The Organization of Eastern Merchant Empires Claudia Rei* January 15, 2011

Department of Economics Vanderbilt University VU Station B#351833 2301 Vanderbilt Place Nashville, TN 37235

[email protected] Phone: (615) 322-2482 Fax: (615) 343-8495

*

I thank Maristella Botticini, Boyan Jovanovic, Bob Margo, and Andy Newman for many helpful discussions and guidance, and Jeremy Atack, Bill Collins, and Andrea Moro for valuable comments on earlier drafts. I also thank Sidarth Chandra, Leonor Costa, Guillaume Daudin, Mauricio Drelichman, Rui Esteves, Eric Hilt, Carol Shiue, Peter Temin, Joachim Voth, Jeff Williamson, the participants at the annual meetings of the Economic History Association, the EGE meeting at the NBER, and the SITE meetings in Stanford. The financial support of the Economic History Association (dissertation award) is gratefully acknowledged.

Abstract In the sixteenth century, European countries engaged in long-distance trade with the East. Despite sharing the same objectives and technology, Portugal opted for a crown monopoly, England, the Netherlands, and Sweden franchised trade to private merchants, whereas in Denmark and France, king and merchants shared control. The …nancial condition of the crown appears to have been relevant for the monarchs’decision. I provide an economic mechanism to illuminate the historical variation in terms of the di¤erences in relative endowments of king and merchants within each country. I also explore the implications of control allocation using archival data on labor compensation and shipping technology. Di¤erences in the long run performance of merchant empires suggest a major impact of organization.

KEYWORDS: Organization, Merchant Empires, Financial Constraints

1

1

Introduction

The choice of how to organize overseas trade arose when Europeans began to venture across the seas in the …fteenth century. After the discovery of the Eastern passage around the Cape of Good Hope in 1487, several European countries engaged directly in the business, halting a long lasting reliance on the intermediaries of the Silk Route. The expansion of international trade beyond European borders did not follow a uniform pattern. In Portugal the king had direct control. In England, the Netherlands, and Sweden, kings chartered monopoly rights of trade to merchants, while in Denmark and France the king and merchants shared control.1 The di¤erent organizational choices by the monarchs are puzzling. All empires traded in exotic products highly demanded in Europe, and all had access to the same transportation technology –the sailing ship –to bring those products from distant locations. Moreover, all kings had similar objectives with regards to their overseas possessions. On the one hand, a clear priority in mercantilism was to access the riches outside of Europe: direct control over newly discovered sources of precious metals was certainly desirable; and the control of trade in highly valued eastern products (e.g. spices) would also attract gold and silver to the realm. Kings also extracted glory from their overseas empires: the international prestige and visibility which would be recognized not only by the peers of their time, but also immortalized in history. This status could be achieved through visible demonstrations of power, for instance, the military dominance in foreign and distant territories, the expansion of the crown’s rule over a larger number of subjects, or the gigantic ships that would dazzle the observer and raise fear in the enemy. From the pharaohs of Ancient Egypt to the Soviet 1

The Spanish empire is not considered in this study due to geographical reasons: it was the only empire established originally in the western hemisphere, which was virtually unknown in the late …fteenth century. It does not seem plausible to analyze the Spanish case with a framework in which kings and merchants negotiate over known outcomes as it was the case in the East. Spices were brought to Europe since the Silk Route emerged in Hellenistic times. Arriving in scarce quantities due to the large number of intermediaries between Asia and Europe, spices were very expensive products only accessible to the elite. Aware of the source of spices in Asia, Europeans had a clear notion of the potential gains of the spice trade, which a¤ected the negotiation for the type of organization when merchant empires were established in the East. Spain set foot on virgin territory with unknown products and grim ex-ante prospects of trade, which makes the emergence of its merchant empire unique and not comparable to that of those established in the East.

2

Republics, history is laden with examples of rulers to whom the extension of power beyond national borders (or the expansion of borders themselves) was greatly cherished. The historical literature is vast in the characterization of merchant empires, their sequence, and their economic performance, but silent on the reasons behind the emergence of the di¤erent organizational forms.2 The initial control of a large number of eastern ports by the Portuguese in the …rst half of the sixteenth century was supplanted by that of the Dutch and the English in the seventeenth. Figure 1 shows a relative measure of the success of empire with the percent tonnage of European shipping to Asia in the sixteenth and seventeen centuries. The initial decline of the Portuguese market share after a long period of standing alone in the market is an expected result of competition. However, it did not take long for Portugal’s share to drop below that of the Dutch and English private companies and by the end of the period, it was at the level of the smaller Danish and French companies.3 The …rms’distinct abilities to maintain territorial control and, thus, guarantee sustained trade, suggest that the choice of organization had an impact on the long term economic performance of the empires. At the establishment of merchant empires, the crowns faced many di¤erent circumstances which may have been in‡uential in their organizational choice. Despite the variety of circumstances, also discussed later, the …scal position of the crown relative to the merchants was potentially critical. In this paper I o¤er an economic rationale in which the distribution of bargaining power between the king and the merchants in each country –stemming directly from their original endowments –has direct implications on the control structure of the empire: if the king were ‡ush with cash (relative to the scale of investment) he would choose to maintain control; if not, the king would franchise the enterprise, thereby delegating control to private merchants.4 The historical evidence is a few centuries old and not very systematic, but it suggests 2

Braudel (1972), Chaudhuri (1978), Furber (1976), Godinho (1969), Hart (2003), Steensgaard (1974). The Portuguese decline is also visible in absolute terms. 4 See Blair and Lafontaine (2005) for a survey on the literature of franchising. 3

3

that divergent …scal constraints on di¤erent monarchs were relevant to the type of empire to emerge. In Portugal, the estimated high rates of return accrued by the king in the early expansion (1415-1498), seem to corroborate the crown’s decision to own the monopoly of trade after Da Gama’s voyage in 1498. In contrast by 1600, when the East India Company (EIC) was founded, the English Royal Treasury was empty: the funds the crown was able to raise not only had more pressing employment (war), but were also insu¢ cient to bear the cost of the voyages. Unable to fund the empire, it comes as no surprise that the English monarch chose instead to charter a monopoly of trade to the EIC, thereby forfeiting full pro…ts in lieu of a tax on the proceeds.5 This simple mechanism suggests that the di¤erences in organizational control have implications on the …rms’business decisions. Ever since the seminal work of Oliver Williamson in 1975, we know that organizational control manifests in the inner workings of …rms as well as their economic performance. In this paper I further investigate the implications of organizational control with the analysis of data on labor compensation and shipping technology for two cases of extreme control: Portugal and the Netherlands. Indeed the two empires seem to have behaved very di¤erently in these areas: Portuguese workers were paid a larger part of their compensation in the form of bonuses than their Dutch counterparts; and Portugal was slower to adopt new and more e¢ cient shipping technology. The paper illuminates the literature of merchant empires with an economics perspective, contributing to the general research about the impact of institutions on economic performance (Davis and North 1970, North 1991, Greif 1993, Acemoglu et al 2005). However, the 5

A dynamic model may have been appropriate to explain the irreversibility of organizational choice across empires and their consequent divergent long run performance. However, given that neither the Portuguese nor the English, Dutch or Swedish empires ever switched the type of organization, the simple static model su¢ ces to analyze the origins of the initial choice. Such organizational persistence was not free from setbacks. In Portugal the king attempted to sell the rights of Asian trade to private merchants in three di¤erent occasions (1587, 1628, and 1697), but none succeeded. In England and the Netherlands, the original private merchant companies of 1600 and 1602 respectively, were very di¤erent from those formally dissolved in 1873 and 1799 in which the state had a more prevalent role. The residual claimant remained the same in each country, which is essential for the purpose of this paper, but this path dependent pattern still needs to be explained. Further research is necessary to investigate the factors that drove the institutional lock-in (as in David 1985:334) of merchant empires, in which the individual histories of all companies (royal and merchant) may have a decisive role.

