Victor Bulmer-Thomas. Summary INTRODUCTION. The Wider Caribbean in the 20 th Century: A Long-Run Development Perspective

The Wider Caribbean in the 20 th Century: A Long-Run Development Perspective Victor Bulmer-Thomas PhD, Director, Royal Institute of International Affairs. Summary This paper takes a long view of economic development and inter-country income inequality in the 28 countries making up the wider Caribbean. Using a specially constructed data base in two parts (1900-1960 and 1960 to 1998), the paper uses exports per head as a proxy for living standards in the first period (1900-1960) and finds that the coefficient of variation widened significantly after 1900. This finding is supported by evidence on public revenue per head, another proxy for living standards. This confirms the hypothesis that inter-country income inequality widened in the first half of the century, as some countries were much more successful than others in diversifying exports. In the later period (1960-1998), for which GDP per head figures have been constructed for all countries, inter-country income inequality narrows as a result of other countries following the example of earlier success stories in introducing new exports, especially services. This finding is supported by the evidence on public revenue per head and years of schooling per head for the same period. However, the gap between the richest countries and the poorest remained enormous, confirming that geography -as far as the Caribbean is concerned- is not destiny. INTRODUCTION This paper explores long-run development patterns in the wider Caribbean over the last 100 years. It uses a specially constructed data base for 28 Caribbean countries divided into two time periods: 1900-60 and 1960-1998. 1 All countries of the Caribbean have been included whether or not they are independent together with the mainland countries of Belize and the three Guyanas. Where country names have changed, the most recent is always used. Where countries have separated (e.g. Aruba from Netherlands Antilles), an effort has been made to treat them consistently as if they had been separate entities I N T E G R A T I O N & T R A D E 5

throughout the last 100 years. The result is a data base covering all the former British colonies (12 countries), former Dutch (Suriname), French (Haiti) and Spanish (Cuba and the Dominican Republic) colonies (4 countries), the British Overseas Territories (5 countries), the US territories (2 countries), the Dutch-speaking territories (2 countries) and the French Departments d Outre Mer (DOM)(3 countries). One hundred years is a long period in economic development. One has only to think of the transformation of the US economy between 1800 and 1900 or the changes in Japan between 1870 and 1970. An annual growth of income per head of three per cent leads to a doubling in just under 25 years and therefore a 16-fold increase in living standards in less than a century. It is possible therefore for countries with even modest growth rates to make the transition from poor to rich status in a relatively short period of time. The wider Caribbean was a poor region in 1900 and it still suffers a hundred years later from a large gap with rich countries. Thus, we are not looking at a region that has transformed itself from less developed to developed in this time period. This is not a success story in contrast to South-East Asia for example. However, the wider Caribbean includes a wide range of cases from very poor (e.g. Haiti) to very rich (e.g. Cayman ). Thus, it is relative performance that needs to be explained as well as absolute performance. This is all the more challenging given the assumption made by many that geography is destiny (Gallup, Sachs and Mellinger [1999]). Clearly, in view of the wide range of living standards, that is not true of the wider Caribbean. In this paper we proceed as follows: the first part is devoted to an examination of the long-run trends found between 1900 and 1960. This section relies heavily on the data base referred to above, although it has been supplemented by secondary sources in view of the gaps that still remain in the data base for this period. The second part of the paper examines the trends in the period after 1960; this also makes use of the data base which for some series includes data on all 28 countries for every year. The third part of the paper provides the conclusions, while the statistical appendix contains a number of tables from the data base. 2 I. FROM 1900 TO 1960 At the beginning of the 20th century, the Caribbean countries had all the characteristics of classic developing countries: high birth and death rates leading to modest demographic expansion; a productive structure orientated towards exports; an export structure dominated by one or two primary products; and a heavy dependence on trade with the colonial power. In the case of the three independent countries (Haiti, Dominican Republic and Cuba), the situation was very similar as the United States was playing (Cuba) or was about to play (Haiti and the Dominican Republic) a de facto colonial role. None of the indicators available to us suggest that there were at this stage any exceptions to the generalised poverty and backwardness of the region. DEMOGRAPHY As in Central and South America, the crude death rate (CDR) began to fall steadily after 1900 as a result of medical advances in treating such tropical diseases as yellow fever and malaria, 3 improvements in sanitation and better social conditions. 4 With unchanged crude birth rates (CBR), 5 this led to a rise in the annual rate of population growth from approximately one per cent in the first third of the century to two per cent in the second 6 I N T E G R A T I O N & T R A D E

