Search

A model with proportional errors in variables arising naturally in microeconomics is considered. Unlike the classical additive errors case, all OLS parameter estimates exhibit attenuation bias that does not depend on the limiting distribution of the data. The distribution of OLS estimators is developed. With no intercept, a simple correction of OLS based on mean predictions is identified that is consistent and asymptotically normal. With an intercept, a readily available additional moment based on sample data identifies the parameters. In neither case are additional restrictions or use of extra-sample data as instruments required as for common errors-in-variables methods.

Consider an observed binary regressor D and an unobserved binary variable D*, both of which affect some other variable Y . This paper considers nonparametric identification and estimation of the effect of D on Y , conditioning on D* = 0. For example, suppose Y is a person¡¯s wage, the unobserved D* indicates if the person has been to college, and the observed D indicates whether the individual claims to have been to college. This paper then identifies and estimates the difference in average wages between those who falsely claim college experience versus those who tell the truth about not having college. We estimate this average effect of lying to be about 6% to 20%. Nonparametric identification without observing D* is obtained either by observing a variable V that is roughly analogous to an instrument for ordinary measurement error, or by imposing restrictions on model error moments.

The purpose of this paper was to realize a multivariate analysis of the monthly unemployment rate in the county of Constanta using the following variables: the average net wage (SAL), the number of the employees (EMPL), the consumption price index (IPC) for the January 2000- march 2012 period. This analysis is based on the correlation study that can be established as a consequence of applying the parametric correlation methods between the mentioned indicator and the influential factors.

The purpose of this research was to identify several features of the unemployment phenomenon in the SE region of Romania, using several of the most significant indicators highlighting the main correlations between specific labour market indicators and an indicator by which we can assess the economic development degree of a country. The resulted regression equations, tested from the point of view of their significance, allow, on the one hand, the assessment of the existing situation and, on the other hand, the performance of the forecasting calculations required to develop different scenarios and strategies for attracting investments and EU funds in order to reduce the number of the unemployed and to absorb higher education graduates on the labour market.

In a translog production function, the number of parameters practically öexplodesö as the number of considered production factors increases. Consequently, the shortcoming in the estimation of the respective production function is the occurrence of collinearity. Theoretically, the collinearity impact is minimum if a single production factor is taken into account. In this case, we can determine not only the output elasticity but also the elasticity of scale related to the respective production factor. In the present paper, we demonstrate that the relationship between the output elasticity and estimated average elasticity of scale depends on the dynamics trajectory of the production factor, underexponential and overexponential, respectively. At the end, a practical example is offered, dealing with the computation of the Gross Domestic Product elasticity and average elasticity of scale related to employed population in the United Kingdom and France during 1999-2009.

In econometric models, linear regressions with three explanatory variables are widely used. As examples can be cited: Cobb-Douglas production function with three inputs (capital, labour and disembodied technical change), Kmenta function used for approximation of CES production function parameters, error-correction models, etc. In case of multiple linear regressions, estimated parameters values and some statistical tests are influenced by collinearity between explanatory variables. In fact, collinearity acts as a noise which distorts the signal (proper parameter values). This influence is emphasized by the coefficients of alignment to collinearity hazard values. The respective coefficients have some similarities with the signal to noise ratio. Consequently, it may be used when the type of collinearity is determined. For these reasons, the main purpose of this paper is to identify all the modeling factors and quantify their impact on the above-mentioned indicator values in the context of linear regression with three explanatory variables.Classification-JEL:C13,C20,C51,C52Keywords:types of collinearity, coefficient of mediated correlation, rank of explanatory variable, order of attractor of collinearity, mediated collinearity, anticollinearity.

This paper reviews the conceptual-theoretical framework related to the analysis of the savings-investments relationship in an open economy and reveals the main trends of the respective relationship in the consolidated market economies and emerging countries during the second part of the 20-th century and first decade of the 21-st century. In this context, the author proposes some improvements in the savings-investments relationship analysis methodology, emphasizing the role of the foreign trade balance stock on the values of discrepancy between the savings and investments and of the Feldstein-Horioka econometric model estimated parameters. The respective methodological improvements are applied in the case of Romania for the period 1990-2007. At the end of the paper some guidelines for the savings-investments relationship modeling are presented in order to avoid recession and assure the premises for a sustainable economic growth.

