It has sometimes been argued that “ globalisation ” benefits merely a little figure of states, and that this leads to greater marginalisation of excluded states. This paper argues that globalisation is non needfully biased towards greater concentration in international trade and investing flows. Marginalization is more likely to be explained by domestic policies in comparatively closed states. The paper shows that among comparatively unfastened economic systems, the concentration of international trade and investing flows has declined over the last two decennaries, whereas the opposite is true among comparatively closed economic systems. Therefore, marginalisation is non intrinsic to globalisation.

( JEL F11, F13, F21 )

Cardinal Wordss: Globalization, international trade and investing flows concentration.

____________________

* We are thankful to Sanoussi Bilal and participants at the CEPR/Venice International University workshop on Globalization, Regional Integration and Development, Venice, 31 January 1998. The positions expressed here are those of the writers and do non needfully reflect those of the establishments with which they are associated.

aˆ Economic Research, World Trade Organization, 1211 Geneva 21, Switzerland, electronic mail: Patrick.Low @ wto.org.

aˆ? Economic Research, World Trade Organization, 1211 Geneva 21, Switzerland, and CEPR, London UK ; e-mail: Marcelo.Olarreaga @ wto.org.

A§ Dpt of Political Economy, University of Geneva, 1211 Geneva 4, Switzerland, electronic mail: Javier.Suarez @ ecopo.unige.ch.

Non-technical sum-up

The rapid addition in international trade and investing flows over the last two decennaries is frequently seen as an of import beginning of efficiency additions and growing. However, it has sometimes been argued that the impressive 5 and 12 per centum one-year growing of international trade and investing flows since the early 1970s has non contributed to overall universe growing, but merely benefited a little figure of states. In other words, the statement is that there has been an in-built prejudice that led to a concentration of trade and investing flows among merely a few states, connoting the marginalisation of others in universe trade and investing. This paper argues that there are no grounds to believe that this is the instance, and empirical grounds at the universe degree tends to demo the antonym. The account for marginalisation of some states or parts resides in the domestic policies of the affected states and should non be seen as a natural effect of rapid additions in international trade.

The averment that merely a few states have benefited from the rapid addition in trade, while others have been marginalized, looks believable at a first glimpse. An frequently quoted illustration of marginalisation in universe trade is Sub-Saharan Africa, which accounted for 3.1 per centum of universe exports in the 1950s and saw its portion autumn to 1.2 by 1990. More by and large, Africa ‘s portion of universe exports, for illustration was half its 1985 degree in 1996. Similarly, Latin America lost 14 per centum of its portion during the same period ( from 5.6 per centum to 4.9 per centum ) , whereas Western Europe increased its portion of universe trade by 11 per centum ( from 40.1 to 44.6 per centum ) . Therefore, there is a feeling that the addition in international trade has been mostly restricted to a smattering of states. Similarly, 85 per centum of FDI influxs to developing states are concentrated in merely 10 states ( China entirely accounts for 40 per centum of FDI influxs to developing states ) . However, these figures give merely give a partial image. The portion of Asiatic states in universe trade has increased by more than 25 per centum between 1985 and 1996. Therefore, a full image of what has happened to the concentration of trade flows during the last two decennaries requires a broader geographical attack.

Two inquiries are asked. First: Be the addition in international trade equally distributed across states or has it been concentrated among merely a few states? Whether trade has been equally redistributed at the universe degree or non, there is grounds that some states have been marginalized. The 2nd inquiry is: What has caused the marginalisation of some states in universe trade?

