Information about Asia and the Pacific Asia y el Pacífico

13 ASEAN in a Regional Perspective

John Hicklin, David Robinson, and Anoop Singh
Published Date:
July 1997
  • ShareShare
Information about Asia and the Pacific Asia y el Pacífico
Show Summary Details
Jeffrey A. Frankel and Shang-Jin Wei 

ASEAN and Its Pattern of Trade

There are two striking pieces of conventional wisdom about the status of regional trading blocs in East Asia. The first is that the only formal regional arrangement in the area, the Association of South East Asian Nations (ASEAN), does not in fact function as an economic bloc. Trade among the members is thought to be very low. The second is that East Asia taken as a whole does, under Japanese direction, function as a trading and investment bloc, and increasingly so over time despite the absence of a formal preferential trading area among these countries. In other words, according to the conventional wisdom, the formal regional trade is not, in reality, a bloc, and the actual bloc is not formally recognized as such. Like much conventional wisdom, this characterization of East Asian trading patterns, although it has some truth in it, is not entirely correct. This paper investigates patterns of trade and direct investment in Southeast Asia, with an eye to these hypotheses regarding blocs.

A Short History

ASEAN was founded in 1967 for political purposes and was declared a preferential trade area in 1977.1 The preferential trade agreement granted 10-15 percent margins of preference on 71 commodities and industrial projects in 1978. This amounted to little at the time because the most important sectors were exempted from the system of preferences that they were supposed to grant each other (Panagariya, 1994). Between 1985 and 1987, the ASEAN leaders agreed to expand the list covered by the preferential trade agreement and to increase the margin of preferences. However, as recently as 1989, only about 3 percent of goods were eligible for regional preferences.

Talks in January 1992 led to the decision to create the ASEA free trade area (AFTA). AFTA sounded more serious than the earlier attempts, calling for the reduction of tariffs and non tariff barriers in phases from 1993 to 2008. At a meeting in 1994, economic ministers moved the date for full implementation forward to 2003. Unlike the earlier agreements, AFTA is to cover nearly all sectors of intra-ASEAN goods trade, including agriculture, although a number of exclusions for nonprocessed agriculture are still under negotiation and the treatment of nontariff barriers is vague.2 Even if fully implemented, the agreement will allow intra-bloc tariffs of up to 5 percent to continue. Thus, the “free trade area” is really a preferential trading arrangement. Some preliminary work has also been done on cooperation in services and intellectual property, but services are far from liberalized (ITR 5/3/95).

Conventional Wisdom on Intra-ASEAN Trade

Conventional wisdom holds that trade within ASEAN is relatively low, despite the formal measures taken. But low compared with what? Trade within ASEAN is considered to be low compared with other regions of the world. In one variant of the conventional wisdom, the low share of intra-ASEAN trade is just what one would expect, given the similarity in the factor endowments of the ASEAN countries. De Rosa (1995, p. 28) offers a typical statement of this view:3

By comparison [with their trade with industrialized partners], intra-ASEAN trade accounted for only about 16 percent of ASEAN exports and imports combined. If intra-ASEAN trade involving Singapore is excluded, the extent of intra-ASEAN trade falls to a level lower than that for ASEAN trade with the East Asian [newly industrializing economies] and the developing countries outside Asia…. ASEAN economies are essentially competitive rather than complementary…. This means that ASEAN comparative advantage and greatest gains from trade lie mainly in trade with the major industrial countries … whose relative endowments of physical and human capital, basic labor, and natural resources are different from those of the ASEAN countries.

The assertion that intra-ASEAN trade is unusually low is not clearly accurate, depending on what metric is considered appropriate. Indeed, by some measures, trade among these countries is high, as we shall see. Two firms in Southeast Asian countries are far more likely to trade with each other than two firms at random locations around the globe. Much of this regional concentration can be explained by natural determinants, such as geographic proximity, especially if one allows a special role for Singapore as an entrepôt, and if one allows for the extra trade orientation of ASEAN countries and Asian countries in general. Nevertheless, the conclusion is that intra-ASEAN trade is not lower than would be expected.

Problem with Using Intraregional Trade Shares

The judgment that intra-ASEAN trade is low is based largely on simple trade share statistics. The denominator of the ratio is total trade undertaken by ASEAN countries, and the numerator is the trade that they undertake with each other. As Table 1 shows (Ratio 1), although the regional trade share increased gradually over time from 14 percent in 1980, it was still only about 21 percent in 1994. (Brunei Darussalam is included. If Indochina is also included in the grouping, then the recent increase in trade within Southeast Asia is just slightly greater.) It is indeed true that most of ASEAN trade takes place with countries outside the group.4

Table 1.Intraregional Trade Shares and Intensity Ratios
Ratio 21.443.175.415.56
ASEAN (incl. Brunei Darussalam)
Ratio 21.891.921.831.61
ASEAN and Indochina
Ratio 21.961.891.67
Australia and New Zealand
Ratio 22.102.422.943.33
Ratio 10.510.550.560.53
Ratio 20.720.710.780.86
European Free Trade Association
Ratio 20.980.900.991.06
Ratio 22.914.346.026.13
Ratio 10.370.400.410.44
Ratio 20.951.001.041.04

Note: Let xij denote exports from country i to country j; define z to be the sum of xij, where both i and j are in the trading bloc; define zx to be the sum of xij, where i is in the bloc, zm to be the sum of xij, where j is in the bloc, and w to be the sum of xij for all i, j in the sample. Then

Note: Let xij denote exports from country i to country j; define z to be the sum of xij, where both i and j are in the trading bloc; define zx to be the sum of xij, where i is in the bloc, zm to be the sum of xij, where j is in the bloc, and w to be the sum of xij for all i, j in the sample. Then

By comparison, the European Union (EU) and the North American Free Trade Agreement (NAFTA) have much higher shares of trade (53 percent and 44 percent, respectively) within the group. However, some free trade areas, such as the Mercado Común del Sur (MERCOSUR) and the Andean Pact, have even lower ratios than ASEAN (19 percent and 10 percent, respectively). One might be tempted to infer from those statistics that free trade areas are likely to be successful only among industrial countries.

The conclusion that free trade areas among developing countries do not fare well would be consistent with the experience of the 1960s, when many regional trading arrangements among poor countries were proclaimed with great fanfare and then came to naught. The history is a story of failure to translate visions into specific plans, of delays in implementation, of rampant sectoral exclusions or escape clauses, and of poor enforcement of nominal agreements.5 But this judgment is not consistent with the experience of the 1990s, when such previously ineffective clubs as the Andean Pact have become much more serious, and new free trade areas, such as MERCOSUR, have been established.

A drawback to the trade share as a measure of intraregional trade concentration can be seen in any table that compares different groupings. The larger the grouping, the higher the intraregional trade ratio. The share is very high for large groupings like the Asia Pacific Economic Cooperation (APEC) group or Western Europe. Is this because APEC has been successful at promoting trade among its members and ASEAN has not? Not necessarily. Rather, it reflects primarily that APEC is a large group of countries, both in the sense of the number of members that belong and in the sense that many of them are quite large trading countries, while ASEAN represents a relatively small group of small countries. It is a necessary property of the intraregional share measure that the larger the set of countries around which one throws the lasso, the higher the apparent concentration of trade within. At the limit, if one throws the lasso around all countries of planet Earth, one would find a ratio of 100 percent. Only after one takes into account APEC’s share of world trade (41 percent) can one consider its intraregional trade to be noteworthy.

The fallacy arises even more often in comparisons across time. The intraregional share of East Asian trade has been rising steadily, from 36 percent in 1980 to 49 percent in 1994 (Frankel, 1996). This increase in trade has often been cited as evidence that Japan is building a trade bloc in East Asia, even without explicit policy steps toward a preferential trading area.6 We shall see that these inferences regarding the speed with which trade is becoming intraregionally concentrated are incorrect.

It is worth noting that levels or shares of intraregional trade are indeed useful for some purposes. Let us say we are interested not in the effects of preferential tariffs and other policy determinants on bilateral trade patterns, but rather in the effects of bilateral trade. Such effects would be of interest, for example, to businesspeople, macroeconomists, and political scientists. Then it would be perfectly appropriate to look at intraregional trade shares.

A businessperson, particularly one in a trade-related industry like shipping, might want to know in what parts of the world bilateral trade is increasing the most rapidly, so that he or she can plan where to invest. For a particular small Southeast Asian economy that depends importantly on the magnitude of its trade links with Japan and the United States, a macroeconomist may wish to know its sensitivity to sudden cyclical fluctuations emanating from these two countries. The old principle that East Asian economies are highly dependent on North American growth is rapidly becoming less true as trade within Asia becomes more important. For that matter, Japan is itself declining in importance as a fourth “growth pole” on the East Asian mainland is gaining.

Intraregional trade may have important political implications as well. Hirschman (1980), in a classic study, pointed out the international influence that arises from trade. In time of political or military conflict, a country may be reluctant to side against a large trade partner. Hirschman made it clear that the trade need not be the outcome of a preferential trading arrangement. “For the political or power implications of trade to exist and to make themselves felt, it is not essential that the state should exercise positive action, i.e., organize and direct trade centrally; the negative right of veto on trade with which every sovereign state is invested is quite sufficient” (pp. 16–17). Thus, to observe that intraregional trade shares for groupings that include such large countries as the United States and Japan will necessarily be large is to observe accurately that the United States and Japan are powerful players. To repeat the central objection to the trade shares, however, they cannot be used to assess whether trade is in any meaningful sense necessarily concentrated or biased toward the United States or Japan, or toward all the members of APEC, beyond what would be expected from the size of these countries.

Intra-ASEAN Trade Compared with That of Other Regional Groupings

To obtain a usable measure of regional concentration, we need to adjust the intraregional trade shares by a measure of each group’s importance in world trade. We want to know if a typical member of ASEAN trades more with other members of the group than does a typical country located anywhere in the world. The simplest way to determine this is to divide each intraregional trade share by that region’s share of world trade, as in the measure reported as Ratio 2 in Table 1. We shall call such numbers “intensity” or “concentration” ratios. The intuitive idea is that if bilateral trade takes place in geographic patterns that are simply proportionate to the distribution of countries’ total trade, then the concentration ratio should be close to 1. If trade is concentrated within a given grouping of countries, that grouping should show a ratio in excess of 1.

As Table 1 shows, the intensity ratio is above 1 for ASEAN, as for most groupings. The conclusion is that, on the one hand, trade is geographically concentrated, although less so for the EU and NAFTA. Suddenly, the trading arrangements among developing countries look more effective than those to which industrial countries belong. On the other hand, there is no upward trend in intra-ASEAN or intra-Asian trade intensity. Rather, the large increase in trade among ASEAN countries, or among Asian countries more generally, is fully in line with the large increase in trade undertaken by these countries with the entire world.7 Thus, the standard intraregional trade shares are misleading with respect to both the level of regional trade concentration and its rate of change.

Various economists have observed the recent regional concentration of trade and have drawn varying inferences from it. The key difference in interpretation centers on whether the evident regional concentration in trade should be attributed to the natural factor of geographical proximity or to the artificial factor of preferential trade policy. Two eminent economists, while admitting that existing trade policies must play a role in such statistics, have asserted that the dominant explanation for the high concentration ratios must be geographical proximity (Krugman, 1991; and Summers, 1991).

At the opposite extreme, other eminent economists have dismissed the role of geographical proximity and asserted that the explanation for the observed concentration must therefore be existing discriminatory trading arrangements (Panagariya, 1994). This issue is important, because each of the two camps engages in a line of reasoning that runs from the positive statements, regarding the effect of policy on trade, to normative statements regarding the desirability of regional trading arrangements.8 Fortunately, it is possible to quantify the extent to which intraregional concentration is attributable to proximity, as in the Krugman-Summers view, versus existing preferences, as in Panagariya.

In the following sections, we shall adjust the bilateral trade figures for the effects of geographical proximity and other nonpolicy variables that naturally link countries. In this way, we hope to isolate the effects on trade of preferential trading policies.