4

paper goes beyond the debate of which institutions may be growth enhancing or hindering, by o¤ering a model that explains the emergence of such institutions. The model draws on the theory of the …rm, in particular, the e¤ect of property rights on …rm ownership (Grossman and Hart 1986, and Hart and Moore 1990). The main thrust is on how the division of surplus, rather than e¢ ciency, a¤ects the result of the negotiation: divergent bargaining positions may give rise to less e¢ cient outcomes (Legros and Newman 2008). Institutions do not emerge because they are e¢ cient, but because they align with the incentives of the players who bring them about. The negotiation between the players contributing to a particular institution, as well as their bargaining power, are crucial for the type of institution to emerge, which is well known to have implications on future economic performance. Whether players’priorities match market e¢ ciency is unclear, as can be illustrated in the history of the organization of merchant empires.

2

A Model of Organizational Choice

The multiple scenarios of the emergence of merchant empires occurring in di¤erent time periods make history hard to understand. Here I provide a simple economic model that helps illuminate the di¤erent choice of organization by the monarchs. Long-distance trade results from the cooperation between merchants and king. The merchants contribute with management skills, a non-contractible and non-transferable e¤ort type represented by e

0.6 The king owns the right to award the monopoly of trade either

to the merchants or to himself.7 Beyond e¤ort and charter, the business also requires "cash" c owned by the merchants cm or by the king ck (c = cm + ck and cm ; ck > 0), which can be invested in long-distance trade (e.g. to buy ships and pay sailors) or other alternative 6

Merchants are a group of agents with the same preferences negotiating with the king as a single entity, which was historically observed. The considerable investment to run such an enterprise made merchants gather their resources together in the form of joint stock, and negociate directly with the monarch for monopoly rights of overseas trade, then a royal prerogative (Cawston and Keane 1896). 7 I also considered the case in which the king exerts e¤ort (e.g. protection from pirates), which does not change the results and is ignored for simplicity.

5

purposes (e.g. to build cathedrals, domestic trade).8 Historically, we observe kings and merchants engaging in long-distance trade, indicating it was an attractive investment. For the purpose of the model, I assume competing outside options for merchants cm , and king ck to have lower returns

< , with

representing the return on long-distance trade.

Long-distance trade also depends on operational decisions d 2 [0; d] (e.g. size of ships), ordered according to the king’s preferences. Although the choice of high d may be favorable to production (larger ships carry more spices), it a¤ects the cost of merchants’e¤ort (larger ships are harder to maneuver and more likely to perish in storms, so merchants prefer smaller and safer ships). Business decisions are far too complex to specify in the contract (let alone enforce), so d is chosen after the realization of the "state of the world"

2 [0; 1], with mean

.9 The right to make such decisions is however contractible. Output from long-distance trade results in

y = c, with probability (e; d) = D e1

and 0 otherwise. The success probability D=1

1 (d 2

,0<

<1

(1)

depends on merchants’ e¤ort e, and a term

)2 re‡ecting the sensitivity of

to d and : the farther d is from the realized

state of the world , the smaller the D and the less likely the venture is to succeed. The utility functions for merchants and king are:

u = sE[y]

e; and v = (1

s)E[y] + d:

(2)

Each party cares for its share of expected output and also for organizational control, which confers the right to choose d. Like the merchants’, the king’s utility is a¤ected by d through 8

An alternative use of cash could be to make transfers between parties in order to achieve the outcome of the bargaining process. The party in control would prefer not to do so because its utility would fall; the party not in control could not do so because of cash constraints (Figure 3 in the Appendix). I therefore assume that cash is non-transferrable and that parties cooperating in long-distance trade invest all their available cash. 9 The state of the world consists of any conditions that may a¤ect d, a few examples: the existence of armed con‡ict, the quality of the crops in a given year, the weather throughout the voyage, local prices, etc.

6

expected output, but in addition, the king cares for royal glory: even with a zero share of output, having control generates positive utility for the king ( > 0, bigger ships as a means to show o¤ military power) who rules over more people, controls a larger territory, etc.10 The timing of the game is as follows: (i) design a contract consisting of two variables over which parties bargain: a sharing rule s 2 [0; 1], and a control structure (king control –KC, or merchant control –MC);11 (ii) realization of the state of the world , and subsequent choice of d by the party in control; (iii) determination of optimal e¤ort level, e, by the merchant. Solving the model by backward induction, I get optimal values for e¤ort e and business decisions d in both control structures.12 The …rst result is that for any sharing rule s merchants choose to put more e¤ort in a …rm they control (eM C > eKC ), which, all else the same, makes the probability of success of the venture smaller when the king is in charge (

KC

<

M C ).

Second, dKC > dM C for any s, which reinforces

KC

<

MC

and is consistent

with the king extracting direct utility from taking business decisions: he will always overshoot relative to the optimal level of d, for any realized state of the world. The two results go in the same direction: choices of e¤ort and business decisions in a king controlled …rm combine into a lower probability of success, relative to the merchant controlled scenario. Given the e¤ort and decision pro…les determined in stages (iii) and (ii), in stage (i) parties bargain over the sharing rule s, and the control structure. Substituting e and d in the utility functions, generates the optimal utility pair for merchants and king (u ; v ) which depicts concave utility possibility frontiers in Figure 2 upon variation of s. There is a continuum of disagreement points [( c; 0); (0; c)] representing both parties’ outside options. If the king owns a lot of cash, in the limit c = ck as it seems to have been the case in Portugal, bargaining takes place to the northeast of the disagreement point 10

I considered a similar term in the merchants’utility function, but the results are left unchanged, so long as the merchants care less for glory than the king. 11 If the sharing rule were non-contractible, for example if the king were unable to commit not to expropriate/con…scate the proceeds of a private company, renegotiation would be allowed at the end of the period. Instead of two frontiers corresponding to di¤erent control structures, we would get two points only, but the results would follow. 12 See appendix for detailed expressions of e , d , u and v under KC or MC.

7

(0; c) and the solution will likely lie on frontier KC (say point A). If on the other hand, the monarch has no cash to invest c = cm , as in England, bargaining occurs to the northeast of ( c; 0) and the outcome will be in the MC frontier. A non-extreme …nancial structure would place the disagreement point in some intermediate location between ( c; 0) and (0; c) and the result of the negotiation could alternate between KC and MC, which is consistent with the shared control cases of Denmark or France. Though resulting from the bargaining process, outcomes on KC are not socially optimal if we consider the social surplus to be the unweighted sum of the utilities of king and merchants.13 The choice of control structure does not derive from social e¢ ciency, but rather from the distribution of bargaining power across parties. The implications of the model are very clear: 1) a cash-‡ush king will control the monopoly of trade (KC case): he does not need merchant money to conduct trade, so he hires merchants to run the royal business giving them a small share of the spoils of trade; 2) a cash-constrained king will give up control to the merchants who gather in the form of joint stock, forcing the king to concede them monopoly rights over trade in exchange for a small share (MC case); 3) organizational control has stark implications on …rms and should be visible in terms of merchants’e¤ort and in the business decisions the …rms adopt.

3

The Emergence of Merchant Empires in the East

In light of the above, the key question for organizational choice is how well endowed monarchs were at the establishment of long-distance trade. I analyze the di¤erent cases according to the control ownership of the empires. 13

The highest indi¤erence curve of slope

1 touches MC, that is, the socially optimal control structure.

8

3.1

King Control

Table 1 shows the progress of the Portuguese expansion in the …fteenth century, which consisted of military expeditions in Northern Africa as well as exploration voyages down the African coast and out to the Atlantic islands. The narrative evidence seems to indicate that this crown-sponsored venture provided signi…cant revenue by the end of the …rst half of the …fteenth century: gold was …rst found in West Africa in 1436, the …rst slaves arrived in Lisbon in 1441, and trade in Guinean pepper started by 1450 (Serrão 1980:194-9). The king’s political and …nancial status can be inferred from the records of cortes, the general assemblies between the king and representatives of the nobility, the clergy, and the general estates. These meetings had formal legislative, political, as well as …scal purposes, but were also the place where the king heard the country’s complaints, and where the …nancing of royal policies was discussed (De Sousa 1990).14 Table 2 shows the average interval between meetings before and during the expansion. If the cortes were a constraint on the king, or if the king were …nancially dependent on taxes raised at this level, there should be no signi…cant change in the average time between meetings. Until 1415, before the start of the maritime expansion, the cortes convened on average every one and a half years, whereas in the period of African expansion (1415-1498), the meetings occurred every three years. After reaching India the interval escalated to nine years. The ever-increasing time between meetings suggests a revenue independent monarch, who could undertake important decisions for the realm without the advice, consent, or funding of his subjects. As suggestive as this evidence may be, knowing that Portuguese monarchs were rich only addresses half the story of the organizational choice in merchant empires. The other half relates to the expenditures faced by those same monarchs in the same time period, especially 14

As De Sousa shows, cortes never convened regularly despite the general estates repeated request: it was their single chance to voice complaints to the monarch, who depended on their taxes. In 1371, the general estates proposed a three-year interval for the assemblies, to which the monarch replied that parliament would be called "whenever he understood necessary to his service and to the good of the land" (De Sousa 1990:111).

