nep-ecm New Economics Papers
on Econometrics
Issue of 2007‒06‒11
thirty-two papers chosen by
Sune Karlsson
Orebro University

  1. Flexible Time Series Forecasting Using Shrinkage Techniques and Focused Selection Criteria By Christian T. Brownlees; Giampiero Gallo
  2. Tikhonov Regularization for Functional Minimum Distance Estimators By P. Gagliardini; O. Scaillet
  3. Instrumental Variables Estimation of Heteroskedastic Linear Models Using All Lags of Instruments By Kenneth D. West; Ka-fu Wong; Stanislav Anatolyev
  4. Local Transformation Kernel Density Estimation of Loss Distributions By J. Gustafsson; M. Hagmann; J.P. Nielsen; O. Scaillet
  5. The Empirical Saddlepoint Approximation for GMM Estimators By Sowell, Fallaw
  6. Forecasting key macroeconomic variables from a large number of predictors: A state space approach By Arvid Raknerud, Terje Skjerpen and Anders Rygh Swensen
  7. A multiple regime smooth transition heterogeneous autoregressive model for long memory and asymmetries By Michael McAller; Marcelo C. Medeiros
  8. Testing for Strict Stationarity By George Kapetanios
  9. On the Uniformly Most Powerful Invariant Test for the Shoulder Condition in Line Transect Sampling By Riccardo Borgoni; Piero Quatto
  10. Dynamic time series binary choice By Robert M. de Jong; Tiemen Woutersen
  11. Panel Unit Root Tests and Spatial Dependence By Badi H. Baltagi; Georges Bresson; Alain Pirotte
  12. Fixed and Random Effects Models for Count Data By William Greene
  13. Robust Subsampling By Lorenzo Camponovo; Olivier Scaillet; Fabio Trojani
  14. Dynamic Stochastic General Equilibrium (DSGE) Priors for Bayesian Vector Autoregressive (BVAR) Models: DSGE Model Comparison By Theodoridis, Konstantinos
  15. Efficiency Bounds for Estimating Linear Functionals of Nonparametric Regression Models with Endogenous Regressors By Thomas A. Severini; Gautam Tripathi
  16. Estimating Probabilities of Default With Support Vector Machines By Wolfgang Härdle; Rouslan Moro; Dorothea Schäfer
  17. Testing the Martingale Difference Hypothesis Using Neural Network Approximations By George Kapetanios; Andrew P. Blake
  18. Stylized Facts of Return Series, Robust Estimates, and Three Popular Models of Volatility By Teräsvirta, Timo; Zhao, Zhenfang
  19. Information criteria for impulse response function matching estimation of DSGE models By Alastair Hall; Atsushi Inoue; James M. Nason; Barbara Rossi
  20. Correlation in Bivariate Poisson Regression Model By William Greene
  21. Testing for cointegration using the Johansen approach: Are we using the correct critical values? By Paul Turner
  22. General Saddlepoint Approximations: Application to the Anderson-Darling Test Statistic By Qian Chen; David E. Giles
  23. On the Interaction between Ultra–high Frequency Measures of Volatility By Giampiero Gallo; Margherita Velucchi
  24. A Saddlepoint Approximation to the Distribution of the Half-Life Estimator in an Autoregressive Model: New Insights Into the PPP Puzzle By Qian Chen; David E. Giles
  25. A Solution Method for Linear Rational Expectation Models under Imperfect Information By Katsuyuki Shibayama
  26. Construction of an Index by Maximization of the Sum of its Absolute Correlation Coefficients with the Constituent Variables By Mishra, SK
  27. Using Meta Analysis for Benefits Transfer: Theory and Practice By John C. Bergstrom; Laura O. Taylor
  28. A Comparative Study of Various Inclusive Indices and the Index Constructed by the Principal Components Analysis By Mishra, SK
  29. Total factor productivity estimation: A practical review By Ilke Van Beveren
  30. Poisson Models with Employer-Employee Unobserved Heterogeneity: An Application to Absence Data By Jean-François Angers; Denise Desjardins; Georges Dionne; Benoit Dostie; François Guertin
  31. The 2 × 2 × 2 case in causality, of an effect, a cause and a confounder. A cross-over’s guide to the 2 × 2 × 2 contingency table By Colignatus, Thomas
  32. The Introduction of Dependent Interviewing on the British Household Panel Survey By Annette Jäckle; Heather Laurie; S.C. Noah Uhrig

  1. By: Christian T. Brownlees (Università degli Studi di Firenze, Dipartimento di Statistica); Giampiero Gallo (Università degli Studi di Firenze, Dipartimento di Statistica "G. Parenti")