In this paper, we characterize and study the efficiency of a representative set of firms in the leather and footwear sector in Bogotá, Bucaramanga, Cúcuta and the Metropolitan Area of Medellin. This study is based on a survey addressing multiple aspects of production in this industry (inputs, integration of activities, administrative management, innovation, exports, etc). Our results suggest that a reasonable percentage of total variance is accounted for by the variance of the inefficiency term. Additionally, the sample exhibits diminishing returns to labor. Even though many firms seem to be efficient, the leather sector as a whole is not operating near its production possibilities frontier.

High-breakdown-point regression estimators protect against large errors and data contamination. We generalize the concept of trimming used by many of these robust estimators, such as the least trimmed squares and maximum trimmed likelihood, and propose a general trimmed estimator, which renders robust estimators applicable far beyond the standard (non)linear regression models. We derive here the consistency and asymptotic distribution of the proposed general trimmed estimator under mild β-mixing conditions and demonstrate its applicability in nonlinear regression and limited dependent variable models.

The generalised method of moments estimator may be substantially biased in finite samples, especially so when there are large numbers of unconditional moment conditions. This paper develops a class of first-order equivalent semi-parametric efficient estimators and tests for conditional moment restrictions models based on a local or kernel-weighted version of the Cressie–Read power divergence family of discrepancies. This approach is similar in spirit to the empirical likelihood methods of Kitamura et al. [2004. Empirical likelihood-based inference in conditional moment restrictions models. Econometrica 72, 1667–1714] and Tripathi and Kitamura [2003. Testing conditional moment restrictions. Annals of Statistics 31, 2059–2095]. These efficient local methods avoid the necessity of explicit estimation of the conditional Jacobian and variance matrices of the conditional moment restrictions and provide empirical conditional probabilities for the observations.

In this paper we study the common impact of increasing VAT and decreasing incomes in consumer models. The considered models are linear ones (see [3], [4] and [8]). It is in fact the extension of the study [2], where there was performed the study of the impact of only one of the mentioned government decision (increasing VAT). We have already noticed that applying the simple three rule is not appropriate. But the problems that arise come from the common impact. It is possible that if it is applied only the decreasing of the wedges (25%) , the incomes from selling products decreases by the ratio b , if we apply only increasing VAT the income decreases by the ratio a (we have obtained in [2] a = 4.01653 ignoring the dependence of quantity on wedges), but if there are applied both the income decreases by the ratio g ¹1−(1−a )(1−b ) . This is the general case, and the explanation of such phenomenon comes from analogous reasons as in [2]: the total income is the sold quantity multiplied by the price, hence we have not linearity. Another explanation comes from the least squares method: in the obtained linear system for estimating the three parameters (intercept, coefficient of prices and coefficient of wedges) both variables influence the result.

We discuss instrumental variables (IV) estimation in the broader context of the generalized method of moments (GMM), and describe an extended IV estimation routine that provides GMM estimates as well as additional diagnostic tests. Stand-alone test procedures for heteroskedasticity, overidentification, and endogeneity in the IV context are also described. Copyright 2003 by Stata Corporation.

Having in view some algebraical properties of the Student Test statistics, in this paper a new indicator is proposed in order to evaluate the quality of estimation in case of a linear regression with three explanatory variables, when OLS method is used. The respective indicator, named Synthesis of Transformed Form of Student Test statistics (STFST) is defined by taking into consideration the arithmetical mean and the coefficient of variation of Transformed Form of Student Test Statistics. Also, an explanation regarding the deviation of the arithmetical mean of Transformed Form of Student Test Statistics from Fisher Test statistics is given. In this context, a factorial analysis model is proposed in order to reveal the modeling factors contribution in obtaining the Student Test statistics. At the end of the paper a numerical example is presented in order to show the practical use of the proposed analysis methodology.