9

inevitable ones, typically wars. Portugal had its fair share on this front: in 1383-85 the Portuguese dynastic crisis issued a civil war, in which Castile unsuccessfully intervened; in 1475-79 Portugal intervened in the war of the Castilian succession with no e¤ect; and from 1580 to 1583 the war of the Portuguese succession of which Spain came out victorious. In addition to wars, Portugal had frequent disputes with Spain on account of newly discovered territories throughout the …fteenth century. The Treaty of Alcáçovas in 1479 concluded the war of Castilian succession ending Portuguese claims to the throne while granting Portugal hegemony over all Atlantic islands excepting the Canaries. To the south of this archipelago Portugal was allowed to continue exploration of unknown land. The disputes were revived after Columbus’1492 voyage: the newly discovered West Indies lied to the south of the Canary Islands, and Portugal claimed the territory. In 1494 the Treaty of Tordesilles formally divided the world along a meridian 370 leagues west of the Cape Verde islands: to the east all territories were Portuguese, and to the west Spanish. This however, was not the end of con‡ict between Portugal and Spain over new lands. The Portuguese discovery of the valuable Moluccas (the Spice Islands) in 1512 led Spain to claim the archipelago since it was on the western hemisphere of the Treaty of Tordesilles. The dispute was …nally settled in 1529 by the Treaty Zaragossa, which enabled Portuguese control of the Spice Islands upon a payment of 350,000 gold ducats to Spain. The continuous wars and territorial disputes with neighboring Spain corresponded to a level of expenditure that cannot be assessed. Nevertheless, Portuguese kings seem to have been …nancially able to support the inevitable expenditures of war and empire, at least until 1529. As the purchase of the Moluccas suggests, money was not a concern for the king of Portugal. Such fortunate …nancial status might have been achieved because early stages of discovery were very pro…table. I now turn to the estimation of (labor and capital) costs of the discoveries listed in Table 1 and compare those to the historical reports of revenues from the same voyages. The purpose of this exercise is to evaluate the king’s …nancial ability in 1498

10

when faced with the decision of owning or chartering out monopoly rights of eastern trade. Costs and Revenues Wage series for …fteenth century Portugal are not available.15 Therefore, I rely on Robert Allen’s (2001) database of wages for laborers and craftsmen of other European cities in the same time period.16 Together with the scattered information on ships and men in maritime expeditions (henceforth events), labor costs per event i at time t are calculated as follows:

wLi;t = daily waget

number of meni

number of daysi

The …rst term is a weighted average of daily wages in grams of silver, for skilled (20%) and unskilled (80%) workers in all available cities for the year of the event.17;18 The labor requirement varied with the nature of the event. Military expeditions demanded more men and ships than exploration voyages, and the latter’s labor requirement varied with distance. The third term is used when known, or estimated by the baseline duration of the well documented events. The estimations are based on men and ships sent, not returned, so that costs are not underestimated.19 Capital costs for each event (rKi;t ) are calculated recursively using a given labor share and applying it to wLi;t . With a Cobb-Douglas20 production function y = AL K 1 15

, and

The 1755 Lisbon earthquake destroyed two-thirds of the city and burried the national archives, making it nearly impossible to …nd data for earlier periods. 16 Cities (Antwerp, Florence, Krakow, London, Oxford, Paris, Strasbourg, Valencia) were weighted by own population (DeVries 1984) in the year of the event. Labor markets need not be integrated, but shared shocks (demographic or technological) possibly yielded similar wage trends in early modern Europe (Allen 2001). 17 The skilled share was practice in the Portuguese ships sailing to India in the late sixteenth century: in a crew of 123 men there were 45 sailors, 48 apprentice sailors, 4 pages, and 11 bombers (Falcão 1859:198). The rest of the crew had occupations associated with a skill, such as mason, or pilot. If bombers are not considered skilled labor the percentage of unskilled workers is 79%, if they are, the …gure changes to 88%. 18 The conversion of silver into gold is necessary to compare estimated costs with revenues from the gold trade. There are three gold-silver ratios: .1080 for 1410 (Wee 1963:127); .0895 for 1457-64 (Munro 1994:6745); and .0897 for 1489 (Munro 1994:674-5). I calculate yearly values by linear interpolation. 19 Refer to section 7.1 in the Appendix for details on assumptions and calculations. 20 The original reduced form problem M ax(2); s:t:(1) is the same as the one in which the production rk wL. Along the utility possibilities frontier, it can be shown function is E[y] = AL K 1 c D e1 that the cost minimizing problem must be solved. The two functional forms are thus equivalent.

11

each factor of production paid at its marginal product w = A( K )1 L

and r = (1

L )A( K ) ,

we can calculate labor share as follows: y wL = , wL = (rK + wL) , rK (1 )y

=

1 . rK 1 + wL

(3)

A constant labor share over time, yields capital costs:

rKi;t =

(1

)

wLi;t .

(4)

The labor share , can be calculated from (3) for any event in maritime history with data on labor and capital costs. I consider two episodes: a trip to India on a Portuguese vessel in 1615, the calculation of which yields industry in 1859, which yields 1859

1615 1859

= :8188; and a trip of a vessel in the American whaling

= :4559.21 Any labor share above

1615

= :8188 or below

= :4559 seems implausible given current estimations for developed and underdeveloped

countries of .65 and .80 respectively (Senhadji (2000) and Gollin (2002)).22 The estimated 1615

and

1859

generate the upper and lower bounds for capital costs on Table 3.

On the revenue side I only observe revenue corresponding to the gold trade in Guinea. As for the two other sources of trade controlled by the crown –slaves and Guinean pepper – unfortunately, prices are only available for the sixteenth century which renders them useless for this exercise. Nevertheless, historians have ranked the …fteenth century sources of royal revenue as slaves, gold, and spices by order of magnitude, which was reversed in the sixteenth century with the India trade (Marques 1982:64). The reported revenues on Table 3 are, thus underestimated by at least one-half.23 In the last column, we see revenues always exceeding total costs by a magnitude of at least three, which indicates an extremely pro…table enterprise 21

Refer to section 7.2 in the Appendix, where I also calculate the labor share for Portuguese shipping to Brazil in 1633, as well as throughout the seventeenth century, and for a voyage of an English merchant ship in 1693 (Tables 10 to 14). All are within the bounds. 22 Underdeveloped countries today are likely to have labor shares that resemble pre-industrial values given the relatively intense use of labor due to the low degree of mechanization in these countries. 23 Moreover, slaves had been arriving in Lisbon since 1441: not only the size of the downward bias is large, but also the king was getting revenue much earlier than reported.

12

for the Portuguese monarch. The return was even higher in the successful military ventures of Northern Africa: though more demanding in ships and men, the revenue in looting was considerably higher than that attained in regular trade. Sensitivity Analysis Even though revenues are underestimated resulting in revenue-cost ratios that are likely to be a lower bound on the crown’s returns, the unusually high magnitudes beg a few robustness checks. Since revenues are …xed, one could attribute the high ratios to the underestimation of costs either via low wages, or the short duration of the events;24 the inappropriate calculation of the labor share could also have depressed costs. First, I compute labor costs using the highest wage in Allen’s database (Valencia for all years). Doing so reduces the ratio of revenues to costs, but it still is between 2:1 and 10:6. I also computed labor costs using only the wages of the highest paid craftsmen (again Valencia’s), that is, assuming all labor on board is skilled, and the returns now vary from 1:5 to 9:3, still quite high. Low wages do not seem to be the driving force of high returns. Second, the duration of 1492 Columbus’voyage relative to the distance traveled was much shorter than that of the Portuguese expedition to Ceuta.25 If instead, Columbus’voyage had been used as baseline, the estimated costs would be much lower. The low duration of the events appears not to result in high rates of return either. Third, a high value of

generates low capital costs, which in turn generate high rates of

return. Omitting the value of cargo, a part of capital costs, contributes to this e¤ect. I am interested, however, in measuring the cost of …fteenth century exploration voyages in which ships were not loaded with bullion.26 Moreover, maintaining a constant labor share (at the 24