Gravity Estimates for Trade Among Five ASEAN Countries

The key to detecting and quantifying a possible intraregional trade bias is to establish a “norm” of bilateral trade volume based on economic, geographic, and cultural factors. A useful framework for this purpose is the gravity model.9 Once the norm has been established by the gravity model, a dummy variable can be added to represent when both countries in a given pair belong to the same regional grouping. The coefficient on this “bloc variable” tells us the extent to which trade within the group has been promoted, whether by explicit preferential trading policies or by less formal sociopolitical forces. One can check, in particular, how the level of trade and time trend in ASEAN compares with that in other groupings.

Gravity Model

The dependent variable in our gravity estimation reported in this section, denoted Tij, is the bilateral volume of total trade (exports plus imports) between countries i and j. The two most important factors in explaining bilateral trade flows are the geographical distance between the two countries and their economic size. These factors are the essence of the gravity model and are the source of the name, by analogy to the formula for gravitational attraction between two heavenly bodies.

A large part of the apparent bias toward intraregional trade is due to simple geographical proximity. Most obviously, proximity reduces shipping costs; it also reduces other costs associated with time lags (e.g., interest charges, spoilage, and obsolescence) and cultural barriers (e.g., ignorance of foreign customs and tastes). Indeed, as already noted, Krugman (1991) and Summers (1991) assert that most of the observed tendency for countries to trade disproportionately with their intraregional neighbors is due to proximity. Krugman uses this proposition to argue that the three trading blocs are welfare-improving “natural” groupings (as distinct from “unnatural” trading arrangements between distant trading partners, such as Malaysia and the United Kingdom under the old Commonwealth preferences). The argument is that natural intracontinental trade blocs are more likely to be trade-creating than trade-diverting; because transportation and other distance-related costs inhibit trade between continents anyway, there is less trade to be diverted.

Theoretical models and empirical studies alike surprisingly often neglect to take into account distance and transportation costs. Our measure is the log of distance between the two major cities (usually the capital) of the respective countries.10 We also add a dummy variable, ADJACENCY, to indicate when two countries share a common land border.

Entering GDP in product form is empirically well established in bilateral trade regressions and can easily be justified by the modern theory of trade under imperfect competition. Intuitively, one will choose to trade more with a larger country than a smaller country because it offers more variety, and consumers like variety. There are also reasons to believe that GDP per capita (GDPpc) has a positive effect for a given size: as countries become more developed, they tend to specialize more and trade more. An important part of this process is that higher-income countries tend to have lower trade barriers.

A common language can facilitate trade partly because it directly reduces transaction (translation) costs and partly because it enhances exporters’ and importers’ understanding of each other’s culture and legal system, which indirectly promotes trade. To capture this effect, we also include a dummy variable, LANGUAGE, that takes the value of 1 if the country pair in question shares a common language or has a previous colonial connection. We consider nine languages: English, French, German, Spanish, Portuguese, Dutch, Arabic, Chinese, and Japanese.

A representative specification is

The last three explanatory factors are dummy variables. ASEANij is an example of the sort of dummy variable we use when testing the effects of membership in a common regional grouping. It is defined as 1 for a given pair when both countries are members of ASEAN, and 0 otherwise. We use the technique of ordinary least squares (OLS) regression, which is capable of testing the effect of each independent variable while holding constant the effects of the others.

Our base data set covers 63 countries (or 1,953 country pairs) for 1980, 1990, 1992, and 1994. In most cases, results are reported separately year by year because there are enough data to do so, and one wants to see how the coefficients change over time. The sources are the United Nations trade matrix for 1980 and the International Monetary Fund’s Direction of Trade Statistics for 1990, 1992, and 1994.

Do ASEAN and East Asia Constitute Regional Trading Blocs?

Table 2 shows ASEAN, alone in the world among the six contemporary free trade areas tested, as having a statistically significant, apparent intraregional bias in every year tested, 1965 through 1992.11 The coefficient estimate in 1992 is 1.8, which also happens to be close to the mean, median, and mode of the yearly estimates. The implication is that two ASEAN countries trade six times more than two otherwise similar countries. (Because trade is expressed in logs, one must take the exponential of the coefficient: [exp(1.8) = 6].) It is in this sense that intra-ASEAN trade can be said to be high, rather than low.

Table 2.Gravity Model of Trade

(Dependent variable: ln (Tij))

ln (GDPi*GDPj)0.637**0.685**0.646**0.702**0.744**0.786**0.775**0.804**0.797**0.834**0.796**0.832**0.930**0.963**
ln (GDPpci*GDPpcj)0.235**0.284**0.377**0.403**0.255**0.294**0.238**0.323**0.247**0.264**0.080**0.128**0.128**0.153**
ln (DISTANCEij)-0.483**-0.447**-0.562**-0.594**0.698**-0.683**0.588*-0.555*-0.715-0.707**-0.572**-0.656**-0.770**-0.733**
EU bloc0.218#0.1430.061-0.0780.140-0.229-0.021-0.0760.277*0.1340.267**0.158-0.083-0.135
NAFTA bloc0.0200.178-0.227-0.050-0.313-0.0280.0980.379-0.2640.1850.1520.367-0.226-0.201
MERCOSUR bloc-0.343-0.0510.3110.4510.2770.2470.561*0.7460.808*0.686#1.918**1.324**0.690**0.934*
ANDEAN bloc-1.310**-1.198**-0.307-0.2830.3110.3510.0820.103-0.1030.046-0.1040.2040.965**1.187**
ASEAN bloc1.621**1.274**2.045**1.570**1.824**1.512**2.272**1.925**1.704**1.487**1.757**1.196**1.766**1.126*
ANZCERTA bloc1.263**1.448**1.399**1.380**1.732**1.768**1.716**1.688
EU openness-1.120-0.159-0.096s-0.0360.077-0.186**-0.132*
NAFIA openness-0.600*-0.664**-0.630**-0.491**-0.616**-0.434**-0.751**
MERCOSUR openness-0.289**-0.091-0.136-0.1320.252**0.818**-0.295**
ANDEAN openness-0.1470.030-0.0320.058-0.087-0.1060.190#
ASEAN openness0.451**0.620**0.392**0.469**0.312**0.640**0.610**
ANZCERTA openness-0.331-0.0720.272-0.154
Number of observations1,1941,1941,2741,2741,4531,4531,7081,7081,6471,6471,5731,5731,5461,546
Adjusted R20.6600.6740.6840.7010.7030.7130.6940.7030.7210.7300.7500.7760.7980.816
Standard error of regression1.0961.0721.1261.0941.2001.1801.2421.2231.2041.1851.1151.0571.1151.085
Note: All variables except dummy variables are in logs; *, **, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.
Note: All variables except dummy variables are in logs; *, **, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

We know that Singapore plays an entrepôt role: its imports and exports are more than 100 percent of GDP. The island nation accounts for almost half of intra-ASEAN trade. It is possible that the apparent intra-ASEAN bias is partly or wholly a reflection of the extreme openness of Singapore. To examine this, we have elsewhere tried adding a dummy to the regression that represents any bilateral trade involving Singapore. The Singapore dummy does indeed have a positive and very significant coefficient, 1.51. The coefficient on the dummy ASEAN is reduced to 1.40, but remains quantitatively large and statistically significant. This suggests that Singapore’s extreme openness does not explain all of the apparent inward bias among the ASEAN countries.

The effect in each year is reduced a little more if one allows for the fact that the entire group of ASEAN countries, and not just Singapore, is more open than are typical countries at the same stage of development. This is accomplished by adding a dummy variable representing observations where either of the two partners is a member of ASEAN (or likewise with any other grouping). A positive coefficient indicates openness.

ASEAN is indeed open. Part of what appeared to be a proclivity to trade with other ASEAN members was really a proclivity to trade with everyone. But, again, some intra-ASEAN trade remains unexplained. The bloc coefficient is still in every year highly significant statistically, equaling 1.1 in 1992. These findings—that ASEAN countries are significantly more open than predicted by the gravity determinants, but that allowing for this openness reduces the strong estimated bloc effect by only a little—are confirmed in other tests as well. When the data from 1970 to 1992 are pooled, the ASEAN coefficient is 2.0, or 1.3 when allowing for ASEAN openness (Frankel and Wei, forthcoming, Table 1).

Allowing for a trend in the coefficient shows no evidence of upward or downward changes over time. If we wish to test the effect of the establishment of a regional trading arrangement on the change in trade, there is no one clear date on which to focus. As already noted, ASEAN negotiated a preferential trading arrangement with in its membership in 1977, but serious progress in removing barriers did not get under way until 1987. It was not until January 1992 that the members proclaimed plans for implementing an ASEAN free trade area by reducing tariffs and nontariff barriers in phases. Thus, we choose 1992 as the key date. A test of the change in intra-ASEAN trade between 1990 and 1992 shows an insignificant point estimate of 0.2. Thus, one cannot attribute the regional concentration, which shows up in the trade numbers of this period, to the agreements proclaimed in 1987 or 1992.

The estimated effect of ASEAN on trade among its members can, however, change radically depending on what other bloc effects are being tested at the same time. When we test for an East Asian bloc effect simultaneously with an ASEAN effect, the latter disappears completely. If one is interested solely in formal regional arrangements, then one can accept at face value the first results reported here, that is, the strong bloc effects for ASEAN. If one considers the larger, less formal blocs to be on an equal footing a priori, then one will want to accept the verdict of the data that ASEAN has no independent effect: Southeast Asian countries trade a lot with each other simply as an example of the phenomenon that Asian countries trade a lot with each other, not out of any special ASEAN effect.

Wang and Winters (1991) in gravity tests found the ASEAN dummy to reflect one of the most significant trading areas in the world. They did not include a broader dummy variable for intra-Asian trade (or for the extra openness of East Asian countries in general, or of Singapore in particular). Thus, their results are consistent with ours.

Continuing the process that began with ASEAN, we consider a sequence of nested candidates for trading blocs in the Pacific.12 The significance of a given bloc effect turns out to depend on what other blocs are tested at the same time. One way to draw the boundaries is to include all the countries with eastern coasts on the Pacific, which includes Australia and New Zealand along with East Asia. We call this grouping “Asia-Pacific.” Its coefficient and significance level are both higher than the East Asia dummy. When we broaden the bloc search and test for an effect of APEC, which includes the United States and Canada, it is highly significant. The significance of the Asia-Pacific dummy completely disappears. The East Asia dummy remains significant, although at a lower level than the initial results that did not consider any wider Pacific groupings.

Let us pause to summarize our results so far. When one takes into account the size of the economies, intraregional trade is high, as much within ASEAN as within East Asia more broadly. The same is true when one takes into account the proximity of the countries. These bloc effects could be due either to formal preferential trading arrangements, that is, the effects of ASEAN, or to informal factors, such as links among Chinese businesspeople. The rate of increase of trade within ASEAN or within East Asia, however, can be entirely explained by the rapid growth of the countries. There is nothing left over to attribute to a bloc that is intensifying over time. Since ASEAN preferences were not operational at the beginning of the sample period, the evidence tends to point more to the informal social forces than to the formal policy measures.

Is ASEAN Open to Trade or Trade-Diverting?

The coefficient on the openness dummy tells how much members of a group trade with other countries in general (regardless of whether they are in the same group). Thus, it reflects the extent to which tariff and non tariff barriers have been removed, as well perhaps as nonpolicy influences on the propensity to trade (excluding, of course, income and the geographical variables for which we control), compared with other countries. If this variable is negative, the members of the group in question trade less with the rest of the world than would be predicted, perhaps because the preferential trade area among them (if there is one) has diverted trade.

After adjustment for income levels, East Asian groupings show up as the most open to trade with the rest of the world. ASEAN shows little or no evidence of trade diversion. To the contrary, given their stage of development, the ASEAN countries consistently show a higher level of openness than other countries in the sample (Wei and Frankel, forthcoming). If 1990 is taken as the key date for ASEAN, the estimated effect on the change in trade with nonmembers is also positive; the same openness is revealed for the broader grouping of East Asia. As already noted, allowing for openness changes the estimates of the bloc effects quantitatively, but not qualitatively.

Gravity Estimates Extended to Focus on Trade with New Southeast Asian Partners

To focus on Southeast Asia, we now add several countries to the base data set used in earlier studies: Brunei Darussalam (a member of ASEAN since 1984), Cambodia, the Lao People’s Democratic Republic, Myanmar, and Vietnam.13 (Trade data for Brunei Darussalam, the Indochinese countries, and Myanmar come from the IMF country desks.) Unfortunately, data for the three Indochinese countries and Myanmar are not available before 1990.

