Hervé Boulhol

The Contribution of Economic Geography to GDP Per Capita

This article examines how much of the dispersion in economic performance across OECD countries can be accounted for by the proximity to areas of dense economic activity. To do so, various indicators of distance to markets and transportation costs are added as determinants in an augmented Solow model, which serves as a benchmark. Measures of distance to markets are found to have a statistically significant effect on GDP per capita. And the estimated economic impact is far from negligible. The reduced access to markets relative to the OECD average could contribute negatively to GDP per capita by as much as 11% in Australia and New Zealand. Conversely, a favourable impact of around 6-7% of GDP is found in the case of two centrally-located countries: Belgium and the Netherlands. The paper provides also some tentative evidence that spending on R&D and human capital might have a stronger effect on GDP per capita in countries with a higher degree of urban concentration.

Have Developed Countries Escaped the Curse of Distance

There is widespread evidence that a better access to markets contributes to raising income levels. However, no quantification of the impact of distance to markets has been made on the basis of a sample restricted to advanced — and therefore more homogeneous — countries. This paper applies the framework developed by Redding and Venables (2004) on a panel data covering 21 OECD countries over 1970-2004, and shows that, relative to the average OECD country, the cost of remoteness for countries such as Australia and New Zealand could be as high as 10% of GDP. Conversely, the benefit for centrally-located countries like Belgium and the Netherlands could be around 6-7%. Second, the paper explains why the key estimated parameter in the Redding-Venables model is biased upwards in cross-section samples that mix both developing and developed countries, because of the inability to adequately control for heterogeneity in technology levels across countries. The paper also provides a detailed discussion of the links between the ?death-of-distance? hypothesis, the evolution of transport costs and that of the elasticity of trade to distance. Les pays developpes ont-ils echappe a la malediction de la distance ? De nombreuses etudes empiriques ont montre qu‘un meilleur acces aux marches contribue a augmenter les revenus. Cependant, aucune quantification de l‘impact de la distance aux marches n‘a ete effectuee a partir d‘un echantillon homogene limite aux pays developpes. Ce papier applique le cadre developpe par Redding and Venables (2004) a des donnees de panel couvrant 21 pays de l‘OCDE entre 1970 et 2004, et montre que, relativement a la moyenne des pays de l‘OCDE, le cout de l‘eloignement geographique pour des pays comme l‘Australie et la Nouvelle Zelande s‘eleve a environ 10% de PIB. Reciproquement, le benefice que tirent les pays ayant une position centrale comme la Belgique et les Pays-Bas serait de l‘ordre de 6-7%. Deuxiemement, cette etude explique pourquoi le parametre-cle dans le modele Redding-Venables est biaise a la hausse dans des echantillons en coupe qui melent pays developpes et en developpement, en raison de l‘incapacite a controler l‘heterogeneite des niveaux technologiques entre pays. Le papier propose egalement une discussion detaillee des liens entre l‘hypothese de la « fin de la distance », l‘evolution des couts de transport et celle de l‘elasticite du commerce a la distance.

Deindustrialisation and the Fear of Relocations in the Industry

The steady decline in the share of industry in total employment currently seems to be accelerated by competition from low wage countries, which depresses international prices for manufactured products and is being translated into offshore outsourcing and defensive innovation. In order to fully address these issues, we define offshore outsourcing in a broader sense, the one which is generally adopted in the public debate, namely trade in goods with countries where offshoring takes place. We replicate and extend the estimations realised by Rowthorn and Ramaswami (1998) and ask what has been the responsibility of outsourcing in the observed deindustrialisation in sixteen OECD countries. Our estimation strategy, relying on a dynamic panel methodology, leads to the conclusion that net trade with low wage countries is associated with a non-negligible average decrease of around 2 points in the manufacturing employment share between 1970 and 2002. However, this contribution, which substantially varies across countries, represents only a fifth of the deindustrialisation over the period. We did not find any increased impact of imports from developing countries in the second half of the period (1986-2002), other than that due to increased trade.

Applying the GLM Variance Assumption to Overcome the Scale-Dependence of the Negative Binomial QGPML Estimator

Recently, various studies have used the Poisson Pseudo-Maximal Likehood (PML) to estimate gravity specifications of trade flows and non-count data models more generally. Some papers also report results based on the Negative Binomial Quasi-Generalised Pseudo-Maximum Likelihood (NB QGPML) estimator, which encompasses the Poisson assumption as a special case. This note shows that the NB QGPML estimators that have been used so far are unappealing when applied to a continuous dependent variable which unit choice is arbitrary, because estimates artificially depend on that choice. A new NB QGPML estimator is introduced to overcome this shortcoming.

