Outside Options, Coercion, and Wages: Removing the Sugar Coating∗

Home , Colony of Jamaica, Plantation economy

Christian Dippel† Avner Greif‡ Dan Treﬂer§

December 14, 2019

Abstract

In economies with a large informal sector firms can increase profits by reducing workers’ outside options in that informal sector. We formalize this idea in a simple model of an agricultural economy with plantation owners who lobby the government to enact coercive policies—e.g., the eviction and incarceration of squatting smallhold farmers—that reduce the value to working outside the formal sector. Using unique data for 14 British West Indies ‘sugar islands’ from 1838 (the year of slave emancipation) until 1913, we examine the impact of plantation owners’ power on wages and coercion-related incarceration. To gain identification, we utilize exogenous variation in the strength of the plantation system in the different islands over time. Where planter power declined we see that incarceration rates dropped, and agricultural wages rose, accompanied by a decline in formal agricultural employment.

Keywords: Labor Coercion, Economic Development, Institutions JEL Codes: F1, F16, N26

∗We are especially indebted to Jim Robinson who, in the initial stages of the project when we were wallowing in case studies drawn from disparate times and places, encouraged us to focus on the British Empire and the under-exploited Colonial Blue Book data. We are also indebted to Elhanan Helpman for his encouragement in exploring the relationship between international trade and domestic institutions. We beneﬁted from discussions with Daron Acemoglu, Lee Alston, Quamrul Ashraf, Magda Bisieda, Kyle Bagwell, Abhijit Banerjee, Stanley Engerman, James Fenske, Murat Iyigun, Sara Lowes, Karthik Muralidharan, Suresh Naidu, Luigi Pascali, Diego Puga, Manisha Shah, Shanker Satyanath, Alan Taylor, Duncan Thomas, Vitaly Titov, Francesco Trebbi and seminar participants at Boulder, CIFAR, ERWIT, Harvard (PIEP) , LSE, Los Andes Namur, the NBER Development Economics Conference, PSE, Ryerson, Stanford, Toronto (Law Faculty), Toulouse, Western, the World Bank, UBC, UC Davis, and UC San Diego. We thank Scott Orr, Nicolas Gendron-Carrier, Jacob Whiton and especially Jake Kantor for fantastic research assistance. A previous version of this paper was circulated under the title “The Rents From Trade and Coercive Institutions: Removing the Sugar Coating” †University of California, Los Angeles, and NBER. ‡Stanford University and “Institutions, Organizations and Growth Program’, Canadian Institute for Advanced Research, Toronto, Ontario, M5G 1Z8, Canada. §University of Toronto, NBER and “Institutions, Organizations and Growth Program’, Canadian Institute for Advanced Research, Toronto, Ontario, M5G 1Z8, Canada. “The fact that the wage level in the capitalist sector depends upon earnings in the

subsistence sector is of immense political importance, since its effect is that capitalists

have a direct interest in holding down the productivity of the subsistence workers.

Thus the owners of plantations, if they are inﬂuential in the government, are often

found engaged in turning the peasants off their lands.” — Lewis(1954)

1 Introduction

Many economists and historians would agree with Acemoglu and Wolitzky’s 2011 assessment that “the majority of labor transactions throughout much of history and a signiﬁcant fraction of such transactions in many developing countries today are coercive”. Indeed, labor coercion is at the heart of much of the literature on long run development and institutional change (Domar,

1970; Engerman, 1999; Acemoglu, Johnson and Robinson, 2001, 2002; Engerman and Sokoloff,

2002; Greif, 2005; Nunn, 2008; Dell, 2010; Naidu, 2010; Nunn and Wantchekon, 2011; Naidu and

Yuchtman, 2013; Bobonis and Morrow, 2014; Ashraf, Cinnirella, Galor, Gershman and Hornung,

2018; Lowes and Montero, 2018). Despite this, rigorous empirical evidence on labor coercion is scarce and mostly focused on relating present-day outcomes to historical labor coercion.1

Focusing on the workings of labor coercion rather than its long run consequences, we exploit a historical setting involving 14 British West Indies sugar colonies from 1838 —the year slaves were emancipated in the British Empire— to 1913. We are thus studying 14 free labor markets at their inception. Before 1838, the 14 colonies we study were exceedingly similar. Economically, all were slave societies and all were completely specialized in sugar cane production. Institutionally, all had the same political and legal systems inherited from Britain and were dominated by a small group of white planters. After Emancipation, “the main fact of life in the free West Indies was that black laborers were unwilling to remain submissive and disciplined cane workers” (Green, 1976,

170). We study how planters over the ensuing 76 years used their inﬂuence over the state to enact coercive policies that kept wages low and secured a steady supply of labor.

In keeping with Lewis’ quote, our focus is on ‘legal coercion’, i.e., the use of the state’s legislative and judicial institutions to manipulate workers’ outside options. This focus is particularly

1Two exceptions are Naidu and Yuchtman(2013) and Bobonis and Morrow(2014), summarized below.

1 pertinent where workers’ wages in the formal sector are determined by outside options generated in the informal sector. Our paper has two empirical objectives: One, we want to test to what extent legal coercion was used to lower wages. Two, we study the importance of the planters’ economic and political inﬂuence over the government in shaping legal coercion.

To guide our empirics, we employ a simple model where workers earn a wage w that is equal to their outside option in the informal sector. Coercion C reduces this outside option, e.g., by evicting smallholders from their plots. Coercive policies C are set by the government to bene-

fit the planters as in Grossman and Helpman’s ‘Protection for Sale’ framework (1994), and the government chooses a higher C when planters’ influence is greater.2 Planter influence depends on the number of plantations on the island (N), which in turn depends on the extent to which

Caribbean plantations offer higher returns to British investors than returns obtainable elsewhere.

In our model a shrinking plantation sector leads to lower levels of coercive activity which raises wages: N C w. Exogenous factors act as an instrument for N. → → The data are collected from the British Colonial Ofﬁce’s Blue Books: wages wit are reported as the annual average of the agricultural spot market wage. Cit is measured by incarceration rates per capita. Nit is measured as the share of sugar in total exports or as the share of all plantation crops in total exports. Our empirical focus is on within-colony over-time panel variation.

Our hypothesis is that where the plantation system went into decline, Cit decreased and wit increased relative to other islands. A generalized difference-in-differences strategy robustly bears this out across a range of different speciﬁcations. We instrument for Nit using exogenous variation in the returns to investing in Caribbean plantations relative to investing elsewhere. Across islands, returns in the Caribbean were lower where smallholders could easily escape work on plantations.

Some colonies had large hinterlands while others had next to none.3 Over time, returns to investing outside of the Caribbean improved rapidly over the 19th century. As the Empire expanded,

2One may think of planter influence as lobbying capacity but this is not explicitly modelled. Our own view is that legal coercion is in practice almost always the result of collective action by an elite that influences the state to regulate coercive labor laws to their benefit, e.g., to reduce rights-at-work, to harass workers in the informal sector or to limit worker mobility, with the ‘Black Codes’ in the post-Bellum U.S. South being a prominent example. In section2, we describe in detail the practical applications of legal coercion in our context. 3 During slavery, this difference did not matter: Land that was unsuitable for sugar lay uncultivated even if it was very fertile. After slavery, the availability of hinterlands became important, eroding Nit by making it easier for freed slaves to evade the plantation system. This basic explanation of the divergent post-Emancipation fortunes of ex-slaves across the islands figures prominently in the literature on Caribbean history and indeed it was anticipated in the 1830s debates surrounding Emancipation, see Merivale(1861, 312–317), Engerman(1984, 137), Richardson(1997, 134–135, 157–8), and Patterson(2013).

2 opportunities to export British manufactures grew much faster than opportunities in Caribbean agriculture. and English investors turned their attentions away from the Caribbean. Combin- ing these exogenous elements, our instrument Oit is the product of the cross-section of islands’ hinterlands and the time-series of English exports to non-Caribbean destinations. In the data, a higher Oit led to a decline in Nit and Cit and a rise in wages. When we instrument for Nit we get two-stage least-squares (TSLS) estimates that are highly statistically significant and economically large: A fifty percent decline in the plantation system—roughly the average over the 76 years we study—would have increased wages by about fifty percent and reduced incarceration rates by close to their mean of one in a hundred people. These results hold up under an extensive set of robustness checks.

We also provide evidence on the mechanisms linking N, C, and w. A causal mediation analysis shows that three-quarters of the negative effect of the plantation system on wages operated via measured coercion, prima facie evidence for our hypothesis. To lend support to our assumption that coercion was ‘legal coercion’ exercised by the government in response to pressure from the plantation system, we study the effects of an exogenous shock which changed the composition of the plantation-owning elite, while keeping the size of the plantation sector constant. This shock strengthened the plantation system’s lobbying power with each island’s government apparatus, but severed the localized ties between the previous plantation owners and the parochial constab- ulary and judges. We ﬁnd that this shock increased coercive legislation, which was set centrally, but decreased incarceration, which depended on these parochial ties. This evidence suggests that legal coercion adjusted when the planters’ form of inﬂuence changed.

We conclude this introduction with a literature review. The importance of using of the legislative, judicial, and policing powers of the state to reduce the outside options of workers and to beneﬁt a small elite is emphasized by Lewis(1954) and also in Acemoglu and Robinson’s (2012, ch. 9) study of Apartheid. In the focus on the manipulation of workers’ outside options, we relate to Alston and Ferrie’s account of Southern Paternalism (1993). The closest empirical study to ours is Bobonis and Morrow(2014) who show that coffee price shocks in Puerto Rico in the mid-1800s led elites to reduce human capital investments so as to depress plantation workers’ outside options. We also connect to a large literature on labor coercion. This literature is more focused on the coercion of workers on the job (Chwe, 1990; Basu, 1999; Bloch and Rao, 2002; Naidu and Yucht-

3 man, 2013); however, in a principal-agent framework, there is a complementary between coercing workers on the job, and reducing their outside options as non-workers (Acemoglu and Wolitzky,

2011).4 In our setting, the latter form of coercion was clearly dominant.

In an interesting counter-point to our ﬁndings, Dell and Olken(2019) show that the location of Dutch colonial sugar plantations on the island of Java had positive long-run effects on local infrastructure and economic development. This line of research suggests that there are potentially positive and non-institutional long-run consequences of colonial extraction. Lastly, our focus on the political consequences of geographic variation in the ability to evade the plantation system connects us to a broader literature on the institutional consequences of geography; see e.g Enger- man and Sokoloff(2002); Acemoglu et al.(2002). 5

2 History

British West Indies plantations were a source of vast proﬁts for British planters and investors.

These proﬁts were dealt a severe blow by the emancipation of slaves in 1838. Emancipation initially led to sharply rising wages as freed slaves rejected plantation life in favour of squatting on abandoned estates or small plots in the hinterland. This ‘ﬂight off the estates’ did not last long.

Within a few years of Emancipation the white planter elite had developed a system of legal coercion over labor that lowered wages and slowed the demise of the plantation system. We now describe the workings of this system.6

2.1 Legal Coercion Cit

Legal coercion in our setting took three main forms. First, coercion restricted access to land. The full force of the law was brought to bear on peasants who attempted to squat on abandoned estates or Crown land. Squatting was so rampant that it seriously undermined the ability of planters to keep peasants on plantations. In Jamaica there were 10,000 squatters by 1844 and this number probably climbed to 40,000 by the mid-1860s (Eisner, 1961, 215–216). The Colonial Blue Books list

4 Lowering wages in the bad state in order to induce effort (i.e. relax workers’ incentive compatibility constraint) is easier if legal coercion simultaneously prevents the worker from walking away (i.e. relaxes the participation constraint). 5 Lastly, we naturally connect to a large literature on the Caribbean’s economic adjustment to Emancipation (Eisner, 1961; Engerman, 1982, 1984). 6 See Merivale(1861, 340–341), Engerman(1984, 134 and Table 2) and Riviere(1972, 13). On the ‘ﬂight’ see Hall (1978, 7), Engerman(1982, 199) and Green(1976, 174–175, 198).

4 the titles of all colonial statutes and a quick perusal shows that every colony repeatedly enacted and strengthened trespass and vagrancy laws in order to prevent squatting. The salience of the squatting-incarceration issue is illustrated by Jamaica’s Morant Bay Rebellion, which left 600 dead and many more imprisoned (Underhill, 1895; Craton, 1988).7 Other examples abound: In Do- minica, the ‘Queen’s Three Chains’ unrest of 1856 was driven by a dispute over whether several black families had implicit title to or were squatting on the land they were farming. It was resolved by the rapid dispatch of troops from nearby Antigua (Honychurch, 1984, 136-138). The ‘Toll Bar

Riot’ and the ‘Florence Hall Riot’ in Jamaica were triggered respectively by objections over limiting smallholders’ access to water sources, and over the sentencing of squatters for trespassing on an abandoned estate. See Cundall(1906, 5–12). 8 A feature of land conﬂicts in the post-Emancipation

Caribbean was that they resulted from planters’ attempts to reduce peasant smallholders’ outside options. Planters were not grabbing the land for themselves. In fact, there was plenty of fallow plantation land during the period we study. For example, in mid-century Jamaica, less than a quarter of all land was under cultivation and, even on the active plantations, less than half of the plantation land was in use. That is, disputes were not over land that could be or ever was used for plantation crops. See Satchell(1990, chapter 4 and especially, 63-64). Land conﬂict in our case was thus different from historical episodes in other parts of the world where planters cultivated land stolen from peasants. See for exampleS anchez,´ del Pilar Lopez-Uribe´ and Fazio(2010) on

Colombia’s land conﬂicts 1850–1925 and Acemoglu and Robinson(2012, ch. 9) on Guatemalan’s coffee land grabs during the same period.

Second, coercion was manifest in asymmetric terms of employment contracts. If a worker started employment without a formal contract the laws of many colonies stated that he or she had implicitly entered into a one-month contract. Failure to work during that month was a ‘breach of contract’ that resulted in ﬁnes or imprisonment (House of Commons Parliamentary Papers, 1839b,

7 By 1865, a number of villages had been established illegally on Crown lands in the hills above Morant Bay. Tensions ran high as the government sought to limit further expansion of these villages. Things came to a head during a trespass case involving a villager who had been pasturing on an abandoned estate (Underhill, 1895, 59). A crowd gathered at the courthouse, violence broke out, and then quickly ignited all of Jamaica. 8 Planters also lobbied for a host of restrictions which limited worker access to affordable land with clear legal title. Large tracts of Crown land were kept off the market, made available only at artiﬁcially high prices, or sold only in large lot sizes, e.g., Craton(1997, 390–393). For example, 83% of Trinidad’s landmass was owned by the Crown, yet was kept off the market for decades after Emancipation (Sewell, 1861, 103, 106, 133). Also, in many colonies peasants were prohibited from pooling their resources to buy plantations and bankrupt planters were pressured not to sell to smallholders (Eisner 1961, 211, Craton 1997, 390).

5 205–206). If in addition to wages a worker also received a cottage and a plot for growing crops, he or she was obligated to work on the plantation for a full year. The law allowed the planter to evict a peasant for absenteeism (‘breach of contract’) and threats of eviction were very effective in forcing peasants back to work.9 Laws additionally allowed planters to burn or conﬁscate cottages and crops of evicted peasants.10 The resulting destruction led to retribution by peasants who could then be sentenced to lengthy imprisonment for ‘malicious injury to property.’

Third, coercion was manifest in a tax system that penalized smallholders. This was most apparent in the biased taxation of imports. Planters imported flour and rice to feed plantation workers and these imports competed directly with smallhold crops, making smallholding less attractive. High tariffs on foodstuffs were therefore “opposed by the estate interests since they tended to deplete labor reserves by driving workers from plantations to the hinterland, where they grew ground provisions” (Rogers, 1970, 96). Green(1976, 186) similarly states that there was much political conflict “over import duties on food, [which] enticed freedmen to abandon estate labor in favor of the production and sale of provisions.” Property taxation was also biased against smallholders. A smallholder with five acres could pay higher taxes than a planter with 500 acres.

Not only did such abusive smallhold taxes reduce the returns to smallholding, they also led to punitive loss of title. For example, Satchell(1990, ch. 4 and Table 4.3) documents that 18,000 acres of Jamaican smallholds were repossessed after 1869 for failure to pay taxes. Many other discriminatory taxes have been documented, including export taxes that were higher on smallhold crops than on sugar, e.g., Underhill(1895, xvii). Dookhan(1975, 156) emphasizes the importance of such regressive taxes, arguing in particular that the 1853 ‘Chateau Belair’ Riots in the Virgin Islands were caused by a new tax on cattle. In his words, “as cattle rearing was primarily the occupation of the rural negro population, this tax fell principally on them.”

Legal coercion was a fact of life in the British West Indies. Its role was simple: Reduce the returns to smallholding so as to encourage peasants to work on the plantations for low wages.

Restated in more theoretical language, legal coercion did not affect plantation workers directly; rather, it affected them indirectly by reducing their outside options.

9 See House of Commons Parliamentary Papers(1839b, 131, 134–136), Bolland(1981, 595), Dookhan(1975, 130), and Brizan(1984, 128). 10 These laws were repeatedly criticized by the Governor of the Leewards in 1838 on the grounds that they were inequitable and unconstitutional, e.g., House of Commons Parliamentary Papers(1839b, 61–62).

6 2.2 The Government Apparatus

The three pillars of the law — lawmakers, judges and police — were all controlled by planters.

To gauge their coercive effects, it is useful to apply Acemoglu and Robinson’s (2008) distinction between officially sanctioned de jure power and unofficially sanctioned de facto power. Planters’ de jure power was exercised primarily through their influence on discriminatory laws passed by legislators in the islands’ legislative assemblies.11 Planters’ de facto power was exercised primarily through the selective application of these laws in different locations: rural magistrates did much of their work on plantations, had little legal training and were frequently former plantation overseers (McLewin, 1987, 85–87). Local police were also beholden to planters. The first post-Abolition laws constituting the police force (‘Police Acts’) stated that rural police were to be appointed by planters. The Leewards Governor complained that this was unconstitutional, but found it difficult to pressure planter-dominated legislatures into changing the laws (House of Commons Parliamen- tary Papers, 1839b, 49).

2.3 Planters’ Relative Political Power Nit

We are interested in the impact of the planter elite’s political power Nit on coercion Cit. The previous section described Cit. We now describe Nit. Planters’ political power was based on their economic power relative to the peasantry. Economic power in the Caribbean hinged on exports.

Marshall(1968, 253–254) concludes from his survey of the British West Indies that the period from roughly 1850 to 1900 was one of “continuing expansion of the number of peasants and, more important, a marked shift by the peasants to export crop production.” This growing peasant participation in exports was clearly accompanied by the declining acreage of plantations, and the rising number of freeholders and squatters (Riviere, 1972, 15-17). Eisner(1961, 235) argues that the “increasing prosperity of the peasantry is thus seen to be mainly due to their growing share in export crops.”12 Our measure of the relative economic power of the plantation system is thus the

11 Carvalho and Dippel(2016) show that planters completely dominated colonial legislatures, especially in the early years of Emancipation. 12Eisner(1961, 220, 221, and 234) has ﬁne-grained data on Jamaican smallholds and peasant exports. Our own calculations show that between 1850 and 1890 the share of Jamaican exports originating from freeholds and squatters rose spectacularly from 10.4% to 39.0%.

7 share of plantation crops in total exports.13

Sugar is consistently identiﬁed with a plantation mode of production and has often been argued to be detrimental to economic and social development, e.g., Sokoloff and Engerman(2000) and Easterly(2007). 14 In the British West Indies, sugar was also by far and away the dominant plantation crop, completely eclipsing all other crops such as cotton or coffee. We therefore use the share of sugar in total exports as one measure of the strength of the planter elite relative to the peasantry.

Figure1 displays the lowess-smoothed share of sugar in total exports by colony. It is best to focus on the two dominant features. First, in 1838 every colony was highly specialized in sugar.

Second, by 1913 there were substantial cross-colony differences in sugar export shares. Colonies roughly divided into three groups. Five colonies remained heavily involved in sugar for the entire period (Antigua, Barbados, Guyana, St. Kitts, and Nevis). Four colonies saw sugar decline to less than half of total exports (St. Lucia, Trinidad, Tobago, and Jamaica). Five colonies exited sugar entirely by the end of period (Virgin Islands, Grenada, Dominca, St. Vincent, and Montserrat). In

Figure1 and the econometric analysis below we use lowess-smoothed export shares because we are interested in capturing long-run changes in the strength of the plantation system rather than short-run agricultural ﬂuctuations.15

The data displayed in ﬁgure1 was intimately linked to the planter’s power and inﬂuence over government, in particular their success in securing plantation labor at low wages.16 Dookhan

(1977, 11) argues that across the Caribbean, laborer “drift from the estates” was lowest in Antigua,

13 An acreage-based measure would be an appealing alternative, but the data are only sporadically available. In On- line Appendix Table 12 we provide some evidence that acreage-based and export-based measures are highly correlated. 14 Sugar was unequivocally a plantation crop (Engerman, 1983). However, there is not complete consensus in the literature about the factors that made it so and such a discussion is beyond the scope of this paper. However, we conjecture that three factors are important. (1) The sugar mill was a major capital asset that was beyond the financial reach of all but the richest members of Caribbean society (Marshall, 1996, 73; Lobdell, 1996, 322, 326). (2) Sugar must be processed within hours of harvesting so that there was always a sugar mill either on the plantation or nearby. See Higman’s (2001, Figure 2.5) map of Jamaican mills. (3) Labor demand during the sugar harvest was physically brutal (e.g., 90-hour work weeks) and conflicted with workers’ needs to harvest their own provision grounds (Higman, 1984, 182–183). These factors favoured a system of production that vertically integrated harvesting with milling at a single location (the plantation) and which, in the racialized post-Emancipation period, used coercion rather than overtime pay as an incentive device, i.e., these three factors favoured a plantation system. 15The lowess smoothing faithfully reproduces the annual data. See Online Appendix Figure 1. 16 While we do not focus on this, it is worth noting that the evolution of plantations displayed in figure1 also correlated tightly with the share of British Whites in the Caribbean at the time. To European observers at the time, the exodus of whites was synonymous with the decline of the plantation system and the decline of the institutions that had until then characterized the British West Indies. We discuss this in detail in Online Appendix B. See also Carvalho and Dippel(2016).

8 Figure 1: The Differential Decline of the Plantation Economy

The Share of Sugar in Total Exports 1.0

0.9 Antigua Barbados 0.8 Guyana 0.7 St. Kitts 0.6 Tobago St. Lucia 0.5

0.4 Jamaica Trinidad 0.3

0.2 Share of Sugar in Total Exports Exports Total of Sugar in Share 0.1 Virgin Is. Grenada Dominica St. Vincent 0.0 Montserrat 1838 1853 1868 1883 1898 1913

Notes: This ﬁgure reports the share of sugar in total exports Nevis is not reported because it stayed above 0.99 in each panel. Also, Nevis merged with larger St. Kitts in 1883 and Tobago merged with larger Trinidad in 1899.

St. Kitts and Barbados, i.e., in precisely the three islands where the plantation system survived the longest. Green(1976, 258) similarly argues that the success of Barbados’ plantations was “not due to new technology but rather due to a large and disciplined labor pool.”

Over time, some planters shifted to other plantation crops, which were similar to sugar in that they also incentivized labor coercion. To handle this shift in coercive crops, we construct the share of all plantation-produced goods in total exports as a second measure of elite power. Using a 76-year panel of exports by colony and crop, we used historical accounts to code up the share of each crop’s production accounted for by wage-paying plantations, P lgit [0, 1], and used this to ∈ calculate the share of plantation-crop exports in total exports. The detailed coding is described in

Online Appendix A.17

17 All available evidence strongly suggests that the share of other plantation crops is indeed strongly and negatively correlated with the number of smallholders and smallholder participation in exports. See, for example, Dookhan(1977, 136).

9 2.4 Exogenous Drivers of the Strength of the Plantation System

The preceding discussion of Caribbean history suggests regressions of wages and coercion on the power of the planters. There is naturally a concern that the varying power of the planters displayed in Figure1 was endogenous to factors that may also have affected wages and coercion directly. It is therefore desirable to identify drivers of Nit that were exogenous in the sense of im- pacting wages and coercion only through Nit. Our ﬁrst candidate for such a driver was a source of cross-sectional variation across colonies that became very salient to peasants after Emancipation.

In a colony like Barbados, all of the land was sugar-suitable and in 1838 only 4% of land was not under cultivation (1838 Barbados Blue Book). By contrast, a colony like Jamaica had sugar-suitable coastal plains but a higher-elevation interior (which was fertile but not sugar-suitable). During slavery this difference between Barbados and Jamaica did not matter because sugar-suitable land was similarly suitable everywhere and the hinterland was largely inaccessible to slaves and not used by many others. After Emancipation, differences in the availability of a fertile but sugar- unsuitable hinterland led to stark differences in the ease with which peasants could evade the plantation system and resist coercion (Engerman, 1984; Richardson, 1997; Patterson, 2013, respectively 137, 134–135, 157–8).18

To measure this differential ability to evade the planters we carefully calculated the share of

19 each colony’s land that is unsuitable for sugar cane, Oi. The relationship between Oi and the historical outside options that peasants actually had is visually displayed in Figure2 for Jamaica, the only colony with historical maps on evolving land-use patterns. In the left panel, the black areas are lands that we have coded as highly suitable for sugar cane. In the right panel, the shaded areas (both black and grey) are plantations in 1790. The two are spatially correlated and, averaging across Jamaica, the share of land under plantation in 1790 is very close to the share of

20 highly sugar-suitable land. This map and all the post-Emancipation history make clear that Oi is a good proxy for the amount of hinterlands available to freed slaves after 1838. We note that Oi is

18Engerman(1984, 137) quotes a Jamaican planter in the 1830s as arguing that Emancipation “will be less mischievous to other colonies than ours. For in Barbados and Antigua and several other Islands the liberated slaves must work for wages or want the necessaries of life.” 19We defer data details to section4 and the appendix sections cited therein. 20The right panel of Figure2 illustrates another point. The grey-shaded areas are plantations that shut down between 1790 and 1890. These were the lands that were most difﬁcult to keep out of peasant hands and were thus a major focus of coercive interventions.

10 Figure 2: Sugar Suitable Land in Jamaica (Left), and Plantations in 1790 and 1890

Notes: The left panel shows the spatial distribution of land that is sugar-suitable (black), only moderately sugar-suitable (dark grey), not sugar-suitable (light grey) and totally sugar-unsuitable (white). See Online Appendix C for details. The right panel shows the extent of sugar plantations in 1790 (black plus grey areas) and 1890 (black areas). There is a good match between our estimate of sugar-suitable land in the left panel, and the historical sugar plantation land in Jamaica. The left panel is based on authors’ calculations. The right panel is a digitized version of Higman’s (2001) remarkable Figure 2.9.

an extremely good predictor of the long-run survival of the plantation system. This point is made visually in Online Appendix Figure 7: The higher is Oi, the smaller is the 1913 sugar export share.

In the 18th century British investors made their fortunes in Caribbean sugar. During the 19th century, however, investment opportunities shifted to other regions. Panel (a) of Figure3 displays the per capita GDPs of Britain’s largest export destinations at the start of our sample. Data are from an update of the Madisson data (Bolt, Inklaar, de Jong and van Zanden, 2018). The only

British West Indies colony with the Madisson data is Jamaica. See the solid line at the bottom of the panel. The ﬁgure shows that while Jamaica languished, the majority of Britain’s largest export destinations rapidly grew richer.

Exogenous events elsewhere around the globe were overtaking Jamaica and drawing away the attention of British investors: The repeal of the Corn Laws in 1846 gave a tremendous impetus to free trade around the world, and started what eventually came to be called the ‘ﬁrst globalization’, which, buttressed by a precipitous decline in ocean freight rates, lasted until 1914 (North, 1958;

O’Rourke and Williamson, 2001, 2002).21 In this period, the Caribbean was left behind partly because the repeal of the Corn included provisions for a phasing-out of preferential tariffs on

West Indies sugar over the period 1846–54 (Curtin, 1954).

This globalization-driven boon and shifting-away of British attention from the Caribbean to-

21 Falling ocean freight rates led to a shift from trade in only high unit-value products like sugar towards trade in many lower unit value products like heavy machinery exports to the U.S. and Europe.