Including Brunei Darussalam as the Sixth Member of ASEAN

Table 3 is the first to include Brunei Darussalam in the definition of ASEAN, for the period 1980-94. The other major respect in which the results from here on differ from those reported in Table 2 is that the income of the importing country, j, is allowed to have a different coefficient from the income of the exporting country, i. This, in turn, requires that the dependent variable be defined as the log of exports from i to j. The estimates for the gravity variables are generally similar to what they were before, although the variables ADJACENCY and LANGUAGE have lost their ability to explain bilateral trade in the 1980 regression. As before, the ASEAN bloc is highly significant in 1980, with a small downward trend subsequently. The magnitude of the bloc effect is in each year somewhat smaller than it was without Brunei Darussalam. (Among the other regional groupings, the EU and NAFTA have gained significance, while MERCOSUR has lost some.)

Table 3.Gravity Model, with Brunei Darussalam, Indochina, and Myanmar in the Data Set

(Dependent variable: ln (exportij))

ln (GDPi)0.642**0.512**0.528**0.692**
ln (GDPj)0.643**0.551**0.497**0.615**
ln (DISTANCEij)-1.073**-1.045**-0.949**-0.724**
ASEAN bloc1.017**0.986**0.4300.777*
EU bloc0.252*0.704**0.537**0.468**
EFTA bloc0.2220.0560.306#-0.293
ANZCERTA bloc1.969**2.582**2.210**1.809**
NAFTA bloc4.290**1.811**1.481**1.193**
ANDEAN bloc-0.328-0.673*-0.718**1.760*
MERCOSUR bloc0.3022.012*1.429-0.231
Number of observations2,8542,9572,9852,557
Adjusted R20.4300.3910.4000.592
Standard error of regression2.1502.4222.1991.547
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

Table 4 adds dummy variables for openness of the various regions with respect to imports and exports, as well as special dummy variables for Singapore’s openness. The Singapore effect is very strong in magnitude, significance, and consistency (although the effect has declined slightly ewer the past 15 years). In 1990, Singapore imported five times as much as a typical country of its size and other characteristics and exported six times as much [exp(1.60) = 4.94 and exp(1.75) = 5.75]. The openness of the other ASEAN countries is no longer strong enough in most years to be statistically significant. As before, the presence of the openness terms (in particular, for Singapore) reduces the significance of the regional bloc effect. In this table, this actually means that the ASEAN bloc loses significance during 1990-94. One possibility is that adding Brunei Darussalam to the set affects the results. As a predominantly oil-exporting country, it naturally trades more with countries outside its region than does a typical country, which would tend to reduce the apparent tendency toward intragroup trade.14

Table 4.Gravity Model with Openness of Regions

(Dependent variable: ln (exportij))

ln (GDPi)0.612**0.472**0.524**0.729**
ln (GDPj)0.605**0.492**0.476**0.620**
ln (DISTANCEij)-1.133**-1.141**-1.073**-0.815**
ASEAN bloc0.790**0.161-0.3330.161
EU bloc-0.851**-0.663**-0.424**0.036
EFTA bloc0.605**0.501*0.700**0.302
ANZCERTA bloc1.259**1.640**1.618**1.706**
NAFTA bloc0.580-0.0210.4070.894**
ANDEAN bloc0.798*1.449**1.304**0.786
MERCOSUR bloc0.852*0.0790.0750.464
ASEAN export openness0.0620.408*0.2270.193
ASEAN import openness-0.1570.3650.0420.163
EU export openness0.650**1.041**0.505**0.179*
EU import openness0.863**0.974**0.489**0.213*
EFTA export openness0.1720.429**-0.222-0.031
EFTA import openness-0.222#-0.320*-0.752**-0.538**
ANZCERTA export openness1.180**1.280**0.692**0.208
ANZCERTA import openness0.2760.488#0.2190.086
NAFTA export openness2.053**1.455**0.630**-0.057
NAFTA import openness1.962**1.167**0.640**0.391**
ANDEAN export openness-0.469**-0.464**-1.089**0.734**
ANDEAN import openness0.192-1.129**-0.725**0.545**
MERCOSUR export openness0.0071.452**1.020**-0.145
MERCOSUR import openness-0.3400.677**0.336-0.443**
SGP export openness1.806**1.752**1.765**1.518**
SGP import openness1.953**1.597**1.432**0.935**
Number of observations2,8542,9572,9852,557
Adjusted R20.5040.4570.4480.618
Standard error of regression2.0062.2892.1101.497
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

We have also tried adding a new variable to measure the remoteness of the exporter and the importer from the world at large. It is computed as the weighted-average distance from trading partners, a variable separate from bilateral distance. (The weights are incomes.) The idea is that remote countries such as Australia and New Zealand will trade more with partners at a given closeness. The remoteness variable itself is not successful here. It does, however, increase the significance level of the ASEAN bloc effect.

We have seen that the level of intragroup trade bias and the trend can be very different. Intra-ASEAN trade is high, adjusting for some factors, but is, if anything, declining over time. If we wish to test the change over time explicitly, it is best to do so by taking first differences of the equation. The price is that such unchanging variables as proximity, common borders, and common languages will be lost. The results will appear to be less precise.

Extending the Tests to Indochina and Myanmar

In Table 5 we focus on the countries of Indochina, plus Myanmar, for 1990-94. (The data are not available for 1980.) These countries have largely been cut off from trade with market economies for the past twenty years, but are now beginning to reintegrate themselves with the world economy, with Vietnam in the lead. An important component of this process is the reestablishment of relations with their southeastern neighbors. Vietnam became a member of ASEAN in 1995. Cambodia, the Lao People’s Democratic Republic, and Myanmar may join as early as 1997.

Table 5.Gravity Model with Indochina Variables

(Dependent variable: ln (exportij))

ln (GDPi)0.473**0.531**0.706**
ln (GDPj)0.489**0.468**0.601**
ln (DISTANCEij)-1.133**-1.069**-0.814**
ASEAN bloc-0.084-1.073*-0.371
EU bloc-0.711**-0.475**-0.020
EFTA bloc0.460*0.675*0.261
ANZCERTA bloc1.700**1.667**1.783**
NAFTA bloc-0.1850.2610.772*
ANDEAN bloc1.498**1.325**0.848
MERCOSUR bloc0.0280.0330.445
ASIND bloc0.1780.6940.462
INDCHA export openness-2.404**-2.762**-1.992**
INDCHA import openness-2.045**-2.578**-2.060**
ASEAN export openness0.354#0.1630.085
ASEAN import openness0.306*-0.0330.066
EU export openness0.988**0.435**0.106
EU import openness0.916**0.432**0.148
EFTA export openness0.373**-0.310*-0.140
EFTA import openness-0.402**-0.833**-0.632**
ANZCERTA export openness1.235**0.640**0.131
ANZCERTA import openness0.4240.1570.001
NAFTA export openness1.412**0.565**-0.098
NAFTA import openness1.135**0.606**0.371*
ANDEAN export openness-0.559**-1.176**0.542**
ANDEAN import openness-1.236**-0.834**0.370*
MERCOSUR export openness1.357**0.930**-0.292*
MERCOSUR import openness0.563*0.204-0.577**
SGP export openness1.770**1.781**1.565**
SGP import openness1.615**1.453**0.975**
Number of observations2,9572,9852,557
Adjusted R20.4670.4630.639
Standard error of regression2.2672.0791.456
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

As one would expect, these countries still show an extreme negative effect for openness.15 As of 1992, their tendency to import was less than 8 percent that of other countries [exp(-2.58) = 0.076], even adjusting (as always) for their levels of income, among other factors. Their tendency to export was even lower, only 6 percent that of other countries [exp(-2.76) = 0.063]. By 1994, some opening had become evident, especially on the export side. Indochina and Myanmar now have an estimated tendency to import that is 12 percent that of other countries [exp(-2.06) = 0.127]. Their tendency to export is 14 percent that of others [exp(-1.99) = 0.136].

There is still enormous room for liberalization. If these formerly autarchic countries restore normal trade relations with the rest of the world over the coming decade, the gravity model predicts that their trade will grow sevenfold from 1994 levels. In addition, their trade will grow in proportion to their incomes. If they grow more rapidly than the worldwide average, their trade levels will grow correspondingly more rapidly.

The gravity model can estimate what projected growth rates will do to Indochinese trade. This requires plausible estimates of growth rates in Indochina as well as in the rest of the world. Vietnam has been growing at about 8 percent a year, about 6 percent a year above the world average, and is forecast to continue to do so.16 The same is true of Thailand (and several other ASEAN countries). In our gravity model estimates that include per capita GDP in the equation, it appears that for every 1 percent increase in a country’s rate of growth of per capita income, relative to the world average, its trade with each partner also grows about 1 percent faster. It follows that Vietnam’s total trade is expected to grow about 6 percent a year faster than the worldwide average (which is about 4 percent a year), and its trade with Thailand to grow about 12 percent a year faster. This is on top of the sevenfold increase predicted for the period during which Vietnam becomes integrated with the world economy. Needless to say, these projections are rough and need to be refined. This is a possible subject for future research.

The term labeled ASIND bloc in Table 5 estimates the effect of a dummy variable for trade within the group that includes Indochina and Myanmar along with the six original ASEAN members. In other words, it tests for a concentration of trade within the group that is scheduled to constitute ASEAN in the year 2000. Its effect is estimated to be positive in all three years, 1990, 1992, and 1994, but is not statistically significant. At the same time, we include a variable for the original ASEAN members. Its point estimate is negative, but again not statistically significant. At this point, we have 26 dummy variables (not counting LANGUAGE and ADJACENCY). Singapore, for example, is counted six times (ASEAN bloc, ASIND bloc, ASEAN export openness, ASEAN import openness, SGP export openness, and SGP import openness), even with-out the East Asia grouping. With this many parameters estimated at once, the reliability of each is diminished.

Australia and New Zealand

Finally, we consider the role of Australia and New Zealand in the region. These two countries underwent thorough liberalization programs in the 1980s. In 1983, the Australia-New Zealand Closer Economic Relations Trade Agreement (ANZCERTA) superseded and expanded a previous accord, to cover all trade: nontariff barriers, subsidies, countervailing duties, antidumping, and government procurement. A 1988 accord expanded this agreement to encompass the use of national treatment for trade in most services between the two countries. The agreement was again slightly expanded in 1992. Thus, it is the deepest integration agreement in the Pacific (World Trade Organization, 1995). Talks were held in March 1995 to discuss closer links between ASEAN countries and ANZCERTA. It has been argued that such links would attract investment to ASEAN and would prod AFTA trade liberalization (Jaggi 1995; Financial Times, July 27, 1995 and December 15, 1995).

Table 6 focuses on the role of the two antipodean countries. As in earlier estimates, they show up as relatively open, particularly on the export side. In earlier results, they were found to trade even more with each other than with third countries, a strong bilateral effect attributable to ANZCERTA. In Table 6, that bloc effect is somewhat diminished, in both magnitude and statistical significance. The reason is plain to see. The equation also includes a dummy variable, ASANZ, for trade within a grouping comprising ASEAN, with Brunei Darussalam, and Australia and New Zealand. The effect of this bloc is very strong, in level as well as in statistical significance. As was true of the East Asia bloc (and, to a lesser extent, the ASEAN-Indochina grouping), the bloc effect for ASEAN seems to lose its significance when the equation simultaneously tests for large groupings in which the ASEAN countries are included.