    Abstract: Nonlinear time series models can exhibit components such as long range trends and seasonalities that may be modeled in a flexible fashion. The resulting unconstrained maximum likelihood estimator can be too heavily parameterized and suboptimal for forecasting purposes. The paper proposes the use of a class of shrinkage estimators that includes the Ridge estimator for forecasting time series, with a special attention to GARCH and ACD models. The local large sample properties of this class of shrinkage estimators is investigated. Moreover, we propose symmetric and asymmetric focused selection criteria of shrinkage estimators. The focused information criterion selection strategy consists of picking up the shrinkage estimator that minimizes the estimated risk (e.g. MSE) of a given smooth function of the parameters of interest to the forecaster. The usefulness of such shrinkage techniques is illustrated by means of a simulation exercise and an intra-daily financial durations forecasting application. The empirical application shows that an appropriate shrinkage forecasting methodology can significantly outperform the unconstrained ML forecasts of rich flexible specifications.
    Keywords: Forecasting, Shrinkage Estimation, FIC, MEM, GARCH, ACD
    JEL: C22 C51 C53
    Date: 2007–05
  2. By: P. Gagliardini (University of Lugano and Swiss Finance Institute); O. Scaillet (University of Geneva and Swiss Finance Institute)
    Abstract: We study the asymptotic properties of a Tikhonov Regularized (TiR) estimator of a functional parameter based on a minimum distance principle for nonparametric conditional moment restrictions. The estimator is computationally tractable and takes a closed form in the linear case. We derive its asymptotic Mean Integrated Squared Error (MISE), its rate of convergence and its pointwise asymptotic normality under a regularization parameter depending on sample size. The optimal value of the regularization parameter is characterized. We illustrate our theoretical findings and the small sample properties with simulation results for two numerical examples. We also discuss two data driven selection procedures of the regularization parameter via a spectral representation and a subsampling approximation of the MISE. Finally, we provide an empirical application to nonparametric estimation of an Engel curve.
    Keywords: MinimumDistance, Nonparametric Estimation, III-posed In-verse Problems, Tikhonov Regularization, Endogeneity, InstrumentalVariable, Generalized Method of Moments, Subsampling, Engelcurve.
    JEL: C13 C14 C15 D12
    Date: 2006–05
  3. By: Kenneth D. West; Ka-fu Wong; Stanislav Anatolyev
    Abstract: We propose and evaluate a technique for instrumental variables estimation of linear models with conditional heteroskedasticity. The technique uses approximating parametric models for the projection of right hand side variables onto the instrument space, and for conditional heteroskedasticity and serial correlation of the disturbance. Use of parametric models allows one to exploit information in all lags of instruments, unconstrained by degrees of freedom limitations. Analytical calculations and simulations indicate that there sometimes are large asymptotic and finite sample efficiency gains relative to conventional estimators (Hansen (1982)), and modest gains or losses depending on data generating process and sample size relative to quasi-maximum likelihood. These results are robust to minor misspecification of the parametric models used by our estimator.
    JEL: C13 C32
    Date: 2007–05
  4. By: J. Gustafsson (Codan Insurance and University of Copenhagen, Copenhagen, Denmark); M. Hagmann (University of Geneva and Concordia Advisors, London, United Kingdom); J.P. Nielsen (Festina Lente and University of Copenhagen, Copenhagen, Denmark); O. Scaillet (University of Geneva and Swiss Finance Institute)
    Abstract: We develop a tailor made semiparametric asymmetric kernel density estimator for the estimation of actuarial loss distributions. The estimator is obtained by transforming the data with the generalized Champernowne distribution initially fitted to the data. Then the density of the transformed data is estimated by use of local asymmetric kernel methods to obtain superior estimation properties in the tails. We find in a vast simulation study that the proposed semiparametric estimation procedure performs well relative to alternative estimators. An application to operational loss data illustrates the proposed method.