11 Figure 3: British Markets and British Exports by Destination (a) per Capita GDP (b) British Exports by Region 10000 Ausl 2% 500

USA 12% 8000 400 Can 6%

Ger 13% 6000 300

Nld 6%

Fra 5% 4000 200 per capita GDP ($) Exports excluding Caribbean Bra 10% British Exports (in 1913 £ millions) 100

2000 Jam 7%

India 10% Exports to Caribbean 0 0

1850 1871 1892 1913 1838 1853 1868 1883 1898 1913

Notes: Panel (a) is per capita GDP (1990 international dollars) from Madisson, 1850–1913. (Earlier data are not consistently available.) The countries included were the top destinations for British exports in 1846 (the earliest date available for exports by country) and accounted for 71% of all British exports. The legend reports 1846 export shares, e.g., Ger- many received 13% of all British exports in 1846. Panel (b) is British exports to the Caribbean and British exports excluding the Caribbean. These exports are in constant 1913 pounds sterling (£). All exports exclude re-exports.

wards the rest of the world can be seen in export volumes. Panel (b) plots British exports to the

Caribbean and to all non-Caribbean destinations. Data are in constant 1913 pounds sterling (£) and purged of re-exports (e.g., Jamaican sugar that is shipped ﬁrst to London and then to France).

The underlying data are from Mitchell(1988) and detailed in Appendix B.5. British exports to non-Caribbean destinations displays periods of rapid growth followed by shorter periods of stag- nation (e.g., the 1890s) and infrequent bursts of decline. However, the overall picture is one of rapid growth, with real exports growing at an annually compounded rate of 3.4%. In comparison, British exports to the Caribbean experienced practically no growth. Judged by exports then,

British investors had largely turned their focus away from the Caribbean in response to exogenously improving opportunities elsewhere.

We have focussed on two exogenous drivers of the strength of the plantation system that will be useful for our identiﬁcation strategy. Across colonies, coercion on plantations was more difﬁcult where the hinterland was large. Over time, the plantation system declined in part because British investors shifted their attention away from Caribbean sugar and towards opportunities in regions that were experiencing rapid growth.

12 2.5 Other Major Drivers of the Plantation System’s Strength

2.5.1 Crop Price Movements

One challenge for identiﬁcation is that pressures on the coercive plantation system came in part from variation in smallhold crop prices. In fact, there was a secular decline in plantation crop prices relative to smallholder crop prices in the Caribbean during this time.22 Higher smallhold crop prices would have raised smallhold exports in revenue terms and therefore reduced Nit, while they would have also increased smallhold proﬁts, and by extension wages, even if planters’ actual power over governance had been unaffected. We therefore carefully control for changes in the export prices of smallhold crops.23 We also note the importance of inspecting the responses of

Cit to Nit in addition to the wage responses.

We are also concerned with the fact that crop choices were endogenous, irrespective of the fact that world crop prices were plausibly exogenous to Caribbean production.24 Smallholders’ proﬁts rose more if they substituted towards crops whose prices increased. The impact of the crop price changes on smallholder proﬁts in turn was biggest in places where the crops with the best geographic suitability and the steepest price increases coincided. We therefore embed a model of crop choice into our theory, and include into our empirics an exogenous crop-suitability-driven export-price basket, estimated as in Costinot, Donaldson and Smith(2016). 25

2.5.2 Labor Supply Shocks

Crop prices were not the only factor shaping plantation labor supply during this period. Fortu- nately, Caribbean history is quite clear on what the other big labor supply shocks were during this period: the immigration of indentured East Indian Immigrants (Laurence, 1971; Riviere, 1972) and the construction of the Panama Canal (Maurer and Yu, 2013, ch.4).

22This decline in plantation prices, particularly in sugar, was important. A previous version of this paper (Dippel, Greif and Treﬂer, 2015) was about the impact of international trade (declining sugar prices) on wages, coercion, and institutions. While we continue to carefully control for prices, they are no longer the focus of this paper. 23We additionally control for other important changes in plantation labor supply in the Caribbean during this time, namely the immigration of East Indian laborers and work opportunities from the construction of the Panama Canal. See Appendix B.6. 24World crop prices were exogenous to Caribbean production. Even for sugar, at their production peak in the early years of the data, the British West Indies produced only 18% of world sugar output and this number fell to 1% by the end of the sample. See Online Appendix Figure 4. 25 See Appendix Appendix B.2.

13 2.5.3 Hurricanes

Finally, we note that the Virgin Islands are an outlier in Figure Online Appendix Figure 7: the plantation system had collapsed by 1913 despite low values of Oi. Nowhere else in the Caribbean did planter power diminish as rapidly as in the Virgin Islands. See, for example, Online Appendix

Table 1. As it turns out, the Virgin Islands plantation system collapsed because of two major hurricanes in 1848 and 1852, which destroyed the colony’s sugar infrastructure and left planters too indebted to rebuild. Hurricanes do two types of damage: They destroy crops and they destroy structures such as sugar mills. Since sugar cane must be processed within hours of harvesting and since cane is difﬁcult to transport, there was always a sugar mill either on the plantation or nearby, e.g., Higman(2001, Figure 2.5). Sugar mills were unique in Caribbean agriculture in that they were expensive and long-lived assets. They were also prone to hurricane damage. In the post-Emancipation Caribbean, an increasing share of planters could only cover their variable costs but not their ﬁxed costs. In other words, it made sense for a marginal planter to operate an existing mill, but not to rebuild a destroyed mill (Marshall 1996, 73; Lobdell 1996, 322, 326). As a result, hurricane landfalls that destroyed mills had long-lasting effects. To control for hurricane damage, we geo-referenced the paths of every major hurricane that hit the Caribbean between 1838 and

1913, and assign each Caribbean hurricane landfall an island-speciﬁc damage index HDIit. Data sources, the list of all hurricanes and their measured impact appear in Online Appendix E.

3 A Simple Model of Coercive Labor Market Institutions

To ﬁx ideas and provide additional motivation for the empirics, we now turn to a simple theory of coercive labor markets.

3.1 Technology and Crop Choice

There is an exogenous measure L of workers (former slaves) and an endogenous measure N of planters (members of the planter elite). There is a continuum of heterogeneous plots indexed by

ω, each of which can be planted in one of g = 1,...,G crops. We follow Costinot et al.(2016) in modelling crop choice by assuming that plot ω planted in crop g has a baseline yield of zg(ω)

−θ Tgz where zg(ω) is a random variable with a Frechet´ distribution: P r zg(ω) < z = e . On a { } −

14 p p plantation, plot ω combined with one worker produces output τg zg(ω) where τg 0 describes ≥ p the efﬁciency of plantation agriculture, e.g., τg is large for sugar and small for livestock. On a

s s smallhold, plot ω combined with one worker produces output τ zg(ω) where τ 0 describes g g ≥ p s the efﬁciency of smallhold agriculture. The crop-speciﬁc τg and τg explain why some crops are better-suited than others for plantation agriculture.

We consider a small open economy so that crop prices p = (p1, . . . , pG) are exogenous. Crops j are chosen to solve maxg pgτg zg(ω) where j = p if it is a plantation plot and j = s if it is a smallhold plot. The optimal choice varies across plots, but on average the expected revenue per plot will be

1 j j j θ θ r(p, τ ) = E max pgτ zg(ω) = ΣkTk(τ pk) Γ , j = p, s (1) g g k where τ j = (τ j, . . . , τ j ) and Γ = Γ(1/θ 1) is the gamma function. See Appendix A for the proof 1 G − or Costinot et al.(2016, 215). r(p, τ j) captures how crop choices respond to prices.

Each smallholder is randomly allocated one plot and each planter is randomly allocated l(N) ≥ 1 plots. Since each plot uses one worker, the maximum number of planters is N = L and when

N = L each planter receives one plot, i.e., l(L) = 1. We also assume that the more planters there ∂ ln l(N)N ∂ ln l(N) are the more land they receive collectively ( ∂ ln N > 0), but not individually ( ∂ ln N < 0). The latter creates a ‘congestion cost’ which ensures that not all agriculture is plantation agriculture.

3.2 The Worker’s Occupational Choice and Coercion

Each smallholder must choose between plantation work and smallholding. Utility from working on the plantation is w.26 Utility from smallholding is r(p, τ s) C where C is the negative impact − of planters’ legal coercion on the returns to smallholding. C is endogenous. It follows that in any equilibrium with both plantation and smallhold agriculture,

w = r(p, τ s) C. (2) − r(p, τ s) captures how wages respond to prices when crop choices are endogenous.

The costs of coercion (e.g., building jails) are given by Cγ where γ > 1. These costs are funded

26By equating utility with income we are implicitly assuming that only the numeraire good is consumed and that all other goods are exported.

15 by a head tax on planters of Cγ/N. Consider planter proﬁts. When there are N planters, each receives l(N) plots, earns per plot revenues of r(p, τ p), pays per plot wages of r(p, τ s) C and is − left with proﬁts of

π(C,N) = l(N)[r(p, τ p) r(p, τ s) + C] Cγ/N . (3) − −

We use Grossman and Helpman’s (1994) ‘Protection for Sale’ framework to determine the level of coercion C. C 0 is chosen to maximize a weighted sum of the proﬁts of the N planters and ≥ the L workers:

W (C) = αNπ + Lw . (4)

α is the weight given to planters’ proﬁts. Our key assumption is that α > 1 so that planters have greater sway over the choice of coercion. Substituting equations (2)–(3) into (4) and maximizing with respect to C subject to C 0 yields the following characterization of optimal coercion. Let N¯ ≥ be the value of N for which αl(N)N L = 0. Under our assumptions, N¯ is unique and 0 < N¯ < L. − The optimal level of coercion is

1 αl(N)N L γ−1 C∗(N) = − for N N¯ (5) αγ ≥ and C∗(N) = 0 for N < N¯. Since the land controlled by planters is increasing in the number of planters [l(N)N is increasing in N], equation (5) implies one of our key results, namely, CN∗ > 0 when N is sufﬁciently large. The insight is simple: The stronger is the planter elite, the greater is its political inﬂuence (as measured by αN) and hence the higher is the level of coercion. Equation

(5) further implies a threshold effect: When the number of planters drops below N¯, there is no coercion. See Appendix A for proofs.27,28

27We note in passing that if α = 1 then N¯ = L so that C∗ = 0 for all N, which reflects the fact that coercion is an inefficient redistributive policy that would never be used if smallholders had equal say in choosing coercion. 28This ‘Protection for Sale’ setup abstracts away from part of the collective action problem in that the level of coercion grows with the number of planters. However, planters do not solve the bigger collective action problem, namely, that of collectively restricting entry into planting and thereby preventing profits from being driven to zero. Historically, in the median colony whites represented only 1.6% of the population so that, in the highly racialized colonial society, whites ‘stuck together.’ Thus empirically, there was no white collective action problem when it came to policies restricting black smallholders.

16 3.3 Worker Resistance and the Planter’s Entry Decision

As discussed in the history section, workers often resisted white rule, which resulted in deaths, property damage and incarceration. Since worker motivations for resistance do not enter into the empirics we model resistance simply as a probability χ that resistance is successful. χ(O) is increasing in the share of non-sugar-suitable land O because it is costly for the police and military to operate in these more remote and often highland areas.29 Without loss of generality we assume

χ(O) = O. If resistance is successful neither planters nor workers are able to plant their crops and returns to each are normalized to 0. If resistance is unsuccessful then workers and planters generate the earnings, proﬁts, and coercion levels that appear in equations (2)–(5).

We next turn to the free entry condition for planters. Each potential planter must choose between (1) staying in England where he can invest, export to non-Caribbean countries and earn W versus (2) moving to the colony where he can invest in sugar, export back to England and earn planter proﬁts. Conditional on no resistance, planter proﬁts are given by equation (3), namely,

π(N) π(C (N),N). Thus, in the colony the planter earns expected proﬁts (1 O)π(N). If ≡ ∗ − (1 O)π(N) < W for all N then no planter moves to the colony and there is only smallholding. − If (1 O)π(N) > W for all N then L planters move, each has one plot and one worker, and there − are no smallholders. We focus on the intermediate case where planters and smallholders coexist.

In that case there is an N such that (1 O)π(N ) = W or ∗ − ∗

W π(N ∗) = . (6) 1 O −

This equation pins down the equilibrium number of planters N ∗. In Appendix A we provide sufﬁcient conditions on the underlying parameters of the model for such an N ∗ to exist and to be stable in the usual sense that πN (N ∗) < 0.

In any stable equilibrium, N ∗ is decreasing in O and W . Further, N ∗ depends on the interaction of O with W . We exploit this interaction in the empirical work.30

29The classic example is the 18th century Maroons operating in the mountainous interior of Jamaica. 30 ∗ ∗ ∗ ∗ From equation (6), ∂N /∂W = πN (N )/(1 − O) < 0 because πN (N ) < 0. Likewise, ∂N /∂O < 0. While we do not need to sign ∂2N ∗/(∂O∂W ), note that if π(N ∗) is linear in N in the neighbourhood of N ∗ then an increase in O ∗ ∗ ∗ increases ∂N /∂W = πN (N )/(1 − O), i.e., a higher probability of revolt makes N more sensitive to the returns from staying in England.

17 3.4 General Equilibrium and Comparative Statics

An equilibrium in our small open economy is a crop choice for each plot ω and mode j = p, s j that solves maxg pgτg zg(ω), a wage w that leaves each smallholder indifferent between plantation work and smallholding (equation2), a level of coercion C∗(N ∗) that maximizes planter-biased so- cietal welfare (equation5), and a mass of planters N ∗ that leaves each planter indifferent between staying in England and moving to the colony (equation6).

Our main comparative statics results are as follows. First, the wage is increasing in an index of prices r(p, τ s) and decreasing in coercion C. See equation (2). Second, coercion is increasing in the number of planters. See equation (5). Third, in the absence of coercion, wages are given by w = r(p, τ s) (equation2) so that we have a benchmark for competitive wages that deals explicitly with the crop substitution problem identiﬁed in section 2.5.1. Fourth, planter strength N ∗ is decreasing in W/(1 O) and W/(1 O) has no direct impact on anything but N .31 − − ∗

4 The Evidence

Our main hypothesis is that the powerful Caribbean planter elite held enough sway over government to institute forms of legal coercion such as the incarceration of squatters, which were aimed at reducing wages. We begin by testing this hypothesis in OLS regressions of coercion (Cit) and wages (wit) on the relative economic power of the planters (Nit):

w w w w w wit = β Nit + γ Xit + λ + λ + ; (7) N · X · i t it C C C C C Cit = β Nit + γ Xit + λ + λ + , (8) N · X · i t it

where λi are colony ﬁxed effects, λt are year ﬁxed effects or year trends, and Xit are controls.

Data: Our wage data comes from the Blue Books, which report wages for ‘praedial’ or agricultural workers. This was the wage paid to plantation workers. These wages were the largest component of the cost of the most important economic activity in the colonies (sugar). It is thus not surprising that wages were a constant subject of discussion in contemporary sources. The Blue

31One could extend the model to include a market for plantation land. We have abstracted from this because land prices are not even remotely consistently available over time or across colonies.

18 Book wage data are identical to the sporadic data from reliable sources discussing wages for sugar cultivation (e.g., West India Royal Commission, 1897, 107). Further, the Blue Books themselves are sometimes the source of data quoted by contemporaries (e.g., Sewell, 1861). In short, the Blue Books reliably reported the well-known and very public data on wages for plantation work. Appendix

B.1 discusses wages at greater length.32, 33

Our coercion data are incarceration rates per capita from the Blue Books. The Books report the daily average number of prisoners, averaged over the year, for 1838–1913. These data contain incarcerations for all reasons, but as discussed below and in Appendix B.3 our best estimate is that two-thirds of incarcerations at the start of our period were associated with legal coercion.34 New incarcerations per capita (expressed as a percent) had a sample mean of 1.1%, indicating that 1.1% of the population entered jail each year. Incarceration rates are admittedly a fairly narrow measure of what was in reality a bundle of policies of legal coercion. The reason we focus on incarceration rates is that it is the only empirical counterpart to legal coercion that we could consistently code for the entirety of our wage sample. In the later years we are able to validate this measure using data on court sentencing that targeted smallholders. See appendix Appendix B.3.

In equations (7)–(8), Xit are controls for observable factors that may have directly affected wages. Smallholder returns, and therefore plantations wages, were a function of exogenous price shocks and islands’ crop-speciﬁc soil productivity; see the wage equation (2). To construct

s ri(pt, τ ), we used the Blue Books to construct a 76-year panel of exports by colony and crop— generating a database containing exports by colony and year for 17 products accounting for 98% of exports—and combined it with ﬁne-grained information on islands’ agro-climactic conditions to develop suitability indexes for the most important crops. We then estimated a Frechet-based´

s structural model of crop choice, as in Costinot et al.(2016), to recover estimates of the ri(pt, τ ).

s ri(pt, τ ) is an index of smallholder revenue based on exogenous crop suitability and it captures endogenous crop-switching in response to changes in the relative price of crops. It is an exogenous, model-based prediction. See Appendix B.2 for details.

32We work with nominal wages. The Blue Books report that the major components of the cost of living were largely imported from Britain (clothing and many staples such as flour and rice) so that all 14 colonies shared a common cost of living. It follows that the cost of living deflator is absorbed in the year fixed effects used in our regressions. 33Over 90% of the wage data are for a daily pay period and do not involve in-kind payments. Nevertheless, in Online Appendix Table 7 we show that our results are not sensitive to adjustments for pay periods or in-kind payments. 34Brizan(1984, 134) arrives at a similar number for Grenada, 1850–1870.

19 Also included in Xit are the major labor-supply shocks discussed in section 2.5.2. For this, we calculated the island-speciﬁc cumulative stock of East Indian immigrant arrivals over time.

It turns out that this ﬂow of migrants was heavily right-skewed. It only really affected Guyana,

Trinidad, and to a lesser extent Jamaica. For the Panama Canal shock, there are no destination- speciﬁc estimates of islands out-migration so we measure the shock with a time-dummy for the years of Canal construction (1881-1889 under the French and 1908–1913 under the Americans) divided by distance from island centroids to Panama. See Appendix B.6 for details.

We control for colony size using the log of population and the log of total export revenues.

We control for each colony’s time-varying British support using the number of times the colony is discussed in the British Parliament.35

OLS Results: Table1 reports OLS estimates of equations (7) and (8), i.e., the effects of Nit on wit and Cit. Nit is measured as sugar’s share of exports. Columns 1–4 show results with wit as the outcome. As a baseline, column 1 includes only a linear and quadratic time trend, and the

s Frechet-based´ index of smallholder export prices ri(pt, τ ). Consistent with the wage equation (2) in the model, we see that smallholder export prices positively impact wages. In column 2 we add the controls for Indian immigration and the construction of the Panama Canal, i.e., the two most important labor supply shocks in the Caribbean during this period. As expected, Indian immigration reduced wages in the islands.36 Proximity to the Panama Canal during the period of its construction had a positive but insignificant effect on wages. Column 3 additionally controls for factors that are likely endogenous: As expected, population growth lowered wages and higher economic activity raised wages. However, their inclusion does not affect the relationship between wages and Nit. The negative coefficient on the Hansard indicates that declines in Parliamentary discussions of a colony correlate with declines in the colony’s wage. This is useful because it rules out a potential source of reverse causality: As a colony’s wage rose, sugar became unprofitable,

Britain lost interest in the colony and stopped subsidizing its white planter elite, and so Nit fell.

That is, the Hansard result is not compatible with wit causing Nit through a British subsidy chan- nel. 35Data are from the Parliamentary Hansard and are available at https://hansard.parliament.uk/search. 36The effect of immigration was statistically signiﬁcant and economically large, but it only really affected two colonies. It depressed wages by 0.31 log points in Guyana (−0.028 × ln(230, 000)) and by about 0.15 log points in Trinidad.

20 Table 1: OLS Effect of the Plantation System (Nit) on Wages (wit) and Coercion (Cit)

Outcome w it : log wage C it : Incarceration (per Cap.) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

N it : Sugar Exports as -0.3667 -0.3633 -0.3613 -0.3831 -0.3804 0.4459 0.4577 0.4918 0.4183 0.4467 Share of Total Exports [0.0067] [0.0077] [0.0116] [0.0080] [0.0119] [0.0318] [0.0258] [0.0273] [0.0321] [0.0299] {0.017} {0.029} {0.017} {0.02} {0.017} {0.077} {0.063} {0.079} {0.071} {0.058} Frechet Smallholder Export 0.3256 0.3263 0.2578 0.1608 0.2149 0.2657

Price Indexit [0.0026] [0.0032] [0.0272] [0.2841] [0.1554] [0.1440] log (Net Immigration)it -0.0266 -0.0244 -0.0224 -0.0179 -0.0013 -0.0099 0.0065 -0.0009 [0.0023] [0.0017] [0.0110] [0.0146] [0.9325] [0.6680] [0.7071] [0.9705] log(1/[Dist to Panama])i * 0.0040 0.0036 0.0265 0.0352 -0.0185 -0.0184 0.3637 0.3860

D(1881-1889)|(1908-1913)t [0.3439] [0.4229] [0.6995] [0.5619] [0.0780] [0.0865] [0.0569] [0.0453] log(Population)it -0.1197 -0.1381 0.1940 0.1412 [0.2864] [0.1480] [0.3893] [0.5839] log(Value Total Exports)it 0.0716 0.0524 -0.0202 0.0185 [0.0628] [0.2298] [0.8467] [0.8569] log Hansard-mentionsit 0.0019 0.0096 0.0076 -0.0157 [0.7089] [0.1592] [0.6690] [0.4407]

Time Controls t + t2 t + t2 t + t2 t-fe t-fe t + t2 t + t2 t + t2 t-fe t-fe Observations 908 908 908 908 908 798 798 798 798 798 R-squared 0.711 0.724 0.728 0.768 0.771 0.511 0.516 0.518 0.572 0.574

s Notes: Column 1 includes a quadratic time trend and the Frechet-based´ index of smallholder export prices ri(pt, τ ). Column 2 adds the most important labor supply shocks in the Caribbean during this period. Column 3 additionally controls for the log of population (a proxy for size), the log of total exports (a proxy for economic size), and Parlia- mentary Hansard mentions (a proxy for British interest in the colony). Column 4 replaces the quadratic time trend with year ﬁxed effects, and drops the export price index. Column 5 adds the same controls as column 3. In columns 6–10, we repeat the same speciﬁcations for incarceration as the outcome. Standard errors are clustered by colony and the corresponding p-values are in square brackets. For the main regressor of interest, p-values for wild-bootstrapped standard errors are shown in braces.

21 In column 4 we use year fixed effects instead of the quadratic time trend. With year fixed effects the export price index is never close to significant. We therefore do not include it in any regressions with year fixed effects. This is done for expositional clarity and its exclusion has no bearing whatsoever on any of the other coefficients. Column 5 again adds the endogenous controls. Summarizing columns 1–5, the partial correlation between changes Nit and wages is highly

ˆw signiﬁcant, very stable and appears economically large. The estimate βN in column 4 says that in a colony like Grenada where sugar’s share in exports had been reduced to zero by the end of the period, wages had increased by about 38% more over the 76 years than in a colony like Barbados where sugar’s share in exports remained close to one at the end of the period.

The reader may worry about inference: We always cluster standard errors at the colony level, as this is intuitively appealing for our panel setting. However, with only 14 clusters we are naturally concerned about the asymptotic theory underlying standard clustering approaches (Moul- ton, 1986). For our core causal coefﬁcients, we therefore also bootstrap standard errors using the wild bootstrap, e.g. Cameron and Miller(2008); Davidson and MacKinnon(2010). Throughout the paper we report p-values for standard errors clustered by colony on all coefﬁcients, and for our core regressors of interest additionally report those for the wild bootstrap in braces.

In columns 6–10, we repeat these specifications for Cit as the outcome. The resulting estimates suggest that a complete collapse of the plantation system (from 1 to 0) is associated with a decrease in incarceration rates of about one person per two-hundred in a year, roughly one half the mean incarceration rate in the data.37 Consistent with the model, the labour-supply (wage) controls have no direct effect on coercion. Only the Panama canal control is significant, but switches signs across specifications. Our model, our intuition and our regression results all tell us that labor supply shocks do not effect incarceration.

In the remainder of the paper, we estimate all results with two core specifications. The first corresponds to columns 2 and 7. The second corresponds to columns 4 and 9. That is, we include either a quadratic time trend or year fixed effects, but omit the potentially endogenous controls for population, total exports and the Hansard. (Including these endogenous controls alters none of the results.)

In Table2 we report the corresponding coefﬁcients on Nit where Nit is now measured as plan-

37To put this number in context, in the United States in 2016, the stock of incarcerated was 0.7 persons per hundred.

22 Table 2: OLS Effect of Sugar-Plantations on Wages (wit) and Coercion (Cit)

Outcome w it : log wage C it : Incarceration (per Cap.) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

N it : Plantation Exports as -0.5076 -0.5131 -0.5084 -0.5504 -0.5391 0.6923 0.6971 0.6963 0.6451 0.6458 Share of Total Exports [0.0030][0.0026][0.0033] [0.0028] [0.0037] [0.0140][0.0148][0.0148] [0.0184] [0.0161] {0.018} {0.031} {0.018} {0.025} {0.026} {0.02} {0.017} {0.02} {0.028} {0.03}

Time Controls t + t2 t + t2 t + t2 t-fe t-fe t + t2 t + t2 t + t2 t-fe t-fe Observations 908 908 908 908 908 798 798 798 798 798 R-squared 0.717 0.730 0.734 0.776 0.778 0.518 0.522 0.523 0.578 0.578

Notes: This table replicates Table1 with Nit instead measured as plantation crops’ export share.

tation crops’ export share. The results are similarly robust and stable in magnitude. Across the board, the coefﬁcients in Panel B are about one third larger than in Panel A, but the economic effect is almost the same because sugar’s share in exports dropped less than the share of all plantation

ˆw crops combined. The estimate βN in column 4 says that in a colony like Grenada where plantation crops’ share in exports had been reduced to about one-quarter by the end of the period, wages had increased by about 38% (0.55 0.75) more over the 76 years than in a colony like Barbados × where the plantation system continued to completely dominate.

Endogeneity of Nit aside, the reader may have a number of concerns with Tables1 and2. One potential concern pertains to the claim (and in the model, our assumption) that agricultural wages are only paid in the plantation sector. We validate this claim in Appendix B.1 where we provide prima facie evidence that wage employment was the domain of the plantation system, and that the only common alternative for plantation workers was independent smallhold farming, as opposed to wage labor outside of agriculture.

Another potential concern pertains to the accuracy of the reported wage data. Wages sometimes included an in-kind component that usually involved a cottage and a small plot. For two- thirds of our sample the Blue Books indicate whether the wage includes in-kind payments. Only 9% of these observations include in-kind payments. Nevertheless, Online Appendix Table 7 shows that our results are robust to an adjustment for in-kind payments. Also, wages are almost always reported as daily, but for 5% of our observations data are reported as weekly or monthly. When this happens, the Blue Books explicit indicate that the data are for 5-day weeks or 20-day months

23 and so are easily converted to daily wages. Nevertheless, Online Appendix Table 7 shows that our results are robust to adjustments for alternative conversions of weekly and monthly wages into daily wages.

Another potential concern pertains to using incarceration rates as our measure of legal coercion. Mitigating this concern is the fact that the historical literature is quite clear that incarceration was indeed often the result of legal coercion aimed at smallholders. In Appendix B.3 we document this by very detailed category of offense for the Leewards for the start of our period. For the end of our period (1871 to 1913) we have court sentences but only by broad categories of offences, and only one of these—offences against property—maps into the legal coercion discussed in section 2.1. In Appendix B.3 we show that this category’s share of total court sentences correlated positively with both Nit and Cit.

Identiﬁcation: The OLS estimates reported thus far should be interpreted with caution because

Nit is likely to be endogenous in both the wage and coercion equations. In the wage equation, for example, productivity growth in plantation agriculture would have increased both Nit and wit.

Unobserved differential productivity growth would thus bias the OLS results against ﬁnding a negative effect of Nit on wit. In the coercion equation, for another example, if, unobserved to the econometrician, peasants in some but not all colonies were actively resisting the plantation system this could have evoked a planter response (increased Cit) while also undermining the plantation

38 system (decreased Nit). This would bias the OLS results against ﬁnding a positive effect of Nit on Cit.

Equation (6) offers a path towards an instrument for Nit. A British investor chooses between investing in the Caribbean, which returns (1 Oi)π(Nit), or investing elsewhere, which returns − W t. Across colonies, those colonies exogenously endowed with more hinterland Oi are less likely to receive the investment. Over time, improvements in non-Caribbean investment opportunities W t make it less likely that any colony receives the investment. Thus, the larger is Oi W t the smaller is · 39 Nit. We already have a hinterland measure Oi. We also argued in section 2.4 that British exports

38 For instance, Caribbean historiography suggests that the unmeasured proliferation and inﬂuence of local missionaries was an important factor that encouraged civil disobedience and undermined the planters; see for example Dookhan(1977, 156), Lewis(1986, ch.3), McLewin(1987, 85–87), and Holt(1992, ch.7). 39 In the model Oi · W t appears in equation (6), though parameterized slightly differently.

