Further simulation evidence on the performance of the Poisson pseudo-maximum likelihood estimator

We extend the simulation results in Santos Silva and Tenreyro (2006, The log of gravity, The Review of Economics and Statistics, 88, 641-658) by considering a novel data-generating process. Our results confirm that the Poisson pseudo-maximum likelihood estimator is generally well behaved, even when the proportion of zeros in the sample is very large.     

Does the Confucius institute impact international travel to China? A panel data analysis

This article examines the impact of Confucius Institutes on inbound travel to China. We estimate a panel gravity model of inbound tourism flows to China between 2004 and 2010. We use a Poisson pseudo-maximum likelihood estimator to control for heteroscedasticity endemic in gravity models (Santos Silva and Tenreyro, 2006). We find that the presence of Confucius Institute(s) in the source country increases overall tourism in general and business and worker tourists in particular.     

Federalism, separatism and international trade

This paper explores whether countries that have a federal Constitution engage in more international trade. We identify two possible mechanisms through which political fragmentation of nation-states, namely federalism, might impact positively on trade globalization processes: domestic market fragmentation and the free trade strategy pursued by certain separatist regions in federal countries. We use a gravity equation running panel regressions to estimate the impact of federalism on trade. The Poisson estimator proposed by Santos Silva and Tenreyro (2006) is used to handle the null trade flows. We test our predictions on a large data set of 148 countries on the 1980–2002 period. After controlling for determinants of trade potentially correlated with federalism, a federalist system is found to increase international trade. We also find that separatism and linguistic fractionalization impact positively on trade openness.     

International student migration to Germany

The past decades have witnessed an impressive growth of international student mobility. This article presents first empirical evidence on international student migration to Germany, one of the most important destination countries for international students worldwide. While previous research in the field has mainly used data on international trade in educational services, I use a novel approach that analyzes student mobility as a form of migration, using data on international student migrants. An augmented gravity equation is the basis for the theoretical and empirical framework. I also provide extensive sensitivity checks of the empirical results and estimates using both the usual log-linearized and a multiplicative specification of the gravity equation, following recent work by Santos Silva and Tenreyro (Rev Econ Stat 88(4): 641–658, 2006). The results provide evidence for the importance of distance—a familiar result from the empirical migration literature. Unlike for international migration on the whole, the importance of disposable income in the home country does not seem to be too big for students, and student migrant flows from politically unfree countries are significantly lower.     

The combined estimator approach to model transferability and updating

The idea of transferability is to employ in model estimation, fitted model parameters computed from a different data set. Thecombined estimator approach to the transferability problem is expressed as a linear combination of the unbiased direct estimators on the two data sets. The major gain is in variance reduction. The combined estimator is shown to have superior accuracy, in a Mean Square Error sense, to a unbiased direct estimator whenever the transfer bias is relatively small. A test that indicates if the combined estimator is superior to the direct estimator is provided. Variances of the direct estimators are assumed to be known. Monte Carlo experiments are performed to assess the quality of the approximations. The results show that the approximations used are highly conservative. An empirical example of the combined estimator applied to a discrete choice problem is presented.     

A Monte Carlo study of growth regressions

Using Monte Carlo simulations, this paper evaluates the bias properties of estimators commonly used to estimate growth regressions derived from the Solow model. We explicitly allow for measurement error, country-specific fixed effects and regressor endogeneity. An OLS estimator applied to a single cross-section of variables averaged over time (the between estimator) performs best in terms of the extent of bias on each of the estimated coefficients. Fixed-effects and the Arellano–Bond GMM estimator overstate the speed of convergence under a wide variety of assumptions, while the between estimator understates it. Finally, fixed effects and Arellano–Bond bias towards zero the slope estimates on the human and physical capital accumulation variables, while the between estimator and the Blundell–Bond system GMM estimator bias these coefficients upwards.      

On Calculation of the Extended Gini Coefficient

The conventional formula for estimating the extended Gini coefficient is a covariance formula provided by Lerman and Yitzhaki (1989). We suggest an alternative estimator, obtained by approximating the Lorenz curve by a series of linear segments. In a Monte Carlo experiment designed to assess the relative bias and efficiency of the two estimators, we find that, when using grouped data with 20 or fewer groups, our new estimator has less bias and lower mean squared error than the covariance estimator. When individual observations are used, or the number of groups is 30 or more, there is little or no difference in the performance of the two estimators.     

How useful is yet another data-driven bandwidth in long-run variance estimation?: A simulation study on cointegrating regressions

This paper investigates how bandwidth choice rules in long-run variance estimation affect finite-sample performance of efficient estimators for cointegrating regression models. Monte Carlo results indicate that Hirukawa's (2010) bandwidth choice rule contributes bias reduction in the estimators.     

