In case the off-farm wage employment is constrained, the "internal wage" plays an important role of equilibrating the demand for labor with its supply within the household. Once this distinct role is recognized, responses of quantities demanded and supplied respectively are decomposed into the direct effects of changes in exogenous variables and the "internal wage effects" of changes in the internal wage caused by those in the same exogenous variables. This method of decomposition proves to be more readily amenable to the standard microeconomic theory to enhance the tractability of agricultural household models under constrained off-farm wage employment.

The purpose of this study is to assess the convergent validity of environmental benefit estimates within a framework of benefit transfer. Benefit transfer is the process to predict a benefit estimate of a new policy site by using benefit estimates or benefit functions of existing studies. If benefit estimates at the policy and study site are not statistically different, the convergent validity of benefit transfer is presumed. Benefit transfer is also an attractive procedure when a policy maker conducts a benefit-cost analysis within the limits of tight budgetary appropriation. In this study, meta-analytic transfer and benefit function transfer were employed to test the convergent validity. Although several off-the-shelf contingent valuation studies on agricultural landscapes and rural amenities were collected, 11 were usable for the meta-analytic transfer and 6 for the benefit function transfer. It was empirically shown that percentage errors resulting from meta-analytic transfers by policy site data were in the range of 1 to 26%. On the other hand, percentage errors resulting from benefit function transfers by policy site data were in the range of 3 to 45%. It can be said that both meta-analytic transfer and benefit function transfer are likely to generate accurate benefit estimates in terms of percentage error. Besides comparative analyses of percentage error, convergent validity was also tested to learn if the estimate from benefit function transfer lied within the confidence interval over the original estimate and vice versa. As a consequence of the hypothesis test, 67% (8 of 12) of the tests rejected the convergent validity of benefit function transfers.

The purpose of this paper is to evaluate the effects of incentives and operating rules of JA's insurance-selling divisions on the characteristics of the divisions by adopting an agency theory. The analysis is based on data from three JAs and their member surveys. The JA's manager as a principal has two types of insurance-selling divisions. One is selling full-time as agent-1, and the other is selling limited term as agent-2 through JA employees outside the insurance division. The Kohonen network model is applied to cluster analysis intended to identify key characteristics of the JA insurance division by using its managerial indicators. Three types of JA managerial indicators of insurance division are classified: the rate of intention to conclude JA's insurance policies in the future is high (T1); the growth rate of policies is high (T2); and some indicators suggest that the type concentrates on the market of insurance against death (T3). The difference of contract menu by which a principal creates incentives and operating rules for agents is estimated by using a probit model. The major findings are as follows. First, if the principal creates a goal that is acceptable to agent-1 and agent-2, such as sales promotions for member-oriented (T1) or high growth rate of policies (T2), then particular incentives for agent-1 aren't designed. Second, if the principal recognizes that the profit of the insurance division reached a high level, then the principal will reward agent-1 by designed particular incentive contracts. Third, if the principal designed to share an agent's information toward to deal with asymmetry of information, it may contribute to the member's customization for JA. The classification model is also estimated by using a radial basis function model. The model partly improves the estimation error and gives a more visual and detailed presentation to the JA classification.

My experience concerning agricultural and rural development research originally started from farm management studies in the late 1950s. I started with a survey on the structure and operation of farming businesses in individual farm households. I then studied the developments in orange production and the factors that resulted in differences in product profitability. Consequently, I studied investment efficiencies and supply response on production in terms of production function, and the cost function models in orange production based on traditional farm management data. When comparing agriculture in Southeast Asia to Japanese paddy agriculture, I reviewed the efficiency of input-output production and other parameters of correlated factors that influence economic performance. The types of farming practices, coefficients and elasticity of structural equations were also estimated. Since managerial process and knowledge situation of small farmers are supposed to create differences in farming performance, surveys of subsistence farmers were also pursued. In the 1970s, agricultural and rural developments in the context of regional agricultural development were important issues in South and Southeast Asia. During my involvement in several agricultural development programs and projects that JICA supported, I made various kinds of interview surveys concerning farming activities. I applied the principal component analysis technique in analyzing the rural survey data. Through the survey, preferences for rural policies were reviewed in different areas. The preferences and problems towards rural development were also discussed. The need for comprehensive and integrated planning efforts were stressed in the report. Since 1980 I was involved in a research program for agricultural and rural development in Bangladesh. This country was in one of the worst socioeconomic condition. It was essential to increase agricultural production because agriculture was the most important source of income for majority of the population. However, agricultural production was insufficient for the whole population in rural areas—because of flooding every year for a period of several months to half of the year. Inland fishing and other kinds of jobs provided supplementary income. Thus, agricultural development program may cause a conflict against other kinds of development programs such as fishing or non-agricultural development programs. According to the results of our pilot study in Bangladesh, the close relationship of the agricultural sector with the inland fishery and other non-agricultural sector programs was identified. Improvements in other livelihood activities were also considered as important for subsistence farmers according to individual household surveys. Here again, surveys concerning perception, intent, and expectation of farmers towards development policies were necessary in order to formulate practical programs for farmers. The methods of data collection were crucial for getting meaningful results. Moreover, in order to prevent further deterioration of present resources and to improve the environment, comprehensive management devices in development planning policy should be formulated and implemented. For the coming century, it is hoped that sustainability in agricultural production balanced between ecology and human needs could be attained. The ongoing developments in information technology will tremendously improve data collection and their analyses. In order to achieve more fruitful research results for practical use, it is necessary to organize multidisciplinary collaborations of natural, agronomic and social scientists.