Lost ships would have the same e¤ect if the calculations included destruction rate, which is not the case: I calculate the costs of ships sent, not returned. 25 In 71 days, Columbus travelled 4,323 miles (7 days from Palos in the South of Spain to the Canary Islands; 28 days stopped for repairs at the Canary Islands; and 36 days from the Canary Islands to the Bahamas), an average of 60.9 miles a day (Columbus, 1989). The 1415 voyage to Ceuta lasted 28 days for a distance of 479 miles, that is, 17.1 miles a day. Columbus could have arrived Ceuta in just 8 days. 26 Similar calculations taking into account cargo have been done for France in the eighteenth century and the labor share falls to about .4 (Daudin 2005, p. 306). Using this lower labor share and adding 4 trips a

13

1615 or the 1859 level) does not reduce the estimates of rKi;t , on the contrary, capital costs during the period of interest are overestimated. If dropped (so

1

changed over time, it could only have

in (4) would be lower in 1415 than 1615) for two reasons: (a) the …fteenth

and sixteenth centuries witnessed dramatic technological improvements in shipping–both the transition from oars to sails in the early …fteenth century (lower L) and the later increase in vessel size (higher K) increased the ratio K=L, which would bring

down if labor were

cheaper than capital ( wr > 1); and (b) even though nominal wages went up in almost all cities, real wages fell everywhere except in Antwerp (Allen 2001). The decline in purchasing power is consistent with the specie in‡ow in Europe since the late …fteenth century, which caused generalized in‡ation. Consequently, r is likely to have risen more than w (stickier wages) over the …fteenth century, which con…rms the condition Fourth,

r w

> 1.

could have been miscalculated due to the inaccurate estimation of a ship’s

useful life. As justi…ed in the appendix,

1615

takes into account a ship’s life of four round

trips (Phillips 1994:102, Steensgaard 1965). If instead we assume the implausible duration of a single round trip, by

1859 .

This last

1615

falls from :8188 to :5305, a value still above the lower bound given

was calculated with capital and labor costs per trip, without further

assumptions on the ship’s life. In sum, the duration of ships does not seem to be responsible for the high rates of return to the Portuguese king. Both the indirect evidence and the computed returns on investment indicate that the early expansion process was pro…table. With such high return rates, it is not surprising that the Portuguese crown chose to maintain the monopoly of trade after arrival in India: the king did not need merchant money to engage in long-distance trade because he could a¤ord it himself. This result was helped by the fact that initial expeditions were fairly cheap. The model’s prediction that a non-cash constrained monarch would opt to take charge of the enterprise seems to illuminate Portugal’s choice of organization. year for the gold trade with the Guinea coast (Vogt 1979) the return rates are reduced, ranging between 1.6 and 2.5, still a high return on the investment.

14

3.2

Merchant Control

Apart from the Portuguese monarch, all other kings establishing merchant empires in the East franchised out trade rights to private merchants. Within this set of countries however, the control pattern was not uniform. Here I focus on the pure merchant control cases. The English Case The foundation of the EIC came at the end of the long and politically agitated reign of Elizabeth I (1558-1603). Internal tensions, such as the northern rebellion of Catholic Earls in 1569, the Nine-Years war (1594-1603) with Ireland, and the dispute of the throne by Mary Queen of Scots, were coupled with the ensuing Anglo-Spanish war (1585-1604). All these events added to the …nancial distress of the crown. Elizabeth’s predecessors had been very successful in increasing government revenue: Henry VII raised taxes, requesting …nancial help from parliament only once during his reign (1485-1509); his successor Henry VIII seized the lands of monasteries after the split from the Roman Church in 1534 (Pollard 1910). But wars with France and Scotland reversed the …nancial advantage and at the death of Henry VIII in 1547, the political divisions of the kingdom resulting from the Wars of the Roses emerged again. The following eleven years were marked by economic problems resulting from political instability, which included wars, rebellions, throne disputes and short reigns. By the time queen Elizabeth ascended to the throne in 1558 she had a lot of people to please and very little money. The incipient nature of national credit in the sixteenth century led the English monarch in 1544 to turn to the Antwerp market, which had abundant resources and was capable of o¤ering the large loans the crown was looking for. Antwerp was however, a credit market originally developed to satisfy the short term credit requirements of merchants, quite di¤erent from the …nancing of long term wars. The increasing di¢ culty of the royal agent to raise loans led to a virtual closure of Antwerp to the English in 1569. Even though all foreign debt was repaid by 1574, new foreign loans were sought in 1575-6 and 1589, but neither was

15

fully subscribed (Outhwaite 1971). Royal …nances remained constrained even after the Antwerp episode.27 From a surplus in 1583-5, the government balance had turned into a de…cit by 1597-1600, demonstrating the greater …nancial needs resulting from increasing wars (Goldsmith 1987:190, 194). But the de…cit around 1600, would not prevent royal engagement in long-distance trade, provided there was scope for government borrowing and debt issue. Such was not the case in the late sixteenth and early seventeenth centuries when the English …nancial system was characterized by the absence of specialized lenders, which had been the main reason behind the earlier switch to the Antwerp market.28 Government borrowing turned to the internal market in the form of forced loans, which were involuntary, usually secured under threat, and had highly unpredictable repayment, never on the initial terms (North and Weingast 1989). Such method of raising capital, was both unwelcome with merchants and unsustainable in the long run. Table 4 shows the history of forced loans in this period. Granted the queen had many competing uses for such funds, but we still need to check if these amounts were su¢ cient to assume control of the East India trade should the crown have decided to do so. Table 5 presents the expenses of the EIC during its …rst …fteen years during which the original merchants were granted exclusive trade to the east of the Cape of Good Hope.29 The last column shows the discounted value of the cost of ships and victuals in 1601 pounds sterling. The cost of the …rst voyage is above the value the monarch could obtain in forced loans and also above the value the merchants could gather in the …rst subscription of the EIC’s stock –£ 30,000 (East India Company 1886:1-4). The limited size of the initial subscription and the slowness with which capital was raised, 27

The …nancial situation of the crown was also negatively a¤ected by falling revenues due to the parliament’s cancelation of patented monopolies, such as the manufacture and importing of playing cards granted by the Queen to Darcy, her groom in 1601 (Letwin 1965:27). 28 Only after the 1688 Glorious Revolution were there institutional arrangements protecting lenders’property rights, which allowed for an expansion of the national credit market. Long-term government borrowing rose to unprecedented levels (Brewer 1989, Dickson 1967, North and Weingast 1989). 29 For the complete text of the original charter, see East India Company (1893).

16

delayed the …rst voyage for a year (Chaudhuri 1965). Nevertheless, the merchants ability to raise capital was higher than that of the king in the early years of the company. The total discounted (at a 10% annual rate) value sums to £ 227,062, which is higher than any amount obtained by the monarch in the form of forced loans. The minimum cost of the production process was high relative to the monarch’s ability to pay it herself. With public …nance under stress and no borrowing possibilities, the English crown chose to franchise out monopoly rights to the EIC in 1600. The Dutch Case Since 1595, Dutch merchants were organized in small companies, which sent ships to the East in order to break the nearly one-hundred year old Portuguese monopoly of Indian spices in Europe. Even though the …rst voyage was not a commercial success, it showed that Dutch shipping to Asia via the Portuguese controlled Cape-route was feasible –a big achievement, compared to the failed attempts of other merchants in the search for the passage to Asia via the North Arctic Ocean (Jacobs 1991). Between 1595 and 1601, eight di¤erent companies dispatched ‡eets to the East, some being extremely successful, while others complete failures. The success soon turned against Dutch entrepreneurs. On the supply side, competition among Dutch ships was e¤ectively rising Asian spice prices whereas in Europe, prices were falling due to too much quantity available. The small companies came to a merger agreement encouraged by the government of the Dutch republic –the States-General. Founded in 1602 the VOC (Verenigde Oost-Indische Compagnie) o¤ered a slightly di¤erent organization form from the EIC’s. There was no body of company shareholders distinct from the chambers’shareholders, which would give no power to smaller shareholders. The main contributions of capital came from Amsterdam’s chamber with 57.2%, which allowed it to dominate the decisions of the company (Furber 1976:188). The VOC had a central management unit of seventeen delegates, coming from all six chambers. These delegates e¤ectively governed the company. According to the 1602 charter,

