Institutional quality and the effectiveness of aid for trade

ABSTRACT

Trade policy barriers and high transaction costs hinder developing countries from taking the full advantages of the global trading system. In order to help developing countries overcome the problem, the World Trade Organization (WTO) launched the Aid for Trade (AFT) initiatives in its Ministerial Conference held in Hong Kong in 2005. We examine the effects of AFT inflows on bilateral trade costs facing 133 developing countries while accounting for differences in their location on the contours of various measures of institutional quality. Our results from the estimation of a mixed effects (random-intercept and random-coefficient) model indicate that institutional quality significantly affects the extent to which AFT reduces bilateral trade costs. An important policy implication of our findings is that an economically robust and sustainable reduction in bilateral trade costs facing developing countries requires the presence of both promulgated and effectively functioning institutions such as regulatory power and the rule of law.

KEYWORDS:

• Institutions
• policy
• aid for trade
• effectiveness
• mixed-effects model

JEL CLASSIFICATION:

• F10
• F35
• C51

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Arvis et al. (2013) note that distance, not only induces transportation costs, but also creates barriers to information and reduces the probability that a trade connection between two countries will take place.

² De Melo and Olarreaga (2016) indicate that both traditional variables (technology and endowments) and the quality of contracting institutions largely determine current-day trade.

³ According to OECD (2015, 23), high trade costs inhibit LDCs from exploiting the market access that the global trading system creates. As a result, market access does not always convert into market presence. Lanz, Roberts, and Taal (2016) and Arvis et al. (2013) indicate that poorer countries have higher levels of trade costs than do richer countries. Using costs for moving a standard container across borders, Königer, Busse, and Hoekstra (2011) further show that developing countries pay, on average, 43% more on exports, and 54% more on imports than developed countries.

⁴ The labelling 'Market enabling’ is due to Rodrik, Subramanian, and Trebbi (2004).

⁵ In countries with weak market enabling institutions, potential partners may find trading too costly due to a lax in contracts enforcements, or payments uncertainty, even if a country lowers its trade barriers such as tariffs.

⁶ As a result, the World Bank has made strengthening good governance in developing countries as both an objective and a condition for development assistance.

⁷ Much of the efforts to develop trade cost measures focus on the direct estimates of the trade cost components (international transport costs, the costs of moving goods from the factory to the dock of a ship at the nearest seaport, the cost of observed tariff and policy-based non-tariff, or opinions of business, or logistics service providers, etc.). However, these approaches do not provide a comprehensive measure of international trade costs, which combine the different measures and indicators into a feasible comprehensive measure.

⁸ To facilitate better understanding of these values, an average ad valorem tariff equivalent trade costs of 305.9% between countries i and j suggest that trading goods involve additional costs as compared to when the two countries trade the goods within their borders, on average.

⁹ Across sectors, bilateral trade costs for agricultural goods facing a typical recipient (340.9%) is significantly higher than for manufactured goods (207.9%), the differences remaining consistent across regions. For example, agricultural trade costs facing recipients in Africa (360.2%), Asia (343.2%), and the Pacific (334%) are significantly higher than the typical recipient in the Americas (326.4%) and Europe (300.2%). Similarly, for manufactured goods, recipients in Africa (223.8%) and the Americas (215.6%) exhibit much higher trade costs than the typical recipient (207.9%), the recipients in Asia (195.9%) and Europe (177.4%) exhibiting relatively lower trade costs, on average.

¹⁰ The estimated correlation coefficient is −0.2077.

¹¹ In addition to suitability for log transformations and comparisons of the values across the measures, the reference countries, and the period possible, there exist perfect positive correlation coefficient of +1 between the normalized indices of each dimension and their respective raw values from the Kaufmann, Kraay, and Zoido-Lobatón (2014). The observation indicates that the transformed values retain all information contained in the original data.

¹² Appendix Table A1 provides the variance structure of bilateral trade costs of recipients in the study.

¹³ We also estimate our model using linear and log-linear versions of the dependent variable series. Results obtained from alternative specifications (pooled OLS and the standard fixed panel data estimation approaches) while consistent with the results reported here are not reported for brevity of space.

¹⁴ The magnitude of the BLUPs indicates the extent to which the actual effect of AFT on bilateral trade costs in the given country is larger or smaller than the average effect computed for the entire sample.

¹⁵ While direct statistical comparisons of the coefficients from separate estimations are not plausible, results presented in columns (b) to (e) of Table 5 specifically suggest that improvements in the quality of institutions reflected by the rule of law and regulatory quality (power) tend to have larger effects.

¹⁶ The anomaly in the results for agricultural goods has more to do with the limited sample size.

¹⁷ A rise in AFT inflow reduces bilateral trade costs. The computed marginal effects are. thus, negative (See, Table 9) in footnote #17 to (See, Table 8). The graphs depicting the corresponding effects are downward slopping. However, we inverted the graphs in order to better depict the ‘pattern’ and ‘magnitudes’ of the observed effects.

¹⁸ Appendix Table A2 provides a summary of recent observations.