Outgrowing the Plantation: A Regional Comparison of the Asset Specificity Problem and Technical Change in the Cuban Sugar Industry, 1899-1929

Alan Dye
Universidad Carlos III de Madrid

I wish to acknowledge those who have made useful comments on various versions of this paper or in discussions related to the work presented here: thanks to Larry Neal, Jeremy Atack, Werner Baer, Laird Bergad, Michael Edelstein, Stanley Engerman, Pedro Fraile, Knick Harley, Leandro Prados, Jordi Maluquer, Rebecca Scott, James Simpson, Antonio Tena, Carl Van Ness, David Weiman, Oscar Zanetti; and participants at the Conferencia sobre la historia de Cuba held at the Universidad Aut—noma de Barcelona, the Seminar for Economic History at Columbia University and the Fundaci—n Ortega y Gasset in Madrid. Also special thanks should go to the Fanjul family for making the Braga Brothers Collection available.

People have become accustomed to thinking of Cuba as isolated from western capitalism, but in the 19th century that was far from the case. Then, Cuba had one of the more innovative and advanced sugar industries in the world Đ a leader in the adoption of both steam power and vacuum-heating technologies in sugar manufacture. After the Cuban War of Independence (1895-98), producers wanted to regain that status, this time with closer political ties to their major market, the United States. For the first time, North American direct investment was attracted in a big way to the Cuban sugar industry, but surprisingly, it did not seem to reinforce the once highly competitive establishment that had been located in the western provinces before the separatist insurrections had begun. Rather, foreign investment went more frequently to the east where in 1899 most existing sugar producers were small and backwards. Cuban prominence in the sugar industry was regained, but the technologically dynamic center of the industry suddenly shifted to the east. The new continuous-process technologies of the second industrial revolution were adopted more quickly and more thoroughly in the east.

The change was dramatic, and it has drawn the attention of contemporaries and historians, who more than anything have emphasized the social discontinuities that resulted. Explanations of foreign investors' attraction to the east have been indirect or implicit, but they have tended to attribute it to easier North American domination of the backward and vulnerable eastern provinces. While domination may have been a consequence, the motive of investors participating in international capital markets was not domination but, rather, maximization of the returns to their investments, and it is not obvious why they chose the backward region over the developed. What was the economic incentive for investing in the east rather than in the west when the west had been the former bastion of Cuban sugar's competitiveness?

This paper argues that the institutional setting in the west, combined with the influence of pre- existing infrastructure, created a disadvantage for western sugar producers in attracting capital to invest in the latest continuous-process technologies. The argument relies on Oliver Williamson's transaction- cost theory of the effects of asset specificity on contractual relations (Economic Institutions of Capitalism, 1985). Using that approach, I show that the western part of the industry suffered a relative disadvantage in the adoption of new technologies because mills in that region were more likely to suffer holdup problems due to the transaction-specificity of canefield assets. Accordingly, institutional persistence in the west seems to have caused its relative decline1; indeed, it may be said that it was the west's very success with the previous century's technology which was at the root of its difficulties in the early 20th century.

Continuous process innovations were stimulated in the 1880s which revolutionized sugar manufacturing. These innovations were similar to those Chandler emphasizes in industrial countries, except that they were scattered in tropical (or semitropical) countries outside the industrial centers. Over the next few decades, further development and adoption of these new techniques in cane sugar manufacturing was widespread. Although the response in Cuba was interrupted and delayed by political conflict and war in the 1880s and 1890s; after 1901, Cuban sugar producers came to be among the most successful at employing the new technologies. In the next three decades, Cuba came to produce over 30% of the cane sugar supplied on the world market, and construction of capital- intensive, high-throughput mills in Cuba lasted until the late 1920s.

How these innovations were assimilated into sugar production in Cuba is interesting to observe from the point of view of its industrial organization. Two consequences are noteworthy. (1) There was an enormous change in the optimal scale of production, as occurred in other industries when continuous-process technologies were adopted. This resulted, interestingly, in the vertical disintegration of cane cultivation and milling into what was known as the central factory system of production. Where the self-contained plantations once reigned, central mills now contracted with nearby cane farmers, known as colonos, to procure the vast quantities of cane they had come to require. (2) Although mill capacities grew significantly in all parts of the island, those in the western part of the island grew less rapidly than those in the eastern part of the island. Cubans who lived through these events commonly noticed the difference in average mill capacities that developed between the sugar industries of the east and the west, and they emphasized the coincidence in the east of the largest mills, the greater presence of North American capital, and the greater concentration of land (latifundismo) in the hands of mills. However, behind these differences must have been economic incentives that made the east more attractive to investors than the west. These incentives need to be explained if we wish to understand why this regional investment pattern formed. In answering the question, we also learn something about the influence of rational economic responses filtered through institutions on regional differentiation.

