Vijil, M., Wagner, L. 2013. Does aid for trade enhance export performance? Investigating on the infrastructure channel. GREAT Insights, Volume 2, Issue 5. July-August 2013. Maastricht: ECDPM
Aid for Trade is intended to boost trade partly through a reduction in internal trade costs. We show that the effectiveness of such assistance as regards to export performance transits via the infrastructure channel.
Reducing domestic constraints through aid for trade
For many developing countries, an increase in trade does not depend on tariff reduction anymore and some of the poorest developing countries such as the Least Developed Countries (LDCs) benefit from free access to major markets. International traders may face other – at the border – and – beyond the border – trade costs, such as burdensome procedures, transit bottlenecks and absence of certification agencies; and these constraints are recognized to have significant impacts on trade volumes. Today, estimates indicate that addressing those issues are likely to have higher pay-offs in terms of trade than a reduction in trade-policy barriers. Therefore, the Aid for Trade initiative was launched in 2005 with the aim to assist developing countries in their attempt to enhance export performance and integration into the global economy, by targeting their own domestic constraints. Expectations were high after the announcement of an increase in assistance directed towards trade-related sectors. Furthermore, the ongoing debate on aid effectiveness following the “Paris Declaration” called for further evidence on its efficacy. In an attempt to investigate the relationship between aid for trade and export performance, we highlight infrastructure as one of the main channels of transmission.
There are few empirical studies assessing the effectiveness of aid for trade on trade performance, mostly because of the lack of sectoral data of sufficient quality and time length. Among the papers seeking to quantify empirically its impact on trade flows, evidence suggest that public investments can have a strong impact on export performance and underline the case for assistance in countries that have weak capacity and limited financial resources to address relevant constraints. Nevertheless, existing works do not test channels by which aid for trade flows spend in programs or projects can boost trade. As the extensive literature on the impact of trade costs on trade exhibit strong results, it seems relevant to focus on the effectiveness of aid flows on reducing these internal constraints.
Trade costs can be widely defined as any costs which increase the price of traded goods during the delivery process from the exporter (or producer) to the final consumer. If international trade costs associated to distance and, for example, overseas transportation costs are beyond the reach of political intervention, at the border and internal trade costs such as weak institutions or a lack of infrastructure can still be targeted and addressed by well-designed aid for trade, as part of the overall Official Development Assistance (ODA).
Following the OECD task force on aid for trade definition, aid for trade can be divided into five categories: (i) technical assistance for trade policy and regulations, proxy for trade-related institutions; (ii) trade-related infrastructure; (iii) productive capacity building; (iv) trade-related adjustment; and (v) other trade-related needs. Considering that the aim of our study is to test the channels (potentially related to trade costs) by which aid for trade can affect trade performance, we only focus on the first two categories, as other channels may be more difficult to comprehend.
Looking for channels of transmission of aid for trade efficacy: trade costs
We address the question of the effectiveness of aid for trade on trade performance using a two-step empirical analysis. Relying on an export performance model, we first test whether institutions and infrastructure, our two potential channels of transmission, are significant determinants of export performance. The quality of institutions is approximated by the number of days needed to export from the Doing Business database, which measures the time required to move a standard cargo from the gate factory in the economic capital to the ship in the most easily accessible port. As three-quarters of the delays seem to be due to administrative constraints, an increase in days can be associated with deterioration in the quality of the institutions related to trade. As for infrastructure, we construct an index which includes kilometres (km) of road and paved road (in total area, in km²), and the number of subscribers to mobile and telephone fixed lines (per 100 people) from the World Development Indicators database. We work with a sample of developing countries and use average values for the period 2002–08 for our estimations by 2SLS. As part of this first step, the infrastructure channel appears as a highly significant determinant of export performance whereas the institutional channel turns out to have a limited positive impact on developing countries’ export performance. Focusing on alternative measures of infrastructure, our results seem to point to the importance of an extended network of paved roads for export performance in developing countries.
