DESARROLLO Y SOCIEDAD

MARZO DE 2004 53
ISSN 1900-7760
(Edición Electrónica)

FTAA and Service Liberalization in Colombia

Miles K. Light*

Abstract
In a previous study we found that accession to the FTAA could be welfare-
worsening for Colombia because exports to the USA will be diverted away
from Colombia in favor of other Latin countries. In this paper we show that
there remain potentially large gains from an FTAA-style agreement. These
gains come from increased factor productivity and product variety as a result
of service liberalization and foreign direct investment (FDI). These benefits
are likely to be large enough to overcome the loss of competitiveness in US
goods markets. We use a computable general equilibrium model of the Co-
lombian economy that includes imperfect competition in order to highlight the
pro-competitive effects from entry as well as productivity effects from in-
creased product variety. In contrast to perfect competition models, such as
the GTAP model, this analysis incorporates productivity effects in both goods
and services markets endogenously, through a Dixit-Stiglitz framework. The
numerical model is innovative as it recognizes that foreign direct investment
or the availability of foreign expertise is necessary to have foreign firms com-
pete in key business services; and it endogenously captures increases in total
factor productivity from foreign direct investment liberalization.

Introduction
Colombia currently enjoys almost full access to US and European markets
under the Andean Trade Preference Act (APTA)1. Because tariffs for Colom-

* Business Research Division, University of Colorado. Email: miles@[Link]

1 The name was changed to the ATPDEA: Andean Trade Preference and Drug Eradication
Act. However, we have chosen here to maintain the orginial acronym, ATPA. The ATPA
offers low or zero tariffs for most Andean exports into US markets in exchange for efforts
to combat drug production.

165
FTAA and Service Liberalization in Colombia
Miles K. Light

bia’s goods are already near zero, further gains from trade appear unlikely
under the Free Trade of the Americas Agreement (FTAA). In particular, there
are two daunting prospects that Colombia must confront when considering
joining the FTAA. First, Colombia and the other ATPA countries (Peru, Bo-
livia, and Ecuador) will face increased competition for their exports as other
Latin countries begin to enter US markets. By the same token, trade-prefer-
ences for Colombia within the Andean Pact will also face increased competi-
tion from non-Andean countries. Second, tariffeliminationinColombiawillgenerate
significant budget shortfalls. Current tariff collections represent about eight per-
cent of total revenues for the government or about three percent of GDP. A
reduction in domestic tariff rates must be accompanied by either an increase in
value-added taxes or higher income taxes in order to replace lost revenues. These
findings are not encouraging for Colombian trade negotatiors or businesses.

Despite these discouraging prospects, we find that Colombia stands to gain

substantially from accession to the FTAA. However, these gains will come
from liberalization in services rather than trade liberalization of goods. Using
a small open-economy model which incorporates increasing returns to scale,
we find that Colombian consumption could increase by 3.9% by joining the
FTAA (Table 1). In the long-run, after capital and investment changes take
effect, the gains are 5.9%. If the government must recover lost tariff revenues
by raising VAT rates, we still find that welfare increases by 2.6%. In contrast,
the traditional gains from trade in goods is small. If we assume constant returns
to scale and consider only trade in goods, then welfare increases by 1.4% in
the central FTAA scenario, and by only 0.2% in the balanced-budget scenario.

These significant differences are indicative of some important characteris-

tics of the Colombian economy. First, services are the largest component
of the economy. Depending upon their definition, services represent be-
tween 40-60% of the Colombian economy (and between 75-85% of the
US economy). Because this sector is so large, even a small improvement in
productivity can lead to a relatively large change in welfare. Second, most
of the Colombian “core’’ manufacturing and production depends upon serv-
ices as an intermediate input. Transportation services, financial services,
consulting services, and communication services are all part of the produc-
tion process. Finally, services are used both as intermediate inputs to pro-
duction and as inputs to final consumption. Since services are so prevalent
in production and consumption, we find that even a small increase in com-
petitiveness or productivity has a large effect upon welfare because it low-
ers production costs and at the same time increases utility.

166
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
Table 1. Welfare Impact of FTAA Accession and Service Liberalization.

Full Access Steady-state Equal Yield

IRTS CRTS IRTS CRTS IRTS CRTS

Consumption (%) 3.9 1.4 5.9 2.6 0.2

GDP (%) 2.5 0.9 3.8 1.7 0.1
Revenues (%) -4.1 -4.4 -8.4 0 0
VAT Change (%) 12.4 11.7

Impact Dimensions under the FTAA Using a small open-economy

(SOE) economic model for Colombia, we can consider several effects
from FTAA accession which are likely to be important. These effects are
listed below.

• An impact upon goods markets which use CRTS technology in light of

existing tariffs and trade preferences. Tariff elimination will improve
resource allocation by eliminating price distortions.
• A shift in tax collections from import tariffs to value-added taxes. The
impact upon tax-efficiency will depend upon the tariff rate structure
and the efficiency of the replacement tax.
• An increase in overall factor productivity derived from increased product
variety. The increases in variety come as more import varieties are
avaialble for IRTS producers and as more domestic varieties are available
as a result of increased Foreign Direct Investment (FDI).
• An increase in productivity coming from better techniques offered
through the service sector. This improvement comes from liberalization
in the service sector and from transfer of technology from multinational
corporations.
• An impact upon the long-run, steady-state level of investment and
production as a result of increased product variety and factor productivity.