To understand the economic incentives, it must be recognized that cane fields were site-specific assets. First, the working of the cane transaction was dominated by technical attributes that tied the cane cultivation site spatially to the mill. This is because cane is very sensitive biologically and deteriorates rapidly once it is cut. Deliveries from the harvest to the mill thus had to be made promptly. Consequently, the cane railroad became an essential component of sugar production. Investment in railroads had to be made by the mills in order to access the cane supplies, and that investment fixed the site for the long term. Given the specificity of the investments made by mills, theoretically one would expect to find threats of "opportunistic behavior" on the parts of colonos attempting to capture the quasi-rents created by the investments, unless contractual mechanisms were in place to preclude them. In the paper, an argument and empirical support is offered for the proposition that such holdup threats were effective when contract renewals were negotiated.

To summarize the argument, because of the site specificity of cane field investments made by the mill, dependence on short term cane supply arrangements was dangerous because of the potential for opportunistic behavior. Once the central mill (or central) and the cane supplier had agreed on a cane price and the central had made the commitment to the site-specific assets, the supplier could afterwards threaten to withdraw the cane supplies, holding out for a higher price than the one agreed upon. The salvage value of the site-specific investments was less than its productive value; therefore, the supplier's threat to withhold the cane supplies could effectively force the central to accept short-term losses of a magnitude up to the difference between the productive and salvage values of the assets. In this way, the supplier could appropriate some of the quasi-rents generated by the investment in railroads. Further, this appropriation was not limited to a redistribution of the profits between participants; it was feasible for the supplier appropriate as well some of the rents intended for the replacement or updating of equipment.2

Explicit contractual stipulations safeguarded against interruptions in cane deliveries during the term of the contract; therefore, the threat of holdup tended to take place during the negotiation for contract renewal. Such holdup threats were most effective if the colono owned the land he cultivated, so land ownership increased the relative bargaining position of the colono with respect to the mill. Consequently, the mills' cane costs tended to depend on the initial distribution of property in land, which differed between the two regions. Table 1 briefly summarizes some of the more salient regional differences. Mills capacities come to be considerably larger in the east where the percentage of cane delivered by landowning colonos had been and remained small relative to the west. (More detailed data as well as explanation about the technical and organizational changes that occurred in sugar production are given in the paper.)

While the introduction of the new technology and the increase in the optimal scale created incentives to invest in larger scales of production, the effective threats of holdup by landowning colonos, who were more prominent in the west, caused existing mills in that region to confront higher cane costs. The figures show that the percentage of cane coming from landowning colonos was two to three times higher in the west than in the east. The reason for this has to do with the histories of the two regions. The west was the center of the older 19th-century plantation industry. Whenever the larger scale technology was adopted, in the west this took place by consolidation of pre-existing plantation properties so that many of the colono contracts that were signed were signed with former plantation owners who ceased to mill their own cane. In contrast, the east consisted principally of new lands into which the sugar industry was being introduced. In these lands there was a greater tendency for the mills to own their cane lands and to contract them out to colonos who occupied them only as long as they were under contract.

This initial institutional difference implied a greater disadvantage, from the western mill owners' perspective, in attracting capital for investment. In contrast, in the eastern part of the island the institutional and infrastructural landscape was pliable so that contractual relations could be established differently based on the logic of the new technological system. This gave the eastern region an relative advantage in attracting scarce capital for investment in new technology and larger mill capacities. Consequently, the east came to be associated with larger, more technically advanced mills and larger concentrations of land holdings. That the east also came to be associated with the destination of inflowing North American capital was endogenous to this incentive structure.

Two kinds of empirical support are presented to support the proposition that this regional institutional difference had an effect on the returns to investment in sugar mills. First, evidence in negotiations of contract renewals between centrales and landowning colonos is presented in which this kind of holdup behavior was apparent at the point of contract renegotiation. This evidence is in the form of correspondence between mill managers describing the holdups in ongoing renegotiations. Second, evidence of higher rates of payment for cane to colonos in the west, where landowning colonos were prominent, is presented. (See Table 2.)