    Keywords: Actuarial loss models, Transformation, Champernowne distribution, asymmetric kernels, local likelihood estimation
    JEL: C13 C14
    Date: 2006–11
  5. By: Sowell, Fallaw
    Abstract: The empirical saddlepoint distribution provides an approximation to the sampling distributions for the GMM parameter estimates and the statistics that test the overidentifying restrictions. The empirical saddlepoint distribution permits asymmetry, non-normal tails, and multiple modes. If identification assumptions are satisfied, the empirical saddlepoint distribution converges to the familiar asymptotic normal distribution. In small sample Monte Carlo simulations, the empirical saddlepoint performs as well as, and often better than, the bootstrap. The formulas necessary to transform the GMM moment conditions to the estimation equations needed for the saddlepoint approximation are provided. Unlike the absolute errors associated with the asymptotic normal distributions and the bootstrap, the empirical saddlepoint has a relative error. The relative error leads to a more accurate approximation, particularly in the tails.
    Keywords: Generalized method of moments estimator; test of overidentifying restrictions; sampling distribution; empirical saddlepoint approximation; asymptotic distribution
    JEL: C5 C12 C13
    Date: 2006–07
  6. By: Arvid Raknerud, Terje Skjerpen and Anders Rygh Swensen (Statistics Norway)
    Abstract: We use state space methods to estimate a large dynamic factor model for the Norwegian economy involving 93 variables for 1978Q2–2005Q4. The model is used to obtain forecasts for 22 key variables that can be derived from the original variables by aggregation. To investigate the potential gain in using such a large information set, we compare the forecasting properties of the dynamic factor model with those of univariate benchmark models. We find that there is an overall gain in using the dynamic factor model, but that the gain is notable only for a few of the key variables.
    Keywords: Dynamic factor model; Forecasting; State space; AR models
    JEL: C13 C22 C32 C53
    Date: 2007–05
  7. By: Michael McAller (School of Economics and Commerce, University of Western Australia); Marcelo C. Medeiros (Department of Economics, PUC-Rio)
    Abstract: In this paper we propose a flexible model to capture nonlinearities and long-range dependence in time series dynamics. The new model is a multiple regime smooth transition extension of the Heterogenous Autoregressive (HAR) model, which is specifically designed to model the behavior of the volatility inherent in financial time series. The model is able to describe simultaneously long memory, as well as sign and size asymmetries. A sequence of tests is developed to determine the number of regimes, and an estimation and testing procedure is presented. Monte Carlo simulations evaluate the finite-sample properties of the proposed tests and estimation procedures. We apply the model to several Dow Jones Industrial Average index stocks using transaction level data from the Trades and Quotes database that covers ten years of data. We find strong support for long memory and both sign and size asymmetries. Furthermore, the new model, when combined with the linear HAR model, is viable and flexible for purposes of forecasting volatility.
    Keywords: Realized volatility, smooth transition, heterogeneous autoregression, financial econometrics,leverage, sign and size asymmetries, forecasting, risk management, model combination.
    Date: 2007–04
  8. By: George Kapetanios (Queen Mary, University of London)
    Abstract: The investigation of the presence of structural change in economic and financial series is a major preoccupation in econometrics. A number of tests have been developed and used to explore the stationarity properties of various processes. Most of the focus has rested on the first two moments of a process thereby implying that these tests are tests of covariance stationarity. We propose a new test for strict stationarity, that considers the whole distribution of the process rather than just its first two moments, and examine its asymptotic properties. We provide two alternative bootstrap approximations for the exact distribution of the test statistic. A Monte Carlo study illustrates the properties of the new test and an empirical application to the constituents of the S&P 500 illustrates its usefulness.