This paper employs three different concentration indexs to research the first inquiry ( Herfindal-Hirschman concentration index, Theil-entropy coefficient and the Mean Logarithm divergence ) . These indexs portion at least two desirable belongingss: foremost, they satisfy the Pigou-Dalton status which implies that any “ transportation ” from a state with a high portion of universe trade to a state with a low portion of universe trade decreases the degree of the concentration index. This may look an obvious belongings, but it is clearly non satisfied when perceivers argue that the portion of Africa ‘s trade in universe trade has declined. Second, they are analyzable, which is a desirable belongings when replying the 2nd inquiry of why some states have been marginalized. We besides allow the indexs to hold different grades of homogeneousness on the degree of universe trade. The thought is to capture the consequence that a rapid addition in universe trade may hold on states ‘ perceptual experiences of their portion of universe trade ( e.g. , a high concentration of international flows may be more burdensome in a universe where few international minutess occur ) . In other words, it may be better to hold a little portion of a big pie than a larger portion of a smaller pie.

The period under scrutiny is 1976-1995 and the sample contains informations for 127 development and developed states. Result show that:

trade and investing concentration indices suggest an equivocal image sing the development of the concentration of international trade and investing flows if we do non account for the important addition in universe trade throughout the period ( i.e. , indexs are homogeneous of grade nothing on the degree of universe trade ) . When indexs suggest an addition in concentration, it appears that this basically occurred among economic systems which have big portions of universe trade and non among little trading spouses. Furthermore, if one corrects the concentration indices to account for the addition in universe trade, so merchandise concentration unequivocally falls throughout the period for any degree of homogeneousness larger than 0.25 ( i.e. , low sensitiveness of the concentration indexs with regard to the degree of universe trade ) .

when spliting the sample of 127 states into unfastened and closed economic systems, it appears that concentration of trade and fiscal flows has unequivocally fallen among unfastened economic systems, whereas it has increased among closed economic systems.

From these consequences, we conclude that marginalisation of some states from universe markets can be largely explained by inward-looking domestic policies. Marginalization in universe trade is non built-in to the globalisation procedure.

1 Introduction

International trade and investing flows have increased more quickly than universe GDP over the last two decennaries.[ 1 ]This rapid growing of international minutess has sometimes been referred to as “ globalisation ” .[ 2 ]Most economic experts would reason that the rapid addition in international minutess may be seen as a beginning of efficiency additions and growing[ 3 ], as states tend to specialise in the production of goods in which they have a comparative advantage.

However, it has sometimes been argued that globalisation has non contributed to overall universe growing, but merely benefited a little figure of states, while many others have failed to harvest the benefits of rapid additions in international trade and investing flows. In other words, the globalisation procedure contains an in-built prejudice that leads to a concentration of trade and investing flows and greater inequality. This paper argues that there are no grounds to believe that globalisation may bring on marginalisation. The account for increasing inequality among states and marginalisation resides in the domestic policies of the affected states.

Section 2 discusses some theoretical and empirical statements to explicate why “ Globalization ” does non needfully take towards greater concentration of international trade and investing flows. It besides reports some grounds on the alterations in the concentration of international trade and investing flows at the universe degree from 1972 to 1995. The grounds is slightly assorted for both investing and international trade flows, and the consequences depend on the type of indexs that are used. When utilizing indexs of concentration that are homogenous of grade larger than 0.25, the concentration of both trade and investing flows have unequivocally fallen during the period 1972-1995. Giving some grade of homogeneousness to the concentration index is justified by the fact that it is better to hold a smaller portion of a large pie than a larger portion of a little pie.

The following measure, undertaken in Section 3, is to sort states into unfastened and closed economic systems in order to place whether alterations in the concentration of international trade and investing flows may be explained by domestic policies. The basic impression of openness is defined in footings of the ratio of trade and investing flows to GDP. These indexs are corrected to account for some of the unfavorable judgments that have been made in the literature by commanding for certain factors, such as the size of the economic system and the portion of non-tradable sectors in entire GDP. For illustration, our rectification shows that one should anticipate big states to hold a comparatively smaller portion of trade in GDP. Therefore, if a big and a little state portion the same trade to GDP ratio, the former should be seen as a more unfastened economic system.

Section 4 estimates the concentration of trade and investing flows from 1972 to 1995, utilizing different indexs of concentration. It shows that there has been a inclination towards a lower degree of concentration of trade and investing flows among unfastened economic systems, whereas the opposite is true for closed economic systems. Section 5 provides some reasoning comments.