Infrastructure as a determinant of export performance: aid to infrastructure matters
In a second step, we test the impact of aid to infrastructure on the previously detected determinant of export performance. We rely on the literature on economic geography, urban economics and the determinants of public investment in infrastructure to build our empirical model. Aid commitments for trade-related infrastructure per capita in constant US$ of 2000 from the OECD-CRS database, averaged over the period 2002-2007, constitutes our explanatory variable.(1) One of our main contributions is to offer a new instrument to tackle the endogeneity issue affecting this aid variable.
Endogeneity has been the main and recurring technical difficulty plaguing the aid effectiveness literature, in particular due to the reverse causality issue. In our case, one must acknowledge that aid for infrastructure is almost certainly allocated towards countries that lag behind. It is very difficult to find a convincing story around an external instrument supposed to explain aid flows but not the realization of the outcomes. Another challenge lies on the fact that estimators for panel data commonly used to deal with endogeneity are not recommended in our context. We deliberately choose to discard those techniques for two reasons. Firstly, those estimators do not seem adequate for series with a short-time span like ours where aid is likely to be highly persistent. Secondly, we believe they would prevent us from using variables of higher quality and precision. Indeed, the most interesting and precise data for some variables are only available for, at best, the most recent years. As a result, we propose a new instrument for aid for infrastructure: the number of privatisation transactions in the infrastructure sector between 2000 and 2007. We believe that this variable explains the aid for infrastructure flows received by a country without directly influencing the infrastructure level.
Indeed, nowadays the public sector is once again seen as the major player in financing many of the expansion needs in the infrastructure sector. Removing the dichotomous choice between public and private involvement in infrastructure investment, the public sector is now expected to retain an important financing role, while the private sector might bring better efficiency to supply and management. Furthermore, because of the high costs and limited capacity to pay by many of the users, the donor community is expected to be a central actor in the scaling-up of the public investment efforts, at least in the poorer countries. Hence, privatization transactions are often followed hand in hand by assistance directed towards sectors that were reformed.
We show, from this second step, that aid for infrastructure, once instrumented, has a strong and positive impact on the infrastructure level. As a result, when combining the two previously mentioned empirical steps, we find that a ten per cent increase in aid for infrastructure commitments per capita in developing countries leads to an average 2.34 per cent increase in the exports over GDP ratio. It is also equivalent to a 2.71 per cent reduction in tariff and non-tariff barriers.
These results highlight the high potential impact of aid for trade on developing countries’ export performance throughout the infrastructure channel. Our analysis seems to support the view that aid for trade might be a powerful instrument for assisting developing countries in their attempt to enhance export performance and integration into the global economy, while the multilateral talks within the Doha Round linger on. They also suggest that the private sector has to be supported by public and international financing in the infrastructure sector, advocating for an increased emphasis on public-private partnerships. Finally, our results do not explain precisely which type of infrastructure really matters, whether it is soft infrastructure related to efficiency considerations or hard infrastructure related to physical investments. Further evidence on the link between precise infrastructure categories, aid for trade and export performance are needed.
This article summarizes the findings of the following paper: Vijil, M., and Wagner, L., 2012, « Does aid for trade enhance export performance? Investigating on the infrastructure channel », The World Economy, Vol 35, Issue 7, pp 838-868.
Mariana Vijil holds a Ph.D. in Economics from Institut National de la Recherche Agronomique (INRA) Agrocampus-Ouest and is an international consultant.
Laurent Wagner holds a Ph.D. in Economics from Auvergne University (CERDI) and is a research officer in Foundation for International Development Study and Research (FERDI).
1. We choose to rely on commitments as sector specific disbursementsdata are of poor quality over the considered period because multilateral donors do not routinely report their activities in the OECD-CRS database. However, robustness checks using disbursements lead to similar conclusions.
This article was published in Great Insights Volume 2, Issue 5 (July-August 2013)