24 40 to non-Caribbean countries is a good measure of non-Caribbean opportunities W t. We expect that a higher Oi W t leads to a lower Nit, i.e., Oi W t has a negagtive sign in the ﬁrst stage. · · The Virgin Islands’ experience—where the plantation system collapsed because of two major hurricanes despite low values of Oi— suggests the need for a second, hurricane-based instrument

HDIit for Nit. (See section 2.5.3.) It is important to be clear about what role HDIit plays: Hurri- canes are not central to our paper but we need them to ﬁt the variation in Nit for the Virgin Islands.

Without the Virgin Islands, hurricanes are no longer needed to ﬁt Nit, but we prefer to retain them in order to reﬂect the full universe of British Caribbean sugar colonies, and because they were a particularly striking example of the mechanisms we highlight.

First Stage and Reduced Form Estimation: In moving on to our TSLS estimation, we propose the following ﬁrst stage equation

N N N N N N First Stage: Nit = β Oit + β HDIit + γ Xit + λ + λ + , (9) O · H · X · i t it as well as the associated reduced form relations

w w w w w w Reduced Form (Wages): wit = β Oit + β HDIit + γ Xit + λ + λ + ; (10) O · H · X · i t it C C C C C C Reduced Form (Coercion): Cit = β Oit + β HDIit + γ Xit + λ + λ + , (11) O · H · X · i t it

where i indexes colonies and t indexes years, the λs are ﬁxed effects, and Xit are the same control variables as in the OLS. Columns 1–4 of Table3 report estimates of the First Stage equation (9) for our two favored speciﬁcations and for the two different measures of planter power. In columns

N 1–4, βˆ , the coefﬁcient on Oi W t indicates that in islands with large hinterlands the plantation O · system declined faster in response to improving non-Caribbean opportunities for British interests.

The hurricane damage index HDIit has a powerful negative impact on the plantation system. In columns 5–8, we report estimates of the two reduced-form equations (10) and (11). The estimate

ˆw βO in column 3 says that compared to Barbados, wages in Grenada (where half the land was hinterlands) rose substantially more in response to an increase in W t. At the same time, columns

7–8 show a signiﬁcant effect on incarceration rates working in the opposite direction.

40We include both the British and non-British Caribbean in order to net out Cuba, which produced about a third of

25 Table 3: First Stage and Reduced Form N : Sugar's Export N : Plantations' C : Incarceration it it w : log wage it Outcome Share Export Share it (per Cap.) (1) (2) (3) (4) (5) (6) (7) (8)

O i x English Exports -0.4819 -0.5238 -0.2365 -0.2611 0.2351 0.2724 -0.4577 -0.4121

to all Non-Caribbeant [0.0000] [0.0001] [0.0012] [0.0030] [0.0153] [0.0075] [0.0056] [0.0287] {0.002} {0.003} {0.012} {0.008} {0.08} {0.108} {0.02} {0.015}

HDIit -0.0680 -0.0751 -0.0625 -0.0646 0.0483 0.0565 -0.0857 -0.0775 [0.0000] [0.0000] [0.0000] [0.0000] [0.0001] [0.0001] [0.0000] [0.0094] {0.085} {0.098} {0.067} {0.053} {0.225} {0.114} {0.124} {0.017}

2 2 2 2 Time Controls t + t t-fe t + t t-fe t + t t-fe t + t t-fe Observations 1,018 1,018 1,018 1,018 908 908 798 798 R-squared 0.853 0.867 0.843 0.851 0.721 0.772 0.526 0.580

Notes: Columns 1–4 present estimates of equation (9), the First Stage relation between the instruments and the two measures of Nit. Columns 5–8 present estimates of equations (10) and (11), the Reduced Form relations between the instruments and the two main outcomes wit and Cit. For each outcome, we present the results for our two preferred specifications which include a quadratic time trend (odd-numbered columns) or year fixed effects (even-numbered columns). These correspond to columns 2 and 4 of Table1. For brevity, we do not report coefficients on any of the controls that appear in Table1. All specifications include colony fixed effects. Standard errors are clustered by colony, p-values are in square brackets, p-values for wild-bootstrapped standard errors are in braces.

Figure 4: The Year-Speciﬁc Effect of Oi on Nit and on wit

-1

1837 1856 1875 1894 1913

Notes: This figure reports the point estimates and 95th-percentile confidence bands on Oit in two estimations of regression-equations (9) and (10), when Oit is replaced by a flexible interaction of Oi with year fixed effects (while also including year fixed effects on their own). The solid thick (red) line reports on the estimated effect on Nit in (9), the solid thin (blue) line reports on the effect on wit in (10); dashed lines are confidence bands.

26 Clearly, the time-varying component of Oit = Oi W t is a strongly trending variable; see · Panel (b) of Figure3. Therefore, a potential concern with the results reported in Table3 is whether the time path of the effect of Oi on the plantation system and on wages actually matches the evolution of W t. The most flexible way to ask this question is to re-estimate equations (9) and (10), replacing Oit with an interaction between Oi and year fixed effects (while continuing to include year fixed effects on their own as in columns 2, 4, 6, and 8 of Table3).

Figure4 plots the coefficients estimates on this interaction that result from doing so, with 1838 omitted because it gets absorbed by the colony fixed effects as the first year. The solid thick (red) line shows a clearly discernible negative effect of Oi on Nit, and this effect sets in around the mid-1860s. The solid thin (blue) line shows a positive effect of Oi on wages kicking in around

41 the same time. We view these time paths as consistent with the time-path of W t, and thus with our instrument: it is natural that the interaction-effect on Nit in Figure4 could not drop below 1 − since Nit is bounded below at zero; and it is therefore equally natural that the positive effect of Oi on wit ﬂattens off.

TSLS Estimation: We now turn to the causal effect of the declining strength of the plantation system on coercion and wages, estimating equations (7) and (8) with TSLS, and using equation (9) as the ﬁrst stage for both. Table4 reports the results for our two preferred speciﬁcations. Columns

1–4 report estimates of equations (7) and columns 5–8 report estimates of equation (8). Columns

1–2 and 5–6 report results when Nit is measured as sugar’s share of exports. Columns 3–4 and

ˆw 7–8 report results when Nit is measured as plantation crops’ share of exports. The estimates βN

ˆC and βN are very robust in magnitude within pairs of columns (1–2, 3–4 etc.), indicating results do not hinge on the particular speciﬁcation of the time controls. The F statistics on the First Stage

ˆw instruments are comfortably above critical threshold levels. The TSLS magnitudes βN in columns 1–2 and 3–4 are about 60% larger than the OLS estimates reported in columns 2 and 4 of Tables

ˆC 1 and2. The TSLS magnitudes βN in columns 5–6 and 7–8 are around twice the OLS estimates reported in columns 6 and 8 of Tables1 and2. In combination, this suggests that the OLS estimates are downward biased, consistent with the discussion at the start of the identiﬁcation section. world sugar cane in our period. 41 There is also a temporary spike in wages associated with a cholera epidemic that spread through the Caribbean in the early 1850s.

27 Table 4: TSLS Effect of the Plantation System (Nit) on Wages (wit) and Coercion (Cit)

Outcome w it : log wage C it : Incarceration (per Cap.) (1) (2) (3) (4) (5) (6) (7) (8)

N it (Share of Exports) -0.5437 -0.5881 -0.8359 -0.9614 0.7887 0.6354 1.1343 0.9499 [0.0009] [0.0077] [0.0000] [0.0001] [0.0036] [0.0310] [0.0000] [0.0048] {0.005} {0.029} {0.062} {0.037} {0.02} {0.012} {0.036} {0.014} Frechet Smallholder Export 0.3361 0.3153 0.1870 0.2197

Price Indexit [0.0001] [0.0004] [0.1686] [0.1026] log (Net Immigration)it -0.0261 -0.0219 -0.0283 -0.0241 -0.0000 0.0073 0.0033 0.0099 [0.0043] [0.0570] [0.0000] [0.0226] [0.9988] [0.7426] [0.8560] [0.6157] log(1/[Dist to Panama])i x 0.0045 0.0542 0.0038 0.0594 -0.0194 0.3327 -0.0178 0.3593

D(1881-1889) | 1908-1913)t [0.2382] [0.2957] [0.3145] [0.2864] [0.0493] [0.0983] [0.0665] [0.0819]

N it measured based on Sugar Plantations Sugar Plantations Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe Observations 908 908 908 908 798 798 798 798 F Statistic (Instruments) 54.8 52.7 9.0 8.4 44.7 51.8 8.1 7.9 Notes: This table reports the TSLS specifications corresponding to the specification in Tables1 and2. Columns 1–4 report the estimates of the wage equation (7). Columns 5–8 report the estimates of coercion equation (8). In columns 1–2 and 5–6, Nit is measured as sugar’s share of exports and plantation crops’ share of exports, respectively. For each outcome, we present the results for our two preferred specifications which include a quadratic trend or year fixed effects. This corresponds to columns 3–4 of Table1. For expositional clarity, we do not report coefficients on any of the controls. All specifications include colony fixed effects. Standard errors are clustered by colony, p-values in square brackets. p-values for wild-bootstrapped standard errors are in braces.

28 While we argue that the agro-climactic and international factors that underlie our instrument

Oi W t have no direct impact on Caribbean wages or coercion, Oi W t may nevertheless fail · · to satisfy the exclusion restriction because it is correlated with omitted variables that inﬂuence wages or coercion. In this section we present a range of speciﬁcations which address this concern.

Given the presence of both colony and year fixed effects, the precise concern is that Oi W t · is correlated with differential trends across colonies. For concreteness consider the following example, chosen because it is particularly damaging to our exclusion restriction. (1) Suppose that in some colonies labour supply trended up, leading to a fall in wages and a resulting expansion of sugar (wit Nit ) while in other colonies labour supply trended down, leading to a rise in ↓→ ↑ wages and a resulting contraction of sugar (wit Nit ). (2) Suppose that we have not controlled ↑→ ↓ for these differential trends in labour supply. (3) Suppose that our instrument is correlated with these differential trends. Under (1)–(3), our exclusion restriction is violated and a TSLS regression of wit on Nit would be more negative than the true coefficient. Thus, differential labour supply trends can potentially explain away our results. In Online Appendix H.1–Online Appendix H.6, we show that controlling for a variety of different trends leaves the point estimates and significance levels of our results qualitatively unchanged.

5 Extensions

In this section, we focus on showing that the negative effect of planter power on wages was in fact explained by the positive effect of planter power on coercion; and that coercion in our setting was of the type described in Sections 2.1–2.2.

The history, theory and econometrics presented thus far all point towards the following causal relations: Oit Nit Cit wit . However, we have not shown that there is a causal ↓ → ↓ → ↓ → ↑ impact of Cit on wit; we have only shown that Nit impacts Cit and wit. To remedy this, we estimate a model that regresses wit on Cit, and instruments Cit with Oit; reported in Online Appendix I

(Extension I). The model establishes a causal impact of coercion on wages (Cit wit). This, → however, still does not quantify the relative importance of the politics-coercion mechanism we postulate (Nit Cit wit) because our estimates of the impact of plantation power on wages → → (Nit wit) reported in Section4 may be capturing our postulated mechanism along with other →

29 political mechanisms through which Nit impacts wit. It is therefore of interest to estimate how much of the impact of Nit on wit operates via coercion as opposed to other political mechanisms.

A decomposition like this is done through a ‘mediation analysis,’ which asks to what extent the ‘total effect’ of a treatment variable (e.g. Nit) on a ‘ﬁnal outcome’ (e.g. wit) is explained by an intermediate outcome or mechanism (e.g. Cit) that is also an outcome of Nit. See for example

Imai, Keele and Yamamoto(2010); Imai, Keele, Tingley and Yamamoto(2011); Heckman, Pinto and

Savelyev(2013); Heckman and Pinto(2015). We use the methodology developed in Pinto, Dippel,

Gold and Heblich(2019) for performing mediation analysis in an IV setting like ours, estimate that upwards of 75% of the estimated effect of Nit on wit operates through coercion Cit; reported in Online Appendix I( Extension II).42 One aspect of legal coercion that is emphasized in the historical discussion in Section 2.2 is the distinction between ofﬁcially sanctioned de jure power and unofﬁcially sanctioned de facto power. We now supplement this historical discussion with econometric evidence by investigating the consequences of the Caribbean Incumbered Estates Act (IEA) of 1854, an exogenous shock to planters’ de facto relative to their de jure power in the exercise of coercion. The IEA originated in the Irish Potato Famine (1845–1849), which left Irish estate owners without operating capital and deeply in debt. In response, the Irish Incumbered Estates Act of 1849 strengthened creditor rights by allowing forced foreclosures (estate sales), which allowed capital to return to the Irish estates. The success of this legislative innovation, the result of a famine that was exogenous to the British West

Indies, led British Parliament to enact a Caribbean IEA in 1854.43

The most important consequence of the Caribbean IEA was a change in the composition and identity of the planter elite, because buyers of encumbered estates were often London-based merchant houses rather than local Caribbean planters (Hall, 2011, p.97). This meant that planters with deep local roots at the parish level were being replaced by London-based merchant houses that set up ofﬁces in colonial capitals, which changed the way legal coercion operated. The de facto power of the traditional planters elite over local parish police and magistrates made incarceration

42 Results reported in Online Appendix Table 15. 43 Before 1838, in the heyday of sugar, many plantations accumulated large encumbrances, that is, ﬁnancial commit- ments to pay annual stipends to family members and to repay loans incurred for capital projects such as mills. After 1838, rising wages and falling sugar prices led many planters to ignore their encumbrances. They could do so with impunity because the law subordinated creditors to both the plantation owners and to the London-based merchant houses who proﬁted from selling sugar. This and the subsequent discussion is based on Cust(1859, p.8–15). Cust was Secretary to the West Indian Incumbered Estates Commission. See also Hall(2011, p.94–97).

30 an easy method of intimidation and coercion. By contrast, the new merchant-house planters that were based in the islands’ metropoles had more direct access to influence the colonial administra- tion and the islands’ legislatures (de jure power). We therefore conjecture that the IEA led coercion to become more legislated, i.e., de jure-based. To test this we need a measure of legislated coercion. Planter elites enacted many coercive laws, such as discriminatory taxes on smallholds and licensing requirements for the transportation and sale of smallhold crops. Most of the coercive laws are very difficult to code up as a regressor.44 A less complex coercive law is the tariff on imports of foodstuffs. As noted in section 2.1, food imports competed with smallhold crops so low foodstuff tariffs made smallholding less attractive. Low foodstuff tariffs were de jure coercive. We also need to code up the IEA as a regressor. Using the full records of IEA court proceedings during the period it was in effect (1858–1889), we identified all 398 cases of court-ordered sales of plantations.45 Our reading of the IEA court cases is that the petitions for forced sale were often initiated by family members, sometimes distant family members, who had randomly fallen on hard times. To be safe, we also instrument court-ordered sales with court petitions. There were 523

IEA petitions in total, of which 398 led to court-ordered sales. Appendix Figure A1 in Appendix

B.7 shows the cumulative sales and petitions by colony over time.46

To summarize, we will regress wages, incarceration rates and foodstuff tariffs on IEA sales and instrument IEA sales with IEA petitions. We will also include Nit as before. Panel B of Table5 reports ﬁrst stages. There is a very clean statistical separation between the instruments for IEA sales and Nit. That is, IEA petitions have a strong impact on IEA sales, but no effect on Nit (see columns 1–4). Likewise, Oit and HDIit have a strong effect on Nit, but no effect on IEA sales (see columns 5–6).

Panel A of Table5 reports the TSLS estimates of the effects of Nit and IEA sales on wages, incarceration rates and foodstuff tariffs. To compare magnitudes, we have scaled IEA sales by

44These laws appear in the Blue Books. For example, most islands had regressive land and property taxes, but some taxes were levied in values, others were levied on acreage, and yet others had ad hoc qualifications such as tax adjustments for windmills, which were aimed at relieving planters. 45 Hall(2011, p.96) gives a slightly lower number of 382. He likely pulled this number from the 1884 Royal Com- mission Report on the IEA, which preceded the actual ending of the IEA by five years (Crossman and Baden-Powell, 1884). To the best of our knowledge we are the first to code up the IEA records since 1884. The sales and petitions were recorded in the Colonial Office Records Series CO 318-282-50 and CO-441-2-11, respectively. 46 The IEA was passed in London but needed to be incorporated on each island to apply there. St Lucia, Guyana, Trinidad and Barbados never incorporated the IEA, and thus no IEA sales took place in those colonies (Hall, 2011, p.97).

31 Table 5: The Effect of the IEA

Panel A. TSLS Estimation of Effects of Nit and IEA-Salesit on Three Outcomes Incarceration (per Import Duties w : log wage Outcome it Cap.) Flour (£) (1) (2) (3) (4) (5) (6)

N it : Sugar Exports -0.7151 -0.5683 1.2239 0.9346 -1.0126 -0.6124 Share of Total Exports [0.0004] [0.0181] [0.0001] [0.0031] [0.3102] [0.5148]

IEA Salesit -0.0388 -0.0589 -0.0662 -0.0547 -0.3593 -0.4226 [0.0495] [0.0258] [0.2909] [0.4050] [0.0026] [0.0298]

Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe Observations 908 908 798 798 942 942 p-value: Test[ 2 coefficients equal] 0.001 0.031 0.000 0.002 0.515 0.835 F Statistic (Instruments) 18.878 15.111 17.581 17.279 21.146 17.397

Panel B. First Stage for Two Endogenous Regressors: Nit and IEA-Salesit N : Plantations' N : Sugar's it it IEA Sales Outcome Export Share Export Share it (1) (2) (3) (4) (5) (6)

O i x English Exports -0.2399 -0.2359 -0.4452 -0.4925 0.0517 0.1095

to all Non-Caribbeant [0.0075] [0.0025] [0.0001] [0.0001] [0.5780] [0.3414]

HDIit -0.0665 -0.0665 -0.0677 -0.0718 0.0076 0.0140 [0.0000] [0.0000] [0.0000] [0.0000] [0.3717] [0.2126] Incumbered Estate Act (IEA) -0.0227 -0.0214 -0.0133 -0.0177 1.0367 1.0522

Petitionsit [0.1734] [0.2438] [0.5498] [0.4786] [0.0000] [0.0000]

Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe Observations 1,018 1,018 1,018 1,018 1,018 1,018 R-squared 0.882 0.884 0.891 0.897 0.991 0.990 Notes:(a) Panel A presents the TSLS Estimation of the effect on three outcomes. Online Appendix Table 16 reports additional robustness checks on Panel A. Panel B presents the First Stage Results for two endogenous regressors, i.e., Nit and IEA-Salesit. Because there are 3 different numbers of observations in Panel A, we report the First Stage in Panel B for all available data. In Online Appendix Table 17, we also report the First Stage for each of the 3 sets of observations in Panel A. This does not materially affect any coefficients. (b) For each outcome, we present the results for our two preferred specifications, i.e., the second and third specifications shown in all results in section4.( c) Standard errors are clustered by colony, p-values in square brackets.

32 47 their mean. The coefﬁcients on IEA sales are thus the average effect of the IEA. The effects of Nit on wages and incarceration rates (columns 1–4) are almost identical to our baseline Table4 results.

The average total impact of the IEA on wages was to lower them by approximately 5.89 percent

(in column 2). In contrast, Nit declined on average by 0.5 and so lowered wages by approximately

28.4 percent (= 0.5 0.568). This suggests that the collapse of the plantation system (Nit) was the × much more fundamental shock to labor markets than the introduction of the IEA, which is fully consistent with the historical literature. This eyeball observation is also borne out in a statistical test of the equality of the two coefﬁcients of interest (on Nit and on IEA sales), which we report at the bottom of Panel A. When it comes to incarceration, the effect of IEA sales was not only much smaller than that of Nit but was in fact statistically not distinguishable from zero (in columns

3–4). This stands in contrast to the effects on foodstuff tariffs, where columns 5–6 show that as the plantation system weakened (Nit fell), foodstuff tariffs rose to the beneﬁt of smallholders, but that this effect is statistically insigniﬁcant.48 In contrast, and as hypothesized, columns 5–6 also show the higher de jure access of the new merchant-planters brought in by the IEA to the islands’ lawmakers led to a much more precisely estimated reduction in foodstuff tariffs.49

In summary, the IEA sales provide evidence on the use of de facto versus de jure coercion. IEA sales did not increase de facto coercion as measured by incarceration rates (the results in columns

3–4 are insigniﬁcant), did increase de jure coercion as measured by lower foodstuff tariffs. The shape of legal coercion in the British West Indies varied with changes in the political power of the planter elite.

6 Conclusion

In this paper, we provide the ﬁrst rigorous empirical evidence on the importance of legal coercion in informal developing-country labor markets. Following Lewis’ famous assessment quoted at the beginning of this paper, we hypothesize that legal coercion aimed at depressing workers’ outside

47 In the full panel data, the average observation (over time and across all islands with any sales) had a cumulative number of IEA sales of 28. See Figure A1. 48 We do not report wild bootstrap standard errors, which blow up when instrumenting two endogenous variables. 49 Seeing as the point estimates of the effect of Nit on foodstuff tariffs are imprecise but large, we do not have the statistical power to reject that the effects of Nit and IEA-Salesit on tariffs are equal. In Online Appendix Table 16 we report additional speciﬁcations where we control for general changes in public ﬁnance (revenues and expenditures) in our data.

33 options in the informal sector is an important factor in determining wages, and that the formal sector’s inﬂuence over government is a key driver of the use of legal coercion.

We focus on a uniquely relevant empirical setting, studying 14 Caribbean plantation islands at the inception of free labor markets, i.e., starting right after the Emancipation of slaves. We measure the formal sector’s inﬂuence over government Nit by the plantation sector’s share of overall production. To gain identiﬁcation, we instrument for Nit with a interaction between the planters’ time-varying reservation wages outside the Caribbean plantation system and variation across islands in the peasants’ ability to evade the plantation system.

Controlling for possibly confounding crop price variation, labor demand and labor supply shocks, we ﬁnd strong support for both hypotheses in the data: The plantation system had powerful effects on wages and legal coercion (see Tables1 and4). A complete collapse of the plantation system raised agricultural wages by 100% and lowered coercion by two times its mean level in the data. At the sub-island parish-level we can also show that it lowered mortality rates. Pushing further towards identifying the causal mechanism linking these patterns, we ﬁnd that upwards of three-quarters of the plantation system’s effect on wages is explained by the plantation system’s effect on coercion (see Online Appendix Table 15). We also provide suggestive evidence that planters shaped legal coercion both through legislation passed in the islands as well as through personal connections to the police and judicial apparatus in the countryside.

In summary, Lewis was right.

References

Acemoglu, Daron and Alexander Wolitzky, “The Economics of Labor Coercion,” Econometrica, 2011, 79 (2), 555–600.

and James A. Robinson, “Persistence of Power, Elites, and Institutions,” American Economic Review, March 2008, 98 (1), 267–293.

and , Why Nations Fail: The Origins of Power, Prosperity, and Poverty, New York, NY: Crown Publishing, 2012.

, Simon Johnson, and James A. Robinson, “The Colonial Origins of Comparative Develop- ment: An Empirical Investigation,” The American Economic Review, 2001.

, , and , “Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution,” The Quarterly Journal of Economics, 2002, 117 (4), 1231–1294.

34 Alston, Lee J and Joseph P Ferrie, “Paternalism in Agricultural Labor Contracts in the US South: Implications for the Growth of the Welfare State,” The American Economic Review, 1993, pp. 852– 876.

Antr`as,Pol and Alonso de Gortari, “On the Geography of Global Value Chains,” Working Paper No. 23456, National Bureau of Economic Research, May 2017.

Ashraf, Quamrul H, Francesco Cinnirella, Oded Galor, Boris Gershman, and Erik Hornung, “Capital-Skill Complementarity and the Emergence of Labor Emancipation,” Technical Report 2018.

Basu, Kaushik, “Child Labor: Cause, Consequence, and Cure, with Remarks on International Labor Standards,” Journal of Economic Literature, 1999, 37 (3), 1083–1119.

Blattman, Christopher, Jason Hwang, and Jeffery G. Williamson, “Winners and Losers in the Commodity Lottery: The Impact of Terms of Trade Growth and Volatility in the Periphery 1870- 1939,” Journal of Development Economics, 2007, 82 (1), 156–179.

Bloch, Francis and Vijayendra Rao, “Terror as a Bargaining Instrument: A Case Study of Dowry Violence in Rural India,” The American Economic Review, 2002, 92 (4), 1029.

Bobonis, Gustavo J. and Peter M. Morrow, “Labor Coercion and the Accumulation of Human Capital,” Journal of Development Economics, 2014, 108 (0), 32–53.

Bolland, O.Nigel, “Systems of Domination After Slavery: The Control of Land and Labor in the British West Indies after 1838,” Comparative Studies in Society and History, 1981, 23 (4), 591–619.

Bolt, Jutta, Robert Inklaar, Herman de Jong, and Jan Luiten van Zanden, “Rebasing ‘Maddison’: New Income Comparisons and the Shape of Long-Run Economic Development,” 2018.

Brizan, George I., Grenada, Island of Conﬂict: from Amerindians to People’s Revolution, 1498-1979, London: Zed Books, 1984.

Cameron, A Colin and Douglas L Miller, “Bootstrap-Based Improvements for Inference with Clustered Errors,” Review of Economics and Statistics, 2008, 90 (3), 414–427.

Carvalho, Jean Paul and Christian Dippel, “Elite Identity and Political Accountability: A Tale of Ten Islands,” Technical Report, NBER Working Paper 22777 2016.

Chwe, Michael Suk-Young, “Why Were Workers Whipped? Pain in a Principal-Agent Model,” The Economic Journal, 1990, 100 (403), 1109–1121.

Costinot, Arnaud, Dave Donaldson, and Cory Smith, “Evolving Comparative Advantage and the Impact of Climate Change in Agricultural Markets: Evidence from 1.7 Million Fields around the World,” Journal of Political Economy, February 2016, 124 (1), 205–248.

35 Craton, Michael, “Continuity Not Change: The Incidence of Unrest Among Ex-Slaves in the British West Indies, 1838–1876,” Slavery and Abolition, 1988, 9 (2), 144–170.

, Empire, Enslavement, and Freedom in the Caribbean, Princeton: Markus Wiener Publishers, 1997.

Crossman, Colonoal C.M.G and George Esq. M.A. Baden-Powell, Report on the West Indian In- cumbered Estate Act, London: Eyre and Spottiswoode, 1884.

Cundall, Frank, Political and Social Disturbances in the West Indies: A Brief Account and Bibliography, London: H. Sotheran & Co., 1906.

Curtin, Philip D., “The British Sugar Duties and West Indian Prosperity,” Journal of Economic History, Spring 1954, 14 (2), 157–164.

Cust, Reginald John Esq., A Treatise on the West Indian Encumbered Estate Acts, London: William Amer, Law Bookseller, 1859.

Davidson, R. and J.G. MacKinnon, “Wild Bootstrap Tests for IV regression,” Journal of Business and Economic Statistics, 2010, 28 (1), 128–144.

Deerr, Noel, The History of Sugar, London: Chapman and Hall, 1950.

Dell, Melissa, “The Persistent Effects of Peru’s Mining Mita,” Econometrica, November 2010, 78 (6), 1863–1903.

and Benjamin A Olken, “The Development Effects of the Extractive Colonial Economy: The Dutch Cultivation System in Java,” The Review of Economic Studies, 03 2019.

Dippel, Christian, Avner Greif, and Daniel Treﬂer, “The Rents From Trade and Coercive Institu- tions: Removing the Sugar Coating,” NBER Working Paper No. 20958 February 2015.

Domar, Evsey D., “The Causes of Slavery or Serfdom: A Hypothesis,” Journal of Economic History, 1970, 30 (1), 18–32.

Dookhan, Isaac, A History of the British Virgin Islands, 1672–1970, Epping, England: Caribbean Universities Press, 1975.

Dookhan, Issac, A Post-Emancipation History of the West Indies, Collins, 1977.

Easterly, William, “Inequality does Cause Underdevelopment: Insights from a New Instrument,” Journal of Development Economics, 2007, 84 (2), 755–776.

Eaton, Jonathan and Samuel Kortum, “Technology, Geography, and Trade,” Econometrica, September 2002, 75 (5), 1741–1779.

Eisner, Gisela, Jamaica, 1830-1930: A Study in Economic Growth, Manchester: Manchester Univer- sity Press, 1961.