Table 6.Gravity Model Including ASEAN-ANZ Trade Links

(Dependent variable: In (exportij))

ln (GDPi)0.610**0.471**0.523**0.728**
ln (GDPj)0.604**0.491**0.476**0.620**
ln (DISTANCEij)-1.099**-1.115**-1.053**-0.801**
ASEAN bloc-0.616#-0.956**-1.253**-0.494
EU bloc-0.838**-0.652**-0.416**0.042
EFTA bloc0.610**0.505**0.704**0.304
ANZCERTA bloc0.0490.6940.829**1.164**
NAFTA bloc0.593-0.0090.4160.898**
ANDEAN bloc0.844*1.483**1.330**0.803
MERCOSUR bloc0.905**0.1200.1060.487
ASANZ bloc1.612**1.276**1.052**0.748**
ASEAN export openness0.0380.1640.153
ASEAN import openness-0.2510.293-0.0200.117
EU export openness0.658**1.048**0.510**0.182*
EU import openness0.871**0.982**0.494**0.216*
EFTA export openness0.1790.434**-0.219#-0.027
EFTA import openness-0.217#-0.314*-0.749**-0.535**
ANZCERTA export openness0.987**1.122**0.566**0.106
ANZCERTA impart openness0.0900.3350.094-0.007
NAFTA export openness2.040**1.446**0.623**-0.060
NAFTA import openness1.950**1.158**0.632**0.387**
ANDEAN export openness-0.489**-0.478**-1.098**0.727**
ANDEAN import openness0.209-1.142**-0.734**0.539**
MERCOSUR export openness-0.0151.435**1.006**-0.155
MERCOSUR import openness0.362*0.659**0.324-0.451**
SGP export openness1.808**1.757**1.775**1.519**
SGP import openness1.955**1.598**1.442**0.942**
Number of observations2,8542,9572,9852,557
Adjusted R20.5060.4570.4480.618
Standard error regression2.0022.2872.1091.496
Note: **, *, and * denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

As already noted, it is natural that the estimates of the bloc effects vary, depending on what groupings are included in the equation. But how should one ultimately interpret the results? On the one hand, if one is interested in testing the hypothesis that formal regional trading arrangements affect trade, then one should focus on the ASEAN (and ANZCERTA) results and ignore equations that feature larger groupings of countries that do not coincide with existing regional trading arrangements. On the other hand, one is also often interested in knowing the strength of trade links between, for example, Southeast Asia and Australia and New Zealand. Looking at the simple magnitude of trade flows or at trade shares is not very useful. Adjusting for such factors as size and proximity produces more informative measures of trade links. The gravity results for these ad hoc groupings are perhaps best viewed as sophisticated versions of descriptive statistics.

Role of Foreign Direct Investment

Total net foreign direct investment (FDI) into East Asian developing countries has been estimated at $43 billion in 1994.17 It has doubled every two years since 1987 and has risen more than 32-fold since 1970 (admittedly, in terms of current dollars). Accounts of the bloc that is said to be evolving in Southeast Asia, or all of East Asia, tend to emphasize FDI almost as much as they emphasize trade. Thus, we devote a substantial portion of this paper to the subject.18

U.S. Foreign Direct Investment

The view of FDI as another form of international investment flowing from capital-rich countries to capital-poor countries has an obvious appeal in the case of Southeast Asia. The mainstream view, however, is that FDI is very different from portfolio investment, which is driven by macroeconomic considerations such as interest rates and exchange rates.19 The mainstream view borrows, rather, from the industrial organization literature, emphasizing that FDI is undertaken by large monopolistic corporations that have a special advantage in technology, management skills, or brand name, and that it goes into host countries that have the attractions of either cheap inputs or a large market that is removed from the rest of the world by either transportation costs or trade barriers.20

Until the 1980s, the United States was the dominant investor in Southeast Asia, and the mainstream view seemed to fit well. Early in the postwar period, the Philippines was the dominant destination, and selling into the local market was the dominant motivation. U.S. FDI increased rapidly in the 1970s, especially in Indonesia (much of it in the oil sector), but also in Singapore (where Americans were attracted by the liberalization of the economy). By 1988, Singapore was said to be the location of more than one-fourth of U.S. FDI in Asian manufacturing, particularly in electronics. Investment in Malaysia went specifically into semiconductors, which have increasingly begun to be sold elsewhere in Asia (Encarnation, 1992).

Japanese Foreign Direct Investment

Japanese FDI has received much attention in the past ten years and is the focus of much of the speculation regarding a yen bloc in East Asia centered on Tokyo. While the Japanese data are subject to measurement problems, some major trends are evident.21

In the aftermath of World War II, Japanese investment was small and (in Asia) concentrated in the extraction of natural resources, particularly in Indonesia. Substantial Japanese investment dates from 1972, when the Japanese government removed controls on outward investment. One contributing factor was the beginning of U.S. irritation with Japan’s balance of payments surpluses. FDI could be expected to reduce the overall balance of payments surplus immediately and perhaps the trade balance subsequently. Direct investment in Asian manufacturing was heavy in the textiles and electronics sectors, with most of the output being exported. By the late 1970s, more Japanese FDI than U.S. FDI was going into East Asia.

The first wave of yen bloc theories matched the big wave of Japanese FDI in the 1980s. Rapid growth in the host countries was a major attraction in Southeast Asia. The very sharp appreciation of the yen against the dollar in 1985-87, and the subsequent bubble in the prices of land and equity in Japan, encouraged many Japanese corporations to locate some operations offshore. Environmental concerns led some polluting industries to relocate.22

While manufacturing received the most attention, Japanese investment in the commercial-financial sector was considerably greater in the 1980s than in manufacturing (worldwide). Investment in real estate was also large, while the share of investment in the primary sector, like that in manufacturing, fell off sharply from the high levels of the 1950s-1970s.

Hong Kong,23 Singapore, Thailand, and Malaysia joined Indonesia as the leading Asian destinations of Japanese FDI in the 1980s, as did China in the early 1990s.24 Since much of Japan’s FDI went to East and Southeast Asia, it is deemed an important component of the yen bloc hypothesis. Around 1987, the stock of Japanese FDI in East Asia surpassed the stock of U.S. FDI there. During 1987-91, Japanese FDI constituted 96 percent of total FDI into Indonesia, 26 percent of FDI into Malaysia, 33 percent into the Philippines, 21 percent into Singapore, and 51 percent into Thailand. Graham and Anzai (1994) point out that Japanese FDI makes up a high percentage of total FDI only in countries where FDI does not make up a high percentage of total fixed investment. For this reason, Japanese FDI is, in all countries, less than 10 percent of gross domestic capital formation.

The share of mining in Japan’s Asian FDI is much lower than in the past, with manufacturing, commerce, and finance constituting the major categories. Within manufacturing, textiles constituted fully one-third in the 1950s and 1960s, but are now down to 7 percent. Electrical goods have risen to 27 percent, followed by chemicals, metals, machinery, transport, and foodstuffs. Within electrical goods, the greatest shares were going to Malaysia and Thailand by 1990, representing very strong growth relative to ten years previously.25

Labor costs are undoubtedly the greatest single factor behind Japanese FDI in Southeast Asia. In 1989, the ASEAN Promotion Center on Trade in Tokyo surveyed a large number of firms. Of those that had already invested in Southeast Asia, 61.2 percent cited low-cost labor as the major reason. The second-most-cited reason (40.1 percent) was access to the local market, which includes tariff jumping. Of those firms contemplating investing in Southeast Asia, exporting back to Japan was the number two reason (36.8 percent), after low-cost labor (58.8 percent); these two reasons are of course entirely consistent (Tokunaga, 1992).

Another motive, relevant in such sectors as textiles and consumer electronics, has been quantitative restrictions on imports into the United States, as companies in Japan (or Hong Kong or Taiwan Province of China) switch production to Southeast Asian countries that are not yet constrained by their quotas. Other relevant factors within the host countries include local tax breaks and subsidies, infrastructure, macroeconomic and political stability, and growth rates.

Japanese FDI fell off in 1990-92 (although by less in Asia than in the rest of the world). Several macroeconomic explanations were evident: monetary policy tightened, Japan went into recession, corporate and bank balance sheets were ravaged by the decline in stock and land prices, and the appreciation of the yen eased slightly.26 In 1991 the ratio of the accumulated Japanese FDI stock in East Asia to the U.S. stock reached its peak, at 1.75.

The decline in Japanese FDI flows leveled off in 1993, however. Japanese multinationals began a renewed expansion in 1994, especially in Southeast Asia and China, where their investments were up 52 percent, responding in part to a renewed appreciation of the yen to unprecedented heights. Thus, the yen bloc hypothesis stays alive.

As noted earlier, the growth of Japanese trade with Southeast Asia, which appears to be extremely rapid, can be entirely explained by the rapid economic growth of Japan (until the 1990s) and of the other Asian countries. It turns out that this is also true of Japanese FDI. While a full analysis should await an application of the gravity model, a simple calculation illustrates the point. If one scales by the host region’s size in world trade, one finds that Japan’s investment in East Asia and Australia is almost exactly in proportion to their size. There is no evidence of regional bias. Japan’s direct investment in the United States and Canada, in contrast, is more than twice what one would expect from their share of world trade. Japan’s investment in Europe is about half the continent’s share of trade. As with trade, there is far stronger evidence of a Pacific-wide bloc that includes North America than of an exclusive East Asian bloc.27

Other Foreign Direct Investment in Southeast Asia

Although the United States and Japan used to dominate FDI in Southeast Asia overwhelmingly (together with some European investment), the pattern has changed recently. The East Asian newly industrializing economies have become major investors in the region. Companies have responded to rising wages and appreciating currencies at home by setting up manufacturing operations in neighboring countries with lower labor costs. In the Republic of Korea and Taiwan Province of China, the rising demand for domestic goods and labor, which the government had for a time succeeded in damming up, burst out in the late 1980s in the form of real currency appreciation and real wage increases (especially in 1987-88). The impetus for FDI followed.

We have already listed political instability as a concern to foreign investors. This is of particular relevance in the case of investment coming from Hong Kong, given the uncertainties surrounding the economic policy environment that will prevail after the territory reverts to Chinese control on July 1, 1997.

During 1985-91, Taiwan Province of China’s new direct investment in ASEAN was almost as great (18 percent of the total inflow) as Japan’s(21 percent). Adding in either the Republic of Korea (5 percent), Hong Kong (6 percent), or Singapore (5 percent) easily puts the newly industrializing economies ahead of Japan (let alone Europe at 14 percent or the United States at 7 percent).

Already, investors in the newly industrializing economies have run into rising labor costs in ASEAN countries, particularly Malaysia and Thailand, and are looking to still-cheaper China or Indochina. Over-loaded infrastructure is said to be another factor pushing multinationals to move on. In 1992, Taiwan Province of China companies’ investment in Indonesia, Malaysia, Singapore, and Thailand declined sharply. They are said to be turning to Vietnam, and the Koreans to Myanmar. Both, along with Hong Kong, are also investing heavily in China.

Taking the cycle to the next stage, these four ASEAN countries are themselves beginning to invest in China and Vietnam, much as the newly industrializing economies began investing in them a decade ago, and Japan in the newly industrializing economies two decades ago. It has been reported that 10-15 percent of foreign investment in China is coming from ASEAN.28

Foreign Direct Investment and Trade

Some readers of our earlier work have responded to our finding of no trend toward an Asia bloc in trade by suggesting that there has been a strong trend toward such a bloc in FDI and that this will show up in trade with a lag. We evaluate this effect in the next subsection.

There is certainly a long tradition of connecting FDI with trade. FDI can lead to (1) higher exports from the source country to the host, especially when the investment is in the retail sector or when the subsidiary has a relative proclivity to import intermediate inputs from the mother country; (2) lower exports from the source country to the host when the aim of the investment is to circumvent trade barriers, so that sales within the host market substitute for shipments from the source country; and (3) higher imports into the source country from the host country, especially when the motive for the FDI is cheap labor in manufacturing or raw materials in extractive industries.29 The experience of U.S. multinationals has been extensively studied; the usual finding is that U.S. FDI abroad leads to increased U.S. exports and an improved U.S. trade balance.

Kojima (1985) claims that Japanese FDI is more trade oriented than U.S. FDI. Part of the theory is that the Japanese corporations doing the investing are smaller and more competitive than the U.S. multinationals. This is the opposite of the Dunning-Hymer-Kindleberger theory of FDI in general, and the opposite of the popular American conception of “Japan Inc.,” in particular.30 Kojima’s characterization does seem to fit some industries, such as electrical machinery, one of the two largest manufacturing sectors for FDI. Japanese companies in this sector report that over two-thirds of the sales of their affiliates are exports (to various destinations). This pattern is not entirely typical, however. The transportation machinery sector (chiefly autos and trucks) is at the other extreme: only 20 percent of affiliates’ sales were exports (Graham and Anzai,1994). Urata (1993), Bergsten and Noland (1993), and Encarnation (1992) argue that Japanese affiliates in East Asia are on the whole less export oriented than American affiliates there.