third. However, the small size of the population in some countries meant that emigration was often just as important as the CBR and the CDR in determining the population increase. These migration patterns took workers not only to the United States, but also to Panama to build the canal, to Central America to develop the banana industry and to Europe during the two world wars. 6 As a result of migration, the growth of population in some of the smaller countries was therefore far lower than might be expected from an examination of CBRs and CDRs. The population of Montserrat, for example, was almost the same in 1960 as it had been in 1900. 7 The same was true of the Turks and Caicos as well as the Cayman. Only in the larger countries, such as Haiti, did the demographic pattern conform to what was occurring in the rest of the developing world. And in a small number of countries there was net inward migration with Cuba in particular receiving a significant number of migrants. 8 The growth of population is shown in Figure 1 for the wider Caribbean. In 1900 there were just under 7 million inhabitants less than London or New York today. Cuba had the largest population with nearly 1.6 million followed by Haiti (1.25 million), Puerto Rico (953,000) and the Dominican Republic (600,000). Jamaica (750,000) was the most populous of the British colonies and Anguilla (3,872) the least. The largest of the French colonies was Martinique with 215,000, while the largest of the Dutch territories was Suriname estimated at just under 80,000. By the end of the period in 1960, the population had trebled to 21.3 million (see Figure 1) not much larger than the conurbation of São Paulo today. This overall growth conceals wide differences at the national level (see Table A.1 in the statistical appendix). While the population of Trinidad & Tobago, for example, approximated the regional trend, the population of Barbados barely increased over the sixty years. By contrast, the population of Cuba increased almost five-fold to nearly 7 million the same size as the Caribbean as a whole in 1900. EXPORTS The small demographic size of all the Caribbean countries has made it virtually impossible to use the domestic market as the main destination for production. Exports have therefore played a crucial role in the productive system and have provided the foreign exchange for the import of both luxury goods and necessities. For that reason, the size of the export sector in relation to the rest of the economy has traditionally been a proxy for the standard of living. In the absence of GDP figures, we can therefore use exports per head as a crude approximation for the trend and variance of GDP per head in the Caribbean. 9 The mean of exports per head 10 is shown in Figure 2. In 1900 its value was just over US$14. Following the First World War, there was a sharp rise that reflected the shortlived boom in sugar prices (by far the most important Caribbean export at that time). The sugar price then fell back and with it the value of exports per head. The 1930s depression led to another steep fall in the value of exports per head, which was only reversed at the end of the Second World War. From then until 1960, the value of exports per head rose rapidly aided by the dollar inflation as well as increases in volume. The mean value of exports per head is heavily influenced by the export performance of a small number of countries in the Greater Antilles (Cuba, Dominican Republic, Haiti, Jamaica and Puerto Rico). Another way of showing the data is to use the unweighted mean, i.e. the average of exports per head for all the countries. This is done in Table 1 for 1900 to 1960, where the unweighted mean is reported alongside the standard deviation, the coefficient of variation, the median and the extreme I N T E G R A T I O N & T R A D E 7

values. The unweighted mean is usually higher than the weighted mean, reflecting perhaps the greater openness of small countries, although this was not true in 1905 nor during the sugar boom at the end of the First World War. The generalised poverty of the region in 1900 is captured by the low standard deviation and coefficient of variation (CV). 11 The median is also very similar to the mean, suggesting an absence of skewness in the distribution. Even in 1900, however, there was a significant gap between the highest and the lowest value for exports per head. The value for Trinidad & Tobago (the highest) exceeded the value for Haiti (the lowest) by a factor of twenty-one. Ten years later (1910) little had changed. The mean, standard deviation and coefficient of variation had all risen (see Table 1), but the gap between the highest and lowest values was still modest. Furthermore, St. Kitts had overtaken Trinidad as the country with the highest level of exports per head suggesting that no country was as yet able to pull away from the rest on a permanent basis. Cuban and Puerto Rican exports, however, had risen rapidly as a result of the tariff preferences on sugar granted to the islands by the United States. At this time there was very little diversification of exports. Bananas had begun to be exported from Jamaica, Belize and a few other countries. Arrowroot was already a staple in St. Vincent. The most important cases were gold from Guyana and asphalt from Trinidad. The British colonies as a whole, however, were still very dependent on the export of sugar, which had been in decline for much of the 19th century as a result of the reduction of imperial preference and the rise of beet sugar in continental Europe. This had reached a crisis in the last two decades of the 19th century forcing the British government to appoint a Royal Commission in 1897. The report had no immediate solutions to offer, but it did argue in favour of ending the system of bounties under which beet sugar exports were subsidised. This was partially achieved by the Brussels Convention of 1902 and the conditions under which sugar was exported from the Caribbean did improve, although not so radically as to lead to a major increase in the value of foreign exchange receipts (Burn [1951], Chapter 6). A brief surge in exports per head followed the First World War. This was made possible by exceptionally high prices for sugar in 1920, which helped to push Cuban and Puerto Rican exports per head in particular to record levels. The sugar boom proved shortlived. Yet the 1920s did bring one important change - the introduction of oil refining in the Netherlands Antilles. Exports per head in the Dutch colony by the end of the 1920s were far higher than in the rest of the Caribbean and remained so for the rest of the period. This had a dramatic effect on the coefficient of variation as well as the ratio between the highest and the lowest values (see Table 1). It also explains the rise in the mean between 1920 and 1930 at the time when the median had fallen back to its 1900 level. During the Second World War both exports and imports were subject to numerous difficulties. In Dominica, for example, the level of exports per head had fallen to $6.5 by 1940 and the average for all British colonies was a mere $30.2. Cuban exports per head, as well as those from the Dominican Republic, were also back to the level in 1900 while exports per head from Suriname were far below what had been achieved at the beginning of the century. Haiti, whose exports per head were almost invariably the lowest throughout the period, saw their value drop below $2 at the beginning of the war. By 1950 the mean value of exports per head had increased significantly from the British colonies and from the other countries as well. Much of this was due to inflation, but the high coefficient of variation (see Table 1) suggests that at least some countries were pulling away from the rest. The figures, however, do not include exports of services which 8 I N T E G R A T I O N & T R A D E