17

if the assembly (of the chambers) did not come to an agreement on a question, this was to be decided by the States General (Glamann 1981). The 1602 charter was conceded for a period of twenty-one years. A renewal was granted in 1623 for another twenty-one years, but afterwards the renewal periods were extended: in 1647 the charter was renewed for twenty-…ve years and in 1696 for forty years more. With the exception of 1623, each renewal corresponded to a payment to the Government, which was considered the price for the monopoly of the India trade even though the Company paid an annual amount to the Government in the form of convoy and license money for the protection it enjoyed in times of war by convoying (Glamann 1981:6-7). The initial payment of 25,000 ‡. (0.4% of the initial capital stock) was waived by the Government as the state’s modest contribution to the assets of the company (Glamann 1981:6). The very small participation of the Government in the initial venture may be indicative of its cash constraint, as expected in light of the Eighty Years War (1566-1648) between Spain and the United Provinces, from which the independent Netherlands emerged. The subsequent payments to the Government for the renewal of the charter (1.5 million ‡. in 1647 and 3 million ‡. in 1696) continue to indicate a constrained central power, which seems to be consistent with the merchant controlled VOC. The Swedish Case From its start, the Swedish East India venture was pressed by …nancial constraints. As early as 1626, a merchant obtained a charter to trade with the East for a period of twelve years. But the Swedish entry into the Thirty Years’War in 1630, consumed all available resources in the kingdom. After the peace treaty of Westphalia in 1648, there were new attempts to re-launch Swedish Asian trade, but again armament was given priority (Hermansson 2004:9). The constant participation of Sweden in armed con‡icts over the late seventeenth and early eighteenth centuries (coinciding with the belligerent reign of Charles XII, 1697-1718) made it di¢ cult to engage in overseas trade.

18

A few petitions for charters of Eastern trade were …led by merchants in the decade after Charles’death, but all were refused by the Parliament – capital was still insu¢ cient (Hermansson 2004:10). Only in 1731 was a petition successful, when a number of …nanciers shifted their money away from the discharged Ostende Company dissolved that same year.30 The charter was valid for a period of …fteen years and granted the company the sole right to carry Swedish trade and shipping to the east of the Cape of Good Hope. It was subsequently renewed four times for periods of twenty years (Koninckx 1999). There were no restrictions on foreign capital …nancing the company, but board members were required to be "Swedish citizens of Protestant faith" (Hermansson 2004:30). Swedish citizenship could be arranged if the person resided in Sweden, was interested in the business, and had the su¢ cient requisite knowledge and capital. The board was composed by four directors: two foreigners of French and Scottish origin, and one of the two other directors had previously worked for the Ostende Company, showing not only physical, but also human capital was in fact very much circulating throughout Europe at the time. Even a small and remote kingdom like Sweden could attract foreign capital with the purpose of Eastern trade. The late entry of the Swedes into Eastern trade due to successive cash constraints and the willingness to overlook citizenship in the higher interest of attracting capital, seem to indicate that the whole kingdom was still cash constrained by 1731. The franchising of long-distance trade to private merchants seems to have been the best option to the Swedish monarch.

3.3

Mixed Control

The cases of mixed control are those in which, despite chartering monopoly rights of trade to merchants, monarchs were still able to keep a non-negligible role in the destinies of the companies, e¤ectively sharing control with merchants. 30

This episode shows not only that merchant capital circulated in Europe by the eighteenth century, but also that kings were relatively more abundant, which conferred them with less bargaining power in the negotiations.

19

The Danish Case Chartered in 1616, the Danish EIC had very limited trade within Asia, when compared to the English or the Dutch counterparts. Danish trade was centered mainly on the southern coast of India and Ceylon. Similar to the English EIC, the Danish company was chartered after a proposal to the king by two Dutch merchants, with the objective of competing with the English and Dutch companies. The company faced capital problems from the start and the monarch had to exert royal pressure to furnish the …rst share of capital (180,000 rd.), which was subscribed to by about 300 shareholders. The royal family contributed with 12.5%, the nobility with 15.5%, citizens of Copenhagen and the university with 35%, citizens of other towns of Denmark with 29.5%, and foreign capital 7.5% (Feldbaek 1981:139-140). The monopoly of trade between Denmark and Asia was chartered to the company for a period of twelve years. Initially the company was very much in‡uenced by its Dutch counterpart as opposed to the English: intra-Asian (rather than Euro-Asian) trade dominated the administration of the company (Subrahmanyam 1989). The king’s role was decisive in maintaining the company alive in the 1630s-40s after the ruinous Danish participation in the Thirty Years’ War, a major disturbance in the company’s operations, as in the Swedish case. By 1630, the king owned half of the shared capital and e¤ectively controlled the company: not only did the board of directors function practically as a group of civil servants, but also the original shareholders withdrew and were substituted by others on whom the king could exert investment pressure (Feldbaek 1981). The reorganization of the company according to the French model in 1670, again showed the king as the chief supplier of capital, while insisting on the support of Copenhagen merchants (Furber 1976). Denmark o¤ers a hybrid case. The …nancial constraint of the king is evident as he encouraged (perhaps forced) private investment in the company, which was not negligible. Despite choosing to franchise out long-distance trade, the Danish king was still able to control the company for periods of time.

20

The French Case The French Company was chartered in 1664 to direct French trade with the Eastern Hemisphere. The government not only took part in the foundation (sometimes even the initiative), but also recruited members, named directors, and procured the capital needed (Coornaert 1967). There was never a body of directors, and high ranked o¢ cials were appointed by the ministries. The initial capital stock of 15 million livres was in fact a forced loan launched by the royal subscription of 3 million – total subscriptions amounted only to 8,179,885 livres, in which merchants were a feeble minority –only 650,000 livres subscribed by the Parisian mercantile community (Furber 1976:203). The Company’s monopoly of French trade with the East was leased to groups of merchants, whose contracts were initially valid for single voyages and later for periods of three and ten years, whereas the company’s charter was valid for a period of …fty years (Manning 1996). The king was such a central …gure in the company that this organizational form is somewhere in the middle of the two extreme cases of private and crown monopoly. Private merchants were allowed to freight goods from the Indies on the company’s ships or on their own, which proved to be a successful fund-raising method for the company. The continuous mixture of private and royal interests can also be seen in other episodes of the company’s history, such as the case of the drained …nances due to participation in the war of Spanish succession when the company turned to the Crown for a loan. The king agreed with the condition that the directors each put up a certain amount and the stockholders 50% of the value of their holdings (Furber 1976:207). France o¤ered special conditions that allowed for this hybrid case to persist overtime: on the one hand, private merchants were reluctant to invest in a company they could never control as there were no shareholder meetings; on the other, the French state was su¢ ciently decentralized, which inhibited the company to merge into a department of state as in the Portuguese case (Manning 1996:24).

21

The involvement of the king in the private enterprise is unusual from the standpoint of the franchised structure. Similar to the Danish case in which the king could not initially supply all capital for the enterprise but had to raise the rest through forced loans, the French monarch was also indeed capital constrained. As a result, long-distance trade was franchised out. France o¤ers another hybrid result because the king was able to exert some control over the destinies of the company even if not owning the monopoly.

4

Discussion of Alternative Hypotheses

In spite of the suggestive evidence on the role of royal …scal capacity, at the establishment of merchant empires crowns faced many di¤erent circumstances which might have in‡uenced their organizational choice. These circumstances could have been induced by diverse external factors such as risk, timing, or competing parties, or by factors intrinsic to each country such as religious, social, or institutional di¤erences. I now address the two groups of alternative explanations for the divergent organization of merchant empires, polarizing the discussion between the extreme cases of crown and private ownership.

