Semiparametric estimation and inference using doubly robust moment conditions

We study semiparametric two-step estimators which have the same structure as parametric doubly robust estimators in their second step. The key difference is that we do not impose any parametric restriction on the nuisance functions that are estimated in a first stage, but retain a fully nonparametric model instead. We call these estimators semiparametric doubly robust estimators (SDREs), and show that they possess superior theoretical and practical properties compared to generic semiparametric two-step estimators. In particular, our estimators have substantially smaller first-order bias, allow for a wider range of nonparametric first-stage estimates, rate-optimal choices of smoothing parameters and data-driven estimates thereof, and their stochastic behavior can be well-approximated by classical first-order asymptotics. SDREs exist for a wide range of parameters of interest, particularly in semiparametric missing data and causal inference models. We illustrate our method with a simulation exercise.

Inference in differences-in-differences with few treated groups and heteroskedasticity

Differences-in-Differences (DID) is one of the most widely used identification strategies in applied economics. However, how to draw inferences in DID models when there are few treated groups remains an open question. We show that the usual inference methods used in DID models might not perform well when there are few treated groups and errors are heteroskedastic. In particular, we show that when there is variation in the number of observations per group, inference methods designed to work when there are few treated groups tend to (under-) over-reject the null hypothesis when the treated groups are (large) small relative to the control groups. This happens because larger groups tend to have lower variance, generating heteroskedasticity in the group x time aggregate DID model. We provide evidence from Monte Carlo simulations and from placebo DID regressions with the American Community Survey (ACS) and the Current Population Survey (CPS) datasets to show that this problem is relevant even in datasets with large numbers of observations per group. We then derive an alternative inference method that provides accurate hypothesis testing in situations where there are few treated groups (or even just one) and many control groups in the presence of heteroskedasticity. Our method assumes that we can model the heteroskedasticity of a linear combination of the errors. We show that this assumption can be satisfied without imposing strong assumptions on the errors in common DID applications. With many pre-treatment periods, we show that this assumption can be relaxed. Instead, we provide an alternative inference method that relies on strict stationarity and ergodicity of the time series. Finally, we consider two recent alternatives to DID when there are many pre-treatment periods. We extend our inference methods to linear factor models when there are few treated groups. We also derive conditions under which a permutation test for the synthetic control estimator proposed by Abadie et al. (2010) is robust to heteroskedasticity and propose a modification on the test statistic that provided a better heteroskedasticity correction in our simulations.

Inference in differences-in-differences with few treated groups and heteroskedasticity

Differences-in-Differences (DID) is one of the most widely used identification strategies in applied economics. However, how to draw inferences in DID models when there are few treated groups remains an open question. We show that the usual inference methods used in DID models might not perform well when there are few treated groups and errors are heteroskedastic. In particular, we show that when there is variation in the number of observations per group, inference methods designed to work when there are few treated groups tend to (under-) over-reject the null hypothesis when the treated groups are (large) small relative to the control groups. This happens because larger groups tend to have lower variance, generating heteroskedasticity in the group x time aggregate DID model. We provide evidence from Monte Carlo simulations and from placebo DID regressions with the American Community Survey (ACS) and the Current Population Survey (CPS) datasets to show that this problem is relevant even in datasets with large numbers of observations per group. We then derive an alternative inference method that provides accurate hypothesis testing in situations where there are few treated groups (or even just one) and many control groups in the presence of heteroskedasticity. Our method assumes that we know how the heteroskedasticity is generated, which is the case when it is generated by variation in the number of observations per group. With many pre-treatment periods, we show that this assumption can be relaxed. Instead, we provide an alternative application of our method that relies on assumptions about stationarity and convergence of the moments of the time series. Finally, we consider two recent alternatives to DID when there are many pre-treatment groups. We extend our inference method to linear factor models when there are few treated groups. We also propose a permutation test for the synthetic control estimator that provided a better heteroskedasticity correction in our simulations than the test suggested by Abadie et al. (2010).

Genetic parameter estimates for buffalo milk yield, milk quality and mozzarella production and Bayesian inference analysis of their relationships

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Bayesian inference in genetic parameter estimation of visual scores in Nellore beef-cattle

The aim of this study was to estimate the components of variance and genetic parameters for the visual scores which constitute the Morphological Evaluation System (MES), such as body structure (S), precocity (P) and musculature (M) in Nellore beef-cattle at the weaning and yearling stages, by using threshold Bayesian models. The information used for this was gleaned from visual scores of 5,407 animals evaluated at the weaning and 2,649 at the yearling stages. The genetic parameters for visual score traits were estimated through two-trait analysis, using the threshold animal model, with Bayesian statistics methodology and MTGSAM (Multiple Trait Gibbs Sampler for Animal Models) threshold software. Heritability estimates for S, P and M were 0.68, 0.65 and 0.62 (at weaning) and 0.44, 0.38 and 0.32 (at the yearling stage), respectively. Heritability estimates for S, P and M were found to be high, and so it is expected that these traits should respond favorably to direct selection. The visual scores evaluated at the weaning and yearling stages might be used in the composition of new selection indexes, as they presented sufficient genetic variability to promote genetic progress in such morphological traits.