36 Engerman, Stanley, Terms of Labor: Slavery, Serfdom, and Free Labor, Stanford University Press, 1999.

Engerman, Stanley L., “Economic Adjustments to Emancipation in the United States and British West Indies,” Journal of Interdisciplinary History, 1982, 13 (2), 191–220.

, “Contract Labor, Sugar and Technology in the Nineteenth Century,” Journal of Economic History, September 1983, 43 (3), 635—659.

, “Economic Change and Contract Labor in the British Caribbean: The End of Slavery and the Adjustment to Emancipation,” Explorations in Economic History, 1984, 21 (2), 133–150.

and Kenneth L. Sokoloff, “Factor Endowments, Inequality, and Paths of Development Among New World Economics,” Technical Report, National Bureau of Economic Research 2002.

Green, William A., British Slave Emancipation: The Sugar Colonies and the Great Experiment 1830- 1865, Oxford: Clarendon Press, 1976.

Greif, Avner, “Commitment, Coercion, and Markets: The Nature and Dynamics of Institutions Supporting Exchange,” in Claude Menard and Mary M. Shirley, eds., Handbook for New Institu- tional Economics, Norwell, MA: Kluwer Academic Publishers, 2005, pp. 727–788.

Grossman, Gene M. and Elhanan Helpman, “Protection for Sale,” American Economic Review, September 1994, 84 (4), 833–850.

Hall, Douglas, “The Flight From the Estates Reconsidered: The British West Indies 1838–1842,” Journal of Caribbean History, 1978, 10 (11), 7–24.

, “The Crisis of the Plantation,” in K.O. Laurence, ed., General History of the Caribbean Volume 4 - the Long 19th Century, UNESCO Books, 2011.

Heckman, James J and Rodrigo Pinto, “Econometric Mediation Analyses: Identifying the Sources of Treatment Effects from Experimentally Estimated Production Technologies with Unmeasured and Mismeasured Inputs,” Econometric Reviews, 2015, 34 (1-2), 6–31.

Heckman, James J., Rodrigo Pinto, and Peter A. Savelyev, “Understanding the Mechanisms Through Which an Inﬂuential Early Childhood Program Boosted Adult Outcomes,” American Economic Review, 2013, 103 (6), 2052–2086.

Higman, Barry W., Slave Populations of the British Caribbean, 1807–1834, Baltimore: John Hopkins University Press, 1984.

, Jamaica Surveyed: Plantation Maps and Plans of the Eighteenth and Nineteenth Centuries, Mona, Jamaica: University of West Indies Press, 2001.

Holt, Thomas C., The Problem of Freedom: Race, Labor, and Politics in Jamaica and Britain, 1832–1938, Johns Hopkins University Press, 1992.

37 Honychurch, L., The Dominica Story: A History of the Island, Roseau: The Dominica Institute, 1984.

House of Commons Parliamentary Papers, Papers Relative to the West Indies, Part II. The Windward Islands’ Government, Printed for the House of Commons, March 1839.

, Papers Relative to the West Indies, Part III. Leeward Islands, Printed for the House of Commons, March & August 1839.

Imai, Kosuke, Luke Keele, and Te Yamamoto, “Identiﬁcation, Inference and Sensitivity Analysis for Causal Mediation Effects,” Statistical Science, 2010, 25 (1), 51–71.

, , Dustin Tingley, and Teppei Yamamoto, “Unpacking the Black Box of Causality: Learning about Causal Mechanisms from Experimental and Observational Studies,” American Political Science Review, 2011, 105 (4), 765–789.

Laurence, Keith O., Immigration into the West Indies in the 19th Century, Epping, England: Caribbean Universities Press, 1971.

Lewis, Gordon K., The Growth of the Modern West Indies, New York: Monthly Review Press, 1986.

Lewis, W Arthur, “Economic Development with Unlimited Supplies of Labour,” The Manchester School, 1954, 22 (2), 139–191.

Lobdell, Richard, “Patterns of Investments and Credit in the British West Indian Sugar Industry, 1838–1897,” in Hilary Beckles and Verene Shepherd, eds., Caribbean Freedom: Economy and Society from Emancipation to the Present, Princeton: Markus Wiener, 1996, pp. 319–329.

Lowes, Sara and Eduardo Montero, “Concessions, Violence, and Indirect Rule: Evidence from the Congo Free State,” 2018.

Marshall, Woodville K., “Notes on Peasant Development in the West Indies since 1838,” Social and Economic Studies, September 1968, 17 (3), 252–263.

, “Metayage in the Sugar Industry of the British Windward Islands, 1838–1865,” in Hilary Beck- les and Verene Shepherd, eds., Caribbean Freedom: Economy and Society from Emancipation to the Present, Princeton: Markus Wiener, 1996, pp. 64–79.

Maurer, Noel and Carlos Yu, The Big Ditch: How America Took, Built, Ran, and Ultimately Gave Away the Panama Canal, Princeton University Press, 2013.

McLewin, Philip J., Power and Economic Change: The Response to Emancipation in Jamaica and British Guiana, 1840-1865, Princeton, NJ: Garland, 1987.

Merivale, Herman, Lectures on Colonization and Colonies, London: Longman, Green and Roberts, 1861.

38 Michalopoulos, Stelios, “The Origins of Ethnolinguistic Diversity,” American Economic Review, 2012, 102 (4), 1508–1539.

Mitchell, B.R., British Historical Statistics, Cambridge, England: Cambridge University Press, 1988.

Moulton, Brent R, “Random Group Effects and the Precision of Regression Estimates,” Journal of Econometrics, 1986, 32 (3), 385–397.

Naidu, Suresh, “Recruitment Restrictions and Labor Markets: Evidence from the Postbellum U.S. South,” Journal of Labor Economics, April 2010, 28 (2), 413–445.

and Noam Yuchtman, “Coercive Contract Enforcement: Law and the Labor Market in Nine- teenth Century Industrial Britain,” American Economic Review, 2013, 103 (1), 107–44.

North, Douglass, “Ocean Freight Rates and Economic Development 1730-1913,” The Journal of Economic History, 1958, 18 (4), 537–555.

Nunn, Nathan, “The Long Term Effects of Africa’s Slave Trades,” Quarterly Journal of Economics, February 2008, 123 (1), 139–176.

and Leonard Wantchekon, “The Slave Trade and the Origins of Mistrust in Africa,” American Economic Review, 2011, 101 (7), 3221–3252.

O’Rourke, Kevin H. and Jeffrey G. Williamson, Globalization and History: the Evolution of a Nineteenth-Century Atlantic Economy, The MIT Press, 2001.

O’Rourke, Kevin H and Jeffrey G Williamson, “When did Globalisation Begin?,” European Re- view of Economic History, 2002, 6 (1), 23–50.

Paget, Hugh, “The Free Village System in Jamaica,” Caribbean Quarterly, 1964, 10 (1), 38–51.

Patterson, O, “Institutions, Colonialism and Economic Development: the Acemoglu-Johnson- Robinson (AJR) Thesis in Light of the Caribbean Experience,” 2013.

Pinto, R., C. Dippel, R. Gold, and S. Heblich, “Mediation Analysis in IV Settings With a Single Instrument,” 2019.

Richardson, Bonham C., Economy and Environment in the Caribbean: Barbados and the Windwards in the Late 1800s, Mona, Jamaica: University of the West Indies Press, 1997.

Riviere, W.Emanuel, “Labour Shortage in the British West Indies after Emancipation,” Journal of Caribbean History, 1972, 4, 1–30.

Roberts, G.W. and J. Byrne, “Statistics on Indenture and Associated Migration Affecting the West Indies, 1834–1918,” Population Studies, July 1966, 20 (1), 125–134.

Rogers, Howard Aston, “The Fall of the Old Representative System in the Leeward and Wind- ward Islands, 1854-1877.” PhD dissertation, University of Southern California. 1970.

39 Sánchez,Fabio, Mar´ıadel Pilar López-Uribe,and Antonella Fazio, “Land Conflicts, Property Rights, and the Rise of the Export Economy in Colombia, 1850–1925,” The Journal of Economic History, 2010, 70 (2), 378–399.

Satchell, Veront M., From Plots to Plantations: Land Transactions in Jamaica, 1866–1900, Mona, Ja- maica: Institute of Social and Economic Research, 1990.

Sewell, William G., The Ordeal of Free Labor in the British West Indies, Harper & brothers, 1861.

Sokoloff, Kenneth L. and Stanley L. Engerman, “History Lessons: Institutions, Factor Endow- ments, and Paths of Development in the New World,” Journal of Economic Perspectives, 2000, 14 (3), 217–232.

Underhill, Edward Bean, The Tragedy of Morant Bay: A Narrative of the Disturbances in the Island of Jamaica in 1865, Alexander & Shepheard, 1895.

West India Royal Commission, Report of the West India Royal Commission, New York, NY: Eyre and Spottiswoode, 1897.

40 Appendix A Mathematical Proofs

Proof of Equation (1): Drop j superscripts, i and t subscripts, and pt arguments. Deﬁne Tbg θ θ ≡ Tg(τgpg) . Let rg(ω) = Tbgzg(ω) be revenue per plot generated by crop g. Then rg(ω) has cu- −θ −θ Tbgrg Tbgr 1/θ mulative distribution e− , density θTbge− and mean Tbg Γ(1 1/θ). As is well known − −θ (Tb Tbg)r G from Eaton and Kortum(2002), P r rk(ω) < r k = g = e− − where Tb = Σ Tbk { ∀ 6 } k=1 and P r choose g = Tbg/Tb. Hence, the joint probability of g being the optimal crop choice and { } rg(ω) = r is

Tb r−θ P r choose g and rg(ω) = r = P r rk(ω) < r k = g P r rg(ω) = r = (θTb e− )(Tbg/Tb) . {{ } { }} { ∀ 6 } { }

The ﬁrst term in parentheses is a Frechet´ density with scale parameter Tb and hence with mean Tb1/θΓ. Hence the expected revenues generated by crop g are

1/θ 1 1 rg(Tbg) E[rg(ω)] = Tb Γ(Tbg/Tb) = TbgTb θ − Γ . (12) ≡

Further, the expected revenues generated by all crops are

1 r(Tbg) Σgrg(Tbg) = Tb θ Γ (13) ≡

θ where Tb Σg=1Tbg. Substituting Tbg Tg(τgpg) into this mean yields equation (1). Q.E.D. ≡ ≡ Characterization of C (N): Substituting equations (2)-(3) into (4) yields W (C) = (αl(N)N L)C ∗ − − αCγ +αl(N)N[r(p, τ p) r(p, τ s)]+Lr(p, τ s), which is concave in C for γ > 1. Hence C is unique. − ∗ ∂W/∂C = 0 implies (C )γ 1 = [αl(N)N L]/[αγ]. From the deﬁnition of N¯ before equation (5), ∗ − − αl(N)N L > 0 N > N¯. Hence the constrained (C 0) optimal solution is C = 0 for N < N¯ − ⇔ ≥ ∗ and equation (5) for N N¯. ≥

Existence, Interior Solutions, and Stability of N ∗: The equilibrium number of planters N ∗ is given by equation (6). We first derive a sufficient condition which ensures an interior solution N (0,L). Start by defining ∆r r(p, τ p) r(p, τ s). A sufficient condition for an interior N ∗ ∈ ≡ − ∗ is π(0) > W/(1 O) > π(L). From equation(5), C∗(0) = 0 so that from equation (3), π(0) = − 1/(γ 1) l(0)∆r. Also, it is tedious but straightforward to show that π(L) < ∆rL − . Hence a sufficient condition on parameters is l(0)∆r > W/(1 O) > ∆rL1/(γ 1). To see stability, consider a plot of − − π(N) and W/(1 O) against N. As we move from N = 0 to N = L, π(N) must cut the horizontal − W/(1 O) line from above, which is the requirement for stability. Finally, at such an intersection − N = N ∗ and πN (N ∗) < 0.

Appendix B Data

We made use of two types of original data sources. The ﬁrst is new dataset of geographic and agro-climatic conditions in the Caribbean. To calculate pre-determined outside options as well as

41 Frechet-based´ structural model of crop choice, we need detailed spatial data on soil suitability for different crops. Standard sources for crop-suitability data are too coarse for our colonies.50 We therefore gathered agro-climatic data on the Caribbean at an unusually ﬁne spatial level and then developed agro-climatic suitability indexes for the 8 most important crops in our data. The second source of novel data is archival. Starting in the mid-1830s, Britain began collecting statistics on colonial conditions. Each colony ﬁlled out an annual Blue Book and sent it to London where it is now stored in the British National Archives. We photographed thousands of the relevant Blue Book pages. As well, we made use of the Statistical Tables Relating to the Colonial and Other Possessions of the United Kingdom, annual Censuses, and various other House of Commons Parlia- mentary Papers. We manually entered the relevant data into spreadsheets and built a panel data set on exports and export prices by crop, race demographics, wages, incarceration rates, coercive taxes, and military expenditures. The panel consists of 14 colonies from 1838 to 1913.

Appendix B.1 Wage and Employment Data

Praedial Wages: The Blue Books report wages for trade, domestic labour, and praedial labour (praedial means agricultural). We treat the praedial wage as the wage paid for plantation work. While the Blue Books do not explicitly state this, it is implicit from the context, namely, that there was no other agricultural activity that paid wages. Aside from plantation work, the only signiﬁcant amount of agricultural work was on smallholds and we now provide evidence that smallholds did not hire workers. Eisner(1961) shows that the majority of Jamaican smallholds were under two hectares, which even today is considered a very small farm 51 and hence requires the peasant to search for off-smallhold work, i.e., it was rare for smallholders to hire workers and hence to pay wages. Anecdotal evidence supports the claim the smallholders were subsistence farmers who did not make wage offers. Paget(1964) provides many contemporary quotes from the 1838–1840 period. For example: “It appears to me that the land which they purchase is ... too little in extent to be looked to as a permanent source of subsistence and that they must calculate either on obtaining additional means of comfort by going out to labour, or on taking more land on lease.” (Paget, 1964, 48). For a later period, the West Indies Royal Commission of 1897 makes the stronger point that, except for plantation work, there were no other wage opportunities in agriculture: “If work cannot be found for the labouring population on estates, they must either emigrate or support themselves by cultivating small plots of land on their own account.” (West India Royal Commission, 1897, 17) This discussion establishes that praedial wages were plantation wages and those paid them were either full-time employees or part-timers with subsistence smallholds.

50For example, each grid cell in the Geographically Based Economic Data (GBED) database (e.g., Michalopoulos, 2012) has a resolution of 0.5 degrees latitude by 0.5 degrees longitude, which, at the equator, is over 3,000 square kilometers. Likewise, the crop suitability data compiled by the FAO GAEZ project (e.g., Costinot et al., 2016) is at the 5 arc-minute level, which, at the equator, is 86 square kilometers. The smallest island in our data, Nevis, is as big as one cell in the FAO GAEZ data. The ten smallest islands in our data together ﬁt into a single cell in the GBED database. 51e.g., “Proﬁle of the Small-Scale Farming in the Caribbean, ”Workshop on Small-Scale Farming in the Caribbean Barbara Graham 2012, IMF, page 13).

42 Employment and Non-Agricultural Outside Options: Were there other, non-agricultural options for peasants? These appear to have been very limited or non-existent, as suggested by the Royal Commission quote above. More systematic evidence to this effect can be gleaned from the Blue Books. The Blue Books list three types of employment: agricultural, manufacturing, and commercial. The Books report that there was virtually no manufacturing employment and commercial employment was typically tiny. The population share of those employed in commerce never exceeded seven percent, while that in agriculture was as high as ninety percent in some colony-years.52 Table A1 shows in columns 1–2 that a complete collapse of the plantation system

(∆Nit = 1) was associated with a 33-percentage-point decline in the population-share in agri- − cultural employment. This exactly equals the average agricultural employment share. By way of a counterfactual argument, if wage labor on smallholds had been an option for plantation workers, we might have expected this correlation to be close to zero given as plantation agriculture was replaced by smallhold agriculture. Columns 3–6 of Table A1 also show no signiﬁcant correlation between Nit and employment in commerce or manufacturing. By way of a counterfactual argument, if wage work in the towns had been an option for plantation workers, we might have expected this correlation to be signiﬁcantly negative. Thus, the Blue Books do not provide any evidence of non-agricultural outside options for plantation workers.

s Appendix B.2 Export Price Index ri(pt, τ )

We used the Blue Books to construct a 76-year panel of exports by colony and crop, generating a database containing exports by colony and year for 17 products accounting for 98% of exports.53 We then built prices as export unit values (export revenues divided by export quantities) for the 17 products. See Online Appendix D for a description of the price data. We validated our series against the sporadic contemporary sources available. For example, for sugar our export unit values are virtually identical to data in Deerr(1950), the seminal work on the subject, and to sugar prices in Blattman, Hwang and Williamson(2007). s We now describe how we estimate ri(pt, τ ) in equation (1). First, except for sugar, no other crop was exclusively a plantation crop, e.g., coffee was typically grown both on plantations and smallholds. We therefore had to determine for each island whether a crop was fully or partly grown by plantations; see discussion in Online Appendix A. Second, we use the same information on agro-climactic conditions that we utilized for section Appendix B.4 and develop the agro- climatic suitability indexes for the 7 most important plantation crops beyond sugar.54

52These data wes reported in the population table. However, employment data was much more sparse because it was only occasionally reported or updated. Furthermore, it was reported at the parish level but with year-to-year variation in which parishes were reported so that aggregation to the colony level was frequently not possible. (For more on parish-level data , see Online Appendix H.5.2.) 53 The products are sugar, livestock, arrowroot, cocoa, lime juice, cotton, bananas, oranges, pimento, coffee, charcoal, lumber, coconuts, ginger, other spices (cloves, mace and nutmeg), balata (a natural tar), and asphalt. 54This entails identifying the agro-climactic factors relevant for each crop. For example, lime has seven factors including average temperature (23–30◦C is very suitable) and soil pH (6.1-6.5 is very suitable). For each crop, the relevant

43 Table A1: Partial Correlation between Nit and Agricultural Employment (1) (2) (3) (4) (5) (6) # Employed in Agriculture / # Employed in Commerce / # Employed in Outcome: Population Population Manufacturing / Population Panel A

N it : Plantation Exports as 0.3319 0.0007 -0.5276 Share of Total Exports [0.0003] [0.9716] [0.4356]

N it : Sugar Exports as 0.2392 -0.0030 -0.1515 Share of Total Exports [0.0002] [0.8117] [0.7084] Observations 125 125 102 102 16 16 R-squared 0.381 0.374 0.291 0.293 0.823 0.762 Panel B

O i x English Exports -0.0955 -0.1363 -0.0004 0.0003 0.0124 0.0124

to all Non-Caribbeant [0.0058] [0.0326] [0.9526] [0.9671] [0.7840] [0.7840]

HDIit 0.1103 -0.0020 [0.2516] [0.8966] Observations 125 125 102 102 16 16 R-squared 0.311 0.323 0.291 0.291 0.745 0.745

Notes: This table presents results of an OLS regressions where the outcome is the total number of people employed in agriculture (columns 1–2), commerce (3–4) and manufacturing (5–6) divided by the total population and the regressor is our Nit. With so few observations, we otherwise only include colony ﬁxed effects. Standard errors are clustered by colony, p-values in square brackets.

Let Agi be a vector of crop suitability characteristics on island i pertaining to crop g. We esti- s mate the parameters of ri(pt, τ ) using an almost standard gravity equation method common to j θ θ Frechet-based´ models. That is, we relate crop-level exports to the Tgi(τg ) (pgt) terms in equation (1). In the standard gravity approach a plot’s productivity is drawn from a single distribution, but p here it is drawn from two distributions, depending on whether the plot is a plantation (τg ) or a s smallhold (τg ). To keep the estimation within a standard framework we assume that a plot’s productivity is drawn from a single distribution with a parameter τg which is the geometric average p s of τg and τg . Speciﬁcally, p P lgit s 1 P lgit τg(P lgit) (τ ) (τ ) − (14) ≡ g g 55 where P lgit is the share of crop g in colony i in year t produced on plantations. We also assume that ln Tgi = αgAgi where αg is a parameter vector. Gravity estimation very roughly boils down to 56 regressing log exports on Agi, P lgit, a crop dummy, and ln pgt. θ Substituting Tbg Tg(τgpg) into equation (12) of Appendix A and adding subscripts, the aver- ≡ factors are ranked in their importance and aggregated into discrete suitability bins that reﬂect meaningful cutoffs in overall suitability for given crops. The details of our crop suitability coding appear in Online Appendix C and online appendix tables Online Appendix Table 4 and Online Appendix Table 5. 55 This is closely related to the more rigorous solution to this problem in Antras` and de Gortari(2017). 56 j p s p s s j θ θ To see this note that ln τg = P lgit ln τg + (1 − P lgit) ln τg = ln(τg /τg )P lgit + ln τg . Hence, ln[Tgi(τg ) (pgt) ] = p s s αgAgi + θ ln(τg /τg )P lgit + θ ln τg Dg + θ ln pgt where Dg is a crop dummy. This explains the regressors Agi, P lgit, Dg, and ln pgt.

44 age earnings per plot from crop g are

1 1 h θi h θi θ − rgi(pt, τ) = Tgi(τgpgt) ΣkTki(τkpkt) Γ . (15)

To estimate the parameters of this equation, recall that we assumed ln Tgi = αgAgi and τg = p P l s 1 P l τg(P lgit) (τg ) git (τ ) git where P lgit is the share of exports produced on plantations. Then ≡ g − differencing equation (15) between crop g and any other crop g0 yields

ln rgi(pt, τ) ln rg0i(pt, τ) − h θi h θi = ln Tgi(τg(P lgit)pgt) ln Tg0i(τg0 (P lg0it)pg0t) − p s p s s s = αgAgi αg0 Ag0i + θ ln(τ /τ )P lgit θ ln(τ 0 /τ 0 )P lg0it + θ ln(τ /τ 0 ) + θ ln pgt/pg0t . − g g − g g g g

We assume that exports per plot are a noisy measure of output per plot: ln xgit/ni = ln rgi + εgit where ni is the number of plots. Then we obtain the estimating equation ln(xgit/xg0it) = βgAgi + βsAg0i + β0 P lgit + β0 0 P lg0it + β00Dg + β000 ln pgt/pg0t + νgit i, t and g = g0 g g g ∀ 6 (16) where βs are regression coefficients, g is fixed (it will be sugar), Dg is a crop dummy and νgit 0 ≡ εgit εg0it. It is easy to verify that the estimates of equation (16) identify all the parameters of (15) − s except τg0 e.g., β000 = θ. Thus, rgi(pt, τ) and ri(pt, τ) = Σgrgi(pt, τ) (see appendix equation 13) are s s known up to the multiplicative constant τg0 . Recall that if P lgit = 0 for all g then τ = τ i.e., if there s is no plantation production then τ takes on its smallhold value τ . Hence, setting P lgit = 0 for all s s s g, ri(pt, τ ) is also known up to the multiplicative constant τg0 . In regressions we use ln ri(pt, τ ) s so that the additive constant ln τg0 is subsumed into intercepts. Equation (16) is our gravity equation, where g0 is sugar. Data on the share of exports produced on plantations (P lgit) are described in Online Appendix A. Data on agro-climactic conditions (Agi) are described Online Appendix C. The bin-level agro-climactic variables are averaged up to the colony level to create the Agi. Data on export prices (pgt) are described in Online Appendix D. Table Online Appendix Table 6 displays the estimates of the key parameters. There are 7126 = 1018 7 observations where 1018 is the number of colony-year pairs and 7 is the number of (non- × sugar) crops. We estimate θ to be 2.30 (t =3.93). This is in line with Costinot et al.(2016) who estimate it to be 2.46. The remaining estimates reported in table Online Appendix Table 6 are the coefficients on

P lgit. Except for livestock they are all positive as expected. There are several things to note. The negative livestock coefficient is largely driven by a comparison of Virgin Island smallholds with Tobago plantation and it is thus not surprising to find that the Virgin Islands did better. Cotton p s p is always a smallhold crop, which means that its coefficient (βg0 = θ ln(τg /τg )) and hence τg are p s 57 not identified; however, we do not need to know τg in order to recover ri(pt, τ ). Likewise, the

57 This is important for another reason: Most other smallhold crops (e.g. garden vegetables, roots and tubers) were perishable, and therefore do not appear in any of our export price series. It is worth noting that suitability for small-

45 sugar coefficient is not identified because sugar is always a plantation crop. This does not matter because sugar does not enter into the Frechet´ smallholder price index. In summary, the gravity p s G equation (16) identifies the parameters αg, τg , τg g=1 and θ needed to recover estimates of the s { } ri(pt, τ ).

Appendix B.3 Coercion Data

In this appendix we address two questions. First, what share of incarcerations or court convictions were associated with legal coercion? Second, what is the relationship between incarcerations and court convictions, given that the latter involved both fines and incarcerations? Incarceration and conviction data were rarely reported in sufficient detail to isolate crimes associated with legal coercion. Fortunately, in the earliest post-Abolition years London was concerned about the abuse of legal coercion by planters and so requested detailed reports on each colony’s Acts, incarcerations, convictions, and evictions of peasants from plantations. Colonial Assemblies did not want to comply and often did not: Sometimes there are no tables and sometimes the tables are very imperfect. Where they are available, they are reported in complicated ways. For example, in some colonies there is a table for each magistrate separately along with details of each case and its outcome (fine, incarceration, eviction). This is a useful source for our fine-grained historical discussion of legal coercion, but not for statistical analysis. Fortunately, in 1838, four of the Leewards (Antigua, St. Kitts, Nevis, and the Virgin Islands) actually filled in Lon- don’s standardized and very detailed tables of convictions by detailed crime. These tables appear amidst hundreds of pages of detailed exegesis on colonial laws, and as a result the detailed crimes can be fully understood and categorized as coercive or not. The tables were combined into a single table that appears as our Table A2. The first column lists the detailed crimes. The first group (‘Coercive’) are the crimes that were described in section 2.1 and are clearly related to legal coercion. Note that breach of contract, vagrancy and trespass were all lumped in the Vagrancy Acts of a number of colonies,58 which illustrates that one cannot classify crimes as coercive or non- coercive without detailed knowledge of the laws of each colony. Turning to the other coercive crimes in Table A2, licensing was used to limit the opportunities for non-plantation activities such as huckstering (selling goods), jobbing (handyman), boating (transporting goods), and deporting (moving peasants to another island, which was illegal in the Leewards).59 Malicious injury to property could stem from the eviction of a peasant from the plantation, the subsequent fight over the peasant’s access to his/her crops, and retribution on both sides for perceived injustices.60 hold crops other than cotton would not have varied much across islands since these are hardy crops that could grow anywhere in the fertile Caribbean climate. This is witnessed by the fact that they were, in fact, grown everywhere; on plantations as well as in the hinterlands. 58 For example, the Vagrancy Act for St. Kitts states that “all persons shall be deemed vagrants who [are] dwelling in any of the houses upon any plantation without the sanction of the owner or director thereof, or shall be found trespassing on the land of any plantation by attempting to cultivate ... [or] have left without sufficient cause any work unfinished, which they have contracted to perform” (House of Commons Parliamentary Papers, 1839b, 19). 59The restrictive provisions of these Acts were condemned by London-appointed governors, e.g., House of Commons Parliamentary Papers(1839b, 41). 60Again, classification is not neat. For example, the Virgin Islands Vagrancy Act covers attempts to burn down

46 Table A2: Convictions (Fines and Incarceration) by Type of Crime Jan. - Sept. 1838 July - Sept. 1838 Number % Number %

Coercive 991 41% 425 45% Breach of contract 479 20% 227 24% Vagrancies 192 8% 91 10% Trespass 137 6% 54 6% Hucksters without licence 2 0% 0 0% Jobbing without licence 44 2% 5 1% Plying unlicenced boat 4 0% 0 0% Deportation without license 4 0% 2 0% Malicious injury to property 129 5% 46 5%

Partially Coercive 886 37% 348 37% Police Act 257 11% 91 10% Riotous, disorderly conduct 77 3% 26 3% Assault and battery 552 23% 231 25%

Not Coercive 516 22% 166 18% Larceny 75 3% 22 2% Miscellaneous 441 18% 144 15%

Total 2,393 100% 939 100%

Notes: Data compiled by authors from House of Commons Parliamentary Papers (1839b), pages 78-79 (Antigua), 225–226 (St. Kitts), 286–287 (Nevis) and 332-334 (Virgin Islands). Hucksters are sellers of small wares. Jobbers are handymen. The Police Act varies from colony to colony. It always covers day-to-day non-coercive offences (startling a horse etc.) and, in some colonies, offences relating to coercion. Riotous and disorderly conduct ranges from public drunkenness to organized peasant protests. Assault and battery includes assaulting an employer. ‘Misc.’ is the authors’ own category and collects crimes that are not related to coercion of smallholders, e.g., petty theft, mutiny and abusive language. The results are very similar for January-September and July-September. See footnote 61 for a discussion of the differences between these two periods.

