Paper for History & Economic Development Group annual workshop,

LSE 7 May 2002

THE LIMITS TO GROWTH IN A PEASANT COLONY ? AFRICAN RURAL CAPITALISM AND COCOA FARMING IN COLONIAL GHANA

Gareth Austin

Department of Economic History, LSE (g.m.austin@lse.ac.uk)

This paper reconsiders the micro-foundations of the export economy of colonial Ghana in the framework of the debate about the significance of peasant colonies for the general issue of whether colonial rule promoted or ultimately retarded capitalist economic development in Africa and in what used to be called the Third World more generally.

The framework: peasant economies within the debate about colonial rule as a pioneer of capitalist development in the Third World

My starting-point is two classic debates and one thought-provoking recent paper. The first debate is the general argument over whether colonial rule was an agent or brake on capitalist development within the colonies: whether, as Karl Marx maintained with reference to British India, colonial rule by an industrial capitalist power was the cruel but progressive pioneer of capitalist development in the non-Western world; or whether, as Dependency writers have argued, colonial rule permitted only the emergence of a half-baked and self-limiting peripheral capitalism lacking its own circuit of accumulation and therefore doomed to self-perpetuating stagnation accompanied by dependence on the Western-dominated capitalist world economy. The second debate concerns a specifically West African edition of the general issue: whether what used to be called the British West Africa Lands Policy protected or preempted the possibility of sustained economic development. That is, whether the policy of supporting African ownership of land, which the British broadly followed in their West African colonies - in contrast to the settler and plantation economies that were so prominent in other regions of colonial Sub-Saharan Africa - was a curse in disguise. Did it avert the alienation of land to European planters and settlers at the cost of perpetuating economically-inefficient land tenure systems and units of production?

The recent paper is by Ralph Austen: presented here last year. My references to Austen s paper are entirely to that earlier formulation; which I think is particularly welcome because it offers an Africanist input into the emerging literature on global economic history , which has so far been overwhelmingly dominated by discussions of comparisons and connections between East Asia and the West (especially the North Atlantic fringes). Austen draws a broad distinction between the earlier and later phases of British (and French) rule in South Asia, Africa and the Caribbean. He argues that in the earlier phase the colonialists were primarily traders and planters, hence their aims were narrowly economic and the export production systems which they created or presided over were highly market-oriented. In the later phase there was much greater reliance on peasant households as the unit of export production. In a sense, as he notes, this seems paradoxical: to make much greater use of a mode of production in which, by definition, factor inputs are drawn at least partly from outside the market appears at odds with any assumption that capitalist development on a world scale involves progressively more extensive domination by market mechanisms. Indeed, one might add that it is during Austen s later colonial phase that, according to Williamson and O Rourke, factor markets finally began to integrate across the world, including in the Third World ; thus, for example, real wage rates in geographically diverse markets moved closer together. In this context it could be that Austen s analysis points us to a reconciliation of the debate about the economically-progressive or self-limiting impact of colonial rule: perhaps the answer is period-specific, with colonialism in the tropics ceasing to be a motor of market expansion and market integration once the industrial revolution actually got under way in Britain.

Then again, perhaps the retreat to the peasantry was no retreat at all. Earlier colonialism may have been highly commercial in its aims, but because it was based to a very large extent on slave plantations, an indispensable part of its micro-foundations was an extra-economic force, namely coercion. Of the three regions which Austen discusses, the Caribbean was dominated (from the seventeenth century) by slave-holding; West Africa featured in that period as a supplier of slaves; while the exception was India. This exception is significant for the issue, because though South Asian production for western markets was not from slave plantations, its social relations were not more market-oriented for that: given the importance of family and caste relations. Conversely, the prominence of peasants in nineteenth and twentieth-century colonial economies could be interpreted as evidence not of a dilution of the capitalist dynamic behind European imperialism, but as a natural - perhaps inevitable - consequence of the extension of that expansionism into territories in which there were, and remained, very large indigenous populations. As Austen emphasises, the acquisition of formal territorial sovereignty over such vast populations was indeed characteristic of later colonialism, as far as Britain and France were concerned: thus he dates later colonialism from the beginning of British rule over Bengal. It is surely not surprising that once colonial rulers were presiding over vast subject populations they tended to give priority to the preservation of their own rule. Hence the colonial bureaucracies that emerged were not only thinly-resourced relative to the size of the populations they sought to control, but were much less focussed on commercial ends that their predecessors. Yet, ultimately, the peasant economies too saw a shift towards market relations, notably in the form of the emergence of widespread wage labour on African-owned cash-crop farms; the workers concerned migrating seasonally from home areas where transport costs and/or soil conditions made production for export uneconomic. For instance, during the colonial period the main source of hired labour in the cocoa farms of the forest zone of southern Ghana was male workers migrating seasonally from the savanna to the north.