I. Services and FDI

A. Theory
The notion of imperfect competition and increasing returns to scale has
been recognized very early on in the international trade literature (e.g., Ohlin

167
FTAA and Service Liberalization in Colombia
Miles K. Light

(1924) and Graham (1923)), but it was not until a set of useful theoretical
constructs were developed that the role of multinational corporations, and
foreign direct investment could be characterized in a meaningful way. Work
by Lancaster (1975), Spence (1976), and Dixit and Stiglitz (1977) was
readily used as an approach to monopolistic competition –that could be
applied to interindustry trade– could recognized that increasing returns to
scale can lead to large-scale production and international specialization.
Since 1985, the focus of international trade literature has been upon in-
creasing returns to scale and imperfect competition. This new literature is
often termed the new trade theory, because it moves beyond the standard
Ricardian theory of comparative advantage to explain why countries with
similar endowments might engage in trade. The impetous for the new trade
theory was a series of empirical observations which contradicts standard
trade theory. The main finding was that 80% of all trade occured between
countries who have similar endowments. Also, most trade in manufactured
and finished products moved between OECD countries, while most of the
trade in primary factors and unfinished goods occured between developed
and developing countries.

B. Quantifying the Barriers to Services and FDI

Unlike tariffs, there is no simple measure of the barriers to Foreign Direct
Investment (FDI) and service provision by foreigners. Often, these barri-
ers are vagely-defined in the law and they are imposed differently across
sectors, often on a case-by-case basis. The non-standard nature of non-
tariff barriers (from here on: NTBs) clearly causes problems when attempting
to quantify the restrictiveness for a country.

Despite these challenges, economists have developed several means to

identify the restrictiveness of trade, at least to a certain order of magni-
tude. For example, estimates of the quantity impact of NTBs have been
derived from econometric trade models (Brown and Stern 2001). The
difference between the observed trade flows and the predicted flows is
assumed to result from NTBs. Of course, this method suffers the major
problem that econometric models do not capture all of the determinants
of trade and therefore some of the difference between predicted and
actual flows is not caused by the presence of NTBs, but instead simply
by a model with low explanatory power. Other methods which have been
considered are:

168
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
• Frequency measures Counting the number of restrictions in each sector.
• Price Differences Attribute the difference in the price of services to
non-tariff barriers.
• Indices Applying weights to various trade and investment restrictions
in order to calculate a “restrictiveness index.’’

For the most part, and for lack of a better methodology, we use the “Trade
Restrictiveness Index’’ (TRI) which applies weights to various barriers to
investment in order to synthesize the barriers into a single number.

The second step in the process is to then quantify the ad-valorem equiva-
lent of the TRI for each sector. We use TRI estimates for 38 other coun-
tries in each of the service and FDI sectors, then regress price against the
RTI in order to compute percentage change in price related to the restric-
tiveness. Obviously, this method is somewhat arbitrary. The index weights
have been chosen based upon personal judgment, only certain countries
have been included in the regression due to data limitations, and the ex-
planatory power of the regressions is necessarily low becuase market
structure is not captured in the regression model. Despite these difficul-
ties, we notice that the regression results are reasonable. As a defense
against spurious calculations, we also compute a sensitivity analysis to
the central ad-valorem estimates in the model. We calculate a range of
scenarios where the TRI ad-valorem equivalents reach the boundries of
their 95% confidence intervals.

Prior research resulting from the GATS negotiations can be used here to
define the major barriers to FDI. Table 2 lists common barriers and classi-
fies them into three main categories:Restrictions on Entry, Ownership
and control restrictions, and Operational restrictions.

A 38-country study of the trade restrictiveness in services was conducted

by the Australian Productivity Comission, and the resulting research find-
ings were presented in an edited volume by Findlay and Warren (2000).
Although the particular focus was trade impediments in Asian Pacific
countries, Colombia appears to have been included as one of the 38
countries. Some of the findings from this research has been consoli-
dated into Table 3.

169
FTAA and Service Liberalization in Colombia
Miles K. Light

Table 2. Common Barriers to FDI.

Restrictions on market entry

• Bans on foreign investment in certain sectors
• Quantitative restrictions (eg limit of 25 per cent foreign ownership in a sector)
• Screening and approval (sometimes involving national interest or net economic benefits
tests)
• Restrictions on the legal form of the foreign entity
• Minimum capital requirements
• Conditions on subsequent investment
• Conditions on location
• Admission taxes

Ownership and control restrictions

• Compulsory joint ventures with domestic investors
• Limits on the number of foreign board members
• Government appointed board members
• Government approval required for certain decisions
• Restrictions on foreign shareholders’ rights
• Mandatory transfer of some ownership to locals within a specified
• time (eg 15 years)

Operational restrictions
• Performance requirements (eg export requirements)
• Local content restrictions
• Restrictions on imports of labour, capital and raw materials
• Operational permits or licences
• Ceilings on royalties
• Restrictions on repatriation of capital and profits

Source: UNCTAD (2002)

Barriers to foreign direct investment have been estimated in a few Colom-

bian service sectors, namely in telecommunications, banking, external mari-
time transportation services, and retail and wholesale distribution services.
The methodology employed is an application of the methodology and data
work of Christopher Findlay and Tony Warren. Findlay and Warren have
employed cross-country data sets in several service sectors where the price
and quantity of services in the sector is regressed on measures of regula-
tory barriers. Findley and Warren then infer from these regressions the
impact of changes in any of the regulatory barriers on the price or quantity
of the service. Using the estimated coefficients listed in Table 3, we find
that the cost of banking services in Colombia are 18% higher due to barri-
ers to FDI and the cost of telecommunications services are 24%higher.

170
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
Table 3. Colombia Trade Restrictiveness and Price-Effect Indices.

Trade Restrictiveness Price Effects

Sector Domestic Foreign Domestic Foreign

Banking 0.2850 0.3997 3.54% 18.35%

Distribution 0.1238 0.1904
Maritime 0.1805 0.4690
Telecommunications 0.2000 0.4600 10.55% 24.26%

Definitions:
• Distribution: Wholesale and retail trade (except motor vehicles and motorcycles) including
commission trade and repair of personal and household goods. (622,63,51-2).
• Banking: Financial intermediation services, except insurance and pension funds.(811).
• Maritime: Water transportation.
• Telecommunications: Telecommunications, including fixed line, mobile, and internet
communications.
Source: McGuire/UNCTAD (2002)

The restrictiveness indices listed on the left-hand side of Table 3 represent

an index value, between zero and one. The difference between the domes-
tic index and the foreign index reflects the discrimination applied to for-
eign firms. On the right-hand side of the table are the estimated price effects.
Although there is some ambiguity regarding whether all prices are raised
by this number, or if this is effectively the ad-valorem equivalent of the
restrictiveness index for Colombia. We adopt the latter meaning, and take
these figures to be the ad-valorem equivalent for non-tariff barriers.

