British Business In Brazil : Maturity and Demise ( 1850-1950 ) *

This paper analyses the long-term trends of "British business" in Brazil since 1850. It covers investment and other manifestations of the British presence such as those related to trade as well as financiaI intermediation. Primary interest is in British involvement in Brazilian private sector activities, whether by direct investment or by the flotation of sterling securities for firms operating in Brazil. The article also considers the role of London as a financiaI market where Brazilian public loans were floated, the relevance of Britain as a market for Brazilian commodities and as a supplier to Brazil, and British intermediation in Brazilian trade with third countries. It is divided in chronological sections: imperial years (1850-1889)j stagnation and boom (1889-1914)j first signs of decline (1914-1930), and the divestment period (1930 to the mid-1950's). The final section presents the conclusions and mentions post-1950 trends.


Introduction
In this article the term "British business" covers investment and other manifestations of the British presence such as those related to trade as well as financiaI intermediation.Primary interest is in British involvement in Brazilian private sector activities, whether by direct investment 01' by the ftotation of sterling securities for firms operating in Brazil.The article will also consider the role of London as a financiaI market where Brazilian pu blic loans were ftoated, the relevance of Britain as a market for Brazilian commodities and as a supplier to Brazil, and British intermediation in Brazilian trade with third countries.
In order to establish an effective chronological framework to deal with a subject of this complexity, it is essential to recapitulate the long-term trend of British investment in Brazil.Estimates of British investment in Brazil from the 1840s to the early 1950s are detailed in tables 1 and 2. In spite of criticisms of the data presented by Stone, suggesting that he overestimated the stock of British capital in Latin America even more than Paish, it seems that, at least in the case of Brazil, some of these criticisms do not apply (Platt, 1986:92;Paish, 1914).With the exception of estimates for 1895, Stone's data are in line with Brazilian officia.! figures on outstanding public foreign debt.Account is taken of repayments of principal, one of the main reasons suggested for the overestimation of British capital in 1913.Another major criticism of the Stone data is that no account is taken of the fact that sterling loans were held by non-British investors and that some series were ftoated outside Britain.While this is a relevant criticism, it is not important in the context of this article, which is concerned with the behaviour of British interests in Brazil rather than the actual distribution of assets between holders of different nationality.In this sense, the criticaI factor was whether a loan was denominated in sterling 01' noto In successive international negotiations from the 1890s to the 1930s, provoked by difficulties experienced by Brazil in obtaining sufficient foreign exchange to cover the foreign debt service and other remittances, British negotiators acted on behalf of ali the holders of sterling-denominated assets, rather than on behalf of British investors alone (Franco, 1990:26-8, 75-7, and 84;Abreu, 1988).For public portfolio, see Bouças (1955); data refer to nominal outstanding debt.For non-public investment  see Stone (1977:695); data refer to paid-in share capital and the face value of debentures.2.9 14.3 3.5 6.9 2.6 10.0 2.8 Sources: For public portfolio, see Abreu (1999:232) and Brasil (several years).For non-public investment, see Rippy (1949) and Spiegel (1949); both use data from the South American Joumal.These are open to criticism as they refer to nominal values, do not include securities sold privately and do not take into account foreign holdings af British securities and, conversely, British holdings af foreign securities.For alternative estimates, see yearly volumes of Bank of England (1950 and1957).Bank of England data also exclude securities not quoted on the London Stock Exchange and have a more limited coverage.See United Nations (1955:156).
British Business in Brazil: Maturity and Demise  Until 1895 practicalIy alI foreign investment in Brazil was British.By 1905 this had decreased to about 75% of total foreign investment, and in 1913 to 65%.In 1930 it was 57%, but estimates of other than British European investment are notoriously unreliable.In 1945 total British and US investments were of similar magnitude (Abreu, 1985:168 and1999:47).Data on the sectoral distribution of British investments are presented in tables 3 and 4. Figures taken from the S outh A merican .J O1trnal for the more recen t period are inadequate, as mentioned in the notes to table 2, but there is no alternative sectoral information.In contrast with the volume of Brazilian Imperial government bond issues in London, British direct investment was rather modest before 1875 and was concentrated in railways and public utilities, which received a profit guarantee issued by the central government and/or the provinces.After 1875, the inflow increased substantially and was sustained until the mid-1890s, as shown in tables 1 and 3.The share of railways increased to more than 80%, the bulk of the remainder corresponding to public utilities as well as industrial and commercial companies.Stagnation between 1895 and 1905 was due partly to the severe economic crisis which affected Brazil and to the conversion of guaranteed railway investments into federal government bonds in a major financiaI operation at the turn of the century (Rodrigues, 1902).Public portfolio British investment, which had been more important than other investments before 1885, maintained a share of about 50% of total investment between the mid-1880s and mid-1890s.But it rose steeply again after 1898 as a result of the funding loan operations, the large 1901loan to finance the rescission of railway guarantees, and new loans floated immediately after the turn of the century for the federal government, several states and major municipalities.By 1905 public portfolio loans corresponded to 2/3 of sterling investment in Brazil (table 1).
The nine years before the World War I marked the heyday of British private investment in Brazil: no less than 1'10 million entered the country on average every year.Two-thirds of the total increase in stock can be ascribed to the purchase of corporate securities.This is an important qualification to arguments which stress the importance of British direct investment in Brazil.The trend for British investment also contrasts with the pattern of US investment, which was to be overwhelmingly direct.The share of railway investment feU quite rapidly to less than 44% of total stock, and was almost ovcrtaken by investment in public utilities.This mainly took the form of purchases of securities of non-British-owned utility companies, most of which were registered in Canada.By 1912 no less than 1:22.9 million had been invested in securities of the Brazilian Traction, Light and Power Company, a major public utility operator incorporated in Canada.1By contrast, public portfolio investment increased at a slower rate to reach a leveI about 50% above that of 1905, and by 1913 it was again slightly below a half of British investment in Brazil.