In this framework two major questions arise. First, when and where later colonial administrations promoted or endorsed peasant production as the basis of the local export economy, did they do so for political reasons - most obviously, to reduce discontent - or because they believed that peasants were the most efficient producers available? The second question is whether the peasant path of economic growth, even where it delivered impressive results over some decades, ultimately set limits to growth. For purposes of this discussion I use the word peasant in the sense most widely used in the peasant studies literature: a family-based, part-subsistence, part-surplus producing farmer. That is, a farmer (or member of a farming household) producing partly for household consumption and partly for outsiders (in this context, for sale), using at least some family labour though possibly supplementing this with additional labour. Conversely, by a capitalist farmer I mean someone whose farm labour force is mainly obtained through the market. In the rest of the paper I will not offer general answers to these questions; but I hope to illuminate them by reflecting on the classic case of relatively successful agricultural export production by African farmers in colonial Africa, namely Ghana.

The cocoa farmers of colonial Ghana: peasants or rural capitalists ?

In Ghana, as is well documented in the literature, cocoa was adopted on a wide scale soon after the beginnings of colonial rule, but largely as a result of African initiative. The British declared political authority over the territory to the south of Asante (later known as the Gold Coast Colony) in 1874; regular cocoa exports began in 1891. Further inland, the kingdom of Asante was put under a form of colonial occupation in 1896; widespread cocoa-planting began there in the years immediately following the suppression of the Yaa Asantewaa revolt in 1900. In 1911 Ghana overtook Brazil as the world s largest exporter of cocoa beans, with 40,000 tons in the year. In 1923 annual output reached and passed the 200,000 ton mark.

Anyone who has read Polly Hill s classic study of the take-off of cocoa farming in the Gold Coast Colony will recall that she powerfully denounced the assumption that this signal economic achievement was the work of peasants . She maintained that the original take-off was the mainly the work of particular groups of cocoa farmers, who she described as rural capitalists . It is essential to note that she characterises capitalist in terms of behaviour and attitude. Thus for her capitalist essentially means entrepreneurial: taking risks, investing for the long term, and dedicated to the continued expansion of the business. Cocoa was a new, imported, exotic cultigen; once planted it did not begin to yield for several years; and in expanding their production, the Akwapim and Krobo pioneers had to migrate to buy suitable land. For Hill this migratory behaviour was proof of entrepreneurship; or, as she put it, of capitalism . There is a further level to Hill s discomfort with the appellation peasant . She identified other groups of Ghanaian cocoa-farmers as non-capitalist. In particular, she characterised the Akyem farmers, whose chieftaincies originally owned the land on which the migrants planted in the 1890s and after, as sedentary rather than migrant farmers. But she seemed reluctant to call even them peasants , because her general position - most fully developed in her later work- is that it is best to avoid the word altogether. Thus, if I understand her correctly, her preferred glossary has capitalists as the only legitimately-definable sub-set with the general range of farmers .

As you have seen, my definition of capitalist is different, being based not on entrepreneurial attitudes and behaviour but on relations to the markets for output and for inputs such as labour. But using either definition Hill s general point, that it was not peasants who were primarily responsible for Ghana s cocoa take-off, and who accounted for the bulk of output in the rest of the colonial period, is correct. Rather, most output came from (usually male) producers who supplemented the labour of themselves and family members (usually wives and sons) with the regular or permanent labour of others: wage workers or, in the early years of the Asante cocoa industry, people who had been brought into the household via lending or purchase, i.e. as pawns or slaves or who were the descendants of slaves.