would acquire special importance after the Second World War. These are included in the 1960 data in Table 1. By 1960 there were a small number of countries with very high levels of exports per head. These included, among the British colonies, the Bahamas (boosted by tourism and financial services) and Trinidad & Tobago (helped by oil). The star performers, however, were found elsewhere. The US Virgin and the Netherlands Antilles benefited from oil refining, while Puerto Rican exports per head reflected the gains made under Operation Bootstrap since 1940 in the export of manufactured goods. There were some countries by 1960 where exports per head were well below the regional average. These included Haiti, where exports per head had consistently lagged behind other countries throughout the century, as well as the Dominican Republic. Two of the French territories (French Guyana and Martinique) also lagged far behind the regional average. Indeed, French Guyana had the second lowest value of exports per head (after Haiti). The worst performing British colony was Montserrat. PUBLIC REVENUE There is a tendency for public revenue to increase as a share of GDP with rising incomes. Thus, public revenue per head is an alternative proxy for the trend and variance in GDP per head. Since public revenue is spent in part on education, health and infrastructure, all of which are assumed to impact positively on living standards, there is a two-way causality running between public revenue and GDP. This makes trends in public revenue per head of special interest. The data on public revenue per head are less complete than for exports. The main results are presented in Table 2 for the period from 1900 to 1960. The unweighted mean in 1900 was seven dollars (the weighted mean - $2.5 was even lower). Even after making allowance for the change in the price level, this suggests a parsimonious attitude to public spending. This is borne out by the low standard deviation and coefficient of variation, which suggest that even the more prosperous colonies were subject to tight restrictions on public spending. By 1910 the position was virtually unchanged. The main exception was Suriname where public revenue per head ($24) was the highest in the region and much higher than the average for the British colonies ($7.4). The sugar boom at the end of the First World War brought a predictable surge in the mean value of public revenue per head (see Table 2). This was especially true of Cuba (by now the highest in the region), but not in Puerto Rico despite the importance of sugar to the island. By 1930, however, the value of revenue per head had failed to increase. This was mainly as a result of the fall in the value of trade (especially imports), on which public revenue depended. Throughout the Caribbean social conditions deteriorated sharply in the 1930s leading to social unrest and in some cases riots. The British colonies were not exempt from these pressures and a Royal Commission was set up in 1938 to make proposals for improvements. The Commission s report was not published until 1945, but its findings echoed the gloomy picture presented by W.M. Macmillan in his Warning from the West Indies [1936]. The Second World War provided little opportunity to raise public spending, although a small number of countries benefited from the opening of US bases as a result of the wartime agreement with Great Britain. Although these bases later became a source of political tension (particularly in Trinidad), they did lead to a substantial transfer of resources, an increase in employment and a market for local producers. I N T E G R A T I O N & T R A D E 9

The value of public revenue per head rose sharply after the Second World War (see Table 2), but the low value of the coefficient of variation suggested that little had changed with regard to colonial spending at least in the British colonies. Public revenue per head in Puerto Rico, for example, was far higher than the average for the British colonies as was the figure for Suriname. GROSS DOMESTIC PRODUCT (GDP) The data on GDP before 1960 are very incomplete. There is a series for Cuba and Puerto Rico for the whole period and for a very small number of the British colonies. In the 1940s, however, Benham estimated GDP per head for several countries and the British authorities collected data also on the structure of the labour force. This is often a useful proxy for the state of development since there tends to be an inverse relationship between the level of GDP per head and the proportion of the labour force in agriculture. This information is presented in Table 3 for the 16 British colonies. Barbados stands out as the country with one of the lowest shares for the agricultural labour force (27.6%) and the highest income per head ( 30). The poorest colonies recorded were St. Lucia and St. Vincent. However, the highest agricultural labour force share was found in Montserrat (59.9%) so that it would not be surprising if GDP per head there had been even lower. The most complete figures on GDP per head are for the final year of the period under examination (1960). For the 28 countries we have data of variable quality giving an unweighted mean of $3,302 at 1995 prices. This figure appears to be high and this perception is correct as the mean has been pulled upwards by a small number of outliers (see Figure 3). The median figure for GDP per head in 1960 was approximately $1,500 and a third of the countries were found in a narrow ($500) range around the median. The two outliers, for whom the GDP figures are not reliable, are the US Virgin and the Netherlands Antilles. The very high figures (greater than $10,000) reflect the impact of oil refining rather than tourism and this was true also of Aruba ($9,484). However, the Bahamas also scores high ($8,171) showing the impact of high service exports on GDP per head as early as 1960. This is confirmed by the figure for the Cayman ($6,944), where service exports were still in their infancy. (Figure 3). The next richest countries were both French territories (French Guyana and Martinique). As these two countries both had low levels of exports per head, this might seem surprising. However, GDP in the DOMs is affected by the pattern of public transfers between mainland France and the Caribbean and these had begun to increase sharply at the end of the 1950s. There were also a small number of countries which were already lagging far behind the others. Haiti was particularly unsuccessful on almost any indicator used. However, no less than seven countries had a GDP per head below $1,000 in 1960 - less than one third of the mean level and two-thirds of the median. These other countries were Montserrat, Suriname, Guyana, Belize, Turks and Caicos and the Dominican Republic. This is a sufficiently heterogeneous group to suggest that low levels of GDP per head had no simple cause since the list included populous and large states as well as thinly populated and tiny countries. It also included English-, Dutch-, Spanish- and French-speaking countries. By 1960, therefore, there were clear signs that a number of countries were deviating sharply from the regional average. Among the successful countries, the explanations appeared to be found either in export diversification or metropolitan transfers. 10 I N T E G R A T I O N & T R A D E

