Content uploaded by Keun Lee
Author content
All content in this area was uploaded by Keun Lee on Dec 11, 2017
Content may be subject to copyright.
1
Forthcoming in Economics of Transition
Asymmetric Trade Protection leading not to Productivity but to Export Share Change:
The Korean Case from 1967 to 1993
Hochul Shina and Keun Leeb
Seoul National University
Abstract
This paper revisits a classical theme in economics, that is, the relationship between trade
protection and economic performance, with an improved treatment of the endogeneity of tariffs
and with consideration of alternative performance criteria. The paper also considers the effects of
the asymmetric protection, such as higher tariffs on consumer goods and lower tariffs on
producer goods. Using sectoral data on Korean manufacturing during the period from 1967 to
1993, this study finds that the effect of trade protection by tariff tends to show up not in terms of
total factor productivity but in terms of revealed comparative advantage and export shares of
sectors. Such an effect tends to be greater in consumer goods, which are the main targets of the
promotion by higher tariffs. This study verifies the potentially positive role of tariffs under
certain conditions, especially under discipline from world markets.
Keywords: industrial policy, tariff, productivity growth, revealed comparative advantage (RCA),
export shares, trade protection, economic growth
Econ Lit classification codes: O19, O20, O24, O25
* An earlier version of this paper was presented at the AsiaPacific Economic and Business
History Conference 2012 held in Canberra, Australia. The authors thank the editor and the two
referees for this journal, as well as Price Fishback, Sokbae Lee, Jee-Hyeong Park, and Elias
Sanidas for their valuable comments. The second author acknowledges the support provided by
the Korean government through the National Research Foundation of Korea (NRF-2010-330-
B00093) and the WCU Program (R32-20055).
aDoctoral student, Seoul National University, bProfessor, Economics Department, Seoul National University
2
1. Introduction
The relationship between trade protection and economic development may be an old
theme in economics, but it remains controversial. The free trade theory proposed by Adam Smith
(1776) and the concept of infant industry protection proposed by Friedrich List (1841) continue
to cause controversy today. The HeckscherOhlin (HO) model, which is a modern standard trade
theory, states that trade protection reduces economic welfare by causing price differences
between the world market and domestic market because a small country cannot
change the terms of trade by tariff. However, if dynamic increasing returns, external economies,
and market failure exist, trade protection may increase welfare. Furthermore, the HO model is a
static model and does not consider economic growth. As the theoretical debate goes on, the
empirical evidence remains incomplete.
Empirical research on trade protection and economic development may be conducted in
the country, industry, or firm level. Among the many previous studies that took a country-level
approach, the studies by Sachs and Warner (1995), Wacziarg and Welch (2008), and Edwards
(1997) are the most notable. These papers used country-level data after World War II to report a
significant and positive relationship between free trade and economic growth. However,
Rodriguez and Rodrik (2001) argued that the studies conducted by Sachs and Warner (1995) and
by Edwards (1997) and
they omitted some important determinants of economic growth. By analyzing the cases of 60
countries from 1975 to 2000, DeJong and Ripoll (2006) also argued that the effect of trade
observed that the negative relationship
between trade protection and economic growth, reported in many previous papers, could be
confirmed only in developed economies, and that an insignificant or a positive relationship
existed in developing countries. Some studies, such as those by ORourke (2000) and by
, reported that high tariff, especially manufacturing tariff, helped
economic growth by facilitating resource reallocation from agriculture to industries in ten
Western countries from 1875 to 1914. Vamvakidis (2002) extended the research period covered
in previous studies by analyzing the relationship between trade protection and economic growth
from 1870 to 1990. He further divided the research period into four. He found that a positive
relationship between free trade and economic growth existed only from 1970 to 1990, and that
no significant relationship existed in other periods. He also found a positive relationship between
trade protection and economic growth in the 1930s.
However, these country-level studies have a common limitation: they did not consider the
fact that the degree of trade protection differs between industries, and that the government may
strongly protect only those industries that it wants to support. In such a case, the effect of trade
protection differs between industries. Thus, an industry-level analysis is needed. For instance,
Ferreira and Rossi (2003) analyzed the Brazilian manufacturing case from 1985 to 1997,
reporting that Brazil trade protection was negative and significant to the growth rates of
industry productivity and industry labor productivity. A well-known industry-level study on the
relationship between trade protection and economic growth is that of Lee (1996). Lee (1996)
classified Korean manufacturing into 38 industries from 1963 to 1983 to analyze the effect of
3
trade protection on the growth rate of industry value added per labor hour and industry total
factor productivity (TFP) growth rate using five-year-term panel data. He reported that nominal
tariff was negative and significant to the growth rate of labor productivity and TFP at the sector
level. Nam and Lee (2005) and Lee (2000) also analyzed the Korean manufacturing case from
1984 to 2000 and from 1969 to 1993, respectively. They reported mixed results on the effect of
trade protection on economic growth.
While the picture appears to be mixed even when industry-level research is used, an
important problem in these studies was that the possible endogeneity of trade protection through
both effective and nominal tariffs was not rigorously addressed. Moreover, these studies used
productivity change as the criterion to assess the policy impact. As Krugman (1994) and Young
(1995) argued, other criteria, such as export growth and share change, may also be considered as
more relevant to developing countries that tend to pursue growth (output) rather than efficiency
(financial efficiency or TFP). These studies characterized Asian growth as input-driven rather
than TFP-driven. This recognition is consistent with the classical literature on industrial policy,
such as that by Johnson (1982), which defined it as an intervention to promote the growth of
certain sectors at the expense of others.
The current study employs the similar data set and periods as those used by Lee (1996) to
revisit the impact of industrial policy with a new and improved treatment of the endogeneity of
tariff using instrument variables. The study also looks into the impact of industrial policy in
terms of alternative criteria for assessment, such as Balassas revealed comparative advantage
(RCA) and the change in the export shares of sectors. Another contribution is sought by
analyzing the effects of the asymmetric nature of protection, wherein consumer goods targeted
by industries for export promotion are charged with high tariffs, and producer goods not targeted
for promotion are charged with lower tariffs.
In sum, this paper examines the effect of trade protection in Korea during its catching-up
growth period from 1967 to 1993 at the manufacturing level divided into 36 sectors.1 The paper
contributes new findings in relation to the impact of trade protection by tariff. This impact tends
to appear not in terms of TFP but in terms of RCA and export shares, and is more prominent in
consumer goods, which were the main target for export promotion from 1967 to 1993.
In Section 2, industrial policy is described to derive empirical hypotheses. A
discussion of methodologies, especially on how to measure trade protection, and an estimation
method are presented in Section 3. Section 4 provides the estimation results, and Section 5
concludes the paper.
2. Overview of Korean Industrial Policy and Hypotheses
1
capital stock data were obtained,
provided data on 36 industries. Except for two industries, the industry classification used in the current
paper is similar to that used by Lee (1996).
4
Industrial policy is a very broad concept. It refers to policies concerning the structure of
the domestic
competitiveness (Johnson 1982). Variants of industrial policies have existed in countries that
have succeeded in terms of industrialization, such as the United Kingdom from the 14th to the
18th centuries, the United States and Germany in the 19th century, Japan in the late 19th century,
and Korea and Taiwan in the late 20th century (Cimoli et al. 2009). While diverse tools may be
used to protect infant industries, tariff protection, which is the focus of the present study, is one
of the most traditional and commonly used tools.
Evolution of Industrial Policy in Korea
After the Korean War and during the 1950s, the Korean economy was supported by aid
from the United States and the United Nations. However, the US began to reduce aid2 after 1957.
Thus, Korea needed to meet foreign exchange demands through exports. Hence, in the 1960s, the
new government, led by the former general and modernizer President Park Chung-hee, adopted a
policy to boost exports from labor-intensive and consumer goods industries.
In the early part of that decade, a controversy arose between the economic growth
strategy driven by consumer goods and that driven by producer goods (Lee 2004). However, the
Korean government adopted the economic growth strategy driven by labor-intensive and
consumer goods industries, unlike some other developing countries such as India or Brazil,
which pursued import substitution industrialization. The Korean government imposed high tariffs
to protect the consumer goods industry. The weighted average tariff, based on Hong
and Kim (1996) industry real production data as weight, was 74.82% in 1966 and 79.74% in
1968, as shown in Figure 2.1(a).
The Korean government implemented many other kinds of export promotion policies in
addition to tariff. It formulated complex support policies for exporting firms and enterprises,
including compensation to encourage exports, reduction in corporate and income taxes,
exemption on tariff for imported raw materials to be used in producing goods for export,
financial support, and reduction in electrical bills and railroad fares (Kang et al. 2008).
The Korean government also determined a target export value based on major
commodities and major export markets (Korean Economy Compilation Committee 2010).
Furthermore, an export promotion conference attended by the president, minister, and the
officers of major export firms was held every month. The Korean government encouraged
business leaders to meet the export target, and during the conference, the president immediately
solved the difficulties faced by firms. Apart from adopting complex support policies, the
government regularly checked industry performance based on a clear export performance
standard. As a result, the government was able to reduce possible unproductive
abuse of the supporting policies.
2 The US began to cut aid as a kind of dollar defense policy after 1957 because of the deterioration in the
international balance of payment of the US. Korean Economy Compilation Committee, The Korean
Economy: Six Decades of Growth and Development, 2010, Vol. 2, p. 180.
5
In the mid-1970s, the main target sectors of the industrial policy became capital-intensive
industries such as electronics, shipbuilding, iron and steel, automobiles, and so on. Some of them
were already supported by an industry promotion law legislated from 1967 to 1971 under the
second five-year economic development plan. However, more intensive support was granted to
these industries after 1973, including loans from the national investment fund and the Korea
Development Bank, exemption from internal tax and tariff, and the establishment of vocational
schools as well as government research institutes to meet the human resource needs of the heavy
and chemical industries and to develop technology, respectively.
The percentage of policy financing that offered preferential interest rates was less than 40%
in 1971 but over 55% in 1976 and 1977, after the heavy and chemical industrialization plan was
implemented. The percentage reached 70% in 1978 when support for the heavy and chemical
industrialization plan was at its peak (Haggard 1990). In addition, the effective tax rate was
generally less than 20% for the capital-intensive industry and nearly 50% for the light industry
(Korean Economy Compilation Committee 2010).
Since the 1980s, the Korean economic landscape has been characterized by gradual trade
liberalization (lower tariff), and the focus has again shifted toward encouraging in-house research
and development (R&D) by private businesses led by large corporations or the so-called
Chaebols (Lee and Kim 2010). The economy experienced a depression in the early 1980s after
the second oil shock in 1979, which weakened some industries. The new military regime that
came into power in 1980 found it necessary to change the previous selective-supporting
industrial policies and introduced the idea of liberalization.
