Estimating the impact of country‐level policy restrictions on services trade

• Date: 01 September 2018
• Author: Antoine Gervais
• DOI: http://doi.org/10.1111/roie.12340
• Published date: 01 September 2018

ORIGINAL ARTICLE

Estimating the impact of country-level policy

restrictions on services trade

Antoine Gervais

University of Notre Dame, Indiana

Correspondence

Antoine Gervais, Department of

Economics, University of Notre Dame,

Notre Dame, IN 46556.

Email: agervais@nd.edu

Abstract

Standard techniques used to estimate gravity equations, rely

on bilateral variation to identify the effect of trade barriers

on bilateral trade flows. In this paper, I develop a method

that estimates the effects of country-level variables on trade

between countries in the presence of firm heterogeneity and

country selection into trade, a natural extension of existing

empirical models of international trade. I implement the

method in services data, where available measures of trade

policy provide information on the average level of regula-

tions restricting entry into a country. The results suggest that

policy barriers are important determinants of services trade

flows.

1

|

INTRODUCTION

The gravity equation, which relates trade flows between countries to their economic size and the trade

resistance between them, is the workhorse empirical model in international trade. It is used to estimate

the impact of a wide range of trade barriers and policy issues such as distance, border effect, tariffs,

preferential trade agreement, and currency unions.

1

Although the gravity equation provides a flexible

framework and generally performs well, the standard estimation techniques have some limitations. In

particular, estimates of structurally motivated gravity equations are typically obtained from fixed-effect

estimators and rely on bilateral variation to identify the impact of trade costs on bilateral trade flows.

As a result, these methods cannot be used to estimate the impact of importer- or exporter-specific

restrictions on trade between countries.

This paper makes two main contributions. The first contribution is to combine and extend the

works of Redding and Venables (2004) and Helpman, Melitz, and Rubinstein (2008) to develop a new

method that provides estimates for the effects of country-level variables on trade between countries in

the presence of firm heterogeneity and country selection into trade. The second contribution is to

implement the methodology in services trade data to evaluate the impact of services trade regulations.

Because available measures of policy restrictions in services provide information on the average level

of regulations restricting entry into an importing country (i.e., bilateral measures of policy restrictions

are unavailable), standard techniques cannot be used.

2

Received: 6 June 2016

|

Revised: 25 October 2017

|

Accepted: 9 November 2017

DOI: 10.1111/roie.12340

Rev Int Econ. 2018;26:743–76 . wileyonlinelibrary.com/journal/roie

|

743

© 2018 John Wiley & Sons Ltd

7

Theoretical foundations for the gravity equation can be derived from a broad range of models. The

analysis in this paper is based on the framework developed by Anderson and Van Wincoop (2003) and

extended by Helpman et al. (2008) to include firm heterogeneity. The model demonstrates that estima-

tion of trade flows requires controlling for firm heterogeneity and “multilateral resistance”(MR) terms,

theoretically appropriate averages of the trade barriers between countries and all their partners. As

noted by Eaton and Kortum (2002) and Feenstra (2004), it is possible to control for the unobserved

MR terms using importer and exporter fixed effects. Nearly every gravity equation since then uses this

approach, including Helpman et al. (2008).

3

A potential drawback of the “fixed-effects”gravity equ a-

tion is that it cannot be used to estimate the impact of country-specific factors because they are sub-

sumed in the country-specific dummy variables. In this paper, I suggest an alternative approach that is

not subject to that limitation. Because the procedure is easy to implement, it can be effectively used in

many applications.

Similar to Anderson and Van Wincoop (2003), my method relies on the theoretical structure of a

model of international trade to obtain empirical measures for the MR terms.

4

My estimation approach

is quite different, however. While Anderson and Van Wincoop (2003) use numerical techniques to esti-

mate a nonlinear system of equations, I build on the work of Redding and Venables (2004) and con-

struct empirical measures for the parameters of the model (e.g., MR terms and fixed production costs)

from the estimates of the fixed-effect gravity equation. Another important distinction with Anderson

and Van Wincoop (2003) is that my model features firm heterogeneity and country selection into trade



alaMelitz (2003). While it is possible to extend the Anderson and Van Wincoop (2003) methodology

to account for heterogeneity and selection, doing so requires assumptions on trade costs that are very

restrictive and inconsistent with the data.

5

In contrast, my estimation approach is derived under more

modest assumptions and provides a straightforward extension of workhorse models that is consistent

with firm heterogeneity and selection into exporting.

The paper is organized as follows. In the next section, I obtain analytical expressions for the struc-

tural parameters of the model developed in Helpman et al. (2008) by imposing the free entry and labor

market clearing conditions. My analysis shows that bilateral trade depends only on country size, fixed

production costs, demand and supply parameters (i.e., MR terms), a measure of firm heterogeneity,

trade barriers, and fixed costs. In Section 3, I demonstrate that it is possible to construct empirical

measures of the unobserved parameters (e.g., MR terms, firm heterogeneity, and fixed costs) from the

estimates of the fixed-effect gravity equation. I then show that, by replacing the country dummies with

the generated measures in the gravity equation, it is possible to obtain estimates for the impact of

country-specific factors on bilateral trade flows. In Section 4, I implement the procedure in services

data to quantify the impact of policy restrictions on international trade.

In the application, I measure regulations using the newly available World Bank’sServicesTrade

Restriction Index. This index is the most comprehensive indicator available on services trade policies.

Nevertheless, it only provides information on regulations at the country level—thereby precluding the

useofstandard“fixed-effect”gravity equations. Using my procedure, I find that policy restrictions are

statistically significant and quantitatively important in predicting services trade flows. The empirical

results also highlight the importance of controlling for unobserved country effects using the generated

controls. When they are not included in the regression, the point estimates are significantly larger

(in absolute value). This suggests that, while informative, the few empirical studies that measure the

impact of policy restrictions on services trade present estimates that are biased owing to omitted varia-

bles (e.g., Marel & Shepherd, 2013; Nordås & Rouzet, 2015).

Similar to Chaney (2008), my analysis demonstrates that fixed export costs, through their general

equilibrium effects on wages and mass of firms, also play a significant role in determining the volume

of trade between countries. An important empirical contribution of this paper is to generate

GERVAIS

744

|