2 Does globalisation do marginalisation?

The averment that merely a few states have benefited from “ globalisation ” , while others have been marginalized, looks believable at first glimpse. Africa ‘s portion of universe exports, for illustration, was half its 1985 degree in 1996. Similarly, Latin America has lost 14 % of its portion during the same period ( from 5.6 % to 4.9 % ) , whereas Western Europe increased its portion of universe trade by 11 % ( from 40.1 % to 44.6 % ) .[ 4 ]

As for FDI, the figures suggest a similar province of personal businesss: nine developing states receive 41 % of entire influxs of FDI to developing states in 1993 whereas they represent merely 17 % of entire developing states ‘ GDP, and these figures excludes China which represents 40 % of developing states entire influxs.[ 5 ]Furthermore, developed states ‘ portion of universe escapes is close to 85 % .[ 6 ]

These tendencies are illustrated in Figure 1 below, which shows the development of the portion of sub-Sahara African states in entire universe trade and investing flows. Both portions tended to fall during the period 1976-1995, though the tendency is more impressive for the portion of trade.

Insert Figure 1: sub-Saharan States: development of portion in universe trade and investing flows

Therefore, there is the feeling that “ globalisation has been mostly restricted to a smattering of states ” .[ 7 ]As universe trade and investing flows addition, the statement is that these be given to be more concentrated among a few states. However, the figures given above merely give a partial image of the narrative. Trade and investing flows have besides allowed some developing states to turn faster. Note that the portion of Asiatic states in universe trade has increased by more than 25 % between 1985 and 1996.[ 8 ]Besides, the portion of FDI from developing states in universe FDI more than doubled from 6 % in 1985 to 14 % in 1996.[ 9 ]Therefore, a planetary image of what has happened to the concentration of trade and investing flows requires a broader attack.

The purpose of this subdivision is to look into whether a careful analysis of the development of trade and investing flows over the last two decennaries can corroborate the thought that international trade and investing flows are more concentrated than they were two decennaries or so ago. We calculate different concentration indices across clip for universe trade and universe investing flows for a sample of 144 states ( including both development and developed states ) . It appears that the grounds is assorted, as reported in subdivision 2.2. Section 2.1 describes the different indices that we employed and their belongingss.

2.1 Concentration Indexs

In order to measure the degree of concentration in universe trade and investing flows we employed 3 different indexs. Each of these indexs has different belongingss. The indexs besides portion, at least, two desirable belongingss: foremost, they satisfy the Pigou-Dalton status which implies that any “ transportation ” from a state with a high portion of universe trade to a state with a low portion of universe trade decreases the degree of the concentration index. This may look an obvious belongings but neither the Rawls standard, nor the Quantile analysis, frequently used to claim that Globalization has merely benefited a few states satisfy this. Second, they are analyzable, which will be a desirable belongings in subdivision 5 when the sample is decomposed into unfastened and closed economic systems.

The first concentration index we employed is besides the most normally used index of concentration, i.e. the Herfindhal-Hirschman concentration index ( H ) . It is given by:

where ( 1 )

where are trade or investing flows of state I ; F are entire universe trade or investing flows ( i.e. ) ; therefore, is the portion of state I ‘s trade or investing flows on entire universe trade or investing flows.

The Herfindhal-Hirschman index additions with the degree of concentration. It reaches its upper-bound of 1 with a maximal degree of concentration and its lower-bound of 0 with a minimal degree of concentration. The Herfindhal-Hirschman index is a flow-weighted concentration index which implies that it can be decomposed harmonizing to the portions of entire flows of each group. Therefore, the weight given to each group depends on the trade portion of each group. The Theil information coefficient ( T ) besides portions this belongings and is given by:

( 2 )

The chief difference between H and T is that the former is a bulging map on the portions of universe flows, whereas the latter is a concave map on the portions. This implies that the former is more influenced by alterations in the portion of big states whereas the latter is more influenced by alterations in the portion of little states. A comparing of the development of these two indices may give us some of import information on which states ( little or big in footings of trade and investing flows ) have experienced alterations in their portions. If, for illustration, T is comparatively changeless through clip, whereas H additions, this implies that the addition in concentration has chiefly occurred within the group of states which have a big portion of international flows. Therefore, in this sense, the concave belongings of T may be of peculiar involvement if we are interested in analyzing the development of states who have a smaller portion of international flows.