Among the unsuccessful countries, the most common reason appeared to be the low level of exports per head. There was no perfect correlation, however, between these variables so that the pattern of GDP per head was hard to predict from any single variable. II. FROM 1960 TO THE PRESENT By the end of the 1950s, the traditional pattern of economic development in the Caribbean had begun to break up. Export diversification was well advanced in a small number of countries and exports of services were beginning to be highly significant in a small number of islands. Independence from the United Kingdom for the larger countries was on the horizon, although the short-lived experiment in Federation was about to come to an end. The Cuban Revolution was a further reminder that the status quo could not be taken for granted. The last four decades of the 20th century accentuated many of these trends. By the end of it there was a huge gap in economic and social terms between the most successful and the least successful countries. Indeed, by the end of the century the gap in income per head between the richest and the poorest country was a factor of 65 - far bigger than between the Caribbean as a whole and the developed countries. Even the gap between the second richest and the second poorest was a factor of 23 suggesting that both virtuous and vicious circles have been at work in the Caribbean in recent decades. POPULATION AND THE ENVIRONMENT The demographic expansion of the Caribbean peaked in the 1960s as in mainland Latin America. However, even at its peak the annual rate of growth was only two per cent. Thereafter, the rate of growth steadily declined, falling to one per cent in the 1990s for a long-run growth after 1960 of 1.6 per cent. Even this figure is misleading as the median rate of growth over the whole period (1960-98) was a mere 1.25 per cent. 12 Outward migration is largely responsible for this contrast with mainland Latin America rather than a difference in CBRs or CDRs. Migration had been a powerful force in the smaller countries even in the first half of the century, but the phenomenon had now become generalised. At least a million Cubans left the island after the Revolution and a similar proportion of the population migrated from Haiti, the Dominican Republic and Suriname. Taking the period as a whole, no less than nine countries had population growth of less than one per cent (see Table 4), although in one case (Barbados) this was mainly due to the low birth rate rather than migration. Migration could also be inwards. The high standard of living and wages paid in hard currency attracted migrants into some of the Caribbean countries. Three countries had population growth rates in excess of three per cent (Cayman, French Guyana and the US Virgin ) and there was inward migration into the Bahamas, the British Virgin and the Turks and Caicos as well. Finally, Belize had both: net outward migration (mainly English-speakers) in the 1960s and 1970s and net inward migration (mainly Spanish-speakers) in the 1980s and 1990s. The different demographic patterns in the Caribbean affected the pressure on land in a predictable fashion. Six countries began the period with abundant land reserves (defined as five hectares or more per person). The extreme case was French Guyana, where 33,000 people each had an average of 275 hectares. By the end of the period, the six had been reduced to four (Belize, Guyana, Suriname and French Guyana). I N T E G R A T I O N & T R A D E 11

Elsewhere the situation was a more familiar one of very limited land availability. Fifteen countries in 1960 had less than one hectare per person - more than half the 28 territories. By 1998 this number had risen to 21 - three-quarters of the total - and thirteen had less than 0.5 hectares per person. This might not matter where economic activity no longer depended on natural resources, but that was true of very few countries. Tourism, for example, is a natural resource-intensive activity even if financial services are not. EXPORTS OF GOODS AND SERVICES By the 1960s the domestic market in all Caribbean countries had increased, but only in a very few was it possible to embark on a programme of import-substituting industrialisation (ISI) as was taking place in mainland Latin America. The most important experiment was in Cuba, but this was driven by the US embargo which had raised the shadow price of foreign exchange to very high levels. Elsewhere, it was hard to justify ISI at the national level since the cost per unit tended to be so much higher than the unit price of imports (net of tariffs). ISI at the regional level, however, was another matter. Once the bitter legacy of Federation had been overcome the Caribbean Free Trade Area (CARIFTA) was launched and was converted into the more ambitious Caribbean Community (CARICOM) in 1973. Regional integration in the Caribbean, however, suffered from two main weaknesses. First, it was politically and constitutionally difficult to extend it to all 28 countries (even today CARICOM only embraces half the countries and several of these do not apply the common external tariff). Secondly, and perhaps in part because of the first reason, intra-regional trade accounts for a very small share of total trade in goods and an even smaller share of trade in goods and services. This has meant that exports of goods and services have continued to be dominated by sales to countries outside the Caribbean - mainly the US and the European Union. Exports of goods and services on a per capita basis have grown rapidly (see Table 5), rising from $350 in 1960 to $3,259 in 1980 and $5,111 in 1998. Although these figures reflect to some extent dollar inflation, the aggregate performance is still impressive. 13 What has not been so impressive has been the sub-regional distribution of the gains from trade. The high (unweighted) mean is to some extent explained by a small number of outliers at the top end. In particular, the oil-refining territories of the US Virgin, Aruba and the Netherlands Antilles distort the mean - especially so in the earlier years. When the median is used rather than the mean, we find a much lower figure for average exports per head (see Table 5). Indeed, the ratio of the median to the mean fluctuates between one quarter (1980) and a little over half (1970). Much of this is explained by the variation in world oil prices. Clearly, therefore, the export performance is affected by a small number of very successful countries. Thus, the coefficient of variation in 1960 was no less than 2.6 (see Table 5), reflecting the high standard deviation. In that year only six countries recorded a figure for exports per head that exceeded the (unweighted) mean. Among these the only non-oil exporters were Puerto Rico, where exports of manufactured goods were responding to fiscal incentives provided by the US federal authorities, and the Bahamas, which was the first to demonstrate the gains from exports of services. Ten years later (1970), Trinidad & Tobago had dropped out of this privileged list, but the Cayman and the British Virgin had joined it. Both were exporting services rather than goods, showing what could be done even by micro-states. 12 I N T E G R A T I O N & T R A D E