The new military regime phased out several industrial policies in accordance with the
global trend toward a more liberal economic order. The Korean government abolished the
existing industry promotion law that supported specific industries and enacted the Industry
Development Law in 1986, which focused on functional intervention neutral to all industries.
Instead, the Korean government put emphasis on technology development, for instance, by
-based Technology Development Project. In addition, it provided tax
exemptions and financial support to enable firms to establish in-house R&D centers (Lee and
Kim 2010). These supporting policies were implemented to enhance a new kind of
competitiveness not based on cheap wages or currencies but on product differentiation and
innovation capabilities in higher-end goods. The same kind of competitiveness was also
necessitated by the increasing wages in the domestic economy and the rise of next-tier catching-
up economies offering lower wage rates.
The upgrade via R&D and innovation since the mid- or late 1980s was a turning point in
Korean economic history because it marked the beginning of efforts to transition from input-
driven growth to productivity-driven growth. As argued in studies such as those by Krugman
(1994) and Young (1995), East Asian economies, including Korea, grew based on its labor and
capital inputs but not so much on TFP. Overall, the above elaboration illustrates that Korea
economic growth was led more by input investment and gross output (export) growth in the
earlier period, and by productivity only at the later period. The discussion also suggests that
while tariffs and other protections led to export expansion earlier, such export growth later
stimulated productivity growth through its discipline and learning-by-exporting effects,
6
combined with the shift of government activism toward the promotion of in-house R&D by tax
exemption.
In sum, the goal of the government the growth of output and
exports at the earlier period, and to enhance productivity at the later period. If that is the case, we
also have to examine the impact of industrial policy and trade protection with these changing
goals in mind. Considering that Jung and Lee (2010) proved that productivity increased at the
later period (19852005) owing to R&D, exports, an oligopolistic market structure, and their
interactions, this study focuses on the earlier period to see if government action led to export
growth rather than to productivity.
Hypothesis One: Not TFP but RCA and Export Shares
The fact that economic growth in East Asia was driven not by TFP but by inputs and
investment is not surprising. The documents on Korean industrial policy emphasized not so
much on the importance of TFP but more on the increasing output and exports and related market
shares in the world export market. In addition, classical literature on industrial policy did not
mention TFP increase as the goal of industrial policy. Rather, it stated that industrial policy is an
intervention bias toward certain sectors, an attempt to increase a particular share
(Johnson 1982). This consideration of alternative criterion for assessment of the impact of
industrial policy is also consistent with some research such as that of Cheong et al. (2010), which
employed a formal model to derive the behavior pattern of latecomer firms, led by business
groups or conglomerates. They proved through empirical analysis that that these companies
pursued growth by aggressive investment rather than by financial efficiency.
Thus, in the current study, we hypothesize that the impact of industrial policy, specifically
tariff protection, would become evident not much in TFP but in export performance that can be
represented by RCA (Balassa and Balassa 1984) and/or changing export shares. RCA is
defined as the share of certain sectors exports by a concerned country in the total world exports
of that sector divided by a countrys share in the worlds total exports. Thus, RCA is a good
measure of a country relative export performance in certain sectors. If index is
greater than one in a certain sector, then that country has a comparative advantage in that
particular sector.
The second measure we are using share in total manufacturing
exports, which is more consistent with the original definition of industrial policy. If a sector is
favored by government intervention in the form of asymmetric tariffs, then that sector is
expected to increase its share in exports.
The next question relates to the exact mechanism of high tariff protection that would lead
to better export performance. We can argue that higher tariff can translate into higher profit
margins under oligopolistic markets, which can then be used as investment funds for expansion
of factories at subsequent rounds. However, the actual link between more profits and more
investment is not automatic but is conditional upon discipline from market competition. This
condition is the crux of debates in economics. Schumpeters classic argument is in favor of big
businesses as innovators, whereas the critics leans toward monopolistic firms facing
7
the danger of being complacent and not being aggressive in investment. In the Korean context,
the favored large firms were enjoying the status of oligopoly but were subject to discipline from
export markets, and the government-provided privileges were linked to their export performance
(Lee 1992). A study by Jung and Lee (2010) also suggested an evidence of the discipline effect
of export orientation. They found that
degree of oligopoly in the sector where these firms belonged to had a positive effect on
productivity growth at the later period or from the mid-1980s to the 2000s.
Analyzing the structures of Korean manufacturing commodities in 1970, 1977, and 1982,
Lee, Urata, and Choi (1986) reported that more than 80% of commodities were produced under a
monopoly or oligopoly. They also reported that in highly protected industries, the price-cost
margin, which is defined as (value added - wage)/sales, was higher than the price-cost margin in
weakly protected industries in 1973, 1978, and 1983. Although firms could have earned higher
profits through trade protection in domestic markets, they had to sell at world market prices to be
able to export. Yang and Hwang (1994) analyzed six Korean manufacturing industriestextiles,
wood and wood products, chemicals, mineral and mineral products, metal and metal products,
and machines and equipmentfrom 1976 to 1990 using the Branson and Marston (1989) model.
They reported that Korean firms behaved like oligopolists in the domestic market, but acted like
price takers in the international market and imposed price discrimination between the domestic
and international markets.3
An example is the first Korean-branded car, Pony. It was developed by Hyundai Motors
and achieved excellent results in 1976, as it earned 44% share of the domestic car market
(Hyundai Motors 1997). Trade protection was one reason for the success of Pony. The Korean
car market was highly protected by an 82% tariff in 1977. A Pony car was sold for 4,500 USD in
Korea in 1976. In comparison, small cars made by Japanese or West German companies were
sold for 2,300 USD in the US market at that time. Hyundai Motors sold Pony for 4,500 USD in
the domestic market and for 1,850 USD in foreign markets because of price competition from
Japanese and West German cars. Thus, Hyundai Motors could start to sell to foreign markets
using domestic profits. Without trade protection, Hyundai Motors would not have been able to
develop a new car with no price competitiveness.
Hypothesis Two: Effects of Asymmetric Tariffs Across Consumer and Producer Goods
One of the most important but unnoticed aspects of industrial policy in Korea was the
asymmetrical structure of tariffs. That is, the government charged a very high tariff for consumer
3 Yang and Hwang (1994) analyzed how the rate of change of the US price index and exchange rate and
the rate of change of petroleum import price affected the growth rate of the domestic price indices and of
export price indices of six Korean manufacturing industries. The growth rates of the domestic price
indices were not affected by the rate of change of the US price index and exchange rate, but was affected
almost one-to-one by the rate of change of the petroleum import price. However, the growth rates of the
export price indices were very sensitive to the rate of change of the US price index and exchange rate, but
they were slightly affected by the petroleum import pricerate of change.
8
goods and a much lower tariff for producer goods. The rationale behind the imposition of low
tariffs for producer or capital goods was that the export-oriented consumer industry needed to
import these goods, and that this particular industry was not targeted for promotion as an engine
of export. The weighted average nominal tariff rate from 1966 to 1993, which is the key
industrial policy in the current paper, is shown in Figure 2.1(a).
[Figure 2.1(a) and 2.1(b)] [Figure 2.2]
As shown in the figure, the weighted average tariff in the consumer goods used to be as
high as 80% in the second half of the 1960s, whereas it was only half of that level in producer
goods, or 36.86% and 42.11% in 1966 and 1968, respectively.4 Although tariffs had been
gradually reduced since the mid-1970s,5 the gap between consumer and producer goods
remained at 60% versus 30% in the mid-1970s, and at 40% versus 20% in the mid-1980s. The
gap narrowed only in the mid-1990s as the government shifted its focus on promoting producer
goods industries largely led by the SMEs, as opposed to the final goods industries claimed by big
businesses, which had become too dominant in the economy (Kim and Lee 2008). The mid-
1990s also marked the period when the Ko openness as measured by tariffs
closely reached the openness level of advanced economies. In 1993, the weighted average tariff
in the producer goods industry was only 8.03%, and the weighted average tariff in the consumer
goods industry dropped to as low as 11.26%. These levels were comparable to those of
developed countries. In 1979, as the Tokyo Round concluded, the weighted average tariff in
manufacturing using import as weight was 6.9% in the European Communities, 6.0% in Japan,
and 5.7% in the US (Balassa 1984; Table 1). In general, big businesses had been leading the final
consumer goods, which had been the main target of industrial policies for export promotion. On
the contrary, SMEs had been leading the intermediate goods, which had been less competitive
internationally, in addition to not being the main engines of exports (Kim and Lee 2008).
This asymmetry in tariff makes it interesting to examine the possibly different impact of
tariffs on consumer versus produced goods. However, we cannot be free from a certain degree of
arbitrariness when it comes to sector classification, which exists in every classification method.
Nevertheless, the examples of tariffs for selected items of consumer and producer goods shown
in Figure 2.1(b) confirm that this asymmetry indeed existed. For instance, for typical export
promotion items such as TV sets and phones, the tariffs were as high as 100% or 70% around
1970. By contrast, the tariff for lathe was as low as 20% in 1970, and those for transportation
shaft, crank, and gear was 50% in the 1970s and decreased to 20% in 1974.
Essentially, our data show that RCA in the consumer goods industry tend to be higher
than that in the producer goods industry. This finding is consistent with the fact that economic
4 The appendix contains the data sources and data-producing methods used in the current study.
5 The trend of the non-tariff barrier was somewhat different. The non-tariff barrier for the producer goods
the non-tariff barrier for the consumer goods industry decreased slightly from 72.79% in 1970 to 71.31%
in 1975.
9
and export growth in Korea used to be led by consumer and labor-intensive industries. Thus, we
may hypothesize that the performance change effect (RCA and export share) would be greater in
high-tariff sectors of consumer goods than in the producer goods industry.
One rationale for such hypothesis is the fact that producer industries, especially those
engaged in the production of machine tools and parts as well as industrial materials, had not been
very export-oriented, whereas consumer or final goods industries had been so, as shown in
Figure 2.2. Then, to the extent that export orientation mattered as a disciplinary device for rent
beneficiaries, we can suppose that the marginal effects of tariffs would have been higher in
consumer goods industries where more of the rents would have been used for investment.
