No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Small Sample Performance of Dynamic Panel Data Estimators in Estimating the Growth-Convergence Equation: A Monte Carlo Study
Abstract
This chapter conducts a Monte Carlo investigation into small sample properties of some of the dynamic panel data estimators that have been applied to estimate the growth-convergence equation using Summers-Heston data set. The results show that the OLS estimation of this equation is likely to yield seriously upward biased estimates. However, indiscriminate use of panel estimators is also risky, because some of them display large bias and mean square error. Yet, there are panel estimators that have much smaller bias and mean square error. Through a judicious choice of panel estimators it is therefore possible to obtain better estimates of the parameters of the growth-convergence equation. The growth researchers may make use of this potential.
... Primero, no existe consenso sobre el método de estimación de la ecuación en (4). Islam (2000) y Hauk y Wacziarg (2009), mediante un estudio de Monte Carlo en regresiones de crecimiento, demuestran que los resultados que se obtienen varían dependiendo del método de estimación, pero que ninguno de los métodos está libre de costos. En otras palabras, existe un trade-off entre el sesgo de heterogeneidad y la medición del error. ...
... Como lúcidamente explica Islam (1995Islam ( , p.1138) "La presencia de la variable dependiente rezagada en la parte derecha de la ecuación hace del estimador MCEF un estimador inconsistente… Sin embargo las propiedades asintóticas de estimadores de datos de panel pueden ser consideradas en la dirección de T" 13 Así mismo, Amemiya (1967) demuestra que cuando T → ∞ el estimador MCEF es asintóticamente equivalente al Estimador de Máxima Verosimilitud. Empíricamente el estimador MCEF parece no presentar mayores problemas cuando es aplicado al estudio de convergencia como ha quedado demostrada en Islam (1995Islam ( , 2000 quien concluye que el estimador MCEF tiene un mejor desempeño en estudios de crecimiento que estimadores más complejos. 14 Finalmente, es importante aclarar que, con datos de panel, siempre existe la posibilidad de asumir que el shock tecnológico tiene un efecto aleatorio en lugar de efecto fijo. ...
A panel spatial approach is adopted to analyze the convergence hypothesis for Bolivia’s regional GDP for the period 1988-2011. Two conclusions emerge from our results. First, the speed of convergence to the steady state for the departments of Bolivia fluctuates around 4 % to 7 %. Second, we could not find statistical evidence of spatial dependence, put another way, the speed of convergence of a department does not appear to be affected by the GDP of its neighbors. Additionally our results are robust to different econometric specifications.
... The slope coefficients are assumed to be equal for all the countries: the rate of convergence is thus identically estimated for all the countries. The empirical studies of Knight, Loayza and Villanueva (1993), Islam (1995Islam ( , 2000, Caselli, Esquivel and Lefort,(1996) or Berthélemy and Varoudakis (1998) are representative of this type of real convergence analysis. However, the homogeneity of the slope coefficients is often an unreasonable assumption, and one can allow for crosssectional heterogeneity. ...
... The panel data analysis focuses on the sample of 28 members of the enlarged EU. ) and the initial income per capita. Barro and Sala-I-Martin (1995) specified the model of absolute convergence rewritten in dynamics for panel data by Islam (1995Islam ( , 2000: ...
The European Integration of the Central Eastern Europe countries (CEECs) is still a
challenging issue. Firstly, the disparities of the per capita revenue between the CEECs
and EU-15 can blame the feasibility of an enlarged Europe with homogeneous growth.
Secondly, the global economic crisis is actually impacting their growth rates. The
objective of this article is to test the existence of the catching-up process between
European countries. If real convergence is empirically checked, at which speed is it
actually done? Thanks to the theoretical and empirical lessons, the β - convergence
hypothesis is tested for the heterogeneous European panel data. In particular, the
heterogeneity among the speeds of convergence contradicts the idea of the uniformity
of the catching-up process in Eastern Europe. Thus, the EU enlargement seems to lead
to a “variable-speed Europe” of real convergence.
... The slope coefficients are assumed to be equal for all the countries: the rate of convergence is thus identically estimated for all the countries. The empirical studies of Knight, Loayza and Villanueva (1993), Islam (1995Islam ( , 2000, Caselli, Esquivel and Lefort,(1996) or Berthélemy and Varoudakis (1998) are representative of this type of real convergence analysis. However, the homogeneity of the slope coefficients is often an unreasonable assumption, and one can allow for cross-sectional heterogeneity. ...
... Method : Iterative Bayesian procedure.Islam N. [2000] 10 proposes to test the following specification for the model of conditional convergence in panel data: ...
The aim of this paper is to examine the macroeconomic implications of remittances in MENA and Mediterranean countries. After a brief review of theoretical and empirical literature devoted to macroeconomic performance of remittances, we select the arguments that can be applied to this region and try to identify a significant relationship between remittances and per capita growth in these countries. The Iterative Bayesian procedure allows us to calculate the rates of convergence for each MENA country (without taking into account the remittances flows). Our results show that only 6 countries out of 13 appear to have a positive impact of remittances on growth.
... Para determinar se o efeito threshold é estatisticamente significante no modelo (13a-13c), testa-se a hipótese nula de ausência de efeito threshold: Anderson e Hsiao (1981, 1982 e o estimador baseado em GMM proposto por Arellano e Bond (1991). Islam (2000) utiliza os dados de Summers-Heston (1991, 1998) e analisa as propriedades de pequenas amostras dos principais estimadores de painel dinâmico. Através de um estudo de Monte Carlo aplicado a um modelo de crescimento para análise de convergência, ele verifica que em pequenas amostras há a possibilidade da presença de viés de estimação. ...
... Além dos estimadores Anderson-Hsiao e Arellano-Bond,Islam (2000) analisa também o de mínima distância sugerido porChamberlain (1982Chamberlain ( , 1983 e os estimadores tradicionais OLS e LSDV. ...
... 69. See for example Alonso-Borrego and Arellano (1999), Kiviet (1995), Harris and Matyas (1996), Islam (2000), Jusdon andOwen (1997), andZiliak (1997). However, it needs to be mentioned that the instrument set used for the Arellano and Bond (1991) estimator in the Monte Carlo study is not exactly the same that Caseli et al. seems to have used. ...
... For example, Amemiya (1967Amemiya ( , 1971 shows that, although LSDV is biased in the direction of N ! 1, it is consistent in the direction of T ! 1. 72. See for example Islam (2000). The dynamic panel data estimators considered in this study include: LSDV, two instrumental variable estimators by Hsiao (1981, 1982), two GMM estimators by Arellano and Bond (1991), 2SLS, 3SLS, generalized 3SLS and Minimum Distance estimators by Chamberlain (1982Chamberlain ( , 1983. ...
This paper surveys the convergence literature. It begins by laying out different definitions of convergence and by showing the link between the convergence issue and the growth theory debate. The paper then follows the convergence research conducted along four different approaches, namely the cross-section, panel, time-series, and distribution approaches. The paper shows the association of these methodological approaches with various definitions of convergence and highlights the connections among the convergence results. It shows that, despite some impressions to the contrary, there is considerable agreement among the results. Although the convergence research might not have solved the growth debate entirely, it has helped both the neoclassical and the new growth theories to adapt and evolve. The research on convergence has established new stylized facts regarding cross-country growth regularities. It has brought to fore the existence of large technological and institutional differences across countries and has given rise to new methodologies for quantifying and analyzing these differences. This is providing a new information base for analysis of technological and institutional diffusion and for further development of growth theory in general. Copyright Blackwell Publishing Ltd, 2003.
... This choice is justified by Islam (2001) findings. In fact, while investigating the small sample properties of dynamic panel estimators for a growth convergence equation and comparing a set of ten different estimators, he surprisingly found that the LSDV estimator proves to be a relatively superior estimator that outperforms more sophisticated ones 6 . ...