Stagnation of the stock of direct investment and holdings of corporate securities folIowed after the World War I and lasted until 1930, but public portfolio investment increased by more than 30%.Decline in both cases was rapid after 1930 and continued until the late 1940s and early 1950s.During the latter decades blocked sterling balances accumulated during World War II were used to purchase British corporate investment and to repatriate public debt (Abreu, 1990a).By the mid-1950s the stocks of both public portfolio and non-public investment were around 25% of their leveIs at the beginning of the 1940s, with public portfolio decreasing earlier than non-public investment.The role of British capital in Brazil had thus beco me rather unimportant.
The chronology based on trends in the stock of foreign investment lS matched by the record of nominal rates of returno Pre-1914 average nominal annual yields on British railway investments in Brazil were generalIy above 5%, reaching almost 5.7% in 1913 (Rippy, 1949, ch. 13).2After World War I, as provisions to counter the impact of foreign exchange devaluation were much less effective, yields felI to 2.5% in 1923, and to an average of 1.5% during the depression.Even the San Paulo Railway, by far the most profitable railway enterprise in Brazil, which had yielded 11.2% on average from 1876 until1930, could do no better than 3.1% in the 1930s and 2.8% in the 1940s, due to the increased cost of servicing debentures.A verage yields on direct investments other than in railways were considerably higher: above 5% before 1933.Returns then dropped below 2% in the 1930s, but rose again to more than 7% in the late 1940s.
The above review indicates the chronological structure of this paper, which is divided into the folIowing sections: pre-1890; 1890-1914; 1914-1930; and 1930 to the early 1950s.The final section presents the conclusions and mentions post-1950 trends.This does not imply that alI turning points are necessarily of the same importance.One major turning point marking the history of British economic and financiaI involvement in Brazil and elsewhere in the world was undoubtedly 1914.Before this watershed, booms in the world economy, such as the one immediately before the war, resulted in a substantial inflow of British capital and rapid increase in bilateral British trade with 2 These yields were not spectacular, but perfectly respectable.For a comparison of realised rates of return in other countries, that is, rates of return calculated taking into account asset market prices and infiation, on Brazilian railway debentures in the period prior to 1914, see Edelstein {1982: 154).In the 1897-1909 period they were beaten only by the performance of US railway debentures.
economies such as Brazil (Marichal, 1989:74-7).While not outstanding, the performance of British trade early in the century was respectable.Sterlingdenominated capital ftows were, on the other hand, very substantial, outpacing dollar-denominated ftows although probably not those denominated in other currenCles.
After 1914, however, booms in the world economy did not have the same consequences.Sterling-denominated capital ftows recovered slightly in the second half of the 1920s, but were no match for dollar debt expansion.Anglo-Brazilian trade continued to contract after the World War 1.The fragility of the British balance of payments became increasingly pronounced and was paralleled by the deteriorating performance of British investments in public utilities in Brazil and elsewhere.Tt was just a question of time before divestment would occur.The financiaI arrangements put in place during World War II provided the conditions for an accelerated withdrawal of British investment on a world-wide basis.

The Imperial Years: 1850s-89
British inftuence in Brazil in the 40 yeaJ's before the proclamation of the Republic was overwhelming.This was firmly rooted in Britain's economic and financiaI pre-eminence in the global economy.Britain was practically the only supplier of capital to Brazil during the Empire and was the major, though rapidly declining, supplier of goods to Brazil.Up to the early 1870s Britain had maintained a share of around half Brazil's total imports, but then the value of British exports to Brazil stagnated, while those of new suppliers such as Germany and the United States of America increased very rapidly.Yet, around the turn of the century Britain was still the leading Brazilian supplier with a market share of about a third of the Brazilian market.Total British trade (imports plus exports), valued at approximately 1'12 million in 1901, was substantial.Total trade figures for Britain's principaIs rivaIs were: USA, 1'20 million; Germany, 1'8 million, France, 1'6.5 million (Brasil, 1990:572-4).
British-owned commercial banks operating in Brazil played a major role in financing trade, agriculture, and industry, although the domestic business of British banks had been eroded by the growth in competition from national houses towaJ'ds the end of the century (Joslin, 1963:162-5).Britain also supplied the bulk of the specialised labour force required by industry, mainly skilled textile workers, as well as those employed in railway construction and operation and other relatively sophisticated activities (Graham, 1968:137-9).British shipping dominated commerciallinks with the world economy (Greenhill, 1977a:1l8-55).
As a market for Brazilian commodities, coffee especially, British trade, which had lost its place of dominance to the USA in the mid-1850s, recovered in the 1860s due to the American Civil War, but collapsed after the early 1870s.The British share in Brazilian exports fell from nearly 40% in 1871 to less than 15% in the early 1900s as the vaJue of Brazilian exports to Britain fell by approximately 25% (Brasil, 1990:571-4).
In this initial phase the ftow of British capital to Brazil was closely linked with the development of the export infrastructure.The provision of modern urban services induced by the rapid growth of the population of Brazilian cities, which was to be the other major source of attraction for foreign capital, was still rather modest in the mid-1880s.
The London firm of N.M.Rothschild, as agents to the imperial government 111 London, hOO the monopoly of ftotation of Brazilian central government loans in London from the mid-1850s to 1907.The first non-British foreign loans were ftoated in Paris for Brazilian provinces and corporations in the late 1880s (Abreu, 1994a).
By 1895 the total British investment in Brazil of 1;78 million was divided into roughly equal parts between public portfolio investment and direct investment, including corporate securities (table 1).In the previous 20 years Brazil had attracted relatively less investment in government bonds -which had roughly doubled -, while direct investment and holdings of private securities grew fourfold.To promote foreign investment in firms operating in Brazil, the imperial government made lavish use of legislation which guaranteed a minimum rate of return, typically of 6 or 7% for up to 60 years, especially in the case of railways and sugar processing plants (Graham, 1968:152;Lewis, 1983:262-3;Eisenberg, 1974:88-92).These guarantees were frequently split between the central and provincial governments.Above a certain rate of return, normally 8%, profits were to be divided between the companies and guarantors.Certain successful railways gave up their guarantees as soon as it became clear that the return on authorised capital was likely to be above the 8% threshold.There is evidence that the system led to abuse.As guaranteed capital was defined on an open per kilometre basis, it encouraged overconstruction by some guaranteed lines (Pinto, 1903:182;Duncan, 1932:26-7).