The chief defect of the Herfindhal-Hirschman and the Theil information indices from our position is that they are sensitive to the figure of observations, in the sense that if in period 0 the universe is divided into two states and each has a portion of 1/2 in universe trade flows, so the index takes the value of 0.5 ; whereas, if in period 1, the universe is divided into 3 states which each has a 1/3 portion of universe trade so the index takes the value of 0.33. This may be a desirable belongings, but it may be misdirecting in our instance, since the figure of states besides varies with the handiness of informations. Therefore our last index is non sensitive to the figure of observations in the sense that regardless of the figure of states in the sample, an equal portion for each state does non impact the value of the index.

The last index is the Mean Logarithm divergence ( L ) which is given by:

( 3 )

where N is the figure of states. Note that L is a population-weighted index which implies that the indexs can be decomposed and the weights given to each group depend on the figure of single ( states ) in each group.[ 10 ]

Note that regardless of the figure of states in the sample, when states have an equal portion in universe flows, L takes the value of 0.

2.1.1 Non-zero-homogeneous concentration indices

The three concentration indices described above are homogenous of degree 0 on entire flows, or in other words, they are invariant to a alteration in the graduated table of the distribution. That is, an addition of 10 % of the trade flow of each state leaves the index unaffected. We may besides desire to look at step which are non zero-homogeneous, which captures the thought that it may be better to hold a little portion of a big pie than a larger portion of a smaller pie. Or instead, that a high concentration of international flows may be more burdensome in a universe where few international minutess occur.

Bourguignon ( 1979 ) proposes two concentration indexs which are non-homogeneous and that generalize the Mean Logarithm Deviation Index and the Theil information coefficient. These are severally given by and below:

( 4 )

where is the mean flow across the universe in a peculiar twelvemonth ; and is the grade of homogeneousness. If, this implies that an addition of 10 % in all states ‘ flows will increase the value taken by the concentration coefficient by 10 % . Note that as for L and T, and are the corresponding population-weighted and flow-weighted analyzable steps of concentration.

Some of the indexs proposed supra have different upper and lower bounds ; therefore, as we are interested in the development through clip of the degree of concentration and non in the degree itself, we report the consequences in regard of each index with a normalized value of 100 in the initial period. An addition in the value of the normalized concentration index corresponds to higher concentration whereas a autumn of the normalized concentration index corresponds to less concentration.

2.2 The Concentration of Flows from 1972 to 1995

Datas are available from 1972 to 1995. Description of the informations can be found in the appendix. To avoid year-specific fluctuations, all variables are taken as a 5-year moving norm. Therefore, our initial observation for 1976 corresponds to the norm from 1972 to 1976, whereas the concluding observation for 1995 corresponds to the norm from 1991 to 1995.

In subdivision 2.1.1 we analyze the development of the concentration of trade flows, and subdivision 2.1.2 we analyze the development of the concentration of investing flows.