Pro-competitive Effect of Trade and Non-decreasing Price-Cost Margins

This study surveys the empirical evidence on the pro-competitive effect of international trade and analyses the determinants of price-cost margins for OECD countries between 1970 and 2003. The main objective was to focus on the quantification of the impact of imports on margins, and understand why, despite trade liberalization, price-cost margins have not fallen overall. On average, imports would have contributed to a large decrease of five percentage points in the price-cost margins. However, these effects seem to have been partially counterbalanced by the impacts of financial deepening and disinflation.

Recent Trends and Structural Breaks in US and EU15 Labour Productivity Growth

This article examines shifts in labour productivity growth in the US and in Europe between 1970 and 2007 based on econometric tests of structural breaks. Additionally, it makes use of time-series-based projections of labour productivity growth up to 2009 in order to detect breaks depending on confidence intervals of the projections. The identification of structural breaks in the US labour productivity growth is far from obvious. A statistically significant break is found in the late 1990s only if at least the 97.5th percentile of forecasts materializes in the future, which means that despite a clear pick up in productivity growth in the second half of the 1990s, the size of the hump is not large enough compared with past variations to make this change a statistically significant break. However, a significant break point is detected in the mid-1990s for the difference in labour productivity growth between the US and the EU15, even when controlling for the convergence of Europe towards the US productivity levels that has contributed to higher European performance in the early catch up phase. Finally, within Europe, the accumulation of Information and Communication Technology (ICT) capital seems to be related to differences in the shifts in structural labour productivity growth across countries.

Employment-Productivity Trade-off and Labour Composition

This paper formalises the analysis of the employment-productivity trade-off by extending the framework developed by Gordon (1997) to account for labour heterogeneity. The extent of the trade-off is determined by the extent of the adjustment of capital to effective labour and by the changes in aggregate labour quality. The main experiment reported in the paper consists of assessing the labour utilisation and productivity impacts in OECD countries of aligning group-specific employment rates to the US levels. Matching the US employment performance defined in that sense would enable low-employment OECD countries to reduce only half of the aggregate employment-rate gap vis-a-vis the United States, the other half being mechanically due to differences in the population structure by age and educational attainment. In this experiment, a 1% gain in employment is associated with a decrease of 0.24% in labour productivity on average across countries, and of 0.35% in low-employment countries. Compromis emploi - productivite et effets de composition Cette etude formalise l’analyse du compromis entre emploi et productivite en etendant le cadre developpe par Gordon (1997) pour prendre en compte l’heterogeneite de la main-d’oeuvre. L’ampleur de ce compromis est determinee par l’etendue de l’ajustement du capital a la main-d’oeuvre effective et par les changements dans la qualite de la main-d’oeuvre. La principale experience rapportee dans l’etude consiste en l’evaluation de l’impact sur l’utilisation de la main-d’oeuvre et sur la productivite du travail de l’alignement, pour chaque pays de l’OCDE, des taux d’emplois par groupe de population sur ceux des Etats-Unis. Repliquant la performance des Etats-Unis ainsi definie permettrait aux pays de l’OCDE ayant un faible niveau d’emplois de reduire seulement la moitie de l’ecart de taux d’emploi agrege vis-a-vis des Etats-Unis, l’autre moitie etant due mecaniquement a la structure de la population par âge et niveau d’education. Dans cette experience, des gains de 1% en termes d’emplois sont associes a une baisse de 0.24% de la productivite du travail en moyenne pour les pays de l’OCDE et de 0.35% pour les pays ayant les niveaux d’emplois les plus bas.

The upward bias of markups estimated from the price-based methodology

Previous studies have emphasized that Roegers methodology generates too high markups. This feature is confirmed on the basis of the unrealistically low capital shares implied by the estimates herein. Theoretically, it is shown that the normalization choice, the slow adjustment of capital and the mismeasurement of capital expenditures, each produces an upward bias. For instance, the price-based estimated markup is in fact the markup adjusted for returns to scale on the variable inputs only. Based on the empirical analysis, each of these sources of overestimation is very likely to play a role.