PurposeThe purpose of this paper is to detect the existence of price bubbles and examine the possible contributing factors that associate with price bubble occurrences in China agricultural commodity markets.Design/methodology/approachUsing recently developed rolling window right-side augmented Dickey–Fuller test, we first detect the dates of price bubbles in China's two important agricultural commodity markets, namely corn and soybeans. Then, we use a penalized maximum likelihood estimation of a multinomial logistic model to estimate the contributing factors of price bubbles in both markets, respectively.FindingsResults from the bubble detection indicate that price bubbles account for 5.48% (3.91%) of the studied periods for corn (soybeans). More importantly, we find that market liquidity and speculation have opposite effects on the occurrences of bubbles in the corn and soybeans market. World stocks-to-use and exchange rates affect the occurrences of bubbles in a different way for each commodity, as well. Price bubbles are more likely associated with strong economic activity, high interest rates and low inflation levels.Originality/valueThis is the first study considering commodity-specific features into the formation of price bubbles. Through accurately identifying the bubble dates and fixing the estimation bias of rare events models, this study enables us to obtain robust results for each commodity. The results imply that China's corn and soybeans market respond differently to the speculative activity and external shocks from international markets. Therefore, future policy regulations on commodity markets should focus on more commodity-specific factors when aiming at avoiding bubble occurrences.

This study elicits the risk preferences of rural households through a field experiment conducted in Shaanxi Province using the Holt-Laury mechanism that considers the effects of implementing the Sloping Land Conversion Program (SLCP) on the risk preferences of farmers. The program has significantly changed the structure of farmers' productive property, which may further influence their risk attitudes. This study reveals that household geographic and demographic characteristics have a significant effect on the risk preferences of participants in the experiment. More importantly, the SLCP has had a significant effect on farmers' risk preferences. Hence, when assessing the outcome of such public policies as the SLCP that may affect the future incomes of farmers, we should consider the socioeconomic characteristics of the households concerned and the public policies implemented in the targeted area in detail.