Development of a Bayesian estimator for non-stationary Markov transition probabilities and its application to EU farm structural change
Final Report Abstract
The projects improve and develop new empirical method to analyze and model farm structural change. Changes of the farm structure are not only important for the sector itself but may have broader economic, social and environmental consequences for a region. Understanding this process is important in order to assess how (agricultural-) policy affects this development. A common approach to analyze farm structural change is the Markov framework. The project provides a Bayesian estimation framework that allows combining individual and aggregated level data in the estimation of non-stationary Markov models more consistently then existing methods. It is shown that the data combination improves estimation precision and the numerical stability of the estimation. Building on this, a Bayesian prediction formwork for EU farm structural change is developed that exploits the available information more fully. Secondly, farm interdependences and their importance for farm structural change are analyzed. It is argued that the assumption of independence between farm behaviors can become problematic in specific applications. Empirical evidence is provided that these interactions are important to consider for a consistent aggregation of farm level results and an assessment of policy effects at regional level. Specifically, it is show for the case of Norway that it is important to consider neighboring characteristic when assessing the influence of direct payments on farm survival. To the knowledge of the author the project is the first to show empirically that spatial interdependence on the farm level are indeed important for farm structural change. With respect to policy assessment, the empirical results indicate that direct payments a farm received itself have a positive influence on farm survival while neighboring direct payments have a negative one. For an overall assessment of the policy effects it is thus necessary to consider the interdependencies between farms. Ignoring these interdependencies might lead to an overestimation of the effects of direct payments.
Publications
- (2011): Bayesian estimation of non-stationary Markov models combining micro and macro data. 2011 AAEA Annual Meetings, Pittsburgh, Pennsylvania, July 24-26
Storm, H. and T. Heckelei
- (2012): Predicting agricultural structural change using census and sample data. 2012 AAEA Annual Meetings, Seattle, Washington, August 12-14
Storm, H. and T. Heckelei
- (2013): Direct payments, spatial competition and farm survival in Norway. 2013 AAEA Annual Meetings, Washington, DC, August 4-6, 2013
Storm, H., K. Mittenzwei and T. Heckelei
- (2014): Bayesian estimation of nonstationary Markov models combining micro and macro data. Working Paper
Storm, H., T. Heckelei and R.C. Mittelhammer