The system, especially in the case of small companies working on a scale incompatible with high profitability, provided "no inducement (-.. ) to do more than make receipts balance expenditure" (Economist, 1902).
The profit-guarantee system meant that a major difficulty faced by foreign capital invested in an economy outside the gold standard was partly circumvented, as the actual rate of return did not vary with fluctuations in the exchange rate.On the other hand, while exchange rate fluctuations in Brazil seemed staggering for a European or North American observer, exchange movements in this period were not as wide as they later became.Between 1865 and 1889 the average annual rate of exchange ranged from a low of 17d per milréis in 1868 to a high of 27 7/32d in 1875 (Brasil, 1987).The depreciation of the exchange rate resulted in higher costs of imported inputs, parts, and components, as well as capital goods.It also had a direct impact on those companies indebted in international financiaI markets.Given the leveI of operating receipts, exchange depreciation would, of course, reduce the profitability in foreign currency, which was what mattered for the foreign investor.In the absence of guarantees, exchange depreciation inevitably resulted in pressure by the foreign companies to increase charges to domestic consumers of public services (Saes, 1981:160).I t is interesting to note that as early as 1885 there were already modest British holdings of sterling securities issued by Brazilian companies, mainly paper offered by railway companies in which coffee growers were the main holders of equity.Although some shares were purchased by Britons resident in Brazil, by far the greater part of British investment was in debenture stocks issued in London by companies registered in São Paulo and Rio de Janeiro.This trend would become even more maJ'ked after 1890 (Lewis, 1991:44-5 and app.).Successful British investment in Brazil before the Republican period was quite diversified: railways, mining, public utilities, and banking (Rippy, 1949, ch. 13).The San Paulo Railway, !inking the port of Santos to the city of São Paulo, enjoyed a de facto monopoly of railway transportation in paulista coffee zones which were rapidly displacing older producing regions in the Paraíba valley during the last decades of the Empire.By the 1890s coffee exports through the port of Santos exceeded for the first time shipments from Rio de Janeiro (Wileman, 1909: 632).Ali rail way !ines in the planalto coffee district connected with the San Paulo.It was a funnel through which the produce of the region was shipped overseas and the most profitable British-owned railway company in Latin America: by the late 1880s the company was paying annual dividends in the region of 10-12% (Graham, 1968:60-1 and 327).Whereas investment in the St. John deI Rey Mining Company, a gold-mining enterprise established in the state of Minas Gerais in 1830, was rather limited -in the region of 1'0.3 million -its average yearly return until the mid-1880s was not very much below 20% (Eakin, 1989:63).Gas companies were also initially rather successful, both in Rio de Janeiro and São Paulo.It is impossible to distribute banking profits among the different Latin American countries where British banks operated but both the London and Brazilian Bank and the English Bank of Rio de Janeiro paid dividends of at least 9% in the 20 years before the early 1890s (Rippy, 1949, ch. 13).
The most important sectoral failure of British investment in Brazil occurred in sugar milling, namely the central mills installed in the North-East of the country in the 1880s.They were depicted as a revolutionary advance that would fully exploit the advantages of large-scale production and as intrinsically more efficient than traditional, estate-based engenhos.But most central mills went bankrupt for reasons which ranged from fraud and gross mismanagement to technical problems and difficulty in obtaining access to supplies of low-cost sugar-cane (Graham, 1968:152-6;Eisenberg, 1974, ch. 5).

Stagnation and Boom: 1889-1914
After stagnating until the middle of the first decade of the century, British direct and private portfolio investment grew rapidly unti11914, but had to face stiff competition from capital exports from continental Europe and, to a lesser extent, from the USA.By 1914 non-British direct investment was substantial.Rough estimates suggest that the stock of continental European capital may have been as high as 1'110 million, up from 1'30 million in 1905.US investment was more than 1'10 miJlion (Abreu, 1985:168 and1999:47).Macroeconomic adjustment, as already mentioned, had resulted in an increase in British public portfolio holdings of 130% in the 10 years before 1905 (table 1).
Stagnation of British direct and private portfolio investment until 1905 was partly a resul t of the difficul ties faced by the economy from the early 1890s, but also reftected a contraction of the supply of capital.The early republican period was marked by a severe acceleration of inftation, partly related to loose monetary policies (Fishlow, 1972), partly to the foreign exchange problems which followed the Baring crisis of 1890 and the consequent reduction in foreign capital infiows, and partly to the fall in coffee prices in the second half of the decade.The milréis exchange rate fell dramatically from an average of 26 7/16d for the year in 1889 to just over 10d in 1894 and 7 3/16d in 1897 (Brasil, 1987(Brasil, :1354)).
Abrupt exchange devaluation adversely affected suppliers of public services and led to the introduction of a sliding scale of utility charges linked to the exchange rate.The system, which became widespread towards the end of the century, at least partly indexed railway and utility earnings.But fiuctuations in the tariffs provoked domestic protests both from producers and exporters, who had benefited from the exchange devaluation, and other economic agents whose incomes were slow to adjust price movements (Duncan, 1932:153-5).