When an applied econometrician calculates regression coefficients or other statistics based on a data sample, there is a moment of truth when the statistical precision or reliability of the estimates is evaluated using critical values of relevant probability distributions. If statistical precision is underestimated, producing confidence intervals that are too wide, the researcher may falsely conclude that no useful or reliable evidence has been provided. Since statistically imprecise results often are regarded as noninformative and may never see the light of day, applied econometricians have a strong incentive to obtain tight confidence bands. But overestimation of statistical precision also is a concern, as it potentially causes rejection of a null hypothesis when in fact no statistically reliable evidence has been presented. Such results are misleading and provide poor guidance for future research and perhaps public policy as well. This article describes two techniques that have been developed by statisticians in the last 20 years to enhance the accuracy of estimated confidence bands and critical values: the bootstrap and multiple imputations. These are computationally intensive methods that rely on repeated sampling from empirical data sets and associated estimates. As such, their development and dissemination have been supported by substantial increases in computing power over the same period. These techniques often can provide accurate estimates of statistical precision when standard analytical estimates are biased or, in some cases, unavailable. Both already

By noting that the Hodrick-Prescott filter can be expressed as the solution to a particular regression problem, we are able to show how to construct confidence bands for the filtered time-series. This procedure requires that the data are stationary. The construction of such confidence bands is illustrated using annual U.S. data for real value-added output; and monthly U.S. data for the unemployment rate.

I present evidence that higher frequency measures of inflation expectations outperform lower frequency measures of inflation expectations in tests of accuracy, predictive power, and rationality. For decades, the academic literature has focused on three survey measures of expected inflation: the Livingston Survey, the Survey of Professional Forecasters, and the Michigan Surveys of Consumers. While these measures have been useful in developing models of forecasting inflation, the data are low frequency measures that are anachronistic in the modern era of high frequency and real-time data. I present a collection of 37 different measures of inflation expectations, including many previously unexploited monthly and real-time measures of inflation expectations. These higher frequency measures tend to outperform the standard three low frequency survey measures in tests of accuracy, predictive power, and rationality, indicating that there are benefits to using higher frequency measures of inflation expectations. Out of sample forecasts confirm the findings.

We study the distribution of hard-, soft-, and adaptive soft-thresholding estimators within a linear regression model where the number of parameters k can depend on sample size n and may diverge with n. In addition to the case of known error-variance, we define and study versions of the estimators when the error-variance is unknown. We derive the finite-sample distribution of each estimator and study its behavior in the large-sample limit, also investigating the effects of having to estimate the variance when the degrees of freedom n-k does not tend to infinity or tends to infinity very slowly. Our analysis encompasses both the case where the estimators are tuned to perform consistent model selection and the case where the estimators are tuned to perform conservative model selection. Furthermore, we discuss consistency, uniform consistency and derive the uniform convergence rate under either type of tuning., minor corrections

In this selective review, we first provide some empirical examples that motivate the usefulness of semi-nonparametric techniques in modelling economic and financial time series. We describe popular classes of semi-nonparametric dynamic models and some temporal dependence properties. We then present penalized sieve extremum (PSE) estimation as a general method for semi-nonparametric models with cross-sectional, panel, time series, or spatial data. The method is especially powerful in estimating difficult ill-posed inverse problems such as semi-nonparametric mixtures or conditional moment restrictions. We review recent advances on inference and large sample properties of the PSE estimators, which include (1) consistency and convergence rates of the PSE estimator of the nonparametric part; (2) limiting distributions of plug-in PSE estimators of functionals that are either smooth (i.e., root-n estimable) or non-smooth (i.e., slower than root-n estimable); (3) simple criterion-based inference for plug-in PSE estimation of smooth or non-smooth functionals; and (4) root-n asymptotic normality of semiparametric two-step estimators and their consistent variance estimators. Examples from dynamic asset pricing, nonlinear spatial VAR, semiparametric GARCH, and copula-based multivariate financial models are used to illustrate the general results.

Care must be taken when interpreting the coefficients of dummy variables in semi-logarithmic regression models. Existing results in the literature provide the best unbiased estimator of the percentage change in the dependent variable, implied by the coefficient of a dummy variable, and of the variance of this estimator. We extend these results by establishing the exact sampling distribution of an unbiased estimator of the implied percentage change. This distribution is nonnormal, and is positively skewed in small samples. We discuss the construction of bootstrap confidence intervals for the implied percentage change, and illustrate our various results with two applications: one involving a wage equation, and one involving the constructions of an hedonic price index for computer disk drives.