Other crimes are for categories that include both coercive and non-coercive crimes. Police Acts varied from colony to colony, always covering day-to-day non-coercive offences such as startling a horse and, in some colonies, covering offences relating to coercion. Riotous and disorderly conduct ranges from public drunkenness to organized peasant protests. Assault and battery includes assaulting an employer. This discussion of Table A2 explains why it is so difﬁcult to identify legal coercion from the crime statistics. It also makes clear that we can do so in 1838 because we have the supporting Acts and legal discussion of them. Assuming that half of the ‘Partially Coercive’ crimes were coercive, 60% of convictions involved legal coercion (60 = 41 + 37/2). Legal coercion was thus an important source plantations (House of Commons Parliamentary Papers 1839b, 310) so that malicious injury is sometimes recorded as vagrancy.

47 of convictions.61 We next turn to the question of what percentage of these court convictions involved incarceration. Recall that convictions led to either fines or incarceration. From inspection of convictions tables from 1838, a year in which such tables were collected and published in a major House of Commons report, several patterns emerge. First, the vast majority of higher court convictions involved incarceration, e.g., House of Commons Parliamentary Papers(1839b, 72). Second, lower court cases involved either mediation (no conviction, one party agrees to indemnify the other), or conviction with fine, or conviction with imprisonment (for more significant crimes such as long or repeat absenteeism and neglect leading to death of livestock).62 The only comprehensive account of lower court cases for 1838 is a single table for Barbados (August 1 to October 15). In that table, 68% of the convictions involved imprisonment (House of Commons Parliamentary Papers, 1839a, 104–105). Summarizing what we could glean from these reports, the majority of court convictions were for offenses that involved legal coercion, and the majority of court convictions resulted in incarceration, albeit usually only for two weeks or one month, often at hard labor. For the end of our period (1871 to 1913) we have an annual panel of court convictions (with no information on what share resulted in incarceration). It would thus be of interest to use these convictions as an alternative measure to legal coercion. To this end, we must identify which convictions are associated with legal coercion. Unfortunately, the conviction data are only available by four broad categories: (1) offences against property, (2) offences against the person, (3) praedial larceny (the theft of livestock and crops), and (4) other. The best we can do is equate offences against property with legal coercion. There are pros and cons to this. The Blue Books state that offences against property include both offences against rights of property and injuries to the sub- jects of property, which means that offences against property include trespass.63 The Blue Books further state that “by praedial larceny is meant the offence—prevalent in the sugar-growing and Cooley-importing colonies—of robbing provision grounds and homesteads.” By 1871, it probably had only a small component of legal coercion, i.e., of evicted peasants recovering their crops.64 Offences against the person is a category that, as with assault and battery, likely had only a small component of legal coercion, i.e., of peasant retribution against plantation overseers for loss of cottage and crops. ‘Other’ could subsume any number of offences listed in Table A2 and, since

61 These data are for January to September 1838 and so span the August 1, 1838 end of slavery. This is not an important issue because half of the convictions occurred in Antigua (the most populous colony), which had abolished slavery outright in 1834. (The remaining colonies had abolished slavery, but were in the transition Apprenticeship period.) Erring on the side of caution we also report the results for the last quarter available (July, August and September, 1838). For this period, the percentage of legal coercion rises slightly to 64%. 62Note that imprisonment for failure to pay a fine occurred, but this is not the same as saying that a prisoner could buy his or her way out of jail. Sentences were typically of the form “sentenced to pay a fine of x dollars; in default, y days imprisonment.” In St. Kitts in the second half of 1838, convictions for breach of contract/vagrancy resulted in 46 fines and only one of these resulted in imprisonment for failure to pay the fine. See House of Commons Parliamentary Papers(1839b, 200–206). 63As noted above, breach of contract is not an offence against property. 64Unfortunately, whether a particular theft is praedial larceny (maximum penalty is a long prison term) or petty theft (maximum penalty is a short prison term) appears to vary across colonies. Petty theft likely falls in the ‘other’ category.

48 it makes up about 55% of the convictions in our data, it is too much of a catch-all category to be called legal coercion. In summary, we conclude that offences against property are the only category that can be linked to the legal coercion discussed in section 2.1, but this must be done with caution. Note that crimes against property only make up about 15% of the convictions in our data. Table A3 shows that this category’s share of all convictions correlates strongly with Nit. Specifically, columns 1–4 of the table report a regression of the share of court convictions for offences against property on our two measures of Nit. The full specification corresponds to our two baseline specifications in Table1. There is also a somewhat weaker positive correlation with Cit in columns 5–6. We view these patterns as validation of incarceration as our measure of coercion.

Table A3: OLS Regressions of Court Convictions on Nit and Cit (1) (2) (3) (4) (5) (6) Outcome Share Court Convictions for Offences Against Property

N it : Sugar Exports as 0.2471 0.2204 Share of Total Exports [0.0372] [0.0898]

N it : Plantation Exports as 0.3298 0.3190 Share of Total Exports [0.0206] [0.0512] Incarceration (per Cap.) 0.0309 0.0325 [0.2650] [0.2100] Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe Observations 502 502 502 502 408 408 R-squared 0.141 0.227 0.140 0.226 0.136 0.226

Notes:(a) From 1871 to 1912 we observe data on local court sentences by categories of offences. There are 4 categories: offences against property, offences against the person, “praedial larceny” (the theft of livestock and crops), and “other.” Only one of these—offences against property—corresponds closely to the legal coercion discussed in section 2.1. In this table we run our set of three core speciﬁcations on three relationships: In columns 1–4, we regress the share of all court sentences that were levied for offences against property, for our two core speciﬁcations. In columns 5-6, we regress it on our incarcerations measure. (b) Standard errors are clustered by colony, p-values in square brackets.

Appendix B.4 Measuring Oi

To measure Oi we carefully calculated the share of each colony’s land that is not suitable for sugar cane. This was a major undertaking, but only insofar as the Caribbean islands small size forced us to collect geographic data at an unusually ﬁne level of disaggregation. See Online Appendix C for details.

Appendix B.5 Measuring W t

Real British exports to non-Caribbean destinations (W t) is constructed from data in Mitchell(1988). Real British exports to all destinations are from Table IX.19. These data exclude re-exports.65 Nominal British exports to the Caribbean are from Tables IX.5 (1838–1847) and IX.16 (1846–1913).

65Britain imported items such as sugar and then re-exported them.

49 Over our period, 83% of British exports were manufactures, so we convert nominal exports to the Caribbean into real exports using the price index for British manufacturing exports. The price index is constructed from data in Tables IX.5 and IX.19. Our instrument W t is real British exports less real British exports to the Caribbean. Data are in millions of 1913 pounds sterling.

Appendix B.6 Additional Labor Supply Shocks

There was relatively little cross-island migration in the Caribbean in the 19th century. During 1838–1913 the most significant movement of people in the Caribbean was the arrival of indentured immigrants from India. We coded up the annualized data as a cumulative stock from Roberts and Byrne(1966). Between 1838 and 1913, cumulative net immigration was 230,000 for Guyana, 124,000 for Trinidad, 37,000 for Jamaica and small amounts for Grenada, St. Lucia, Antigua and Dominica. The ratio of cumulative net immigration to 1913 population exceeded 0.15 for only two colonies, Trinidad where it was 0.37 and Guyana where it was 0.77. Restated, immigration was important in only two colonies but in those two it was indeed important. The other labour supply shock was much smaller. British West Indies workers left for Panama during the building of the canal by the French (1881-1889) and Americans (1908–1913) (Maurer and Yu, 2013, ch.4). No destination-specific estimates exist of Caribbean emigration but the official numbers for overall emigration reported in the Colonial Blue Books were small throughout. We control for the Panama canal shock with a time-dummy for the years in question interacted with inverted distance as a measure of exposure to this shock.

Appendix B.7 IEA Sales and Petitions

This section consists only of Figure A1.

50 Figure A1: Incumbered Estate Act Petitions and Sales

Antigua Dominica 15 100

50 5

0 0 Petitions (Dashed) & Sales Petitions (Dashed) & Sales 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 Grenada Jamaica

40 200

30 150

20 100

10 50

0 0 Petitions (Dashed) & Sales 1838 1853 1868 1883 1898 1913 Petitions (Dashed) & Sales 1838 1853 1868 1883 1898 1913 Montserrat Nevis 6 40

30 4

20 2 10

0 0 Petitions (Dashed) & Sales Petitions (Dashed) & Sales 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 St Kitts St Vincent 40 40

30 30

20 20

10 10

0 0 Petitions (Dashed) & Sales Petitions (Dashed) & Sales 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 Tobago Virgin Islands 60 1

0 -1 Petitions (Dashed) & Sales Petitions (Dashed) & Sales 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 Notes: This figure shows colony-specific time series of the cumulative number of IEA petitions and sales. In total, there were 523 IEA petitions (by claimants) to force sales and 398 plantations were actually sold through the IEA between 1858–1889. The IEA was passed in London but needed to be incorporated on each island to apply there. St Lucia, Guyana, Trinidad and Barbados are not in the figure because they never incorporated the IEA, and thus no IEA sales took place in those colonies (Hall, 2011, p.97).

51 Online Appendix – Not for Publication

Online Appendix

“Outside Options, Coercion, and Wages: Removing the Sugar Coating” Online Appendix – Not for Publication

Online Appendix A Additional Information on Plantation Crops

Online Appendix Figure 1 shows that the lowess-smoothing in Figure1 faithfully reproduces the raw annual data. Online Appendix Figure 2 plots the share of plantation crops in total exports by colony. There are two difficulties in measuring the share of all plantation-produced goods in total exports. First, except for sugar, no other crop was exclusively a plantation crop, e.g., coffee was typically grown both on plantations and smallholds.66 To know which crops were plantation crops, we must therefore use detailed histories of each colony. Second, we must construct a 76-year panel of exports by colony and crop, using the Blue Books. The final database contains exports by colony and year for 17 products accounting for 98% of exports. The products are sugar, livestock, arrowroot, cocoa, lime juice, cotton, bananas, oranges, pimento, coffee, charcoal, lumber, coconuts, ginger, other spices (cloves, mace and nutmeg), balata (a natural tar), and asphalt. For the purposes of identifying plantation crops we focus on crops that were important in the sense of accounting for at least 15% of any one colony’s exports in any one year. There are eight such crops and they account for 95% of exports.67 The time series of export shares by colony and crop appear in Online Appendix Figure 3. Let P lgit [0, 1] be the share of a given crop g grown in colony i in year t that was produced ∈ on plantations. It is coded as follows. Sugar is always and everywhere a plantation crop. Hence P lsugar,it = 1 for all i and t. Livestock in the Virgin Islands (1842–1913)68 was exclusively a smallhold crop (Harrigan and Varlack 1975, 64–65; Dookhan 1975, 138) so that P lgit = 0. Livestock in Tobago (1886–1892) was primarily a plantation crop (Craig-James, 2000, 266-267) so that P lgit = 3/4. Arrowroot in St. Vincent (1858–1913) started as a smallhold crop, but was adopted by planters after 1875. By 1885, planters accounted for about half of production (Richardson, 1997, 156–157; Handler, 1971). Hence P lgit = 0 for t 1875, P lgit = 1/2 for t 1885 and we linearly interpolate ≤ ≥ 1875–1885. Lime in Montserrat (1867–1913): Sugar collapsed very early on and was not replaced by another export crop until much later when lime came on the scene. As a result, the government allowed former slaves to control both their cottages and their unusually extensive provision grounds. Furthermore, the largest landowner in Montserrat, Edmund Sturge, owner of the Montserrat Lime Juice Company, happened to be a Quaker and anti-slavery activist. Thus, even though lime was a plantation crop, provision grounds provided former slaves with an excellent outside option. See Hall(1971, 49–53). Hence P lgit = 0 for Montserrat lime and cotton. Lime in Dominica (1887–1913) was sharecropped (Trouillot, 1988), which we conservatively code as P lgit = 3/4. Oranges, pimentos and bananas (‘fruit’) in Jamaica (1853–1913): Oranges and pimentos were smallhold crops. Hence P lgit = 0 for oranges and pimentos. Bananas in Jamaica were initially a smallholder crop (Lewis, 1986, 72). In 1883–84, 90% of land holdings in the core of banana country were smallholds and during the 1880s there was a major increase in the number of local bank accounts, an indication of the rising prosperity of smallholders (Soluri, 2006, 146). However, in the 1890s, the United Fruit Company emerged first as a monopsony buyer and then as a grower.

66See Nugent and Robinson(2010) for a discussion of the primacy of politics over nature for understanding why coffee is sometimes a plantation crop (El Salvador, Guatemala) and sometimes a smallhold crop (Colombia, Costa Rica). 67The crops are sugar, cocoa, coffee, cotton, arrowroot, livestock, lime, and fruit. Fruit is largely a Jamaican aggregate consisting of bananas, oranges and pimentos. 68Years in parentheses indicate the years in which the crop accounted for at least 15% of total exports. This information is not used in coding P lgit. Online Appendix – Not for Publication

Hence for bananas P lgit = 0 for t 1885, P lgit = 3/4 for t 1895 and we linearly interpolate ≤ ≥ 1885–1895. Cocoa in Grenada (1861–1913), Dominica (1878–1913), Trinidad (1850–1913) and St. Lucia (1886–1913): Cocoa was primarily a plantation crop in Trinidad, St. Lucia and Tobago (Richard- son, 1997, 194; Sewell, 1861, 102; Bekele, 2004; Harmsen, Ellis and Devaux, 2012, 240–241). Hence P lgit = 3/4 for Trinidad, St. Lucia and Tobago. In Grenada, cocoa was “the ideal crop for smallholders” (Richardson, 1997, 194). While exact numbers on smallholder production are hard to come by, the 1853 Blue Book and the 1897 Royal Commission report that smallholder production was substantial. What is clearer is that the same highland geography that allowed cocoa to thrive also fostered the rapid growth of smallholder farms and villages, which in turn provided ample work off the plantation. See Brizan(1984, chapter 10), Richardson(1997, chapter 6), and the 1911 Blue Book for the Leewards (Great Britain, 1911, chapter 9). Turning to cocoa grown by planters, these usually used sharecropping contracts rather than a pure plantation form. In Dominica, cocoa was primarily a smallhold crop. Trouillot(1988, 95) cites testimony given to an 1884 Royal Com- mission that seven-eighths of the cocoa crop was produced on smallholds. Though this strikes us as an exaggeration, it does indicate a substantial smallholder presence. Hence P lgit = 1/4 in Grenada and Dominica. Coffee in Jamaica (1838–1896): Two thirds of production was by smallholds (Eisner 1961, 217; American and Foreign Anti-Slavery Society 1849, 97; Lewis 1986, 72). Hence P lgit = 1/3.A small amount of coffee was grown in Guyana and Dominica prior to 1860 and this was grown on plantations. Hence P lgit = 1 for t < 1860 and P lgit = 0 for t 1860 in these two colonies. ≥ Cotton in Monserrat (1905–1913), St. Kitts (1907–1913), St. Vincent (1905–1913) and Virgin Islands (1908–1913): This high-quality sea cotton was a smallhold crop that was promoted by English cotton manufacturers and government subsidies (Blue Books; Dookhan, 1975, 227-228). P lgit = 0. Finally, P lgit = 0 in all cases not listed above. Online Appendix Figure 2 visualizes the alternative series of Nit. The differences between ﬁgure1 and Online Appendix Figure 2 are mainly due to cocoa. Cocoa was a major export crop for Trinidad and St. Lucia (where plantations produced three quarters of all cocoa exports) and Grenada and Dominica (where plantations produced one quarter of all cocoa exports). The main smallhold crops varied by colony but include livestock in the Virgin Islands, arrowroot in St. Vincent, lime in Montserrat, cocoa in Grenada and Dominica, and fruits and coffee in Jamaica. This can be seen in more detail in Online Appendix Figure 3. Online Appendix – Not for Publication

Figure Online Appendix Figure 1: Smoothed and Unsmoothed Nit

Antigua Barbados Dominica 1 1 .95 1 .9 .95 .9 .85 .85 .8 .8 .8

.4 Petitions (Dashed) & Sales Petitions (Dashed) & Sales .2

0 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913

Grenada Br Guyana Jamaica 1 1 1 .95 .9 .85 .8 .8 .8 .75

.6 .6

.4 .4 Petitions (Dashed) & Sales Petitions (Dashed) & Sales .2 Petitions (Dashed) & Sales .2

0 0 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913

Montserrat Nevis St Lucia 1 .991 1 .96.97.98

.8 .8

.6 .6

.4 .4 Petitions (Dashed) & Sales Petitions (Dashed) & Sales .2 Petitions (Dashed) & Sales

0 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 St Kitts St Vincent Tobago 1 1 1

.8 .8 .8

.6 .6

.4 .4 Petitions (Dashed) & Sales Petitions (Dashed) & Sales Petitions (Dashed) & Sales .2

0 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 Trinidad Virgin Islands 1 1

.8 .8

.6 .6

.4 .4 Petitions (Dashed) & Sales .2 Petitions (Dashed) & Sales .2

0 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 Notes: This ﬁgure shows that the lowess-smoothed data in Figure1 everywhere faithfully reproduces the annual data Online Appendix – Not for Publication

Figure Online Appendix Figure 2: The Differential Decline of the Plantation Economy

The Share of Sugar in Total Exports 1.0 Antigua Barbados 0.9 St. Lucia Guyana 0.8 Tobaggo 0.7 St. Kitts Trinidad 0.6 Jamaica Dominica 0.5 0.4 St. Vincent 0.3 Grenada 0.2 Virgin Is. 0.1 Montserrat

Share of Plantation Crops in Total Exports Exports Total Crops in of Plantation Share 0.0 1838 1853 1868 1883 1898 1913

Notes: This ﬁgure reports the share of plantation crops in total exports. Nevis is not reported because it stayed above 0.99 in each panel. Also, Nevis merged with larger St. Kitts in 1883 and Tobago merged with larger Trinidad in 1899. Online Appendix – Not for Publication

Figure Online Appendix Figure 3: Major Export Crops of Colonies Where Sugar Declined

Virgin Is. - Livestock (solid) Cotton (dash) Montserrat - Lime (solid) Cotton (dash) St Vincent - Arrowroot (solid) Cott. (dash) 1 1 1 .8 .8 .8 .6 .6 .6 .4 .4 .4 .2 .2 .2 0 0 0 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 Dominica - Cocoa (solid) Lime (dash) Grenada - Cocoa Jamaica - Fruit (solid) Coffee (dash) 1 1 1 .8 .8 .8 .6 .6 .6 .4 .4 .4 .2 .2 .2 0 0 0 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 St. Lucia - Cocoa Tobago - Livestock Trinidad - Cocoa 1 1 1 .8 .8 .8 .6 .6 .6 .4 .4 .4 .2 .2 .2 0 0 0 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913 Notes: The ﬁgure displays the major export crops for colonies where sugar declined. The vertical axis is a crop’s share in total exports. The horizontal axis is time (1838–1913). The thin dashed black line is the export share of sugar. The thick solid red line is the export share of the most important non-sugar crops. A thick dashed green line is added where there is a second important non-sugar crop. The title of the panel names the colony and crops. Data are from the Colonial Blue Books. Online Appendix – Not for Publication

Figure Online Appendix Figure 4: World Sugar Production by Region and the British West Indies’ Share 20% BWI$Share$of$World$ 18%

16% 3,500,000 14% World Cane Sugar 12%

10% Beet Sugar British West Indies 8% 350,000 Cane Sugar 6%

2% BWI's Share of World Sugar Production Production Sugar World of BWI's Share Tons of Sugar (Log Scale) (Log Scale) of Sugar Tons 35,000 0% 1838 1853 1868 1883 1898 1913 1838 1853 1868 1883 1898 1913

Notes: The left-hand panel is the log output of sugar (measured in tons) by source: (1) cane sugar grown in our sample of 14 British West Indies sugar colonies, (2) cane sugar grown worldwide, and (3) beet sugar. The right-hand panel is the British West Indies’ share of world sugar output, i.e., (1) divided by (2)+(3). Data are from Deerr(1950). Online Appendix – Not for Publication

Online Appendix B Race and Institutions

Online Appendix B.1 Discussion and Data This paper is about the use of coercive policies. These policies were the product of a set of institutions that characterized post-slavery British West Indies society and the evolution of coercive policies were embedded in a broader process of institutional change. The reason this paper is about policies rather than institutions is because the distinction is subtle and because Caribbean institutions and their evolution cannot be readily quantified as variables in a regression. The hallmarks of these institutions have largely been touched on. The economic hallmark was sub-tropical plantation agriculture (see Engerman and Sokoloff(1997); Sokoloff and Engerman(2000) and footnote 14 above). The political hallmark was an executive, legislature, and judiciary that were dominated by a small number of planters.69 The social hallmark was a highly racialized environment. We develop this last point in some detail because it is amenable to quantification, and because it serves to buttress our main measure of planters’ relative power Nit. One cannot understand British West Indies institutions without understanding the demographics of race. In the U.S. Deep South on the eve of the Civil War, one in two people were white (Gibson and Jung, 2005). In contrast, for our 14 colonies in 1838, whites were a tiny mi- nority (6%) and white power stemmed solely from sugar profits and military support from the British government. As the plantation economy declined so too did the white population. By 1913 only 3% of the population was white. This decline was prominent in the minds of contemporaries. Froude(1888, ch.XVII) lamented that “the English of those islands are melting away. Families who have been for generations on the soil are selling their estates everywhere and are going off. Lands once under high cultivation are lapsing into jungle ... The white is relatively disappearing, the black is growing.”70 To European observers at the time, the exodus of whites was synonymous with the decline of the institutions that had until then characterized the British West Indies. Online Appendix Table 1 shows the share of whites by colony for the first and last years of our sample.71 Interestingly, almost all of the colonies where whites were more than 3% of the population in 1913 were colonies that stayed in sugar. Likewise, almost all of the colonies where whites were less than 1.5% of the population in 1913 were colonies with almost no sugar exports in 1913. Panel B of Online Appendix Table 1 shows that the share of whites correlated very strongly with Nit within almost every one of the colonies. In Online Appendix Table 3 we explore this in more detail, regressing each of our main variables on either the ranked share of whites in the population (0 for the smallest white share and 1 for the largest white share), or on quartile dummies of the distribution of white shares. These measures correlate strongly positively with the strength of the plantation system and with coercion, and negatively with wages. Carvalho and Dippel(2016) study the importance of race in the context of what was arguably the colonies’ most important institution, i.e., their powerful legislative assemblies. Planters completely dominated the assemblies in the early post-Emancipation years. In 1837, twenty-two of twenty-five Antigua assemblymen were planters, and similar proportions could be observed across the colonies (Green, 1976, 73). Carvalho and Dippel(2016) show that the planters’ ability to push coercive legislation through the assemblies began to erode with the emergence of colored

69 This is signiﬁcant because this small, cohesive group was able to overcome free-rider problems in the collective provision of coercion. 70Indeed, in almost every colony white populations declined absolutely, not just relatively. 71 Standard sources of demographic data are scarce. Of necessity, we therefore collected all of the colonial decadal censuses and built what is by far the most comprehensive database on Caribbean race demographics ever assembled. We interpolate linearly between decadal censuses, we note that there are no race data for Trinidad. See Online Appendix Table 2 for details. Online Appendix – Not for Publication

Table Online Appendix Table 1: Whites as a Share of the Population, 1838 and 1913

Bar- St. St. Guy- Anti- Jamai- St. Domi- Mont- Gre- Virgin bados Vincent Kitts ana gua ca Nevis Lucia nica serrat nada Islands Tobago 1913 6.9% 5.2% 4.9% 4.1% 3.0% 1.8% 1.7% 1.6% 1.2% 1.1% 1.0% 0.7% 0.5% 1838 12.4% 4.7% 7.3% 5.5% 5.3% 4.5% 5.2% 6.0% 3.7% 3.9% 2.5% 8.4% 2.0%

Panel B. Coefﬁcient of by-colony regression Nit = α + βWhite-Shareit Coeff. 1.5*** 8.4** 9.2*** 1.3*** 1.6*** 16.9*** 0.0 3.4*** 16.0*** 29.1*** 63.5*** 9.1*** 9.6*** [0.000] [0.022] [0.000] [0.000] [0.000] [0.000] [0.248] [0.000] [0.000] [0.000] [0.000] [0.000] [0.000]

Notes: Authors’ calculations based on census data. Nevis and Tobago data listed as 1913 are for 1882 and 1898, respectively. Panel B shows regression results for each colony’s time-series separately. p-values for robust s.e. in square brackets planters who, the authors argue, were more accountable to the local population. This discussion suggests that our ﬁnding of a causal impact of declining planter power (as measured by export shares) on coercive policies was part of a larger process of institutional change. Online Appendix – Not for Publication

Online Appendix B.2 Data Construction All the white share dummies and white share ranks have the expected signs. In terms of magnitudes for our preferred wage speciﬁcation (column 3, top panel), if a colony moves from the ﬁrst quartile (many whites) to the second, third or fourth quartiles (few whites), its log wage rises by 0.07, 0.20, and 0.35 log points, respectively. Thus, the white share dummies not only have the expected sign, but also have the expected increasing pattern. Turning to incarceration rates, incarceration rates fall by 0.86 percentage points (which is large compared to the mean incarceration rate of 1.1%).72

72 Interpreting the bottom panel, if a colony moves from the highest share of whites (rank of 1) to the lowest share (rank of 0) its log wage rises by 0.46 log points. Online Appendix – Not for Publication

Table Online Appendix Table 2: Share of Whites in the Population by Colony

White Share (%) Year Pre- Early post- Early 1838 1840s 1851 1861 1871 1881 1891 1901 1911 1913 Pre-1838 1840s post-1913 ATG 5.3 5.2 5.1 3.2 1821 BRB 12.8 11.6 10.9 10.2 9.3 8.6 7.0 1832 DOM 4.0 1.3 1.2 1.2 1833 GRD 2.6 1.3 0.9 1834 1921 GUY 3.5 3.6 2.8 2.4 1.8 1.5 1.3 2.3 1829 1841 1946 JAM 4.2 3.1 2.6 2.5 2.3 1.9 1844 MON 4.3 2.8 2.4 1.7 1.1 1828 NEV 5.4 2.6 1.8 1.4 0.8 1836 SLU 6.5 5.0 4.1 3.5 2.6 2.3 0.2 1836 1843 1946 STK 7.3 7.5 7.6 5.1 1837 STV 4.7 4.7 7.4 6.6 6.6 6.0 4.5 1825 1844 1931 TOB 2.3 1.6 0.8 0.6 - 1833 1844 VIR 8.4 8.0 5.6 1.8 1.0 0.7 0.7 1838 1841 1921

TRI 9.2 1837 Notes: 1. Grey shaded areas are colonies that disappear (Nevis merges with St. Kitts in 1883 and Tobago merges with Trinidad in 1899). 2. Censuses before 1838 and in the 1840s did not occur at regular times. The years of these censuses appear in the ‘Year’ columns. 3. We linearly interpolate all missing data. For data at the end points (e.g., 1838 and 1913) we must extrapolate and do so as follows. For Jamaica in 1838 we extrapolate back using the 1844–1861 change. For colonies missing 1913 but having 1911 data, we extrapolate using 1891–1911 growth rates. For colonies missing 1913 and 1911, we interpolate using post-1913 data, the year for which is indicated in the ‘Years’ column. For Tobago, we extrapolate to 1899 (when it merged with much-larger Trinidad) using 1861–1881 data. 4. Dominica, Grenada, Montserrat and St. Kitts each share a long gap in the data and it is possible that the decline of whites was not linear over these gaps. However, in the case of Grenada the censuses additionally report ‘birth by country’ data and these data change linearly. 5. There are no data for Trinidad between 1837 and 1946. In 1946 the white share was 2.7%. Online Appendix – Not for Publication

Table Online Appendix Table 3: The Exodus of British Whites and the Plantation System

N it : Plantation N it : Sugar Wages Incarceration (2) (2) (3) (4)

D(Share Whites Bottom Quartile)i -0.31*** -0.32*** 0.35*** -0.86*** (3.31) (3.29) (3.22) (3.86)

D(Share Whites 3rd Quartile)it -0.04 -0.10 0.20** -0.65*** (0.45) (1.06) (2.22) (3.83)

D(Share Whites 2nd Quartile)it -0.05 -0.10 0.07 -0.26* (0.47) (1.03) (0.73) (2.00)