I would want to go further than Hill in arguing that there was a continuum of farmers , in that I am not convinced that she was right to contrast migrant capitalists with sedentary farmers. From my own primary research in Asante, whether farmers planted only on land belonging to the chief they served , or whether they planted on neighbouring chiefs lands, or whether they migrated - i.e. travelled - to establish farms much further afield was largely a question of opportunity and cost. Many farmers did all three, moving through the stages as their careers developed and as fresh land became scarce in the earliest cocoa-growing centres.

Hill was evidently much struck by the fact that the Akyem farmers, despite being the indigenous land-holders, did not seize the opportunity to plant their own lands to the maximum of their resources. But this needs to be seen in the context of Akyem choices of economic specialisation before cocoa, and of contemporary European perceptions of those choices. In the precolonial nineteenth century Akyem producers grew food crops only for subsistence, but were extremely active in extra-subsistence production: gold-mining, the production and trading of kola nuts and, when the rubber market emerged in the 1880s, of rubber too. In 1897 a British commissioner, H. M. Hull, wrote I have been pretty well all over Akim . He proceeded to assert that Akim people (men?) are not energetic ... prefer sitting to standing, and doing nothing to work . Yet the sentence continues: they are keen as petty traders and of course collect large quantities of the Kola nut which abounds in their forests . What prompted these remarks was that Hull was now visiting Sefwi, in western Ghana, where he found large numbers of Akim men tapping rubber.

it is a quite ordinary occurrence to find Akim towns where all the able-bodied men are absent in this pursuit. Arrived at the site selected, they hire a stretch of land for usually about 5 months. When the time is up they take their produce to the Coast, sell it, buy goods and return to sell these and get more rubber.

Thus Hull s perception of Akims as economically unenergetic is contradicted from his own pen (and he conceded that Gathering rubber is fairly hard work ). Like some of the Basel missionaries of his time, this British official seems to have equated energy in this context with what he regarded as honest agriculture, as distinct from specifically money-making forms of work. Hull s account documents Akims migrating seasonally many miles to hire land, and reinvesting the profits in the continued expansion of his business (to borrow the phrase Hill precisely did not apply to the Akim). Hull was reporting during exactly the time when the Akwapim and Krobo farmers were buying and planting Akim land. Comparing the migrant cocoa-farmers with the migrant rubber-tappers, the latter may have misjudged future profit opportunities: but in no basic structural or attitudinal sense do they seem to be less capitalist than the farmers rightly celebrated by Hill.

It may be added that this pattern of behaviour by Akyems was parallelled, in important respects, earlier in the century, and was thus no innovation of the era of colonization. In 1839 the Basel missionary Andreas Riis visited the Kibi, the Akyem Abuakwa capital, to find it emptied of its usual population, who had gone to the gold pits . We will return to this observation shortly.

Thus I am sceptical of the notion of a fundamental cultural difference between capitalist and peasant rural producers in southern Ghana. Again, the evidence does not show a qualitative divide between capitalist and peasants if the criterion is the use of extra-familial labour. Even small cocoa-farmers tended to hire causal labour to help with particularly heavy tasks; slightly larger ones (after the demise of slavery and male pawning) hired one or two regular labourers; still larger ones hired more. At the end of the colonial period, in 1956-7, the first relatively comprehensive survey of cocoa-farming families in Asante (i.e. what are now the Ashanti and Brong-Ahafo regions) found that the ratio of hired labourers (wage and sharecrop combined) to farmowners was nearly two to one. The same source indicates that the farm-size threshold at which farmers hired one regular, full-time labourer was when it was producing about 26 loads a year: implying about three acres of bearing cocoa, though the farmers would, on average, also have had some immature cocoa acreage too. Thus even pretty small farmers, definitely in a size band that one would intuitively associate with peasants, relied to a significant extent on hiring labour. Larger farmers, in a size band averaging 152 loads a year, employed on average 3.67 labourers (wage or sharecrop) each. Only a small minority of farmers seem to have had double-figure numbers of regular employees. Thus most of the farmers who were capitalist by the criterion of obtaining the majority of their labour inputs from the market, were small capitalists. More precisely, at any one time farmers, mainly according to the size of their enterprises, variously occupied different points on a range that straddled the categories of family-labour based and wage-labour based enterprise.