In this paper we give a generalized model of the interest rates term structure including Nelson-Siegel and Svensson structure. For that we introduce a continuous m-factor exponential-polynomial form of forward interest rates and demonstrate its considerably better performance in a fitting of the zero-coupon curves in comparison with the well known Nelson-Siegel and Svensson ones. In the sequel we transform the model into a dynamic model for interest rates by designing a switching dynamical system of the considerably reduced dimension n

The instrumental variables strategy is commonly employed in empirical research. For correct inference using this econometric technique, the instruments must be perfectly exogenous and relevant. In fact, the standard t-ratio test statistic used in this context yields unreliable and often inaccurate results even when there is only a slight violation of the exclusion restriction. It is crucial to realize that to make reliable inferences on the structural parameters we need to know the true correlation between the structural error and the instruments. The main innovation in this paper is to identify an appropriate test in this context: a joint null hypothesis of the structural parameters with the correlation between the instruments and the structural error term. Since correlation cannot be estimated, we propose a test statistic involving a grid search over correlation values. To address inference under violations of exogeneity, significant contributions have been made in the recent literature by assuming some degree of non-exogeneity. We introduce a new approach by deriving a modified t-statistic that corrects for the bias associated with non-exogeneity of the instrument. A key advantage of our approach over that of the previous literature is that we do not need to make any assumptions about the degree of violation of exogeneity either as possible values or prior distributions. In particular, our method is not a form of sensitivity analysis. Since our modified test statistic is continuous and monotonic in correlation it is easy to conduct inference by a simple grid search. Even though the joint null may seem to be limiting in interpreting rejection, we can still make accurate inferences on the structural parameters because of a feature of the grid search over correlation values. The procedure for calculating the modified coefficients and statistics is illustrated with two empirical examples.

A regression model is considered where earnings are explained by schooling and ability. It is assumed that schooling is measured with error and that there are no data on ability. Regressing earnings on observed schooling then yields an estimate of the return to schooling that is subject to positive omitted variable bias (OVB) and negative measurement error bias (MEB). The effects on the OVB and the MEB from using family background variables as proxies for ability are investigated theoretically and empirically. The theoretical analysis demonstrates that the impact on the OVB is uncertain, while the MEB invariably increases in magnitude. The empirical analysis shows that the MEB generally dominates the OVB. As the measurement error increases and/or more family background variables are added, the total bias rapidly becomes negative, driving the estimated return further and further away from the true value.

This study sheds new light on the question of whether or not sentiment surveys, and the expectations derived from them, are relevant to forecasting economic growth and stock returns, and whether they contain information that is orthogonal to macroeconomic and financial data. I examine 16 sentiment surveys of distinct respondent universes and employ the technique of principal components analysis to extract the common signals from the surveys. I show that the ability of different population groups to anticipate correctly economic growth and excess stock returns is not identical, implying that not all sentiment is the same, although there exist some common components. I demonstrate that sentiment surveys have significant predictive power for both GDP growth and excess stock returns, and that the results are robust to the inclusion of information pertaining to the macroeconomic environment and momentum. Furthermore, the findings reject the conventional wisdom that the effect of sentiment is apparent exclusively in small-capitalization stocks.

Consider an observed binary regressor D and an unobserved binary variable D*, both of which affect some other variable Y . This paper considers nonparametric identification and estimation of the effect of D on Y , conditioning on D* = 0. For example, suppose Y is a person's wage, the unobserved D* indicates if the person has been to college, and the observed D indicates whether the individual claims to have been to college. This paper then identifies and estimates the difference in average wages between those who falsely claim college experience versus those who tell the truth about not having college.We estimate this average returns to lying to be about 7% to 20%. Nonparametric identification without observing D* is obtained either by observing a variable V that is roughly analogous to an instrument for ordinary measurement error, or by imposing restrictions on model error moments.