This paper provides a literature review and an empirical study of the particular independent growth impact of various capital inflows in the GCC countries during the period 2005-2018. It augments the standard growth determinants with inflows of foreign direct investment (FDI), foreign portfolio investment (FPI) and other investments (OI) and estimates this equation with the LSDV estimator. Main results are that capital inflows as a whole don't exert any effect on economic growth. However, when it is decomposed in different inflows kinds, other investment and FDI don't affect growth while foreign portfolio investment has a significant negative effect.
Keywords: Capital flows, FDI, Portfolio investment, other investment, economic growth, GCC countries.
JEL Classifications : F21, F32, F43
DOI: https://doi.org/10.32479/ijefi.10319
... Baseline Regression Analysis Islam (2001) found that the least square dummy variable (LSDV) estimation method is even better than the IV estimation and GMM estimation for samples with large numbers of individuals but limited time periods. Since the data used are short term and the number of companies is relatively large, we first conduct the LSDV method to investigate the effects of environmental management and debt financing on sustainable financial growth. ...
Environmental management and sustainable financial growth are currently hot topics in academic research. This article examines the relationships among environmental management, debt financing, and sustainable financial growth in the Chinese tourism industry. The results show that environmental management and debt financing have promoted sustainable financial growth, and the overall effect of debt financing on sustainable financial growth has been affected by environmental management. After employing different methods of controlling the endogeneity, these conclusions are still robust. We also examine the mediating effect and threshold effect on environmental management, debt financing, and sustainable financial growth, and the results reveal that debt financing can mediate the effect of environmental management on sustainable financial growth and that there is nonlinear impact of debt financing on sustainable financial growth in different thresholds of environmental management. The analysis results show that the presented policy proposals promote the development of tourism companies from the aspects of debt financing and environmental management.
... However, taking into consideration the properties of the LSDVC estimator, we decided to use it to evaluate causality between structural change and economic growth 1 . On the other hand, Islam (2001) compares different estimators for a dynamic panel growth-convergence equation based on a small sample and states that the LSDV estimator 'proves to be a relatively superior estimator for the problem' although T is small and that small-sample properties of estimators may depend on the problem analysed and data considered. The LSDV estimator is also used by Dietrich (2012) to assess the causality of structural changes vs. economic growth with the aid of a VAR model. ...
The objective of this study is to test empirically the relationship between structural changes (changes in gross value added and employment) and economic growth. We used a panel Granger-causality analysis based on annual data for eight transition countries, covering the period 1995–2011. The main finding is that the causality relations analysed are heterogeneous processes and are identified more often when we measure structural changes by value added than by changes in employment. Among the countries analysed, we separate a subgroup of economies with very strong bilateral causality (small countries such as Latvia, Lithuania, and Estonia), a subgroup in which no causal relationships are observed (e.g., Hungary in the case of employment), and a group with a one-directional relationship (e.g., Poland, where GDP changes cause employment changes, but not vice versa). The research results point to the necessity of taking into account different relationships, whether one- or two-directional, between growth and structural changes in government economic policy. The paper presents a verifiable methodology, which was originally used to identify the analysed relationship in transition countries.
... In general, to test a real convergence, we need to specify the correlation between the GDP per capita growth rate and the initial GDP per capita (see Barro and Sala-I-Martin 1995). Here, we follow the model of real convergence proposed by Islam (1995Islam ( , 2001, who developed the original cross-section model of absolute β-convergence in dynamic panel data model: ...
The convergence and the business cycle synchronisation in the European Union (EU-28) remains a topical issue as recent economic crisis has affected GDP growth of each member country. The aim of the paper is to test a real convergence between the European countries and to analyse the synchronisation of their GDP per capita. Firstly, the speed of β-convergence is tested using Bayesian Shrinkage Estimators allowing for heterogeneity in panel data model. Secondly, we estimate the cluster dendrogram, which offers the categorisation of countries into clusters according to the similarities of their GDP growth rates. Thirdly, a distance-based method is applied in order to analyse a temporal dynamics of GDP per capita synchronisation in the whole EU-28. Bayesian Shrinkage Estimators reveal rather the heterogeneity of the speed of convergence. The distance-based approach concludes that the overall convergence in the EU-28 prevailed up to 2007, but the crisis caused a divergence of their GDP per capita. Despite a recent temporal convergence, average distances between GDP per capita of all EU-28 countries remain important and are still higher compared to the pre-crisis period.
... for all i and q = 2...4, whereσ 2 qt is the sample variance of log x q,it in period t. 28 We proceeded in exactly the same way for the income variable log y it : ...
... In the appendix, I try reducing the endogeneity of the instruments by using their most distant lags only or by implementing Arellano and Bond (1991)'s first-difference GMM, where only the instruments in level are used; the Sargan and Hansen tests remain highly significant. 62 See also Islam (2001) and Hauk and Wacziarg (2009), who cast doubt on GMM as the standard estimation technique in the growth literature. They compare various dynamic panel estimators on simulated growth data and find that least-square estimators (e.g. ...
Does inequality matter for economic growth? The question is controversial. A recent trend of the literature has suggested viewing income inequality as a composite measure of two more relevant kinds of inequalities: inequality of opportunity and inequality of effort. These inequalities are thought to be more relevant than income inequality, both normatively and in terms of their effects on medium to long run growth. Inequality of opportunity is unfair in that it is derived from factors over which individuals have no control, and is thought to be detrimental to economic growth. Inequality of effort, on the other hand, is fair in that it is a result of individuals’ responsible choices, and is expected to be beneficial for economic growth. This hypothesis, according to which inequality matters for growth to the extent that it reflects equity issues, offers a promising explanation to the mixed evidence found by empirical studies about the effect of inequality on growth: the effect of income inequality would depend on which of its ethical component dominates. The hypothesis has received strong evidence from Marrero and Rodríguez (2013) in the US. However, it needs to be validated on different samples, especially in developing countries, in order to gain legitimacy. This is the purpose of this paper, which investigates the effect of inequality of opportunity and inequality of effort on Brazilian municipalities’ growth from 1980 to 2010 using the decennial national censuses ran by the IBGE. Results from various OLS and FE regressions provide strong support in favor of a negative effect of inequality of opportunity and a positive effect of inequality of effort.
... i t i t yy ) and the initial income per capita. Here, we follow the model of Nazrul Islam (1995Islam ( , 2000 who specified the model of absolute convergence in dynamics for panel data by: log( ) it y . In this case, the poor economies tend to grow faster than the rich ones, which imply an absolute convergence. ...
The catching-up process in the European Union (EU-28) remains a very topical issue as the global economic crisis has impacted the growth rates of each member country. The objective of this paper is to test a real convergence between the European countries. Firstly, the β-convergence hypothesis is tested for the heterogeneous panel data using Bayesian Shrinkage Estimators. Secondly, we estimate the cluster dendrogram. Thirdly, a distance-based method is applied in order to reveal a temporal dynamics of real convergence in the whole EU-28. The panel data model confirms the heterogeneity among the speeds of convergence. The distance-based approach concludes that convergence prevailed up to 2007, but the crisis caused a divergence of GDP per capita. Despite a recent temporal convergence, distances between GDP per capita remain still important. Thus, the results confirm the idea of " variable-speed Europe " of real convergence.
... Islam, 1995;Durlauf, Quah, 1999. Dla porównania, wyniki symulacji metodą Monte Carlo wskazują, że dla małych prób estymator GMM charakteryzuje się istotnym obciążeniem (Islam, 2001). w tym samym stopniu (Rassekh, 1998, s. 96). ...
The study presents not only the definition of income convergence and its classification
according to the nature of the analysed economic phenomenon, research method
used, extent of the study and the place of income convergence in selected growth theories,
but also the characteristics of various methods used to explore income convergence
as well as a comparative analysis of the empirical results obtained.
The analyses is concentrated on the EU member states which comprise a relatively
homogenous group of objects in terms of socio-economic development, characterised
by similar propensities to save, fertility, capital depreciation rates and levels of technology.