4.1

Di¤erent External Factors

Countries faced di¤erent levels of risk at the emergence of their merchant empires. The Portuguese empire established in the last decade of the …fteenth century, was preceded by the discovery of the Cape Route, a long lasting and costly investment – sponsored by the crown – which involved venturing into the unknown world at the time. The public good nature of the discoveries and the high degree of risk and uncertainty of the world at the time, could have repelled private merchants from such a volatile business as long-distance trade, leaving the door open for royal intervention. In Portugal, the king’s decision to control the monopoly of trade in eastern spices could have been necessary to compensate his prior investment in exploration. All other chartered …rms were founded in the seventeenth century

22

(or later) when the map of the world was already drawn. By that time, the merchants would be willing to enter the long distance trade market as they could only then expect positive and less volatile returns from it. This explanation is unconvincing on two grounds. First, perceptions of risk are historically unobservable, but even if it were possible to measure risk, economic theory gives implausible predictions in the presence of moral hazard. We could have a plausible scenario in which the relatively poorer merchants should control the …rm and bear all risk because the king would be much harder to incentivize (Newman 2007). Second, the arguable nature of public good in the discoveries does not match historical evidence. If discoveries had the nature of a public good, no merchant would be interested in its provision. Even though since 1415 Portuguese exploration voyages were indeed encouraged by Prince Henry the Navigator, between 1469 and 1474 the crown rented the discovery of the Guinea coast as well as its trade to a private Lisbon merchant. The alternative of franchise, not only of trade, but also of coastal exploration, was therefore available to the Portuguese king, suggesting it could have been an option after the successful voyage of Da Gama in 1498 in exchange for customs taxes to repay for the prior investment in exploration voyages. Timing can also be a relevant external factor a¤ecting the organizational choice of merchant empires. Early entry could have been cheap because Portugal established its empire in unexplored areas, whereas late entry could have been costly because English and Dutch had to share territories and …ght the established Portuguese empire. If later entrants were also crowns they would be stealing revenue from another monarch, which could constitute a casus belli and result in major war expenses.31 It is however unclear how the cost of discovery supported by Portugal relates to the cost of displacement supported by England 31

The threat of a casus belli seems far more credible a century before the establishment of the English and Dutch merchant companies. The discovery of the all-sea route to Asia in 1487 opened the Portuguese Golden Age in the East. Portugal’s exclusive control of the Cape Route remained intact as con…rmed by Columbus’voyage to the East via West and by the unsuccessful searches of the Northern passage to Asia. By the seventeenth century however, Portugal’s military strength was greatly diminished as the attacks by the Dutch in the Indian Ocean attest. To the Netherlands, Portugal was no longer a feared power to avoid, but rather the enemy since the Iberian Union in 1580.

23

and the Netherlands, as none of these costs has been measured. Early entry may not have been that cheap given that Portugal’s venture into the unknown consisted in drawing the map of the world as we know it today. On the other hand, late entry might not have been as costly given the fast displacement of the Portuguese in the East, as shown on Figure 1. For ninety years the Portuguese crown was alone in this market, but in the late sixteenth century, prior to the foundation of the English and Dutch private companies, the Portuguese share dropped abruptly to 53%: even the individual Dutch merchants organized in the small companies that preceded the VOC, or the private English interlopers without a chartered company, caused great damage to Portugal’s trade. Competition may be another factor explaining monarchs’ choices with regard to the organization of merchant empires. Initially kings would choose to be in charge if alone in the market reaping all bene…ts (as in the case of Portugal in 1498), but as more agents entered the business, competition decreased bene…ts and kings would prefer not to enter (as in the case of England in 1600). This was unlikely the case in the far from exhausted markets of eastern spices, which seems to be con…rmed by the foundation of merchant companies for long-distance trade with the east throughout the seventeenth and eighteenth centuries. From a revenue maximizing perspective it would be best for the king to get full share of pro…ts than a small share of pro…ts in the form of customs taxes either earlier or later on.

4.2

Intrinsic Country Di¤erences

More than external factors associated with risk, timing, or competition, the di¤erences across empires may have had more to do with the fact that the empire building countries were inherently very di¤erent from each other in as varied aspects as religion, society, and institutions. Beyond their trade objectives, the Portuguese were also fervent converters of souls. The English did not share the latter purpose possibly because there was no king controlling the enterprise, in which case religion was a consequence of organizational control rather than a causal di¤erence. The merchant controlled English EIC never exhibited religious concerns, 24

but the English crown founded its own religion in 1534 prosecuting followers of Rome shortly after. It would thus be natural to see an interest in religion in the English empire had the English monarch been in charge of Eastern trade. Another intrinsic di¤erence between countries is society. It can be argued that Portugal and England did not have the same social development level, which allowed for merchant initiative to emerge in the latter country but not the former. If so, Portuguese and English monarchs could not have chosen over the same menu of options when deciding how to organize their overseas empire. While the English monarch could have owned or franchised the monopoly of trade, the Portuguese king would have been forced to get involved in business due to the lack of entrepreneurial merchants in the country. In fact, Portugal lacked the background of the English Magna Carta which, since 1215, imposed voluntary restrictions on royal initiative and allowed English subjects to engage in private market oriented activities. Despite the more centralized scenario in Portugal, the lease of the exploration and trade o¤ of the Guinea coast in 1469-74 shows that franchising was indeed an option for the Portuguese king. The same held true in 1587, 1628 and 1697 when the monarch sold the rights of Asian trade to private ownership. Moreover, history provides evidence of private corporations of craftsmen and artisans in medieval Lisbon (Marques 1982, Serrão 1980). If Portugal had capable merchants to handle trade and exploration, perhaps the problem was the king who preferred an old-fashioned/medieval solution to a more modern one, in which case the royal institution is backward. It is an arduous, and somewhat arbitrary task to justify one monarch’s myopia and another monarch’s unusual visionary insight, as well as it is to consider merchants in a given country sluggish and in another gifted.32 In the model, kings are no di¤erent across countries, nor are merchants. Whether kings chose medieval forms of enterprise (or merchants engage in entrepreneurial activities) they do so responding to incentives whose knowledge is vital to understand the course of history. 32

Such assessment may be fairly vindicated by the time of the industrial revolution in England, a process which Portugal only took-up in the late nineteenth century. In the eve of merchant empires however, it seems plausible that both countries shared a common development level on account of the similar overseas trade opportunities unfolding to both.

25

In early modern Europe, international trade was a royal prerogative: the king had the privilege of exploration or, alternatively, the king could grant rights of exploration. Even if the king decided to explore business himself, he could not travel to the East and negotiate in person the spice agreements with local rulers. Cooperation between king and merchants was thus necessary. Whether the institution was controlled by one or the other party would result from the relative bargaining position of king and merchants at the emergence of each empire. In this context, the bargaining power of king and merchants can be a¤ected by two complementary factors. First, the relative scarcity or abundance of kings and merchants could have played a role at the negotiation table, which history seems to con…rm: initially, when there were few kings willing to engage in long-distance trade, merchants had little room for negotiation and would settle for work in a royally controlled …rm; as more kings became involved in the long-distance trade, bargaining power favored merchants who would then be able to choose the monarch who best tailored their interests, and would thus be able to control the enterprise. Second, the …nancial ability of each party and the availability of its outside options, would give leverage in the negotiation for control. In this case, bargaining power becomes historically observable in the contribution of kings and merchants to eastern ventures. The incentive behind the backward institution is thus materialized, justifying the di¤erent organizational choice of merchant empires.

5

Di¤erent Firms?

The diverse …nancial situations of the monarchs at the emergence of merchant empires resulted in di¤erent control choices. But did this imply any noticeable di¤erences across …rms trading analogous exotic products, in similar distant locations, and using the same technology? According to the last prediction of the model, the control structure was not a mere label: an enterprise controlled by the king should have been fundamentally di¤erent from an

26

enterprise with the merchants in charge. In light of the model I now turn to the analysis of the surviving historical data to evaluate the potential implications of …rm ownership.