II. The Model

A. Conceptual Framework
The trade model employed in this analysis is different from previous studies for
Colombia. First, we do not use the multiregional framework and dataset from
the trade-in-goods study by Rutherford and Light (2003). Instead, we model
Colombia as a small open economy based upon 1997 national accounting
data. The second major departure is the inclusion of increasing returns to scale
technology (IRTS) based upon the Dixit-Stiglitz product variety framework.

There are 17 sectors in the model that are listed in Table A.1. This model
can also be applied to a more detailed, 57-sector dataset for Colombia.
However we found that the nature of the gains and the trade effects are
likely to be similar between the 17-sector and the 57-sector aggregations.
For computational simplicity and logical transparency, we use the 17-sec-
tor aggregation during this analysis.

171
FTAA and Service Liberalization in Colombia
Miles K. Light

The 17 sectors are listed in three separate categories. One category of

sectors is those goods or services that are produced under constant re-
turns to scale and perfect competition. In these sectors, competitive do-
mestic firms face competition from foreign producers where goods are
differentiated in the demand functions of Colombian consumers and firms.
This is known as the Armington assumption.

A second category of sectors is those goods that are produced under in-
creasing returns to scale and imperfect competition. These goods are char-
acterized as Dixit-Stiglitz composites of domestic and import varieties with
firm-level product differentiation. The efficiency gains associated with an
increased number of varieties accrue to both consumers and firms using
these goods as intermediate inputs. Foreign firms supply the Colombian
market with production facilities abroad, but the number of foreign firms
that are present in the Colombian market depends on quasi-rents available
in the Colombian market, which in turn depends on the tariff rate.

The third category of sectors contains services which are produced under
increasing returns to scale and imperfect competition. For these services,
two types of firms operate: domestic and multinational. Multinational service
firm providers must establish a domestic presence in order to compete in the
Colombian market. They must import some of their technology or manage-
ment expertise. They cannot supply the Colombian market from abroad as
goods providers can do. Thus, their cost structure differs from goods pro-
viders. They incur costs related to both imported inputs and domestic good
and factor inputs. Domestic service providers do not import foreign technol-
ogy or management expertise. Hence, domestic service firms incur costs
related to domestic goods and factor inputs only. These services are charac-
terized by firm-level product differentiation. Restrictions on foreign direct
investment, right of establishment, the movement of business personnel, and
lack of intellectual property protection and contract enforcement have major,
direct impacts on multinational firms providing services to the market.

B. Overview of the Model Formulation

The model algebra is very similar to the Jensen, Tarr, Rutherford (2003)
model for Russia’s accession into the World Trade Organization. Natu-
rally, the model structure was altered in order to reflect special character-
istics within the Colombian economy.

172
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
Primary factors include capital, skilled and unskilled labor, and sector-spe-
cific workers. Twenty-five percent of the labor in all sectors is assumed to
be sector specific.

Goods produced subject to increasing returns to scale are differenti-

ated at the firm level; firms in these industries set prices such that mar-
ginal cost equals marginal revenue; and there is free entry, which drives
profits to zero. We employ the standard Chamberlinian large group
monopolistic competition assumption, which results in constant markups
over marginal cost.

Aggregate productivity is affected by the number of varieties using the stand-

ard Dixit-Stiglitz formulation. The effective cost function for users of goods
produced subject to increasing returns to scale declines in the total number
of firms in the industry.

For simplicity we assume that the composition of fixed and marginal cost is
identical in all increasing returns to scale sectors. This implies that the ratio
of fixed to marginal cost is a constant. This assumption in a large-group
model assures that output per firm for all firm types remains constant, i.e.,
the model does not produce rationalization gains or losses.

We assume that manufactured goods are either produced domestically or

imported, and the cost structure of domestic firms is defined by observed
primary factor and intermediate inputs to that sector in the base year data.
The CIF import price of foreign goods is simply defined by the import
price and by the zero profits assumption. In equilibrium, the import price
must cover fixed and marginal costs of foreign firms.

We assume that in the IRTS service sector, there are two types of firms
providing services to the Colombian economy: (i) Colombian firms, who
employ primary factors and intermediate inputs and (ii) multinational firms
who provide services using imported inputs (FDI and foreign expertise)
together with primary factors and intermediate inputs.

We assume that the structure of both the marginal and fixed costs of serv-
ices firms are identical, so that output per firm is fixed and there are no
rationalization gains. This assumption lies parrallel to the cost structure for
IRTS producers of goods.

173
FTAA and Service Liberalization in Colombia
Miles K. Light

For multinational service providers, both the fixed and variable costs of
service supply are assumed to be a convex combination of the domestic
supply price in the same sector and the cost of imported inputs.

Comparative Steady State Formulation In this version of our model, we

allow the capital stock to adjust to its steady state equilibrium along with all of
the model features we employ in our FTAA reference case, i.e., we allow for
tariff and FDI liberalization with endogenous productivity effects as above.
We call this our comparative steady state model. In the comparative static
model, we assume that the capital stock is fixed and the rental rate on capital
is endogenously determined. In the comparative steady state model, the logic
is reversed. We assume that the capital stock is in its initial steady state equi-
librium in the benchmark dataset, but that the capital stock will adjust to a
new steady state equilibrium based on a fixed rate of return demanded by
investors. That is, if the trade policy shock happens to induce an increase in
the rate of return on capital so that it exceeds the initial rate of return, inves-
tors will invest and expand the capital stock. Expansion of the capital stock
drives down the marginal product of capital, i.e., it drives down the rental
rate on capital, until the rate of return on capital falls back toward the long-
run rental rate. To analyze trade policy, this comparative steady approach
has been employed by many authors, including Harrison, Rutherford and
Tarr (1997) and Francois et. al. (1997). The approach, however, dates back
to the 1970s, when both Koopmans and Manne used it.