The time span covers the years 1995–2006. The data used for the purpose of the
study has been sourced mostly from The Conference Board and Groningen Growth and
Development Centre, Total Economy Database, supported by the World Oeconomic
Outlook and Eurostat’s databases.
The empirical evidence leads to an unambiguous conclusion that in the years
1993–2006 the EU member states experienced an absolute β income convergence. The
rate of convergence of the real GDP per capita was equal to 3.83% annually and the
convergence half-time approached 18 years. It is noteworthy, however, that the economic
transition in new members states and their integration with the “old” EU member
states – which involved more freedom of mobility of new technology, capital, goods,
products and labour – not only affected the higher evaluation of convergence rate as
compared to the results obtained by other authors but may also significantly shorten the
full time period of convergence. This conclusion seems to be supported by the results of
the marginal horizontal β convergence, both in its forward and backward approach.
The results of the conditional β convergence based on the equation estimated for
a selected subgroup of less developed countries (including the new EU member states
next to Portugal, Spain and Greece), where the competitive effect calculated by means
of the shift-share method was used as an additional explanatory variable, indicate
nearly the same income convergence rate as the results obtained for this group of countries
by means of the absolute β income convergence analysis. Different results, however,
are delivered by the conditional convergence approach when the differences between
individual economies related to their ITC expenditure are taken into account. In
that case an increase in the convergence half-time is observed.
The hypothesis about the real convergence in the EU member states is supported
also by the results generated by the absolute γ income convergence analysis and by the
dispersion and concentration measures. The use of the methods of analysis of stochastic
processes provides yet other evidence to support the existing view about a relatively
significant ambiguity of those results both in relation to the method and the type of the
unit root test used. It should be noted, however, that the evidence supporting the existence
of the convergence phenomenon is provided more frequently by the KwiatINCOME
CONVERGENCE IN EU MEMBER STATES ECONOMETRIC ANALYSIS
268
kowski-Phillips-Schmidt-Shin test rather than the Dickey-Fuller and the augmented
Dickey-Fuller tests.
The evaluation of the real convergence by means of the vertical marginal β convergence
method reveals that Estonia and Latvia have contributed to the income convergence
in the European Union whereas Poland and Lithuania have had a negative
impact on the speed of convergence. In the case of the vertical marginal σ convergence,
negative impact was reported for Portugal, Latvia, Poland and Slovenia in 1993.
In 2006 this group excluded Latvia, and Portugal’s impact was much weaker. It should
also be mentioned that in all those four approaches using the Mt, t + 1 measure, real convergence
in the EU member states was observed for the entire 1993–2006 period.
The obtained results may serve not only as indications for policy makers in relation
to the cohesion of the EU member states but also as a sound foundation for further
efforts concerning income convergence oriented at the diagnosis of the effects of the
crises and recessions on the convergence of socio-economic development in the EU
member states. It is also noteworthy to observe that the findings of the income convergence
analysis, particularly those of methodological character, may become indications
for the enhancement of the existing growth models and may set the development directions
for new research methods to be used in the analysis of this phenomenon.
... Ahora utilizamos el modelo de convergencia condicional cuya escritura dinámica ha sido propuesta por Islam (2000): ...
... Tables 5 and 6 present the corresponding estimations for both the one-step and the two-step GMM estimator. While the two-step estimator should theoretically be preferred, experimental evidence reports problems concerning a downward bias in its estimates of the standard errors (Arellano and Bond, 1991; Islam, 2000). Although in our case, both procedures appear to produce very similar outcomes, we nevertheless follow the recommendation made by Arellano and Bond, and use only the one-step results for inferences regarding the coefficients. ...
The paper aims for an empirical validation of the impact of industrial structure on aggregate income and growth. Various mechanisms for the linkage between meso-structure and macro-performance are identified: the income elasticity of demand, the structural bonus versus burden hypotheses, differential propensities towards entrepreneurial discovery, and producer or user related spillovers. After discussing detailed results from conventional shift-share analysis, dynamic panel estimations are applied to a standard growth model augmented by structural variables. Based on data for 28 OECD countries, the results confirm that industrial structure has been a significant determinant of macroeconomic development and growth in the 1990s.
... Nezrul Islam (1995Islam ( , 2000) proposes to test the following specification for the model of conditional convergence in panel data: ...
This paper aims to illustrate the impact of financial variables on the process of convergence between selected European Union countries and the Balkan countries. Following a delay in the realization of structural changes resulting from the historical legacy and circumstances in which the transition process took place, Balkan countries began essential reforms in their financial systems at the end of 1990s. This included the adoption of concrete measures directed towards the growth and increase of the financial sector efficiency. Given this we use panel data over the period 1999-2007 for a sample of 21 countries, to test the convergences hypothesis by the Bayesian iterative estimation method. Here two financial variables are introduced to control the differences in steady-state. Our empirical results sustain the importance of the domestic credit and the market capitalization in the catching-up process by a significant increase in the speed of convergence.
... To the extent that some of the regressors (for example, the growth rate of the physical capital stock) are likely to be jointly determined with the output growth rate, an endogenity bias may arise. However, given the poor small sample performance of the IV and GMM estimators in estimating growth regressions, it is not clear whether the use of these frameworks is worthwhile: the reduction in the endogenity bias may well be outweighed by the introduction of a small sample bias (Islam, 2000). The issue of endogenity of regressors in estimating growth regressions is not yet comprehensively addressed (Temple, 1999; Islam, 2003). ...
This paper studies the effects of industrialization on growth and convergence dynamics in Italy. We propose a semi-parametric procedure linking growth theory with density-type studies on convergence. Our analysis suggests that absolute convergence processes do not necessarily exclude dynamics of club convergence and vice-versa. Growth and convergence processes should be thought as the composition of contrasting economic influences, in which industrialization matters. Our results suggest that the convergence process between the South and the North of Italy ended due to a slowing down both in the industrialization process in Southern regions and in the steady state convergence process of all regional economies.
... The problem with this line of research is that Monte Carlo studies using Summers-Heston data shave shown large small sample bias in the estimates of β . The concern is that any reduction in the endogeneity bias could be outweighed by the introduction of small sample bias (see Islam, 2000 for more details). ...
The issue of convergence has been hotly debated since the mid 1980s. Only recently certain consensus has arisen around some of the most fundamental issues. It seems hardly surprising, then, to find a large variation in how those issues are taught to undergraduates. This paper is an attempt at clarifying the different concepts of convergence, and their relation to both the neoclassical model of growth and available cross-country evidence. Evidence on some of the contradictory ways in which the issues are taught is provided, a simple way to teach the relevant concepts of convergence by means of numerical examples is presented, and relevant examples drawn from historical evidence are shown. Also, the implications of the neoclassical growth model are presented in a slightly different, but clearer way, and some of the difficulties to interpret the cross-country empirical evidence are reviewed.
... for all i and q = 2...4, whereσ 2 qt is the sample variance of log x q,it in period t. 28 We proceeded in exactly the same way for the income variable log y it : ...
Using Monte Carlo simulations, this paper evaluates the bias properties of common estimators used in growth regressions derived from the Solow model. We explicitly allow for measurement error in the right-hand side variables, as well as country-specific effects that are correlated with the regressors. Our results suggest that using 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. The fixed-effects estimator and the Arellano-Bond estimator greatly overstate the speed of convergence under a wide variety of assumptions concerning the type and extent of measurement error, while between understates it somewhat. Finally, fixed effects and Arellano-Bond bias towards zero the slope estimates on the human and physical capital accumulation variables.
Several empirical studies explore the effects of regulatory enforcement on environmental behavior and performance. Within this literature, extremely little empirical research examines the role of fairness, which we interpret broadly to include multiple dimensions, e. g. similar treatment of similarly situated regulated entities. Our study empirically examines the effect of perceived enforcement fairness on the extent of compliance with wastewater limits imposed on chemical manufacturing facilities regulated under the Clean Water Act. Our study also explores the influence of perceived fairness on the effectiveness of enforcement efforts – government inspections and enforcement actions – at inducing better compliance. For our analysis, we use a subjective measure of the degree of “fair treatment” of regulated facilities by environmental regulators, as perceived by facilities and reported as survey responses. Results reveal that a more (perceived) fair enforcement approach raises compliance, but only under limited enforcement conditions; in most instances, perceived more fair enforcement lowers compliance. As important, results show that greater perceived fairness improves the effectiveness of federal inspections and informal enforcement, but undermines the effectiveness of state inspections and formal non-penalty enforcement.