In this context, referring back to the general framework of this paper, we should distinguish the issue of the economic efficiency of peasants from the issue of the economic efficiency of the West African Lands Policy , i.e. of agriculture being maintained in the hands of indigenous farmers.

The origins of the West African Lands Policy : the Ghana case

So was it for political reasons - to reduce discontent - that the colonial administration in what is now Ghana accepted and endorsed the status quo, in the sense of eventually confirming African ownership of virtually all the cultivable land, or was because it came to believe that this was economically optimal in any case? I have addressed this, from different angles, in a 1996 chapter and in part of a current book manuscript. Given that time and space are short, here I will largely assert my view in summary form, inviting questions if time allows.

It is established that the West African Lands Policy was not decided at the start of colonial rule in what is now Ghana, even if was in Nigeria. Indeed, in the Gold Coast Colony and in Asante the colonial government initially permitted quite large-scale alienation of agricultural land to Europeans. As it turned out, only two European plantations were established in Asante, and both failed in competition with African farmers, just as happened with the rather larger number of European plantations in the Gold Coast Colony. It was the success of indigenous cocoa producers that enabled the pro-peasant lobby within the colonial administration, led by Clifford, to defeat their opponents. Two further points should be added.

One is that this colonial-era expansion was based in part on the re-investment of assets accumulated in the preceding decades. The precolonial nineteenth-century had already seen the emergence (or in some areas, such as among the Krobo, the consolidation) of extra-subsistence production by a multitude of rural producers. Both in Asante and further south much of this involved the recruitment of labourers from outside the family (pawns and slaves), who were acquired for money (whether as loan or payment), albeit in markets based upon the coercive assertion of property in people. My work on this has focussed on Asante, so it may be appropriate here to give a brief example from further south. Riis, quoted by Marion Johnson, indicated that Akyem household heads also invested in the acquisition of labour (slaves) to expand their gold operations. In Akyem the livelihood of the people, according to Riis, was:

working the land, hunting and gold-mining. Which brings in most I could not say with certainty; one seems as necessary to the people for survival as the other. But goldmining must be the most profitable as it is only practised once each year . . . For this work each householder not only uses his whole family, but also often buys as many slaves as he has money for; they pay 30 thaler, or sometimes more, for them . . .

The general implication is that if there was a process of peasantization - of moving from purely subsistence to part-subsistence, part-surplus production - in southern Ghanaian history, it had already happened before colonial rule.

The other point is that, as Phillips has shown for the interwar period, even after the notion of a colonial economy based on European plantations had been abandoned within the colonial administration, there were articulate voices within that administration arguing for the introduction of compulsory land titling. This would have amounted to an attempt to shift the balance among African cocoa farmers in the direction of fewer peasants and more capitalists. This policy debate continued to the eve of the process of formal transfer of power to Ghanaian politicians; without the proposal ever being implemented. I will comment on this further below.

The limits to economic growth in a peasant (and rural capitalist) cocoa-exporting economy

Let us now ask whether the dominance of relatively small producers in Ghanaian export agriculture constrained growth: if not over a few decades - given the very fast expansion of cocoa output in the first thirty years of regular cocoa exports - then by setting limits to subsequent growth. It is necessary to preface this by briefly assessing the record.

In terms of the volume of investment and output, let me underline what Ghanaian cocoa farmers achieved during the colonial period as a whole, and in its immediate aftermath. As we have seen, the early growth was remarkable. True, the colonial government facilitated the remarkable scale of the early growth, primarily by bringing about railway construction, and later contributed technical skills in the making of motor roads. But the important point is that the opportunities provided by the attractive producer prices of the early decades were seized by Ghanaian farmers rather than by European plantations: as is too often overlooked, a number of European planters were established, but none were commercially successful. Also commonly overlooked is the fact that there was a second massive wave of cocoa planting. The growth of output tailed off in lagged response to the lower prices of the interwar period, peaking at 311,000 tons in 1936. But in the middle and late 1950s, responding to the then boom in world commodity prices, there was extensive new investment, which created the capacity that made possible a new peak and world record of 572,000 tons in 1964-65.