C. Algebraic Formulation
The model includes the standard general equilibrium consistency features.
Final demand arises from a representative household who earns income
from the sale of primary factors (capital, skilled and unskilled labor). The
government levies direct and indirect taxes and purchases a vector of goods
and services. In this section we outline the key features of the model in
terms of the objectives and constraints facing various agents.

Consumer Behavior Private consumption in the model arises from budget-

constrained utility maximization. Preferences are represented as a Cobb-
Douglas aggregate of goods and services:

174
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
in which θi > 0 and = 1. Associated demand functions are defined in
terms of goods prices pi, consumption tax rates and aggregate income, M:

Income is defined in terms of sources of factor income:

The right side of the budget constraint includes wage income from both
mobile and sector-specific labor, and capital earnings. Investment demand
is fixed when k = 1. In a steady-state equilibrium, both the capital stock
and the level of investment adjust to a level k > 1 which equates the cost of
capital formation and the discounted present value return to a unit of new
capital. The final term on the right-hand side is the level of lump-sum tax
adjust which is used to balance the government budget and hold public
output constant (see below).

Domestic Supply Goods and services are produced for sale in the do-
mestic and international markets. A constant elasticity of transformation
(CET) function shows the transformation possibilities in a given period be-
tween domestic (Di) and export (Ei) sales for a given composite output
level (Yi). The shares of sales at home and abroad are determined by rela-
tive prices given that firms produce the final good to maximize profit sub-
ject to the CET constraint:

In this equation parameters bar and bar are the base year output for
the domestic and export markets, respectively, and θD is the baseline value
share of domestic sales in total sales (the base year production level is
scaled to unity) and is the elasticity of transformation.

175
FTAA and Service Liberalization in Colombia
Miles K. Light

Production is associated with a nested production function of intermediate

inputs xji, labor services left (Lli and LSli), and capital (Ki). Given prices of
intermediate goods and labor, the aggregate production sector operates so
to minimize the costs of producing a given output subject to the constraint:

(1)

in which aji represents the intermediate input of good j to sector i. In this

function, skilled and unskilled labor (both mobile and sector-specific) and
capital enter in a Cobb-Douglas aggregate with value shares determined
by base year demands.

Differentiated Goods Goods produced subject to increasing returns to

scale are characterized as differentiated products of domestic and
foreignfirms. Effective supply of all firms in a given sector is described by:

(2)

nDi and nFi are the numbers of domestic and foreign firms/varieties and XDi
and XFi represent composites of domestic and foreign goods:

(3)

In the final expression χki is output of a representative type k firm, and

is resource inputs at marginal cost of all type k firms. The out-
put of domestic firms, characterized by (3), is therefore equal to domestic
supply less fixed costs of domestic firms, i.e.

(4)

176
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
and the output of importing firms is defined by imported resource less the
fixed cost of those firms:

Holding total output constant, effective supply of either domestic or for-

eign varieties of commodity i increases with , which is the “variety
k
effect multiplier.’’The multiplier increases with n i and increases as the elas-
ticity of substitution decreases toward 1.

The supply of good i equals aggregate demand, the sum of intermediate

demand, consumer demand, investment demand and government demand:

(5)

The number of domestic and foreign varieties determine the effective sup-
ply index, Xi, and we thereby assume that the Dixit-Stiglitz productivity
has an symmetric impact on both intermediate and final demand. Changes
in the number of domestic and foreign varieties are reflected through changes
in the price index of the commodity associated with Xi.

Differentiated Services Services supplied under conditions of increas-

ing returns to scale are characterized as the differentiated products of do-
mestic and multinational firms. Effective supply of all firms in a given sector
is described by:

(6)

nDi and nMi are the numbers of domestic and multinational firms / varieties,
and χDi and χMi are output per firm of those two types of firms. In the final
equation χDi and χMi represent composites of domestic and multinational
services, i.e.:

177
FTAA and Service Liberalization in Colombia
Miles K. Light

(7)

In the final expression χki is output of a representative type k firm, and

is resource inputs at marginal cost of all type k firms.

The crucial distinction between differentiated goods and differentiated serv-

ices is that in the case of goods, domestic supply, characterized by (1),
equals the value of domestic goods while for services this quantity is split
between resources used in producing domestic services and resources em-
ployed by multinational firms. In the case of services, we may then define:

(8)

where represents domestic resources used in the supply of services by

domestic firms:

while represents domestic resources used in the supply of services by

multinational firms.

We assume that multinational firms use domestic inputs in fixed proportion

to imported inputs. Hence,

and

( ) (
Di = 1 − θ iM niM χ iM + f i M )
in which represents the benchmark value share of imported inputs to
multinational service supply.

As in the case of differentiated goods, holding total output of either domes-

tic or multinational services constant, effective supply of either domestic
or multinational varieties of service commodity increases with ,
which is the “variety effect multiplier”. The multiplier increases with and
increases as the elasticity of substitution decreases toward 1.

178
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
Likewise, the supply of differentiated service i equals aggregate demand,
the sum of intermediate demand, consumer demand, investment demand
and government demand:

(9)

The number of domestic and multinational varieties determine the effective

supply index, Xi, and we thereby assume that the Dixit-Stiglitz productiv-
ity has a symmetric impact on both intermediate and final demand. Changes
in the number of domestic and foreign varieties are reflected through changes
in the price index of the commodity associated with Xi.

Tax Revenue and the Public Budget In the model, the government collects
a variety of indirect taxes. These taxes and the associated ad-valorem rates
include the taxes on output tyi, tariffs tMi, taxes on exports by trading partner
txir, and taxes on consumption tCi. The government budget constraint is then:

in which Tk represents revenue from tax instrument k, and TLS represents

direct (lump-sum) taxes. The model features a constant level of public pro-
vision, which is achieved through adjustment of the level of lump sum tax.