The growth literature has identified four channels of transmission by which the abundance of natural resources can negatively affect economic growth. In this article, I suggest ideology as a fifth transmission channel. To test this hypothesis, I exploit the geography of Bolivia whose western regions have natural resources that differ considerably from its eastern regions. I find that regions with predominantly extractive natural resources tend to choose redistributive and interventionist rather than laissez-faire policies. Additionally, I identify two effects on growth depending upon the type of natural resource that a region possesses in abundance.
Objective. The purpose of the study is to estimate the impact of the modern monetary policy of the Bank of Russia on economic growth in the country. Methodology. The authors applied correlation and regression analysis, which allows constructing a regression equation with the values of independent variables (factors). Findings. The paper analyzes the main indicators characterizing the effectiveness of monetary policy implementation in Russia in 2000-2016. The undertaken correlation and regression analysis of the impact of monetary policy tools on economic growth in Russia according to empirical evidence (2006-2016) revealed the dependence of GDP on funds of credit institutions on correspondent accounts at the Bank of Russia. Application area. The obtained results can be used by the Central Bank of Russia in the development of the main directions of monetary policy and by the Ministry of Economic Development of the Russian Federation in preparing a forecast of socioeconomic development in the long run.
The study examines the influence of macroeconomic variables interest rate, economic growth, inflation and exchange rate on stock market performance in G7 countries over the period 1980-2015. We employ the estimating technique of Pooled mean group estimation (PMG) to reveal the short run and long-run impact of explanatory variables on the dependent variable. The study established that long run interest rate and GDP affect the stock market performance positively. Moreover, in short run, all the variables are found to be statistically insignificant. The outcomes suggest formulating comprehensive and effective policies to ensure a stable macroeconomic environment for G7 countries.
This paper analyses the effects of inflation and its volatility on Gross Domestic Product (GDP) per capita for the regions of Bolivia over the period 1989–2011. Results show the existence of a positive relationship with inflation but a negative one with respect to volatility. Moreover, we identify a threshold for volatility and inflation. Our conclusions imply that economic agents are more tolerant to relatively high levels of inflation than to high inflation volatility.
The aim of this article is to prove the hypothesis of absolute and conditional convergence of private (per capita) income for the 32 states of the Mexican Republic, using the econometric technique of panel data. On the basis of iterative Bayesian estimates we demonstrate that, contrary to the standard empirical results, Mexican states do not converge at the same velocity. Besides, the distribution of the absolute convergence velocity allows us to show, in terms of the dynamics of development, the analogies which may exist among some states and the heterogeneity which may exist among others. Thus, we are able to group those states into homogeneous regions in terms of their respective convergence rates.
Efficiency of full information approaches for the estimation of a system of equations is well known for large samples. However, less is known about the estimators’ small sample properties. This is especially true for panel data settings, although the latter are by now a standard approach to model static and dynamic economic processes. Here extensive finite sample evidence based on Monte Carlo (MC) simulations is solely given for the single equation case. In this paper, we provide new results for the multiple equation case by testing multiple equations extensions for the standard fixed effects type models (FEM) as well as familiar IV/GMM-style estimators, which have recently been proposed in the literature. Building upon the MC simulation evidence, the paper then applies dynamic simultaneous equation estimation to a small-scale economic model for regional output growth, private and public capital investment among German states. The model is used to identify the likely two-way effects among the variables and to perform policy analysis. The latter is done by augmenting the baseline model by variables measuring interregional spillover effects from public capital as well as transfer payments from regional equalization schemes. We find positive but insignificant effects from interregional spillovers in transport infrastructure, while spillovers from science infrastructure even tend to be negative. The latter result is likely to originate from specific locational advantages of science infrastructure, which allows regions to poach production factors from their neighborhood. For regional equalization transfers we find mixed results, crucially depending on the specific policy programme.
The current financial crisis is likely to reduce investment in Research and
Development (R&D). Yet endogenous growth models underline the importance of the
technological capital in the growth process: the accumulation of knowledge through
R&D activities is one of the factors of self-sustained growth because it generates
positive external effects. This paper aims to show the impact of the investment in R&D
on the process of convergence between selected OECD countries. Using panel data
over the period 1983-2004 for a sample of 16 countries, we test the convergence’s
hypothesis by the Bayesian iterative estimation method; a technological variable is
introduced to control the differences in steady-state. Our empirical results validate the
effect of expenditure in R&D in the catching-up process. Consequently, a decrease in
these investments would be detrimental to growth.
The issue of convergence has been hotly debated since the mid 1980s. Only recently certain consensus has arisen around some of the most fundamental issues. It seems hardly surprising, then, to find a large variation in how those issues are taught to undergraduates. This paper is an attempt at clarifying the different concepts of convergence, and their relation to both the neoclassical model of growth and available cross-country evidence. Evidence on some of the contradictory ways in which the issues are taught is provided, a simple way to teach the relevant concepts of convergence by means of numerical examples is presented, the implications of the neoclassical growth model are presented in a new way, and the difficulties to interpret the cross-country empirical evidence are reviewed.
Purpose
The purpose of this paper is to examine empirically the sources of economic growth based on an augmented Mankiw, Romer, and Weil's model which considers human capital in the forms of both health and education for a group of East Asian economies including China.
Design/methodology/approach
Empirical results are based on the analysis of a panel dataset from 1961 to 2007. Sub‐sample estimation for the post‐1997 Asian Financial Crisis period is also considered for comparison purposes.
Findings
The impact of the stock of health and education on economic growth is statistically significant for both the whole sample and sub‐sample period. However, the impact of investment in education on economic growth is a little “fragile”. The statistical results show that the statistical impact of health on economic growth is stronger than that of education. It seems that it is more plausible for the policymakers in East Asia to invest more in health than educational human capital.
Originality/value
This paper is one of the first empirical studies to analyze the effect of human capital in the form of both health and education on economic growth in East Asia.
Using panel data from 118 developing countries in 1971–2000, this paper explores the channels linking social spending, human capital, and growth and compares the effects of alternative economic policy interventions. With separate modeling for education and health capital, explicit control for governance, and incorporation of nonlinearity, the paper finds that both education and health spending have a positive and significant impact on education and health capital, and thus support higher growth. Also, other policy interventions, such as improving governance and taming inflation, can achieve similar results. Hence, higher spending alone is likely insufficient to achieve the Millennium Development Goals.
This paper deals with maximum likelihood estimation with singular systems of equations. We propose to estimate the singular systems by convoluted-likelihood functions. The consistency and asymptotic normality of the estimator are also established.
Designing appropriate environmental and energy policies, in order to meet the Kyoto protocol's carbon dioxide (CO2) reduction targets in the European Union (EU), requires a detailed examination and thorough understanding of CO2 emission trends across the EU member states. This paper investigates whether CO2 emissions have converged across 22 European countries over the 1971 to 2006 period. The Bayesian shrinkage estimation method is employed to do this work and the results reveal the following: first, the hypothesis of absolute convergence in per capita CO2 emissions is supported and a slight upward convergence is observed; second, the fact that countries differ considerably in both their speed of convergence and volatility in emissions makes it possible to identify different groups of countries; third, the results with respect to convergence do not vary much once the share of industry in GDP is accounted for in a conditional convergence analysis. However, a decreasing share of industry in GDP seems to contribute to a decline in per capita emissions. These findings may carry important implications for both national and EU environmental policies.