By 1990 this list of successful exporters (defined as those where exports per head exceeded the mean) had expanded to ten, this time with the addition of Antigua & Barbuda, Anguilla and French Guyana. All three were exporting services, although in the case of French Guyana this had little to do with international competitiveness and everything to do with strategic decisions taken by the metropolitan French government. By the end of the period (1998), the list had expanded to 11 (nearly 40 per cent of the territories) with the inclusion of Barbados - another exporter of services. The virtuous circle implied by these above average export per head figures must be set against the vicious circle of failure by several states. The most dismal has been Haiti, which has almost invariably occupied the last position in terms of the value of exports per head. If this poor performance had been matched by a dynamic internal market, it might not have mattered. However, this was far from the case. The other under-performers have not remained the same, suggesting that it is possible to break out of the vicious circle (and also unfortunately to break into it). In 1960 the second lowest figure was recorded by French Guyana, in 1970 by the Dominican Republic, in 1980 by Dominica, in 1990 by Guyana and in 1998 by Montserrat (where the volcanic eruption in 1995 had played havoc with the economy). The third lowest figure for exports per head has also rotated, being occupied by Cuba in 1998. The rapid growth of tourism had not yet compensated the island for the collapse of exports to the former Soviet Union. INSTITUTIONS AND HUMAN CAPITAL FORMATION While there is broad agreement about the role of trade in economic development in the Caribbean, there is much less consensus on the role played by institutions and human capital formation. While few would depart from the premise that the quality of institutions and human capital are important determinants of long-run growth, it is much more difficult to find proxies that capture accurately what one wants to measure. In the case of human capital, the trend in public expenditure on education per person is shown in Table 6. 14 This is a measure of inputs rather than outputs; it ignores private expenditure and it ignores the differential impact of expenditure on the three branches of educational spending: primary, secondary and tertiary. Nevertheless, it is all we have for the complete set of countries and, despite, its imperfections, it still contains much useful information. Table 6 shows, once again, the imbalance between countries this time in terms of spending on public education. The (unweighted) mean has been consistently higher than the median, pulled up by very high figures for a small number of countries. The wide spread in spending per head on education has been a feature of the whole period and the coefficient of variation (ratio of the standard deviation to the mean) reflects this. The highest expenditure per head in 1960 ($115) was in the US Virgin and the lowest ($1) was Haiti, where public spending on education has never been a priority. 15 The US Virgin was also the highest in 1970, but was temporarily relegated to third place in 1980 by Martinique and Guadeloupe. By 1990 the US Virgin had regained the first place, which it still occupies today. Public expenditure on education per person in Cuba has not been particularly high by comparison with other parts of the Caribbean. Even in 1980, when spending was not affected by the subsequent economic crisis, the Cuban figure was only equal to the median for the whole group of countries and therefore less than the mean. 16 The lowest figure by far has always been recorded by Haiti, followed by the Dominican Republic until 1998 when it was replaced by Guyana. Other low spenders have I N T E G R A T I O N & T R A D E 13

been Montserrat and Antigua. The dollar value of spending per head in Suriname also fell sharply in the 1990s, although it had exceeded the median in all previous decades. A proxy for the quality of institutions is even harder to find. One possibility is the stability of the nominal exchange rate on the grounds that ceteris paribus fixed exchange rates bring low inflation, which leads in turn to macroeconomic stability and secure public finances. Falling exchange rates, by contrast, in small open economies tend to be associated with high inflation, undermining public finances and weakening institutions. It is not difficult to criticise this proxy, but it is interesting to explore the picture that emerges. 17 If we consider the period since 1960, six countries have experienced a collapse of the nominal exchange rate; Guyana (after 1986), Jamaica (after 1983), Suriname (after 1993), Haiti (after 1990) and the Dominican Republic (after 1984) with Cuba experiencing a collapse of the black (later parallel) rate after 1990. None of these countries was able to achieve much by way of real exchange rate devaluation as a result of currency fluctuations so that the gains for the tradeable sector were largely illusory (with the possible exception of the Dominican Republic) while the costs were very real. Both Guyana (1971-86) and Jamaica (1968-83) managed controlled devaluations before collapse, but the only Caribbean country to have achieved a controlled depreciation on a permanent basis has been Trinidad & Tobago (since 1984). No doubt oil exports have been part of the reason why depreciation did not ultimately lead to currency collapse as in the other countries. The other 21 countries stand out for the stability of the exchange rate. They include the Eastern Caribbean states where the nominal exchange rate for all eight states is underpinned by the regional central bank: five countries where the US dollar circulates freely (Bahamas, British and US Virgin, Turks & Caicos and Puerto Rico): three countries where the local currency is pegged to the US dollar (Barbados, Belize and the Cayman ): two where it is pegged to the Dutch guilder (Netherlands Antilles and Aruba): and three where the French franc is legal tender (Martinique, Guadeloupe and French Guyana). These 21 countries have had stable rates for a very long time. Furthermore, most of the other countries have taken steps to stabilise their nominal rates in recent years. Thus, fluctuations in the nominal exchange rate are not viewed with much sympathy in the Caribbean despite their widespread use in much of mainland Latin America. Stable nominal exchange rates has not eliminated inflation in the Caribbean, but the inflation rate has tended to track consumer prices in the main trading partners (the United States and the European Union). During the mid-1970s, inflation surged as a consequence of imported inflation combined with the impact of the commodity boom on the domestic money supply. In the 1990s, however, inflation has been reduced to international levels in all countries except those with depreciating exchange rates. Only Guyana (1991) and Suriname (1993-5) experienced three-digit annual inflation. 18 GROSS DOMESTIC PRODUCT (GDP) Although the early estimates are less reliable than those for the later years, a series for GDP per head in US dollars (at 1995 prices) for all 28 countries has been constructed covering the years from 1960 to 1998. The picture that emerges is once again very different depending on whether one uses the aggregate figures or looks at the individual countries. 19 The weighted mean for GDP per head is shown in Figure 4. It shows a steadily rising trend and a doubling in value over the four decades. However, the openness of the 14 I N T E G R A T I O N & T R A D E