    Keywords: Covariance stationarity, Strict stationarity, Bootstrap, S&P500
    JEL: C32 C33 G12
    Date: 2007–06
  9. By: Riccardo Borgoni; Piero Quatto
    Abstract: In wildlife population studies one of the main goals is estimating the population abundance. Line transect sampling is a well established methodology for this purpose. The usual approach for estimating the density or the size of the population of interest is to assume a particular model for the detection function (the conditional probability of detecting an animal given that it is at a given distance from the observer). Two common models for this function are the half-normal model and the negative exponential model. The estimates are extremely sensitive to the shape of the detection function, particularly to the so-called shoulder condition, which ensures that an animal is almost certain to be detected if it is at a small distance from the observer. The half-normal model satisfies this condition whereas the negative exponential does not. Therefore, testing whether such a hypothesis is consistent with the data is a primary concern in every study aiming at estimating animal abundance. In this paper we propose a test for this purpose. This is the uniformly most powerful test in the class of the scale invariant tests. The asymptotic distribution of the test statistic is worked out by utilising both the half-normal and negative exponential model while the critical values and the power are tabulated via Monte Carlo simulations for small samples. .
    Keywords: Line Transect Sampling, Shoulder Condition, Uniformly Most Powerful Invariant Test, Asymptotic Critical Values, Monte Carlo Critical Values
    JEL: C12
    Date: 2007–05
  10. By: Robert M. de Jong; Tiemen Woutersen
    Abstract: This paper considers dynamic time series binary choice models. It proves near epoch dependence and strong mixing for the dynamic binary choice model with correlated errors. Using this result, it shows in a time series setting the validity of the dynamic probit likelihood procedure when lags of the dependent binary variable are used as regressors, and it establishes the asymptotic validity of Horowitz’ smoothed maximum score estimation of dynamic binary choice models with lags of the dependent variable as regressors. For the semiparametric model, the latent error is explicitly allowed to be correlated. It turns out that no long-run variance estimator is needed for the validity of the smoothed maximum score procedure in the dynamic time series framework.
    Date: 2007–06
  11. By: Badi H. Baltagi (Center for Policy Research, Maxwell School, Syracuse University, Syracuse, NY 13244-1020); Georges Bresson; Alain Pirotte
    Abstract: This paper studies the performance of panel unit root tests when spatial effects are present that account for cross-section correlation. Monte Carlo simulations show that there can be considerable size distortions in panel unit root tests when the true specification exhibits spatial error correlation. These tests are applied to a panel data set on net real income from the 1000 largest French communes observed over the period 1985-1998.
    Keywords: Nonstationarity, panel data, spatial dependence, cross-section correlation, unit root tests
    JEL: C23
    Date: 2006–12
  12. By: William Greene
    Date: 2007
  13. By: Lorenzo Camponovo (University of Lugano); Olivier Scaillet (University of Geneva and Swiss Finance Institute); Fabio Trojani (University of St. Gallen)
    Abstract: We compute the breakdown point of the subsampling quantile of a general statistic, and show that it is increasing in the subsampling block size and the breakdown point of the statistic. These results imply fragile subsampling quantiles for moderate block sizes, also when subsampling procedures are applied to robust statistics. This instability is inherited by data driven block size selection procedures based on the minimum confidence interval volatility (MCIV) index. To overcome these problems, we propose for the linear regression setting a robust subsampling method, which implies a su±ciently high breakdown point and is consistent under standard conditions. Monte Carlo simulations and sensitivity analysis in the linear regression setting show that the robust subsampling with block size selection based on the MCIV index outperforms the subsampling, the classical bootstrap and the robust bootstrap, in terms of accuracy and robustness. These results show that robustness is a key aspect in selecting data driven subsampling block sizes.
    Keywords: Subsampling, bootstrap, breakdown point, robustness, regression
    JEL: C12 C13 C15
    Date: 2006–11
  14. By: Theodoridis, Konstantinos (Cardiff Business School)
    Abstract: This Paper describes a procedure for constructing theory restricted prior distributions for BVAR models. The Bayes Factor, which is obtained without any additional computational effort, can be used to assess the plausibility of the restrictions imposed on the VAR parameter vector by competing DSGE models. In other words, it is possible to rank the amount of abstraction implied by each DSGE model from the historical data.