Estimation of genetic parameters for milk yield in Murrah buffaloes by Bayesian inference

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Genetic parameters for buffalo milk yield and milk quality traits using Bayesian inference

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Genetic correlations between heifer subsequent rebreeding and age at first calving and growth traits in Nellore cattle by Bayesian inference

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

5S rDNA variation and its phylogenetic inference in the genus Leporinus (Characiformes : Anostomidae)

5S rDNA sequences have proven to be valuable as genetic markers to distinguish closely related species and also in the understanding of the dynamic of repetitive sequences in the genomes. In the aim to contribute to the knowledge of the evolutionary history of Leporinus (Anostomidae) and also to contribute to the understanding of the 5S rDNA sequences organization in the fish genome, analyses of 5S rDNA sequences were conducted in seven species of this genus. The 5S rRNA gene sequence was highly conserved among Leporinus species, whereas NTS exhibit high levels of variations related to insertions, deletions, microrepeats, and base substitutions. The phylogenetic analysis of the 5S rDNA sequences clustered the species into two clades that are in agreement with cytogenetic and morphological data.

Body size and flight distance in stingless bees (Hymenoptera: Meliponini): inference of flight range and possible ecological implications

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

A fuzzy inference system for identification of crossflow filtration processes

The crossflow filtration process differs of the conventional filtration by presenting the circulation flow tangentially to the filtration surface. The conventional mathematical models used to represent the process have some limitations in relation to the identification and generalization of the system behavior. In this paper, a system based on fuzzy logic systems is developed to overcome the problems usually found in the conventional mathematical models. Imprecisions and uncertainties associated with the measurements made on the system are automatically incorporated in the fuzzy approach. Simulation results are presented to justify the validity of the proposed approach.

Mapping lightning processes using fuzzy inference system

This paper describes a novel approach for mapping lightning processes using fuzzy logic. The core regarding lightning process is to identify and to model those uncertain information on mathematical principles. In fact, the lightning process involves several nonlinear features that our current mathematical tools would not be able to model. The estimation process has been carried out using a fuzzy system based on Sugeno's architecture. Simulation results confirm that proposed approach can be efficiently used in these types of problem.

Psychophysical power functions for apparent area in perceptive, memory, and inference conditions for observers of different age groups

The purpose of this research was to verify the effect of age on the exponent of the power function in Perceptive, Memory, and Inference experimental conditions. In the Memory condition the intervals of 2 min., 8, 24, and 48 hr. and 1 wk. were used between acquisition of information and remembering. For each experimental condition the ages of observers ranged between 17 and 35 years (Group I), 40-55 years (Group II), and 60-77 years (Group III), and education ranged from high school to graduate school. The observers estimated the areas of the Brazilian states using the psychophysical method of magnitude estimation. No significant differences were obtained for Groups I, II, and III for each experimental condition, except in the Memory Condition with the 24-hr. interval. Analysis for experimental conditions and ages showed a significant difference between the Perceptive Condition and each of the others, but no difference between the Inference and Memory Conditions. These results indicated that in the remembering processes there is no loss of information as a function of age. From the small variability in the power function exponents for the three ages, we may assume that age could be related to amount of education of the observers, which suggests study is important.

Genetic parameters by Bayesian inference for dual purpose Jaffarabadi buffaloes

Knowledge of genetic parameters is essential for improved reproductive management and increased yield. Quantitative analysis of genetic parameters is lacking for many breeds of buffaloes. This article provides the first estimate of genetic parameters for dual purpose (meat and milk) Brazilian Jaffarabadi buffaloes, using Bayesian inference. Data on milk yield (MY), lactation length (LL), weight at 205 days (W205) and 365 (W365) days of age, and average daily gain (ADG) from 205 to 365 days of age were collected in two herds. Bivariate analyses (using the program MTGSAM) were performed with the Gibbs sampler to obtain estimates of variance and covariance. Average lactation milk yield and lactation length were 1 620.2 +/- 450.9 kg and 257.6 +/- 46.8 days, respectively, and the mean values for weight traits (kg) were 181.6 +/- 63.3 (W205), 298.04 +/- 116.1 (W365), and 0.73 +/- 0.35 (ADG). Heritability estimates (modes) were 0.16 for MY, 0.10 for LL, 0.43 for W205, 0.48 for W365 and 0.32 for ADG. There was a high genetic correlation (0.96) between milk yield and lactation length and very high genetic correlations (0.99) between the three growth traits. Our data suggest that both milk production and growth traits have clear potential for yield improvement through direct selection in this dual purpose breed. The selection for weight at an early age would be successful and selection for MY can be performed in the first lactation.