2.2.1 The concentration of trade flows

Trade flows for state Is are defined as the amount of exports ( ) and imports ( ) of state I, i.e. . Figure 2 below illustrates the development of the concentration of trade flows 1976 to 1995 for the three zero-homogenous indexs, i.e. H, T and L. The grounds from figure 2 seems equivocal. When regressing the three concentration indices on a clip tendency over the whole period, merely L indicates a positive and important correlativity. H has a positive but undistinguished correlativity, whereas T has a negative and undistinguished correlativity.[ 11 ]However, in figure 2, an addition in trade flow concentration is observed in the late eightiess. The Herfindhal-Hirschman concentration index ( H ) suggests that trade concentration was comparatively stable until the beginning of the 1990s and it has increased since so. In 1995, H was 20 % higher than in 1976. A similar decision can be drawn from analyzing the development of the Mean Logarithm Deviation index ; L was 17 % higher in 1995 than in 1976. However, the Theil information coefficient seems to propose that concentration has remained comparatively stable through clip.

Insert Here Figure 2: Trade Concentration from 1976 to 1995

As antecedently suggested, comparing the development of T and H may be of involvement, given that the former is concave on trade portions and the latter convex. Therefore, the fact that T is comparatively changeless through clip and H additions by 20 % over the period implies that the addition in trade concentration has non occurred among states that have a little portion of universe trade, but among economic systems that have a big portion of universe trade. This information is of import in itself, since it means that smaller trading spouses are non needfully going comparatively smaller through clip.

As discussed above, the three concentration indexs reported in figure 2 are homogeneous of degree 0. Therefore, the fact that Globalization has implied an of import addition in universe trade during the last two decennaries does non impact the concentration index. During this period existent trade flows have increased by 135 % . Had this increased in entire trade flows been every bit shared, and our concentration index L homogeneous of degree 1 ( i.e. ) , so the concentration index would hold fallen by 118 % ( 118=135-17 ) , and trade flows would hold been much less concentrated.

However, the addition in trade flows has non been every bit shared as shown in figure 3 where is reported for different values of. When is homogenous of degree 1 ( i.e. ) , it appears that trade concentration falls by 25 % during the period 1976-1995 ( and non 118 % ) . This was expected, and confirms the thought that if our concentration step is non nonsubjective, in the sense that it accounts for additions in the size of universe trade, so merchandise concentration falls throughout the period for any degree of homogeneousness larger than 0.25, as shown in figure 3. When regressing these three indexs on a clip tendency, we obtained that for homogeneousness grades of 0.5 and 1, the relationship is negative and important, whereas for a grade of homogeneousness of 0.25 the relationship is non important.

Insert Here Figure 3: Trade Concentration and Trade Growth ( )

More distinct decisions can be drawn from figure 4, where is reported for different values of. As T is changeless through clip, it is clear that as universe trade has increased, universe trade concentration has fallen for any with a degree of homogeneousness larger than zero ( i.e. ) . This was confirmed when we regressed these three indexs on a clip tendency, as we obtained a negative and important relationship for all three indexs.

Insert Here Figure 4: Trade Concentration and Trade Growth ( )

To sum up, trade concentration has seemingly increased if we do non account for the important addition in universe trade throughout the period. This addition in inequality has occurred basically among economic systems which have big portions of universe trade. But, if one corrects the concentration indices to account for the addition in universe trade, so merchandise concentration falls for any degree of homogeneousness larger than 0.25.

2.2.2 The concentration of investing flows

Investing flows of state Is are defined as the amount of inward and outward Foreign Direct Investment ( FDI ) , and inward and outward Portfolio Investment Abroad ( PIA ) . Therefore, where inferiors i, o refer to inward and outward flows severally.[ 12 ]To smooth the tendencies in investing flows, a moving norm of these flows is besides taken. Datas are discussed in the appendix.

Figure 5 reports the development of H, T and L throughout the period. When regressing the three concentration indexs on a clip tendency, we found that H is negatively and significantly correlated with a clip tendency, whereas the other two are positively but non significantly correlated with the clip tendency. Therefore, the grounds here is once more equivocal. The first index shows that fiscal flows tend to be less concentrated throughout the period whereas the other two indexs suggest that there has been no important alteration in the concentration of fiscal flows between 1976 and 1995. Figure 5 shows that L has increased throughout the 1980s to a flat 30 % higher in 1989 than in 1976. From at that place on, L tends to fall to a flat 20 % higher in 1995 than in 1976. Therefore, harmonizing to the concentration step L, trade concentration additions until 1989 and falls thenceforth but remains at a flat 20 % higher than in 1976. A similar tendency can be observed for the Theil coefficient T, although the degree of concentration in 1995 harmonizing to T is about equal to the degree bing in 1976. The Herfindhal-Hirschman concentration index shows a autumn of 40 % in the degree of concentration throughout the period, once more with an addition during the 1980s.