We study the distribution of the adaptive LASSO estimator (Zou (2006)) in finite samples as well as in the large-sample limit. The large-sample distributions are derived both for the case where the adaptive LASSO estimator is tuned to perform conservative model selection as well as for the case where the tuning results in consistent model selection. We show that the finite-sample as well as the large-sample distributions are typically highly non-normal, regardless of the choice of the tuning parameter. The uniform convergence rate is also obtained, and is shown to be slower than $n^{-1/2}$ in case the estimator is tuned to perform consistent model selection. In particular, these results question the statistical relevance of the `oracle' property of the adaptive LASSO estimator established in Zou (2006). Moreover, we also provide an impossibility result regarding the estimation of the distribution function of the adaptive LASSO estimator.The theoretical results, which are obtained for a regression model with orthogonal design, are complemented by a Monte Carlo study using non-orthogonal regressors., revised version; minor changes and some material added

This paper examines the ways in which structural systems can yield observed variables, other than the cause or treatment of interest, that can play an instrumental role in identifying and estimating causal effects. We focus speciÖcally on the ways in which structures determine exclusion restrictions and conditional exogeneity relations that act to ensure identification. We show that by carefully specifying the structural equations and by extending the standard notion of instrumental variables, one can identify and estimate causal effects in the endogenous regressor case for a broad range of economically relevant structures. Some of these have not previously been recognized. Our results there create new opportunities for identifying and estimating causal effects in non-experimental situations. Our results for more familiar structures provide new insights. For example, we extend results of Angrist, Imbens, and Rubin (1996) by taking into account an important distinction between cases where Z is an observed exogenous instrument and those where it is a proxy for an unobserved exogenous instrument. A main message emerging from our analysis is the central importance of sufficiently specifying the causal relations governing the unobservables, as these play a crucial role in creating obstacles or opportunities for identification. Because our results exhaust the possibilities for identification, we ensure that there are no other opportunities for identification based on exclusion restrictions and conditional independence relations still to be discovered. To accomplish this characterization, we introduce notions of conditioning and conditional extended instrumental variables (EIVs). These are not proper instruments, as they are endogenous. They nevertheless permit identification and estimation of causal effects. We analyze methods using these EIVs either singly or jointly.

We extend our 2003 paper on instrumental variables (IV) and GMM estimation and testing and describe enhanced routines that address HAC standard errors, weak instruments, LIML and k-class estimation, tests for endogeneity and RESET and autocorrelation tests for IV estimates.

We consider the problem of estimating the unconditional distribution of a post-model-selection estimator. The notion of a post-model-selection estimator here refers to the combined procedure resulting from first selecting a model (e.g., by a model selection criterion like AIC or by a hypothesis testing procedure) and then estimating the parameters in the selected model (e.g., by least-squares or maximum likelihood), all based on the same data set. We show that it is impossible to estimate the unconditional distribution with reasonable accuracy even asymptotically. In particular, we show that no estimator for this distribution can be uniformly consistent (not even locally). This follows as a corollary to (local) minimax lower bounds on the performance of estimators for the distribution; performance is here measured by the probability that the estimation error exceeds a given threshold. These lower bounds are shown to approach 1/2 or even 1 in large samples, depending on the situation considered. Similar impossibility results are also obtained for the distribution of linear functions (e.g., predictors) of the post-model-selection estimator.

We consider estimating the linear regression model’s coefficients when there is uncertainty about coefficient restrictions. Theorems establish that the mean squared errors of combination estimators, formed as weighted averages of the ordinary least squares and one or more restricted least squares estimators, depend on finding the optimal estimator of a single normally distributed vector. Our results generalize those of Magnus and Durbin (1999) [Magnus, J.R., Durbin, J. 1999. Estimation of regression coefficients of interest when other regression coefficients are of no interest. Econometrica 67, 639-643] and Danilov and Magnus (2004) [Danilov, D., Magnus, J.R. 2004. On the harm that ignoring pretesting can cause. Journal of Econometrics 122, 27-46].

This paper introduces a new class of robust regression estimators. The proposed twostep least weighted squares (2S-LWS) estimator employs data-adaptive weights determined from the empirical distribution, quantile, or density functions of regression residuals obtained from an initial robust fit. Just like many existing two-step robust methods, the proposed 2S-LWS estimator preserves robust properties of the initial robust estimate. However contrary to existing methods, the first-order asymptotic behavior of 2S-LWS is fully independent of the initial estimate under mild conditions; most importantly, the initial estimator does not need to be pn consistent. Moreover, we prove that 2S-LWS is asymptotically normal under B-mixing conditions and asymptotically efficient if errors are normally distributed. A simulation study documents these theoretical properties in finite samples; in particular, the relative efficiency of 2S-LWS can reach 85–90% in samples of several tens of observations under various distributional models.