The record on productivity, however, is not so remarkable. The indications are that yields remained roughly static during the period in relation to inputs of labour, of land, and of capital. It was only in the last few years before independence that useful innovations such as higher-yielding varieties and effective insecticides became available, and they do not seem to have made a significant difference by 1957. Rather, what made possible the sustained expansion of output was the continued availability, somewhere in the forest zone, of land which had not been permanently cultivated before. Here let me introduce the very important work of François Ruf, who has coined the term forest rent to describe the difference between the cost of planting cocoa on fresh land and the cost of re-planting it on land previously under cocoa. His empirical proposition, which I think bears up well in the comparative literature, is that cocoa world-wide has so far only been really successful (in terms of scale and growth of output) when is planted in fresh land. This brings us to the issue of the longer-term growth potential of a cocoa economy.

Viewed in the perspective of staple theory , subsequent growth would depend greatly on the number and nature of the linkages between cocoa and the rest of the cocoa: linkages that could facilitate the emergence of new, higher value-added, forms of production. If such links are weak, the result would be a staple trap : in which the economy was unable to improve on cocoa as the main product in which it has a comparative advantage in the world market. The need to avoid this is made worse in the Ghanaian cocoa case by two things. One is population growth, which puts downward pressure on cocoa earnings per head; the other is the progressive exhaustion of the forest rent. One positive qualification to this is that it seems likely that in the twenty-first century it will become profitable to replant cocoa on a large scale somewhere in the world, simply because suppliers of fresh land are diminishing fast world-wide. That would require much higher inputs of capital and labour than has been involved in Ghanaian cocoa production so far: but the availability of higher-yielding, early-maturing varieties makes this not inconceivable.

Diversification from cocoa was tried by Kwame Nkrumah s government, and to a lesser extent by some of his successors. The era of statist development policies, characterised by increasing substitution of administrative for market mechanisms, failed almost completely to achieve sustainable diversification, while more than half-throttling the goose that had laid the golden eggs. That is to say, there was negligible per capita income growth over the period 1951-83, and cocoa output declined after 1964-65 to 180,000 tons (plus a very few tens of thousands of tons smuggled out) in 1982-83.

For this paper, the question is whether the peasant and small capitalist pattern of cocoa-farming was in itself a trap, reducing the chances of reviving and strengthening cocoa production in a context in which land was beginning to be scarce, and inhibiting diversification around cocoa. If it was, in principle the constraints could take various forms. They might arise from the nature of the cocoa-farming production units or enterprises themselves: these could be cultural (as in the notion of a distinct peasant economic ethic) or structural (stemming from the characteristics of the family labour farm). Alternatively, they might reside in the collective consequences of individual ownership and behaviour: most obviously, with perhaps 300,000 cocoa farmers in 1937, and even more by the mid-1950s, the distribution of cocoa-farming income, unequal though all surveys show that it was, was much less uneven than in a plantation or settler economy. In theory, that could have a macroeconomic downside, in the form of low savings rates: on the assumption that richer people save a higher proportion of their income than poorer people. Again, it is conceivable that economic growth would be constrained, not necessarily by any of the above, but by the institutions - the rules, formal and informal - surrounding the wide distribution of cocoa farms. If compulsory registration of farms - or of the land they stood on - had been introduced, to take perhaps the most obvious example, it should have lowered the risks and transactions costs of buying and selling farms.

Each of these issues could use an essay in itself, but let me be brief. I have argued above that it is hard to sustain the notion of a fundamental difference in attitude or behaviour between peasants and capitalists among Ghanaian cocoa farmers: rather, what I see is a continuum, along which individuals moved according to the stage of their farming career and according to the opportunities available to them at any one time and place. The difference between a family-labour enterprise, and one based largely on wage-labour, matters in terms of the size and structure of labour costs (as in Millar s neo-classical version of Chayanov s theory). But it does not matter nearly as much when there is a labour market, such that wage work is an alternative to family labour. This was precisely the case in the Ghanaian cocoa industry, from early on. This meant that family labour inputs had an implicit cost, an opportunity cost, because they meant foregoing the opportunity to earn wages. Thus the structure of labour costs in family-labour farms in Ghana was not dissimilar to that of capitalist enterprises. Thus I doubt that differences in the nature of the peasant enterprise, as opposed to capitalist enterprise, were a brake on economic growth.

