Estimating the bilateral impact of nontariff measures on trade

• Published date: 01 November 2017
• Date: 01 November 2017
• Author: Michael Bratt
• DOI: http://doi.org/10.1111/roie.12297

ORIGINAL ARTICLE

Estimating the bilateral impact of nontariff measures

on trade

Michael Bratt

Nyon, Switzerland

Correspondence

Michael Bratt, Nyon, Switzerland.

E-mail: bilateral.aves@hotmail.com

Abstract

This paper seeks to estimate how the impact of nontariff

measures (NTMs) on trade can vary across exporter–

importer pairs. Covering data for the early 2000s, regres-

sions are run at a disaggregate tariff line level and the

estimated results are converted into ad valorem equivalents

(AVEs). The results underline the importance of condition-

ing conclusions on trading partners and products and

demonstrate that the same NTM can have different—even

opposite—effects across exporting countries. One general

pattern that emerges is that low-income importers impose

more restrictive NTMs, but that the capacity for exporting

countries to address NTMs increases with GDP per capita.

1

|

INTRODUCTION

In 1989, there was a shift in Australia’s national standards for vehicles—the Australian Design Rules

(ADRs). Whereas the previous (second) edition of the ADRs had primarily been based on U.S. vehicle

requirements, the Australian government’s policy in the new (third) edition was that the ADRs should

be harmonized with the prevailing international standards (Australian Government, 2011), namely, the

UN Regulations adopted by the World Forum for Harmonization of Vehicle Regulations (WP.29)—a

body of the United Nations Economic Commission for Europe (UNECE).

1

One long-term implication

of this shift was that it became comparatively easier for vehicle exporters that already met UN Regula-

tions to comply with any revised or new ADR. It is, for instance, reasonable to expect that Australia’s

policy shift benefited, other things being equal, car producers in the European Union (which is a con-

tracting party to the relevant UN Agreement) more than car manufacturers in the United States (which

is not a contracting party).

2

This example illustrates two points with respect to nontariff measures

3

(NTMs) such as vehi-

cle regulations. First, the impact of an importer’s NTM on the exports of its trading partners can

be asymmetric: the same NTM can affect different exporters differently. Even, as in the case of

EU and U.S. vehicles, when the goods of the exporters are of similar quality. Second, the impact

of NTMs on trade flows is ambiguous; in some trading relationships it can be positive, while in

others it can be negative. It follows that aggregate imports do not necessarily have to fall with the

Rev Int Econ. 2017;25:1105 –1129. wileyonlinelibrary.com/journal/roie V

C2017 JohnWiley & Sons Ltd

|

1105

Received: 6 July 2015

|

Revised: 29 January 2017

|

Accepted: 5 April 2017

DOI: 10.1111/roie.12297

introduction of an NTM, as is clearly the case when total exports that benefit from an NTM out-

weigh the total exports that do not.

There are several empirical studies that demonstrate that the asymmetries of NTMs are real and

that the effects on trade flows can be positive as well as negative. For instance, Crivelli and Gr€

oschl

(2012) analyze the impact of sanitary and phytosanitary (SPS) measures and find that these lower the

probability of trade (the extensive margin), but have an overall positive impact on trade conditional on

market entry (the intensive margin). The authors also find that NTMs can have asymmetric effects, as

the impact at the intensive margin of trade is heterogeneous across exporters: the exports of some pro-

ducers tend to increase, while the exports of others tend to decrease. Bao and Qiu (2012), who focus

on the impact of technical barriers to trade (TBTs), similarly find a negative effect at the extensive mar-

gin and a positive effect at the intensive margin. In their case, the effects of TBTs on trade are contin-

gent on whether the trading partners are developing or developed countries.

The present paper is similar to these studies as it relates to the asymmetric and ambiguous features of

NTMs, but it takes a more detailed approach by asking a seemingly straightforward question: what are

the bilateral effects of NTMs on trade flows? The problem with addressing such a question is that

NTMs,unlike(ad valorem) tariffs, do not readily lend themselves to be represented as simple trade costs.

Indeed, any empirical study dealing with the role of NTMs in international trade faces the challenge of

how to quantify them and researchers have to resort to various tools to estimate their impact on trade.

In essence, two broad approaches can be distinguished: price-based approaches and quantity-based

approaches (for reviews, see Deardorff & Stern, 1997; Ferrantino, 2006, 2010). Although the means of

the two approaches differ, they share the same ultimate aim: to estimate the trade costs associated with

NTMs in terms of ad valorem equivalents (AVEs) (Berden & Francois, 2015). The price-based

approach does this directly by using detailed price data to estimate the impact of NTMs on price gaps,

that is, between import prices affected by NTMs and an NTM-free world price. For instance, Bradford

(2003) uses retail prices from an OECD survey in combination with data on transport and retail mar-

gins. He then attributes any price differences between domestic prices and export prices to nontariff

measures (in a broad sense of the term). Dean, Signoret, Feinberg, Ludema, and Ferrantino (2009)

adopt a different methodology, whereby both data on prices and on NTM incidence is used to econo-

metrically estimate the impact of NTMs on price gaps. The quantity-based approach, by contrast, is

more indirect and makes use of trade data to estimate the impact of NTMs on trade. Gravity-type equa-

tions are used to regress the value

4

of trade flows on NTM and other explanatory variables, where the

coefficient of the NTM variable is then used to estimate AVEs. Studies that follow this approach

include Kee, Nicita, and Olarreaga (2009), Beghin, Disdier, and Marette (2014), and Egger, Francois,

Manchin, and Nel son (2015).

Naturally, there are advantages and disadvantages with each approach. The strength of the price-

based approach lies in its direct comparison of prices, but it is a method that requires considerable

amount of detailed data across products and countries that is not usually available. It is therefore pri-

marily suited for quantifying NTMs in specific sectors and countries. By contrast, the quantity-based

approach is amenable to multi-sector and multi-country analysis because it uses detailed trade data that

is both available and easily accessible. However, the reliance on trade data is also the weakness of the

approach, since the results depend on the validity of the econometric specification (Ferrantino, 2010).

Because the concern in this paper is on how the effects of NTMs vary across a range of sectors

and country pairs—something that requires highly detailed data—it follows that the quantity-based

approach is the more suitable method. More specifically, the analysis draws on the work of Kee et al.

(2009), who estimate the importer-specific impact of NTMs at the six-digit level of the harmonized

system (HS 6-digit level) and compute AVEs per importing country and good. In a similar vein, AVEs

are here estimated based on cross-country regressions run at the same disaggregated product level, but

1106

|

BRATT