The small sample properties of several limited information estimators in interdependent structures with scalar and vector autoregressive errors

This study is concerned with an examination of the finite sample behaviour of several limited information estimators in interdependent structures with error terms related over time and in certain specifications across equations. The Monte Carlo or simulation approach is adopted and applied to computationally manageable structures containing lagged dependent variables. The analysis of the Monte Carlo experiments is formulated in terms of estimating response functions, the dependent variables of which are the first two moments of target model estimators. In addition to the impact of simultaneity, autocorrelation and lagged dependent variables on the estimators, evidence is also accumulated on the small sample effects of misspecification in terms of the faulty inclusion and deletion of regressors. The results of the experiments revealed the substantial impact which autocorrelation can have on ordinary least squares (OLS) and two-stage least squares (2SLS) in terms of efficiency loss. Averaging over all the coefficients in the models, estimators which take account of both autocorrelation and simultaneity had a relative efficiency factor of about 1.5 to 1.9. Many of the parameters in the Monte Carlo model (including misspecification errors, multicollinearity) had qualitatively the same effect on bias and dispersion properties of the estimators.     

Monte Carlo studies on the effectiveness of the bootstrap bias reduction method on 2SLS estimates

If first moments exist, two stage least squares estimators are consistent although biased. In this paper several bias correction methods are compared including bootstrap two stage least squares, Nagar's k-class and jackknife estimators for both parametric and non-parametric cases. Monte Carlo experiments on several models investigate the non-large sample properties of these estimators. The results strongly favor the bootstrap procedure judged by the amount of bias reduction and comparative variances.     

Small sample properties in a heterogenous Weibull model

Neglected heterogeneity implies bias for estimators in, e.g., the Weibull duration model. In a Monte Carlo experiment the proper maximum likelihood estimator is better than a new least squares estimator. The likelihood estimator neglecting heterogeneity is inferior.     

Rational expectations estimation of macroeconomic models: some Monte Carlo results

This paper analyzes the relative performance of alternative estimation methods for rational expectations macroeconomic models using a Monte Carlo approach. The methods studied include a single equation instrumental variable method most often attributed to McCallum, a full information substitution method proposed by Taylor and an efficient full information technique developed by Wickens. In general, the results of our Monte Carlo experiments indicate that although the full information methods tend to perform bettern than the single equation techniques, the gains of efficiency are relatively modest. However, in some experiments involving misspecification errors, the single equation method outperforms the full information estimators.     

A monte carlo study of the small sample properties of various estimators of Solow's distributed lag model

The results of some Monte Carlo experiments investigating the small sample properties of estimators ofSolow's distributed lag model are presented. In general, the relative efficiency of an estimation technique depends on the true values of the parameters being estimated. But if the assumptions of the model are not violated, then the maximum and mean likelihood and the Bayesian estimators are best.     

Spatial econometric Monte Carlo studies: raising the bar

We discuss Monte Carlo methodology that can be used to explore alternative approaches to estimating spatial regression models. Our focus is on models that include spatial lags of the dependent variable, e.g., the SAR specification. A major point is that practitioners rely on scalar summary measures of direct and indirect effects estimates to interpret the impact of changes in explanatory variables on the dependent variable of interest. We argue that these should be the focus of Monte Carlo experiments. Since effects estimates reflect a nonlinear function of both \(\beta \) and \(\rho \), past studies’ focus exclusively on \(\beta \) and \(\rho \) parameter estimates may not provide useful information regarding statistical properties of effects estimates produced by alternative estimators. Since effects estimates have recently become the focus of inference regarding the significance of (scalar summary) direct and indirect impacts arising from changes in the explanatory variables, empirical measures of dispersion produced by simulating draws from the (estimated) variance–covariance matrix of the parameters \(\beta \) and \(\rho \) should be part of the Monte Carlo study. An implication is that differences in the quality of estimated variance–covariance matrices arising from alternative estimators also plays a role in determining the accuracy of inference. An applied illustration is used to demonstrate how these issues can impact conclusions regarding the performance of alternative estimators.     

Asymptotic Properties Of The Estimator Of The Long-Run Coefficient In A Dynamic Model With Integrated Regressors And Serially Correlated Errors

In this paper we examine the asymptotic properties of the estimator of the long-run coefficient (LRC) in a dynamic regression model with integrated regressors and serially correlated errors. We show that the OLS estimators of the regression coefficients are inconsistent but the OLS-based estimator of the LRC is superconsistent. Furthermore, we propose an alternative consistent estimator of the LRC, compare the two estimators through a Monte Carlo experiment, and find that the proposed estimator is MSE-superior to the OLS-based estimator.     