Whereas coffee prices were sustained in the early 1890s, increased output resulted in price falls from the mid-1890s and fostered the growth of a consensus in favour of the adoption of an adjustment programme in 1898.Coffee growers were willing to make sacrifices in the immediate future in return for stability in the long run.The funding loan of 1898 was a watershed in the financial and monetary history of Brazil: it involved a renegotiation of the foreign debt -funding interest payments for three years and postponing repayments of the principal for 13 years -and conditionalities which entailed a domestic adjustment programme that was designed to balance the budget and stabilise the exchange rate.Taxes were increased and targets for the contraction of the money supply established as the conditions which Brazil had to fulfil in order to assure the refinancing of interest payments.The London and River Plate Bank played a crucial role in these negotiations. 3ollowing earlier attempts to fight the effects of the loose monetary policy pursued during the initial years of the republican period -the encilhamento, the application of a stringent monetary policy provoked a sharp recession that was followed by a severe banking crisis in which the largest Brazilian commercial bank, the Banco da República, went to the wall.The failure of the Banco da República was partly due to its attempt to follow the speculative lead of the London and River Plate Bank, which had tried to force the appreciation of the milréis.This move had been countered by the other foreign banks fearing a flight of deposits (Fritsch, 1988:8-9;Joslin, 1963:145).Nevertheless, renewed capital inflows, initialIy associated with the funding operation and subsequently the general growth in international liquidity, led to a recovery of the exchange rate which reached an average of 12d per milréis for the year in 1903and almost 16d in 1905(Brasil, 1987:1354)).
The onset of recovery was marked by the successful conversion of prac-ticalIY alI remaining railway guarantees into 4% bonds.Shareholders and bondholders of guaranteed railways were offered federal bonds in exchange for scrip.Most of these nationalised lines were then leased to private operators.The Economist was not surprised that Brazil had "got tired of this very one-sided condition of affairs", which entailed paying year after year a guaranteed income to smalI and wasteful railways.Having obtained control of the railways by "spending a little more than guarantee liability", the government could then "amalgamate" the smalI lines into larger concerns to be leased to "responsible people", in exchange for a fixed percentage of gross revenue (Economist, 1902).As a result, the Great Western of Brazil Railway rented some lines in the North-East while the Soro cabana Railway (which had gone bankrupt) was first leased by, and later transferred to, the Brazil Railway Company.The Sorocabana subsequently became the operating base of the American adventurer-entrepreneur Percival Farquhar, who developed diversified interests in Brazil and attempted to create a "railway empire" in South America.The Brazil Railway Company went bankrupt in 1914 (Duncan, 1932).4 As the exchange rate recovered after 1905, at a time when coffee output was expected to increase substantialIy, growers became interested in coffee valorisation, that is, in stabilising export prices on the world market through the accumulation of stocks.But high coffee prices would result in exchange appreciation and a falI in producer profits expressed in milréis.Hence, domestic and externaI conditions were ripe for the Brazilian federal government, after some initial difficulties, to stabilise the exchange rate by adopting the gold standard.The Conversion Office (Caixa de Conversão) was established in 1907, as the federal government decided to move towards a fixed exchange system (Joslin, 1963:158).
FinanciaI backing for coffee valorisation carne initialIy from non-British sources, through the direct involvement of continental European and North American coffee traders, because Rothschilds were extremely reluctant to become involved in the operation.This episode marks the beginning of the end of British pre-eminence in federal government financiaI operations, opening the field for French competitors who had previously been confined to lower class ftotations on behalf of Brazilian states and municipalities.
Despite problems in 1907, the performance of the Brazilian economy then reftected the boom in the international economy and in the international fi-nanciaI markets which lasted until 1912.This is shown by data on the inftow of foreign capital.During the period the stock of British investment in Brazil more than doubled.One-third of the increase in this stock corresponded to public portfolio holdings and 2/3 to either direct investment or holdings of corporate securities (table 1).Until the pre-war boom began to weaken in 1912-13, Brazil appeared to be an increasingly attractive environment for British firms.The economy grew rapidly.
One of the striking features of British investment in sectors other than public portfolio assets has already been noted: the relatively modest increase in direct investment (less than twofold) compared to the increase in holdings of private securities (more than sevenfold).A significant proportion of these holdings related to securities of companies registered in Canada, the most prominent being the Brazilian Traction, Light and Power Company Limited.The company remained the main foreign operator of public utility concerns in Brazil until the late 1970s.The importance of Canadian activities in the supply of public services in Cuba, Mexico and Brazil has been ascribed to the lack of investing outlets for Canadian entrepreneurs given the country's modest urbanisation (Drummond, 1959, quoted in Stone, 1977:717: McDowall, 1988).This characteristic of British involvement in Brazil certainly contributed to the relative aloofness displayed by British officials and financiaI authorities when the decline in Britain's stake in Brazil became evident later in the century.Another feature which was to weaken the British position in the long run was the concentration of investment in the provision of traditional public services, in sharp contrast with the sectoral distribution of emerging US investment, which was concentrated in distribution, manufacturing and public utilities other than railways (Lewis, 1938, app. D).
The British position in the financiaI and commercial sectors remained strong.British banks operating in Brazil -the London and Brazilian Bank, the British Bank of South America (formerly the English Bank of Rio), and the London and River Plate Bank -continued to play an important role, but had to face increasing competition from continental European banks.German overseas banks were particularly aggressive, and successful, in their attempt to obtain a large share of domestic business (Moller, 1988:223-44).Joslin estimated that British banks controlled a third of the deposits of the banking system in 1914.The proportion would have been even higher around the turn of the century (Jones, 1993:57).British insurance companies faced similar conditions but still had 25% of the Brazilian market in 1910 (Graham, 1968:91).The previously important role played by British exporting houses, however, was almost certainly undermined by the decreasing importance of Britain as a market for Brazilian commodities in the last third of the 19th century and, more particularly, by British unwillingness to drink coffee.British firms such as Johnstons and Naumann Gepp, who had controlled perhaps 1/3 of Brazilian exports of coffee early in the century and, along with other foreign houses, had successfully eliminated competition from domestic intermediaries (the comissários), were in an unfavourable position to maintain their position given the geographical structure of the market.Involvement in the coffee trade implied, in certain cases, the ownership of coffee plantations as these passed to the hands of creditors, as a result of bankruptcy of their owners: this may have represented one of the few areas of expansion for British export houses (Greenhill, 1977b:214).