Caribbean economies renders them vulnerable to external shocks. The 1970s oil shock caused GDP per head to stagnate for a number of years, while the debt crisis at the start of the 1980s led to a small fall in GDP per head. However, the greatest drop in GDP per head came at the beginning of the 1990s. This is heavily influenced by the sharp drop in GDP in Cuba after the collapse of the former Soviet Union. (Figure 4) Using the unweighted mean (see Table 7), the figure for GDP per head in 1960 was much higher ($3,302). This suggests a highly skewed distribution and this is confirmed by the median income per head of $1,479, which was a mere 45 per cent of the unweighted mean. The skewness is due as we have already seen to a small number of outliers with very high values of GDP per head. Once again, as for exports of goods and services per head, this draws our attention to the skewed distribution. Thus, in 1960 the gap between the poorest country (Montserrat) and the richest (US Virgin ) was a factor of 26, while even the gap between the second poorest (Haiti) and the second richest (Netherlands Antilles) was 21. By 1960 nine countries had an income per head above the unweighted mean. These countries had already broken away from the traditional pattern of development and they were all to remain among the elite group in the next four decades. They included the three DOMs, the two Dutch dependencies, the virtual US dependencies (Puerto Rico and the US Virgin ) and two British colonies (Bahamas and the Cayman ). During the last four decades, the unweighted mean rose from $3,302 to $7,666, an annual average growth of 2.24 per cent. This impressive figure was dwarfed by the growth in median income from $1,479 to $6,827 - an annual average growth rate of 4.11 per cent. This raised the median as a percentage of the mean to nearly 90 per cent, implying a narrowing of the skewness of the distribution from what it had been in 1960. This was indeed the case. Furthermore, the coefficient of variation was also falling steadily reaching 0.77 by 1990 (the same as in 1998) compared with 1.06 in 1960. The example of the earlier successes was followed by others so that by 1998 13 countries - almost half the total - had an income per head above the unweighted mean. In addition to the nine countries that enjoyed this privileged status in 1960, the list now included Barbados, Antigua, Anguilla and the British Virgin. All these countries had invested heavily in the export of services. Despite the example offered by these relatively successful countries, a handful of states failed to break out of the vicious circle. Haiti is the best known case with GDP per head falling by 1998 to 72 per cent of what it had been in 1960 (even then a very low figure). However, the seven countries in 1960 with income per head less than $1,000 (Montserrat, Haiti, Suriname, Guyana, Belize, Turks & Caicos and the Dominican Republic) had been reduced to three by 1998 (Haiti, Guyana and Suriname). Five countries achieved annual growth in GDP per head above four per cent (St. Kitts, Anguilla, British Virgin, Turks & Caicos, and Montserrat), including some of the poorest states in 1960. This suggests that catch-up is a real possibility in the Caribbean. By contrast, some of the slowest growing countries (e.g. the Netherland Antilles and the US Virgin ) were those that had already achieved by 1960 a high standard of living. Nearly 20 countries managed growth in excess of two per cent per year and only Haiti was unable to avoid negative growth in GDP per head. III. CONCLUSIONS The outstanding feature of the wider Caribbean is the small size of the average unit. At the beginning of the 20 th century, the average size of population was a mere I N T E G R A T I O N & T R A D E 15