Insert Here Figure 5: investing Concentration from 1976 to 1995

As with trade concentration, it is utile to compare the development of T and H, given that the former is concave on the investing flows portions, whereas the latter is bulging. Therefore, the fact that T has remained changeless, while H has fallen, tends to bespeak that the autumn in the concentration of investing flows has basically occurred among states that had a big portion of universe flows. This is the mirror image of what has happened with trade flows, as discussed in the old subdivision.

If the image looks slightly equivocal, the ambiguity disappears when we allow the concentration index to take history of the big addition in universe investing flows that has occured ( existent universe investing flows have increased by 794 % over the period ) . As reported in Figures 6 and 7, the concentration of investing flows falls for any degree of homogeneousness larger than 0.25 in the concentration indices and. This is confirmed when regressing these six indexs on a clip tendency ( except for, where the relationship is negative but undistinguished ) .

Insert Here Figures 6 and 7

To sum up, the development of the concentration of investing flows is comparatively equivocal and depends on the index that is chosen. However, it appears that the degree of concentration has fallen among states that had a big portion of universe investing flows. Furthermore, if we correct the concentration indexs to account for the addition in investing flows at the universe degree, so the concentration of investing flows has fallen irrespective of the concentration step we use.

Therefore, contrary to what has been sometimes suggested, it appears that the addition in universe trade and investing flows has non ( merely ) been limited to a few states. Before pulling more decisions, we consider whether the marginalisation of some states may be explained by domestic policies. We measure the development of the concentration of international flows among two set of states: rapid and slow integration economic systems ( i.e. quickly opening and easy opening economic systems ) This is done in subdivision 5. First, in subdivision 4, we build a trade and investing openness index to sort states into quickly and easy integrating economic systems.

3 Rapid and Slow Integrating economic systems: 1972-1995

Many writers have already done the sort of state categorization contemplated here, so one may inquire why we should reiterate the exercising once more. At least two grounds can be given: foremost, for internal consistence within the paper ; 2nd, because the authoritative openness indexs have been criticized on several evidences and we will seek to rectify at least for some of these reviews.[ 13 ]

In subdivision 3.1 we build the openness index for trade and in subdivision 3.2 we focus on international investing flows.

3.1 International Trade Openness Indicator

The basic trade openness index we are utilizing is the authoritative ratio of trade to GDP. Thus for state I this is given by:

( 5 )

where is the basic trade openness index in state I and is the Gross Domestic Product of state I.

As one is interested in existent effects and non monetary value effects, all these variables are estimated in changeless 1987 dollars. This controls for alterations in trade to GDP ratios that are simply due to alterations in dollar monetary values. This is of peculiar importance for states that trade goods which have a high volatility of monetary values and states with high rising prices. Similarly, GDP in different states is estimated in 1987 US monetary values so that we do non undervalue the GDP of low-price states. This is done utilizing the World Bank buying power para index.[ 14 ]Finally, all variables are once more taken as a 5-year moving norm from 1976 to 1995 to command for year-specific exogenic dazes. This allows us to concentrate on chief tendencies.