We can also discuss another possible reason for the lower investment propensity of
producer goods. As discussed by Lee and Lim (2001) and Kim and Lee (2008), the technological
regime of intermediate capital goods, such as machine tools, made it difficult to promote them
through industrial policy. In many of the intermediate goods industries, tacit knowledge that
accumulated from the interface between producer and customer firms is very important (Pavitt
1984). Thus, the latecomer firms were not able to catch up simply by importing production
equipment and buying licenses for product design and engineering. Additionally, because the
quality of the machine tools and parts directly employed determines the quality of the final goods,
the producers of these goods (typically big businesses in consumer goods) were sensitive to the
quality of their own products and did not want to use domestically produced machine tools. In
such a situation, even government policies that encouraged the use of domestic products were
not effective (Kim and Lee 2008). Furthermore, the weak domestic market, not to mention the
poor export markets, provided no opportunity to accumulate tacit knowledge by expanding
production and interacting with more customer firms. This aspect is one of the fundamental
differences between the intermediate goods industry and other final goods industries, such as
those engaged in automobiles or consumer electronics (Lee and Lim 2001). Thus, given this
lower expected rate of return to investment in intermediate goods, the same amount of rents
generated by tariffs would lead to less investment, and consequently less export expansion.
Actually, this difference in the rate of return is one reason all large businesses prefer to be
involved in final goods production, leaving the intermediate production to the SMEs.
3. Methodology and Estimation Models
Measuring the Level of Protection
Representing the degree of trade protection can be done in two ways: by nominal tariff
rates, and by the effective rate of protection (ERP). We explain in the following paragraphs why
we did not use the ERP, which was what Lee (1996) employed. Then, we discuss the issue of the
possible endogeneity of tariffs and the ways to handle this endogeneity by choosing proper
instruments.
The ERP is widely used as a proxy for trade protection, but a high-quality data set is
required to estimate it properly. A critical issue is correctly reflecting the difference between
10
international and domestic prices. Furthermore, this difference is affected not only by tariffs but
also by diverse factors including government subsidies, finance, R&D, geography, and so on. To
that extent, the ERP is not a precise, separable measure of trade protection.
Given the data requirement, an estimation of effective rates of protection in Korean
manufacturing is possible only for the period after 1975, as in that by Hong (1997).6 Existing
studies tend to produce an unstable and unreliable estimation of ERP. For example, when using
applied the Cohen method, the following ERPs were obtained: -
17,819.2% for leather footwear in 1993, -13,926.1% for seasoning in 1980, 18,140% for base
products in the petrochemical industry in 1978, and -34,440.2% for copper in 1975. The absolute
values of these estimations were larger than 10,000%, probably because of errors in the
calculation of value added using international and domestic prices.
Therefore, following Lee (1996), we decided to use nominal tariff rates, as these are more
stable. For instance, measured in terms of three-year sub-periods, the standard deviation of
which was five times as high as the
standard deviation of the industry nominal tariff. Maximum industry ERP was at 1,846%, and
minimum industry ERP was at -844%, which was more unstable than nominal tariff (maximum
industry nominal tariff was at 150%, and minimum nominal tariff was at 3.43%).
However, the estimation using nominal tariff may be biased because of reverse causality
or omitted variables. As argued by Muendler (2001), the government may impose high tariffs on
a low-productivity industry. Lee (2007) also pointed out the possibility that industries with
higher productivity levels and higher concentrations may obtain trade protection through
lobbying. Karacaovali (2006) reported that based on an analysis of Colombian industries, trade
protection is stronger in sectors with high productivity growth.
Considering the problem of endogeneity, Lee (1996) used one-period-lagged tariffs as the
instrument for current period tariffs. However, a one-period lag may not satisfy
exogeneity conditions because it can be correlated with omitted variables such as non-tariff
barriers. Thus, the current study used other instruments. Harding and Rattso (2010) analyzed the
effects of tariff on labor productivity growth at the industry level in South Africa using the tariffs
of Latin America and the Caribbean, the Middle East and North Africa, and South Asia as
instruments.
The current work employed the Argentine and Kenyan tariffs as instruments for the
Korean industrial tariff. Kenya as well as Argentina and Korea joined the General Agreement on
Tariffs and Trade (GATT) in 1964 and in 1967, respectively. All three countries, which were
treated as developing economies, participated in several GATT negotiations. They were also
affected by international negotiations on tariff reduction. Thus, Argentine and Kenyan tariffs may
be related to the Korean industrial tariff. The current study used the rate of import duty derived
from dividing import duties by imports for Kenyan industries and nominal tariff data for
Argentine industries. Data for Kenyan tariff rates by industry were obtained from Mwega (1994),
and those for Argentine tariff rate by industry were obtained from Brambilla et al. (2010),
6 provided an estimation of the ERP in Korean manufacturing from 1975
to 1995.
11
Galiani and Sanguinetti (2000), and Wogart and Marques (1984). F-statistic (FE2SLS) or chi-
square statistic (RE2SLS) in the first-stage regression was used for the relevance test of the
instruments. The instrument is usually considered relevant if the value is over ten.
The Argentine and Kenyan tariffs can affect Korean manufacturing exports to Argentina
and Kenya as well as the Korean nominal tariff. However, Argentina and Kenya were not major
markets for Korean exports, and their shares were negligible compared with the total Korean
exports. The graph in Figure 3.1 shows the ratio of Korean manufacturing exports to Argentina
and Kenya to Korean manufacturing exports to the world from 1967 to 1993. During that period,
the share of Korean manufacturing exports to Argentina and Kenya was no more than 0.75%.
Thus, Argentine and Kenyan tariffs are not likely to affect the Korean economy through Korean
exports to Argentina and Kenya.
[Figure 3.1]
However, Argentine and Kenyan tariffs can also affect the Korean economy because
these tariffs can influence the export performance of respective countries if the Argentine and
Kenyan market structures are oligopolistic (a similar point was tackled in the Korean case). Thus,
export performances of Argentina and Kenya might affect that of Korea in the world market if
these three countries have similar export structures. Thus, we checked the export structures of the
three countries, and classified Korean, Argentine, and Kenyan manufacturing exports to the
world in 1970, 1980, and 1990 into 36 industries. The export shares of each industry to total
manufacturing exports were calculated using . Next, the correlation
coefficient between the export shares of 36 industries of any two countries was calculated. The
results are shown in Table 1.
[Table 1]
Taiwan and Japan were included for comparison. Table 1 shows that the correlation
coefficients between Korea and Argentina as well as between Korea and Kenya were extremely
small at 0.01 to 0.1777. The correlation coefficients gradually became smaller because of Korean
industrialization. Meanwhile, the Korean export structure was very similar to that of Taiwan, and
was gradually becoming more similar to that of Japan. However, the correlation coefficient
between Argentina and Kenya was high, ranging from 0.6 to 0.97, because the two countries
mainly exported food. The three main export industries of Korea, Argentina, and Kenya are
shown in Table 2.
[Table 2]
Shares of the food industry in Argentine manufacturing exports ranged from 47% to 79%,
and shares of the food industry in Kenyan manufacturing exports ranged from 30% to 70% in
1970, 1980, and 1990. However, the food industry in Korean exports only ranged from
0.9% to 4.6% from 1967 to 1993. As shown in Table 2, the major export industries of Kenya and
Argentina, including food, leather, and petroleum refineries, were resource-based. However,
12
major export industries were labor-intensive industries, such as apparel and textiles, and
technology-intensive industries such as sound, image, and communication equipment. Thus,
major export industries were different from those of Argentina and Kenya.
In our estimation, the Korean tariff rate by industry was used as the suspected
endogenous variable, and the Argentine and Kenyan tariff rates by industry were used as the
instrumental variables. The SarganHansen statistic was used to test the over-identifying
restrictions of instruments. If the p-value of the SarganHansen statistic is over 0.1, the
instrument is usually considered exogenous. A two-stage random-effect model and a two-stage
fixed-effect model were used as estimation methods. The Hausman test was conducted to select
the appropriate model between the two. If the Hausman test rejected the consistency of the two-
stage random-effect model, the two-stage fixed-effect model was selected. Otherwise, the two-
stage random-effect model was chosen.
Estimation Model
The estimation method was similar to that specified in the panel estimation of Lee (1996).
However, we modified the model by including more dependent variables. The basic estimation
equation is as follows:
′
The duration for analysis was 27 years, from 1967 to 1993. This period was divided into
nine three-year terms. Thus, t is the time subscript, which can range from 1 to 9, and i is the
industry subscript, which can range from 1 to 36. All variables were industry-level ones. The
estimation method was the two-stage fixed effect model (FE2SLS) or the two-stage random
effect model (RE2SLS).
refers to the three dependent variables we considered, namely, industry-level TFP
growth rate, industry-level standard RCA index7, and the share of each industry in total
manufacturing exports (export share). Thus, three estimation equations were used. Data for
industry TFP growth rate were obtained from Hong and Kim (1996). The industry TFP index was
calculated through the growth accounting method using the translog method.8 To calculate the
RCA, the National Bureau of Economic Research (NBER) database compiled by Feenstra and
Lipsey was used to obtain the world export data and the Korean export data. In calculating the
export share of industries, Korean industry-level export data were sorted from the Statistical
Yearbook of Foreign Trade of Korea based on our 36 industry classifications.9
7 Industry RCA Index was calculated using the ratio of Korean specific industry exports to world specific
industry exports divided by the ratio of Korean manufacturing exports to world manufacturing exports.
8 A detailed description of the data is in the Appendix.
9 We used this data because we considered it more reliable than Korean export data in the NBER.
However, the calculated export shares of 36 industries using two sources were very similar. Their
13
is a constant term and refers to industry-level nominal tariff rate in Korea.
Data were obtained from the tariff schedules of Korea (see Appendix for details). First, we
calculated a simple-averaged value of the legal or temporary tariff rate on imports at the CCCN
or HS 24 digit level. Then, we calculated the industry-level tariff rate using the value of the
output in the input-output table of Korea as a weight.
refers to the control variables vector, which includes the variables of the ratio of
bank debt to total asset (Finance), ratio of private R&D expense to sales (Research), logarithm of
initial value added per worker, and the logarithm of initial capital stock per worker when the
dependent variable is the industry TFP growth rate. The ratio of bank debt to total assets was
included to control capital mobilizing ability, as capital was allocated mainly by banks in Korea
from the 1960s to the 1990s. The ratio of private R&D expense to sales was included to control
innovation capability, and the logarithm of initial capital stock per worker was included to
control initial capital stock difference between industries. The logarithm of initial value added
per worker was included because an industry with low initial productivity can grow faster.
Lawrence and Weinstein (1999) analyzed Japanese industries and argued that the TFP growth
rate of Japanese industries that have a huge productivity gap in relation to US industries is higher
than the TFP growth rate of Japanese industries that have a productivity level similar to that of
US industries. To minimize the endogeneity problem, the initial year used was the year prior to
the period studied. For example, 1990 is the initial year of the logarithm of the initial value
added per worker variable from 1991 to 1993.