Observations 942 942 844 737 2 R 0.871 0.844 0.744 0.533

Rank(Share Whites)it -0.38* -0.46** 0.46** -1.44** (1.90) (2.64) (2.52) (2.48)

Observations 942 942 844 737 2 R 0.849 0.815 0.734 0.532

Notes:(a) Each column reports on a regression with colony and year fixed effects. In the bottom panel the measure of the strength of the white planter elite is the ranked share of whites in the population (0 for the smallest white share and 1 for the largest white share). In the top panel the strength of the white planter elite is measured by quartile dummies of the distribution of white shares. The first quartile (white share > 6.1%) is omitted. (b) Trinidad is omitted for lack of data, which reduces the sample size relative to the baseline results. (c) Standard errors are clustered by colony. ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. t-statistics are in parentheses. Online Appendix – Not for Publication

Online Appendix C Soil Suitability

We worked with an agronomist to develop a suitability index at the fine spatial resolution described in section Appendix B.2 (8,144 equally sized cells). For each crop we construct a suitability index following the agronomic literature. For sugar we follow the index suggested in Jayasinghe and Yoshida(2010), which includes six factors important for sugar cane (temperature, rainfall, elevation, slope, soil pH, and soil texture). For example, rainfall in the range of 1100–1500 mil- limetres per year is highly suitable for sugar cane, rainfall in the ranges of 950–1100 or 1500–1990 is moderately suitable for sugar cane, rainfall in the ranges of 800-950 or 1990–2500 is moderately unsuitable for sugar cane, and rainfall in the ranges below 800 or above 2500 is highly unsuitable for sugar cane. For each factor, categorical suitability bins are constructed for comparability and the factors are then ranked in their importance, temperature, rainfall, elevation, slope, soil pH, and soil texture in declining order. The first four categories correlated quite strongly with each other, creating a clean spatial separation of suitability. For example, Online Appendix Figure 5 shows the topography of Jamaica and compares it with the distribution of suitability. Finally, a weighting is applied to them that reflects this ranking, with the weights summing to 1. We use the weights from Jayasinghe and Yoshida’s (2010) artificial neural network model. These weights make no use of Caribbean land use patterns (weights are based on Sri Lankan data). For each cell, the weights are used to aggregate the six factors into one of four characterizations of sugar cane suitability: highly suitable, only moderately suitable, unsuitable and totally unsuitable. The vast majority of the cells in our colonies are either highly or moderately suitable. Very little Caribbean land was totally unsuitable for sugar since the climatic conditions were so ideal. (The Jamaican highlands in Figure2 form an exception because they get exceptionally cold by Caribbean standards.) We therefore focus in the paper on land that is highly suitable or rather the share of land that is highly suitable. We understand well which agro-climatic conditions determine suitability for all of our other crops. Both the agro-climatic conditions and their mapping into suitability are crop-specific. This information is off-the-shelf available. Online appendix tables Online Appendix Table 4 and Online Appendix Table 5 show the suitability-ranges for each input for our four most important crops, as well as the sources for this coding. The variables given the climactic and soil conditions for each plot of land are, for cocoa, minimum temperature, maximum temperature, precipitation, number of months with precipitation less than 100mm, soil depth, soil pH, forest cover (cocoa trees prefer to be shaded), and elevation. For lime these variables are temperature, rainfall, soil pH, soil depth, soil texture (degree of sand and clay), slope, and elevation. For arrowroot, temperature, rainfall, soil pH, soil texture, and elevation. For livestock, meaning for the grasses used by livestock, temperature, rainfall, soil pH, elevation, slope, and humidity. For cotton, temperature, rainfall, soil pH, elevation, slope, and humidity. For fruit, temperature, rainfall, soil pH, soil depth, soil texture, slope, and elevation. The challenge for the West Indies was obtaining all necessary agro-climatic data at a fine enough spatial resolution. With the help of an agronomist we were able to do so. Data on elevation and slope are from the SRTM 90m Digital Elevation Database at http://www.cgiar-csi.org/ data/srtm-90m-digital-elevation-database-v4-1. Data on temperature and rainfall are from the GPCP Version 2.2 Combined Precipitation Data Set at http://www.esrl.noaa. gov/psd/data/gridded/data.gpcp.html. Data on soil texture and pH levels are from the FAO/UNECO World Soil Dataset http://www.fao.org/climatechange/54273/en/. As a point of comparison, we ended up with 61,421 cells of 604 square-meters for the British West Indies. In Online Appendix Table 4, we describe how each island i’s suitability for crop g is estimated Online Appendix – Not for Publication from i’s agro-climatic conditions as they affect g, with the sources for these codings listed in Online Appendix Table 5. Key is the large variation across islands in some geographic characteristics like elevation and soil. The importance of elevation for sugar suitability is illustrated by comparison of top and bottom panel of Online Appendix Figure 5, which shows that sugar suitable land in Jamaica was largely on the low-elevation coastal plains. The Caribbean is divided into three island chains: Most British Caribbean colonies–Dominica, the British Virgin Islands, Grenada, Montser- rat, Nevis, St. Kitts, St. Lucia, and St. Vincent–belonged to the inner chain of the Lesser Antilles, which is volcanic and mountainous. Jamaica was the only British colony in the Greater Antilles (the others are Cuba, Haiti and the Dominican Republic), which are large islands also with mountainous interiors. The outer chain of Lesser Antilles–Anguilla, Bahamas, Barbados, Turks and Caicos–consists of flat limestone. Limestone had the great advantage that it was flat, but most of the outer chain – with the exception of Barbados–was of such low elevation that it was very vulnerable to salination from storm surges. For our identification strategy, we proxy plantation land in 1838 with each island’s share of highly sugar suitable land. For this, we coded sugar suitability directly based on agricultural science literature. Specifically, we follow the artificial neural network method described in Jayas- inghe and Yoshida(2010). We do this because this estimation method makes no use of Caribbean land use patterns at all (it is based on Sri Lankan data) so that the weights and therefore the index are entirely exogenous to our data. The method boils down to ranking the six inputs and as- signing them exponential weights: The most important characteristic is assigned weight 0.408 = 1 1 1 1 1 1 1 1 1 1 1+ 2 + 3 + 4 + 5 + 6 2 + 3 + 4 + 5 + 6 6 ; the second-most important is assigned weight 0.242 = 6 , and so on. These weights sum to one so that the aggregate index lies between 1 and 4. Following convention, we then ‘re-bin’ the data by rounding it to the nearest integer so that the index takes on the values 1 (highly suitable), 2 (marginally suitable), 3 (marginally unsuitable), and 4 (completely unsuitable). The cross-sectional part of our first instrument Outside Option Caribbean Naval Expenditure i × t is an island’s land share that is not sugar suitable. See Figure2 for Jamaica. 73

73Guyana is the only colony that is not an island so that in calculating shares of land care must be given to the denom- inator. Guyana had a large hinterland of dense jungle and swampland, most of which was unsuitable for agriculture. We deﬁne Guyana’s historical border using the map in Higman(2000, Figure 1.8). We geo-coded this map and calculated Guyana’s sugar suitability share based on these borders. The original map and our geo-coding of it is displayed in Online Appendix Figure 6. Online Appendix – Not for Publication

Figure Online Appendix Figure 5: Sugar Suitability, and Topography in Jamaica

Notes: The top panel shows elevation for Jamaica. The bottom panel shows the spatial distribution of land that is sugar- suitable (black), only moderately sugar-suitable (dark grey), not sugar-suitable (light grey) and totally sugar-unsuitable (white). Online Appendix – Not for Publication

Table Online Appendix Table 4: Suitability Conditions by Crop

Sugar - Suitability Bins: (4) (3) (2) (1) (2) (3) (4) Temp (°C) <19 20-22 23-25 26-30 31-35 36-38 >38 Rainfall (mm) <750 750-1000 1000-1200 1200-1500 1500-2000 2000-2500 >2500 Soil ph <5 5.0-5.5 5.6-6.0 6.1-6.9 7.0-7.5 7.6-8.4 8.5 Elevation (m) <0 0-500 500-1000 >1000 Slope (%) <15 15-30 30-60 >60 Humidity 60-70 70-80 80-90

Lime - Suitability Bins: (4) (3) (2) (1) (2) (3) (4) temp (°C) <13 13-18 18.1-23 23.1-30 30.1-34 34.1-38 >38 rainfall (mm) <700 701-800 801-900 901-1200 1201-1900 1901-2500 >2500 soil ph <4 4-5.5 5.6-6 6.1-6.5 6.6-7.5 7.6-9 >9 soil depth (%) <50 50-69 70-89 90-100 loam (clay-; sandy clay-; silt sandy clay; silt soil texture sand loamy sand loam; sandy loam silt loam; silt clay-) clay; clay slope (degrees) 0-7 8-15 16-20 >20 elevation (m) <500 501-1000 1001-1800 >1800

Cocoa - Suitability Bins: (4) (3) (2) (1) (2) (3) (4) Min Temp (°C) <9 - restricted 9-13 13.0-17.99 18.0-20.99 21.0-23.99 24.0-24.03333 Max temp (°C) <24 24-26.99 27-29.99 30-32 32< Rainfall (mm) <800 800-999 1000-1199 1200-2499 2500-2999 3000-3999 4000< Rainfall: months with < 100mm <4 3-4 1-2 0 soil depth (%) <50 50-69 70-89 90-100 soil ph 4.5-5.99 6-6.5 6.51-7.00 7.0-8.0 8.0<

0, 1, 3, 4, 7, 11, land cover (classes) 13, 16 - restricted 5, 6, 9, 10, 12, 14, 8 2 elevation (m) <0 - restricted 0-300 301-500 501-1000 >1000 - restricted

Arrowroot - Suitability Bins: (4) (3) (2) (1) (2) (3) (4) Temp (°C) <17 17-20 20-23 23-29 30-32 32-34 >34 Rainfall (mm) <700 700-1000 1000-1500 1500-2000 2000-3000 3000-4000 >4000 soil ph <4 4-5 5-5.5 5.5-6.5 6.5-8 >8

Silt, Clay loam, Loam, Sandy Sandy clay loam, Sandy clay, Silty soil type Sand Loamy sand loam Silty loam Silty clay loam clay, Clay elevation (m) 0-90 90-500 500-900 >900

Notes: This table shows the agro-climatic inputs that determine suitability for our four most important crops (in four panels): sugar, lime, cocoa, and arrowroot. Each agro-climatic condition is inputs into bins determining what range is ideal (=1), relatively suitable (=2), less suitable (=3), or completely unsuitable (=4). For most inputs, the ideal range is in the middle and suitability is dropping off away from it in both directions, hence the arrangement of columns. Online Appendix – Not for Publication Table Online Appendix Table 5: Sources for Table Online Appendix Table 4 Jayasinghe, P.K.S.C. and Masao Yoshida, “Development of Two GIS-based Modeling Frameworks to Identify Suitable Lands for Sugarcane Cultivation,” Tropical Agriculture and Development, 2010, 54 (2), 51–61. 1. https://www.daf.qld.gov.au/plants/fruit-and-vegetables/fruit-and-nuts/other-fruit-crops/growing-cocoa 2. http://www.academia.edu/9129782/A_Multi- Criteria_GIS_Site_Selection_for_Sustainable_Cocoa_Development_in_West_Africa_A_case_study_of_Nigeria 3. http://link.springer.com/article/10.1007/s10531-007-9183-5 4. http://www.icco.org/about-cocoa/growing-cocoa.html 5. http://www.xocoatl.org/tree.htm 1. http://www.fao.org/nr/water/cropinfo_citrus.html 2. https://www.daf.qld.gov.au/plants/fruit-and-vegetables/fruit-and-nuts/citrus/citrus-land-andclimate-requirements 3. http://afghanag.ucdavis.edu/a_horticulture/fruits-trees/citrus/factsheets/FS_Fruit_Citrus_site_selection.doc 4. http://www.crec.ifas.ufl.edu/academics/classes/hos6545/pdf/mendel_1969.pdf 5. http://www.nda.agric.za/docs/Infopaks/CultivatingCitrus.pdf 1. http://jeromehandler.org/wp-content/uploads/2009/07/Arrowroot-History-71.pdf 2. http://www.prota4u.org/protav8.asp?h=M4&t=Maranta,arundinacea&p=Maranta+arundinacea 3. http://rfcarchives.org.au/Next/Fruits/Vegetables/Arrowroot9-96.htm 4. http://tropical.theferns.info/viewtropical.php?id=Maranta+arundinacea : This table lists the main sources that determined the coding of suitability bins in online appendix table Online Appendix Table 4 . Sugar: Cocoa: Lime: Arrowroot: Notes Online Appendix – Not for Publication

Figure Online Appendix Figure 6: Guyana

Notes: The left panel shows the historical boundaries of Guyana as dashed lines. The panel is from Higman(2000, ﬁgure 1.8). The right panel shows the results of geo-coding the map.

Figure Online Appendix Figure 7: Initial Conditions and Sugar’s End-of-Period (1913) Export Share

1.0 Antigua Barbados 0.8 Guyana Nevis St. Kitts 0.6

0.4 St. Lucia Tobago Trinidad

0.2 Jamaica

Share of Sugar in Total Exports, 1913 Total in of Sugar Share Virgin Is. St. Vincent Grenada Montserrat Dominica 0.0 0.0 0.2 0.4 0.6 0.8 1.0 Potential Outside Option (share of sugar-unsuitable land)

Notes: Each point is a colony’s Nit in 1913 plotted against its share of land that is unsuitable for sugar cane cultivation, which is a proxy for the share of all land that was unused in 1838. The −45◦ line provides a benchmark. Plantation shares for Nevis (Tobago) are extrapolated out to 1913 using 1882 (1898) data and growth rates from St. Kitts (Trinidad). Online Appendix – Not for Publication

Online Appendix D Export Price Data

We need crop prices for the gravity estimation in in Appendix B.2. For a given crop, prices (export unit values) are based on the export data of the largest exporter of the crop. This ensures that the prices are based on large volumes. Livestock: There are many types of livestock. We use the most important of these, namely cattle and horses. Using Virgin Islands exports we build price series separately for cattle and horses. Arrowroot prices are based primarily on export data from Jamaica (1838-1854) and St. Vincent (1855-1913). The St. Vincent Blue Book data are augmented by St. Vincent data reported in the Journal of the Royal Society of Arts (1873). Cocoa prices are based on export data from Granada and Trinidad. Cotton prices are based on export data from Guyana (1838–1844) and Granada (1856–1913).74 Bananas, oranges, pimento, coffee, lumber, and ginger prices are based on export data from Jamaica. For other spices (predominantly cloves, mace and nutmeg), prices are based on export data from Granada. Charcoal prices are based on export data from the Virgin Islands. Asphalt and balata (a natural tar) are based on export data from Trinidad and Guyana, respectively. Lime juice prices are based on export data from Jamaica (1838–61), Monserrat (1862–78), Dominica (1872–87), Monserrat (1872–78), and Jamaica (1891–1912). Because lime was exported in many different forms, we made extensive use of contemporary surveys of prices in order to develop a meaningful price series. Coconut prices are based on export data from Jamaica (1844–58), Tobago (1862–63), Tobago and Trinidad (1868–81), and Tobago, Trinidad and Jamaica (1882–1912). Finally, Online Appendix Table 6 reports the coefﬁcients from estimating the gravity equation (16) in Appendix B.2.

74This is the only crop that appears in Jacks(2005). Our price series is is highly correlated with the Jacks(2005) series except during the Civil War years. Online Appendix – Not for Publication

s Table Online Appendix Table 6: Estimates of the Gravity Equation (16) — Recovering ri(pt, τ )

θ ln(p gt /p st ) 2.30*** (3.93)

p s θ ln(τ g /τ g ) Pl git

g = cocoa 20.36*** g = arrowroot 16.43*** (3.24) (4.40) g = coffee 3.40*** g = lime 3.79 (5.45) (1.50) g = fruit 9.14 g = livestock -5.83 (1.71) (1.45)

Observations 7,126 R-squared 0.863

Notes: This table reports the parameters of interest from equation (16). The dependent variable is ln(xgit) − ln(xsit) where the s subscript stands for sugar. The first row is the co- p s efficient θ on ln pgt/pst. The remaining rows are the coefficients θ ln(τg /τg ) on P lgit. The number of observations is non-sugar crops (7) times colony-year pairs (1018). Online Appendix – Not for Publication

Figure Online Appendix Figure 8: Hurricanes and the Decline of Virgin Islands Sugar 1 .8 .6 .4 .2 (sugar exports) / exports) (total exports) (sugar .. . . 0 1838 1848 1852 1867 1871 1889 Notes: Vertical lines are hurricane dates. Sugar did not recover after 1889.

Online Appendix E Hurricanes

Online Appendix E.1 The Impact of Hurricanes Figure Online Appendix Figure 8 vividly displays the long-lasting effects of hurricane landfalls for the Virgin Islands, showing the evolution of sugar as a share of exports over time. Vertical lines are major hurricane dates. Destructive hurricanes in 1848 and especially 1852 decimated the industry. Because of the long-lasting effects of these early hurricanes, later ones could not further weaken an already collapsed plantation system. To operationalize the impact of hurricanes we need to capture the permanence of their impact and also the fact that a hurricane that hit an already decimated plantation economy could not do much further damage. We therefore constructed an instrument HDIit that assigned each Caribbean hurricane landfall a damage index. HDIit measures a hurricane’s effect on the sugar-economy as the two-year log change in sugar exports around the landfall year t and then stays constant thereafter.75 One concern with using hurricanes as an instrument is that they plausibly could have had direct effects on wages. How- ever, such direct effects would have been only contemporaneous to the years of or immediately after a hurricane, whereas we code the effect of HDIit as persistent by construction in order to capture hurricanes’ persistent effects discussed above. It turns out that we need to include hurricanes in our estimation in order to ﬁt the variation in Nit only for the Virgin Islands.

Online Appendix E.2 Details of Hurricane Damage Construction We coded up all hurricanes landfall on the islands during our period of study from the annual hurricane track maps published by the United States National Hurricane Center(2014) and, for

75The logic for using two-year changes is as follows. We know from our data that crops almost always bounced back within a year after a major storm. In the ﬁrst year there are declines in sugar exports both because of crop damage and because of infrastructure damage. In the second year, the crop comes back only to the extent allowed by the infrastructure damage so it is two years of depressed exports that speaks to the infrastructure damage. Let xsit be colony i’s sugar exports in year t and let t0 be the date of a hurricane so that ∆i(t0) ≡ ln xsi,t0−1 − ln xsi,t0+1 is the two-year log change in sugar exports. If sugar exports increased after the hurricane (∆i(t0) < 0) then we set ∆i(t0) to zero. Our hurricane damage index instrument is HDIit ≡ ∆i(t0) for t ≥ t0 and HDIit ≡ 0 for t < t0. Online Appendix – Not for Publication the pre-1851 period, by Tannehill(1938). Hurricane tracks were digitized and GIS software used to identify landfalls.

The hurricane damage index is max[ 0 , (ln xsi,t0 1 ln xsi,t0+1)] where t0 is the year of the − − hurricane and xsi,t0 1 and xsi,t0+1 are sugar exports by colony i in the years before and after the − hurricane. While the year of the hurricane is known, the resulting fall in exports is sometimes reported in the next Blue Book year because Blue Books do not necessarily report data on a calendar basis. Where the fall in exports happened in the year after the hurricane, we increment t0 by one year. This occurs for Jamaica (1874, 1880, 1896), Montserrat (1851), St. Lucia (1875, 1894), and St. Kitts (1910). For example, the Jamaican hurricane of 1896 is associated with sugar exports of £290,000 (1896), £219,000 (1897) and £261,000 (1898) so we set t0 = 1897. What follows is a list of the 28 hurricanes along with the year (t0), the two-year log change in sugar exports (∆i(t0)), and an indicator for whether it was a major hurricane (*). We group colonies into three groups:

1. Complete collapse of sugar: Virgin Islands 1848 (1.18*), 1852 (1.51*), 1867∗ (6.18*), 1871∗ (0.00), 1889 (0.00); Montserrat 1851 (0.64), 1889 (0.25), 1899∗ (0.00), 1909 (0.00); Grenada 1856 (0.00); Dominica 1883∗ (0.23*), 1903 (0.68); St. Vincent 1886 (0.48), 1898 (0.75). 2. Moderate decline of sugar: Tobago none; St. Lucia 1875 (0.00), 1894 (0.09); Trinidad 1878 (0.08); Jamaica 1874 (0.00 ), 1880 (0.00), 1886 (0.00), 1896 (0.10), 1903∗ (0.38*), 1910 (0.05). 3. No decline of sugar: Antigua 1910 (0.18); Barbados none; Guyana none; St. Kitts 1859 (0.0049), 1889 (0.08), 1908 (0.00), 1910 (0.15).

Looking across the 28 hurricanes that made landfall in the British West Indies during 1838–1913, what stands out is the Virgin Islands: It was hit repeatedly, it was hit early on, and the hits did major damage. There are four hurricanes with historical ambiguities. (1) Virgin Islands 1848: Dookhan(1975) claims that there was a hurricane in the Virgin Islands in 1842 and makes no mention of the 1848 hurricane. Using his hurricane dating rather Tannehill’s (1938) makes no difference to our results. (2) Barbados and Montserrat 1899: This hurricane was extremely powerful, causing many deaths in Montserrat and Barbados. The National Hurricane Centre track for this hurricane shows that it made landfall in Montserrat, but not Barbados. The hurricane had little impact on Blue Book exports in either location. To investigate potential measurement error we experimented with coding this hurricane as having a large damage index (1.00) in both colonies. This improves our results. (3) Antigua 1871: This hurricane overlapped with the disastrous drought of 1870–1874. See Berland, Metcalfe and Endﬁeld(2013, Figure 4). Contemporaries commented much more on the drought than on the hurricane (Berland et al., 2013, 1338). Indeed, there is no mention of the hurricane in the 1871 Antigua Blue Book listing of Parliamentary Acts, but there is a listing of an Act en- titled “An Act to raise the sum of £2500 for the Antigua Water Works.” Finally, when the drought ended, agricultural output completely rebounded: Sugar exports were £228,000 (1870), £239,000 (1871), £136,000 (1872), £153,000 (1873), £96,000 (1874, the worst year of the drought), and £243,000 (1875). In short, the cause of the decline in sugar exports was the drought, not the hurricane. We therefore code it as a 0. (4) Grenada 1856: This hurricane is assigned a small damage index because its impacts are confounded with the massive cholera epidemic that began in mid-June 1854 (too late to affect the 1854 sugar crop) and carried on through 1855. Sugar exports were £130,000 (1854), £82,000 (1855), £90,000 (1856), and £148,000 (1857). Hence ln xis,1856 1 ln xi,s,1856+1 < 0. − − If we assign this hurricane a large damage index (1.00) our results improve slightly. Online Appendix – Not for Publication

Online Appendix F The Impact of Productivity Growth on N ∗ and w

As discussed in detail in Online Appendix H.3, productivity grew much more rapidly in sugar than in smallhold crops. We therefore assume that Caribbean productivity increased in a way that raised plantation revenues from r(p, τ p) to ϕr(p, τ p) and raised smallhold revenues from r(p, τ s) to ϕδr(p, τ s) for some ϕ > 1 and δ (0, 1). Equation (2) now becomes ∈ w = ϕδr(p, τ s) C (17) − and equation (3) now becomes h i π(C,N; ϕ) = l(N) ϕr(p, τ p) ϕδr(p, τ s) + C Cγ/N . (18) − − It follows that

h p δ 1 s i p s πϕ = l(N) r(p, τ ) δϕ − r(p, τ ) > l(N)[r(p, τ ) r(p, τ )] > 0 − − δ 1 where the ﬁrst inequality follows from δϕ − < 1 and the second inequality follows from the fact that in an equilibrium with N ∗ > 0, revenues per standard plot size are higher on plantations than smallholds. Plugging this into the protection-for-sale welfare function (equation4) and differenti- ating shows that the optimal level of coercion C∗(N) does not depend on ϕ. From equation (17), ∂w/∂ϕ > 0. From equation (18), πN (N )(∂N /∂ϕ) + πϕ = 0 or ∂N /∂ϕ = πϕ/πN (N ) > 0 ∗ ∗ ∗ − ∗ where we have used πN (N ∗) < 0. Thus, both N ∗ and w rise. Online Appendix – Not for Publication

Online Appendix G Measuring Wages, Wage Frequency, and In-Kind Payments

In Montserrat, wages appear to have been commonly accompanied by access to a cottage or use of provision lands, for which we do not know the monetary value (Fergus, 1994). All of our readings suggested that this norm in Montserrat was in place throughout our period so that it is absorbed in the Montserrat colony fixed effect. Wages sometimes included an in-kind component that usually involved a cottage and a small plot. For two-thirds of our sample the Blue Books indicate whether the wage includes in-kind payments. Only 9% of these observations include in-kind payments. Panel A of Online Appendix Table 7 shows that our results are robust to an adjustment for in-kind payments. In a handful of cases, wages were reported as weekly or monthly, in which case we divided them by 5 or 20, as this was the explicit conversion stated in the Blue Books. Panels B and C of Online Appendix Table 7 show that none of our results depend on this specific conversion factor. Online Appendix Table 7 shows that none of our results depend on this specific conversion factor. In a few other cases, wages were reported as a range, in which case we used the midpoint. Wages are sticky so we also considered moving averages of 1 to 3 years. This strengthens our results. Online Appendix – Not for Publication

Table Online Appendix Table 7: Robustness Checks on Wage Variable h&g Panel A. Multiplying wit 1.2 when ”home and ground” provided h&g × h&g log wage - home & ground OLS IV log wage - home & ground OLS IV log wage - home & ground (1) (2) OLS (3) (4) (5) (6) IV (7) (8) (1)(1) (2)(2) (3)(3) (4)(4) (5)(5) (6)(6) (7)(7) (8)(8) Nit: Sugar Exports as -0.3714 -0.3839 -0.5971 -0.5891 Nit: Sugar Exports as -0.3714 -0.3839 -0.5971 -0.5891 Nit : ShareSugar of Exports Total Exports as [0.0078]-0.3714 -0.3839[0.0112] -0.5971[0.0043] -0.5891[0.0066] Share of Total Exports [0.0078] [0.0112] [0.0043] [0.0066] Share of Total Exports [0.0078] [0.0112] [0.0043] [0.0066] Nit: Plantation Exports as -0.5229 -0.5483 -0.9306 -0.9700 Nit: Plantation Exports as -0.5229 -0.5483 -0.9306 -0.9700 Nit : SharePlantation of Total Exports Exports as -0.5229[0.0027] -0.5483[0.0034] -0.9306[0.0011] -0.9700[0.0015] Share of Total Exports [0.0027] [0.0034] [0.0011] [0.0015] Share of Total Exports 2 [0.0027]2 [0.0034] 2 [0.0011]2 [0.0015] Time Controls t + t 2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe ObservationsTime Controls t908 + t 908t-fe t 908+ t t-fe908 t 908+ t t-fe908 t 908+ t t-fe908 Observations 908 908 908 908 908 908 908 908 FObservations Statistic (Instruments) 908 908 908 908 23.2832908 21.6359908 24.7038908 22.0467908 F Statistic (Instruments) 23.2832 21.6359 24.7038 22.0467 R-squaredF Statistic (Instruments) 0.724 0.771 0.731 0.779 23.2832 21.6359 24.7038 22.0467 R-squared 0.7240.724 0.7710.771 0.7310.731 0.7790.779 lower Panel B. Discounting Monthly and Weekly Wages More lowerlower loglog wagewage -- lowerlower boundbound OLS IV log wage - lower bound (1) (2) OLS (3) (4) (5) (6) IV (7) (8) (1)(1) (2)(2) (3)(3) (4)(4) (5)(5) (6)(6) (7)(7) (8)(8) Nit: Sugar Exports as -0.2957 -0.3093 -0.5208 -0.5292 Nit: Sugar Exports as -0.2957 -0.3093 -0.5208 -0.5292 Nit: Sugar Exports as -0.2957 -0.3093 -0.5208 -0.5292 ShareShare ofof TotalTotal ExportsExports [0.0527] [0.0711] [0.0133] [0.0211] Share of Total Exports [0.0527][0.0527] [0.0711][0.0711] [0.0133][0.0133] [0.0211][0.0211] Nit: Plantation Exports as -0.4434 -0.4635 -0.8564 -0.9261 Nit: Plantation Exports as -0.4434 -0.4635 -0.8564 -0.9261 Nit : SharePlantation of Total Exports Exports as -0.4434[0.0210] -0.4635[0.0288] -0.8564[0.0026] -0.9261[0.0037] Share of Total Exports [0.0210] [0.0288] [0.0026] [0.0037] Share of Total Exports 2 [0.0210]2 [0.0288] 2 [0.0026]2 [0.0037] Time Controls t + t 2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe ObservationsTime Controls t908 + t 908t-fe t 908+ t t-fe908 t 908+ t t-fe908 t 908+ t t-fe908 Observations 908908 908908 908908 908908 908908 908908 908908 908908 FF StatisticStatistic (Instruments)(Instruments) 23.2832 21.6359 24.7038 22.0467 R-squaredF Statistic (Instruments) 0.682 0.765 0.689 0.772 23.2832 21.6359 24.7038 22.0467 R-squared 0.6820.682 0.7650.765 0.6890.689 0.7720.772 upperupper Panel C. Discounting Monthly and Weekly Wages Less upper log wage - upper bound OLS IV loglog wagewage -- upperupper boundbound OLSOLS IVIV (1) (2) (3) (4) (5) (6) (7) (8) (1)(1) (2)(2) (3)(3) (4)(4) (5)(5) (6)(6) (7)(7) (8)(8) Nit: Sugar Exports as -0.4018 -0.4191 -0.6121 -0.5812 N it:: Sugar ExportsExports asas -0.4018-0.4018 -0.4191-0.4191 -0.6121-0.6121 -0.5812-0.5812 it Share of Total Exports [0.0040] [0.0047] [0.0038] [0.0075] ShareShare ofof TotalTotal ExportsExports [0.0040][0.0040] [0.0047][0.0047] [0.0038][0.0038] [0.0075][0.0075] Nit: Plantation Exports as -0.5576 -0.5845 -0.9419 -0.9459 N it:: Plantation ExportsExports asas -0.5576-0.5576 -0.5845-0.5845 -0.9419-0.9419 -0.9459-0.9459 it Share of Total Exports [0.0016] [0.0013] [0.0010] [0.0019] ShareShare ofof TotalTotal ExportsExports [0.0016][0.0016] [0.0013][0.0013] [0.0010][0.0010] [0.0019][0.0019] 2 2 2 2 2 2 2 2 TimeTime ControlsControls tt ++ tt2 t-fet-fe t + t2 t-fe t + t2 t-fe t + t2 t-fet-fe ObservationsTime Controls t908 + t 908t-fe t 908+ t t-fe908 t 908+ t t-fe908 t 908+ t t-fe908 Observations 908908 908908 908908 908908 908908 908908 908908 908908 F Statistic (Instruments) 23.2832 21.6359 24.7038 22.0467 F Statistic (Instruments) 23.283223.2832 21.635921.6359 24.703824.7038 22.046722.0467 R-squared 0.728 0.777 0.735 0.785 R-squared 0.7280.728 0.7770.777 0.7350.735 0.7850.785