[Link]
We use 1997 national accounts which have been compiled and documented
by the Colombian national statistics office called “DIAN.’’ This dataset is
well-documented and has been used in several tax-reform studies in 20022.

The core input-output model is the 1997 table produced by the Colombian
Ministry of Finance. The official table contains 17 sectors, five factors of
production (capital and 5 labor types), five major tax streams, a single
government agent, ten households distinguished by income level, imports,
exports and rest-of-world net savings.

2 See Rutherford, Light and Hernandez (2002) and Rutherford and Light (2002).

179
FTAA and Service Liberalization in Colombia
Miles K. Light

Table 4. Services as part of the Production Process: Production and Services

in the Colombian SAM.

OUTPUT % OUTPUT % SERVICE

COF 2,284 1.2 15.8

CRO 8,084 4.3 12.4
LVS 6,420 3.4 8.9
FFH 1,032 0.5 8.0
OIL 6,258 3.3 30.6
MIN 1,188 0.6 13.0
THR 554 0.3 10.4
FOD 19,855 10.5 12.3
NRI 1,062 0.6 29.1
NSI 9,506 5.0 24.0
HTC 16,388 8.7 51.3
CON 15,429 8.2 27.8
TRN 10,972 5.8 35.3
ELE 7,477 4.0 17.1
COM 3,105 1.6 27.8
SER 54,107 28.6 27.7
GOV 25,398 13.4 24.4
TOTAL 189,127 100.0 25.9

Output values: Billions of Pesos

% Service: the percentage of services as a share of total intermediate inputs to production.
Sectors included in services are: SER,COM, and TRN.

The households have been combined into a single, representative-agent for

simplicity. Futher analysis may include a disaggregation which can distin-
guish the effects of liberalization upon the poor. The structure of the 1997
SAM is consistent with current efforts to develop a new SAM for the year
2001. Future anlysis may incorporate these developments, but in the mean-
time we believe that most of the action can be captured with the current
model implementation.

The second step in the process is to then quantify the ad-valorem equiva-
lent of the TRI for each sector. We use TRI estimates for 38 other coun-
tries in each of the service and FDI sectors, then regress price against the
RTI in order to compute percentage change in price related to the restric-
tiveness. Obviously, this method is somewhat arbitrary. The index weights
have been chosen based upon personal judgment, only certain countries
have been included in the regression due to data limitations, and the ex-
planatory power of the regressions is necessarily low because market
structure is not captured in the regression model. Despite these difficul-
ties, we notice that the regression results are reasonable. As a defense

180
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
against spurious calculations, we also compute a sensitivity analysis to
the central ad-valorem estimates in the model. We calculate a range of
scenarios where the TRI ad-valorem equivalents reach the boundaries of
their 95% confidence intervals.

Barriers to foreign direct investment have been estimated in a few Colom-

bian service sectors, namely in telecommunications, banking, external mari-
time transportation services, and retail and wholesale distribution services.
The methodology employed is an application of the methodology and data
work of Christopher Findlay and Tony Warren. Findlay and Warren have
employed cross-country data sets in several service sectors where the price
and quantity of services in the sector is regressed on measures of regula-
tory barriers. Findley and Warren then infer from these regressions the
impact of changes in any of the regulatory barriers on the price or quantity
of the service. We use the estimated coefficients listed in Table 3 as the
domestic and foreign barriers to FDI in Colombia.

Table 5 reports the market share controlled by domestic and foreign firms
in services. For foreign firms, the last column in table 5 lists the share of
production that is imported from outside Colombia. This captures the
amount of services and goods that the foreign firm that come from head-
quarters. Barriers that limit the use of these imported goods, such as
limitations on foreign residence, taxes upon special machinery, or the ban
on foreign-purchased cellular handsets, will limit the degree of new firm
entry as well.

Table 5. IRTS Market participation of foreign firms in Colombia.

θ)
Market Share (θ Import Share for FDI (θ M)
Sector Domestic Foreign Foreign

Banking 95% 5% 10%

Distribution 91% 9% —
Maritime 97% 3% —
Telecommunications 85% 15% 10%

Source: Colombian Department of National Planning (2003).

A. Tariff Data and Trade Volumes

Tariff and export tax rates are imputed from total tariff collections and total
imports and exports. Tax collections divided by total trade values produce

181
FTAA and Service Liberalization in Colombia
Miles K. Light

the imputed rate. Because Colombia is part of the Andean Community,

they share a common external tariff and similar institutional requirements
for foreign investment. The imputed rates listed in Table 6 represent the
average rates among several hundred rates. The actual rates for any par-
ticular commodity can vary from 0% to 200%. In particular, the case of
commodity food imports are subject to a price-band system. In this system
the import tariff for all Andean Community members is adjusted until food
import prices are as high as local prices.

We do not attempt to capture all of the intricacies of specific tariff lines in

this paper. Instead, we describe a simple tariff structure among the sectors
listed below. More detailed import tariff rates at the tariff line level can be
obtained from official government decrees which are available online3.

Like most countries in Latin America, import tariffs are highest in the food
and agriculture sector (11%), while most other tariffs are relatively low.
Extensive trade reforms from 1990 and 1991 have sucessfully lowered the
average tariff rate from 25-30% to their current levels which are 7%-10%.
We also observe a current account deficit of about 4,000 billion pesos.
This trade deficit is held constant for all of the scenarios in the model.

It is worth noting that import tariffs in the service sector (SER) and the
transport sector (TRN) are both near zero. These goods are typically non-
traded and require a local presence. On the other hand, the average import
tariff for ‘HTC,’ high-technology and capital-intensive goods is relatively
high, 6.4%, and trade in this sector is the largest in the economy. We model
HTC as a “Dixit-Stiglitz’’ sector, where increased firm entry or interna-
tional competition will improve productivity in this sector.