Again, on the question of savings rates, there is strong evidence that cocoa-farmers had high a propensity to save. This was shown most vividly by the repeated re-investment of profits in new cocoa farms, during the two great waves of cocoa planting. It is also shown by the 1956-57 survey of cocoa-farming families in Ashanti, as Barbara Ingham has noted. The question of whether the land tenure system significantly constrained cocoa production is answered, I think, by the fact that it proved compatible with those two massive planting booms. The colonial government s ultimate inaction on the proposal to introduce compulsory land titling was probably the right one as far as the cocoa economy was concerned, given the very high costs of implementation.

So far I have not identified any real sources of economic retardation arising from the peasant and small capitalist pattern of the colonial cocoa economy. But there remain two aspects that I want to raise, where the picture is at least more mixed. One is the transfer of resources, by market as opposed to fiscal mechanisms, from cocoa to other, potentially more value-adding, sections of the economy. Here the absence of clear, legally secure rights over farms was perhaps a serious potential problem, because (to exaggerate a little for brevity) it meant that farms were acceptable as security for loans only within the informal sector. Had bank loans been available to farmers, in principle at least it would have been easier to use mortgaging as a means for transferring cocoa profits into off-farm investments. On the other hand, in one important respect the characteristics of the crop encouraged successful producers to diversify. This was the absence of significant economic advantages of scale in cocoa production. There is a recurrent pattern that people who made lots of money in cocoa chose not to move from say employing ten cocoa labourers to employing thirty or a hundred labourers in cocoa. Rather, they would invest in transport (lorries) or urban property: the former, in particular, providing an area in which there were scale advantages which their success in cocoa meant that were now able to seize. In that sense cocoa generated incentives for its most successful farmers to re-invest elsewhere. So on transfers the story is ambiguous.

To end this section, let me return to the failure of statist economic policies from the decolonization era through to the adoption of Structural Adjustment. The policies in question clearly exploited the cocoa farmers: the policy that really hurt being not the marketing board s manipulation of the producer price per se, but the extreme over-valuation and relative non-convertibility of the currency that emerged in the 1970s. In the early 1980s Robert Bates argued that peasants were structurally vulnerable to such outcomes, because with a very large number of small producers there was a virtually insuperable free-rider problem which prevented them from acting collectively to assert their interests politically. I have argued before that this is too simple in the case of Ghanaian cocoa farmers: fundamentally because they included capitalists - and chiefs - as well as peasants. This combination made possible the relatively successful cocoa hold-up of 1937-38. It seems to me that it was not structural inevitability, but rather the contingencies of the party struggle over the terms of decolonization, that fatally weakened the farmers movement in the mid-1950s, dividing it between the federalist, Asante-based National Liberation Movement and Nkrumah s Convention Peoples Party, thereby creating the political conditions in which statist policies could be maintained even when, in the medium term, they damaged economic growth and the welfare of most citizens.

Conclusion

This paper is an attempt to give a Ghana cocoa perspective on the debate about the significance of peasant economies in the debate about colonial rule as an agent of or brake on capitalist economic development. I reach two general conclusions. The first is that the shift in British colonialism towards reliance on peasantries - or rather, on a spectrum of peasant and small capitalist farmers - was, in this case at least, an eventual acceptance of an existing indigenous reality. Specifically, it was in the fiscal, the broad economic, and the political interests of Britain to permit the continuation of the pattern of indigenous economic accumulation that had been established during the nineteenth century at least.

The second conclusion is that while I have some doubts about the prospects for diversifyinUse of uninitialized value in concatenation (.) or string at E:\listplex\SYSTEM\SCRIPTS\filearea.cgi line 451, line 61. g around cocoa, I am basically impressed by the efficiency of - again - not peasants necessarily, but rather the peasants and small capitalists that achieved two world-record setting cocoa-farming booms. For the duration of the colonial period at least, the West African Lands Policy was a help not a hindrance to economic growth.