Teaching Regressions with a Lagged Dependent Variable and Autocorrelated Disturbances

The author attempts to rectify the unsatisfactory textbook treatment of the finite-sample properties of estimators of regression models with a lagged dependent variable and autocorrelated disturbances. He contends that the bias of the OLS estimator of a regression model with a lagged dependent variable and autocorrelated disturbances is determined by two effects, the dynamic effect and the correlation effect, which may be reinforcing or offsetting. The implications of these two effects are explored within a theoretical and a Monte Carlo framework.     

Some simulation studies of nonparametric estimators

This paper constructs a number of Monte Carlo studies to assess the quality of various nonparametric estimators that have been proposed recently for the estimation of nonlinear econometric models. We consider both kernel and Fourier series based methods of estimation, and also examine techniques that have been suggested to improve the bias properties of the kernel estimator. The two models examined are a production function and a model emphasising the effects of risk. The Fourier estimator does very well in estimating the first of these, but not the second, while the kernel estimator shows substantial bias for the first, which is only partially alleviated by the procedures advocated for bias correction, and good results for the second.     

Endogeneity bias and growth regressions

The problem of regressor endogeneity stemming from reverse casuality is one that has plagued economists working in the field of empirical economic growth for some time. This paper attempts to address the relevant magnitude of this issue in the context of growth regressions based on the Solow growth model. The paper develops a method of running Monte Carlo simulations that allows us to generate simulated data that match the moments of observed real-world data typically used in such regressions while simultaneously allowing us to impose arbitrarily high correlations between the steady-state determinants of the Solow model and the unobserved residual term of the data-generating process. After running simulations that represent a wide sample of the mathematically-possible correlations, we conclude that a between estimator or a random effects estimator will deliever a lower average absolute bias across all coefficients than alternative estimators in almost all of our simulations. Conversely, estimators that use within-country variation will generate lower biases when looking solely at rates of convergence. Furthermore, we conclude that these results are robust when restricting our sample of simulations to several subsets of the assumed parameters and to changing our assumptions about country fixed-effects terms.     

Small Sample Properties Of Existing And New Estimators Of The Autocorrelation Coefficient And Regression Coefficients In The Context Of Ar(1) Errors

This study systematically and comprehensively investigates the small sample properties of the existing and some new estimators of the autocorrelation coefficient and of the regression coefficients in a linear regression model when errors follow an autoregressive process of order one. The new estimators of autocorrelation coefficient proposed here are based on the jackknife procedure. The jackknife procedure is applied in two alternative ways: first to the regression itself, and second to the residuals of the regression model. Next, the performance of the existing and new estimators of autocorrelation coefficient (thirty-three in total) is investigated in terms of bias and the root mean squared errors. Finally, we have systematically compared all of the estimators of the regression coefficients (again thirty-three) in terms of efficiency and their performance in hypothesis testing. We observe that the performance of the autocorrelation coefficient estimators is dependent upon the degree of autocorrelation and whether the autocorrelation is positive or negative. We do not observe a direct link between the bias and efficiency of an estimator. The performance of the estimators of the regression coefficients also depends upon the degree of autocorrelation. If the efficiency of regression estimator is of concern, then the iterative Prais-Winsten estimator should be used since it is most efficient for the widest range of independent variables and values of the autocorrelation coefficient. If testing of the hypothesis is of concern, then the estimators based on jackknife technique are certainly superior and are highly recommended. However, for negative values of the autocorrelation coefficient, the estimators based on Quenouille procedure and iterative Prais-Winsten estimator are comparable. But, for computational ease iterative Prais-Winsten estimator is recommended.

     

Implications of Bid Design and Willingness-To-Pay Distribution for Starting Point Bias in Double-Bounded Dichotomous Choice Contingent Valuation Surveys

We examine starting point bias in double-bounded dichotomous choice contingent valuation surveys. We investigate (1) the seriousness of the biases for the location and scale parameters of the willingness-to-pay (WTP) in the presence of starting point bias; (2) whether or not these biases depend on the distribution of WTP and on the bid design; and (3) how well a commonly used diagnostic for starting point bias—a test of the null that bid set dummies entered in the right-hand side of the WTP model are jointly equal to zero—performs under various circumstances. Monte Carlo simulations suggest that the effect of ignoring starting point bias depends on the bid design and on the true distribution of WTP. A well-balanced, symmetric bid design may result in very modest biases even when the anchoring mechanism is very strong. The power of bid set dummies in detecting starting point bias is low. They tend to account for misspecifications in the distribution assumed by the researcher for the latent WTP, rather than capturing the presence of starting point bias.