Based on panel data for the manufacturing and retail industries in a sample of 2645 US counties spanning the last two decades of the twentieth century, we find that firm size matters for industry growth. Contrary to previous research, our results suggest that there is a positive linkage between the average size of manufacturing and retail firms on the one hand and industry growth on the other. Our results are consistent with a Schumpeterian growth model, where bigger firms are needed to carry out effective Research and Development, leading to higher growth. The results, therefore, do not support the idea of small firm gardening as a local development strategy.
This paper analyzes the relationship between fiscal adjustment and real GDP growth in a panel of 26 transition economies during 1992–2001. It finds a positive and statistically significant relationship between fiscal adjustment and real GDP growth that is robust to different model specifications and estimation methods when initial conditions, stabilization policies, and institutions are controlled for. Fiscal adjustment promotes growth in both the short and the long run in countries that have not yet attained macroeconomic stability. However, for countries that have achieved credible stabilization, additional fiscal adjustment is unlikely to yield large benefits in growth.
El propósito de este artículo es de probar la hipótesis de convergencia absoluta y condicional del ingreso per-cápita, para los 32 estados que conforman a la republica mexicana, a partir de la aplicación de la técnica econométrica de estimación con datos de panel. Gracias a la obtención de estimadores bayesianos iterativos mostramos que contrariamente a los resultados empíricos estándar los estados mexicanos no convergen a la misma velocidad. Además que las distribuciones de las velocidades de convergencia absoluta nos permiten de poner en evidencia, en términos de la dinámica del crecimiento, las analogías que pudieran existir entre algunos estados y la heterogeneidad que pudiera existir entre otros. De este modo podemos agrupar a los estados en regiones que sean homogéneas en términos de sus respectivas tasas de convergencia.
Abstract:
The aim of this article is to prove the hypothesis of absolute and conditional convergence of private (per capita) income, for the 32 states of the Mexican Republic, from the use of the econometric technique of calculation with panel data. Due to the obtainment of iterative Bayesian estimates we demonstrate that, contrary to the standard empirical results, Mexican states do not converge at the same velocity. Besides, the distribution of the absolute convergence velocity allows us to give away, in terms of the dynamics of development, the analogies which may exist among some states and the heterogeneity which may exist among others. Thus, we are able to group those states in regions homogeneous in terms of their respective convergence rates.
Submitted to the Graduate School of Business. Copyright by the author. Thesis (Ph. D.)--Stanford University, 2005.
This paper focuses on the influence of military spending on European economic growth. The estimated regressions are based on Barro’s (19912.
Barro , R.J. 1991. Economic growth in a cross section of countries. Quarterly Journal of Economics, 106: 407–443. [CrossRef], [Web of Science ®]View all references) growth model, which controls for economic institutional variation across countries. The cross‐section and panel data analyses cover the period 1960–2000. The empirical findings indicate that military spending has an overall net negative influence on economic growth. Furthermore, the magnitude of this negative impact tends to increase over time, as cross‐section regression results indicate. Given the development of a Common European Security and Defence Policy (CESDP), these findings suggest that enhanced defence spending may hinder European economic growth.
Recent advances in the growth literature have proposed that difficult-to-quantify concepts such as social capital may play an important role in explaining the degree of persistent income disparity that is observed among countries. Other recently explored possibilities include institutional mechanisms which generate barriers to aggregate production. An important limitation for empirical work in this area stems from the fact that it is difficult to distinguish sources of heterogeneity when direct observations are not available. In this study, we show how developments in the analysis of nonstationary panels can aid in this endeavor. In contrast to traditional dynamic panel data analysis, this approach focuses explicitly on low-frequency behavior. Under relatively mild assumptions, the approach can be used to infer properties of aggregate production which are robust to the presence of large classes of unobserved features. In this framework we are able to estimate and test the distribution of production function parameters that would be required in order to generate conditional forecast convergence of per capita incomes even when some of the key factors required to explain growth are unobserved. The results indicate that in order to fully explain the observed persistence in the disparity of per capita incomes, the manner in which unobserved mechanisms influence production must go beyond merely accounting for differences in the trending behavior of aggregate productivity. Specifically, if such mechanisms are to be successful empirically, they must also be able to account for cross-country heterogeneity in steady-state capital shares. This adds to a growing literature that provides support for models with multiple production regimes. Copyright © 2007 John Wiley & Sons, Ltd.
Recent research shows that productivity differences are more important than differences in accumulation rates in explaining per capita income differences across countries. So far static differences in productivity have been mainly computed and analyzed in large samples of countries. This paper extends the research by focusing on productivity dynamics . It uses the panel approach to compute productivity indices for a large sample of countries for two time periods, namely an initial period of 1960-75 and a subsequent period of 1975-90. This allows computation of ordinal and cardinal changes in productivity between the two periods. The results show considerable variation in productivity dynamics across countries. The task ahead is to find out what accounts for the observed dynamics. Copyright 2003 by the International Association for Research in Income and Wealth.
Health spending, the largest component of provincial government spending, has risen significantly over the past decade. It has been asserted that larger health expenditures have caused provincial governments to spend less on other types of government services. Using a panel of province-level data for the period 1988/89 to 2003/04, this study provides a test of the hypothesis that health spending has crowded out other types of spending. The results indicate that, for the period studied, there is no evidence that increased provincial government health expenditures resulted in lower levels of spending on other categories of government provided goods and services.
In this chapter we study GMM estimation of linear panel data models. Several different types of models are considered, including the linear regression model with strictly or weakly exogenous regressors, the simultaneous regression model, and a dynamic linear model containing a lagged dependent variable as a regressor. In each case, different assumptions about the exogeneity of the explanatory variables generate different sets of moment conditions that can be used in estimation. This chapter lists the relevant sets of moment conditions and gives some results on simple ways in which they can be imposed. In particular, attention is paid to the question of under what circumstances the efficient GMM estimator takes the form of an instrumental variables estimator.
This paper is concerned with the estimation of a dynamic panel data model with individual fixed effects and a linear trend term with heterogenous coefficients. We expand the available methods for the standard dynamic panel data model to this case, and discuss modified OLS, IV and GMM methods. We hint at the potential of LIML estimation to reduce small sample bias. Although the methods pertain to a more complex model than the standard dynamic model, the simplicity of the various approaches can also be brought to bear on the standard model.
This paper implements a panel data approach of the Solow model to study the phenomenon of growth convergence for 22 OECD countries. It shows that although the derived estimable Solow model is probably underspecified from an econometric point of view, it is still possible to conclude that there is a likely convergence to a steady state rate of about 2-4%.
This paper examines whether the Solow growth model is consistent with the international variation in the standard of living.
It shows that an augmented Solow model that includes accumulation of human as well as physical capital provides an excellent
description of the cross-country data. The paper also examines the implications of the Solow model for convergence in standards
of living, that is, for whether poor countries tend to grow faster than rich countries. The evidence indicates that, holding
population growth and capital accumulation constant, countries converge at about the rate the augmented Solow model predicts.
The Penn World Table displays a set of national accounts economic time series covering many countries. Its expenditure entries
are denominated in a common set of prices in a common currency so that real quantity comparisons can be made, both between countries and over time. It also provides information about relative prices
within and between countries, as well as demographic data and capital stock estimates. This updated, revised, and expanded
Mark 5 version of the table includes more countries, years, and variables of interest to economic researchers. The Table is
available on personal computer diskettes and through BITNET.
It is well known that the usual techniques for estimating random and fixed effects panel data models are inconsistent in the dynamic setting. As a consequence, numerous consistent estimators have been proposed in the literature. However, all such estimators rely on certain well defined assumption, which in practice my be violated.The purpose of this paper is to ascertain how robust the available estimators are to such misspecifications, thus providing guidance to applied researcher as to an appropriate choice of estimator in such situation.
In this paper two new estimators are offred (one each for the fixed random effects specifications), and small sample performance compared with that of all the existing estimators.
This paper provides an overvie of topics in nonstationary panels: panel unit root tests, panel cointegration tests, and estimation of panel cointegration models. In addition it surveys recent developments in dynamic panel data models.