5.1

Labor Compensation

Whether a king or a board of merchants, the organizer of overseas ventures faced a classic agency problem: how to provide future hires with incentives to work hard in distant locations where they could not only shirk, but also smuggle merchandise thus reducing the venture’s returns. Standard labor theory tells us that when monitoring is costly principals implement less of it while inducing workers’e¤ort by linking compensation to performance, such as with the use of bonuses, but when monitoring is free they pay …xed wages. In merchant empires, a royally owned …rm would have less delegation than that controlled by the merchants, making monitoring harder and resulting in eKC < eM C in the model. This moral hazard implication can be assessed historically across empires with di¤erent control structures. I have collected two datasets on the labor compensation of Portuguese and Dutch workers overseas from the sixteenth to the eighteenth centuries. The Portuguese dataset was extracted from a document ca. 1582, which provides a survey of all 198 job posts and respective compensations for the complete network of Portuguese strongholds around the Indian Ocean (Luz 1960).33 The Dutch dataset was collected from Lequin (1982), which provides complete career records for all 115 workers belonging to the directory of Bengal in the eighteenth century, not necessarily exclusively. The dates of the two datasets are not coincident, but workers’ compensation is still comparable, since both start when the two empires are already well established in the East. Moreover, during the time period intercontinental communication remained slow, maintaining the problems of hiring honest workers and monitoring them. All 313 Portuguese and Dutch workers were hired in their home country.34 33

The Portuguese loss of independence to Spain in 1580 merged the overseas Eastern and Western Iberian possessions. The document is addressed to the Spanish monarch with the purpose of providing information about the long-standing conditions of the newly acquired eastern empire. 34 The English and Dutch companies di¤ered quite substantially in terms of the incentives of top managers

27

Labor compensation is divided into two parts: a wage paid by the Portuguese Royal Treasury or by the VOC; and a variable bonus paid in kind but reported in the value of the exports of spices on the worker’s account.35 Table 6 shows the bonus shares of total compensation in the Portuguese and Dutch …rms, a convenient variable to analyze because it abstracts from the di¤erent monetary units in the two samples. The average Portuguese share is larger than the Dutch by a magnitude of 1:65. Median ratios and standard deviations show the same tendency, indicating a more dispersed distribution in Portugal. The data certainly deserves further study, but this brief analysis indicates that companies with di¤erent control structure varied quite substantially with respect to labor compensation.

5.2

Shipping Technology

Long-distance trade companies faced the most varied business decisions, which, in the model, should vary according to the party in charge (dKC > dM C ). We can evaluate this prediction by analyzing the history of shipping technology in merchant empires. In the early …fteenth century discoveries, the Portuguese used small and easily maneuverable ships suited for navigation in unmapped coastlines, but the increased volume of trade gave place to large cargo vessels in the sixteenth century. The …rst attempts to introduce ships larger than 500 tons in the Portuguese India trade date from the 1520s. By the 1550s, large galleons of 900 to 1,000 tons were not uncommon. From 1551 to 1570 the steady increase in tonnage lead sea captains to pressure for limitations on ship size, which led the monarch to decree that all ships in the Asian trade should be between 300 and 450 tons. Under Spanish rule (1580-1640) however, the 1570 law was relaxed and it became a yearly practice to build two or three ships over 1,000 tons each (Boxer 1968:13). In 1588 upon royal located in Europe. Unlike the English, the Dutch board of governors’ earnings not only depended on the company’s pro…t, but also on its turnover (Irwin 1991). The sample however, includes only overseas workers, who are the subject of analysis here. 35 In the Portuguese case, the value is that which each worker could import over the three-year term of the contract. In the Dutch case, it is the value each worker actually imported over his career, which lasted twenty-two years on average.

28

initiative, Lisbon witnessed the construction of large galleons of 700 to 900 tons.36 Larger ships carried more merchandise but were also more di¢ cult to maneuver under storms, which made them less seaworthy and ultimately resulted in higher wreckage rates.37 In the seventeenth century, smaller and more e¢ cient Dutch cargo ships were being used throughout Europe (Unger 1978). As a result of its specialization in naval construction, Holland was supplying ships to the French India Company, England, Hamburg and Ostende, as well as Denmark and Sweden (half of whose ‡eets were Dutch-built), and even the Spanish colonial trade (Barbour 1930:286-7).38 If the decisions on shipping technology in Portugal were not following a standard business perspective but were subject to the whims of the monarch, then the slower adoption of the more e¢ cient ships should have resulted in a higher wreckage rate. This pattern seems to emerge from the historical evidence gathered for the Portuguese and Dutch merchant empires. Falcão (1859) presents a detailed list of all Portuguese ships sailing out of Lisbon to India from 1497 to 1612.39 Table 7 summarizes this information in two periods 1497-1579 and 1580-1612, before and after Spanish rule. Out of 806 ships sent over the complete period, 52.7% returned safely to Lisbon. Of those that did not return to Lisbon, 35.4% ships stayed in the East for defense purposes. Summing the two rates, we get 88.1% of successful voyages. Out of the 11.9% ships that ended in failed voyages, 2.5% returned after aborting the voyage, 8.2% were lost in shipwreck, 0.5% were taken by enemies, and 0.7% were voluntarily burned.40 The most substantial rise in the percentage of failed voyages is that of shipwrecks: before 36

The timing coincides with the adventures of the Spanish Armada, in which the Spanish – and at the time, also Portuguese –king was heavily involved. It is therefore not surprising that the crown intervened in the construction of large war ships. For an account on the participation of Portuguese ships in the Spanish Armada, see Costa 1997:146. 37 The emergence of the popular literature of tragic voyages in the second half of the sixteenth century is no coincidence (Brito 1959, Brito 1968, Burman 1967, Lanciani 1990). 38 The pervasiveness of Dutch ships all over Europe is consistent with the assumption that these technologically advanced ships were generally available to all merchant empires and not just the Netherlands. 39 Royal involvement in the Asian trade as the main merchant and armor lead to the specialization of the Lisbon yards in the construction of large vessels (Costa 1997). 40 Though rare the voluntary burning of ships was not unheard of. It would occur in extreme situations of crew depletion on account of battles or scurvy.

29

1580 one in every twenty ships was lost, whereas after 1580 one in every …ve su¤ered that same fate. Assuming Portuguese navigation malpractices –delayed departures, ship overcrowding and overloading –remained unchanged throughout the period (if anything these should have been reduced with learning), then the increase in ship size seems to be the only other factor capable of raising the loss rate. On the Dutch side, the VOC purchased some ships from private shipyards and built some ships itself. Table 8 shows the sizes of the ships built by the VOC in the seventeenth century. Though the share of large ships of 800 tons or more rises from the …rst (10%) to the second half of the seventeenth century (14%), the overwhelming majority (86%) of vessels built in the company’s yards have smaller tonnage. This value is likely to be understated, given the purchase of ships from private shipyards (mostly ‡uits and other small vessels). Given the approximate composition of the VOC’s ‡eet I now turn to Table 9, which summarizes the losses of in VOC voyages throughout the seventeenth century. The percentage of losses due to wreckage or capture rises from the …rst to the second half of the century for both outbound and inbound voyages, however, the increase is rather small (0.5% in outbound voyages, and 1% inbound). Overall losses throughout the period vary between 3% and 4.2%, a much lower number than that veri…ed for Portuguese ships. Handling a larger volume of trade, the Dutch conducted more voyages with a much lower loss rate, which seems to be associated with technology adoption, another di¤erentiating factor across royal and merchant controlled empires.41

6

Concluding Remarks

When European overseas expansion began, monarchs were faced with a choice of how to organize it. Portugal, the …rst European country to establish direct trade connections with 41

Granted the Portuguese had a period of adaptation to, and learning of, the Cape route that may have increased the loss rate, but so did the Dutch. Throughout the sixteenth century, Portuguese ships could be seen (and copied) in ports throughout Europe, but navigation knowledge was treated with the most secrecy. Moreover, the Dutch specialized in a slightly di¤erent route in the Indian Ocean. The VOC headquarters were located in Batavia (now Jakarta) so the Dutch also had some learning to acquire.