The pre-existing dense railroad network in the west contributed as well to this disadvantage. Railroad regulatory policies from the turn of the century tended favor the use of private railroads (offering service exclusively to particular mills) over public railroads (offering services to the public). Public railroads were heavily regulated by the Cuban Railroad Commission while few restrictions were placed on the construction of private, company railroads. This tended to encourage the construction and use of private railroads for hauling cane, and a number of mills invested in rail systems that allowed their cane to be hauled exclusively on private company railroads. However, a 1902 US War Department restriction on the construction of private railroads created a disadvantage for the western mills. Any private railroad could connect to, but could not cross, a public railroad Đ if a company wished to build a railroad that crossed a public railroad, it had to operate (to offer services) as a public railroad. Therefore, wherever the pre-existing system of public railways was dense, the construction of private rail networks was restricted. In the 19th century, a relatively dense public railroad network had been built to service the western sugar plantations so that under the new regulations, the exclusive use of private railroads to haul cane was restricted, and the cost of cane transport was likely to be higher. In contrast, in the east few public railroads had been built, and the mills were freer to construct private railway networks without encountering this regulatory restriction.

In conclusion, the combination of greater numbers of landowning colonos and the denser network of public railways in the west caused cane costs to be higher in the west. The higher rates of payment for cane that had to be offered in the west inhibited investment in the mills in the west because preferable alternatives for investment in sugar mills could be found in the east. If one observes the paths of growth of the mills in the island, it is apparent that many mills were constrained from growing to the enormous size of the largest mills in Cuba. The scarcity of capital was certainly an important influence in these constraints. Much of the capital on the island was destroyed during the War of Independence. The reconstruction of the Cuban sugar industry after the war depended greatly on North American capital, which could not be attracted with equal abundance to all mills. The pattern of flow of investment funds affected the abilities of mills to invest in the newest technologies, given the large increases in scale which the new technologies implied. But is was the competitiveness of each mill that determined its ability to attract scarce capital. It was difficult for single proprietors anywhere to accumulate the capital necessary for the modern industrial enterprise. Sugar had outgrown the plantation. The technology and the scarcity of capital in Cuba dictated that sugar producers needed North American capital markets. Their attraction to the land abundant east has been associated with latifundismo. But from the point of view of the central, it was driven by their concern to maintain control over the site-specific investments necessary to remain competitive as the world sugar industry adopted continuous processing.

TABLE 1								
REGIONAL COMPARISON OF MILL CAPACITIES AND CANE SUPPLY
SOURCE, 1904-29

(Mill capacities are expressed in thousands of bags (of 325 lbs. each).)
								
grinding season:	1904		1917		1929
			west	east	west	east	west	east
average capacity	40.9	45.2	129.0	188.7	207.2	417.4
capacity range		3-201	9-241	25-450	15-600	50-600	30-1300

share of cane
supplied by
landowning
colonos			43.2	15.3	--	--	22.8	10.3
no. active mills	143	31	138	59	96	67

Notes: The west is defined to include the provinces of Pinar del Rio, Havana, Matanzas, and Santa Clara. The east in defined to include the provinces of Camagźey and Oriente. (Figure 4 indicates the geographical locations of the provinces.) All active mills are reported in 1904 and 1929; 99% are reported in 1917. For 1917 and 1929, the mill capacities reported are engineering estimates. In 1904 actual production is used to proxy mill capacity. Sources: Computed using Cuba, Secretar’a de Agricultura, Comercio, y Trabajo, Memoria de la zafra (Havana, 1916/17, 1930); Cuba, Secretar’a de Hacienda, Industria azucarera (Havana, 1903/04, 1904/05).

TABLE 2

LIQUIDATION RATES AND SUCROSE CONTENTS FROM A SAMPLE OF WESTERN AND EASTERN MILLS

sucrose content of
cane:
mean
standard deviation

Sources: Computed using Braga Brothers Collection, series 127; Cuba, Secretar’a de Agricul-tura, Comercio, y Trabajo, Memoria de la zafra.

1 Persistence or path dependence arises because the formation of institutions in the past had a longlasting effect on current economic constraints. This aspect of the analysis is reminiscent of a study by Paul David, "The Landscape and the Machine," in Innovation, Technical Choice and Economic Growth (Cambridge, 1975), in which the incompatibility of pre-existing infrastructure hindered the adoption in Britain of innovations that had been successfully implemented elsewhere. S.a. Marvin Frankel, "Obsolescence and Technological Change, American Economic Review 45 (1955), pp. 215-49. However, in this case the pre-existing institutional framework, rather than the infrastructural layout, is seen to have affected current contractual relations.

2 See Oliver E. Williamson, "Credible Commitments: Using Hostages to Support Exchange," American Economic Review 73 (1983), pp. 519-40, and The Economic Institutions of Capitalism; and Benjamin Klein, Robert Crawford and Armen Alchian, "Vertical Integration, Appropriable Rents, and the Competitive Contracting Process," Journal of Law and Economics 21 (1978), pp. 297-326.