    Keywords: BVAR; DSGE Model Evaluation; Gibbs Sampling; Bayes Factor
    JEL: C11 C13 C32 C52
    Date: 2007–06
  15. By: Thomas A. Severini (Northwestern University); Gautam Tripathi (University of Connecticut)
    Abstract: Consider a nonparametric regression model Y=mu*(X) + e, where the explanatory variables X are endogenous and e satisfies the conditional moment restriction E[e|W]=0 w.p.1 for instrumental variables W. It is well known that in these models the structural parameter mu* is 'ill-posed' in the sense that the function mapping the data to mu* is not continuous. In this paper, we derive the efficiency bounds for estimating linear functionals E[p(X)mu*(X)] and int_{supp(X)}p(x)mu*(x)dx, where p is a known weight function and supp(X) the support of X, without assuming mu* to be well-posed or even identified.
    Keywords: Efficiency bounds, Linear functionals, Nonparametric regression, Endogenous regressors
    JEL: C14
    Date: 2007–05
  16. By: Wolfgang Härdle; Rouslan Moro; Dorothea Schäfer
    Abstract: This paper proposes a rating methodology that is based on a non-linear classification method, the support vector machine, and a non-parametric technique for mapping rating scores into probabilities of default. We give an introduction to underlying statistical models and represent the results of testing our approach on German Bundesbank data. In particular we discuss the selection of variables and give a comparison with more traditional approaches such as discriminant analysis and the logit regression. The results demonstrate that the SVM has clear advantages over these methods for all variables tested.
    Keywords: Bankruptcy, Company rating, Default probability, Support vector machines.
    JEL: C14 G33 C45
    Date: 2007–06
  17. By: George Kapetanios (Queen Mary, University of London); Andrew P. Blake (Bank of England)
    Abstract: The martingale difference restriction is an outcome of many theoretical analyses in economics and finance. A large body of econometric literature deals with tests of that restriction. We provide new tests based on radial basis function neural networks. Our work is based on the test design of Blake and Kapetanios (2000, 2003a,b). However, unlike that work we can provide a formal theoretical justification for the validity of these tests using approximation results from Kapetanios and Blake (2007). These results take advantage of the link between the algorithms of Blake and Kapetanios (2000, 2003a,b) and boosting. We carry out a Monte Carlo study of the properties of the new tests and find that they have superior power performance to all existing tests of the martingale difference hypothesis we consider. An empirical application to the S&P500 constituents illustrates the usefulness of our new test.
    Keywords: Martingale difference hypothesis, Neural networks, Boosting
    JEL: C14
    Date: 2007–06
  18. By: Teräsvirta, Timo (Dept. of Economic Statistics, Stockholm School of Economics); Zhao, Zhenfang (Dept. of Economic Statistics, Stockholm School of Economics)
    Abstract: Financial return series of sufficiently high frequency display stylized facts such as volatility clustering, high kurtosis, low starting and slow-decaying autocorrelation function of squared returns and the so-called Taylor effect. In order to evaluate the capacity of volatility models to reproduce these facts, we apply both standard and robust measures of kurtosis and autocorrelation of squares to first-order GARCH, EGARCH and ARSV models. Robust measures provide a fresh view of stylized facts which is useful because many financial time series can be viewed as being contaminated with outliers.
    Keywords: GARCH; EGARCH; ARSV; extreme observations; autocorrelation function; kurtosis; robust measure; confidence region.
    JEL: C22 C52
    Date: 2007–06–01
  19. By: Alastair Hall; Atsushi Inoue; James M. Nason; Barbara Rossi
    Abstract: We propose a new information criterion for impulse response function matching estimators of the structural parameters of macroeconomic models. The main advantage of our procedure is that it allows the researcher to select the impulse responses that are most informative about the deep parameters, therefore reducing the bias and improving the efficiency of the estimates of the model’s parameters. We show that our method substantially changes key parameter estimates of representative dynamic stochastic general equilibrium models, thus reconciling their empirical results with the existing literature. Our criterion is general enough to apply to impulse responses estimated by vector autoregressions, local projections, and simulation methods.
    Date: 2007
  20. By: William Greene
    Date: 2007
  21. By: Paul Turner (Dept of Economics, Loughborough University)
    Abstract: This paper presents Monte Carlo simulations for the Johansen cointegration test which indicate that the critical values applied in a number of econometrics software packages are inappropriate. This is due to a confusion in the specification of the deterministic terms included in the VECM between the cases considered by Osterwald-Lenum (1992) and Pesaran, Shin and Smith (2000). The result is a tendency to reject the null of no cointegration too often. However, a simple adjustment of the critical values is enough to deal with the problem.