IV. Policy Results

The role of service-liberalization and increasing returns to scale is impor-
tant because services are widely used as an intermediate input to produc-
tion. The potential gains from service liberalization are significant, especially
when compared to traditional CRTS gains, for two reasons: 1) Production
costs are lowered because of pro-competitive and efficiency effects. These

3 See the the FTAA Hemispheric Database. The data are current for Colombia as of 2002.

182
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
Table 6. Average Import Tariffs, Collections, Imports and Exports.

RATE (%) FDI-NTB (%) REVENUES IMPORTS EXPORTS

COF 6.3 0.0 0 2,602

CRO 11.0 117.9 1,190 2,552
LVS 2.6 0.8 30 20
FFH 7.0 3.6 54 11
OIL 5.9 36.8 656 3,253
MIN 3.0 3.9 135 556
THR 3.8 0.0 0 188
FOD 6.5 105.5 1,739 902
NRI 5.7 0.1 1 978
NSI 4.6 159.2 3,588 1,532
HTC 6.4 798.3 13,294 2,951
TRN 0.4 25 799
ELE 2 0.4
COM 25 23.2 271
SER 0.2 25 0.3 129 230
TOTAL 1226.4 20,847 16,851

Values are: Billions of 1997 Colombian Pesos.

Rate: Imputed tariff rate (percentage).
FDI-NTB Calculated Non-tariff barrier to FDI.

cost-savings translate directly into lower prices for households and they do
not require lower input prices in order to lower the costs, as is the case in
CRTS technologies. 2) Existing distortions between goods and services
are removed which increases efficiency of the overall economic system.
Table 8 reveals, substantial gains from service liberalization in both goods
and services. Consumption increases by 3.9%, and GDP increases by 2.5%
relative to benchmark levels.

Table 7. Description of central scenarios using IRTS and CRTS technology

and investment

Scenarios:

FULL: Full trade concessions for goods, services, and FDI with Increasing Returns to Scale
Technology.
HALF: Partial concessions, where Colombian import tariffs are reduced by 50% and barriers
to FDI are similarly reduced. IRTS technology is assumed.
CRTS: Full trade concessions for goods, but all sectors are assumed to have Constant
Returns to Scale technology.
FDI: Role of investment. FDI barriers are removed while import tariffs remain in place.
SS: Full liberalization scenario results under the Steady-state (long-run) assumption.
CRTSYLD: Equal-yield tax calculation. Impact of FTAA when tariff revenues must be replaced
by value-added taxes. Trade in goods only.
IRTSYLD: Equal-yield tax calculation under the IRTS technology and trade in services model.

183
FTAA and Service Liberalization in Colombia
Miles K. Light

Table 8. Summary Results Table.

FULL HALF CRTS FDI SS CRTSYLD IRTSYLD

EV (% CONS) 3.9 2.9 1.4 2.1 5.9 0.2 2.6

EV (% GDP) 2.5 1.9 0.9 1.4 3.8 0.1 1.7
REAL EX 3.3 2.6 1.2 2.0 2.9 0.8 2.8
GOV REVENUES ($) -863.9 -371.4 -919.0 50.3 -1767.4 0.0 0.0
GOV REVENUES (%) -4.1 -1.8 -4.4 0.2 -8.4 0.0 0.0
VAT (%-CHANGE) 0.0 0.0 0.0 0.0 0.0 12.4 11.7

Descriptions:
EV (% CONS): Equivalent Variation as a percentage of current consumption.
EV (% GDP): Equivalent variation as a percentage of original GDP.
REAL EX: Percentage change to the price of foreign exchange. This is computed
using the price of foreign exchange (COP per USD, for example) and the
domestic price index. The formula is: 100 .
GOV REVENUES ($): Change in government revenues denoted in Billions of 1997 Cololombian
Pesos.
GOV REVENUES (%): Percentage change in government revenues.
VAT (%-CHANGE): Percentage change in value-added tax rates required to replace lost import
tariff revenues. Only used in the ‘equal-yield’ calculations. A 12%increase
in the VAT is equivalent to raising rates from 16% to 18%.

In order to identify where the gains from trade lie, we conduct several
separate scenarios. These scenarios are each described in Table 7. We
identify the role of FDI and IRTS technology, then compare these effects
with the standard CRTS technology.

A. Trade in Goods with CRTS Technology

The impact of FTAA under CRTS production technology has been con-
sidered using the GTAP database in Rutherford and Light (2002). We cal-
culate the ramifications using the SOE model here, but our results are
necessarily biased. They do not reflect changes in comparative advantage
and competition from foreign producers. They also do not reflect increased
import demand from foreign agents. In this treatment FTAA implies full
unilateral tariff elimination only.

In order to consider the role of trade-diversion and existing trade prefer-

ences, we have included results taken from Rutherford and Light and in-
cluded them into the report. These results show that trade diversion is
ultimately a very important component of FTAA accession for Colombia
because the benefits of the ATPA agreement will be diluted under the FTAA.

184
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
The central findings from our SOE model are less optomistic when a full
multi-regional trade model is used. When we consider the GTAPinGAMS
model, which has been tailored for Colombia under the FTAA negotia-
tions, we find that Colombian consumption falls by about 0.6% instead of
rising. This change in welfare reflects the gains from eliminating domestic
tariffs and distortions, but also captures a large loss in exports as competi-
tion to US markets increases.

FTAA enactment essentially eliminates any existing trade preferences by

setting US import tariffs to zero for all countries in the agreement. In this
respect, Colombia and other ATPA4 countries will face increased compe-
tition for exports into the US, which in turn lowers domestic production.

An important finding using the current Dixit-Stiglitz model is that accession

into the FTAA is not necessarily harmful to Colombia. The expected gains
in productivity are expected to offset losses from trade diversion. Using the
GTAPinGAMS model, we found that Colombian consumption fell 0.6%,
using the model presented here, CRTS FTAA (CRTSYLD) consumption
increased 0.2%, but with IRTS and FDI liberalization (IRTSYLD), wel-
fare increases 2.6% (under equal-yield assumptions).