Another issue is whether Japanese affiliates are more prone than those of other countries to import intermediate inputs from the source country. Kreinen (1987), in a study of affiliates in Australia, claims that they are. Others respond that the tendency to import inputs from the mother country is simply an attribute of recent FDI and that Japanese investment in the Asia-Pacific is recent.31

Influence of Accumulated Bilateral Foreign Direct Investment on Trade

The claim that there has been a trend toward an FDI bloc in East Asia and that this can be expected to show up in trade patterns with a lag is well worth investigating. It can be broken into its two constituent propositions: a regionalization of FDI in East Asia and an effect of FDI on bilateral trade. We consider each of these propositions in turn, beginning with the latter.

Thus, in this section, we add bilateral FDI as a variable to explain bilateral trade in the gravity equation. Before doing so, we must acknowledge two serious problems. First is the likely endogeneity of FDI. We address this problem, in a preliminary way, by putting only the lagged cumulative stock of bilateral FDI on the right-hand side of the equation. The idea is that this variable is predetermined, although in a cross sectional study that is not a complete solution. Later, we will address the endogeneity of FDI more fully.

Second is the problem of data. We have already noted the poor quality of the Japanese data, which at least have the virtue of being available. For most pairs of countries, data are simply unavailable. There is no multilaterally gathered universal data bank for bilateral FDI as there is for bilateral trade. Most empirical studies focus on a few key investors. There are comparisons of U.S. FDI by partner, of Japanese FDI by partner, and of U.S. versus Japanese FDI into East Asian countries. But there are no multilateral studies, so far as we are aware.

We have what we think may be the most extensive collection of FDI data among pairs of countries. Most important, for present purposes, it includes not just FDI into ASEAN by the United States, Japan, and European countries, but also FDI by the Republic of Korea and Taiwan Province of China. It also includes FDI undertaken by Thailand, which, unfortunately, is the only Southeast Asian country on which we have data as a source of bilateral FDI going into other ASEAN countries. All the major East Asian countries are included as destinations. (Table 7 lists countries for which data are available.)

Table 7.Countries in the Sample for FDI

United KingdomXXXX
United StatesXXXX
New ZealandX
South AfricaX
Iran, Islamic Rep. of
Saudi ArabiaX
Taiwan Prov. of ChinaXXXo
Hong KongXooo
Korea, Rep. ofXXoo
Brunei Darussalam
Lao P.D.R.X
Source: Organization for Economic Cooperation and Development.Notes: All countries are in the trade regressions. Among the source countries, those denoted by an “o” only have data on FDI going into the Lao People’s Democratic Republic, Myanmar, or Vietnam.
Source: Organization for Economic Cooperation and Development.Notes: All countries are in the trade regressions. Among the source countries, those denoted by an “o” only have data on FDI going into the Lao People’s Democratic Republic, Myanmar, or Vietnam.

Here we focus on the impact of FDI on exports from the source country to the host country. In the future, we hope to look at possible effects of imports entering the source country. Because we are forced to drop all observations for which the FDI data are missing, we must drop the observations on exports from ASEAN countries (except Thailand) to other ASEAN countries.

In Wei and Frankel (forthcoming, Table 4), we looked at the effect of the 1990 stock of bilateral FDI on bilateral 1992 exports. Some important changes in the gravity estimates come about simply from the reduction in sample size (to 347) brought about by discarding observations without FDI data. While the GDP, distance, and adjacency terms are similar, the estimated effect of language falls somewhat, and the coefficient of the exporting country’s GDP per capita becomes negative, although insignificant. The East Asian bloc effect becomes insignificant (although the APEC bloc effect remains, as do the blocs in the Americas and Europe). But East Asia remains by far the most open of the major parts of the world.

The stock of FDI has a positive effect on exports. The effect is extremely high in significance (t-ratios of 7 or 8) and in magnitude (between 0.14 and 0.17, depending on whether one controls for the openness of the East Asian countries and the other groupings). Each 1 percent addition to the stock of investments in a country leads to an increase in exports to that country of about 0.17 percent. There is little change in the other coefficients, except that the coefficient on the GDP per capita of the exporting country now becomes a highly significant negative number (-0.52 to -0.53). The addition of the FDI variable does nothing whatsoever to perk up the East Asia bloc effect (although it does cause a big increase in the Western Hemisphere bloc effect).

Now in Table 8, we extend the tests to 1990 and 1994 and focus on the narrowly defined, or formal, regional trading arrangements rather than the broader grouping. The key point is that the effect of the stock of FDI is again highly significant statistically, estimated at 0.09, 0.07, and 0.19, respectively. It appears, again, that FDI helps promote exports from the source country to the destination country.

Table 8.Bilateral FDI as a Determinant of Trade

(Dependent variable: In (exportij))

ln (FDI stockij)0.09**0.07**0.19**
(previous year’s)(0.02)(0.02)(0.02)
ln (GDPi)0.55**0.55**0.74**
ln (GDPj)0.41**0.39**0.60**
ln (DISTANCEij)-0.49**-0.59**-0.61**
Number of observations307390367
Adjusted R20.640.550.80
Standard error of regression1.031.200.87
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

In the 1994 regression, the stock of FDI in 1992 (the latest available) is used.

Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

In the 1994 regression, the stock of FDI in 1992 (the latest available) is used.

The results for the gravity variables are similar, as before, although the coefficient on adjacency loses some significance, while that on common language loses all significance, probably because of high multi-collinearity with FDI. (Income per capita is omitted from these results.) Alone among the blocs, ANZCERTA remains highly significant. We do not have enough data on intra-ASEAN investment for a true test of an ASEAN bloc effect. The coefficient on Thai exports to the rest of ASEAN in 1990 is positive, but not significant.

Table 9 adds openness of the various groupings as explanatory variables. ASEAN still appears open to foreign products, even holding constant for the large amount of FDI it has been receiving. Some of the other coefficients diminish in size and significance, such as language and the Thai-ASEAN bloc effect.

Table 9.Bilateral FDI and Openness as Determinants of Trade

(Dependent variable: ln (exportij))

ln (FDI stockij)0.10**0.09**0.24**
(previous year’s)(0.03)(0.03)(0.02)
ln (GDPi)0.64**0.62**0.83**
ln (GDPj)0.46**0.43**0.64**
ln (DISTANCEij)-0.64**-0.74**-0.62**
ASEAN export0.70
ASEAN import0.54*0.31#0.16
EU export-0.34-0.49**-0.46**
EU import-0.030.02-0.42**
EFTA export-0.31-0.75-0.29#
EFTA import-0.310.49**-0.54**
ANZCERTA export-0.02-0.100.04
ANZCERTA import0.37#0.28-0.08
NAFTA export-0.56*-0.71**-1.16**
NAFTA import0.390.52*-0.59**
ANDEAN import-0.79*-0.010.59**
MERCOSUR import0.891.27*-0.72**
SGP import0.93**0.97**0.85**
Number of observations307390367
Adjusted R20.680.610.85
Standard error of regression0.971.120.75
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

Table 10 adds, in addition to the usual openness variables, a dummy variable for the openness of Indochina (including Myanmar). Indochina still has a high negative openness coefficient. More surprisingly, the coefficient for Thai exports to other ASEAN countries has now turned significantly negative. (We did not include a dummy variable for Thai ex-ports to Indochina in addition to the members of ASEAN.) Table 11 adds a dummy variable for the ASEAN-ANZCERTA bloc, denoted ASANZ as before. Its coefficient is positive, but not significant.

Table 10.FDI and Openness as Determinants of Trade with Indochina—ASEAN Links

(Dependent variable: ln (exportij))

ln (FDI stockij)0.10**0.11**0.20**
(previous year’s)(0.02)(0.02)(0.02)
ln (GDPi)0.65**0.48**0.76**
ln (GDPj)0.33**0.29**0.52**
ln (DISTANCEij)-0.61**-0.67**-0.72**
INDCHA import-4.19**-4.02**-1.92**
ASEAN export1.44**
ASEAN import0.13-0.29*-0.15
EU export-0.43*-0.85**-0.63**
EU import0.25-0.46**-0.74**
EFTA export-0.35-1.21**-0.65**
EFTA import-0.62**-0.94**-1.02**
ANZCERTA export0.020.070.10
ANZCERTA import0.08-0.19-0.24#
NAFTA export-0.57**-0.74**-1.05**
NAFTA import0.370.30-0.53**
ANDEAN import-1.26**-0.87**-0.19
MERCOSUR import0.040.21-0.85**
SGP import0.89**0.93**0.88**
Number of observations307390367
Adjusted R20.750.760.87
Standard error of regression0.850.870.69
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

In the 1994 regression, the stock of FDI in 1992 (the latest available) is used.

Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

In the 1994 regression, the stock of FDI in 1992 (the latest available) is used.

Thus, FDI has a clear effect on exports throughout that appears to undermine some of the other effects, such as the common language variable and the ASEAN and ANZCERTA bloc effects. Even if one takes the equation at face value, there are insufficient data to pinpoint the effects of specific regional groupings. But one must also worry that bilateral FDI may be endogenous, that it is determined by the same factors that determine bilateral trade, and that it is improperly usurping some of their role in these tests. We now turn to the determination of FDI.

Determination of Bilateral FDI

We now estimate a gravity model of the determination of FDI, analogous to the standard one for trade.32 The results presented here are meant to be only a start and omit many of the possible determinants that have been identified in the literature on FDI. Nevertheless, the gravity framework has its attractions, notably a much larger number of observations than in the typical study of FDI.

To the extent that the motive for FDI is to sell into the local market, one might expect distance and transport costs to have, if anything, a positive effect on FDI, thus reversing a key plank of the gravity model. How-ever, to the extent that the motive is exporting back to the source country, distance should have a negative effect, just as it does for trade. The same is true if distance matters because it breeds unfamiliarity with local culture.33

Table 12 is the baseline gravity model of bilateral FDI. The coefficient on distance is even more significant and negative than it is in the gravity model of trade. Similarly, the coefficient on language (which also includes former colonial links) is extremely high and significant. In 1992, the existence of linguistic links raised the stock of FDI about nine fold [exp(2.24) = 9.4]. In this light, it is not surprising that the addition of FDI to the trade equation in the previous section deprived the language variable of its statistical significance. The coefficients on GDP are also highly significant. But one knows that there is probably a bad misspecification in the equation in this regard. We have not yet included terms for GDP per capita, which would capture the fact that rich countries tend to be the source of FDI.

Table 11.FDI and Openness as Determinants of Trade with ASEAN - ANZCERTA Links

(Dependent variable: ln (exportij))

ln (FDI stockij)0.10**0.09**0.24**
(previous year’s)(0.03)(0.03)(0.02)
ln (GDPi)0.64**0.63*0.83**
ln (GDPj)0.46**0.43**0.64**
ln (DISTANCEij)-0.63**-0.73**-0.62**
ASEAN export0.70
ASEAN import0.54*-0.29-0.16
EU export-0.34-0.49**-0.46**
EU import-0.03-0.02-0.42**
EFTA export-0.31-0.74*-0.29#
EFTA import-0.31-0.49**-0.54**
ANZCERTA export-0.05-0.150.02
ANZCERTA import0.37#0.25-0.09
NAFTA export-0.56*-0.71**-1.16**
NAFTA import0.390.52-0.59**
ANDEAN import-0.79*0.000.59**
MERCOSUR import0.891.26-0.72**
SGP import0.93**0.95**0.84**
Number of observations307390367
Adjusted R20.680.610.85
Standard error of regression0.971.120.75
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

In the 1994 regression, the stock of FDI in 1992 (the latest available) is used.

Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.

In the 1994 regression, the stock of FDI in 1992 (the latest available) is used.

Table 12.Determinants of Bilateral FDI

(Dependent variable: ln (FDI stockij))

ln (GDPi)1.40**0.77**0.37**
ln (GDPj)0.40**0.45**0.85**
ln (DISTANCEij)-0.86**-0.81**-0.70**
Number of observations366301373
Adjusted R20.360.320.45
Standard error of regression2.982.552.02
Note: ** and * denote significance at the 99 percent and 95 percent levels, respectively.