    Keywords: Cointegration, Johansen Test.
    JEL: C15 C32
    Date: 2007–05
  22. By: Qian Chen (China Academy of Public Finance and Public Policy, Central University of Finance & Economics); David E. Giles (Department of Economics, University of Victoria)
    Abstract: We consider the relative merits of various saddlepoint approximations for the c.d.f. of a statistic with a possibly non-normal limit distribution. In addition to the usual Lugannani-Rice approximation we also consider approximations based on higher-order expansions, including the case where the base distribution for the approximation is taken to be non-normal. This extends earlier work by Wood et al. (1993). These approximations are applied to the distribution of the Anderson-Darling test statistic. While these generalizations perform well in the middle of the distribution’s support, a conventional normal-based Lugannani-Rice approximation (Giles, 2001) is superior for conventional critical regions.
    Keywords: Saddlepoint approximation, goodness-of-fit, Anderson-Darling test
    JEL: C12 C46
    Date: 2007–05–24
  23. By: Giampiero Gallo (Università degli Studi di Firenze, Dipartimento di Statistica "G. Parenti"); Margherita Velucchi (Università degli Studi di Firenze, Dipartimento di Statistica)
    Abstract: We analyze several measures of volatility (realized variance, bipower variation and squared daily returns) as estimators of integrated variance of a continuous time stochastic process for an asset price. We use a Multiplicative Error Model to describe the evolution of each measure as the product of its conditional expectation and a positive valued iid innovation. By inserting past values of each measure and asymmetric effects based on the sign of the return in the specification of the conditional expectation, one can investigate the information content of each indicator relative to the others. The results show that there is a directed dynamic relationship among measures, with squared returns and bipower variance interdependent with one another, and affecting realized variance without any feed-back from the latter.
    Keywords: Volatility, Multiplicative Error Models, Realized Variance, Bi-power Variance, Squared Returns, Jumps.
    JEL: C22 C51 C53
    Date: 2007–05
  24. By: Qian Chen (China Academy of Public Finance and Public Policy, Central University of Finance & Economics); David E. Giles (Department of Economics, University of Victoria)
    Abstract: We derive saddlepoint approximations for the density and distribution functions of the half-life estimated by OLS from an AR(1) or AR(p) model. Our analytic results are used to prove that none of the integer-order moments of these half-life estimators exist. This provides an explanation for the unreasonably large estimates of persistency associated with the purchasing power parity “puzzle”, and it also explains the excessively wide confidence intervals reported in the empirical PPP literature.
    Keywords: Saddlepoint approximation, half-life estimator, PPP puzzle
    JEL: C13 C22 F31 F41
    Date: 2007–05–28
  25. By: Katsuyuki Shibayama
    Abstract: This paper has developed a solution algorithm for linear rational expectation models under imperfect information. Imperfect information in this paper means that some decision makings are based on smaller information sets than others. The algorithm generates the solution in the form of k_t+1 = Hk_t + Jx^t,S f_t = Fk_t + Gx^t,S where k_t and f_t are column vectors of crawling and jump variables, respectively, while x^t,S is the vertical concatenation of the column vectors of past and present innovations. The technical breakthrough in this article is made by expanding the innovation vector, rather than expanding the set of crawling variables. Perhaps surprisingly, the H and F matrices are the same as those under the corresponding perfect information models. This implies that if the corresponding perfect information model is saddle path stable (sunspot, explosive), the imperfect model is also saddle-path stable (sunspot, explosive, respectively). Moreover, if the minimum information set in the model has all the information up to time t-S-1, then the direct effects on the impulse response functions last for only the first S periods after the impulse. In the subsequent dates, impulse response functions follow essentially the same process as in the perfect information counterpart. However, imperfect information can significantly alter the quantitative properties of a model, though it does not drastically change its qualitative nature. This article demonstrates, as an example, that adding imperfect information to the standard RBC models remarkably improves the correlation between labour productivity and output. Hence, a robustness check for information structure is recommended.