From table 8, it is fairly clear that welfare improvements come mostly from
increased foreign direct investment, and less so from tariff elimination. Tar-
iff elimination increases the number of varieties for IRTS sectors producing
goods. A comparison between only-FDI, where barriers to investment are
removed against only-tariff elimination, the results differ substantially. Wel-
fare in the FDI scenario increases 2.1% versus 0.5% in gains from tariff
elimination with IRTS sectors included. Most of the difference comes from
the revenue-replacement requirement. A similar calculation which elimi-
nates import tariffs, but does not replace the revenues yields an increase in
welfare of 2.1%. Part of the gains are simply lower tax collections.

4 Under the ATPA, Andean countries including Colombia, Equador, Peru, and Bolivia enjoy
low or zero import tariffs for goods entering the United States. Consequently, the US is
the largest single trading partner for each of these countries. ATPA benefits are given to
these countries in exchange for efforts to combat drug production.

185
FTAA and Service Liberalization in Colombia
Miles K. Light

B. Factor Market Impacts

In general, real wages and the return to capital rise unambiguously across
most scenarios. Tariff elimination and service liberalization both tend to
lower import prices and lower production costs. Lower consumption
prices and higher factor productivity in turn increase the real-return to
labor and capital, even low-skilled labor. Table 9 lists the change in the
real return to factors.

Table 9. Factor returns under FTAA.

FULL HALF CRTS FDI SS CRTSYLD IRTSYLD

UFS 3.2 2.5 0.1 2.9 6.9 -0.2 2.8

UFN 4.4 2.8 1.6 2.5 11.9 -0.3 2.4
UTC 4.4 2.7 1.8 2.3 11.8 -0.2 2.3
UMC 1.7 1.7 -1.2 2.7 3.7 0.1 2.9
RSW 3.8 2.8 1.4 2.3 3.5 1.0 3.4
RNW 4.4 2.9 2.1 2.1 4.2 1.0 3.3
K 4.7 3.2 1.5 2.8 0.6 0.2 3.3

Definitions:
UFS Urban, formal, salaried workers.
UFN Urban, formal, non-salaried workers.

UTC Urban traditional contract workers.

UMC Urban modern contract workers.

RSW Rural salaried workers.
RNW Rural non-salaried workers.
K Return to Capital.

V. Conclusions and Further Work

The government of Colombia is currently in a position to help domestic
producers participate in the global market for goods and services. Although
most of the traditional gains from trade have already been utilized, there
remain large gains from trade in services and from market liberalization and
standardization. We estimate that Colombia has been able to enjoy a 1%
increase in household consumption as producers take advantage of ATPA
preferences in the US. A bilateral agreement would ensure that these pref-
erences are not lost in the medium term, but it would also open Colombian
markets to US goods. The elimination of Colombian tariffs is a mixed bag:
prices for final goods and for intermediate inputs will fall, helping to raise

186
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
the standard of living, but producers will face increased competition from
US imports, which could lower production and raise unemployment. The
combined effect of keeping ATPA preferences and lowering Colombian tar-
iffs is positive, leading to net consumption gains of approximately 0.7%-1%.

The most important component of a bilateral trade agreement has to do

with the services sector and foreign direct investment. If done correctly,
liberalization of financial, shipping, professional, and communications serv-
ices may have a dramatic effect upon economic growth. Using a Dixit-
Stiglitz product variety approach to increasing returns to scale technology,
we find that gains from elimination of barriers to firm entry and to FDI will
increase GDP by 3.2% in the medium term and by as much as 5.8% in the
long run, after the capital stock has had time to adjust to a new, higher
marginal product. These gains come mostly from improved factor produc-
tivity and lower costs of production.

We note that some caution is warranted because there exist several poten-
tial impediments to firm entry and foreign investment that have not been
included in the model. One example is the high incidence of violence in
Colombia. A higher level of domestic uncertainty will limit investment com-
pared to a region where property rights and civil law are well-enforced.
Also, a poorly designed reform program can substiantially undercut the
benefits of liberalization. In a report to the Brazilian government, Fink et.
al. (2003) discuss common pitfalls:

...For example, if privatization of state monopolies to pri-

vate owners is conducted without the introduction of com-
petition, the result may be merely transfers of monopoly rents
to private (sometimes foreign) owners. This has partly been
the case in the Colombian wireless communications mar-
ket. If increased entry into financial sectors is not accom-
panied by adequate supervision and full competition, the
result may be insider lending and poor investment decisions.
If policies to ensure universal service are not put into place,
liberalization need not improve access to essential services
for the poor. Managing reforms of these service sectors
requires complementing trade-openess with the appropri-
ate regulation.

187
FTAA and Service Liberalization in Colombia
Miles K. Light

Agriculture Agricultural subsidies and protection are usually the most dif-
ficult barriers to dismantle. This study finds that there are gains from trade
in agriculture, but that a cautious and measured pace of tariff reduction will
is not costly to the overall economy. That is because the remaining pro-
tected agricultural sectors do not constitute a large component of overall
GDP. The agri-chemical industry may deserve more attention.

A. Directions for Further Research

Multiple Households and Poverty Reduction A key consideration for de-
veloping countries is the effect of free trade upon poverty. A microsimulation
analysis which combines the IRTS (macro)scale effects with a standard
household living survey (microeffects) would improve the understanding of
the trade-poverty interaction substantially. Development of this technique
is currently underway in the Trade Research Division of the World Bank.

Service Export Potential Existing data for Colombian exports reports very
limited exports in the service sector. While this may have been the case in
the past, the current surge in services trade between OECD countries and
developing countries like India and China suggest that there is a large po-
tential market for exports of Colombian services. An analysis of the Co-
lombian comparative advantage in professional services could help identify
how Colombia can take advantages of technilogical improvements for
tranporting information.