We correct the basic trade openness index to account for differences in state size and degrees of development. Indeed, it has frequently been argued that big states in footings of GDP and/or population tend to merchandise less, as there is larger range for trade within the state.[ 15 ]Similarly, it has been argued that states with high degree of GDP/capita may besides be biased toward holding a lower degree of trade to GDP ratio.[ 16 ]The ground is that as states develop, the portion of the service sector tends to increase, and the service sector is mostly non-tradable. To account for differences in state size and degrees of development, we considered the undermentioned arrested development:

where inferior I is for states and T for clip ; the squared footings control for possible u-type relationships ( this may happen, for illustration, if as economic systems get richer the services sector portion becomes larger, but besides at really high degrees of development ( high GDP/capita ) economies start trading services and therefore the non-tradable sector becomes smaller ) . Obviously, the arrested development can non be run as such due to multicollinearity jobs. Besides the squared of the population turned out to be undistinguished ( though it had the right negative mark ) . Thus the forced arrested development we run in panel ( 2540 observations ) is given by:[ 17 ]

where. Consequences of field OLS calculator are reported in table 1.[ 18 ]

Insert Here Table 1: Correcting the trade openness index

The coefficients tend to hold the expected marks and indicate that trade openness falls with population and that the relationship between openness and GDP and GDP/capita has a u-shape. That is, for little degrees of GDP and GDP/capita, the higher the GDP or the GDP/capita the lower the degree of openness. This confirms our anticipations. For sufficiently high degrees of GDP and GDP/capita, the relationship is reversed. In the instance of GDP/capita, this may be explained by the fact that one time a state becomes sufficiently rich it besides starts to merchandise services, as argued before. Because of the restraints due to multicollinearity, it is impossible to place at which degrees of GDP and GDP/capita the relationship alterations. However, as the coefficient of the squared GDP term is comparatively little, we presume that this occurs at comparatively high degrees of GDP, whereas as the coefficient of the squared GDP/capita term is comparatively big, the relationship alterations at comparatively low degrees of GDP/capita.[ 19 ]In amount, consequences tend to corroborate that larger and poorer states tend to merchandise less.

We so construct the fitted value of TI from the above arrested development, which tells us what is the “ normal ” grade of openness of a state with a given GDP and GDP/capita. That is:

( 6 )

Finally, our corrected trade openness index, denoted by, is given by:

( 7 )

Therefore, the corrected trade openness index tell us what the openness divergence is of state I with regard to the “ normal ” openness of a state with the same GDP, population and GDP/capita. If, so state I is more unfastened than norm, whereas if state I is more close than norm.

As we are interested in the development through clip of states ‘ openness we calculate the rate of trade integrating between 1976 and 1995. This is given by:

( 8 )

Therefore when, the state has become more unfastened in the period 1976-1995 whereas when, the state has tended to go less unfastened during the given period. To rectify for the fact that extremely unfastened economic systems in the initial period ( i.e. , 1976 ) may see more trouble in fostering unfastened their economic systems than states that were comparatively closed in 1976, we run the undermentioned arrested development:

Therefore, we control for the initial grade of openness by running the above OLS arrested development across states ( 112 observations ) . The consequences can be found in table 2.

Insert here table 2

We now build the corrected rate of trade integrating index by taking the fitted value of the above arrested development ( ) and comparing it with the existent value of. Thus, the corrected rate of trade integrating, denoted by is given by:

( 9 )

Therefore, when, state I has been opening more rapidly than the mean state with the same degree of trade openness in 1976. Table 4 studies the ranking of states harmonizing to. Postpone 3 besides reports the categorization of states into fast trade incorporating states and slow trade incorporating states harmonizing to whether is larger or smaller than one. This will let us in subdivision 5 to cipher the development of trade concentration among fast and slow trade incorporating states.

Insert here table 3

3.2 International Investment Openness Indicator

The basic investing openness index we are utilizing is the authoritative ratio of international investing flows to GDP. Thus for state I this is given by:

( 10 )

where is the basic investing openness index in state I.

To sort states into fast and slow investing incorporating states we proceed in the same manner as for trade integrating.

We foremost command for size and GDP/capita. Results of the restraint arrested development are given in Table 4. As for trade openness, the larger the population of a state, the lower its degree of investing openness. The relationship between investing openness and GDP and GDP/capita is besides U-shaped. Therefore for low degrees of GDP and GDP/capita, the higher GDP or GDP/capita, the lower the degree of investing openness, whereas for high degrees of GDP and GDP/capita, the relationship is reversed.