    Keywords: Linear rational expectations models, imperfect information
    JEL: C63 C65 C68
    Date: 2007–01
  26. By: Mishra, SK
    Abstract: On many occasions we need to construct an index that represents a number of variables. Cost of living index, general price index, human development index, index of level of development, etc are some of the examples that are constructed by a weighted (linear) aggregation of a host of variables. The weights are determined by the importance assigned to the variables to be aggregated. The criterion on which importance of a variable (vis-à-vis other variables) is determined may be varied and usually has its own logic. In many cases the analyst does not have any preferred means or logic to determine the relative importance of different variables. In such cases, weights are assigned mathematically. One of the methods to determine such mathematical weights is the Principal Components analysis. In the Principal Components analysis weights are determined such that the sum of the squared correlation coefficients of the index with the constituent variables is maximized. The method has, however, a tendency to pick up the subset of highly correlated variables to make the first component, assign marginal weights to relatively poorly correlated subset of variables and/or to relegate the latter subset to construction of the subsequent principal components. If one has to construct a single index, such an index undermines the poorly correlated set of variables. The index so constructed is elitist in nature that has a preference to the highly correlated subset over the poorly correlated subset of variables. Further, since there is no dependable method available to obtain a composite index by merging two or more principal components, the deferred set of variables never finds its representation in the further analysis. In this paper we suggest a method to construct an index by maximizing the sum of the absolute correlation coefficients of the index with the constituent variables. We also suggest construction of an alternative index by maximin correlation. Our experiments suggest that the indices so constructed are inclusive or egalitarian. They do not prefer the highly correlated variables much to the poorly correlated variables.
    Keywords: Index; weighted linear aggregation; principal components; elitist; inclusive; egalitarian; sum of absolute correlation coefficients; maximin correlation; Human development index; cost of living index; level of development index; Differential Evolution; Particle Swarm optimization
    JEL: C43 C10
    Date: 2007–05–25
  27. By: John C. Bergstrom; Laura O. Taylor
    Abstract: Meta-analysis, or the "study of studies", attempts to statistically measure systematic relationships between reported valuation estimates for an environmental good or service and attributes of the study that generated the estimates including valuation methods, human population and sample characteristics, and characteristics of the good or service itself. In this paper, we discuss the general theory behind and practice of the emerging use of meta-analysis for benefits transfer. If carefully conducted following systematic protocols for model development, data collection, and data analysis and interpretation, we believe that meta-analysis may prove to be a useful tool for benefits transfer in particular applications. However, before widespread application of this method, more convergent validity tests are needed. One of the greatest strengths of using meta-analysis for benefits transfer is the ability to combine and summarize large amounts of information from previous studies. This strength can also lead to one of the greatest weaknesses of this method which is the loss of important valuation details across time and space in the aggregation process. Thus, application of this method to policy questions and issues should always proceed with caution.
  28. By: Mishra, SK
    Abstract: Construction of (composite) indices by the PCA is very common, but this method has a preference for highly correlated variables to the poorly correlated variables in the data set. However, poor correlation does not entail the marginal importance, since correlation coefficients among the variables depend, apart from their linearity, also on their scatter, presence or absence of outliers, level of evolution of a system and intra-systemic integration among the different constituents of the system. Under-evolved systems often throw up the data with poorly correlated variables. If an index gives only marginal representation to the poorly correlated variables, it is elitist. The PCA index is often elitist, particularly for an under-evolved system. In this paper we consider three alternative indices that determine weights given to different constituent variables on the principles different from the PCA. Two of the proposed indices, the one that maximizes the sum of absolute correlation coefficient of the index with the constituent variables and the other that maximizes the entropy-like function of the correlation coefficients between the index and the constituent variables are found to be very close to each other. These indices alleviate the representation of poorly correlated variables for some small reduction in the overall explanatory power (vis-à-vis the PCA index). These indices are inclusive in nature, caring for the representation of the poorly correlated variables. They strike a balance between individual representation and overall representation (explanatory power) and may perform better. The third index obtained by maximization of the minimal correlation between the index and the constituent variables cares most for the least correlated variable and in so doing becomes egalitarian in nature.