Depending on the dependent variables, different controls were added, and the details are
summarized in Table 3A. The basic descriptive statistics are shown in Table 3B. For instance,
when the dependent variable was the RCA index, the following control variables were included:
ratio of bank debt to total asset (Finance), ratio of private R&D expense to sales (Research), real
import growth rate of the US industry, domestic inflation, and industry TFP growth rate. The real
import growth rate of the US industry was included because the US was the largest export
market for Korea from the 1960s to the 1990s. The ratio of Korean exports to the US to total
Korean exports was 49.8% in 1971, 26.3% in 1980, and 29.8% in 1990. Domestic inflation was
included because the increase of domestic price reduces the competitiveness of Korean products
in the international market. Industry TFP growth rate was included because productivity growth
can increase its comparative advantage. When the dependent variable was the export shares, the
control variables included were similar to those in the RCA estimation, except when TFP growth
rate was excluded.
is the dummy variable capturing the period effect. It was included to control factors
such as oil shocks, which affect all industries at specific periods.
was included to control industry-specific effects, which may be a random or fixed
effect depending on the estimation method used. is the error term.
To reduce the endogeneity problem, we used all one-period lagged values of the variables,
correlation coefficient is 0.9381. To calculate the RCA index, we used Korean export data from the NBER
because it is the same source of our world export data.
14
except for the period dummy variable (), the logarithm of the initial capital stock per worker
variable, and the logarithm of the initial value added per worker variable. Finally, tax policies
and non-tariff barriers were not considered because of a lack of data, which may lead to an
omitted-variable bias. The possibility of this type of bias was another reason we used the
instrumental variable estimations.
[Table 3A and B]
4. Estimation Results
Estimation results using the period data are presented in Tables 4-A and 4-B. The results
in Table 4-A indicate our use of the one-period lagged tariff as an instrument, following Lee
(1996). Although we believe this method is less satisfactory in handling tariff endogeneity, we
present them here for the purpose of comparison and information. The analysis period used
covers Period 2 (1970 to 1972) to Period 9 (1991 to 1993). The regressions in Table 4-B indicate
our use of the tariffs in Argentina and Kenya as instruments. In this case, the observation number
is reduced to 204 or 219 because of the availability of Argentine and Kenyan tariff data. The
analysis period used covers Period 2 (1970 to 1972) to Period 8 (1988 to 1990). Regression
numbers 1-1, 1-3, and 1-5 are the models that omitted R&D expense per sales variable to
increase the observation number. All regressions in Table 4-A and 4-B satisfy the
relevance condition, and the regressions in Table 4-B satisfy the exogeneity condition.
Based on our estimation, the tariff is insignificant to the TFP growth rate when using one-
period lagged tariff as an instrument, but is significant and positive when using Kenyan and
Argentine tariffs as instruments. These results are quite different from the findings of Lee (1996),
which revealed a negative and significant impact of tariffs. On the contrary, the tariffs are
positive and significant in all RCA index and export share regressions in both Tables 4-A and 4-
B. Furthermore, the coefficients of tariff in the RCA regression and export share regression are
smaller when using one-period lagged tariff compared with those when using Kenyan and
Argentine tariff as instruments. This difference might be related to the endogenous nature of the
one-period lagged tariff.10
[Table 4-A] and [Table 4-B]
To check the robustness of the estimation results, estimations are also done using annual
data, and the results are reported in Table 5. Year dummy variables are used instead of period
10 Omitted variables, such as the relative wage of Korea compared with the wage of competitors, may
negatively affect export shares and the RCA and positively correlate with Korean tariff. Firms in
industries protected by high tariffs may earn higher profits. Thus, the relative wage of Korean industries
protected by high tariffs can be higher, or the Korean government might lower tariffs in industries with
high export shares and strong comparative advantages because of international pressures to reduce tariffs.
15
dummy variables. All regressions in Table 5 are models employing Kenyan and Argentine tariff
as instruments.
The regression models of 2-1, 2-3, 2-5 are without R&D to sales variables to increase
the number of observations. Given the limited availability of Argentine and Kenyan tariff data,
the analysis period is from 1970 to 1988. Moreover, the number of observations is reduced from
607 to 462 in regression models 2-2, 2-4, and 2-6 because of the limited availability of R&D data
before 1976. When the TFP growth rate is used as the dependent variable, the variables of log of
initial value added per worker and log of initial capital stock per worker are excluded.11 All the
regressions in Table 5 also satisfy the exogeneity and relevance conditions.
Our results indicate that the tariff is negative but insignificant to the TFP growth rate, and
is significant and positive to the RCA index and the export share. Furthermore, the coefficients of
tariff in the RCA regression and export share regression are greater than those in the period
estimations. These differences seem to be affected by the fact that more observations are used in
annual data estimates. To check the robustness of the estimation results further, estimation using
the seemingly unrelated regression (SUR) method is also employed, the results of which are
presented in Table 6. SUR allows correlation between error terms from different equations.
Period data are used, and 36 industry dummy and period dummy variables are employed to
control industry and period fixed effects. The estimation results are consistent with those
reported in Table 5. The tariff is insignificant to the TFP growth rate, but is significant and
positive to the RCA index and export shares. The coefficients of tariff in the RCA regression and
the export share regression are smaller than those in Tables 4-B and 5. This difference may
suggest that the SUR cannot control the endogeneity of tariff variables, and thus lead to a
downward bias in the tariff coefficient.
[Table 5 and Table 6]
Based on the estimation results presented in Tables 4, 5, and 6, we can conclude that the
effects of the tariff on RCA and export shares are positive and significant, and are also robust to
the annual versus period estimations as well as the different estimation methods of instrumenting
and the SUR. However, the impact of tariff on TFP growth is often insignificant, unstable, and
not robust. More interestingly, it is never negative and significant, which is a finding that is
contrary to that by Lee (1996). Thus, while the results are somewhat inconclusive toward the
impact on TFP, they generally support our first hypothesis and reasoning that stable and higher
profits associated with higher tariffs had led to better export performance rather than TFP growth
during the study period.
The second hypothesis deals with the impact of the asymmetric tariff structure between
consumer goods and producer goods. In Section 2, we hypothesized that the effect of tariff on
11 We have also tried to add the one-year lagged log of value added per worker variable and the one-year
lagged log of capital stock per worker variable as control variables in other regressions (not reported).
The estimation results were the same. Tariff was insignificant to the TFP growth rate.
16
RCA and export share would be larger in the consumer goods industry than in the producer
goods industry, given that rents from tariffs would have led to less investment because of the
possibly low rate of return in the latter than in the former. Separate estimation results for
consumer and producer goods, using both period and annual data, are presented in Tables 7-A
and 7-B, respectively. As before, Argentine and Kenyan tariffs are used as instruments in all
estimations.
[Tables 7-A and B]
As some of the Hausman statistics in Table 7-B are negative, we provide the regression
results using both the two-stage fixed effect and the two-stage random effect model. However,
only little difference is observed. Most regressions in Tables 7-A and 7-B also satisfy the
exogeneity and relevance conditions of the instruments. Chi-square statistics in the first-stage
regression are 8.36 and 8.17 in regression numbers 3-1 and 3-3, respectively, and are close to 10.
Estimation results presented in Tables 7-A and 7-B are similar. Tariff is significant and
positive to RCA only in the consumer goods industry. It is significant and positive to export
share in both the consumer goods and producer goods industries, but its coefficient in the
consumer goods industry is higher than in the producer goods industry. These regression results
suggest that the significance of tariff to RCA, and partly to export share as well (as shown in
Tables 46) may have been driven by the consumer goods industry. This finding is consistent
with Hypothesis Two, which supposes that during the study period, higher tariff in the consumer
goods industry, such as in clothes manufacturing, helped the participating firms to earn greater
profit and encouraged the industry to invest more, leading to higher comparative advantage in
exports.
So far, we have argued that an important source of the different effects of tariffs across
sectors is the varying extents of discipline associated with the degree of export orientation. Now,
Table 8 aims to test these effects directly by adding the interaction term between tariffs and
export orientation (exports-to-sales ratio). The results show that this interaction term is positive
and significant in the RCA equation, along with the variables of tariffs and export orientation. In
export share equations, the variables of export orientation and tariffs remain significant but the
interaction term loses significance. Although this difference seems to be associated with the close
similarity between the dependent variable (export share) and the explanatory variable (export
orientation), it implies that the interaction effect of export orientation more prominently affects
international comparative advantages rather than export structure at the domestic level.
Finally, throughout the regressions, R&D does not significantly and robustly affect TFP
or RCA. This observation is consistent with that about the changing sources of performance at
the different stages of the Korean economy because R&D turns out to be important at a later
period, as shown by Jung and Lee (2010).
[Table 8]
5. Summary and Concluding Remarks
17
This paper revisits a classical theme in economics, that is, the relationship between trade
protection and economic development, with an improved treatment of tariff endogeneity and
with a consideration of alternative performance criteria. The paper also considers the effects of
asymmetric protection via high tariffs on consumer goods and via low tariffs on producer goods.
Using the sectoral data on Korean manufacturing over its growth period from 1967 to 1993, this
study finds that the impact of trade protection by tariffs tends to appear not in terms of TFP but in
terms of RCA and export shares. Such an impact is found to be greater in consumer goods, which
were the main target for export promotion via higher tariffs. These findings can be regarded as a
contribution because they improved the work of Lee (1996), which tried to capture the impact of
industrial policy only from the TFP side, without considering sectoral differences and using only
lagged values to treat endogeneity.
The results of the current study are consistent with the interpretation that when the
concerned firms or industries are export-oriented and are thus subjected to discipline from the
world market, extra rents associated with higher tariffs tend to be used as investment funds to
expand production. In this sense, this paper verifies the possibly positive role of tariffs under
certain conditions. The condition is that protection through industrial policy should not only be
confined to a certain period but also, and more importantly, be accompanied by disciplinary
devices such as those from the global market. Otherwise, it would be just for protection itself,
but without learning. Industrial policy is not intended merely to protect the economy but to give
it sufficient time to learn.
This study also suggests that tariff protection is better to be sector specific, targeting
only the sectors whose exports are to be promoted, but excluding others. The asymmetric
structure of tariffs in Korea, which provides higher protection to consumer goods being
promoted but charging lower tariffs against imported intermediate goods, is consistent with this
reasoning. In this light, country-level studies using the aggregate, average tariffs are misleading
because they cannot reflect such specificities at the sectoral level. While this study has focused
on proving the impact of trade protection, the question of how long it will take for infant
economies to learn to be competitive remains a topic for future research.