Insert here table 4

We so construct the corrected investing openness index, denoted, which is given by

( 11 )

where is the fitted value of.

As in the instance of trade, we are interested in the development through clip of the degree of investing openness so we build an index of the velocity of investing integrating, denoted, which is given by:

( 12 )

As earlier, we corrected for the initial status by running the OLS arrested development of on. Consequences are reported in Table 5.

Insert here table 5

From there we build the corrected rate of investing integrating index, which is given by:

( 13 )

where is the fitted value of the arrested development reported in Table 5.

Table 6 studies the ranking of states harmonizing to and sort them into fast and decelerate investing incorporating states.

Insert here table 6

4 The Concentration of International Trade and investing Flows among Fast and Slow Integrating Economies

In the old subdivision we classified states into fast and slow integration states. This allows us to break up our sample into these two classs. In order to capture the thought that domestic policies instead than Globalization itself have been the cause of the diminution of some states in footings of trade and investing portions, we will cipher the different concentration indices presented in subdivision 2 with regard to these two sets of states.

4.1 Trade and Investment Concentration among fast integration states

Figures 8 and 9 show the development of the trade and investing concentration indices over the period 1976-1995 for fast integrating states. It appears that both trade and investing concentration degrees have fallen between 1976 and 1995 for fast integrating states. When regressing these six concentration indices on a clip tendency, we obtained a negative and important correlativity at the 99 % degree for two of the trade concentration indices ( H and T ) and a negative and important correlativity at the 90 % degree for the staying index L. Concerning investing flows, the correlativity is negative and statistically important at the 99 % degree for L, negative and important at the 90 % degree for T, and negative but undistinguished for H.

One should observe that there has been an addition in the investing concentration indices during the period 1985-1990 ( this may be due to the early 80 ‘s debt crisis ) . But the concluding degree in 1995 of all three indices is lower than in 1976. The autumn in the trade concentration indices has been about monotone to make a degree of trade concentration 20 % lower harmonizing to the Herfindhal-Hirschman index in 1995 than in 1976 ( 8 % and 5 % harmonizing to the Theil coefficient and Mean Logarithm divergence severally ) .

Insert here figures 8 and 9

4.2 Trade and Investment Concentration among slow integration states

All concentration indices for both trade and investing show that concentration in trade and investing flows have increased over the period 1976-1995, as shown in Figures 10 and 11. This is confirmed when regressing the six indexs on a clip tendency as they are all positively and significantly correlated at the 99 % degree with the clip tendency ( except for L in the instance of fiscal flows, which is negatively and significantly correlated at the 90 % degree ) .

Insert here figures 10 and 11

A comparing of figures 8 to 11 suggests that if trade and investing flows may be more concentrated at the universe degree, this may be merely explained by the fact that some states remain comparatively closed and do non take part in the Globalization procedure. Therefore, Globalization does non inherently create marginalisation. Rather, states marginalize themselves.

5 Concluding Remarks

The purpose of this paper was to find whether trade and fiscal flows have tended to be concentrated among a few states during the period 1972-1995. The indexs of concentration that we have used in this paper tend to bespeak that there has been no clear tendency towards more concentrated trade and investing flows. Furthermore, when the concentration indexs are adjusted to take history of the addition in universe fiscal and trade flows ( i.e. , for the size of the pie ) , it appears that both trade and fiscal flows are less concentrated today than in the early 1970s.

We ranked states into fast and slow-integrating states and calculated the assorted concentration indexs for each of these groups of states. It appears that the concentration of trade and fiscal flows has fallen among quickly incorporating states, whereas it has increased among slow-integrating states. We argue this shows that marginalisation of single states from universe markets can be largely explained by inward-looking domestic policies and hence that marginalisation is non built-in to the globalisation procedure.