The effects for ANZCERTA and the other blocs are high in magnitude and significance. But for present purposes we are most interested in intra-ASEAN investment. Unlike the other bloc effects, the Thai-ASEAN link in 1990 was a large and statistically significant negative number. Reports of important inter-ASEAN FDI tend to be more recent than 1990. The lack of data on FDI by other countries in the region, or even on Thai FDI after 1990, seriously limits this investigation.

Table 13 introduces dummy variables to capture the openness to FDI of various groupings. ASEAN is particularly welcoming to “imports” of FDI, as are ANZCERTA, NAFTA, and MERCOSUR. Major sources of FDI, even after taking into account their size, are North America, the two European groups, and ANZCERTA. Thailand also has a positive propensity to export FDI, although it is only marginally significant. But the coefficient on Thailand’s FDI in the rest of ASEAN remains very negative, as before. The results are similar when allowing for the level of openness to FDI in Indochina—or the lack thereof (Table 14). There is also no support for an FDI bloc among ASEAN, Australia, and New-Zealand (Table 15).

Table 13.Determination of FDI, with Openness

(Dependent variable: ln (FDI stockij))

ln (GDPi)1.61**0.85**0.28**
ln (GDPj)0.38**0.45*0.77**
ln (DISTANCEij)-1.54**-1.56**-1.40**
ASEAN export1.43#
ASEAN import1.27**1.03**1.05**
EU export3.10**1.72**1.16**
EU import0.610.58-0.03
EFTA export4.32**3.01**0.47
EFTA import-0.37-0.74**-0.90*
ANZCERTA export4.98**2.59**1.85**
ANZCERTA import2.33**1.99**2.23**
NAFTA export3.61**3.31**2.89**
NAFTA import-0.43-1.16**-1.01**
ANDEAN import-0.43-1.16**-1.01**
MERCOSUR import3.77**4.822.10**
SGP import0.880.921.37**
Number of observations366301373
Adjusted R20.610.570.57
Standard error of regression2.332.031.79
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.
Table 14.Determination of FDI, with Openness, Including Indochina

(Dependent variable: ln (FDI stockij))

ln (GDPi)1.65**0.77**0.15
ln (GDPj)0.44**0.39**0.58**
ln (DISTANCEij)-1.54**-1.54**-1.45**
INDCHA import1.21#-0.84#-2.53#
ASEAN export1.31#
ASEAN import1.51**0.83#0.57
EU export3.18**1.61**0.84**
EU import0.78#0.41-0.48
EFTA export4.46**2.83**-0.08
EFTA import-0.19-0.90#-1.51**
ANZCERTA export4.91**2.52**1.80**
ANZCERTA import2.46**1.85*1.90**
NAFTA export3.57**3.31**2.89**
NAFTA import2.99**2.76**1.48**
ANDEAN import-0.09-1.48**-2.05**
MERCOSUR import4.15**4.51**1.80**
SGP import0.870.941.34*
Number of observations366301373
Adjusted R20.610.570.59
Standard error of regression2.322.021.74
Note: **, *, and # denote significance at the 99 percent, 95 percent, and 90 percent levels, respectively.
Table 15.Determination of FDI, with Openness and ASEAN-ANZ Links

(Dependent variable: ln (FDI stockij))

ln (GDPi)1.61**0.85**0.28*
ln (GDPj)0.38**0.42*0.77**
ln (DISTANCEij)-1.56**-1.59**-1.42**
ASEAN export1.46#
ASEAN import1.37**1.25**1.06
EU export3.11**1.72**1.16**
EU import0.610.57-0.05
EFTA export4.33**3.03**0.48
EFTA import-0.37-0.73-0.90*
ANZCERTA export5.28**3.08**2.03**
ANZCERTA import2.37**2.05**2.25**
NAFTA export3.62**3.33**2.90**
NAFTA import2.96**2.84**1.48**
ANDEAN import-0.42-1.14**-1.01**
MERCOSUR import3.80**4.87**2.12**
SGP import0.910.731.54*
Number of observations366301373
Adjusted R20.610.570.57
Standard error of regression2.332.021.79
Note: **, *, and # denote significance at the 99 percent. 95 percent, and 90 percent levels, respectively.

To sum up the results on the determination of bilateral FDI, they seem to be similar to the determinants of bilateral trade, which explains the effect that lagged FDI had in the earlier trade equations in the preceding section. Without more data, we cannot yet draw many more conclusions.

Future Extensions of the Analysis on FDI and Trade

Three extensions are required, in increasing order of complexity. We have not yet used the gravity model to test whether Japan (or the United States) has an extra propensity to invest in ASEAN beyond what would be predicted by, among other indicators, incomes. Nor have we tested whether ASEAN countries’ exports are more sensitive to inward FDI from Japan or from the United States.

We have not yet estimated a complete system consisting of bilateral export equations and bilateral FDI equations side by side. That trade and FDI seem to depend on the same set of variables is not a problem until we start to worry about simultaneous causality between the two.34 We think that the stock of FDI influences trade, and yet the stock of FDI is endogenous. To estimate the simultaneous relationship, we need an instrumental variable. The one we have come up with is a dummy variable representing the existence of an investment tax treaty between two countries. This variable is a significant determinant of bilateral investment and yet, we hope, is uncorrelated with the nongravity determinants. Although have had some success with this variable,35 we have not yet applied the instrumental variables technique to the question of ASEAN per se.

High priority also goes to including per capita incomes in the equation.36 The next step would be to include factor endowments or perhaps wage rates. We have tried entering international differences in factor endowments (ratios of capital to labor and of land to labor and schooling levels) as determinants of bilateral trade in the gravity model. This was not very successful. The tendency for rich countries to engage in more trade than poor countries is far more powerful in the data than any tendency for capital-rich countries to trade with capital-poor countries. This supports the predictions of the modern model of trade in imperfect substitutes, as found by Helpman and Krugman, versus the classical theory of factor-endowment trade of Heckscher and Ohlin (1991). The model of FDI as exploiting cheap inputs—unskilled labor and natural resources—seems particularly relevant to Southeast Asia. Perhaps differences in factor endowments would determine FDI better than they seem to determine trade.37

One might expect that, eventually, the flow of capital (human as well as nonhuman) from rich to poor would equalize capital-labor ratios and thus equalize income levels. (The same would be true of trade, under the Heckscher-Ohlin theory, if the same production technology applied everywhere. But it does not, and FDI is probably a more rapid conduit of technology transfer than is trade.) The equalization would of course take a very long time. But, in any case, two types of shorter-term cycle are driving FDI in Southeast Asia for the time being.

The first is the product cycle (Vernon, 1966). Innovation in the United States yields a new product, say, semiconductors. Initially, it is produced in the United States, with a technological process that is intensive in physical and human capital. Then, after a while, the engineers figure out how to produce the same product with less capital and more labor and land, and the company relocates the manufacturing operation to countries where those factors are more abundant.

The second cycle evident in Asia is that of the “flying geese” metaphor. Just as cheap labor was in the past a major motive behind U.S. and Japanese investment in Singapore and the other newly industrializing economies, when the cost of labor and land there rose, cheap labor became a reason for Singapore and the others to invest in Malaysia, Thailand, and the other ASEAN countries. Now rising costs in those countries are motivating FDI into Indochina. This cascading of FDI is an important component of the famous flying geese pattern.38 To model this process would require making income endogenous, as in the “conditional convergence” growth literature, at the same time that factor endowments are introduced.39

Strategies for ASEAN Trade Policy

What would be the effects of the ASEAN free trade area if it came to full fruition? DeRosa (1993b) uses a computable general equilibrium model to find that AFTA would be trade creating, and that it would expand trade within ASEAN by as much as 21 percent. But most-favored-nation liberalization on the part of ASEAN members (even nonreciprocal) would raise trade by three times as much. In his view, the problem with purely intraregional liberalization, according to such models, is that the Southeast Asian countries mostly produce the same sorts of things. It is necessary to promote trade with outsiders, especially developing countries, to get larger welfare gains. “Overall, the findings … cast substantial doubt on the desirability of pursuing regional economic arrangements …” (pp. 5-6). Another computable general equilibrium study, by Lewis and Robinson (1996, p. 23), reaches similar conclusions: “Creation of an ASEAN FTA based on free trade among ASEAN economies alone offers only very modest gains.”40

Potential Economic Gains of ASEAN Regional Integration

The logic is that the countries are too similar to reap large gains from trading among themselves. The notion that only policies that promote trade with industrial countries are worth pursuing is out of date, however. First, such free trade areas as MERCOSUR are now doing well. Second, the modern theories of trade say that differences in income levels are more likely to have a negative effect on trade than a positive effect (especially for manufactures, which are the basis of rapid growth in East Asian countries, even if they started from a low base). While trade in Asia has in the past fit the factor endowments story better than other regions of the world, the rising traffic in intermediate products within the region—what Krugman (1995) calls the “slicing up” of the value-added chain—suggests the large potential for intraindustry trade. Third, if trade is thought to require large disparities in levels of income or factor endowments, it should be noted that the ratio of per capita incomes between Singapore and Indonesia (approximately 23) is larger than between the United States and Mexico (7) and between Germany and Greece (3). Singapore is only one small country, of course. But even Malaysia has a per capita income that is four times that of Indonesia and roughly ten times that of Indochina.41

Making these comparisons brings to mind a different, opposing conventional wisdom about free trade areas—that in the past they have not worked well between countries of very different income levels. This last judgment is a statement about politics rather than economics. The choice between multilateral and regional liberalization ultimately requires a political context (as does the decision about unilateral liberalization). It is unquestionably true that ASEAN’s trade relations with countries outside the region are more important to the determination of growth and economic welfare of the ASEAN countries than are intra-ASEAN trade relations. The problem is that ASEAN countries have little control over the policies of the United States, Japan, or other major external trading partners. As always, the argument for regional liberalization is a case of the theory of the second best, which takes some distortions as given exogenously.

Regional Versus Unilateral Liberalization

The ASEAN countries do, of course, have control over their own trade policies vis-à-vis other countries; that is, they can pursue unilateral liberalization. According to economic theory, studies of trade-growth links, and recent simulations of ASEAN policy options, they would benefit from removing their own barriers. In most cases, the gain would be much greater than the gain from regional integration. The political process, however, does not offer simple, mutually exclusive choices between unilateral, regional, and multilateral liberalization. Sometimes, leaders mus tmake a choice on regional integration, with unilateral and multilateral policies taken as given.

We suspect that the economic scope for regional integration in Southeast Asia, the potential gains from trade and economics of scale, are greater than others have allowed. The region will increasingly produce interna-tionallyfinished products, rather than only inputs. In the automobile industry, there is the potential for different countries to specialize in different parts of the production process, resulting in a finished product that is interna-tionally competitive. With ten members (after the Indochinese countries are admitted), an ASEAN free trade area would be the largest in the world in terms of population: 450 million people. Already, the 1990s have demonstrated that Asian economic growth is to an extent selfsustaining, no longer hostage to the performance of the U.S., Japanese, or European economies.

Role of Regional Integration in More General Liberalization

This is not the place for a full consideration of the political economy aspects of regional integration. There are myriad respects in which regional liberalization can either undermine liberalization more generally or help build support for it, depending on the circumstances.42 But we will make note of several arguments that run in the optimistic direction.

First is what at the time of the 1994 summit of the leaders of the Asia-Pacific Economic Cooperation (APEC) in Bogor, Indonesia, was called “competitive liberalization”: regional leaders sought to outdo each other in demonstrating their forward-looking vision (Bergsten, 1994). More concretely, competition for investment encourages each country to keep up with the others’ pace of liberalization. Second, regional solidarity can sometimes be invoked to remove barriers that would otherwise be politically sacrosanct. Third, regional agreements can sometimes be used to lock in unilateral liberalization. Fourth, regional liberalization can build export constituencies and thus create domestic political momentum for further liberalization. Fifth, it has been argued that in global negotiations smaller countries can benefit from grouping themselves into larger units.