Data Improvements for the Services Sector A major difficulty in assessing

the potential welfare and production gains for Colombia in the services
sectors is the lack of concrete data. This is not by accident. For the most
part, there are no standards for accounting and tracking service provision
and trade. Any advances in the understanding of service provision and
trade by the Colombian government will in turn improve the precision of
economic analysis dramatically. Many standards have been proposed un-
der the General Agreement for Trade in Services (GATS), should be im-
plemented by the Colombian government.

188
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
References:
Brown, Drusilla and Robert Stern (2001), “Measurement and Modeling of
the Economic Effects of Trade and Investment Barriers in Services,”
Review of International Economics, 9(2): 262-286.

Dixit, A. and J. Stiglitz (1977), “Monopolistic Competition and Optimum

Product Diversity,” American Economic Review, 76(1):297-308.

Ethier, W.J. (1982), “National and International Returns to Scale in the

Modern Theory of International Trade,” American Economic Review,
72(2):389-405.

Findlay, Christopher and Tony Warren (eds), Impediments to Trade in

Services: Measurement and Policy Implications, (London:
Routledge), 2000.

Francois, Joseph F. (1990a), “Trade in Producer Services and Returns

due to Specialization under Monopolistic Competition”, Canadian
Journal of Economics, 23:109-124.

Francois, J.F., McDonald, B.J., and Nordström, H. “Capital Accumulation

in Applied Trade Models,’’ in Dynamic issues in applied
commercial policy analysis: A handbook, Cambridge University
Press, 1997.

Harrison, G.W. , Rutherford, T.F., and Tarr, D.G. “Economic Implications

for Turkey of a Customs Union with the European Union’’, European
Economic Review, 41(3-5):861–870, April 1997.

Jensen, J., D. Tarr, and T. Rutherford (2004) “Russian Accession to the

WTO: Economic Impacts with Increasing Returns to Scale’’, World
Bank Economic Archives, forthcoming

Hansen, T. and Koopmans, T.C., “On the Definition of Computation of a

Capital Stock Invariant under Optimization,’’ Journal of Economic
Theory, Vol. 5, No. 3, December 1972, pp. 487-523.

Krugman, P. (1991), Geography and Trade, Cambridge, Massachusetts:

MIT Press.

189
FTAA and Service Liberalization in Colombia
Miles K. Light

McGuire, Greg (2002), “Trade in Services – Market Access Opportunities

and the Benefits of Liberalization for Developing Economies,’’,
UNCTAD Policy Issues in International Trade and Commodities
Study Series No. 19. United Nations, New York and Geneva.

Rutherford, Thomas F. and David Tarr (2002), “Trade Liberalization and

Endogenous Growth in a Small Open Economy,” Journal of
International Economics.

Rutherford, Thomas F. (1999), “Applied General Equilibrium Modeling

with MPSGE as a GAMS Subsystem: An Overview of the Modeling
Framework and Syntax”, Computational Economics.

Rutherford, T., M. Light and G. Hernandez, (2002a) “A General Equilib-

rium Model for Tax Policy Analysis in Colombia’’, Colombian
Department of National Planning Archives.

Rutherford, T., M. Light, “A Dynamic General Equilibrium Model for Tax

Policy Analysis in Colombia’’, Colombia Department of National
Planning Archives.

190
 DESARROLLO Y SOCIEDAD
MARZO DE 2004 53
Data Tables
Table A-1. Sectoral Definitions.

Identifier Description

Constant Returns Sectors

COF Coffee
CRO Other crops
LVS Livestock
FFH Forestry fishing and hunting
OIL Oil
MIN Other Minerals
THR Coffee Threshing
FOD Foodstuffs
NRI Natural Resources Intensive Industries
NSI Non-skilled Labor Intensive Industries
CON Construction
ELE Electricity Gas and Water
COM Communications
GOV Government Services

Increasing Returns Goods

HTC Capital and High Technology Industries

Increasing Returns Services

COM Communications
TRN Transport
SER Private Services

Table A-2. Labor Categories.

Identifier Description

UFS Urban formal salaried work

UFN Urban formal non-salaried work
UTC Urban traditional contract work
UMC Urban modern contract work (consulting)
RSW Rural salaried work (organized farming work)
RNW Rural non-salaried work (farming)

191
FTAA and Service Liberalization in Colombia
Miles K. Light

Table A-3. Labor Allocation Between Sectors for Each Labor Type.

%-Y UFS UFN UTC UMC RSW RNW

SER 29 32 56 59 28 11 16
GOV 13 33 62 15
FOD 10 5 5 2 0 2 2
HTC 9 7 5 3 1 1 1
CON 8 5 12 10 4 2 2
TRN 6 5 12 17 0 2 3
NSI 5 4 5 4 0 1 1
CRO 4 28 32
ELE 4 3 0 0 0 1
LVS 3 22 25
OIL 3 3 1 1 1 1 1
COM 2 1 2 3 0 0 1
COF 1 10 11
MIN 1 1 1 1 1 2 1
NRI 1 0 0 0 0 0 0
FFH 1 2 2
THR 0 0 0 0 0 0 0
TOTAL 100 100 100 100 100 100

Table A-4. Structure of Value-Added in Colombia.

%-Y %-VA %-L/VA %-K/VA %-SKL/L

SER 29 64 63 37 70
GOV 13 65 81 18 80
FOD 10 26 51 48 76
HTC 9 33 51 46 82
CON 8 47 52 48 71
TRN 6 43 89 10 64
NSI 5 31 64 33 74
CRO 4 81 98 2
ELE 4 56 25 73 80
LVS 3 69 94 6
OIL 3 46 44 54 71
COM 2 66 35 63 58
COF 1 84 96 4
MIN 1 84 88 12 55
NRI 1 63 31 67 53
FFH 1 81 87 13
THR 0 32 34 65 76

Key:
%-Y: Percentage of Gross Output - Industry Size
%-VA: Value added as a percentage of total product value.
%-L/VA: Labor’s share in total value-added.
%-K/VA: Capital’s share in total value-added.
%-SKL/L: Skilled labor as a share of total labor for a given industry.

192