Page 121 - DECO503_INTERNATIONAL_TRADE_AND_FINANCE_ENGLISH

Unit 10 : Political Economy of Non-tariff Barriers : and Their Applications

A model of non-tariff barriers to trade Notes
Our initial model, then, is :
(
⋅
NTB t + 1 = AB SIZE+ 1 t + B 2 (logCONST ) + BSIZE logCONST ) + (1)
3
t
t
(
BUNEM + B 5 (UNEM log CONST⋅ ) + BUNEM SIZE log CONST⋅ ⋅ ) t
4
6
t
t
+B REER t + B 8 (REER log CONST ) + B 9
⋅
7
t
(REER SIZE log CONST⋅ ⋅ ) + B TARIFF + t . e
10
t
t
The dependent variable, NTB t + 1 , is the proportion of imports subject to NTBs in each state in year t
+ 1 based on the United Nations Conference on Trade and Development’s (UNCTAD) “inventory
list” of NTBs. As Sam Laird and Alexander Yeats observe, this list includes “Variable import levies
and product specific charges (excluding tariff quotas); Quotas; Prohibitions (including seasonal
prohibitions) : non-automatic import authorisations including restrictive import licensing
requirements; quantitative ‘voluntary’ export restraints; and trade restraints under the Multifibre
Arrangement.” It is, of course, difficult to gauge the extent of NTB protection within or across states.
Unlike tariffs, NTBs have no natural measure of intensity, and calculations of advalorem equivalents
of NTBs have proven to be unreliable. Some analysts have treated NTBs as a dichotomous variable
based on whether or not a given sector is protected by them. Others have estimated the level of NTBs
based on the difference between the predicted and observed openness of an economy.
In contrast to these approaches, we examine the incidence of NTBs. This measure is chosen because
the UNCTAD trade coverage ratios are viewed by many experts as the most reliable estimates of
NTBs across states and because it is the most appropriate variable with which to test our theory. For
example, polities characterized by many (and therefore small) parliamentary constituencies may be
especially prone to pork-barrel politics. Under these conditions, logrolling is likely to be pervasive
and the preferences of many different interest groups are therefore likely to be reflected in trade
policy. Since the extent and variety of interest-group demands reflected in trade policy bear directly
on the incidence of protection, we focus on explaining the incidence of NTBs. The coverage ratios
that we analyze measure the proportion of a state’s imports that are subject to NTBs.
Turning to the independent variables, SIZE is the economic size of each state in year t. It was pointed
t
out earlier that both the ratio of national imports to total global imports and the ratio of national GDP
to global GDP are used to measure relative state size. Thus, we estimate the parameters in equation
(1) using both variables. In addition, log CONST is the natural logarithm of the number of
t
parliamentary constituencies in each state in year t based on Rogowski’s data; UNEM is the
t
unemployment rate in each state in year t; REER is an index of the real exchange rate in each state in
t
year t; TARIFF is the average national post-Tokyo Round offer rate for each state; and e is an error
t
t
term. The remaining variables are included in order to determine whether, as we hypothesized above,
the interaction between factors that regulate the provision of protection (SIZE· log CONST), and the
interaction between factors that govern demands for protection and those that regulate its supply,
are important determinants of cross-national patterns of NTBs.
Data limitations led us to focus on explaining NTBs in 1983 and 1986. UNCTAD provides data on
NTBs for fourteen advanced industrial states in these years. The fourteen states are : Belgium-
Luxembourg, Denmark, Finland, France, Greece, Ireland, Italy, Japan, the Netherlands, Norway,
Switzerland, the United Kingdom, the United States, and West Germany. Although UNCTAD also
provides NTB data for New Zealand, our results indicated that this country was a statistical outlier.
We therefore excluded New Zealand and focus on the aforementioned fourteen states in the following
analysis. Because NTBs are measured in 1983 and 1986 (years t + 1), the independent variables in
equation (1) (except for TARIFF) are measured in 1982 and 1985 (years t). The observations for 1983
and 1986 are initially pooled; however, the extent to which the incidence of NTBs varied between
1983 and 1986 is also examined below.

LOVELY PROFESSIONAL UNIVERSITY 115