As a supplier of the Brazilian market, Britain faced difficulties due to US and German competition, despite a rapid growth in purchases from the USA and Germany, Britain remained the principal source of Brazilian imports during the early years of the 20th century, although the margin was narrowing.In 1901, combined US and German exports to Brazil were about 2/3 the value of purchases from Britain.By 1912, imports from Britain were around 3/4 the combined value of imports from Germany and the USA (Brasil, 1990:572-4).But most of the damage seems to have occurred either before 1900 01' after 1914.
The British record concerning direct investment in industry was not impressive, in spite of inducements provided by the extremely high tariff levied in Brazil.There was a relatively minor involvement in the production of shoes, matches and textiles (including sewing thread), and a very successful experience in the production of wheat fiour.The ordinary dividends distributed by the Rio de Janeiro Flour Mills and Granaries were 10.9% on average between 1893 and 1902, rising then to 19.8% in the following decade (Rippy, 1949:157).The production of matches was undertaken by two factories in Niterói and São Paulo, controlled by Bryant and May; its Brazilian arm, Cia.Fiat Lux, has held the position of major match producer in Brazil ever since.5Shell entered the country with a capital of 1'150,000 as the Anglo-Mexican Petroleum Products Company Limited in 1913 and was to become the largest foreign company operating in Brazil (A história da Shell Brasil, n.d.:1).
It is possible to detect during this period the beginning of a change in the attitude of a number of political groups regarding the role and contribution of direct foreign investment to the country's long-term economic development.Adverse criticism of the effects of government borrowing had, of course, a long history, given the scandalous circumstances which marked the flotation of certain imperial loans in the London market (Mendonça, 1968, ch. 12).Nationalist criticisms of direct investment, however, were rare before the turn of the century.The monopoly position of the San Paulo Railway gave rise to disapproval, but it was most probably the activities of Percival Farquhar that increased the weight of such arguments (Topik, 1987:68-9).By means of an intricate series of financiaI coups de main, Farquhar gained control of transport systems in the Amazon region (the harbour at Belém; shipping lines; and the Madeira-Mamoré Railway), and in the South (the harbours of Rio, Paranaguá and Rio Grande; the Paulista, Sorocabana, Mogyana and São Paulo-Rio Grande railways) between 1904 and 1913.He aIs o purchased farms and slaughterhouses and made other minor investments.The sources of the group's capital were diversified, including French and US funds, but Farquhar relied significantly on the London financiaI market.In addition, he had strong, if sometimes acrimonious, relations with the Canadian syndicate, which was to play such an important role in Brazil.Farquhar's buccaneering activities raised sharp criticisms, but these were possibly not always disinterested, since Brazilian entrepreneurs had been squeezed by his hectic activities. 6Public animosity was aIs o provoked by Farquhar's erstwhile Canadian allies, Brazilian Traction, which had a virtual monopoly of public services -especially the distribution of electricity -in Rio de Janeiro and São Paulo.The antics of Brazilian Traction stimulated criticisms and helped consolidate the image of the "Canadian Octopus" in the popular mind (Gauld, 1972).
It is important also to understand that the Brazilian agricultural oligarchy was not committed to "economic liberalism" to the same degree as exporters in other primary producing countries, such as Argentina, as was made clear by the mobilisation of government support in order to interfere with market forces in the coffee trade.This made resistance to high tariff protection by rural producers and exporters at best half-hearted and helped cement an alliance between export-agriculture and import-substituting industry, a community of interests already indicated by the investment diversification strategy employed by many coffee growers in the last decades of the 19th century.Market power in the world coffee market and the price inelasticity of coffee demand allowed the transfer of some of the costs entailed by protectionism to coffee importers (Abreu, 1994b;Abreu, Bevilacqua & Pinho, 1996).The erosion of liberal principIes may have strengthened nationalist criticism of the role of foreign investment.

Signs of Decline: 1914-30
World War I marked the beginning of a period of serious difficulties for Britain.The liquidation of foreign assets during the conflict, the burden of war debts, and a loss of international competitiveness by sectors of British industry made the continued role of Britain as the most important source of capital for nations such as Brazil impossible in the post-war period.These problems were aggravated by the measures taken to enable the sterling to return to its pre-war parity against the US dollar in 1925.Taking the decade as a whole, government dollar loans were significantly greater than sterling loans and the New York market became particularly important for Brazil in the late 1920s.However, in spite of moral suasion by the Bank of England against loans to countries outside the empire, London continued to play an important role as a source of finance for the Brazilian government.This was particularly important in the case of funding for coffee price support in the second half of the 1920s, as US finance could not be raised due to anti-trust legislation (Moggridge, 1972).7 The stock of British direct investment and holdings of private securities stagnated during this period, while dolIar direct investments and public bonds increased more than fourfold.However, in 1930 the stock of US assets was still only about 1/3 of the British.This trend was, of course, related to the main characteristics of British investment in 1914, noted above, namely the concentration in portfolio holdings and direct investment in railways and public utilities.Significant devaluation of the milréis, from 16d in 1913, to 12d during the war, then to a low of 4d in the early 1920s, triggered a dramatic falI in the profitability of railways and public utilities, underlying the vulnerability of their position compared to investment in other sectors which did not depend on negotiating tariffs with the government.The result was conflict between foreign-owned railways and utilities and government, a deterioration in services and increased public support for greater State control (Duncan, 1932).
The report of the influential Montagu financiaI mission in 1924 is a good indication of the backward looking priorities of the British in Brazil as it considered the difficulties faced by British firms providing public services (Montagu, 1924).One of the main suggestions of the report was the establishment of a tribunal to deal with the fixing of railway tariffs with, unsurprisingly, a large British membership on the board of experts.There were aIs o discussions about the disposal of the State-owned Central Railway -to be purchased by the San Paulo Railway -and of Lloyd Brasileiro.In similar vein, the report recommended the abandonment of proposed federal government involvement in a scheme to establish an iron and steel industry in Minas Gerais (Fritsch, 1980:735-6).
