Estimating a mixture of stochastic frontier regression models via the em algorithm: A multiproduct cost function application

Researchers have become increasingly interested in estimating mixtures of stochastic frontiers. Mester (1993), Caudill (1993), and Polachek and Yoon (1987), for example, estimate stochastic frontier models for different regimes, assuming sample separation information is given. Building on earlier work by Lee and Porter (1984), Douglas, Conway, and Ferrier (1995) estimate a stochastic frontier switching regression model in the presence of noisy sample separation information. The purpose of this paper is to extend earlier work by estimating a mixture of stochastic frontiers assuming no sample separation information. This case is more likely to occur in practice than even noisy sample separation information. In order to estimate a mixture of stochastic frontiers with no sample separation information, an EM algorithm to obtain maximum likelihood estimates is developed. The algorithm is used to estimate a mixture of stochastic (cost) frontiers using data on U.S. savings and loans for the years 1986, 1987, and 1988. Statistical evidence is found supporting the existence of a mixture of stochastic frontiers. First version received: 3/13/01/Final version received: 6/17/02 RID="*" ID="*"  I am grateful to Ram Acharya, Janice Caudill, and especially James R. Barth for several helpful comments on an earlier version of the paper. During the revision process I benefitted greatly from the suggestions of the Associate Editor and three anonymous referees.     

Working time, employment, and work sharing: Evidence from Sweden

We study three questions which are important for work sharing to increase employment. First, is there a negative long-run relation between working time and employment? Second, are hours per worker exogenous with respect to wages and employment? Third, can policy makers influence actual hours per worker? We formulate a theoretical model for employment, hours per worker, production, and real wages. A VAR model with cointegrating constraints is estimated by maximum likelihood using Swedish private sector data 1970:1–1990:4. We find (i) no long-run relation between hours per worker and employment, (ii) that hours per worker are endogenous with respect to the estimation of long-run parameters, and (iii) that legislated working time and hours per worker are related to each other in the long run. First version received: September 1997/final version accepted: June 1999     

Public-private sector wage differentials in Germany: Evidence from quantile regression

This paper measures and decomposes the differences in earnings distributions between public sector and private sector employees in Germany for the years 1984–2001. Oaxaca decomposition results suggest that conditional wages are higher in the public sector for women but lower for men. Using the quantile regression decomposition technique proposed by Machado and Mata (2004), we find that the conditional distribution of wages is more compressed in the public sector. At the low end of wages, differences in characteristics explain less than the raw wage gap when it is the opposite at high wages. Separate analyses by work experience and educational groups reveal that the most experienced employees and those with basic schooling do best in the public sector. All these results are stable over the 80s and 90s.I thank Michael Lechner and the Centre for European Economic Research (ZEW), Mannheim, for letting me work with the full sample of the German Socio-Economic Panel (GSOEP). I am grateful to an anonymous referee and particularly to the editor, Bernd Fitzenberger, for providing me detail comments and suggestions that have significantly improved the paper. I have also benefited from discussions with Michael Lechner and Ruth Miquel.First version received: April 2002/Final version received: June 2004     

Participation and hours of work: Two stages in the life-cycle of married women

A 1979 cross-section of Dutch households is used to estimate an integrated labor force participation and hours of work labor supply model for women. Young women and older women are found to behave differently. Wage and income elasticities are always higher for older than for younger women (in absolute terms), and the relative importance of participation and hours elasticities is reversed as the average woman grows older, thereby implying that younger women tend to be more preoccupied with the participation decision whereas older women are more concerned with the hours decision. Econometrically the paper builds on Heckman's (1974) participation cum-hours model by adapting the estimation method to allow for the fact that hours worked are only observed in intervals. A maximum likelihood model in which we integrate over parts of the hours distribution is used. This technique, though computationally involved, gave satisfying results. We compare the results with estimates of a similar but much simpler probit model to identify the possible biases in the simpler approach.     