High breakdown-point regression estimators protect against large errors and data con- tamination. We generalize the concept of trimming used by many of these robust estima- tors, such as the least trimmed squares and maximum trimmed likelihood, and propose a general trimmed estimator, which renders robust estimators applicable far beyond the standard (non)linear regression models. We derive here the consistency and asymptotic distribution of the proposed general trimmed estimator under mild B-mixing conditions and demonstrate its applicability in nonlinear regression and limited dependent variable models.

The binary-choice regression models such as probit and logit are used to describe the effect of explanatory variables on a binary response vari- able. Typically estimated by the maximum likelihood method, estimates are very sensitive to deviations from a model, such as heteroscedastic- ity and data contamination. At the same time, the traditional robust (high-breakdown point) methods such as the maximum trimmed like- lihood are not applicable since, by trimming observations, they induce the separation of data and non-identification of parameter estimates. To provide a robust estimation method for binary-choice regression, we con- sider a maximum symmetrically-trimmed likelihood estimator (MSTLE) and design a parameter-free adaptive procedure for choosing the amount of trimming. The proposed adaptive MSTLE preserves the robust prop- erties of the original MSTLE, significantly improves the infinite-sample behavior of MSTLE, and additionally, ensures asymptotic efficiency of the estimator under no contamination. The results concerning the trim- ming identification, robust properties, and asymptotic distribution of the proposed method are accompanied by simulation experiments and an application documenting the infinite-sample behavior of some existing and the proposed methods.

No foolproof method exists to fit nonlinear curves to data or estimate the parameters of an intrinsically nonlinear function. Some methods succeed at solving a set of problems but fail at the others. The Differential Evolution (DE) method of global optimization is an upcoming method that has shown its power to solve difficult nonlinear optimization problems. In this study we use the DE to solve some nonlinear least squares problems given by the National Institute of Standards and Technology (NIST), US Department of Commerce, USA and some other challenge problems posed by the CPC-X Software (the makers of the AUTO2FIT software). The DE solves the test problems given by the NIST and most of the challenge problems posed by the CPC-X, doing marginally better than the AUTO2FIT software in a few cases.

In this paper, a general Blinder-Oaxaca decomposition is derived that can also be applied to non-linear models, which allows the differences in a non-linear outcome variable between two groups to be decomposed into a part that is explained by differences in observed characteristics and a part attributable to differences in the estimated coeffcients. Departing from this general model, we show how it can be applied to different models with discrete and limited dependent variables.

State price density (SPD) contains important information concerning market expectations. In existing literature, a constrained estimator of the SPD is found by nonlinear least squares in a suitable Sobolev space. We improve the behavior of this estimator by implementing a covariance structure taking into account the time of the trade and by considering simultaneously both the observed Put and Call option prices.

In this paper, a general Blinder-Oaxaca decomposition is derived that can also be applied to non-linear models, which allows the differences in a non-linear outcome variable between two groups to be decomposed into a part that is explained by differences in observed characteristics and a part attributable to differences in the estimated coeffcients. Departing from this general model, we show how it can be applied to different models with discrete and limited dependent variables.

The principal purpose of this paper is to adapt to the conditional moment context the GEL unconditional moment methods described in Smith(1997, 2001) and Newey and Smith(2004). In particular we develop GEL estimators which achieve the semiparametric efficiency lower bound. The requisite GEL criteria are constructed by local smoothing and parallel the local semiparametric efficient EL method formulated by Kitamura, Tripathi and Ahn (2004) for conditional moment restrictions. A particular advantageof these efficient local methods is the avoidance of the necessity of providing explicit estimators for the Jacobian and conditional variance matrices. The class of local GEL estimators admits a number of alternative first order equivalent estimators such as local EL, local ET and local CUE as in the unconditional moment restrictions case. The paper also provides a local GEL criterion function test statistic for parametric restrictions.

The validity of the key behavioral parameters used in the calibration process of computable general equilibrium (CGE) models remains a debated issue in the CGE literature. CGE modelers prefer to borrow from the handful of estimates available in the literature rather than estimating these parameters empirically. The dearth of data is often mentioned as the major reason for compromises to the empirical basis for the parameters used in CGE models. While the empirical literature on demand elasticities based on household expenditure surveys has been relatively available for both developed and developing countries, it remains lacking for African countries. This paper uses a seemingly unrelated regressions method to estimate own-price and income elasticities, as well as Frisch parameters for households whose consumption behavior is described by a Linear Expenditure System (LES) demand function. All the parameters estimated are intended for use in a Lesotho CGE model. The estimation results are generally consistent with the theory predictions.