The iron ore question was to dominate the political debate about the role of foreign direct investment in the 1920s.Farquhar made a modest comeback on the Brazilian scene after the World War I by associating himself with British interests in the ltabira Iron Company, which controlIed a sizeable portion of the rich iron ore deposits in ltabira do Mato Dentro and the unfinished Vitória a Minas Railway.TypicalIy, Farquhar tried to put together an ambitious project to export iron ore, but faced the opposition of a powerful nationalist lobby led by Arthur Bernardes, then governor of the State of Minas Gerais, and subsequently president of the Republic (1922-26).When, after many wrangles, the project was approved by the Brazilian federal government, it was finalIy sunk by the impossibility of raising fresh capital in international financiaI markets after mid-1928. 8 British interest in railways, shipping, and banking underlines the relative obsolescence of the British negotiating agenda, if compared with that of the USA.When matters relating to direct investment were addressed, British demands were focused on the declining railway sector, as investment in the more dynamic public utilities was mainly indirect, namely minority ownership of securities of Canadian and other foreign concerns.Interest in shipping and banking might suggest a British emphasis on intermediation, but this interpretation does not fit well with relatively low priority accorded to trade matters.Finance was the first priority among British officials and businessmen, then the public utilities and r ailways , and finally trade and direct investment in other branches of activity.This was to become even clearer after 1930, but it was already well delineated in the period following World War I (Abreu, 1999).
In the 1920s the British share of Brazilian exports was never more than 10% and it reached lows of almost 3% late in the decade.By the mid 1920s Argentina, Italy and the Netherlands, besides the USA, Germany, and, occasionally, France, had become more important markets for Brazil than Britain.While in the early 1920s some of the graund lost by British exports during the war had been recaptured, immediately before the depression the USA had become by far the single largest supplier of the Brazilian market, and Britain was relegated to a precarious second position, providing no more than a fifth of total Brazilian imports.In 1929 British exports to Brazil stood at 1'16.6 million while imports fram the USA and Germany were 1'26 million and 1'10 million respectively (Brasil, 1990:572-4).
British capital was still singularly absent fram industry, in spite of limited diversification after 1914.Firms such as Frigorífico Anglo, which was already operating in the River Plate republics, invested in meat packing in Brazil.Cia. de Cigarros Souza Cruz was taken over by the British American Tobacco Company, which rapidly established a dominant market position in the tobacco (cigarette) industry which it continues to defend with success (Souza Cruz, 1983:1).Limited British direct investment in manufacturing was a consequence of the well-known lack of competitiveness of British firms, especially if compared to those from the USA.North American, and, to a lesser extent, continental Eure pean investments were concentrated in branches of industry such as electrical and transport equipment, shoe machinery, meat packing, cement and steel, dynamic sectors in which global British performance was mediocre or indifferent. 9British banks, in line with the fate of other foreign banks, steadily lost ground in Brazil.In 1923 Lloyds Bank took over the London and River Plate, which had already absorbed the London and Brazilian Bank, and changed its name to Bank of London and South America.Amalgamation was completed in 1936 when Bolsa took over the Anglo-South American Bank which had acquired the British Bank of South America in 1920 (Joslin, 1963:243-5).By the end of the decade, British banks may have retained an edge over foreign competitors, but they had lost their formeI' dominant position.German competition had circumscribed or eroded their business (Moller, 1988:248).A greater challenge, however, was posed by domestic houses.In São Paulo, for example, foreign banks had held more than half total deposits in 1910.By 1915 the proportion had fallen to 46% and in 1928 stood at 25% (Saes, 1986:104,107 and lll).With the collapse in Brazilian exports after 1929, and the imposition of exchange control, opportunities for British banking business would be further limited.

Divestment as a Policy: 1930 to the Mid-1950s
The impact of the Depression on the profits of direct foreign investment and the service of corporate securities held by the British as portfolio investments was massive.This was due to a general decline in the leveI of economic activity in Brazil and to the extent of the country's balance of payments crisis, which resulted in the imposition of foreign exchange controls that directly affected remittances of profits and interest.Whereas the impact of a fall in export prices and the interruption of foreign capital inftows was partially accommodated by exchange depreciation, there were limits to such a process, given its impact on government finances and the weakening of world coffee prices, since Brazil remained the major supplier of the world market.
The government introduced foreign exchange controls and a single "official" exchange rate.Since this was overvalued, the Bank of Brazil, acting in the name of the government, had to introduce criteria to distribute foreign exchange, for which there was excessive demando Profit remittances occupied a relatively low place in this list of priorities, coming after official and foreign debt payments, imports thought to be essential, and small private remittances.But, in spite of these arrangements, exchange arrear-s accumulated, affecting the settlement of commercial debts and profit remittances which were then periodically funded through the issue of specific medium-term loans (Abreu, 1990b:75, 84-90).
Public portfolio service payments were sharply reduced or interrupted and the greater part of future British financiaI effort in Brazil would be concentrated on the maintenance of debt service.British financiaI involvement in Brazil between 1930 and the early 1950s was to be restricted to certain operations designed to finance the thawing of financiaI and commercial arrears, the transfer of which had been made impossible due to the lack of foreign exchange cover, and to the financing of capital goods exports such as those related to the contract won by Metropolitan Vickers in 1934 to electrify Rio de Janeiro's suburban railways, operated by the government-owned Central do Brasil.10 The stock of British investment other than that in government loans contracted significantly in the 1930s.Private utility companies and railways were particularly vulnerable to foreign exchange fluctuations, the acceleration of inflation, and poli ti cal frictions with regulatory bodies about acceptable minimum rates of returno Moreover, as elsewhere in Latin America, railways generally faced mounting pressure from competi tive road transportation.The outstanding stock of government loans decreased more slowly, since service was concentrated on interest payments and relatively little amortisation occurred.
