Red Rive Delta of Vietnam, under the economic reform plan called "Doi Moi," which was formulated in 1986 but did not take effect until 1988 through the enforcement of Resolution 10.
　Resolution 10 stated that households were once again the basic production units in the agricultural sector and farmlands were allocated to farming households rather than cooperatives for a period of fifteen years. However, this policy did not ensure legal land-use rights for farmers. In 1993, the government undertook a major revision of the land law to complete the land reform contemplated under Resolution 10. Under the new land law, farming households are provided with stable and long-term rights to use land, 20 years for annual crop lands and 50 years for perennial crop lands, and were given Land Use Right Certificates (LURCs).
　This paper shows that when farmlands were allocated to farming households in 1988, the fifteen-year tenure right was not realized at the grassroots level. Furthermore, the farmland allocations created an inefficient land system characterized by fragmentation of land plots. When land re-allocations were executed before the issuance of LURCs, an attempt was made by rural communities to consolidate lands. It was found that this attempt succeeded under the following two conditions: the availability of off-farm jobs and the capabilities of community leaders. In villages where farmers obtained high incomes from off-farm activities, the average number of plots per household was smaller than in the villages where farmers have fewer opportunities to get off-farm jobs. Although their economic environment was not favorable, some villages managed to consolidate farmlands by building a consensus among farmers.

　Demand analysis has been one of the most major research fields in empirical economic studies. However, there are several neglected issues. Retail level demand analysis is among them. Most econometric studies on demand have been based on aggregate data of commodity groups in, for example, the Family Income and Expenditure Survey by the Japanese government. To analyze the economic behavior of each company or to analyze factors influencing a selected brand, demand analysis at the brand level has been needed.
　This paper analyzed demand for margarine at the brand level. The author has estimated a Linearly Approximate, Almost Ideal Demand System (LA/AIDS) model for margarine utilizing scanner data (POS data) from 65 regional markets in the United States in 1992.
　The author obtained good estimates of own and cross price elasticities, expenditure elasticities, and elasticities of some exogenous variables. The main results are as follows: 1) Own price elasticities at brand level are larger than expected from the past analyses of commodity groups of food; 2) cross price elasticity has shown substitution between two brands produced by a single company; 3) private brands have a somewhat isolated market position with other brands; 4) the top brand has the highest expenditure elasticity, suggesting its characteristic as a superior good; and 5) some exogenous variables are successful to capture factors influencing the demand for each brand. These results seem to be interesting for not only researchers but also practitioners.
　This paper has shown the usefulness and significance of demand analysis at the brand level, and discussed some problems for conducting this type of research using Japanese scanner data.

　The purpose of this research is an econometric analysis of Japanese agriculture in the postwar period. The convergence problem in economic growth theory is also investigated for agricultural growth in Japan. The results are as follows: First, neither of the variance coefficients of agricultural value added or farm income decreased; therefore, we could not see a σ convergence. Second, we extended the Barro & Sala-i-Martin model and checked if there was a β convergence; however, we could not observe a β convergence. We did obtain a conditional β convergence if we included the capital labor ratio, the ratio of agricultural budget, the degree of agricultural dependency, and regional and trade liberalization dummy variables. This means that a region with a low-level agricultural income grows faster than that of one with high income. Third, we also extended the Mankiw, Romer & Weil model but could not get a good result. Therefore, we included almost the same variables as the above Barro & Sala-i-Martin case. Then, we obtained a plausible result which shows that the growth rate of agricultural labor productivity increases 0.0029% and 0.0154%, respectively, if the agricultural budget and agricultural dependency each increase 1%. Also, the trade liberalization of oranges and beef damaged many prefectures. Fourth, the factor shares of land, physical capital, human capital and labor are 7.53, 33.01, 27.37 and 32.09%, respectively. Therefore, the labor share, which includes both labor and human capital, is 59.46%, and the capital share, which includes both physical capital and land, is 40.54%. These are reasonable results.