The introduction of seasonal unit roots and cointegration to test index aggregation optimality: An application to a Spanish farm price index

In this paper we present a methodological proposal of the way integration and cointegration analysis can best be used to test if the level of aggregation of an index is adequate. Using this proposal, we enquire the extent to which a Spanish aggregate farm price index captures the behavior of its components. First version received: October 1994/final version received: September 1998     

Conditional mean functions of non-linear models of US output

We compare a number of models of post War US output growth in terms of the degree and pattern of non-linearity they impart to the conditional mean, where we condition on either the previous period's growth rate, or the previous two periods' growth rates. The conditional means are estimated non-parametrically using a nearest-neighbour technique on data simulated from the models. In this way, we condense the complex, dynamic, responses that may be present in to graphical displays of the implied conditional mean. First version received: Feb. 1999/Final version received: June 2001     

The minimum wage, employment, and the AS-IF methodology: A forecasting approach to evaluating the minimum wage

Out-of-sample employment forecasts for 33 U.S. industries which are likely to be sensitive to the federal minimum wage are, more often than not, more accurate when information about the minimum wage is not taken into account. This is true even in instances where this information improves wage forecasts. When employment forecasts conditional on the minimum wage are better, the improvement is typically small. These results are invariant to the number of workers previously making less than the new minimum wage, and to the value of the minimum wage relative to industry average wages. First version received: August 1999/Final version received: July 2000     

Adoption of safer irrigation technologies and cropping patterns: Evidence from Southern Ghana

The new irrigation technologies introduced in sub-Saharan Africa aim at ensuring safer vegetable production when untreated waste water is used as irrigation water. This paper examines the adoption of safer irrigation technologies and crop choices among vegetable farmers, using cross-section data from urban Kumasi in Ghana. The study employed a two-stage conditional maximum likelihood approach to examine the impact of crop choices and farm-level characteristics on the adoption of safer irrigation technologies. The empirical results indicate that apart from household and farm characteristics such as access to extension agents, belonging to a farmer's organization and education, cropping patterns tend to influence adoption of irrigation technologies for safer vegetable production.     

The effects of off‐farm work on fertilizer and pesticide expenditures in China

This study examines the effects of participation in off‐farm work on farm expenditures on fertilizer and pesticide, using farm household survey data from China. Simple mean value comparisons reveal no statistically significant differences in fertilizer and pesticide expenditures between off‐farm work participants and nonparticipants. However, econometric estimation with a treatment effects model shows a negative selection bias. After controlling for this bias, the empirical results show that participation in off‐farm work exerts a positive and statistically significant impact on fertilizer and pesticide expenditures. Our findings generally suggest that the income effect of off‐farm work stimulates agricultural production by increasing investments in productivity‐enhancing inputs.     

Structural Change Tests in Tail Behaviour and the Asian Crisis

This paper explores tests of the hypothesis that the tail thickness of a distribution is constant over time. Using Hill's conditional maximum likelihood estimator for the tail index of a distribution, tests of tail shape constancy are constructed that allow for an unknown breakpoint. The recursive test is shown to be inconsistent in one direction, and only a one-sided test is recommended. Specifically, the test can be used when the alternative hypothesis is that the tail index decreases over time. A rolling and sequential version of the test is consistent in both directions. The methods are illustrated on recent stock price data for Thailand, Malaysia and Indonesia. The period covers the recent Asian financial crisis and enables us to assess whether breakpoints in domestic asset return distributions are related to known changes in institutional arrangements in the foreign currency markets of these countries.     

Who is ruling Europe? Empirical evidence on the German Dominance Hypothesis

The aim of this paper is to analyze the hypothesis of German dominance in the European Monetary System (EMS). For this purpose, we use monthly interest rates for nine European countries from January 1979 to the second half of 1997. In particular, we test the stability of the implied long-run relationships to assess whether there has been a significant change in the EMS performance. The econometric methodology is based on Johansen's maximum likelihood procedure and several tests for parameter instability. The most important finding is that, although Germany has a significant influence on the monetary policy of the other European countries, it is not possible to accept the existence of dominance. First version received: May 1999/Final version received: November 2000     