There are two sources of inconsistency in existing cross-country empirical work on growth: correlated individual effects and endogenous explanatory variables. We estimate a variety of cross-country growth regressions using a generalized method of moments estimator that eliminates both problems. In one application, we find that per capita incomes converge to their steady-state levels at a rate of approximately 10 percent per year. This result stands in sharp contrast to the current consensus, which places the convergence rate at 2 percent. We discuss the theoretical implications of this finding. In another application, we perform a test of the Solow model. Again, contrary to prior results, we reject both the standard and the augmented version of the model. Copyright 1996 by Kluwer Academic Publishers
Recent empirical studies have examined the determinants of economic growth using country-average (cross-sectional) data. By contrast, this paper employs a technique for using a panel of cross-sectional and time series data for 98 countries over the 1960-85 period to determine the quantitative importance for economic growth of both country-specific and time-varying factors such as human capital, public investment, and outward-oriented trade policies. The empirical results support the view that these factors exert a positive and significant influence on growth. They also provide estimates of the speed at which the gap between the real per capita incomes of rich and poor countries is likely to be reduced over the longer term.
New results on the exact small sample distribution of the instrumental variable estimator are presented by studying an important special case. The exact closed forms for the probability density and cumulative distribution functions are given. There are a number of surprising findings. The small sample distribution is bimodal. with a point of zero probability mass. As the asymptotic variance grows large, the true distribution becomes concentrated around this point of zero mass. The central tendency of the estimator may be closer to the biased least squares estimator than it is to the true parameter value. The first and second moments of the IV estimator are both infinite. In the case in which least squares is biased upwards, and most of the mass of the IV estimator lies to the right of the true parameter, the mean of the IV estimator is infinitely negative. The difference between the true distribution and the normal asymptotic approximation depends on the ratio of the asymptotic variance to a parameter related to the correlation between the regressor and the regression, error. In particular, when the instrument is poorly correlated with the regressor, the asymptotic approximation to the distribution of the instrumental variable estimator will not be very accurate.
The continued-fraction technique has been widely used in solid state physics. It provides a useful theoretical framework for calculating the densities of states (of electrons, phonons...) mainly in disordered systems. The continued-fraction coefficients are computed either from the moments of the density of states or more directly by the recursion method. The former has the advantage of the linearity upon the density of states and the latter is numerically optimized. The generalized-moments method is an interpolation between both methods that, if suitably used, combines their advantages. A version of the method could be seen as a perturbation expansion from the Bethe lattice. The asymptotic limits of the continued-fraction coefficients are shown to be accurately determined from the generalized-moments associated to the recursion method. Computations of electronic densities of states illustrate the method.
I examine the empirical performance of instrumental variables estimators with predetermined instruments in an application to life-cycle labor supply under uncertainty. The estimators studied are two-stage least squares, generalized method-of-moments (GMM), forward filter, independently weighted GMM, and split-sample instrumental variables. I compare the bias/efficiency trade-off for the estimators using bootstrap algorithms suggested by Freedman and by Brown and Newey. Results indicate that the downward bias in GMM is quite severe as the number of moment conditions expands, outweighing the gains in efficiency. The forward-filter estimator, however, has lower bias and is more efficient than two-stage least squares.
In this paper, we consider two basic aspects of demand analysis, with application to the demand for natural gas in the residential and commercial market. The more fundamental one consists in the formulation of a demand function for commodities--such as natural gas--whose consumption is technologically related to the stock of appliances. We believe that in such markets, the behavior of the consumer can be described best in terms of a dynamic mechanism. Related to this is the more specific problem of estimating the parameters of the demand function, when the demand model is cast in dynamic terms and when observations are drawn from a time series of cross sections. Accordingly, this paper is centered around these two major themes, although, as the title suggests, the emphasis is placed on the second one. In Section 1, we present the theoretical formulation of the dynamic model for gas. In Section 2, the results of the estimation of the gas model by ordinary least squares methods are presented. These results, together with more fundamental theoretical considerations, suggest a different approach. The essence of this approach, which is not restricted to the gas model, is discussed in Section 3, while two alternative procedures for estimating the coefficients of the dynamic model in the light of this new approach are proposed in Section 4. It is subsequently shown that the application of these procedures to the gas data produces results that are reasonable on the basis of a priori theoretical considerations.
When a model for panel data includes lagged dependent explanatory variables, then the habitual estimation procedures are asymptotically valid only when the number of observations in the time dimension (T) gets large. Usually, however, such datasets have substantial sample size in the cross-section dimension (N), whereas T is often a single-digit number. Results on the asymptotic bias (N → ∞) in this situation have been published a decade ago, but, hence far, analytic small sample assessments of the actual bias have not been presented. Here we derive a formula for the bias of the Least-Squares Dummy Variable (LSDV) estimator which has a approximation error. In a simulation study this is found to be remarkably accurate. Due to the small variance of the LSDV estimator, which is usually much smaller than the variance of consistent (Generalized) Method of Moments estimators, a very efficient procedure results when we remove the bias from the LSDV estimator. The simulations contain results for a particular operational corrected LSDV estimation procedure which in many situations proves to be (much) more efficient than various instrumental variable type estimators.
The paper examines the relationship between heterogeneity bias and strict exogeneity in a distributed lag regression of y on x. The relationship is very strong when x is continuous, weaker when x is discrete, and non-existent as the order of the distributed lag becomes infinite. The individual specific random variables introduce nonlinearity and heteroskedasticity; so the paper provides an appropriate framework for the estimation of multivariate linear predictors. Restrictions are imposed using a minimum distance estimator. It is generally more efficient than the conventional estimators such as quasi-maximum likelihood. There are computationally simple generalizations of two- and three-stage least squares that achieve this efficiency gain. Some of these ideas are illustrated using the sample of Young Men in the National Longitudinal Survey. The paper reports regressions on the leads and lags of variables measuring union coverage, SMSA, and region. The results indicate that the leads and lags could have been generated just by a random intercept. This gives some support for analysis of covariance type estimates; these estimates indicate a substantial heterogeneity bias in the union, SMSA, and region coefficients.
This chapter discusses the models that are static conditional on a latent variable. The panel aspect of the data has been primarily used to control for the latent variable. Much work needs to be done on models that incorporate uncertainty and interesting dynamics. Exploiting the martingale implications of time additive utility seems fruitful. There is, however, a potentially important distinction between time averages and cross-section averages. A time average of forecast errors over T periods should converge to zero as T→ ∞. But an average of forecast errors across N individuals surely need not converge to zero as N→ ∞; there is a common component in those errors, due to economy-wide innovations. The same point applies when considering covariances of forecast errors with variables that are in the agent's information sets. If those conditioning variables are discrete, one can think of averaging over subsets of the forecast errors; as T→ ∞, these averages should converge to zero but not necessarily as N → ∞.
Research on economic growth has exploded in the past decade. Hundreds of empirical studies on economic growth across countries have highlighted the correlation between growth and a variety of variables. Determinants of Economic Growth, based on Robert Barro's Lionel Robbins Memorial Lectures, delivered at the London School of Economics in February 1996, summarizes this important literature. The book contains three essays. The first is a survey of the research on the determinants of long-run growth through the estimation of panels of cross-country data. The second essay details the interplay between growth and political freedom or democracy and finds some evidence of a nonlinear relationship. At low levels of political rights, an expansion of rights stimulates growth; however, once a moderate level of democracy has been obtained, a further expansion of rights reduces growth. The final essay looks at the connection between inflation and economic growth. Its basic finding is that higher inflation goes along with a lower rate of economic growth.
Panel data models have become increasingly popular among applied researchers due to their heightened capacity for capturing the complexity of human behavior as compared to cross-sectional or time series data models. As a consequence, richer panel data sets also have become increasingly available. This 2003 second edition is a substantial revision of the highly successful first edition of 1986. Advances in panel data research are presented in a rigorous and accessible manner and are carefully integrated with the older material. The thorough discussion of theory and the judicious use of empirical examples make this book useful to graduate students and advanced researchers in economics, business, sociology, political science, etc. Other specific revisions include the introduction of the notion of strict exogeneity with estimators presented in a generalized method of moments framework, the notion of incidental parameters, more intuitive explanations of pairwise trimming, and discussion of sample selection dynamic panel models.