This study presents a dynamic picture of the Korean economy, which is consistent with
Young (1995) work showing that economic growth was led more by input investment and gross
output (export) growth in the earlier period. This study on the earlier period complements Jung
and Lee (2010) research on the later period. Together, the two studies suggest that tariffs and
other protections led to export and output expansion through fixed investment during the early
period, whereas R&D investment and export growth stimulated productivity growth later. For
both periods, the disciplinary impact of export orientation was important because it pushed the
rents associated with tariffs (earlier period) and with an oligopolistic market structure (late period)
to be used for fixed investment (earlier period) and R&D investment (later period), respectively.
Of course, another source of rents during the later period was tax exemption for R&D investment
(Lee 2012). Clearly, the form of government activism in Korea has evolved from trade policy to
technology policy.
18
References
Balassa, Bela and Balassa, Carol, 1984, Industrial Protection in the Developed Countries, The
World Economy, Vol. 7, Issue 2, pp.179-196
Bank of Korea, 1970, 1980, 1990, Input-Output Table of Korea (In Korean)
Bank of Korea, Various Issues, Financial Statement Analysis of Korea (In Korean)
Brambilla, Irene, Galiani, Sebastian, and Porto, Guido, 2010, Argentine Trade Policies in the XX
Century : 60 Years of Solitude, Social Science Research Network Working paper
(http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1680606)
Branson, William H. and Richard C. Marston, 1989, Price and Output Adjustment in Japanese
Manufacturing, NBER Working Paper 2878
Bureau of Labor Statistics Datadase (http://www.bls.gov/ppi/data.htm)
Cheong, KwangSoo, Kineung Choo, and Keun Lee, 2010, Understanding the Behavior of
Business Groups: A Dynamic Model and Empirical Analysis, Journal of Economic Behavior
and Organization, Vol. 76, Issue 2, pp. 141-152.
Choi, Byung-chul, 1991, Study of Economic Cost of Protective Trade and Non-tariff Barrier,
Social Science Study in Ulsan University, Vol. 1, No. 2, pp. 19-38 (In Korean)
Cimoli, Mario., Dosi, Giovanni., and Stiglitz, Joseph E., 2009, Industrial Policy and
Development, Oxford University Press
Dejong, David N. and Ripoll, Marla. 2006, Tariff and Growth : An Empirical Exploration of
Contingent Relationships, The Review of Economics and Statistics, Vol. 88, No. 4, pp. 625-
640
Edwards, Sebastian, 1997, Openness, Productivity and Growth : What Do We Really Know?,
NBER Working paper 5978
Feenstra, Robert C. and Lipsey, Robert E., 2005, World Trade Flows : 1962~2000, NBER
Working Paper Series 11040
Ferreira, Pedro Cavalcanti and Rossi, Jose Luiz. 2003, New Evidence from Brazil on Trade
Liberalization and Productivity Growth, International Economic Review, Vol. 44, Issue 4, pp.
1383-1405
Galiani, Sebastian and Sanguinetti, Pablo, 2000, Wage Inequality and Trade Liberalization :
Evidence from Argentina ( The Impact of Trade Liberalization on
Wage Inequality: Evidence from Argentina, Galiani, Sebastian and Sanguinetti, Pablo, 2003,
Journal of Development Economics, Vol 72, Issue 2, pp. 497-Universidad Torcuato Di
Tella(http://epge.fgv.br/files/1057.pdf)
Haggard, Stephan, 1990, Pathways from the Periphery, Cornell University Press
Harding, Torfinn and Rattsø, Jørn, 2010, Industrial labor productivities and tariffs in South
Africa: Identification based on multilateral liberalization reform, Economics of Transition, Vol.
18, Issue 3, pp. 459-485
Hong, Sung-duk, 1997, Extended Estimation of the Rate of Industry Nominal Protection and the
Effective Rate of Industry Protection(1975~1995), Research Paper 97-06 of Korea
Development Institute (In Korean)
19
Hong, Sung-duk and Kim, Jung-ho, 1996, The Long-Term Change of Total Factor Productivity
of Korean Manufacturing from 1967 to 1993, Research Paper 96-04 of Korea Development
Institute (In Korean)
Hyundai Motors, 1997, Challenge for 30 Years and Vision 21th Century : History of Hyundai
Motors for 30 Years : 1967~1997, Hyundai Motors (In Korean)
Johnson, Chalmers, 1982, MITI and the Japanese Miracle: The Growth of Industrial Policy :
1925-1975, Stanford University Press.
Jung, Moosup, and Lee, Keun, 2010, Sectoral systems of innovation and productivity catch-up:
determinants of the productivity gap between Korean and Japanese firms, Industrial and
Corporate Change, Vol. 19, Issue 4, pp. 10371069.
Kang, Kwang-ha., Lee, Young-hun., and Choi, Sang-oh., 2008, The Korean System of Policy
Decision in High Growth Period, Research Paper of Korea Development Institute (In Korean)
Karacaovali, Baybars. 2006, Productivity matters for trade policy: Theory and evidence, World
Bank Research Working Paper 3925.
Kim, Kwang-suk., 1988, The Economic Effect of Import Liberalization and Policy of Regulating
the Industrial Structure, Research Paper 88-07 of Korea Development Institute (In Korean).
Kim, Yoon-zi, and Lee, Keun, 2008, Sectoral Innovation System and a Technological Catch-up
in Capital Goods Industry: the case of Korea, Global Economic Review, Vol. 37, Issue 2, pp.
135-155.
Korea Customs Service, Various Issues, Statistical Yearbook of Foreign Trade (In Korean)
Korea Customs Service, Various Issues, Tariff Schedules of Korea (In Korean)
-78.
Lawrence, Robert Z. and Weinstein, David E. 1999, Trade and Growth : Import-led or Export-led?
Evidence from Japan and Korea, NBER Working Paper Series 7264.
Lee, Chung H., 1992, The Government, Financial System, and Large Private Enterprises in the
Economic Development of South Korea, World Development, Vol. 20, Issue 2, pp. 187-197.
Lee, Jong-wha. 1996, Government Interventions and Productivity Growth, Journal of Economic
Growth, Vol. 1, No. 3, pp. 391-414.
Lee, Keun, 2012, How Can Korea be a Role Model for Catch-up Development: A Capability-
Based View, in Augustin K. Fosu (editor), Achieving Development Success: Strategies and
Lessons from the Developing World, Oxford: Oxford Univ. Press.
Lee, K. and Y. K. Kim, 2010, IPR and Technological Catch-Up in Korea, edited by Hiroyuki
Odagiri, Akira Goto, Atsushi Sunami, Richard R. Nelson, Intellectual Property Rights,
Development, and Catch Up: An International Comparative Study, Oxford University Press.
Lee, K. and C. Lim, 2001, Technological Regimes, Catching-up and Leapfrogging: Findings
from the Korean Industries, Research Policy, Vol. 30, Issue 3, pp. 459-483.
Lee, Keun-mi, 2004, The Great Choice which Decide National Destiny, Chosun Iibo (In Korean)
Lee, Kyu-uck., Urata, Shujiro., Choi, Inbom., 1986, Recent Development in Industrial
Organizational Issue in Korea, The World Bank Working Paper 8609
Lee, Si- Focused on the Effect
of Decrease of Import Duties, KDI Policy Forum, No. 184, Korea Development Institute (In
Korean)
20
Lee, Tae-yul, 2000, Trade Protection and Change of Productivity in Korean Manufacturing,
International Economic Journal, Vol. 6, No. 2, pp.61-80 (In Korean)
Lehmann,
the Late Nineteenth Century, Review of Economics and Statistics, Vol. 93, No. 2, pp. 606-616.
Ministry of Science and Technology of Korea, Various Issues, Annual Science and Technology of
Korea (In Korean)
Muendler, M, 2001, Trade, Technology and Productivity : A Study of Brazilian Manufacturers :
1986~1998, Mimeo, University of California, Berkeley
Mwega, Francis M. 1994, Trade and Macroeconomic Policies and the Industrialization
Experience in Kenya in the 1970s and 1980s, in G. K. Helleiner(Ed.), Trade Policy and
Industrialization in Turbulent Times, London and New York
Nam, Jong-hyun and Lee, Jae-ho. 2005, Trade and Productivity Growth : Korean Manufacturing
Case, International Economic Journal, Vol. 11, No. 1, pp.49-72 (In Korean)
NBER Database(http://cid.econ.ucdavis.edu/data/undata/undata.html,
http://cid.econ.ucdavis.edu/usixd/usixd4sic.html).
Vol. 110, Issue 463, pp. 456-483.
Pavitt, K. , 1984, Sectoral Patterns of Technical Change: Towards a Taxonomy and a Theory,
Research Policy, Vol. 13, No. 6, pp. 343-373.
Guide to the Cross-National Evidence, In NBER Macroeconomics Annual 2000. Ed. Ben
Bernanke and Kenneth S. Rogoff. Cambridge : MIT Press
Sachs, Jeffrey and Warner, Andrew. 1995, Economic Reform and the Process of Global
Intergration, Brookings Papers on Economic Activity, Vol. 1995, No. 1, pp. 1-118.
Vamvakidis, Athanasios. 2002, How Robust is the Growth-Openness Connection? Historical
Evidence, Journal of Economic Growth, Vol. 7, No. 1, pp. 57-80.
Wacziarg, Romain and Welch, Karen Horn. 2008, Trade Liberalization and Growth : New
Evidence, The World Bank Economic Review, Vol. 22, No. 2, pp. 187-231.
Wogart, Jan Peter and Marques, Jose Silverio, 1984, Trade Liberalization, Tariff Redundancy and
Inflation : A Methodological Exploration Applied to Argentina, Review of World Economics,
Vol. 120, No. 1, pp. 18-39.
Yang, Yung Y. and Hwang, Min, 1994, Price Behavior in Korean Manufacturing, The Review of
Economics and Statistics, Vol. 76, Issue 3, pp. 461-470.
Young, Alwyn. 1995, The Tyranny of Numbers : Confronting the Statistical Realities of the East
Asian Growth Experience, The Quarterly Journal of Economics, Vol. 110, No. 3, pp. 641-680
21
22
Figure 2.1(a). Weighted average nominal tariff in consumer goods and producer goods
Source: alculations based on data from the tariff schedules of Korea
0.00
10.00
20.00
30.00
40.00
50.00
60.00
70.00
80.00
90.00
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
Weighted Average Tariff of Producer goods industry(%)
Weighted Average Tariff of Consumer goods industry(%)
Weighted Average Tariff of Total industry(%)
23
Figure 2.1(b). Tariff of selected items in consumer goods industry and producer goods industry
from 1964 to 1993 (in %)
Source: Tariff schedules of Korea
0
20
40
60
80
100
120
1964
1966
1968
1970
1972
1973
1974
1977
1978
1979
1980
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
Lathe
Bearing
Transmission shaft,
Crank, Gear
Telephone
Television
Air conditioner
24
Figure 2.2. Export orientation in consumer goods and producer goods, 1967 to 1993
Note: Export orientation is defined as real export value divided by real gross output value.