Notes: In Panel A, to check the robustness to the provision of home and ground of this assumption, we multiply wages by a factor of 1.2 in all colony-years in which the Blue Books explicitly mentioned that ”home and ground” were included in the wage payments. The factor 1.2 comes from Sewell(1861, 32). Weekly wages in the Blue Books are consistently reported as being paid for weeks of ﬁve days, and monthly wages as being paid for months of twenty days. In Panels B and C we check for robustness to variations on those rules. It is possible that ﬁve-day work weeks and 20-day work months paid lip-service to Abolitionists but that effective works weeks and months were longer. In Panel B, we therefore divide weekly Wages by 1/6 and monthly wages by 1/24 instead of the 1/5 and 1/20 conversion rates suggested in the Blue Books. On the other hand, it may be that weekly and monthly contracts came with unreported in-kind payments such as right to a garden plot that would not have applied to day workers. In that case, real weekly and monthly wages would have been higher than daily wages. In Panel C, we therefore experiment with dividing weekly wages by 1/4 and monthly wages by 1/16 instead of the 1/5 and 1/20 conversion rates suggested in the Blue Books. Online Appendix – Not for Publication

Online Appendix H Additional Results and Robustness

Online Appendix H.1 The Historical Record on Differential Trends We begin with a simple question: “Have we missed any cross-colony differences emphasized by historians that might predict observed or unobserved differential trends?” The answer is yes. Pop- ulation density is the colony characteristic that is universally identiﬁed as central for understanding the post-Emancipation evolution of the British West Indies economies (Engerman, 1984, 133). To understand why, recall that the slave trade was abolished in 1807, more than a generation before the abolition of slavery in 1838. The lack of a slave trade limited the chief mechanism for equaliz- ing slave prices and marginal productivities across colonies. For example, Trinidad and Guyana turned to slave-based sugar cultivation only after becoming British in 1797 and 1803, respectively. Their post-1807 slave populations were therefore smaller than desired by planters. Correspond- ingly, their population densities were the lowest among our 14 colonies and their slave prices were the highest. As a result, pre-Emancipation population density and slave prices were widely understood to predict how wages and hence plantation success would evolve post-Emancipation, e.g., Merivale’s famous 1841 lectures (Merivale, 1861, 312–317). Modern historians have also focussed on population densities and slave prices, e.g., Engerman(1984) and Eltis, Lewis and Richardson (2005). The literature identiﬁes population density and slave prices, then, as the key colony characteristics for understanding the post-Emancipation evolution of our colonies. Importantly for our purposes, these initial colony characteristics are the most obvious source of unobserved differential trends in labour supply. Colonies with initially abundant labour may have had unobserved differential declines in labour supply while colonies with initially scarce labour may have had unobserved differential increases in labour supply.76 This observation means that the main factors emphasized by contemporaries and historians are labour supply shocks and so are factors that are likely to explain away are results. Since, as we will show, our results are not sensitive to controlling for these factors, it is unlikely that our results are sensitive to other unobserved differential trends in labour supply.

Online Appendix H.2 Differential Trends in Labour Supply We measure population density as the pre-Emancipation population divided by the area of the colony. We denote this by (Pop Density)1836. Data are from Martin(1839) and are based on pre-1838 Blue Books. Slave prices are for head slaves and are denoted by (Slave Price)1836. Data are also from Martin(1839) and are based on the reliable 1836 assessment of slave prices. We consider the possibility that our instrument Oi W t is correlated with (Slave Price) W t and · 1836 · (Pop Density) W t so that, as in point 3 above, Oi W t is correlated with a predictor of differ- 1836 · · ential trends in labour supply and TSLS is negatively biased. To eliminate this bias we include (Slave Price) W t or (Pop Density) W t in both the first 1836 · 1836 · and second stages. The new TSLS results appear in Panels A and B of Online Appendix Table 8. The first observation is that these variables are statistically insignificant. The second observation is that the estimated TSLS coefficients on Nit are very similar to our baseline results in Table4. This suggests that there is no conditional correlation between Nit and these predictors of differential trends in labour supply. It follows that our results are not explained away by the most important historical candidate for unobserved differential trends in labour supply.

76For example, over time labour may have illegally migrated from slave-abundant to slave-scarce locations. Online Appendix – Not for Publication

Noting that W t trends, it is possible that population density and slave price would have per- formed better with a more flexible specification of the differential trends. To investigate, we add flexibility by interacting (Pop Density)1836 and (Slave Price)1836 with a linear time trend t and including (Pop Density) t or (Slave Price) t in the first and second stages. The new TSLS 1836 · 1836 · results appear in Panels C and D of Online Appendix Table 8. The first observation is that these variables are also statistically insignificant. The second and more important observation is that the inclusion of these variables has no impact on our TSLS estimates of the coefficients on Nit. Again, there appears to be no correlation between Nit and these predictors of differential trends in labour supply, making it unlikely that our results are explained away by unobserved differential trends in labour supply. Online Appendix – Not for Publication

Online Appendix H.3 Other Possible Differential Trends The growth of industry, commerce, and urbanization all potentially drew labour away from the plantation economy. Differential growth in these acts as a differential plantation labour supply trend that could explain away our results. However, the historical record on these trends is very clear: Industry, commerce, and urbanization remained at very low levels throughout our period in every colony. We document this in Appendix B.1. To further examine this, Table A1 reports a regression of the share of the population employed in commerce on Nit and colony fixed effects. The coefficient on Nit is statistically insignificant. Likewise when the dependent variable is the share of the population employed in manufacturing. Panel B repeats the exercise with Nit replaced by our instrument Oi W t and the conclusions are the same: Employment in commerce and manu- · facturing are not correlated with our instrument. (See point 3 at the start of this section.) Thus, there is no evidence in the historical literature or data that differential trends in commerce and manufacturing could bias our IV results. As for urbanization, it did not begin until the very end of our sample, if not later. While urbanization statistics are scarce, in the most populous colony of Jamaica, Eisner(1961, 181–186) carefully documents that there was no trend in urbanization before 1913. The final potential differential trend we examine is productivity. There is a large literature on productivity growth in sugar. During our period all aspects of sugar processing became mech- anized, from cane crushing to the evaporation of juice into sugar crystals. This mechanization explains why labour productivity in sugar grew throughout our period, while there was no com- parable mechanization in any other Caribbean crop.77 What impact does unobserved productivity improvements in sugar have on our estimates? If there were unobserved differential trends in sugar productivity, these would work against us finding a negative impact of planter power on wages. Sugar productivity growth leads to more sugar output per unit of labour. This in turn shifts out labour demand, which increases both wages wit and the equilibrium quantity of labour. Higher productivity and more labour inputs both push for higher Nit. That is, both Nit and wit must rise. This is proved in Online Appendix F. The productivity- induced positive correlation between Nit and wit makes the TSLS estimate of the coefficient on Nit more positive than the ‘true’ coefficient. That is, it works against our finding. In summary, going through the list of candidate differential trends across colonies, we can in all cases show that they either did not exist, or do not affect our results, or would bias us away from finding a negative effect of Nit on wit.

77Important plantation crops all experienced productivity gains associated with improved varieties, identiﬁcation of ideal agro-climactic conditions, disease control, transportation, and post-harvest processing. However, the only explosive productivity growth was in post-harvest sugar processing. Compare the long list of inventions related to sugar in Deerr(1950, 534–590) to the inventions for cocoa and arrowroot in Knapp(1920) and Handler(1965). Note that sugar innovations all originated in Europe and the United States, largely in response to the 19th century development of beet sugar. Thus, the innovations were exogenous to the Caribbean. Online Appendix – Not for Publication

Table Online Appendix Table 8: Examining the Exclusion Restriction

Outcome w it : log wage C it : Incarceration (per Cap.)

N it : Sugar N it : Plantation N it : Sugar N it : Plantation t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe (1) (2) (3) (4) (5) (6) (7) (8) Panel A

N it (Share of Exports) -0.5518 -0.5617 -0.8449 -0.8777 1.2288 0.8559 1.5875 1.1269 [0.0001] [0.0109] [0.0000] [0.0006] [0.0004] [0.0299] [0.0000] [0.0059] {0.010} {0.039} {0.005} {0.035} {0.001} {0.012} {0.008} {0.02}

(Pop Density)1836,i x 0.0158 0.0042 0.0357 0.0278 0.0542 -0.0098 -0.0152 -0.0708 English Exports [0.5239] [0.9056] [0.0524] [0.2793] [0.1329] [0.8705] [0.4898] [0.1555] F Statistic (Instruments) 22.9 30.3 28.2 34.1 19.0 31.6 19.2 29.7 Panel B

N it (Share of Exports) -0.5605 -0.5564 -0.9413 -0.994 1.0563 0.8941 1.6755 1.4941 [0.0014] [0.0026] [0.0001] [0.0002] [0.0001] [0.0075] [0.0000] [0.0009] {0.064} {0.139} {0.047} {0.046} {0.004} {0.023} {0.009} {0.024}

Slave Price1836,i x -0.0004 -0.0008 0.0008 0.0008 0.0011 0.0021 -0.0024 -0.0016 English Exports [0.7992] [0.6609] [0.5067] [0.6173] [0.6686] [0.4259] [0.4686] [0.6214] F Statistic (Instruments) 24.5 22.1 17.7 13.7 21.9 21.8 16.8 15.7 Panel C

N it (Share of Exports) -0.5516 -0.5666 -0.8253 -0.8866 1.0544 0.7802 1.2852 1.0607 [0.0060] [0.0135] [0.0001] [0.0009] [0.0200] [0.0643] [0.0035] [0.0159] {0.0197} {0.0165} {0.0115} {0.0167} {0.0358} {0.1488} {0.0219} {0.1306}

(Pop Density)1836,i x t 0.0003 0.0001 0.0009 0.0007 -0.0001 -0.0009 -0.0024 -0.0025 [0.8015] [0.9502] [0.2053] [0.3617] [0.9621] [0.6880] [0.1539] [0.1502] F Statistic (Instruments) 26.7 28.9 36.2 35.5 21.3 28.3 24.0 30.5 Panel D

N it (Share of Exports) -0.5622 -0.5543 -0.9689 -0.9959 1.0568 0.8946 1.6753 1.4907 [0.0095] [0.0164] [0.0029] [0.0051] [0.0034] [0.0071] [0.0019] [0.0010] {0.096} {0.0824} {0.064} {0.0612} {0.016} {0.0382} {0.0143} {0.0339}

Slave Price1836,i x t -0.0000 -0.0000 0.0000 0.0000 0.0001 0.0001 -0.0000 -0.0000 [0.9226] [0.7069] [0.4448] [0.6846] [0.2004] [0.3410] [0.7991] [0.7652] F Statistic (Instruments) 21.8 21.0 14.4 13.2 20.0 20.8 15.0 15.1 Notes: Each panel in this table replicates Table4, but with the indicated additional regressor included both here in the second stage and in the ﬁrst stage (not reported). The additional covariates in Table4 are also included but not reported. Online Appendix – Not for Publication

Online Appendix H.4 Time Interactions with Other Geographic Characteristics In Table Online Appendix Table 9, we investigate the robustness of our main results in Table4 to allowing for a variety of geographic features having time-varying effects on our outcomes. To this end, we add in each panel an interaction between one geographic feature at a time (slope in Panels A and D, elevation in B and E, mean temperature C and F) with either the time-varying part of our instrument W t (in Panels A–C), or with a time trend (Panels D–F). Online Appendix – Not for Publication

Table Online Appendix Table 9: Time Interactions with Other Geographic Characteristics

Outcome w it : log wage C it : Incarceration (per Cap.)

N it N it : Sugar N it : Plantation N it : Sugar N it : Plantation Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe (1) (2) (3) (4) (5) (6) (7) (8) Panel A

N it (Share of Exports) -0.5184 -0.5933 -0.7553 -0.9404 0.8253 0.7056 0.8936 0.7863 [0.0760] [0.1078] [0.0001] [0.0003] [0.0209] [0.0662] [0.0398] [0.0849]

Slope i x 0.0043 0.0031 0.0032 0.0001 0.0038 0.0069 0.0002 0.0037 English Exports [0.6942] [0.8349] [0.6265] [0.9931] [0.7671] [0.6484] [0.9900] [0.7858] F Statistic (Instruments) 65.6 53.3 18.5 14.2 33.8 35.5 10.5 9.0 Panel B

N it (Share of Exports) -0.5000 -0.5837 -0.7533 -0.9388 0.8241 0.6397 0.9347 0.7880 [0.0928] [0.1160] [0.0001] [0.0004] [0.0165] [0.0938] [0.0161] [0.0837]

Elevation i x 0.0002 0.0001 0.0002 0.0000 0.0001 0.0000 0.0000 -0.0000 English Exports [0.5927] [0.7975] [0.4895] [0.9286] [0.7472] [0.9836] [0.9409] [0.9589] F Statistic (Instruments) 55.8 51.6 14.8 13.5 29.1 34.5 8.3 8.3 Panel C

N it (Share of Exports) -0.6820 -0.7835 -0.7781 -0.9588 0.7976 0.6941 0.9156 0.7977 [0.0043] [0.0183] [0.0000] [0.0000] [0.0618] [0.0979] [0.0382] [0.0804]

Mean Temperature i x -0.0004 -0.0005 -0.0001 -0.0002 0.0001 0.0002 -0.0003 -0.0001 English Exports [0.3305] [0.3593] [0.6476] [0.5241] [0.9346] [0.7794] [0.6076] [0.8325] F Statistic (Instruments) 39.8 25.1 7.3 5.8 30.1 24.3 5.9 5.1 Panel D

N it (Share of Exports) -0.6647 -0.7781 -0.7825 -0.9588 0.7850 0.6183 0.9208 0.7818 [0.0053] [0.0177] [0.0000] [0.0000] [0.0646] [0.1797] [0.0376] [0.1163]

Slope i x t -0.0000 -0.0000 -0.0000 -0.0000 0.0000 -0.0000 -0.0000 -0.0000 [0.4121] [0.3739] [0.7543] [0.5175] [0.9770] [0.9258] [0.5713] [0.6150] F Statistic (Instruments) 28.2 23.4 6.5 5.8 23.0 23.0 5.4 5.0 Panel E

N it (Share of Exports) -0.5440 -0.7643 -0.8388 -1.0035 0.7773 0.6208 1.1298 0.8183 [0.0009] [0.0067] [0.0000] [0.0001] [0.0025] [0.0656] [0.0000] [0.0359]

Elevationi x t -0.0023 0.0608 -0.0055 0.0116 -0.0186 0.0053 -0.0137 0.0509 [0.6518] [0.2573] [0.2380] [0.7735] [0.0030] [0.9484] [0.0353] [0.4583] F Statistic (Instruments) 48.3 48.8 8.6 8.9 41.3 100.6 7.7 13.3 Panel F

N it (Share of Exports) -0.6839 -0.7555 -0.8627 -1.0024 0.7340 0.5540 0.9521 0.7561 [0.0005] [0.0067] [0.0000] [0.0001] [0.0304] [0.1278] [0.0094] [0.0785]

Mean Temperaturei x t 0.0015 0.0017 0.0002 0.0003 0.0006 0.0009 0.0021 0.0020 [0.2476] [0.2799] [0.8044] [0.7706] [0.7921] [0.7276] [0.2567] [0.3406] F Statistic (Instruments) 52.6 47.1 9.0 8.4 94.6 102.3 12.8 12.9 Notes: Each panel in this table replicates Table4, adding as a robustness check an interaction between one geographic feature at a time (slope in Panels A and D, elevation in B and E, mean temperature C and F) with either the time-varying part of our instrument W t (in Panels A–C), or with a time trend (Panels D–F). Online Appendix – Not for Publication

Online Appendix H.5 Parish-Level Results Online Appendix H.5.1 Parish-Level Equivalent of Main Results Pushing skepticism further, it is possible that our results are spuriously driven by differential trends across colonies which can only be captured by a full set of colony-year fixed effects. This would mean that nothing can be identified with our colony-year data. Of course we do not believe this: The point of our British West Indies sample of sugar colonies is exactly that the colonies were virtually identical in all but a quantifiable set of agro-climactic factors and subject to common shocks. Nevertheless, we suspend our disbelief and now move to parish-level, sub-colony data so that we can include a full set of colony-year fixed effects.78 All of our colonies have parish-level administrative units. We observe data for just over 100 parishes in 11 colonies for a total of 2,240 observations.79 There are no data for Guyana, Trinidad or the Virgin Islands. The number of parishes by colony and year appear in Online Appendix Table 10. Because our agricultural suitability data are at the sub-parish level, we can compute Oi at the parish level simply by geocoding parish boundaries. For parish p in colony i, let Opi be the parish counterpart to Oi. While we do not have wage or coercion data at the parish level, we do have mortality rates. We will find that the weakening of the plantation system led to a fall in mortality rates and interpret this to mean that the weakening led to some combination of higher wages and lower coercion. Mortality rates are defined as deaths divided by population. Data are from the Blue Books. During the course of the 1860s, possibly at the behest of London, colonies switched from reporting burials to reporting deaths, the latter being more comprehensive. We therefore always include a burial dummy Dpit which equals 1 if parish p in colony i reports burials in year t. We measured the colony-level strength of the plantation system using colony-level exports (Nit). We measure the parish-level strength of the plantation system using the parish-level share of the population employed in plantation agriculture. This is reported in the Blue Books, but only imperfectly. Online Appendix H.5.2 describes how the employment data were partially interpo- lated. We would like to know that the parish-level plantation share of employment is correlated with our country-level Nit, but our cleanup is sufficiently conservative that we can only recover 121 colony-year employment totals. For these observations employment shares and Nit are highly correlated. See Table A1 and the surrounding discussion in Appendix B.1. As a second check on the plantation employment data, we exploit the more limited parish-level Blue Book data on the share of cultivated acres that are planted in sugar. This share is also highly correlated with the plantation share of employment. See Online Appendix Table 12. With parish-level data on mortality rates (Mpit), plantation employment as a share of the population (Npit), and the burial dummy Dpit we estimate

Mpit = βNpit + Dpit + λit + λpi + εpit (19) and instrument Npit with (Opi W t). Note crucially that we are now able to include colony-year · ﬁxed effects (λit). We also include parish ﬁxed effects (λpi) and use two-way clustering at the colony-year and parish levels.

78 To ﬁnd effects at the parish-level, it also needs to have been the case that planter power impacted conditions locally at the parish-level within a colony. The discussion in Sections 2.1–2.2 provides evidence for this ‘parochial’ nature of legal coercion. 79 Parish boundaries of a few colonies change over time; however, it is easy to amalgamate parishes so that parish boundaries are time-invariant. Online Appendix – Not for Publication

Online Appendix Table 11 reports the OLS, first-stage, reduced-form and IV estimates. Even- numbered columns include colony-year fixed effects. For comparability with previous results the odd-numbered columns report the specification with linear and quadratic time trends. Columns 1–2 report the OLS estimates of equation (19). These results are strong and indicate that a weakening of the plantation system within a parish led to a reduction in mortality rates. The table also reports the first stages (columns 3–4), the reduced forms (columns 5–6) and TSLS (columns 7–8). Online Appendix – Not for Publication 3 3 3 333333333 333333333333333333333 5 555 55555555 777777777 7 7777777 21 21 14 14 14 14 14 14 14 14 14 14 14 14 14 9 999999999999999999999 666 6 6 6 6666 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 77777777433333339 33333333 933333333333 3 96666666666666666666666666 3 3 977777777777777777 7777777777 333 3333333333333333333333333 5555555555555555555555555555555 999999999999999999999999999 9 77777 444444 99999999 4444 555555544 1101111111 116 11111111111111 111110 111111 3333333333333 3 3 333 33333333 1111 11 11 8 11 11 11 11 11111111111111111111111111111111111111111111 11 1111119 11 11 1111111111 11 11 11 1111 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 Table Online Appendix Table 10: Number of Observations per Colony and Year for Parish-Level Results 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 5555555555555555555555555554999999999999 44444444444 4 9 5 4444444444444 555555 4 9 3 9 9 66666666666666 6 6 656 10 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 20 20 20 20 21 21 21 21 21 21 21 21 21 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 : This table lists the number of parishes (or groups of parishes) per colony per year used in the parish-level results in Table Online Appendix Table 11 . Jamaica St Kitts St Vincent Tobago Antigua Barbados Jamaica Montserrat Nevis St Lucia St Kitts St Vincent Tobago Nevis Dominica Grenada Montserrat Nevis St Lucia Dominica Grenada St Lucia St Kitts St Vincent Tobago Dominica Grenada Antigua Barbados Antigua Barbados Jamaica Montserrat Notes Online Appendix – Not for Publication

Table Online Appendix Table 11: Parish Level Results

OLS First Stage Reduced Form TSLS

Outcome (Deaths / Pop)pit N pit (Deaths / Pop)pit (Deaths / Pop)pit (1) (2) (3) (4) (5) (6) (7) (8)

N pit : (Plantation Emp)pit 0.0383 0.0377 0.0843 0.1077

/ (Parish Pop)pit [0.0000] [0.0000] [0.0854] [0.0714]

O pi x (English Exports -0.1774 -0.1717 -0.0150 -0.0152

to Non-Caribbean)t [0.0695] [0.0669] [0.0536] [0.1066]

Burials D pit -0.0095 -4.5527 0.0954 -124.8752 -0.0060 -8.7848 -0.0141 [0.0000] [0.0000] [0.0006] [0.0000] [0.0018] [0.0000] [0.0053]

Time Controls t-fe col*t-fe t-fe col*t-fe t-fe col*t-fe t-fe col*t-fe Observations 2,145 2,184 2,145 2,184 2,145 2,184 2,145 2,104 R-squared 0.618 0.733 0.830 0.889 0.596 0.719 0.580 0.679 F Statistic (Instruments) 3.379 6.051

Notes:(a) Columns 1–2 report the OLS regression of equation (19), i.e., mortality rates Mpit on plantation employment shares Npit . Columns 3–4 regress Npit on the parish-year-varying instrument (Opi · W t). Columns 5–6 present the corresponding reduced form. Columns 7–8 present the TSLS estimates. (b) All specification include parish fixed effects λpi.(c) For each outcome, we present the results for our two preferred specifications, i.e., the second and third specifications in the preceding tables. In even-numbered columns these include colony-year fixed effects. In the odd-numbered columns these include time trends so we include (but do not report) the colony-year varying covariates that appear in Table1.( d) Standard errors are two-way clustered by parish and by colony-year. p-values appear in square brackets. Online Appendix – Not for Publication

Online Appendix H.5.2 Parish-Level Results on Employment Parish-Level Employment and Mortality Data: As a way of further validating the use of agricultural employment as a measure of planter power, we can turn to land use data, which is also reported at the parish-level in the Blue Books. The Blue Books reported information on land-use by crop as well as total acreage under crop. Sugar is the crop most consistently reported. The share of sugar-acres in total acres under cultivation is 41 percent on average in our data. Across columns, Online Appendix Table 12 uses the parish- level-reported acreage under crop as an outcome. In columns 1–2 and 4–5, this is aggregated to the colony-level, in columns 3 and 6, this is done at the parish-level. In columns 1 and 4 of Online Appendix Table 12 we relate Nit to the share of sugar in all acres under cultivation, conditional on only colony fixed effects. In columns 2 and 5, we relate the employment share in agriculture to the share of sugar in all acres under cultivation, conditional on only colony fixed effects. Columns 1 and 4 relate a colony-level-reported measure (Nit) to a parish-level-reported one that is aggregated up to the colony level. Like the employment data, land-use is often reported for some parishes and not others so that we can only infrequently aggregate land-use to the colony level, hence the number of observations is roughly the same as in Table A1. Columns 2 and 5 have the lowest number of observations because they relate two variables that are parish-level-reported but then aggregated to the colony. The problem of incomplete parish-coverage therefore gets compounded. Columns 3 and 6 have the highest number of observations because we use the parish as the unit of observation. We deleted repeat-reporting of the same values, leaving only actually updated data. In total, we observe 543 reported employment figures in eleven of the fourteen colonies, while Guyana, Trinidad and the Virgin Islands had no usable parish data. We then in- terpolated the data between the first and last data-entry, using STATA’s ipolate command. We never extrapolated which is important because employment data sometimes started late in our sample period. For example, the first reported employment data in Jamaica is from 1871 and we do not extrapolate to earlier years. In columns 1 and 4, a fifty percentage point decline in Nit is associated with a 28 percentage point reduction in sugar’s share in cultivated acreage, i.e., roughly two-thirds of its mean. Alter- natively, in log terms it is associated with an 80 percent reduction in column 4. In columns 2–3 and 5–6, a fifty percentage point decline in the population-share of those employed in agriculture lowers sugar’s share in cultivated acres by around 18 percentage points. Partial correlations are much more significant in columns 3 and 6 than in columns 2 and 5 because of the higher sample size, but the point estimates are reassuringly similar across these columns. Online Appendix – Not for Publication

Table Online Appendix Table 12: Partial Correlation between Nit, Agricultural Employment, and Land-Usage

(1) (2) (3) (4) (5) (6) Outcome: Share: Sugar / All Acres in Crop log (Sugar Acres)

N it : Sugar Exports as 0.5696 1.6790 Share of Total Exports [0.0000] [0.0081] Share: Agr Empl / Pop. 0.3162 0.3685 1.0162 1.2672 [0.3608] [0.0000] [0.4405] [0.0004] Unit of Observation: colony-year parish-year colony-year parish-year fixed effects: colony colony parish colony colony parish Observations 106 28 816 106 28 816 R-squared 0.898 0.931 0.842 0.978 0.876 0.815

Notes: Columns 1–2 and 3–4 report correlate variables at the colony level, columns 5–6 at the parish-level. In columns 1–4 standard errors are clustered at the colony level, in columns 5–6 at the parish-level. P-values are reported in square brackets. Online Appendix – Not for Publication

Online Appendix H.6 Further Robustness Checks In this section we report on a number of additional robustness checks. Population weighting: Comparing the 1838 populations of the largest and smallest colonies, Ja- maica was 48 times larger than Montserrat. We therefore re-estimate the TSLS speciﬁcations of Table4 weighted by population. The results appear in Panel A of Online Appendix Table 13. Population weighting leads to only minor changes. Hurricane Alternative: One may be concerned that our hurricane damage index HDIit should be constructed independently of any economic data. In Panel B of Online Appendix Table 13 we 80 therefore replace HDIit with a simple set of indicators for when a major hurricane hits an island. We obtain almost identical IV results with this alternative coding. Dynamic Panel: To accommodate the dynamic aspects of our panel we modify the wage equation (7) by adding the lagged dependent variable on the right hand side:

w w w w w w wit = ρ ln wi,t 1 + βN Nit + γX Xit + λi + λt + it . (20) − · · Likewise for the coercion equation. In the NBER version of this paper we presented most of our results using this dynamic speciﬁcation. Here we report it only in Panel C of Online Appendix Table 13. For comparability with our other results, we report the long-run coefﬁcients βw /(1 ρw) N − and βc /(1 ρc). These are about one-third smaller than our baseline results. N −

80 These are cumulative: If a major hurricane hits an island, the variables turns from 0 to at least 1 for every year thereafter. If a second major hurricane hits, it turns to 2, etc. ‘Major’ is deﬁned by the United States National Hurricane Center and corresponds to class 3, 4 and 5 hurricanes. See Online Appendix E. Online Appendix – Not for Publication

Table Online Appendix Table 13: Robustness Checks

Outcome w it : log wage C it : Incarceration (per Cap.)