This article develops a framework for efficient IV estimators of random effects models with information in levels which can accommodate predetermined variables. Our formulation clarifies the relationship between the existing estimators and the role of transformations in panel data models. We characterize the valid transformations for relevant models and show that optimal estimators are invariant to the transformation used to remove individual effects. We present an alternative transformation for models with predetermined instruments which preserves the orthogonality among the errors. Finally, we consider models with predetermined variables that have constant correlation with the effects and illustrate their importance with simulations.
This paper presents a statistical analysis of time series regression models for longitudinal data with and without lagged dependent variables under a variety of assumptions about the initial conditions of the processes being analyzed. The analysis demonstrates how the asymptotic properties of estimators of longitudinal models are critically dependent on the manner in which samples become large: by expanding the number of observations per person, holding the number of people fixed, or by expanding the number of persons, holding the number of observations per person fixed. The paper demonstrates which parameters can and cannot be identified from data produced by different sampling plans.
I consider estimation of the autoregressive panel model with fixed effects y it =c i +βy i,t-1 +ε it . I investigate the estimation method developed by R. Blundell and S. Bond [ibid. 87, No. 1, 115-143 (1998; see the preceding entry, Zbl 0943.62112)] which makes use of the stationarity of the initial levels. I do it by numerically comparing the semiparametric information bounds for the case that incorporates the stationarity of the initial condition and for the case that does not. It is found that the efficiency gain can be substantial.
In this paper we consider a dynamic model for panel data. We show that, under standard assumptions, there are more moment conditions than are currently exploited in the literature. Some of these are linear, but others are quadratic, so that nonlinear GMM is required. We also show that exogenous regressors generate a larger number of relevant moment conditions in a dynamic model than they would in a static model.
Estimation of the dynamic error components model is considered using two alternative linear estimators that are designed to improve the properties of the standard first-differenced GMM estimator. Both estimators require restrictions on the initial conditions process. Asymptotic efficiency comparisons and Monte Carlo simulations for the simple AR(1) model demonstrate the dramatic improvement in performance of the proposed estimators compared to the usual first-differenced GMM estimator, and compared to non-linear GMM. The importance of these results is illustrated in an application to the estimation of a labour demand model using company panel data.
This paper considers the estimation of dynamic models for panel data. It shows how to count and express the moment conditions implied by a variety of covariance restrictions. These conditions can be imposed in a GMM framework. Many of the moment conditions are nonlinear in the parameters. We derive a simple linearized estimator that is asymptotically as efficient as the nonlinear GMM estimator, and convenient tests of the validity of the nonlinear restrictions.
This paper presents specification tests that are applicable after estimating a dynamic model from panel data by the generalized
method of moments (GMM), and studies the practical performance of these procedures using both generated and real data. Our
GMM estimator optimally exploits all the linear moment restrictions that follow from the assumption of no serial correlation
in the errors, in an equation which contains individual effects, lagged dependent variables and no strictly exogenous variables.
We propose a test of serial correlation based on the GMM residuals and compare this with Sargan tests of over-identifying
restrictions and Hausman specification tests.
A key economic issue is whether poor countries or regions tend to grow faster than rich ones: are there automatic forces that lead to convergence over time in the levels of per capita income and product? The authors use the neoclassical growth model as a framework to study convergence across the forty-eight contiguous U.S. states. They exploit data on personal income since 1840 and on gross state product since 1963. The U.S. states provide clear evidence of convergence, but the findings can be reconciled quantitatively with the neoclassical model only if diminishing returns to capital set in very slowly. Copyright 1992 by University of Chicago Press.
This paper comments on recent developments in the literature on the econometric analysis of international growth and convergence. It notes that panel estimates of the neoclassical model, accommodating level effects for individual countries through heterogeneous intercepts, deal with some of the econometric difficulties arising in some of the earlier cross-sectional studies. But it notes that, in dynamic panels, heterogeneity in growth effects and in speeds of convergence renders this estimator inconsistent also. The pervasiveness of such heterogeneity is demonstrated in three samples of countries, and the effects of (incorrectly) imposing homogeneity on estimated parameters are illustrated and discussed. © 2000 the President and Fellows of Harvard College and the Massachusetts Institute of Technology
A panel data approach is advocated and implemented for studying growth convergence. The familiar equation for testing convergence
is reformulated as a dynamic panel data model, and different panel data estimators are used to estimate it. The main usefulness
of the panel approach lies in its ability to allow for differences in the aggregate production function across economies.
This leads to results that are significantly different from those obtained from single cross-country regressions. In the process
of identifying the individual “country effect,” we can also see the point where neoclassical growth empirics meets development
economics.
In this paper we consider the problem of making inference on a structural parameter in instrumental variables regression when the instruments are only weakly correlated with the endogenous explanatory variables. Adopting a local-to-zero assumption as in Staiger and Stock (1994) on the coefficients of the instruments in the first stage equation, the asymptotic distributions of various test statistics are derived under a limited information framework. We show that Wald-type test statistics are not pivotal, thus (1-a)*100% confidence intervals implied by those test statistics can have zero coverage probability if the standard asymptotic distribution theory is used. In contrast, the likelihood type test statistics are pivotal when the model is just identified, thus providing valid confidence intervals. Even the model is overidentified, we show that the distributions of the likelihood type test statistics are bounded above by a Chi-Square distribution with degrees of freedom given by the number of instruments. Hence, we can always invert the likelihood type test statistics to obtain valid, although conservative, confidence intervals. The confidence intervals obtained by using this bounding distribution are compared with those obtained by using the standard Chi-Square 1 asymptotic distribution and an alternative bounding distribution, a transformation of the distribution of the Wilks statistic, suggested by Dufour (1994) . Confidence intervals based on our Chi-Square bounding distribution are shown to be tighter than those based on the Wilks bounding distribution by Monte Carlo experiments.
The paper considers international per capita output and its growth using a panel of data for 102 countries between 1960 and 1989. It sets out an explicitly stochastic Solow growth model and shows that this has quite different properties from the standard approach where the output equation is obtained by adding an error term to the linearized solution of a deterministic Solow model. It examines the econometric properties of estimates of beta convergence as traditionally defined in the literature and shows that all these estimates are subject to substantial biases. Our empirical estimates clearly reflect the nature and the magnitude of these biases as predicted by econometric theory. Steady state growth rates differ significantly across countries and once this heterogeneity is allowed for the estimates of beta are substantially higher than the consensus in the literature. But they are very imprecisely estimated and difficult to interpret. The paper also discusses the economic implications of these results for sigma convergence. © 1997 John Wiley & Sons, Ltd.
We use a Monte Carlo approach to investigate the performance of several different methods designed to reduce the bias of the estimated coefficients for dynamic panel data models estimated with the longer, narrower panels typical of macro data. We find that the bias of the least squares dummy variable approach can be significant, even when the time dimension of the panel is as large as 30. For panels with small time dimensions, we find a corrected least squares dummy variable estimator to be the best choice. However, as the time dimension of the panel increases, the computationally simpler Anderson-Hsiao estimator performs equally well.
The author examines the empirical performance of instrumental variables estimators with predetermined instruments in an application to life-cycle labor supply under uncertainty. The estimators studied are two stage least squares, generalized method-of-moments (GMM), forward filter, independently weighted GMM, and split-sample instrumental variables. The author compares the bias/efficiency trade-off for the estimators using bootstrap algorithms suggested by D. A. Freedman (1984) and B. W. Brown and W. K. Newey (1995). Results indicate that the downward bias in GMM is quite severe as the number of moment conditions expands, outweighing the gains in efficiency. The forward filter estimator, however, has lower bias and is more efficient than two stage least squares.