250,000; even by the end of the century it was not much greater than one million. Thus, it is not surprising that the wider Caribbean has experienced a pattern of development in the last 100 years that in a number of respects distinguishes the region from other parts of the developing world. With its geographical proximity to North America and colonial ties to Europe, migration has played an important part in the demographic patterns of the last 100 years. After the 1995 volcanic eruption, for example, over half the population of Montserrat moved to the United Kingdom. Outward migration has also brought inward remittances. These have not been studied here, but they constitute an increasingly important part of foreign exchange receipts in most Caribbean countries. The small size of the domestic market in almost all countries has placed a special burden on the export sector. The ratio of exports to GDP has always been very high in a number of countries. For the region as a whole, in the mid-1990s, it exceeded 50 per cent -very high by the standards of developing countries in general and Latin America in particular. In the first half of the 20 th century, exports consisted mainly of primary commodities and these were dominated by sugar. This changed dramatically in the second half of the century with the growth of service exports. In many cases, service exports are now more important than commodity exports and this is no longer true only of the smaller countries. Tourism in Cuba, for example, has overtaken all other commodity exports in gross value and will soon overtake sugar in terms of net value. 20 The variance of economic performance appears to have widened significantly in the last 100 years. The generalised poverty in 1900 has given way to a much more complicated picture in which some countries have reached levels of GDP per head equal to those of the Organisation for Economic Co-operation and Development (OECD) average. At the other end of the spectrum are a small number of countries where GDP per head has stagnated or even fallen. This pattern is too complicated to be explained by any single variable, but there does appear to be a high correlation between GDP per head and exports per head (see the paper by Shelton Nicholls in this collection). Other seemingly relevant variables, such as the management of environmental resources, do not appear to be closely linked with movements in GDP per head (see the paper by Elizabeth Thomas-Hope in this collection). The case studies in this collection are designed to shed light on why some countries have been more successful than others in increasing exports. What emerges is very complicated, but several general themes are apparent. The quality of institutions and the stability of the macroeconomic framework is of major importance, but this is not necessarily linked to constitutional status. Some independent countries, such as Barbados, score highly in this regard as do some colonies, such as the Cayman. Investment in human and social capital is more problematic. In some cases, such as Suriname, it appears to have led to a rise in emigration rather than an increase in productivity. The increases in exports and GDP per head are evidence of the ability of at least some Caribbean countries to shape their own destiny. Yet nothing can be taken for granted. In the next few years, the privileged access to the European Union for many traditional exports will be lost while the future of offshore financial services is now in doubt. Tourism is also under threat from domestic violence on the one hand and environmental disasters on the other. Maintaining the dynamism of the export sector even in the hitherto successful countries will not be easy. 16 I N T E G R A T I O N & T R A D E

Figure 1 POPULATION OF THE WIDER CARIBBEAN, 1900-1960 2500 2000 1500 1000 500 0 1900 1910 1920 1930 1940 1950 1960 Figure 2 EXPORTS OF GOODS PER HEAD FOR THE WIDER CARIBBEAN, 1900-1960 (in US Dollars) 140 120 100 80 60 40 20 0 1900 1910 1920 1930 1940 1950 1960 I N T E G R A T I O N & T R A D E 17

Table 1 EXPORTS PER HEAD, 1900-1960 (in US dollars) Y EAR U NWEIGHTE D M EAN ($) S D TANDARD EVIATION CV M DIAN E HIGHEST LOWEST 1900 21. 6 12. 3 0.57 23. 5 46. 8 2. 2 1910 28. 5 18. 2 0.63 22. 3 68. 6 9. 2 1920 72. 2 69. 3 0.96 62. 8 269. 6 1. 8 1930 138. 7 484. 3 3.49 23. 8 2194. 0 5. 9 1940 74. 0 253. 9 3.43 21. 7 1237. 5 1. 9 1950 135. 9 342. 2 2.52 46. 2 1677. 5 12. 0 1960 209. 1 418. 9 2. 0 75. 0 1741. 7 9. 8 Source: 1900-50 derived from Table A.4 in the statistical appendix and refers goods; 1960 from Table A.6 and refers to exports of goods and services. to exports of Table 2 PUBLIC REVENUE PER HEAD, 1900-1960 (in US dollars) Y EAR U NWEIGHTE D M EAN ($) S D TANDARD EVIATION CV M DIAN E HIGHEST LOWEST 1900 7. 0 2. 9 0.42 6.99 12. 6 2. 7 1910 8. 6 5. 7 0.67 6. 6 24. 3 2. 4 1920 13. 6 9. 8 0.72 10. 6 36. 7 3. 2 1930 12. 6 8. 6 0.69 9. 6 37. 4 3. 6 1940 13. 1 8. 0 0.61 11. 7 30. 1 1. 9 1950 31. 7 14. 9 0.47 29. 9 56. 1 7. 0 1960 69. 6 21. 5 0.31 59. 9 106. 6 46. 9 Source: derived from Table A.5 in the statistical appendix. 18 I N T E G R A T I O N & T R A D E

Table 3 STRUCTURE OF THE LABOUR FORCE AND GDP PER HEAD C. 1945 (POUND STERLING) IN BRITISH COLONIES YEAR LABOUR FORCE AGRICULTURE % TRADE % DOMESTICS % MANUFACTURE % PU BLIC/ PROFESSIONAL % FORESTRY NATIONAL INCOME PER HEAD ( POUN D - YE AR) Bahamas 1943 34,762 39. 4 25. 8 18. 1 4. 6 Barbados 1946 91,369 27. 6 13. 2 15. 6 20 6. 1 30 (1942) Guyana 1940 93,523 40. 1 6 7. 9 14. 6 11. 1 29 (1942) Belize 1946 20,133 29. 1 8. 5 5. 5 12. 7 11. 1 10. 1 Jamaica 1943 505,092 43. 8 7. 8 14. 1 12 4 0. 5 27 (1942) Cayman Island Turks and Caicos British Virgin 1943 2,531 11. 4 4. 9 18. 8 17. 4 2. 7 1943 2,689 38. 2 4. 1 16. 1 7. 6 1. 7 1946 1,994 51. 2 1. 7 5. 6 9. 8 3. 6 16 (1946) Antigua 1946 18,605 42. 3 5. 8 14. 6 11. 8 5. 5 St. Kitts 1943 20,809 51. 8 5. 5 10. 9 8. 9 4. 0 Montserrat 1946 6,617 59. 9 3. 5 8. 6 9. 3 3. 4 Trinidad 1946 213,093 25. 3 8. 8 10. 6 17. 8 3. 6 1. 6 Dominica 1948 21,300 54. 0 4. 5 8. 5 11. 7 4. 5 1. 4 Grenada 1946 27,606 45. 0 8. 3 14. 8 14. 4 NA 0. 5 24 (1942) St. Lucia 1946 31,891 50. 2 4. 1 5. 0 14. 9 5. 7 0. 7 14 (1942) St. Vincent 1946 22,691 49. 8 6. 1 7. 9 10. 7 5. 8 1. 5 14 (1942) Source: derived from British Colonial Reports. I N T E G R A T I O N & T R A D E 19