These arguments were in part behind the proposal of Malaysian Prime Minister Mohammed Mahathir to form an East Asian economic caucus. There are fears among Asian developing countries that an Asia bloc would be dominated by Japan and other fears that an APEC bloc would be dominated by the United States. Here, such subgroups as ASEAN might play a role. Currently, the individual members of these groups have little bargaining power vis-à-vis the world’s two largest economies. But a more unified and integrated ASEAN, perhaps even with a common external tariff and speaking with a common voice, would command more attention. The idea, for Southeast Asian countries, would be to use AFTA as leverage in order to be taken more seriously in APEC and in global negotiations. In the words of an IMF study, “ASEAN may be more important as a forum for collectively voicing the concerns of this group of small and relatively open economies on global economic issues, particularly on world trade” (Robinson and others, 1991, p. 38).


    AkamatsuK.1962“A Historical Pattern of Economic Growth in Developing Countries,”The Developing Economies (March-August).

    AraseDavid1991“U.S. and ASEAN Perceptions of Japan’s Role in the Asian-Pacific Region,”in Japan ASEAN and the United Statesed. byH.Kendall and C.Joewono (Berkeley: University of California, Institute for East Asian Studies).

    • Search Google Scholar
    • Export Citation

    BergstenC. Fred1994“APEC and World Trade; A Force for Worldwide Liberalization,”Foreign AffairsVol. 73 (May-June) pp. 2026.

    • Search Google Scholar
    • Export Citation

    BergstenC. Fred and MarcusNoland1993Reconcilable Differences? United States-Japan Economic Conflict (Washington: Institute for International Economics).

    • Search Google Scholar
    • Export Citation

    CavesRichard1982Multinational Enterprise and Economic Analysis (Cambridge, United Kingdom; New York: Cambridge University Press).

    Center for Research and Communications1994Facing the Asia-Pacific Challenge (Manila, Philippines).

    CookeKieran and others1993“ASEAN Free Trade Zone Sputters into Action,”Financial TimesJanuary26.

    DeardorffAlan1995“Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical World?”NBER Working Paper No. 5377 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    DeRosaDean1993a“Sources of Comparative Advantage in the International Trade of the ASEAN Countries,”ASEAN Economic BulletinVol. 10 (July) pp. 4151.

    • Search Google Scholar
    • Export Citation

    DeRosaDean1993b“The ASEAN Free Trade Area Plan and Intra-Regional Trade in Developing Asia” (Washington: International Food Policy Research Institute).

    • Search Google Scholar
    • Export Citation

    DeRosaDean1995“Regional Trading Arrangements Among Developing Countries: The ASEAN Example,”Research Report No. 103 (Washington: International Food Policy Research Institute).

    • Search Google Scholar
    • Export Citation

    DohertyEileened.1994Japanese Investment in Asia: International Production Strategies in a Rapidly Changing World (San Francisco: Berkeley Round table on the International Economy; Asia Foundation).

    • Search Google Scholar
    • Export Citation

    DonerRichard1991Driving a Bargain: Automobile Industrialization and Japanese Firms in Southeast Asia (Berkeley: University of California Press).

    • Search Google Scholar
    • Export Citation

    DonerRichard1993“Japanese Foreign Investment and the Creation of a Pacific Asian Region,” in Regionalism and Rivalry: Japan and the United States in Pacific Asiaed. byJ.Frankel and M.Kahler (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    DunningJ.H.1979“Explaining Changing Patterns of International Production: In Defence of the Eclectic Theory,”Oxford Bulletin of Economics and StatisticsVol. 41 (No. 4).

    • Search Google Scholar
    • Export Citation

    EatonJonathan and AkikoTamura1994“Bilateralism and Regionalism in Japanese and U.S. Trade and Direct Foreign Investment Patterns,”Journal of the Japanese and International EconomiesVol. 8 (December) pp. 478510.

    • Search Google Scholar
    • Export Citation

    EatonJonathan and AkikoTamura1996“Japanese and U.S. Exports and Investment as Conduits of Growth,” in Financial Deregulation and Integration in East Asiaed. byT.Ito and A.Krueger (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    EncarnationDennis1992Rivals Beyond Trade: America Versus Japan in Global Competition (Ithaca, New York: Cornell University Press).

    FrankelJeffrey1993“Is Japan Creating a Yen Bloc in East Asia and the Pacific?” in Regionalism and Rivalry: Japan and the United States in Pacific Asiaed. byJ.Frankel and M.Kahler (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    FrankelJeffrey1996Regional Trading Blocs (Washington: Institute of International Economics).

    FrankelJeffrey and M.Kahlereds.1993Regionalism and Rivalry: Japan and the U.S. in Pacific Asia (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    FrankelJeffreyDavidRomer and TeresaCyrus1996“Trade and Growth in East Asian Countries: Cause and Effect?”NBER Working Paper 5732 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    FrankelJeffreyErnestoStein and Shang-JinWei1995“Trading Blocs and the Americas: The Natural, the Unnatural, and the Super-Natural,”Journal of Development EconomicsVol. 47 (June) pp. 6195.

    • Search Google Scholar
    • Export Citation

    FrankelJeffrey and Shang-JinWei1994“Yen Bloc or Dollar Bloc? Exchange Rate Policies or the East Asian Economies,” in Macroeconomic Linkage: Savings Exchange Rates and Capital Flowsed. byT.Ito and A.Krueger (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    FrankelJeffrey and Shang-JinWei1995a“Is a Yen Bloc Emerging?” in Economic Cooperation and Challenges in the Pacificed. byR.Rich (Washington: Korea Economic Institute of America).

    • Search Google Scholar
    • Export Citation

    FrankelJeffrey and Shang-JinWei1995b“The New Regionalism and Asia: Impact and Options,”paper presented at the Asian Development Bank conference “The Emerging Global Trading Environment and Developing Asia,”ManilaMay 29-31. Forthcoming inA.PanagariyaM.G.Quibria and N.Raoeds.

    • Search Google Scholar
    • Export Citation

    FrankelJeffrey and Shang-JinWeiforthcoming“Regionalization of World Trade and Currencies; Economics and Politics,” in The Regionalization of the World Economyed. byJ.Frankel (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    FrankelJeffreyShang-JinWei and ErnestoStein1995“APEC and Regional Trading Arrangements in the Pacific,” in Pacific Trade and Investment: Options for the 90s: Proceedings of a Conference Toronto 6-8 June 1994ed. byW.Dobson and F.Flatters (Kingston, Ontario: International and Development Studies Institute, Queen’s University).

    • Search Google Scholar
    • Export Citation

    FrootKenneth1993Foreign Direct Investment (Chicago: University of Chicago Press).

    FryMaxwell1993Foreign Direct Investment in Southeast Asia: Differential Impacts (Singapore: ASEAN Economic Research Institute, Institute of Southeast Asian Studies).

    • Search Google Scholar
    • Export Citation

    GrahamEdward M.1994“U.S. Direct Investment Abroad and U.S. Exports in the Manufacturing Sector: Some Empirical Results Based on Cross-Sectional Analysis;” revised as “On the Relationship Among Direct Investment and International Trade in the Manufacturing Sector: Empirical Results for the United States and Japan” (Washington: Institute for International Economics1996) Forthcoming in Dennis Encarnation, ed., Does Ownership Matter: Japanese Multinationals in East Asia (London, United Kingdom: Oxford University Press).

    • Search Google Scholar
    • Export Citation

    GrahamEdward M. and NaokoAnzai1994“Is Japanese Direct Investment Creating an Asian Economic Bloc?” (Washington: Institute for International Economics). A condensed version appears in Doherty (1994).

    • Search Google Scholar
    • Export Citation

    HeckscherEli F. and BertilOhlin1991Heckscher-Ohlin Trade Theorytrans., ed., and introduced byH.Flam and J.Flanders (Cambridge, Massachusetts: MIT Press).

    • Search Google Scholar
    • Export Citation

    HelpmanElhanan and Paul R.Krugman1985Market Structure and Foreign Trade: Increasing Returns Imperfect Competition and the International Economy (Cambridge, Massachusetts: MIT Press).

    • Search Google Scholar
    • Export Citation

    HirataYoshihiko1994“Characteristics of the East Asian Economic ‘Integration’: Is the Asian Economic Bloc Emerging?”paper presented at the second APEC: NAFTA/ASEAN/SAARC conferenceSeptember 1Bali, Indonesia.

    • Search Google Scholar
    • Export Citation

    HirschmanAlbert O.1980National Power and the Structure of Foreign Trade (Berkeley: University of California Pressexpanded ed.).

    HufbauerGaryDariusLakdawalla and AnupMalani1994“Determinants of Direct Foreign Investment and Its Connection to Trade,”UNCTAD Review pp. 3951.

    • Search Google Scholar
    • Export Citation

    HymerStephen H.1976The International Operation of National Firms: A Study of Foreign Direct Investment (Cambridge, Massachusetts: MIT Press).

    • Search Google Scholar
    • Export Citation

    Institute of Southeast Asian Studies1992Regional Outlook: Southeast Asia 1992–93 (Singapore).

    JacksonTom1991“A Game Model of ASEAN Trade Liberalization,”Open Economies ReviewVol. 2 (No. 3) pp. 23754.

    JaggiGautam1995“Association of Southeast Asian Nations (ASEAN) and Asean Free Trade Area (AFTA),”Asia Pacific Economic Cooperation Working Paper Series No. 95–4 (Washington: Institute for International Economics).

    • Search Google Scholar
    • Export Citation

    KatzensteinPeter and MartinRouse1993“Japan as a Regional Power in Asia,” in Regionalism and Rivalry: Japan and the U.S. in Pacific Asiaed. byJ.Frankel and M.Kahler (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    KindlebergerCharles1969American Business Abroad (New Haven, Connecticut: Yale University Press).

    KojimaKiyoshi1985“Japanese and American Direct Investment in Asia: A Comparative Analysis,”Hitotsubashi Journal of EconomicsVol. 26 (June) pp. 135.

    • Search Google Scholar
    • Export Citation

    KomiyaRyutaro and RyuheiWakasugi1991“Japan’s Foreign Direct Investment,”Annals of the American Academy of Political and Social ScienceVol. 513 (January) pp. 4866.

    • Search Google Scholar
    • Export Citation

    KreinenMordechai1987“Japanese FDI in Australia” (unpublished; East Lansing, Michigan: Michigan State University).

    KrugmanPaul1991“The Move Toward Free Trade Zones,”Economic ReviewFederal Reserve Bank of Kansas CityVol. 76 (November-December) pp. 525.

    • Search Google Scholar
    • Export Citation

    KrugmanPaul1995“Growing World Trade: Clauses and Consequences,”Brookings Papers on Economic Activity: 1Brookings Institution pp. 32777.

    • Search Google Scholar
    • Export Citation

    KwanC.H.1994Economic Interdependence in the Asia-Pacific Region: Towards a Yen Bloc (London; New York: Routledge).

    LeeHiro and DavidRoland-Holst1993International Trade and the Transfer of Environmental Costs and Benefits Technical Papers No. 91 (Paris: Organization for Economic Cooperation and Development, Development Centre).

    • Search Google Scholar
    • Export Citation

    LewisJeffrey and ShermanRobinson1996Partners or Predators? The Impact of Regional Trade Liberalization on Indonesia (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    MenonJayant1996Adjusting Towards AFTA: The Dynamics of Trade in ASEAN (Singapore: Institute of Southeast Asian Studies).

    PanagariyaArvind1994“East Asia and the New Regionalism in World Trade,”World EconomyVol. 17 (November) pp. 81739.

    PlummerMichael1994“Asian Regionalism and U.S. Interests,”paper presented at the conference of the National Centre for Development Studies, Australian National University, “Challenges and Opportunities for East Asian Trade,”July.

    • Search Google Scholar
    • Export Citation

    RamstetterEric1991a“Regional Patterns of Japanese Multinational Activities in Japan and Asia’s Developing Countries,”Economic and Political Studies Series No. 74 (Osaka: Kansai University).

    • Search Google Scholar
    • Export Citation

    RamstetterEric1991b“An Overview of Multinational Firms in Asia-Pacific Economies: An Introduction to the Commonplace Ignorance” (Osaka: faculty of Economics, Kansai University).

    • Search Google Scholar
    • Export Citation

    RobinsonDavid and others1991Thailand: Adjusting to Success—Current Policy Issues Occasional Paper No. 85 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    SteinLeslie1995“Japan’s Direct Foreign Investment,”Center for Japanese Economic Studies Working Paper Series No. 95–1 (Sydney, Australia: Macquarie University).