Dynamics of female labour force participation in France

This article formulates and estimates a structural intertemporal model of labour force participation. Relying on theoretical characterizations derived from an economic model of lifetime behaviour, we estimate a dynamic probit model with correlated random effects using longitudinal data to allow for a dynamic structure. The model is applied to a panel of married women drawn from the 1997–2002 French Labour Force surveys in order to represent their participation behaviour. It is estimated by maximum simulated likelihood. Our results show that women’s decisions to go out to work are characterized by significant state dependence, unobserved heterogeneity and negative serial correlation in the transitory error component. In addition, we find that the presence of young children in a woman’s household reduces her labour participation, but unequally according to their age and number. As expected, educational level has a positive impact on women’s participation. Last, a rise in the husband’s wage is found to raise female participation initially and to lower it subsequently. This empirical finding suggests that leisure may not systematically be a normal good, contrary to what is frequently assumed in the literature.     

Dynamic spatial modelling of regional convergence processes

Econometric analysis of convergence processes across countries or regions usually refers to a transition period between an arbitrary chosen starting year and a fictitious steady state. Panel unit root tests and panel cointegration techniques have proved to belong to powerful econometric tools if the conditions are met. When referring to economically defined regions, though, it is rather an exception than the rule that coherent time series are available. For this case we introduce a dynamic spatial modelling approach which is suitable to trace regional adjustment processes in space instead of time. It is shown how the spatial error-correction mechanism (SEC model) can be estimated depending on the spatial stationarity properties of the variables under investigation. The dynamic spatial modelling approach presented in this paper is applied to the issue of conditional income and productivity convergence across labour market regions in unified Germany.First version received: December 2002/Final version received: June 2003We would like to thank an anonymous referee for his helpful comments.     

On uniqueness of the conditional maximum likelihood estimation for a binary panel model

This paper considers an estimation method for a binary panel model with incidental parameters as individual effects. The necessary condition for the conditional maximum likelihood approach proposed by Andersen (1970) is investigated and we show that unique sufficient statistics exist only for logit models in a two-wave panel.     

Forecasting industrial production in the Euro area

The creation of the Euro area has increased the importance of obtaining timely information about short-term changes in the area's real activity. In this paper we propose a number of alternative short term forecasting models, ranging from simple ARIMA models to more complex cointegrated VAR and conditional models, to forecast the index of industrial production in the euro area. A conditional error-correction model in which the aggregate index of industrial production for the area is explained by the US industrial production index and the business confidence index from the European Commission harmonised survey on manufacturing firms achieves the best score in terms of forecasting capacity. First version received: Jan. 2000/Final version received: March 2000     

Institutions and economic development: How strong is the relation?

This paper investigates the relationship between institutions and economic development (output per worker). As in Hall and Jones (1999), we find that a 1% improvement in institutions (as we measure them) generates on average a 5% increase in output per worker. However, this relationship is not linear and the data have important heterogeneity. Countries with the same value of institutions have different levels of income per worker. We ask whether the “returns to institutions” are the same across countries conditional on the level of institutions. Using quantile regression methods, we show that for countries at the top of the conditional distribution of international incomes, the “returns to institutions” are lower (around 3.8%,) than for countries at the bottom of this distribution (around 6.2%). We show that this result is robust for different model specifications and definitions of institutions. We also provide evidence that, conditional on the level of institutional development, the distribution of output per worker tends to become less disperse as countries improve their institutional framework. In other words, having better institutions is essential in order to close the output-per-worker gap across countries. Finally, we provide the rationale behind the results through a modified version of a Neoclassical Growth Model with time varying wedges, representing policy distortions and institutions.We thank Lee Alston, Roger Koenker, and Stephen L. Parente for helpful discussion, Werner Baer for useful comments, and Chad Jones for facilitating access to the data set. We are also indebted to an anonymous referee and an associate editor for important suggestions that improved the final paper. The analysis, opinions and findings expressed herein represent the views of the authors, they are not necessarily those of the Banco de Portugal. Any remaining errors are our responsibility.First version received: May 2001/Final version received: August 2003     