Britain's relative commercial position in Brazil improved marginally, as earlier adverse trends in Anglo-Brazilian trade were to some extent reversed.In the 1930s Britain recovered part of the ground lost as a market for Brazilian commodity exports, especially cotton.But the British share of Brazilian imports was halved as a result of American and, to a lesser extent, German competition induced by bilateral arrangements with Brazil (Hilton, 1975;Abreu, 1991:382-4).During the war Britain absorbed a much increased proportion of Brazilian exports -a peak of 16% was reached in 1942 -, but almost disappeared as a supplier to Brazil due to export controls and payment arrangements (Abreu, 1999, ch. 9-10).After the war Britain was to become a minor trading partner of Brazil, accounting for only a small share of total exports and imports.
Britain abandoned its traditional liberal stance on commercial policies in the early 1930s, with the adoption of Imperial Preference, and made widespread use of whatever bargaining power it possessed as a buyer to foster exports in such countries as Argentina.Export promotion in countries such as Brazil, where Britain generally ran a commercial deficit, was badly affected.The paramount importance of financiaI over commercial issues was made clear during every negotiation involving Brazil and Britain in the 1930s: on the public foreign debt, on commercial and financiaI arrears, and on trade matters.British authorities in the first instance wished to maximise the rate of return in sterling ar sterling-denominated assets in Brazil and, once the war created the appropriate conditions, became interested in maximising the prices that they could obtain by selling such assets (Abreu, 1984:157-9).
An important consequence of the imposition of foreign exchange controls in 1930 was that many commercial and financiaI negotiations necessarily involved representatives of the governments of Brazil's creditors, since the Brazilian government controlled the distribution of foreign exchange.This was even more clearly the case after 1939, when exchange and trade controls were also imposed by Britain.
With the benefit of hindsight it is tempting to detect in the British attitude and policies relative to Brazil a singular lack of perception of the economic potential of the country.This was reflected in the sluggish response of even the more efficient British firms confronted by competition in the Brazilian market, whether in terms of import participation or willingness to invest in local manufacture, when it was clear that market share could only be sustained by developing production in Brazil.It was as if the dominant link with Argentina had resulted in an impairment of the British capacity to re-assess the relative importance of its economic objectives in Latin America as a whole.
After 1930 there was a change of attitude in Brazil regarding the role of foreign capital.Policy-makers began to assess the net balance of payments contribution of foreign firms.This had an unfavourable effect on British investment, given the importance of their holdings of securities, and in view of the fact that the bulk of British firms clearly had no intention of increasing its operations in the country.On the contrary, most British firms were more interested in maximising remittances.With the increasing strength of extreme left-and right-wing groups in the mid-1930s, denunciations of the political and economic costs entailed by foreign investment became common, if not particularly well formulated. 11uring the first half of the 1930s it is difficult to detect a clear Brazilian government policy concerning private foreign investment, in spite of the many problems faced by foreign companies trying to remit profits and interest abroad, especialIy before 1934.As a declaration of intent, the 1934 Constitution, as welI as the new Mining and Water Codes, included quite radical provisions concerning foreign investment: concessions related to mining and water power were to be restricted to Brazilian citizens, and foreign banks and insurance companies were to be nationalised.The 1937 Constitution included similar provisions, but in a more radical form, as it was stated that mining and "essential" industries should be nationalised.Nationalist legislation was introduced, regulating the proportion of Brazilian citizens employed by each firm, as welI as the ownership and operation of Brazilian shipping companies.A de facto State monopoly in the exploration and refining of oil was created, and reinsurance with a new State-owned insurance agency became compulsory (Abreu, 1999:260-3).
But no action was taken and further legislation mitigating the more radical features of the 1937 Constitution concerning foreign involvement in banking and insurance activities was introduced.12In spite of this, the share of foreign banks in total banking sector assets, which had already decreased from 25.4% in 1929to 17.8% in 1939, felI to 5.2% in 1945(Brasil/Ministério do Trabalho, 1932:50;Brasil, 1987Brasil, :1353;;Brasil/Conselho Nacional de Estatística, 1947:265-6;Brasil/Ministério da Fazenda, 1940-7;Brasil/Ministério da Fazenda/Tesouro Nacional, 1947).The relative importance of foreign insurance companies was aIs o reduced, especialIy in the case of reinsurance business, which had been nationalised just before the war.At end of the first Vargas period, in 1945, an American business offensive began, backed by the State Department, to end the Brazilian policy which prohibited foreign involvement in oil exploration. 13The US campaign would not cease until the early 1950s, when Petrobras, a government-owned concern with monopoly rights in oil prospecting, exploitation and refining, was created.N ationalist legislation faced no important internaI opposition, not even from more conservative political circles, but there was an outcry over the alIeged excesses of legislation in the areas of mining and water power.It was argued that the limited availability of domestic capital would hinder the development of these resources (Amaral, 1938:240).
While it is almost impossible to evaluate the impact of such legislation on actual aggregate flows of foreign investment, it is unlikely to have had any significant impact on the decisions of the representative British investor, who, as a rule, wanted to get out of Brazil as soon as possible, provided certain minimum conditions were meto This attitude, which was the outcome of a long period of declining profitability and worsening relations with the government, was to find a perfect answer when a use had to be found for Brazilian sterling reserves which had been accumulated between 1941 and 1947 and which could not be used to purchase goods in Britain.
Britain, during World War lI, was able to attract important financiaI contributions to its war effort from other countries.British current payments were financed by exports, the sale of assets abroad and the accumulation of sizeable debts.Some debts were related to lend lease and gifts from Commonwealth countries, but an important part took the form of blocked sterling balances owed to other countries.In addition to members of the sterling area, several countries were persuaded to hold increasing amounts of inconvertible sterling (Sayers, 1956:15-9, ch. 8-9 and 14).Brazil was no exception, and claims against Britain rose rapidly as the war proceeded.An important early use of blocked balances was in the settlement of claims related to British-held assets which had already been taken over by the Brazilian government.At the beginning of the war, British policy was to avoid those claims which involved wider issues, like the Brazil Railway, and concentrate on well defined cases which were already pending, such as the nationalisation of the Sou thern San Paulo Railway, the Sorocabana Railway and the Brazilian Land and Cattle Company (Abreu, 1990a).