In this paper, I develop a quasi empirical likelihood estimator that has good finite-sample properties when there are many moment conditions. I show that the quasi empirical likelihood estimator, which uses semiparametric efficient estimation, is an approximation to the empirical likelihood estimator, which has been shown to have good statistical properties. The quasi empirical likelihood estimator is a consistent estimator and has a normal asymptotic distribution. As with the full-blown empirical likelihood estimator, the quasi empirical likelihood estimator reduces finite-sample bias, but is much simpler to compute than the empirical likelihood estimator. Monte Carlo experiments and a quick validation exercise confirm my theoretical results

At the beginning of the nineties labor productivity in France slowed down strongly. Yearly labor productivity (per worker) growth in the non-farm business sector fell from 2% during the 1982-1992 period to 0.7% between 1993 and 2000. This productivity slowdown seems to have stimulated employment, especially in the tertiary sector. Several papers deal with this subject, but none can claim to give a proper explanation of the whole productivity slowdown. This paper investigates the productivity slow-down in France using quarterly microeconomic data. We estimate an employment equation, paying special attention to the long run specifications and to the cointegration problems. The equations are estimated on the non-farm business sector. We also concentrate on the non-manufacturing business sector, since it was the most affected by the labor productivity slowdown. We conclude that the drop in the number of hours worked, the change in the labor cost and in that of unskilled workers relative to skilled workers have contributed to the growth of employment in France. Moreover, we cannot exclude a slackening of the long run productivity growth.

We take a fresh look at Theil's BLUS residuals and ask why they have gone out of fashion.All our simulation experiments indicate that tests based on BLUS residuals have higher power than those based on the more popular recursive residuals, even in those cases (structural breaks) where intuition would favour the recursive residuals.

This paper studies the Ordinary Least-Squared (OLS) and Instrumental Variables (IV) estimates of the returns to schooling for male workers in Spain. OLS estimates are often biased due to the endogeneity of schooling, measurement errors or omitted variables. Proper IV estimates correct this bias. The reliability of family background, natural experiments (based on changes in the education system and season of birth) and the availability of a college in the province is checked using Spanish data. The results suggest that background and college availability are valid instruments and that the IV estimates of the returns to schooling are higher than OLS estimates. These results are in line with the majority of previous results in the literature.

This paper condiders an extension of Tran Van Hoa's family of 2SHI (two stage hierarchical information) estimators for the coefficient vector of a linear regression model and derives the conditions for the dominance of 2SHI estimator over the OLS and Stein rule estimators under a Generalised Pitman Nearness (GPN) criterion when the disturbance variable is small.

Lam and Schoeni (1993) consider an equation where earnings are explained by schooling and ability. They assume that ability data are lacking and that schooling is measured with error. The estimate obtained by regressing earnings on schooling thus contains omitted variable bias (OVB), which is positive, and measurement error bias (MEB), which is negative. Adding a family background variable to proxy for ability is claimed to: i) decrease the OVB towards, but not below, zero and ii) make the MEB even more negative. This note claims that while ii) is true, even in the context of multiple family background variables, i) is in general incorrect. The OVB may increase in magnitude and/or change sign. Conditions are provided under which i) holds. A simulation procedure is suggested that will yield consistent estimates of the total bias and its components, conditional upon values on the true return and the measurement error variance.

Our paper addresses the correction of the aggregation bias in linear rational expectations models when there is some unobserved micro-parameter heterogeneity and only macro data are available. Starting from Lewbel (1994), we propose two new consistent estimators, which rely on a flexible parametric specification of the cross-sectional parameter distributions and account for the dependence across coeffcients inherent in such models. A Monte-Carlo study reveals that the finite-sample and asymptotic properties of the proposed estimators correct the aggregation bias found with the maximum-likelihood and generalized-method-of-moments approaches. As a by-product, we can also infer the cross-sectional distribution of the parameters. Finally, we reassess the empirical evidence about the New Keynesian Phillips curve and explain the apparent discrepancy between micro- and macro-based estimates of the average persistence of inflation.