Abstract: Given a market for land circulation, cropland abandonment due to lack of labor forces will be alleviated. Therefore, the correlation between agricultural laborer and cropland abandonment at the farm household level will be weakened to some degree. In this paper, we aimed to identify a reasonable level for analysis of agricultural laborer and size of cropland abandonment. Because land use property transfer primarily occurs within the village unit, we supposed this relationship may be more pronounced at the village level. To test this hypothesis, this study examines to what extent agricultural laborer impacts the size of cropland abandonment at the village level and at the farm household level, and compares these results. The data from a survey of 308 farm households in 40 villages from Wulong County of Chongqing Municipality were used in the analysis. Data for the two variables (agricultural laborer per unit of farm land area and abandoned cropland area in the household) were collected through the survey for each household, and they were aggregated to form the village level data for the corresponding village units. We first conducted two Logistic regressions for the two levels respectively. The results show that the size of agricultural laborer can effectively influence the cropland abandonment decision at the village level, but it had no significant influence at the farm household level. This finding confirms our hypothesis that the size of agricultural laborer affects cropland abandonment much stronger at the village level than the household level. We further explored the reason of this relationship by using a partial correlation analysis between agricultural laborer per unit of land area and the size of abandoned land at both the household level and the village level. Furthermore, the results are compared between with and without controlling for the rate of land transfer. The Spearman correlation between agricultural laborer for each unit of farm land (mu) and abandoned cropland area are not significant at the household level (the correlation coefficient is -0.027), but this relationship is remarkably strong when controlling the land circulation rate, with a coefficient of -0.273. The result at the village level shows that the two variables are significantly correlated with or without controlling for land circulation, and the two correlation coefficients are very close: -0.273 for the controlled analysis and -0.279 without land transfer control. That is, the correlation between the two variables is slightly enhanced at the village level with a market for land circulation. The Pearson partial correlation analysis shows that the correlation between agricultural laborer and cropland abandonment was weakened by land circulation at the farm household level, but enhanced at the village level. Through this analysis, we conclude that village level is a reasonable level for analysis of agricultural laborer affecting cropland abandonment. Due to the land circulation effects, at the household level agricultural laborer cannot be an effective explanatory variable for cropland abandonment. We recommend that the relationship between agricultural laborer and cropland abandonment should be analyzed at the village level.

Key words: agricultural laborer, cropland abandonment, land circulation, level, Wulong County, Chongqing City