Source: alculations based on export data from the Statistical Yearbook of Foreign Trade; gross
output data from Hong and Kim (1996).
0
5
10
15
20
25
30
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
Producer goods industry(%)
Consumer goods industry(%)
25
Figure 3.1. Ratio of Korean manufacturing exports to Kenya and Argentina to Korean
manufacturing exports to the world (in %)
Source: alculations based on data from NBER compiled by Feenstra and Lipsey
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
26
Table 1. Correlations among export shares of 36 industries in pairs of countries,
1970, 1980, and 1990
Correlations between export shares of 36 industries, 1970
Korea
Argentina
Kenya
Taiwan
Japan
Korea
1
Argentina
0.1612
1
Kenya
0.1777
0.9264
1
Taiwan
0.7387
0.6251
0.6063
1
Japan
0.177
0.0652
0.0266
0.3507
1
Correlations, 1980
Korea
Argentina
Kenya
Taiwan
Japan
Korea
1
Argentina
0.1591
1
Kenya
0.0116
0.6076
1
Taiwan
0.7646
0.3327
0.1374
1
Japan
0.2348
-0.0365
-0.1286
0.1105
1
Correlations, 1990
Korea
Argentina
Kenya
Taiwan
Japan
Korea
1
Argentina
0.0233
1
Kenya
0.0374
0.9736
1
Taiwan
0.6911
0.0201
0.0531
1
Japan
0.4795
-0.0762
-0.0943
0.4799
1
Source: alculations based on data from NBER compiled by Feenstra and Lipsey
27
Table 2. Three main exports industries of Korea, Argentina, and Kenya in 1970, 1980, 1990
1970
1980
1990
Korea
Miscellaneous
Manufacturing
(including
wig)
15.96%
Wearing
Apparel
13.43%
Sound, Image
and
Communication
equipment
19.04%
Fiber Yarn,
Textile
Fabrics
14.4%
Fiber Yarn,
Textile
Fabrics
11.09%
Leather and
Footwear
9.82%
(Footwear
9.06%)
Knitted Goods
13.95%
Iron and
Steel
10.79%
Wearing
Apparel
9.23%
Argentina
Food
79.33%
Food
62.91%
Food
47.18%
Leather and
Footwear
7.02%
(Leather
7%)
Leather and
Footwear
7.53%
(Leather
7.38%)
Iron and Steel
8.07%
Industrial
Chemicals
2.79%
Industrial
Chemicals
4.53%
Petroleum
Refineries
7.16%
Kenya
Food
44.62%
Petroleum
Refineries
42.64%
Food
70.62%
Petroleum
Refineries
16.12%
Food
30.59%
Leather and
Footwear
11.4%
(Leather
11.2%)
Industrial
Chemicals
7.28%
Industrial
Chemicals
3.93%
Industrial
Chemicals
5.41%
Source: alculations based on data from NBER
28
Table 3. Variables in regression models and their summary statistics
A: List of the Variables in Regressions
Dependent variables
in regression models
Control variables
(all one-period lagged values except log of initial value added per
worker and log of initial capital stock per worker)
Industry-level
TFP growth rate
Debt per asset (Finance)
R&D expense per sales (Research)
Log of initial value added per worker
Log of initial capital stock per worker
RCA index
Debt per asset
R&D expense per sales
US real import growth rate
Domestic inflation
TFP growth rate
Share of each sector
in total manufacturing export
Debt per asset
R&D expense per sales
US real import growth rate
Domestic inflation
B: Descriptive Statistics
Variable
Observations
(in yearly
regressions)
Mean
Standard
deviation
Maximum
Minimum
Tariffs
1008
36.39%
28.74%
196.25%
0%
Industry
TFP growth rate
936
2.06%
3.57%
48.83%
-36.01%
RCA index
1008
1.23
1.85
15.79
0
Share in
manufacturing exports
1008
2.78%
4.03%
20.88%
0%
Debt per asset
875
28.37%
9.79%
82.4%
1.1%
R&D expense per sales
769
1.25%
1.07%
5.54%
0.02%
US real import
Growth rate
972
9.72%
20.51%
213.04%
-80.21%
Domestic inflation
971
7.64%
12.49%
170.03%
-21.11%
Log of initial value
added per worker
324
(Period data)
3.73
1.3
7.68
0.4
Log of initial capital
stock per worker
324
(Period data)
2.87
1.01
5.88
0.49
29
Table 4A. Regression results using period data
(Instruments: one-period lagged tariff, following Lee (1996))
Dependent variable
TFP Growth rate
RCA
Export share
Estimation method
FE2SLS
FE2SLS
RE2SLS
Tariff
0.014
0.036***
0.062***
(.012)
(.005)
(.012)
One-period lagged Finance
0.011
0.000
-0.003
(.018)
(.007)
(.019)
One-period lagged Research
0.062
0.203**
0.506**
(.211)
(.089)
(.223)
Log Initial Value Added per
Worker
-3.720***
(.452)
Log Initial Capital Stock per
Worker
1.433***
(.483)
One-period lagged TFP
growth
0.049**
(.024)
One-period lagged US import
growth
-0.008
-0.011
(.006)
(.016)
One-period lagged Inflation
-0.001
0.008
(.01)
(.025)
Hausman P value
0.0000
0.0962
0.9999
R Square
0.2014
0.1478
0.0542
F statistic or Chi-Square
Statistic in first-stage
regression
707.4
780.97
3401.32
Observations per industry
7.4
7.4
7.4
All observations
260
259
259
Note: *** p<0.01, ** p<0.05, * p<0.1
Period dummies and constants have been omitted.
Standard errors of coefficient estimates appear in parentheses.
30
Table 4B. Regression results using period data (Instruments: Tariffs in Argentina and Kenya)
Regression number
1-1
1-2
1-3
1-4
1-5
1-6
Dependent variable
TFP Growth rate
RCA
Export share
Estimation method
FE2SLS
FE2SLS
RE2SLS
RE2SLS
RE2SLS
RE2SLS
Tariff
0.065*
0.081**
0.043***
0.039***
0.075***
0.067**
(.035)
(.036)
(.01)
(.01)
(.028)
(.028)
One-period lagged Finance
0.008
0.010
-0.002
-0.002
-0.020
-0.022
(.021)
(.023)
(.007)
(.007)
(.019)
(.02)
One-period lagged
Research
-0.309
0.155
0.413
(.319)
(.105)
(.283)
Log Initial Value Added per
Worker
-4.174***
-4.651***
(.635)
(.666)
Log Initial Capital Stock per
Worker
1.476**
1.824***
(.616)
(.65)
One-period lagged TFP
growth
0.048**
0.057**
(.023)
(.024)
One-period lagged US
import growth
-0.007
-0.006
-0.019
-0.019
(.006)
(.006)
(.016)
(.017)
One-period lagged
Inflation
-0.004
0.001
-0.006
0.005
(.009)
(.01)
(.025)
(.027)
Hausman P value
0.0002
0.0000
0.996
0.9892
0.9991
0.9999
R Square
0.0982
0.1171
0.1572
0.1617
0.0489
0.0483
F statistic or Chi-Square
Statistic in first-stage
regression
16.39
14.69
48.81
48.96
47.58
47.27
Sargan-Hansen statistic p-
value
0.201
0.1104
0.9752
0.9014
0.667
0.5971
Observations per industry
6.3
5.8
6.3
5.8
6.3
5.8
All observations
219
204
219
204
219
204
Note: *** p<0.01, ** p<0.05, * p<0.1
Period dummies and constant have been omitted.
Standard errors of coefficient estimates appear in parentheses.
31
Table 5. Regression results using annual data (Instruments: Tariffs in Argentina and Kenya)
Regression number
2-1
2-2
2-3
2-4
2-5
2-6
Dependent variable
TFP Growth rate
RCA
Export share
Estimation method
RE2SLS
RE2SLS
RE2SLS
RE2SLS
RE2SLS
RE2SLS
Tariff
-0.009
-0.010
0.047***
0.040***
0.094***
0.091***
(.014)
(.014)
(.006)
(.007)
(.017)
(.019)
One-year lagged Finance
0.018
0.004
0.001
-0.001
-0.005
-0.006
(.012)
(.013)
(.003)
(.004)
(.009)
(.011)
One-year lagged Research
0.159
0.073
0.101
(.175)
(.065)
(.185)
One-year lagged TFP growth
0.001
0.016
(.009)
(.012)
One-year lagged US Import
growth
-0.000
0.001
0.001
0.002
(.002)
(.002)
(.004)
(.005)
One-year lagged Inflation
-0.006*
-0.003
-0.006
-0.007
(.003)
(.004)
(.008)
(.012)
Hausman P value
1.0000
0.7373
1.0000
1.0000
1.0000
1.0000
R Square
0.1601
0.1862
0.1707
0.1184
0.057
0.0303
F statistic or Chi-Square Statistic
in first-stage regression
159.5
137.11
158.68
137.59
159.16
136.33
Sargan-Hansen statistic p-value
0.9292
0.5802
0.3731
0.3744
0.1028
0.3087
Observations per industry
19
14.4
19
14.4
19
14.4
All observations
607
462
607
462
607
462
note: *** p<0.01, ** p<0.05, * p<0.1
Year dummies and constant have been omitted.
Standard errors of coefficient estimates appear in parentheses.
32
Table 6. Results of the seemingly unrelated regression
Estimation method
Seemingly Unrelated Regression
Dependent variable
TFP Growth rate
RCA
Export share
Tariff
0.013
0.030***
0.055***
(.01)
(.004)
(.01)
One-period lagged Finance
0.010
0.000
-0.007
(.016)
(.007)
(.017)
One-period lagged Research
0.054
0.222***
0.547***
(.195)
(.08)
(.209)
Log Initial Value Added per
Worker
-3.889***
(.406)
Log Initial Capital Stock per
Worker
1.364***
(.433)
One-period lagged TFP growth
0.040**
(.016)
One-period lagged US import
growth
-0.007
-0.010
(.006)
(.014)
One-period lagged Inflation
-0.002
0.006
(.009)
(.023)
R Square
0.5114
0.839
0.799
BreuschPagan test of
independence p-value
0.0000
All observations
259
note: *** p<0.01, ** p<0.05, * p<0.1
Period dummies, industry dummies, and constant have been omitted.