Designing a payments for ecosystem services (PES) program to reduce deforestation in Tanzania: An assessment of payment approaches

Payments for ecosystem services (PES) programs aim to improve environmental outcomes by providing direct incentives to land managers for the provision of ecosystem services. Participation in PES programs is voluntary, so effective program design requires careful consideration of farmers' preferences. This study quantifies such preferences using a choice experiment. The study site is the East Usambara Mountains, Tanzania, an internationally recognized ‘biodiversity hotspot.’ We assess preferences for four payment approaches: constant and variable annual cash payments to individual farmers, a constant annual cash payment to a village fund on behalf of farmers, and an upfront manure fertilizer payment. We find that the manure fertilizer payment was statistically significant in motivating farmer participation while the group payment was non-significant. In addition, the relationship between the likelihood of participation and the stringency of conditionality is surprisingly non-linear. In a test of external validity, average willingness to accept (WTA) values are found to be similar to the average opportunity cost of maintaining land uses consistent with conservation objectives.     

Measuring the effectiveness of non-price export promotion using a supply-side approach

In this paper we assess the effectiveness of the Market Promotion Program (MPP) in increasing U.S. exports and benefitting U.S. agricultural producers and food processors. Export shipments are linked to producer welfare using Kohli's (1978) profit maximization (GNP function) approach to modeling international trade. Using estimated profit functions in conjunction with a synthetic export demand function for processed agricultural products, we compute changes in farm and food processing sector profits that result from alternative own-price and advertising elasticities of export demand with and without the MPP subsidy. This approach allows us to investigate aggregate welfare effects of nonprice promotion without requiring the difficult task of estimating the export demand effects of market promotion activities for numerous commodities and importing countries. First version received: April 1999/Final version received: June 2000     

Doing it all: Women’s employment and reproductive work in Tajikistan

Central Asia has low labor force participation rates for women, despite high levels of poverty in the non-oil producing countries. Female labor force participation is related to competing uses for women’s time, especially in a context of poor infrastructure and limited availability of purchased replacements for household labor. We examine factors affecting women’s participation in employment and reproductive household work in Tajikistan, drawing on the 2003 Tajikistan Living Standards Survey. We incorporate specific conditions in Central Asia, including the prevalence of extended family households, norms about how women share household work, an absence of market substitutes for caring and reproductive labor, employment in family enterprises and poor infrastructure, especially in rural areas. We estimate the system using a Generalized Maximum Entropy (GME) approach. We find that few individual and household characteristics are related to time in employment. Time in noncare reproductive work decreases if a woman receives direct remuneration for her employment (compared to working in a family farm or enterprise but receiving no direct remuneration), but increases if the woman has less access to infrastructure. Rural women spend more time in both employment and non-care reproductive work than urban women, but less time in care work. Lack of infrastructure may leave women with few choices with respect to non-care reproductive work, while overall levels of time poverty create pressures on care time.     

Does women’s participation in agricultural technology adoption decisions affect the adoption of climate‐smart agriculture? Insights from Indo‐Gangetic Plains of India

Increased participation of women in the agricultural technology adoption decision by farm households is one of the key indicators of gender empowerment in the agricultural sector. This study examines whether women’s participation in the household decision to adopt agricultural technology affects the adoption of climate‐smart agriculture (CSA), using data collected from 1,267 farm households from two Indian states of Bihar and Haryana. When we considered the sex of the household head (using a dummy variable for male‐headed vs female‐headed household) as a basis of analyzing the role of gender in the adoption of CSA, we found that women in Haryana had no role in the adoption of CSA. On the contrary, when we considered women’s participation in technology adoption decisions as a basis of gender analysis, we found that women’s participation in technology adoption decisions in Haryana is much higher as compared to Bihar. Consequently, the likelihood to adopt CSA is higher in Haryana than in Bihar. We also found that wealth, training, and access to extension and market positively influenced CSA adoption. Qualitative analysis shows that women farmers prioritize family food security rather than farm income, and therefore, they are more likely to focus on CSA to ensure food security.