    Keywords: Principal components analysis; weighted linear combination; aggregation; composite index; egalitarian; inclusive; elitist; representation; under-developed systems
    JEL: C63 C43 C61
    Date: 2007–06–01
  29. By: Ilke Van Beveren
    Abstract: This paper aims to provide empirical researchers with an overview of the methodological issues that arise when estimating total factor productivity at the establishment level, as well as of the existing techniques designed to overcome them. Apart from the well-known simultaneity and selection bias; attention is given to methodological issues that have emerged more recently and that are related to the use of deflated values of inputs and outputs (as opposed to quantities) in estimating productivity at the firm level, as well as to the endogeneity of product choice. Using data on single-product firms active in the Belgian food and beverages sector, I illustrate the biases introduced in traditional TFP estimates and discuss the performance of a number of alternative estimators that have been proposed in the literature.
    Keywords: Total factor productivity; Imperfect competition; Endogenous product choice; Semiparametric estimator; Demand
    JEL: C13 C14 D24 D40
    Date: 2007
  30. By: Jean-François Angers; Denise Desjardins; Georges Dionne; Benoit Dostie; François Guertin
    Abstract: We propose a parametric model based on the Poisson distribution that permits to take into account both unobserved worker and workplace heterogeneity as long as both effects are nested. By assuming that workplace and worker unobserved heterogeneity components follow a gamma and a Dirichlet distribution respectively, we obtain a closed form the unconditional density function. We estimate the model to obtain the determinants of absenteeism using linked employer-employee Canadian data from the Workplace and Employee Survey (2003). Coefficient estimates are interpreted in the framework of the typical labor-leisure model. We show that omitting unobserved heterogeneity on either side of the employment relationship leads to notable biases in the estimated coefficients. In particular, the impact of wages on absences is underestimated in simpler models.
    Keywords: Absenteeism, Linked Employer-Employee Data, Employer-Employee Unobserved Heterogeneity, Count Data Models, Dirichlet Distribution
    JEL: J22 J29 C23
    Date: 2007
  31. By: Colignatus, Thomas
    Abstract: Basic causality is that a cause is present or absent and that the effect follows with a success or not. This happy state of affairs becomes opaque when there is a third variable that can be present or absent and that might be a seeming cause. The 2 × 2 × 2 layout deserves the standard name of the ETC contingency table, with variables Effect, Truth and Confounding and values {S, -S}, {C, -C}, {F, -F}. Assuming the truth we can find the impact of the cause from when the confounder is absent. The 8 cells in the crosstable can be fully parameterized and the conditions for a proper cause can be formulated, with the parameters interpretable as regression coefficients. Requiring conditional independence would be too strong since it neglects some causal processes. The Simpson paradox will not occur if logical consistency is required rather than conditional independence. The paper gives a taxonomy of issues of confounding, a parameterization by risk or safety, and develops the various cases of dependence and (conditional) independence. The paper is supported by software that allows variations. The paper has been written by an econometrician used to structural equations models but visiting epidemiology hoping to use those techniques in experimental economics.
    Keywords: Experimental economics; causality; cause and effect; confounding; contingency table; Simpson paradox; conditional independence; risk; safety; epidemiology; correlation; regression; Cornfield’s condition; inference
    JEL: C10
    Date: 2007–05–30
  32. By: Annette Jäckle (Institute for Social and Economic Research); Heather Laurie (Institute for Social and Economic Research); S.C. Noah Uhrig (Institute for Social and Economic Research)
    Abstract: This paper documents the introduction of dependent interviewing in wave 16 of the British Household Panel Survey (BHPS). Dependent interviewing is a method of designing questions on longitudinal surveys where substantive information, available to the survey organisation prior to the interview, is used to tailor the wording and routing of questions to the respondent's situation or to enable in-interview edit checks. The decision to introduce dependent interviewing in the BHPS was motivated by data quality issues and the paper discusses the reasoning behind this decision. A particular aim was to reduce measurement error that leads to cross-wave inconsistencies and hence biases in estimates of change, such as 'seam effects' in histories of employment or benefit receipt. The paper provides documentation for BHPS data users and outlines the implications of the changes made when using the data. The paper also provides information about the questionnaire design, testing process and technical aspects of the implementation, for survey practitioners and methodologists who may be considering implementing dependent interviewing on a longitudinal survey.
    Keywords: CAPI design
    Date: 2007–05

This nep-ecm issue is ©2007 by Sune Karlsson. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.