Figure 3 GDP PER HEAD AT CONSTANT PRICES (1995 dollar) 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 Table 4 POPULATION GROWTH, 1960-1998 A NNUAL RATE OF GROWTH % C OUNTRIES < 0.5 Barbados; Grenada; St. Kitts; Montserrat 0.5 1.0 Suriname; St. Vincent; Dominica; Antigua; Martinique 1.0 1.5 Trinidad & Tobago; Guyana; Jamaica; Antilles; Aruba; Puerto Rico; Cuba Guadeloupe; Netherland 1.5 2.0 St. Lucia; Haiti 2.0 2.5 Anguilla 2.5 3.0 Bahamas; Belize; British Dominican Republic Virgin ; Turks & Caicos ; > 3.0 Cayman ; French Guyana; US Virgin Source: derived from Table A.2 in the statistical appendix. 20 I N T E G R A T I O N & T R A D E

Table 5 EXPORTS OF GOODS AND SERVICES PER HEAD, 1960-1998 (US dollars) U M NWEIGHTED EAN SD CV MEDIAN M M E AN/ EDIAN H IGHEST LOWEST 1960 350 911 2.60 133 2.62 4,902 14 1970 542 798 1.47 314 1.73 4,159 12 1980 3,259 8,298 2.55 825 3.95 44,485 59 1990 4,306 5,536 1.29 1,894 2.27 27,070 74 1998 5,111 6,449 1.26 2,329 2.19 26,834 58 Source: Derived from Table A.6 in the statistical appendix. Table 6 PUBLIC EXPENDITURE ON EDUCATION PER HEAD, 1960-1998 (US dollars) M EAN SD CV MEDIAN M M E AN/ EDIAN H IGHEST LOWEST 1960 22. 8 29. 4 1.29 8. 0 2.85 115 1 1970 57. 3 54. 2 0.95 36. 0 1.59 227 1 1980 185. 9 190. 1 1.02 108. 0 1.72 676 4 1990 348. 5 324. 0 0.93 195. 0 1.79 1,209 7 1998 712. 3 1449 2.03 209. 0 3.41 7,917 6 Source: Derived from Table A.7 in the statistical appendix. I N T E G R A T I O N & T R A D E 21

Figure 4 GDP PER HEAD (WEIGHTED), 1960-1998 3500 3000 2500 2000 1500 1000 500 0 Table 7 GDP PER HEAD, 1960-1998 (US dollars) M EAN SD CV MEDIAN M M E AN/ EDIAN H IGHEST LOWEST 1960 3,302 3,496 1.06 1,479 2.23 13,778 514 1970 4,157 3,891 0.94 2,048 2.03 15,109 471 1980 5,004 3,957 0.79 3,278 1.53 12,924 607 1990 6,898 5,342 0.77 5,066 1.36 21,123 481 1998 7,666 5,939 0.77 6,827 1.12 23,966 370 Source: Derived from Table A.10 in the statistical appendix. 22 I N T E G R A T I O N & T R A D E

Notes 1 Funding for the data base was provided by the Nuffield Foundation in the United Kingdom and the SSRC. 2 These are also available on the website of the Social Science Research Council (SSRC) in New York (www.ssrc.org). 3 There is a good discussion of the impact of these diseases on the wider Caribbean in Jones [1931], Chapter 3. 4 By way of illustration, the CDR (expressed per thousand) in Trinidad & Tobago fell from 25 in 1900 to 19 in 1930 and 8 in 1960. 5 The CBR (also expressed per thousand) was not of course constant, but there was much less sign of a long-term trend. Again, using Trinidad & Tobago by way of contrast, we may note that the CBR was 37 in 1900, 31 in 1930 and 38 in 1960. 6 The population data for this period (1900-1960) for all countries is given in Table A.1 in the statistical appendix. 7 It was 12,217 in 1900 and 12,000 in 1960. 8 One of these was the father of Fidel Castro, who arrived from Spain in the first decade of the 20th century. 9 The close correlation between GDP per head and exports per head for the period since 1960 is demonstrated by Shelton Nicholls in his paper in this collection. 10 The mean is calculated as the sum of exports for all countries divided by the sum of population. 11 The CV measures the ratio of the standard deviation to the (unweighted) mean and therefore can be compared easily over time. 12 The detailed population figures for this period can be found in Table A.2 in the statistical appendix. 13 The detailed figures for exports per head can be found in Table A.6 in the statistical appendix. 14 The detailed figures for public spending on education per head can be found in Table A.7 in the statistical appendix. 15 The proportion of children enrolled in Haiti in 1900 has been estimated at 2% (see Miller [1992], Chapter 7). It was still only 36.5% in 1980. 16 The Cuban wage system for public sector workers is so different from the rest of the Caribbean that too much should not be read into this surprising result. I N T E G R A T I O N & T R A D E 23