N it : Sugar N it : Plantation N it : Sugar N it : Plantation t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe (1) (2) (3) (4) (5) (6) (7) (8) Panel A

N it (Share of Exports) -0.5729 -0.5627 -0.9225 -0.9595 1.0847 0.8771 1.6613 1.4357 [0.0067] [0.0104] [0.0021] [0.0029] [0.0022] [0.0047] [0.0005] [0.0003] {0.043} {0.071} {0.044} {0.068} {0.006} {0.009} {0.011} {0.016} Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe Observations 908 908 908 908 798 798 798 798 F Statistic (Instruments) 24.2 22.1 24.7 22.5 23.1 24.4 19.5 21.9 Panel B

N it (Share of Exports) -0.5437 -0.5881 -0.8359 -0.9614 0.7887 0.6354 1.1343 0.9499 [0.0069] [0.0076] [0.0001] [0.0001] [0.0150] [0.0309] [0.0012] [0.0048] {0.059} {0.059} {0.033} {0.029} {0.042} {0.075} {0.011} {0.036}

Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe Observations 908 908 908 908 798 798 798 798 F Statistic (Instruments) 55.7 52.7 9.2 8.4 45.5 51.8 8.2 7.9 Panel C

Lagged Dep Var: w i,t -1 or C i,t -1 0.7679 0.7652 0.7610 0.7537 0.6294 0.6385 0.6219 0.6328 [0.0000] [0.0000] [0.0000] [0.0000] [0.0000] [0.0000] [0.0000] [0.0000]

Long-Run Coefficient on N it -0.4316 -0.3865 -0.6360 -0.6209 0.6666 0.4480 0.9827 0.7015 [0.0448] [0.0416] [0.0394] [0.0448] [0.0204] [0.0372] [0.0233] [0.0346]

Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe Observations 859 859 859 859 725 725 725 725 F Statistic (Instruments) 16.5 15.6 12.2 9.8 17.8 19.3 11.5 11.9 Notes: Each panel in this Table replicates Table4 with a robustness check added. In Panel A we weight by population. In Panel B we replace the hurricane instrument HDIit with a simple set of indicators. In Panel C, we estimate a dynamic panel regression, adding the lagged dependant variable on the right-hand side. See equation (20). The ‘Lagged Dep w c w w c c Var’ row reports ρ or ρ . The ‘Long-Run Coefficient’ row reports βN /(1 − ρ ) or βN /(1 − ρ ). Significance levels for long-run coefficients in Panel C use score-based F tests. The square root of these F -tests appear in square brackets. Online Appendix – Not for Publication

Online Appendix I Extensions

Extension I: Dispensing with Nit altogether and instead instrumenting coercion Cit with Oit, we obtain the causal effect of outside-option-reducing coercion on wages under the assumption that Oit impacts wit only through Cit:

W W W W W wit = β Cit + γ Xit + λ + λ + , (21) C · X · i t it where Cit is instrumented with Oit. The associated first-stage equation is exactly the reduced-form relation in equation (11), which is reported in columns 7–8 of Table3. Similarly, the reduced-form equation corresponding to (21) is the same as equation (10), which is reported in columns 5–6 of Table3. Online Appendix Table 14 reports the results of estimating (21) using IV. The number of observations for which we observe both Cit and wit is 750. In both specifications, the results imply that an increase in incarceration rates of about one person per hundred in a year (slightly less than the mean incarceration rate in the data) decreases wages by about sixty percent. This is consistent with the results reported in Table4, where the estimated coefficient of the declining plantations system on wit was about sixty percent of that on Cit (i.e., .05736 in columns 2 over 0.8869 in − column 6 of Table4). Online Appendix – Not for Publication

Table Online Appendix Table 14: IV Effect of Coercion (Cit) on Wages (wit)

Outcome w it : log wage (1) (2)

C it : Incarceration (per Cap.) -0.5427 -0.6535 [0.0008] [0.0003] {0.046} {0.022}

2 Time Controls t + t t-fe Observations 750 750 F Statistic (Instruments) 23.762 14.751

Notes:(a) This table reports on the IV results of estimating equation (21). 750 is the number of observations for which we observe both Cit and wit. All controls that were included columns 2, and 4 in Table1 are also included here. The ﬁrst-stage relation corresponding to above results is that reported in columns 7–8 of Table3.( b) Standard errors are clustered by colony, p-values in square brackets; p-values for wild-bootstrapped standard errors are shown in curly brackets. Online Appendix – Not for Publication

Extension II: Instead of dispensing with Nit, we now take the opposite direction by attempting to estimate what portion of the effect of Nit on wit works through Cit and what portion affects wit ‘directly’ (i.e., through other channels). A decomposition like this is done through a ‘mediation analysis,’ which asks to what extent the ‘total effect’ of a treatment variable (e.g. Nit) on a ‘final outcome’ (e.g. wit) is explained by an intermediate outcome or mechanism (e.g. Cit) that is also an outcome of Nit. See for example Imai et al.(2010, 2011); Heckman et al.(2013); Heckman and Pinto(2015b). We fix ideas using Online Appendix Figure 9. The Base-Model in Online Appendix Figure 9 represents the IV estimations reported in Table4 of Section4. In them, the relationship between Nit and the outcomes Cit and wit is confounded by unobservables (the V ), but Oit is both plausibly independent of these unobservables and a powerful predictor of Nit, thus constituting a valid instrument. Extension I in Online Appendix Figure 9 disregards Nit altogether and instead uses Oit as an instrument for Cit on wit. Extension II applies a mediation analysis to this IV setting. Under certain assumptions, Extension II can be estimated using just the instruments at hand and we can ask to what extent the plantation system reduced wages because it raised coercion.81 It is important to note that Extension I and Extension II are both consistent with the standard IV identification assumptions of the Base-Model in Section4. However, Extension I and Extension II are not consistent with each other. This is because Oit can only be a valid instrument in Extension I under the exclusion restriction that it impacts wit only through Cit, which is the assumption that is being relaxed in Extension II. Extension I and Extension II should therefore be viewed as extending the empirical analysis from Section4 into different directions, by making different assumptions about unobservables.

81 Without additional dedicated instruments for Cit, identiﬁcation in Extension II requires that confounders that bias the relationship between Nit and wit work only through Cit. In other words, the relationship between the plantation system and wages is biased only because of confounders that impact wages through coercion. While it is a strong assumption that there are no unobservables that directly impacted wages and the plantation system, it is not implausible given that we control for all historically prominent labor demand and supply shocks during this period. Online Appendix – Not for Publication

Table 1: General Mediation Model and Our Solution to the Identiﬁcation Problem Figure Online Appendix Figure 9: Extending the Analysis A. Directed Acyclic Graph (DAG) Representation

Base-Model: Extension I: Extension II: IV for N C,W IV for C W IV Mediation Model → →

V V 0 V V 0

O N C,W O C W O N C W

B. Structural Equations

N = fN (O, V, N ), N = fN (O, V, N ) C = fC (O,V 0, C ), C = fC (N, V, C ), C = fC (N,V,V 0, C ) W = fW (C,V 0, W ), W = fW (N, V, W ), W = fW (N,C,V 0, W ) O,V 0, 0s Stat. Indep. O,V, 0s Stat. Indep. O,V, V 0, 0s Stat. Indep.

TheNotes first: column The first of column PanelA of gives Panel the A, labeled DAG representation the Base-Model of, gives the two the IVDirected Models, Acyclic which Graphs enable(DAG) the identification representation of of the causalthe two effects IV estimations of T on M inand sectionY . The4. Extension third column I is the of IV Panel model A presents where C theis instrumented General Mediation directly Model with O. withExtension an instru- II mentalis anvariable IV MediationZ. In Model the third that column, allows for identification the identification can only of the be causal achieved effect with of N anon additionalW as it is mediated designated by instrumentcoercion forCM. Extension(not depicted). I and Extension The fourth II are column both consistent of Panel A with presents the identification the Restricted assumptions Mediation of Model the Base-Model with an. instrumental However, variableExtensionZ. This I and modelExtension enables II are the not identification consistent with of each the total, other: theO can direct only and be the a valid indirect instrument effect of forTConinY.ExtensionThe model I alsounder enables the exclusion the identification restriction of that the it causal impacts effectW only of T throughon M Cas. wellPanel as B presentsthe causal the effect nonparametric of M on Y. equationsPanel B of presents each themodel. nonparametric We refer to structural Heckman equations and Pinto( of2015a each) for model. a recent Conditioning discussion on variables causality are and suppressed DAGs. for sake of notational simplicity. U, discussed in the text, is included in the equations but omitted from the DAGs for expositional clarity. The full DAGs, which include U, are in section ??. We refer to Heckman and Pinto(2015a) for a recent discussion on causality and directed acyclic graphs.

“The fact that the wage level in the capitalist sector depends upon earnings in the

subsistence sector is of immense political importance, since its effect is that capitalists

have a direct interest in holding down the productivity of the subsistence workers.

Thus the owners of plantations, if they are inﬂuential in the government, are often

found engaged in turning the peasants off their lands.” — Lewis(1954)

1 Introduction

1 Online Appendix – Not for Publication

We estimate Extension II using the following equation

w|N w|N w|N w|N w|N w|N wit = β Cit + β Nit + γ Xit + λ + λ + . (22) C · N · X · i t it w|N Equation (22) allows coercion to affect wages (βC < 0) and allows the plantation system to affect w|N wages through channels other than coercion (β = 0). It cannot be estimated by OLS when Nit N 6 and Cit are endogenous. However, Pinto et al.(2019) show that (22) can be estimated using IV, as long as the assumptions on the error term that are laid out in Extension II of Online Appendix Figure 9 hold. In that case, the causal effect of Cit on wit can be estimated by instrumenting for Cit 82 with Oit while including Nit in both the second-stage and first-stage equations. An intuition for the identifying variation in Extension II can be gained by observing that conditional on Nit, the unobserved confounder V is in principle an instrument for Cit in Extension II because it enters the system only through Cit. Furthermore, although Oit has to be uncorrelated with V (in order to have been a valid instrument in the Base-Model), this need not be true once we 83 condition on Nit, an insight that is basically an application of Simpson’s Paradox. To identify the causal effect of Cit on wit in Extension II, we estimate equation (22) using IV. The ˆw|N resulting estimate βC in Online Appendix Table 15 suggests that when the plantation system increased incarcerations by one person per hundred this lowered agricultural wages by between 30-40 percent (e.g. 0.33 = e 0.4070 1 in column 1). − − − The question that is ultimately posed in Extension II is how much of the effect of Nit on wit is ˆw|N explained by the effect of Nit on Cit. To answer it, we need only combine βC from equation (22) ˆC ˆw with βN and βN , the coefficient estimates from equations (7) and (8), which we reported in Table4. C w|N w The share of the effect of Nit on wit that operates via coercion Cit is defined as (βˆ βˆ )/βˆ . N × C N This share is reported at the bottom of Online Appendix Table 15. For example, we estimated ˆw ˆC βN = 0.5736 and βN = 0.8869 in columns 2 and 6 of Table4. Combining this with the estimate w|N− of βˆ in column 2 of Online Appendix Table 15, we get 0.856 = (0.8869 .0.5496)/( 0.5736). C × − Across specifications, this type of calculation suggests that between 75 and 90 percent of the effect of the declining plantation system on rising wages wit operates via lowered coercion as measured by Cit. Summarizing the additional insights gleaned in this section, Extension II makes stronger identifying assumptions than Extension I but also sheds additional light on the workings of coercion, enabling us to uncover evidence that is consistent with the motivating quote by Lewis(1954). Further, while Extension I and Extension II make different identifying assumptions, the evidence generated by the two models is nonetheless highly consistent: Extension I, which assumes that Oit impacts wages only through measured coercion Cit, implies that an increase in incarceration rates of one person per hundred in a year decreases wages by about 55–65 percent. Extension II, which allows Oit to impacts wages through channels other than Cit, implies that this same increase in

82 The ﬁrst-stage equation is

C|N C|N C|N C|N C|N C|N Cit = βO · Oit + βN · Nit + γX · Xit + λi + λt + it . (23)

83 For example, suppose that—as we argued in footnote 38 — the unmeasured inﬂuence of local missionaries is a good candidate for V . Missionaries fostered resistance against the plantation system, thereby increasing incarcerations Cit and depressing Nit, potentially leading to downward-biased OLS estimates for equation (8). While the unobserved presence of missionaries is assumed to be uncorrelated with Oit, a high residual value of Oit conditional on Nit may in fact imply a low presence of missionaries, and should therefore cause lower incarceration rates. This is because observing a high natural ability to evade the plantation system given its observed strength Nit is the same as a ‘surprisingly’ strong plantation system given the natural ability to evade it. We refer to Dippel, Gold, Heblich and Pinto(2017) for a detailed treatment and proofs. Whether Oit in fact has a signiﬁcant impact on Cit conditional on Nit is an empirical question of instrument power, not an econometric question of instrument validity. Online Appendix – Not for Publication

Table Online Appendix Table 15: TSLS Effect of Coercion (Cit) on Wages (wit), Conditional on Nit

Outcome w it : log wage

N it : Sugar N it : Plantation t + t2 t-fe t + t2 t-fe (1) (2) (3) (4)

C it : Incarceration (per Cap.) -0.4070 -0.5496 -0.4525 -0.6191 [0.007] [0.004] [0.014] [0.009] {0.0149} {0.0608} {0.0939} {0.1393}

N it (Share of Exports) -0.1789 -0.1219 -0.0877 -0.1209 [0.3062] [0.6759] [0.8500] [0.7582]

% effect of Nit on wit operating via Cit 0.772 0.856 0.823 0.940 F Statistic (Instruments) 17.335 7.463 6.547 3.269 Observations 750 750 750 750

Notes:(a) This table reports on the estimation of equation (22). In columns 1–3 we condition on Nit measured as sugar’s export share. In columns 4–6 we condition on Nit measured as plantation crops’ export share. 750 is the number of observations for which we observe both Cit and wit.(b) % effect of Nit on wit operating via Cit reported at the bottom ˆC ˆw|N ˆw of Panel A is the ratio (βN × βC )/βN .(c) The additional covariates in Table4 are also included but not reported. Standard errors are clustered by colony, p-values in square brackets. For the main regressor of interest, p-values for wild bootstrap standard errors are shown in curly brackets.

incarceration rates decreases wages by about 40–55 percent, and furthermore estimates that 75–85 percent of the effect of Nit (and therefore Oit) on wages operates though Cit. Online Appendix – Not for Publication

Table Online Appendix Table 16: Robustness Checks on Table5

(1) (2) (3) (4) (5) (6) (7) (8) t + t2 t-fe t + t2 t-fe t + t2 t-fe t + t2 t-fe

Panel A: w it , log wage

N it : Sugar Exports -0.7151 -0.5683 -0.7269 -0.5658 -0.7138 -0.5671 -0.7275 -0.5662 Share of Total Exports [0.0004] [0.0181] [0.0002] [0.0158] [0.0002] [0.0163] [0.0001] [0.0148]

IEA Salesit -0.0388 -0.0589 -0.0379 -0.0627 -0.0408 -0.0620 -0.0393 -0.0640 [0.0495] [0.0258] [0.0518] [0.0092] [0.0281] [0.0173] [0.0391] [0.0083] log Total Expenditure 0.0083 -0.0305 0.0195 -0.0200 [0.7766] [0.5433] [0.5052] [0.7010] log Total Revenue -0.0133 -0.0183 -0.0167 -0.0151 [0.4429] [0.2398] [0.3592] [0.3203] Observations 908 908 908 908 908 908 908 908 p-value: Test[ 2 coefficients equal] 0.001 0.031 0.000 0.028 0.001 0.028 0.000 0.026 F Statistic (Instruments) 18.878 15.111 28.058 24.472 25.056 20.703 28.129 26.115

Panel B: C it , Incarceration

N it : Sugar Exports 1.2239 0.9346 1.1463 0.8962 1.2304 0.9557 1.1495 0.9025 Share of Total Exports [0.0001] [0.0031] [0.0001] [0.0025] [0.0000] [0.0030] [0.0001] [0.0019]

IEA Salesit -0.0662 -0.0547 -0.0551 -0.0404 -0.0613 -0.0509 -0.0566 -0.0433 [0.2909] [0.4050] [0.4239] [0.5612] [0.3246] [0.4394] [0.4107] [0.5360] log Total Expenditure 0.2500 0.2343 0.2682 0.2651 [0.0014] [0.0202] [0.0003] [0.0035] log Total Revenue 0.0559 0.0335 -0.0241 -0.0413 [0.2380] [0.5763] [0.5651] [0.4559] Observations 798 798 798 798 798 798 798 798 p-value: Test[ 2 coefficients equal] 0.000 0.002 0.000 0.002 0.000 0.002 0.000 0.002 F Statistic (Instruments) 17.581 17.279 33.914 32.816 33.944 35.809 43.140 47.738

Panel C: Import Duties (£)

N it : Sugar Exports -1.0126 -0.6124 -1.7386 -1.3130 -1.4885 -1.1082 -1.8472 -1.4236 Share of Total Exports [0.3102] [0.5148] [0.1791] [0.3043] [0.2095] [0.3577] [0.1709] [0.2808]

IEA Salesit -0.3593 -0.4226 -0.3263 -0.3645 -0.2811 -0.3287 -0.2776 -0.3168 [0.0026] [0.0298] [0.0305] [0.0858] [0.0068] [0.0639] [0.0350] [0.1161] log Total Expenditure 0.9063 0.9131 0.5184 0.4231 [0.3910] [0.4292] [0.4597] [0.5507] log Total Revenue 0.6282 0.6968 0.5081 0.5991 [0.3067] [0.3054] [0.2615] [0.2619] Observations 942 942 942 942 942 942 942 942 p-value: Test[ 2 coefficients equal] 0.515 0.835 0.239 0.388 0.284 0.472 0.212 0.337 F Statistic (Instruments) 21.146 17.397 38.895 37.474 32.708 28.679 41.823 43.956

Notes: This table checks the robustness of the Second Stage in Table5 to including proxies for changes in the the general structure of public ﬁnance in our data. Online Appendix – Not for Publication

Table Online Appendix Table 17: The Effect of the IEA

N : Plantations' N : Sugar's Export it it IEA Sales Outcome Export Share Share it Time Controls t + t2 t-fe t + t2 t-fe t + t2 t-fe (1) (2) (3) (4) (5) (6) Panel A

O i x English Exports -0.2294 -0.2243 -0.4259 -0.4684 0.0571 0.1145

to all Non-Caribbeant [0.0095] [0.0033] [0.0001] [0.0001] [0.5714] [0.3641]

HDIit -0.0643 -0.0642 -0.0635 -0.0672 0.0086 0.0150 [0.0000] [0.0000] [0.0000] [0.0000] [0.3880] [0.2596] Incumbered Estate Act (IEA) -0.0187 -0.0168 -0.0059 -0.0088 1.0386 1.0538

Petitionsit [0.2593] [0.3811] [0.7967] [0.7396] [0.0000] [0.0000] Observations 942 942 942 942 942 942 R-squared 0.882 0.884 0.895 0.901 0.990 0.990 Panel B

O i x English Exports -0.2414 -0.2453 -0.4419 -0.4978 -0.0137 0.0264

to all Non-Caribbeant [0.0094] [0.0042] [0.0001] [0.0001] [0.7642] [0.5696]

HDIit -0.0591 -0.0590 -0.0594 -0.0634 0.0004 0.0048 [0.0000] [0.0000] [0.0000] [0.0000] [0.9085] [0.2031] Incumbered Estate Act (IEA) -0.0251 -0.0227 -0.0126 -0.0162 1.0479 1.0592

Petitionsit [0.3103] [0.3850] [0.6730] [0.5991] [0.0000] [0.0000] Observations 908 908 908 908 908 908 R-squared 0.889 0.891 0.897 0.904 0.995 0.995 Panel C

O i x English Exports -0.2342 -0.2430 -0.4421 -0.4942 -0.0242 0.0094

to all Non-Caribbeant [0.0176] [0.0119] [0.0002] [0.0002] [0.5413] [0.8245]

HDIit -0.0585 -0.0590 -0.0575 -0.0616 0.0005 0.0049 [0.0000] [0.0000] [0.0000] [0.0000] [0.8856] [0.2707] Incumbered Estate Act (IEA) -0.0180 -0.0149 -0.0052 -0.0089 1.0458 1.0570

Petitionsit [0.3211] [0.4716] [0.8167] [0.7204] [0.0000] [0.0000] Observations 798 798 798 798 798 798 R-squared 0.891 0.895 0.904 0.912 0.995 0.995

Notes: This table re-reports the First Stage in Table5 for the 3 different numbers of observations in that table’s Panel A. Online Appendix – Not for Publication

References

American and Foreign Anti-Slavery Society, The Annual Report of the American and Foreign Anti- Slavery Society Presented at New York ... with the Addresses and Resolutions, New York: American and Foreign Anti-Slavery Society, 1849.

Bekele, Frances L., “The History of Cocoa Production in Trinidad and Tobago,” in L.A. Wilson, ed., Re-vitalisation of the Trinidad & Tobago Cocoa Industry: Proceedings of the APASTT Seminar, Trinidad and Tobago: Association of Professional Agricultural Scientists of Trinidad and To- bago, 2004, pp. 4–12.

Berland, Alexander J., Sarah E. Metcalfe, and Georgina H. Endﬁeld, “Documentary-derived Chronologies of Rainfall Variability in Antigua, Lesser Antilles, 1770-1890,” Climate of the Past, 2013, 9 (3), 1331–1343.

Brizan, George I., Grenada, Island of Conﬂict: from Amerindians to People’s Revolution, 1498-1979, London: Zed Books, 1984.

Carvalho, Jean Paul and Christian Dippel, “Elite Identity and Political Accountability: A Tale of Ten Islands,” Technical Report, NBER Working Paper 22777 2016.

Craig-James, Susan, The Social and Economic History of Tobago (1838–1990): Persistent Poverty in the Absence of Adequate Development Strategy, Tobago, Scarborough: PRDI, 2000.

Deerr, Noel, The History of Sugar, London: Chapman and Hall, 1950.

Dookhan, Isaac, A History of the British Virgin Islands, 1672–1970, Epping, England: Caribbean Universities Press, 1975.

Eisner, Gisela, Jamaica, 1830-1930: A Study in Economic Growth, Manchester: Manchester Univer- sity Press, 1961.

Eltis, David, Frank D. Lewis, and David Richardson, “Slave Prices, the African Slave Trade, and Productivity in the Caribbean, 1674–1807,” The Economic History Review, 2005, 58 (4), 673–700.

Engerman, Stanley L., “Economic Change and Contract Labor in the British Caribbean: The End of Slavery and the Adjustment to Emancipation,” Explorations in Economic History, 1984, 21 (2), 133–150.

and Kenneth L. Sokoloff, “Factor Endowments, Institutions, and Differential Paths of Growth Among New World Economies: A View from Economic Historians of the United States,” in Stephen Harber, ed., How Latin America Fell Behind, Stanford: Stanford University Press, 1997, pp. 260–304.

Fergus, Howard A., Montserrat: History of a Caribbean Colony, Hong Kong: Macmillan Caribbean, 1994. Online Appendix – Not for Publication

Froude, James Anthony, The English in the West Indies: Or, The Bow of Ulysses, London: Longmans, Green, 1888.

Gibson, Campbell and Kay Jung, Historical census statistics on population totals by race, 1790 to 1990, and by Hispanic origin, 1970 to 1990, for large cities and other urban places in the United States, U.S. Census Bureau, 2005.

Great Britain, The Colonial Blue Book for the Leewards Islands, London: H.M.’s Stationery Ofﬁce, 1911.

Green, William A., British Slave Emancipation: The Sugar Colonies and the Great Experiment 1830- 1865, Oxford: Clarendon Press, 1976.

Hall, Douglas, Five of the Leewards: 1834–1870, Epping, England: Caribbean Universities Press, 1971.

Handler, Jerome S., “The History of Arrowroot Production in Barbados and the Chalky Mount Arrowroot Growers’ Association, a Peasant Marketing Experiment that Failed,” Journal of the Barbados Museum and Historical Society, 1965, 31 (3), 1–22.

, “The History of Arrowroot and the Origin of Peasantries in the British West Indies,” Journal of Caribbean History, May 1971, 2, 46–93.

Harmsen, Jolien, Guy Ellis, and Robert J. Devaux, A History of St Lucia, Vieux Fort, St Lucia: Lighthouse Road Publications, 2012.

Harrigan, Norwell and Pearl Varlack, The Virgin Islands Story, Epping, England: Caribbean Uni- versities Press, 1975.

Heckman, James J. and Rodrigo Pinto, “Causal Analysis after Haavelmo,” Econometric Theory, 2015, 31 (1), 115–151.

Higman, Barry W., “The Sugar Revolution,” Economic History Review, 2000, 53 (2), 213–236.

Imai, Kosuke, Luke Keele, and Te Yamamoto, “Identiﬁcation, Inference and Sensitivity Analysis for Causal Mediation Effects,” Statistical Science, 2010, 25 (1), 51–71. Online Appendix – Not for Publication

Jacks, David S., “Intra- and International Commodity Market Integration in the Atlantic Economy, 1800–1913,” Explorations in Economic History, 3 2005, 42, 381–413.

Jayasinghe, P.K.S.C. and Masao Yoshida, “Development of Two GIS-based Modeling Frame- works to Identify Suitable Lands for Sugarcane Cultivation,” Tropical Agriculture and Develop- ment, 2010, 54 (2), 51–61.

Knapp, Arthur W., Cocoa and Chocolate: Their History from Plantation to Consumer, London: Chap- man and Hall, 1920.

Lewis, Gordon K., The Growth of the Modern West Indies, New York: Monthly Review Press, 1986.

Lewis, W Arthur, “Economic Development with Unlimited Supplies of Labour,” The Manchester School, 1954, 22 (2), 139–191.

Martin, Robert M., Statistics of the Colonies of the British Empire, London: WH Allen and Company, 1839.

Merivale, Herman, Lectures on Colonization and Colonies, London: Longman, Green and Roberts, 1861.

Nugent, Jeffrey B and James A Robinson, “Are Factor Endowments Fate?,” Journal of Iberian and Latin American Economic History, 2010, 28, 45–82.

Pinto, R., C. Dippel, R. Gold, and S. Heblich, “Mediation Analysis in IV Settings With a Single Instrument,” 2019.

Richardson, Bonham C., Economy and Environment in the Caribbean: Barbados and the Windwards in the Late 1800s, Mona, Jamaica: University of the West Indies Press, 1997.

Sewell, William G., The Ordeal of Free Labor in the British West Indies, Harper & brothers, 1861.

Soluri, John, “Bananas Before Plantations. Smallholders, Shippers, and Colonial Policy in Jamaica, 1870–1910,” Iberoamericana, 2006, 6 (23), 143–159.

Tannehill, Ivan Ray, Hurricanes: Their Nature and History: Particularly those of the West Indies and the Southern Coasts of the United States, Princeton: Princeton University Press, 1938. Trouillot, Michel-Rolph, Peasants and Capital: Dominica in the World Economy, Baltimore: Johns Hopkins University Press, 1988.

United States National Hurricane Center, “NHC Data Archive,” October 2014. Retrieved October 2013.