Project

Bayesian Econometrics

Updates
0 new
0
Recommendations
0 new
0
Followers
0 new
2
Reads
0 new
8

Project log

Andres Ramírez
added 2 research items
This paper presents several "ex ante" simulation exercises of the 2014 FIFA World Cup. Specifically, we estimate the probabilities of each national team advancing to different stages, using a basic Bayesian approach based on conjugate families. In particular, we use the Categorical-Dirichlet model in the first round and the Bernoulli-Beta model in the following stages. The novelty of our framework is given by the use of betting odds to elicit the hyperparameters of prior distributions. Additionally, we obtain the posterior distributions with the Highest Density Intervals of the probability to being champion for each team. We find that Brazil (19.95%), Germany (14.68%), Argentina (12.05%), and Spain (6.2%) have the highest probabilities of being champion. Finally, we identify some betting opportunities with our simulation exercises. In particular, Bosnia & Herzegovina is a promising, whereas Australia shows the lowest betting opportunities return.
Andres Ramírez
added 2 research items
The interplay between the Bayesian and Frequentist approaches: a general nesting spatial panel-data model. Spatial Economic Analysis. An econometric framework mixing the Frequentist and Bayesian approaches is proposed in order to estimate a general nesting spatial model. First, it avoids specific dependency structures between unobserved heterogeneity and regressors, which improves mixing properties of Markov chain Monte Carlo (MCMC) procedures in the presence of unobserved heterogeneity. Second, it allows model selection based on a strong statistical framework, characteristics that are not easily introduced using a Frequentist approach. We perform some simulation exercises, finding good performance of the properties of our approach, and apply the methodology to analyse the relation between productivity and public investment in the United States.
We introduce a Bayesian instrumental variable procedure with spatial random effects that handles endogeneity, and spatial dependence with unobserved heterogeneity. We find through a limited Monte Carlo experiment that our proposal works well in terms of point estimates and prediction. We apply our method to analyze the welfare effects generated by a process of electricity tariff unification on the poorest households. In particular, we deduce an Equivalent Variation measure where there is a budget constraint for a two-tiered pricing scheme, and find that 10% of the poorest municipalities attained welfare gains above 2% of their initial income.