Reduction of blocked sterling balances after 1947, however, resulted essentially from payments related to the take-over of British investments in public utilities and railways.In early 1948 the Empresa de Armazéns Frigoríficos, the Cia.Indústria Brasileiras de Papel and the Southern Brazil Lumber and Colonization Co. were purchased for /;2 million; in the third quartel' of 1948 the San Paulo Railway was purchased for /;7.6 million, and at the end of 1951 the Leopoldina Railway was bought for /;10.3 million.Some other purchases cannot be dated precisely but were certainly made before the end of 1951: these included the Great Western Railway (1.;3.7 million), the State of Bahia Railway (1.;0.6 million), additional Leopoldina payments (1.;2.3 million), and additional San Paulo Railway payments (1.;1.0 million).14With few exceptions -essentially the San Paulo Railway -these were rather rundown concerns.Prevented from increasing tariffs, the majority had to confront exchange losses and capital scarcity.Unable to maintain profitability in sterling or secure new funds overseas, they cut depreciation provision to the bone.The consequence was a sharp deterioration in the quality of services provided and further aggravation of the difficulties between the companies and the government.The British quite lamented having to part with the San Paulo Railway, in the words of sir Otto Niemeyer, "much our best asset in Brazil ( ... ) and far from dead" .15However, there was much rejoicing in London over the prices agreed for the purchase of the Leopoldina Railway and the Great Western: the offers were considered very advantageous to the companies.16 British willingness to divest had an important if somewhat unexpected consequence in the mining sector.In the context of Brazil-US-UK supply negotiations in 1942, the British government agreed to transfer to the Brazilian government at no cost the former rights of the ltabira Iron Ore Company in return for a Brazilian commitment, which remained unfulfilled, to supply highgrade ore.Helped by US Eximbank loans, these mining and transportation concessions were developed by Cia.Vale do Rio Doce, which was to become one of the major and most efficient State-owned Brazilian companies and a world market leader in the iron ore business (Gauld, 1972:330-1).
Little can be listed to compensate for the divestment of Britain's interests in railways and public utilities.Established firms such as British American Tobacco and Shell did well, but only a few cases can be found of new direct investment in the 1940s and 1950s.A good example of the extreme reluctance of British business to enter the Brazilian market is the judgement of the head of ICI in Brazil in 1934: "[as] importations are looked upon as a drain on the national economy ( ... ) I consider it, therefore, expedient that we establish a small vested interest in Brazil, so as to create the impression in official quarters that we are interested in the country's industrial development".The deception was taken perhaps too seriously as ICI became the most active major British firm in Brazil in the 1930s.Through joint ventures with Dupont it took over Cia.Brasileira de Cartuchos, a cartridge factory, and established Duperial, a chemical concern (Reader, 1975:226).The other major British firm to enter Brazil was Unilever, whose interest began with the acquisition of Atkinsons and the establishment of a factory in 1930 to produce standard Lever products.It lost money until the late 1930s, but then became profitable, and by 1939 was the market-leader in soap (Wilson, 1954:360 and1968:258-60).17Souza Cruz excepted, though there is no instance of a large wholly British manufacturing or commercial concern entering the Brazilian market before 1960.The most active firms were Anglo-Dutch or acted at least initially through joint ventures as in the case of ICI.It was not a question of failing to hold the ground against competition: British firms were reluctant to enter the market.Yet those which decided to enter fared rather well.

Remnants of the Past: the Period After the 1950s
British firms play a relatively secondary role in Brazil today.Souza Cruz (BAT), Royal Dutch Shell, HSCB Bamerindus, Unilever, and Lloyds were amongst the 50 largest foreign groups in 1998, but only Souza Cruz could be classified as a wholly British firm in the manufacturing sector, given the duality which characterises management of the two Anglo-Dutch multinationals.18 The history of the decline of British business in Brazil does not differ greatly from what occurred in many other countries.Indeed, what is striking is its predictability.There was an important British presence until 1930 as a source of government finance.Of the other assets, 40% were direct investment and the rest were corporate securities of which only 40% were floated for concerns registered in Britain.The overwhelming share of British direct investment was in railways and public utilities.
As the depression of the 1930s interrupted foreign loans related to government finance, this modality of business was to concentrate on defensive targets 17 In 1960 Lever merged with Gessy, its main domestic competitor.
18 These firms were placed in the 10th, 12th, 13th, 14th, and 47th position among foreign groups in terms of total net assets.HSBC Bamerindus is a recent entrant in the Brazilian banking sector and has assets more than five times those of Lloyds (Gazeta Mercantil, 1999).such as collecting service payments or the repatriation of capital.Domestic financiaI intermediation was increasingly performed by Brazilian-owned financiaI intermediaries.The industrial sector, which had an impressive performance in Brazil after the early 1930s, with average output growth of over 10% annually and facing very little adverse interference by government, was precisely the sector which had attracted very little interest from British firms.
War financing dictated the repatriation of British assets in Brazil in the 10 years following 1941/42.Had not the blocked sterling balances been available in the 1940s, there is little doubt that Anglo-Brazilian relations would have been dominated for a long time by wrangles about profitability, access to foreign exchange, and quality of services provided by British railways and public services.Decline was inevitable: there was uncertainty only in relation to timing.The actual chronology of British divestment in Brazil was the result of the fragility of the British investment position (U'ising from a concentration in sectors of low profitability and the availability of sterling balances which made possible the outright purchase of such assets.____ , .Anglo-Brazilian economic relations and the consolidation of American pre-eminence in Brazil, 1930-45.In: Abel, Christopher & Lewis, Colin M. (eds.).Latin A merica: economic imperialism and the State.London, 1991.