    • Search Google Scholar
    • Export Citation

    SummersLawrence1991“Regionalism and the World Trading System in Policy Implications of Trade and Currency Zones a symposium sponsored by the Federal Reserve Bank of Kansas City Jackson Hole Wyoming August 22–24 (Kansas City, Missouri: Federal Reserve Bank of Kansas City).

    • Search Google Scholar
    • Export Citation

    TokunagaShojiro1992“Japan’s FDI-Promoting System and Intra-Asia Networks: New Investment and Trade Systems Created by the Borderless Economy,” in Japan’s Foreign Direct Investment and Asian Economic Interdependence: Production Trade and Financial Systemsed byS.Tokunaga (Tokyo, Japan: University of Tokyo Press).

    • Search Google Scholar
    • Export Citation

    UrataShujiro1993“Changing Patterns of Direct Investment and the Implications for Trade and Development,” in Pacific Dynamism and the International Economic Systemed. byC.F.Bergsten and M.Noland (Washington: Institute for International Economics).

    • Search Google Scholar
    • Export Citation

    U.S. International Trade Commission1993East Asia: Regional Economic Integration and Implications for the United States (Washington).

    • Search Google Scholar
    • Export Citation

    VernonRaymond1966“International Investment and International Trade in the Product Cycle,”Quarterly Journal of EconomicsVol. 80 (May) pp. 190207.

    • Search Google Scholar
    • Export Citation

    WangZhen Kun and L.Alan Winters1991“The Trading Potential of Eastern Europe,”Discussion Paper No. 610 (London, United Kingdom: Centre for Economic Policy Research).

    • Search Google Scholar
    • Export Citation

    WeiShang-Jin1996“Foreign Direct Investment in China: Sources and Consequences,” in Financial Deregulation and Integration in East Asiaed. byT.Ito and A.Krueger (Chicago: University of Chicago Press)

    • Search Google Scholar
    • Export Citation

    WeiShang-Jin and JeffreyFrankelforthcoming“Open Versus Closed Regional Trade Blocs,” in NBER East Asia Seminar on Economics Vol. 6ed. byT.Ito and A.Krueger (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation

    World Bank1993World Bank Atlas 1992 (Washington).

    World Trade Organization1995Regionalism and the World Trading System (Geneva).

    YamazawaIppei1990Economic Development and International Trade: The Japanese Model (Honolulu, Hawaii: Resource Systems Institute, East-West Center).

    • Search Google Scholar
    • Export Citation

Note: Jeffrey A. Frankel is Professor of Economics, University of California, Berkeley, and, since writing this paper, Member of the President’s Council of Economic Advisers. Shang-Jin Wei is Assistant Professor of Public Policy at Harvard University’s Kennedy School of Government and Faculty Research Fellow at the National Bureau of Economic Research. The authors would like to thank the Japan-United States Friendship Commission, a U.S. government agency, for research support.


International trade regulation (ITR) 1/13/93, 5/3/95. After advancing the date for creation of the free trade area from 2008 to 2003, the ministers, at a December 1995 summit, put forward a plan to further accelerate the date—specifically, to reduce 89 percent of the tariffs below the 5 percent level by the year 2000 (Asian Wall Street Journal, July 31, 1995; Financial Times, July 22, 1996).


Also, Center for Research and Communications (1994), Cooke and others (1993), De Rosa (1993a), Institute of Southeast Asian Studies (1992), Lewis and Robinson (1996), Menon (1996), and others. Analogous conventional wisdom applies to trade-within Latin America and within Africa.


This statistic is computed by counting the trade between each pair of ASEAN countries twice. This is necessary if the intragroup trade share calculated for individual members is to be comparable with the intragroup share calculated for the group as a whole (e.g., Frankel, 1996). Another way of computing the statistic is to count trade between each pair of members only once (as in Frankel, 1993; or Frankel and Wei, 1994). It makes little difference for cross-group comparisons. The latter measure shows an increase in intra-ASEAN trade (including Brunei Darussalam) from 7.7 percent in 1980 to 9.7 percent in 1990 and 11.8 percent in 1994. If one includes Indochina in the grouping, the data (not available for the 1980s) show an increase from 10.2 percent in 1990 to 11.8 percent in 1994. In both cases, the upward trend is wiped out if one normalizes for the weight of the region in world trade, as explained below.


In this context, Hummer (1994) argues that ASEAN has eschewed timetables and commitments—rather than making and breaking them, like other regional clubs—which is a strength.


Despite widely held fears of a new yen bloc in Asia (e.g., Arase, 1991), Japan is the only industrial country that does not have reciprocal preferential trading arrangements with any neighbors. Proponents of the yen bloc hypothesis argue that Japan is forming an economic bloc in the same way that it runs its economy—through policies that are implicit, indirect, and invisible. This is a hypothesis to be tested with the gravity model (sec next section).


This is not to say that intra-ASEAN trade has increased at only the same rate as ASEAN global trade. Rather, when two economies are growing more rapidly than others, the best benchmark for the growth in trade between them (relative to the world) is the sum of their individual growth rates (relative to the world). This principle emerges from the gravity model.


The issue is fully explored in Frankel (1996); and Frankel, Stein, and Wei (1995).


Not long ago, the gravity model was said to be lacking in theoretical foundations. Then the proposition that trade is proportional to the product of partner sizes was shown to follow naturally from models of trade in imperfect substitutes, as shown by Helpman and Krugman (1985). Today, it seems that the model has an embarrassment of riches: theories compete for the honor of being designated its foundation (Deardorff, 1995). The state of play is perhaps best summed up by pointing out that, if one sets out to explain bilateral trade, one is hound to end up with some version of the gravity model.


We have also tried our tests with a more thorough measure of distance that takes into account land and sea routes, the data generously supplied by Wang and Winters (1991). The results tend to be similar (Frankel, Wei, and Stein, 1995).


The other tree trade areas considered in the statistical analysis reported in Table 2 included the EU, NAFTA, MERCOSUR, the Andean Pact countries (ANDEAN), and Australia-New Zealand (ANZCERTA). A seventh tree trade area, the European Free Trade Association (EFTA), is included in the statistical analysis reported in later tables.


These results are reported in Frankel (1993, Tables 2.2–2.4)


The addition of these countries is one way that the trade results reported here differ from those in earlier studies of ours, such as Frankel (1993), Frankel and Wei (1994,1995a, b), and Frankel, Stein, and Wei (1995).


One cannot take comfort in the idea that a country as small as Brunei Darussalam will not have a major effect on the econometric estimates. A small country counts at least as much as a medium-sized one. (To the extent that large-country data are thought to be more informative than small-country data, heteroscedasticity may be a problem. We have tried an appropriate remedy and found that the basic results do not change. But we have not specifically done this with Brunei Darussalam in the sample.)


As indicated by the negative signs on the dummy variables labeled INDCHA (Indochina and Myanmar) in Table 5.


Merrill Lynch, Global Economics and Currencies, November 15, 1995.


World Bank, World Debt Tables, 1994–95. The strong upward trend continued in 1995 as well.


Recent writings on FDI and possible blocs in Fast and Southeast Asia include Katzenstein and Rouse (1993) and Hirata (1994).


Fry (1993), however, models FDI into Southeast Asia as determined by macroeconomic factors, much like portfolio investment.


’Classic citations include Kindleberger (1969), Dunning (1979), Hymer (1976), and Caves (1982). A recent collection on FDI is Froot (1993).


Ramstetter (1991a, b), Graham (1994), and Stein (1995) explain the data problems. Other general reviews of Japanese FDI include Komiya and Wakasugi (1991) and Encarnation 1992).


See Lee and Roland-Hoist (1993) for Indonesia.


Through June 30, 1997, Hong Kong was administered by the United Kingdom. It was returned to the People’s Republic of China as of July 1, 1997, and became a special administrative region of China.


FDI with China is discussed and analyzed in Wei (1996). The Republic of Korea’s share of Japanese FDI has dropped sharply since 1973.


Patterns of horizontal division of labor across Southeast Asian countries in specific industries are described in Doner (1991,1993) and in the contributions to Doherty (1994).


Stein (1995) finds that credit market conditions in Japan and the exchange rate are Strong explanatory factors for HDI into Asia.


Frankel (1993). The statement extends also to monetary and financial links. The dollar is still by far the leading currency of Asia, not the yen (Frankel and Wei, 1994).


“Japan, NIEs Target China, Indochina,” The Nikkei Weekly, January 11, 1992, p. 20.


The logical fourth possibility, lower imports into the source country from the host country, is dismissed by Graham and Anzai (1994) as not relevant. But Kwan (1994) points out that Indonesia’s exports of raw materials (crude petroleum) to Japan, for example, may fall if Japanese affiliates take over the processing (refining) of the materials on location.


Kwan (1994). In one respect, everyone agrees that Japanese FDI in Southeast Asia fits the “Japan, Inc.” mold: it has been helped along by the sogo shosha, the large trading companies, especially in textiles. Whether the Japanese government in any sense centrally directs the operations of the multinationals is much more controversial, however. The volume-edited by Frankel and Kahler (1993) contains some debate on the subject between economists and political scientists.


Froot (1993). Hufbauer, Lakdawalla, and Malani (1994), in a study of FDI by the United States, Japan, and Germany find, surprisingly, that Japan is the only country where outward FDI consistently raises imports more than exports. But Fry (1993) finds that this is also true of aggregate FDI into Southeast Asia.


Eaton and Tamura (1994, 1996) estimate bilateral gravity models for FDI, but they include only two source countries: the United States and Japan. They find that features of a country associated with more trade with the United States or Japan are also associated with more FDI from those countries.


Eaton and Tamura (1996) find in their gravity model that distance inhibits FDI muchless than it inhibits trade.


Graham (1994) points out the simultaneity, estimates a gravity equation for both FDI and trade (U.S. bilateral), and shows that the residuals are correlated. Eaton and Tamura (1994) do the same, using both Japan and the United States as source countries. But with-out the benefit of instrumental variables, which are excluded from each equation, one cannot disentangle the causality.


Frankel (1996). The apparent effect of FDI on trade vanishes with the instrumental variables technique.


Or, equivalently, population. Eaton and Tamura (1996) argue that population is a key determinant of the extent to which corporations exploit a technological advantage through either FDI or exports.


Militating against this outcome is the huge amount of FDI from Japan to the United States.


Akamatsu (1962), Yamazawa (1990), and Kwan (1994). The inverted “V” pattern was intended by the originators of the (lying geese metaphor simply to describe the rise and fall of a given Asian country’s comparative advantage in a given industry (say, textiles or toys, followed by chemicals, steel, autos, and advanced technology). We think that the same metaphor can be made more vivid if Japan is envisioned as the lead goose in a horizontal “V,” flanked by Singapore and Hong Kong, and the Republic of Korea and Taiwan Province of China, then followed by Malaysia and Thailand, and Indonesia and the Philippines, and finally China and Indochina bringing up the rear (and India as well; some geese are very much larger than others). The lead goose ascertains which economic territory is the most rewarding to enter, and the others in sequence follow the lead of those that went before.


Frankel, Romer, and Cyrus (1996). Our results indicate that the observed effect of trade-on growth and the observed effect of growth on trade each survive the attempt to take into account their simultaneous existence. Thus, for example, simultaneity bias attributable to the endogeneity of income does not appear to have affected our estimates of the gravity model.


Panagariya (1994), like the other, more formal, studies, argues that ASEAN countries would be better off liberalizing unilaterally or multilaterally than through a tree trade area. Kwan (1994) and Plummer (1994) argue that the members of ASEAN are too similar to form a successful free trade area, but that expansion to include Indochina, or to include the newly industrializing economies and Japan, might give the group the necessary economic complementarity.


The figures are for 1992 and use current exchange rates to compare countries’ incomes (World Bank, 1993). The differences are compressed if purchasing power parity rates are used. For example, the ratio of Singapore’s income level to Vietnam’s drops from 71 to 13(for 1994, from Merrill Lynch, Global Economics and Currencies, November 15, 1995).


Chapter 10 of Frankel (1996) is an extended survey of the topic.

    Other Resources Citing This Publication