Standard errors of coefficient estimates appear in parentheses.
33
Table 7-A. Regression results for consumer and producer goods industry using period data
Regression number
3-1
3-2
3-3
3-4
Dependent variable
RCA
Export share
Classification
Consumer
goods industry
Producer goods
industry
Consumer
goods industry
Producer goods
industry
Estimation method
RE2SLS
RE2SLS
RE2SLS
RE2SLS
Tariff
0.059**
-0.001
0.132*
0.077**
(.029)
(.008)
(.075)
(.032)
One-period lagged Finance
-0.014
0.001
-0.049
0.025
(.015)
(.005)
(.042)
(.019)
One-period lagged Research
0.044
0.266***
-0.333
0.101
(.261)
(.101)
(.668)
(.402)
One-period lagged TFP growth
0.048
0.018
(.036)
(.018)
One-period lagged US import
growth
-0.007
-0.006
-0.011
0.018
(.009)
(.005)
(.026)
(.021)
One-period lagged Inflation
0.011
-0.005
0.059
0.023
(.018)
(.007)
(.051)
(.025)
Hausman P value
1.000
1.000
Negative
statistics
0.9997
Hausman P value (sigmamore++)
1.000
1.000
0.9494
1.000
R Square
0.035
0.040
0.026
0.327
F statistic or Chi-Square Statistic
in first-stage regression
8.360
39.890
8.17
33.020
Sargan-Hansen statistic p-value
0.244
0.996
0.5661
0.4357
Observations per industry
5.3
6.4
5.3
6.4
All observations
79
70
79
70
note: *** p<0.01, ** p<0.05, * p<0.1
Period dummies and constant have been omitted. The same instruments are used.
Standard errors of coefficient estimates appear in parentheses
++ The Hausman test, which uses variance estimation of the two-stage random-effect model error term. We use these
statistics because normal Hausman statistics were negative in Regression 3-3. If the variance estimation of error term by two
stage random effect error term is determined, the possibility that the matrix is positive
definite in the Hausman statistics is high.
34
Table 7-B. Regression results for consumer and producer goods industry using annual data
Regression number
3-5
3-6
3-7
3-8
3-9
3-10
3-11
Dependent variable
RCA
Export share
Classification
Consumer goods
industry
Producer goods
industry
Consumer goods
industry
Producer
goods
industry
Estimation method
RE2SLS
FE2SLS
RE2SLS
FE2SLS
RE2SLS
FE2SLS
RE2SLS
Tariff
0.059***
0.057***
0.006
0.004
0.166***
0.167***
0.056**
(.015)
(.015)
(.006)
(.006)
(.047)
(.048)
(.022)
One-year lagged
Finance
0.001
-0.001
0.001
0.001
-0.011
-0.017
0.012
(.007)
(.007)
(.003)
(.003)
(.022)
(.022)
(.01)
One- year lagged
Research
0.147
0.157
0.129**
0.150**
-0.389
-0.420
0.099
(.142)
(.142)
(.06)
(.06)
(.429)
(.438)
(.212)
One- year lagged
TFP growth
0.010
0.009
0.013
0.014
(.022)
(.022)
(.009)
(.009)
One- year lagged US
import growth
0.001
0.001
-0.000
-0.000
0.008
0.008
-0.000
(.003)
(.003)
(.002)
(.002)
(.008)
(.008)
(.006)
One- year lagged
Inflation
-0.013
-0.013
-0.002
-0.002
-0.025
-0.026
-0.009
(.009)
(.009)
(.002)
(.002)
(.027)
(.026)
(.009)
Hausman P value
Negative statistics
Negative statistics
Negative statistics
0.9999
R Square
0.0118
0.0106
0.0107
0.0005
0.002
0.0017
0.2802
F statistic or Chi-
Square Statistic in
first-stage regression
41.06
20.33
106.58
38.2
36.6
17.24
102.94
Sargan-Hansen
statistic p-value
0.2408
0.2072
0.1999
0.2111
0.7618
0.5636
0.7689
Observations per
industry
14.4
14.4
14.6
14.6
14.4
14.4
14.6
All observations
173
173
161
161
173
173
161
Note: *** p<0.01, ** p<0.05, * p<0.1
Year dummies and constant have been omitted. The same instruments are used.
Standard errors of coefficient estimates appear in parentheses.
35
Table 8. Regression results using interaction term of tariff and export orientation
Dependent variable
RCA
Export share
Data type
Period
Annual
Period
Annual
Estimation method
RE2SLS
RE2SLS
RE2SLS
RE2SLS
Tariff
0.023
0.027**
0.070
0.107***
(.018)
(.01)
(.053)
(.029)
Tariff * Export orientation
0.033*
0.019**
0.004
-0.032
(.018)
(.009)
(.052)
(.026)
Export orientation
1.728***
1.659***
7.045***
7.470***
(.54)
(.265)
(1.543)
(.733)
One-period (or year) lagged Finance
0.001
0.000
-0.005
0.001
(.006)
(.003)
(.018)
(.009)
One-period (or year) lagged
Research
0.012
0.007
0.026
-0.073
(.086)
(.052)
(.241)
(.144)
One-period (or year) lagged TFP
growth
0.041**
0.020**
(.02)
(.01)
One-period (or year) lagged US
import growth
0.002
0.001
0.001
0.002
(.005)
(.001)
(.014)
(.004)
One-period (or year) lagged
Inflation
0.003
-0.002
0.015
-0.004
(.008)
(.003)
(.023)
(.009)
Hausman P value
0.996
1.000
1.000
1.000
R Square
0.567
0.4909
0.3477
0.2866
F statistic or Chi-Square Statistic in
first-stage regression
12.87
48.56
12.17
48.56
Sargan-Hansen statistic p-value
0.6124
0.987
0.7562
0.398
Observations per industry
5.8
14.4
5.8
14.4
All observations
204
462
204
462
Note: *** p<0.01, ** p<0.05, * p<0.1
Period dummies, year dummies, and constant have been omitted. The same instruments are used.
Standard errors of coefficient estimates appear in parentheses.
36
Appendix
A. Korean industry classification and industry classification code
Industry classification criteria were derived from the 5th Korea Standard Industry
Classification Code because Lee (1996) and Hong and Kim (1996) used these criteria. This study
uses the same criteria even though they are dated. If one industry has similar shares between the
consumer goods and producer goods industries, the industry is categorized as an obscure
classification industry.
Knitted apparel and knitted fabrics were categorized under the knitted goods industry.
Thus, this industry is categorized under obscure classification because knitted apparel is also
classified as part of the consumer goods industry and knitted fabric is classified as part of the
producer goods industry. The leather, wood, rubber, ceramics, and glass industries were
classified based on the materials used, so these industries fall under both producer and consumer
goods. The sound, image, and communication industry falls under the obscure classification
industries because it has both producer goods such as semiconductors, and consumer goods, such
as TVs and radios, that have similar shares.
Appendix Table 1. Korean industry classification table and industry classification codes
Industry
numbers
Industry Name
Classification
5th KSIC code
1
Food
Consumer goods industry
311, 312
2
Beverages
Consumer goods industry
313
3
Tobacco Products
Consumer goods industry
314
4
Fiber Yarn, Textile Fabrics
Producer goods industry
3211-3213
5
Knitted Goods
Obscure classification industry
3215
6
Other Fabricated Textiles
Consumer goods industry
3214, 3216-3219
7
Wearing Apparel
Consumer goods industry
322
8
Leather and Footwear
Obscure classification industry
323, 324
9
Wood and Furniture
Obscure classification industry
331, 332
10
Paper and Paper Products
Consumer goods industry
341
11
Printing and Publishing
Consumer goods industry
342
12
Industrial Chemicals
Producer goods industry
351
13
Other Chemicals
Consumer goods industry
352
14
Petroleum Refineries
Producer goods industry
353
15
Other Petroleum and Coal Products
Producer goods industry
354
16
Rubber Products
Obscure classification industry
355
37
17
Plastic Products
Consumer goods industry
356
18
Ceramics
Obscure classification industry
361
19
Glass and Glass Products
Obscure classification industry
362
20
Other Nonmetal Mineral Products
Producer goods industry
369
21
Iron and Steel
Producer goods industry
371
22
Nonferrous Metal
Producer goods industry
372
23
Fabricated Metal
Consumer goods industry
381
24
Engine and Turbine
Producer goods industry
3821, 38411
25
Metalworking and Industrial Machinery
Producer goods industry
3822-3824
26
Office and Other General Machinery
Producer goods industry
3825-3826, 3829
27
Electrical Industrial Apparatus
Producer goods industry
3831
28
Sound, Image, and Communication
equipment
Obscure classification industry
3832, 3834
29
Household Electrical Appliances
Consumer goods industry
3833
30
Other Electrical Equipment
Obscure classification industry
3839
31
Shipbuilding and Repair
Consumer goods industry
3841(Excluding
38411)
32
Railroad Vehicles
Consumer goods industry
3842
33
Motor Vehicles
Consumer goods industry
3843
34
Aircraft and Other Transport Equipment
Consumer goods industry
3844-3845, 3849
35
Measuring, Medical, and Optical
Instruments
Obscure classification industry
385
36
Miscellaneous Manufacturing
Consumer goods industry
390
B. Data Sources
Standard Balassa RCA Index The NBER database prepared by Robert Feenstra and
Robert Lipsey was used to obtain the world export data and the Korean, Argentine, Kenyan,
Japanese, and Taiwanese export data.
(http://cid.econ.ucdavis.edu/data/undata/undata.html).
This database provides world import and export data and country import and export data
by commodity. Commodity classification criteria were derived from the Standard International
Trade Classification Rev. 2 four-digit level. This database also provides world import and export
data until 1983 and the import and export data of 72 major countries since 1984. The percentage
of the import-export of the 72 major countries to the import-export of the world was 98% from
2000 to 2004. A lack of information exists only on the trade between non-major countries, thus
the error is negligible (Feenstra and Lipsey 2005).
38
Ratio of industry export to Korean manufacturing export and real production by
industry Korean industry export data were obtained by sorting export data from the Statistical
Yearbook of Foreign Trade based on the proposed classification. Real production data were
obtained from Hong and Kim (1996). They provide an estimation of real industry production
using the input-output table of Korea and the Report on Korean Mining and Manufacturing
Survey.