The authors consider the problem of making asymptotically valid inference on structural parameters in instrumental variables regression with weak instruments. Using local-to-zero asymptotics, they derive the asymptotic distributions of likelihood ratio (LR) and Lagrange multiplier (LM) type statistics for testing simple hypotheses on structural parameters based on maximum likelihood and generalized methods of moments estimation methods. In contrast to the nonstandard limiting behavior of Wald statistics, the limiting distributions of certain LR and LM statistics are bounded by a chi-square distribution with degrees of freedom given by the number of instruments.
This paper develops asymptotic distribution theory for instrumental variables regression when the partial correlations between the instruments and the endogenous variables are weak, here modeled as local to zero. Asymptotic representation are provided for various statistics, including two-stage least squares and limited information maximum likelihood estimators, Wald statistics, and statistics testing overidentification and endogeneity. The asymptotic distributions provide good approximations to sampling distributions with ten-twenty observations per instrument. The theory suggests concrete guidelines for applied work, including using nonstandard methods for construction of confidence regions. These results are used to interpret J. D. Angrist and A. B. Krueger's (1991) estimates of the returns to education.
This paper considers the instrument variable and method-of-moments estimators, as generalizations of two-stage least squares and limited information maximum likelihood, of the coefficients of a single equation. Motivated by a simple principle, asymptotic distributions are derived based on a parameter sequence where both the number of instruments and the sample size increase. The approximations to the distributions provided by this sequence are more accurate than traditional ones. The instrument variable estimator appears to be inconsistent. The asymptotic covariance matrix of the method-of-moments estimator can be consistently estimated under the alternative parameter sequence. The resulting approximate confidence regions have exact levels very close to their nominal levels. Copyright 1994 by The Econometric Society.
I assess the role of wealth and systemic risk in explaining future asset returns. I show that the residuals of the trend relationship among asset wealth and human wealth predict both stock returns and government bond yields. Using data for a set of industrialized countries, I find that when the wealth-to-income ratio falls, investors demand a higher risk premium for stocks. As for government bond returns: (i) when they are seen as a component of asset wealth, investors react in the same manner; (ii) if, however, investors perceive the increase in government bond returns as signalling a future rise in taxes or a deterioration of public finances, then investors interpret the fall in the wealth-to-income ratio as a fall in future bond premia. Finally, I show that the occurrence of crises episodes (in particular, systemic crises) amplifies the transmission of housing market shocks to financial markets and the banking sector.
[eng] Transportation costs and monopoly location in presence of regional disparities. . This article aims at analysing the impact of the level of transportation costs on the location choice of a monopolist. We consider two asymmetric regions. The heterogeneity of space lies in both regional incomes and population sizes: the first region is endowed with wide income spreads allocated among few consumers whereas the second one is highly populated however not as wealthy. Among the results, we show that a low transportation costs induces the firm to exploit size effects through locating in the most populated region. Moreover, a small transport cost decrease may induce a net welfare loss, thus allowing for regional development policies which do not rely on inter-regional transportation infrastructures. cost decrease may induce a net welfare loss, thus allowing for regional development policies which do not rely on inter-regional transportation infrastructures. [fre] Cet article d�veloppe une statique comparative de l'impact de diff�rents sc�narios d'investissement (projet d'infrastructure conduisant � une baisse mod�r�e ou � une forte baisse du co�t de transport inter-r�gional) sur le choix de localisation d'une entreprise en situation de monopole, au sein d'un espace int�gr� compos� de deux r�gions aux populations et revenus h�t�rog�nes. La premi�re r�gion, faiblement peupl�e, pr�sente de fortes disparit�s de revenus, tandis que la seconde, plus homog�ne en termes de revenu, repr�sente un march� potentiel plus �tendu. On montre que l'h�t�rog�n�it� des revenus constitue la force dominante du mod�le lorsque le sc�nario d'investissement privil�gi� par les politiques publiques conduit � des gains substantiels du point de vue du co�t de transport entre les deux r�gions. L'effet de richesse, lorsqu'il est associ� � une forte disparit� des revenus, n'incite pas l'entreprise � exploiter son pouvoir de march� au d�triment de la r�gion l
This note develops a slightly different formulation of one of the basic results presented in a recent paper by Wallace and Hussain [5] on error components models for disturbances in relationships designed to explain cross-sectional observations over time. In their discussion, Wallace and Hussain derive the inverse of the variance-covariance matrix of the disturbances by trial and error. Unfortunately, their formulation does lead to a "natural" interpretation of the generalized least squares estimates, or of the relationships of these estimates to other estimates in the same way diagonalization of the variance-covariance matrix by means of an appropriate orthogonal transformation does. The characteristic roots of the variance-covariance matrix for the disturbances in a three component model which has been studied by Wallace and Hussain are derived here. It is shown how knowledge of these roots and the characteristic vectors associated with them leads to a form of the inverse matrix which may be more readily interpreted, as well as a number of other useful results, including an interpretation of the poor small sample properties of estimates which incorporate dummy variables for each individual.
Availability of data on a large number of individuals, but on each individual only over a very short period of time, has become increasingly common in a number of different fields in economics. Very often we would like to use such data to study behavioral relationships that are dynamic in character, i.e., that contain a distributed lag or other form of autogressive relationship. Since only a few observations are available over time, but a great many observations are available for different individuals at a point in time, it is exceptionally important to make the most efficient use of the data across individuals to estimate that part of the behavioral relationship containing variables that differ substantially from one individual to another, in order that the lesser amount of information over time can be used to best advantage in the estimation of the dynamic part of the relationship studied. As it turns out, the problem is far from simple: obvious devices such as the pooling of all observations and estimation by ordinary least squares, or the introduction of dummy variables for individuals, produce estimates having serious small sample bias. In earlier papers, the author and others have formulated a simple variance components model for the disturbance term in a relationship to be estimated from cross section data over time. This paper presents a series of Monte Carlo studies designed to explore the small sample properties of various types of estimates within this context. Not only is the bias of the obvious methods of estimation mentioned above confirmed, but certain serious deficiencies of the maximum likelihood approach which had been suspected earlier are also confirmed. A two-round estimation procedure is proposed which appears to work well for a wide variety of parameter values.
A GLS estimator of a triangular model with covariance restrictions is proposed that is asymptotically equivalent to the optimal minimum distance estimator regardless of nonnormality. The estimator used for the error covariance matrix Σ is a matrix-weighted average of the efficient restricted and unrestricted estimators of Σ. The relative weights are determined by a measure of the departure from normality.
Monetary policies of the ECB and US Fed can be characterised by Taylor rules, that is both central banks seem to be setting rates by taking into account the output gap and inflation. We also set up and tested Taylor rules which incorporate money growth and the euro-dollar exchange rate, thereby improving the fit between actual and Taylor rule based rates. In general, Taylor rules appear to be a much better way of describing Fed policy than ECB policy. Simulations suggest that the ECB's short-term interest rates have been at a much lower level in the last two years compared with what a Taylor rule would suggest.
Observations on N cross-section units at T time points are used to estimate a simple statistical model involving an autoregressive process with an additive term specific to the unit. Different assumptions about the initial conditions are (a) initial state fixed, (b) initial state random, (c) the unobserved individual effect independent of the unobserved dynamic process with the initial value fixed, and (d) the unobserved individual effect independent of the unobserved dynamic process with initial value random. Asymptotic properties of the maximum likelihood and “covariance” estimators are obtained when T → ∞ and when N → ∞. The relationship between the pseudo and conditional maximum likelihood estimators is clarified. A simple consistent estimator that is independent of the initial conditions and the way in which T or N → ∞ is also suggested.
This paper uses a natural experiment approach to identify the effects of an exogenouschange in future pension benefits on workers’ training participation. We use uniquematched survey and administrative data for male employees in the Dutch public sectorwho were born in 1949 or 1950. Only the latter were subject to a major pension reformthat diminished their pension rights. We find that this exogenous shock to pension rightspostpones expected retirement and increases participation in training courses amongolder employees, although exclusively for those employed in large organizations.