Zhang, Yinan: Conceptualising and estimating rationalised agricultural optimisation models. - Bonn, 2018. - Dissertation, Rheinische Friedrich-Wilhelms-Universität Bonn.
Online-Ausgabe in bonndoc: https://nbn-resolving.org/urn:nbn:de:hbz:5n-51198
@phdthesis{handle:20.500.11811/7353,
urn: https://nbn-resolving.org/urn:nbn:de:hbz:5n-51198,
author = {{Yinan Zhang}},
title = {Conceptualising and estimating rationalised agricultural optimisation models},
school = {Rheinische Friedrich-Wilhelms-Universität Bonn},
year = 2018,
month = jul,

note = {Computational modelling for quantitative agricultural policy assessment in the EU employs more farm level oriented approaches in recent years. This follows policy instruments that increasingly target the farm level and have effects varying with farm characteristics. At the same time, methodological advances such as Positive Mathematical Programming (PMP) increased the acceptance of farm level modelling for policy analysis. By introducing non-linear terms into the objective function of programming models, PMP offers an elegant calibration property and smooth simulation response. This thesis addresses the lack of economic rationalisation of PMP and the econometric estimation of alternative model formulation.
First, this dissertation analyses the economic rationality of the most often used quadratic PMP model. One potential rationalisation of non-linear terms in the objective function discussed in the literature is a non-linear capacity constraint (CC) representing some aggregate of labour and capital stock. Results show that the equivalence between a quadratic CC formulation and PMP model is limited to the calibration property of the programming model. In terms of simulation behaviour and estimation, the two models differ. Therefore, a quadratic capacity constraint cannot fully rationalise a quadratic PMP model. Nevertheless, it could effectively connect supply models to market models in order to exchange information on primary factor. Second, the thesis examines the consistency of Econometric Mathematical Programming (EMP) models. They allow estimating parameters of non-linear technologies using multiple observations and first-order conditions as estimating equations. The chosen EMP model is a single farm optimisation model with Constant Elasticity of Substitution production functions. A Monte Carlo setup evaluates the consistency of the estimation procedure under different error structures. Results show that the estimated parameters converge to the true values with increasing sample sizes. Finally, the dissertation addresses the lack of statistical inference procedures for EMP models in the literature. Bootstrapped confidence intervals are suggested here and evaluated with respect to the accuracy of the coverage probabilities, again using a Monte Carlo approach. The simulated confidence intervals generally succeed in approximating correct coverage probabilities with sufficient accuracy but results differ somewhat by sampling approach and choice of confidence interval calculation.
Keywords: positive mathematical programming, capacity constraint, econometric mathematical programming model, errors in optimisation, bootstrapped confidence intervals.},

url = {https://hdl.handle.net/20.500.11811/7353}
}

The following license files are associated with this item:

InCopyright