Korean tariff rate by industry Data from the tariff schedules of Korea were used. A
simple-averaged value of the legal or temporary tariff rate on imports at the CCCN or HS 24
digit level was used. Industry tariff rate was calculated using the value of the output in the input-
output table of Korea as a weight. Data from the 1970 input-output table of Korea were used to
calculate tariff rates from 1966 to 1976. Data from the 1980 input-output table of Korea were
used to calculate tariff rates from 1977 to 1988. Data from the 1990 input-output table of Korea
were used to calculate tariff rates from 1989 to 1993. When the weighted average tariffs of
consumer goods, producer goods, and total industry were calculated, the real production
estimation proposed by Hong and Kim (1996) was used as a weight. However, the tariff
schedules of Korea in 1967, 1969, 1971, 1975, 1976, and 1981 were unavailable, so a simple-
averaged value of the prior-year tariff rates and proceeding year tariff rates were used to derive
tariff rates in the missing years because the Korean tariff rates changed gradually.
Kenyan tariff rate by industry Data from Mwega (1994), which sort Kenyan
manufacturing industries into 13 classifications and provides the rates of import duty (import
duty divided by import value) in 1970, 1974, 1978, 1980, 1982, 1984, 1986, and 1988, were used.
As data were lacking, those from 1970 had to be used for the second period (1970 to 1972), those
from 1974 for the third period (1973 the 1975), those from 1978 for the fourth period (1976 to
1978), those from 1980 for the fifth period (1979 to 1981), simple average data from 1982 and
1984 for the sixth period (1982 to 1984), data from 1986 for the seventh period (1985 to 1987),
and data from 1988 for the eighth period (1988 to 1990). When using annual data, Kenyan tariff
was assumed to change linearly to fill the gap between selected years. Given that there were only
13 industries, the same tariff rate was assumed for all the Korean sectors within each broad
Kenyan industry. The Kenyan industry classification is presented in Appendix Table 2.
Appendix Table 2. Kenyan industry classification table
Industry Numbers
Industry Classification by Francis M. Mwega (1994)
1 to 3
Food, beverages, tobacco
4 to 7
Textiles and clothing
8
Leather and footwear
9
Wood, cork, and furniture
10 to 11
Paper and printing
39
12 to 13, 16 to 17
Chemicals and rubber
14 to 15
Petroleum, coal products
18 to 20
Building materials and the like
21 to 22
Basic metal industries
23
Metal manufactures
24 to 30, 35
Machinery
31 to 34
Transport equipment
36
Miscellaneous manufacturing
Argentine tariff rate by industry Data from Brambilla et al. (2010), Galiani and
Sanguinetti (2000), and Wogart and Marques (1984) were used. Brambilla et al. (2010) provided
the nominal tariff data of 11 industries with the 1966 to 1970 average, the 1971 to 1976 average,
the 1977 to 1979 average, the 1980 to 1990 average, and the 1991 to 2005 average. Galiani and
Sanguinetti (2000) provided the nominal tariff data of 21 industries in 1990 and 1991. Wogart
and Marques (1984) provided the nominal tariff data of 40 selected products in 1977 and 1979.
As data were incomplete, the period data of Argentine tariff were obtained by the rules described
in Appendix Table 3.
Appendix Table 3. Source of Argentine tariff by industry
Period
Data Source
Description
1967 to 1969
Brambilla et al. (2010)
Average tariff data from 1966 to 1970
1970 to 1972
Simple average of the average tariff data from 1966 to 1970 and
from 1971 to 1976
1973 to 1975
Simple average tariff data from 1971 to 1976
1976 to 1978
Brambilla et al. (2010),
Wogart and Marques
(1984)
Brambilla et al. (2010) data: simple average of the average tariff
data from 1971 to 1976 and from 1977 to 1979
Wogart and Marques (1984): 1977 data
1979 to 1981
Brambilla et al. (2010) data: simple average of the average tariff
data from 1977 to 1979 and from 1980 to 1990
Wogart and Marques (1984): 1979 data
1982 to 1984
Brambilla et al. (2010)
Average tariff data from 1980 to 1990
1985 to 1987
Average tariff data from 1980 to 1990
1988 to 1990
Galiani and Sanguinetti
(2000)
1990 data
1991 to 1993
1991 data
The Argentine industry classifications in these papers were revised because they were
different from the 36 industry classifications used in the current work. The Argentine industry
classification is shown in Appendix Table 4.
Appendix Table 4. Argentine industry classification
40
Industry
numbers
Industry Name
Classification by
Brambilla et al.
(2010)
Classification by Wogart and
Marques (1984)
Classification by Galiani and
Sanguinetti (2000)
1
Food
Processed Food
Food and Beverages
2
Beverages
Processed Food
Food and Beverages
3
Tobacco Products
Processed Food
Tobacco
4
Fiber Yarn, Textile Fabrics
Textiles
Yarn, Cloth
Textile Products
5
Knitted Goods
Textiles
Socks
Textile Products
6
Other Fabricated Textiles
Textiles
Textile Products
7
Wearing Apparel
Textiles
Shirts
Apparel
8
Leather and Footwear
Footwear, Leather
Leather, Footwear
9
Wood and Furniture
Wood
Wood Production (non-furniture)
10
Paper and Paper Products
Pulp, Paper
Paper Production and Paper
Products
11
Printing and Publishing
Printing and Publishing
12
Industrial Chemicals
Chemical
Liquid & Compressed Gas,
Tanning Materials, Basic Chemical
Substances, Fertilizers, Pesticides,
Synthetic Fibers
Chemical Products
13
Other Chemicals
Chemical
Paints and Varnishes
Chemical Products
14
Petroleum Refineries
Chemical
Petroleum Distillery
15
Other Petroleum and Coal
Products
Chemical
Chemical Products
16
Rubber Products
Chemical
Tires
Rubber and Plastic Products
17
Plastic Products
Plastics
Synthetic Resins
Rubber and Plastic Products
18
Ceramics
Mineral
Nonmetal Mineral Products
19
Glass and Glass Products
Mineral
Glass & Glassware
Nonmetal Mineral Products
20
Other Nonmetal Mineral
Products
Mineral
Cement
Nonmetal Mineral Products
21
Iron and Steel
Metals
Iron and Steel
Basic Metals
22
Nonferrous Metal
Metals
Nonferrous Metals
Basic Metals
23
Fabricated Metal
Metals
Metal Containers, Cans, Machine
Tools
Metal Products (Non-machinery
and Equipment)
24
Engine and Turbine
Machinery
Motor and Turbines
Engines and Electric Equipment
25
Metalworking and Industrial
Machinery
Machinery
Agricultural Machinery, Metal &
Woodworking Machinery
Machinery and Equipment
26
Office and Other General
Machinery
Machinery
Ovens and Stoves, Office
Machines, Elevators, Tractors
Computer, Accounting, and Office
Machinery
27
Electrical Industrial
Apparatus
Machinery
Electrical Machinery
Engines and Electric Equipment
28
Sound, Image and
Communication equipment
Machinery
Radio and TV, Communication
Equipment
Engines and Electric Equipment
41
29
Household Electrical
Appliances
Machinery
Refrigerators & Air Conditioning
Engines and Electric Equipment
30
Other Electrical Equipment
Machinery
Batteries, Electrical Bulbs &
Tubes, Electrical Conductors
Machinery and Equipment
31
Shipbuilding and Repairing
Transport
Other Transportation Equipment
32
Railroad Vehicles
Transport
Railway Equipment
Other Transportation Equipment
33
Motor Vehicles
Transport
Auto Engines
Motor Vehicles and Equipment
34
Aircraft and Other Transport
Equipment
Transport
Motorcycles & Bicycles
Other Transportation Equipment
35
Measuring, Medical and
Optical Instruments
Machinery
Scientific Equipment
Medical, Ophthalmic, Watches,
Clocks, and others
36
Miscellaneous
Manufacturing
Furniture and Manufacturing
Industries
The average tariff data from Brambilla et al. (2010) were used as annual data. Thus,
several years of each period (1967 to 1970, 1971 to 1976, 1977 to 1979, 1980 to 1990, and 1991
to 1993) have the same values in the same industry.
Value added per worker by industry, TFP growth rate by industry, number of
industry workers, and capital stock by industry The data for these variables were obtained
from Hong and Kim (1996). The industry value added was calculated by subtracting the industry
intermediate input from the industry real production. The industry TFP index was calculated by
the growth accounting method using the translog method. Industry TFP indices were provided
from 1967. Thus, the industry TFP growth rate during the first period (1967 to 1969) was
substituted by the industry TFP growth rate from 1968 to 1969.
Ratio of bank debt to total asset Data from the
, so the
o
the railroad vehicles industry as well as the aircraft and other transport equipment industry. The
data provided were from 1969, so the first period values (1967 to 1969) were substituted by the
1969 values. Data on the tobacco industry were not provided.
Ratio of private R&D expense to sales
The same ratio for all the sectors within
each broad industry was assumed because the estimate was only available for broad industry
classifications, which varied between periods. The classification shown in Appendix Table 5 is
the broadest.
Appendix Table 5. Classification in the Science and Technology Annual of Korea
Industry Numbers
Classification in the Science and Technology Annual of Korea
1~3
Food, Beverage, and Tobacco
4~8
Textile and Apparel
42
9~11
Wood, Paper, and Printing
12~17
Chemicals
18~20
Nonmetal Mineral Products
21~22
Primary Metal
23
Fabricated Metal
24~26
Machine
27~30
Electrical and Electronic Equipment
31~34
Vehicles
35
Measuring, Medical, and Optical Instruments
36
Miscellaneous Manufacturing
In addition, the only data available before 1976 were from 1968, 1970, 1971, and 1975.
Hence, the data from 1968 had to be used for the first period (1967 to 1969). The simple average
data from 1970 and 1971 were used for the second period (1970 to 1972). Data from 1975 were
used for the third period (1973 to 1975). Data for 1975 were only available for 18 industries.
Data from 1988 were used for the eighth period (1988 to 1990) of the miscellaneous
manufacturing industry.
Real import growth rate of the US industry and the price index of the US industry –
The NBER database was used to obtain the industry-level data, which show the level of US
imports from the world (http://cid.econ.ucdavis.edu/usixd/usixd4sic.html). This database
provides the US import-export data classified by the 1972 Standard Industrial Classification
four-digit level from 1958 to 1994. The data were sorted, and the nominal import of the US
industries was derived. The price index of the US industry was calculated using the weighted
average of producer price indices from the Bureau of Labor Statistics.
(http://www.bls.gov/ppi/data.htm) The value that the Bureau of Labor announced in 2010
was used as the weight. The real import value was obtained by dividing the nominal import value
by the price index.
Domestic inflation by industries The industry price indices were obtained from the
producer price indices of the Bank of Korea, Kim (1988), and Hong and Kim (1996). The
domestic inflation rates of industries were obtained by calculating the growth rates of industry
price indices.