30

the East, chose to keep control of trade, with the king as the ultimate residual claimant on pro…ts. In England, the Netherlands and Sweden, the rights of eastern trade were franchised to private merchants, who e¤ectively controlled the business. In Denmark and France there were also private companies but the monarchs were still able to exert strong in‡uence upon them. The di¤erent organizational choices appear puzzling since the goals of empire should be similar across countries, but diverse circumstances may have a¤ected royal decisions. In this paper I discussed many relevant circumstances and their possible channels of in‡uence, but focused on the …scal position of the crown: if the monarch was not cash constrained ownership was the best alternative; but a …nancially constrained monarch would sell monopoly rights to private merchants. To help illuminate the historical facts, I presented a simple model in which the decision faced by monarchs with respect to the choice of organizational form is determined by surplus division rather than maximization, so that bargaining power matters. The historical evidence on all …ve eastern empires seems to agree with the above rationale. The necessary investment costs in the Portuguese expansion were relatively small compared to expected returns, so it is not surprising that the Portuguese king chose to keep control of the spice trade. On the contrary, …nancially pressed by heavy borrowing to …ght wars, English and Swedish monarchs chartered monopoly rights of eastern trade to the respective East India Companies. In the Netherlands the situation was similar with the States-General encouraging the merger of small trading companies into the VOC, but staying out of business after a long war of independence against Spain. In Denmark and France, the monarchs seem to have been somewhat …nancially constrained: though they …nancially contributed to the companies, they pressured (and sometimes forced) private merchants to do the same with the added incentive of private ownership; nevertheless these two countries show evidence of mixed control between king and merchants. Beyond the implications on the …nancial structure of the companies, the model also suggests that di¤erences in organizational control have implications on the …rms at di¤erent

31

levels, providing further insights on the historical evidence of merchant empires. Using archival data on labor compensation, I …nd that workers employed in a royally controlled …rm were more likely to receive a higher fraction of their compensation in the form of bonuses than workers employed in a private company, which is consistent with more di¢ cult monitoring in a more centralized …rm. Using historical records of shipping technology across merchant empires, I also …nd that royal ownership often led to the choice of larger ships, more suited to …ghting wars than to conducting trade. These were less seaworthy ships, which can be seen in a higher wreckage rate in the crown controlled …rm. Studying the economic incentives behind the organizational choice of merchant empires is a step forward in understanding the historical developments of the seventeenth and eighteenth centuries which set the stage for the industrial revolution.

32

References [1] Acemoglu, Daron, Simon Johnson, and James Robinson. 2005. "The Rise of Europe: Atlantic Trade, Institutional Change, and Economic Growth." American Economic Review 95(3): 546-579. [2] Allen, Robert. 2001. "The Great Divergence in European Wages and Prices from the Middle Ages to the First World War." Explorations in Economic History 38: 411-441. [3] Barbour, Violet. 1930. "Dutch and English Merchant Shipping in the Seventeenth Century." Economic History Review 2(2): 261-290. [4] Barros, João. 1988. Ásia de João de Barros. Lisboa: Imprensa Nacional – Casa da Moeda, (Orig. pub. 1552). [5] Blair, Roger D. and Francine Lafontaine. 2005. The Economics of Franchising. Cambridge, UK: Cambridge University Press. [6] Boxer, Charles R. 1968. Fidalgos in the Far East 1550-1770. Hong Kong, London [etc.]: Oxford U.P. [7] Braudel, Fernand. 1972. The Mediterranean and the Mediterranean world in the age of Philip II. New York: Harper & Row. [8] Brewer, John. 1989. The sinews of power: war, money, and the English state, 1688-1783. New York: Alfred A. Knopf. [9] Brito, Bernardo Gomes de. The Tragic History of the Sea 1559-1565. Translated and edited from the original Portuguese by Charles R. Boxer. Cambridge: Published by the Hakluyt Society at the University Press, 1968 [1735-6]. [10] Brito, Bernardo Gomes de. The Tragic History of the Sea 1589-1622. Edited from the original Portuguese by Charles R. Boxer. Cambridge: Published by the Hakluyt Society at the University Press, 1959 [1735-6]. 33

[11] Burman, Jose. Great shipwrecks o¤ the coast of Southern Africa. Cape Town: Struik, 1967. [12] Bruijn, J. R., F. S. Gaastra and I. Schö¤er editors, with assistance of E. S. van Eyck van Heslinga, ed. 1987. Dutch-Asiatic shipping in the 17th and 18th centuries. Vol. I, Introductory Volume. The Hague: Martinus Nijho¤. [13] Cawston, George and A. H. Keane. 1896. The Early Chartered Companies (A.D. 12961858). London, New York: E. Arnold. [14] Charles Blunt of London: Account book. 1684-1722. Manuscript held by the British National Archives (Ref. PRO, C114/165). [15] Chaudhuri, K. N. 1965. The English East India Company: The Study of an Early JointStock Company 1600-1640. London: Frank Cass & Co Ltd. [16] Chaudhuri, K. N. 1978. The Trading World of Asia and the East India Company – 1660-1760. Cambridge [Eng.]; New York: Cambridge University Press. [17] Columbus, Christopher. c1989. The Diario of Christopher Columbus’s …rst voyage to America, 1492-1493. Abstracted by Fray Bartolomé de las Casas. Transcribed and translated into English, with notes and a concordance of the Spanish by Oliver Dunn and James E. Kelley, Jr. Norman, OK : University of Oklahoma Press. [18] Coornaert, E. L. J. 1967. "European Economic Institutions and the New World; the Chartered Companies." In The Cambridge Economic History of Europe, ed.M.M. Postan and H.J. Habakkuk, Vol. 4, 222-274. Cambridge, MA: at the University Press. [19] Costa, Leonor Freire. 1997. Naus e galeões na ribeira de Lisboa: a construção naval no século XVI para a Rota do Cabo. Cascais: Patrimonia.

34

[20] Costa, Leonor Freire. 2002. O Transporte no Atlântico e a Companhia Geral do Comércio do Brasil (1580-1663). Vol. 1. Lisboa: Comissão Nacional para as Comemorações dos Descobrimentos Portugueses. [21] Daudin, Guillaume. 2005. Commerce et prospérité: La France du XVIIIe siècle. Paris: PU Paris-Sorbonne. [22] David, Paul A. 1985. "Clio and the Economics of QWERTY." American Economic Review, 75(2), Papers and Proceedings of the Ninety-Seventh Annual Meeting of the American Economic Association (May 1985): 332-337. [23] Davis, Lance and Douglass North. 1970. “Institutional Change and American Economic Growth: A First Step Towards a Theory of Institutional Innovation.” Journal of Economic History 30(1): 131-149. [24] Davis, Lance, Robert Gallman and Karin Gleiter. 1997. In Pursuit of Leviathan –Technology, Institutions, and Pro…ts in American Whaling, 1816-1906. Chicago: The University of Chicago Press. [25] De Sousa, Armindo. 1990. As Cortes Medievais Portuguesas (1385-1490). Vol. I. Porto: Instituto Nacional de Investigação Cientí…ca – Centro de História da Universidade do Porto. [26] DeVries, Jan. 1984. European urbanization, 1500-1800. Cambridge, Mass.: Harvard University Press. [27] Dickson, Peter George Muir. 1967. The Financial Revolution in England: a study in the development of public credit, 1688-1756. New York: St. Martin’s Press. [28] Dietz, Frederick. 1964. Charles. English Public Finance, 1485-1641. Vol. I. 2nd ed. London: Frank Cass and Co. Ltd.

35

[29] East India Company. 1893. The register of letters &c. of the Governour and Company of Merchants of London trading into the East Indies, 1600-1619, ed. Sir George Birdwood. London: Quaritch. [30] Elbl, Martin. 1994. "The Caravel and the Galleon –The Caravel." In Cogs, Caravels and Galleons: The Sailing Ship 1000-1650, ed. Robert Gardiner, 91-98. London: Conway Maritime Press. [31] Falcão, Luiz de Figueiredo. 1859. Livro em que se contém toda a fazenda real e património dos reinos de Portugal, India e das ilhas adjacentes e outras particularidades. Lisboa: Imprensa Nacional, (Orig. pub. ca1607). [32] Feldbaek, Ole. 1981. “The Organization and Structure of the Danish East India, West India and Guinea Companies in the 17th and 18th Centuries.”In Companies and Trade, ed. Leonard Blussé and Femme Gaastra, 135-158. Leiden: Leiden University Press. [33] Furber, Holden. 1976. Rival Empires of Trade in the Orient, 1600-1800. Minneapolis: University of Minnesota Press. [34] Glamann, Kristof. c1981. Dutch-Asiatic Trade: 1620-1740. Den Haag: Nijho¤. [35] Godinho, Vitorino Magalhães. L’Économie de l’Empire Portugais aux XVe et XVIe Siècles. Paris:S.E.V.P.E.N., 1969. [36] Goldsmith, Raymond W. 1987. Premodern Financial Systems: a historical comparative study. Cambridge, New York: Cambridge University Press. [37] Gollin, Douglas. 2002. "Getting income shares right." Journal of Political Economy, 110 (2): 458-474. [38] Greif, Avner. 1993. “Contract Enforceability and Economic Institutions in Early Trade: The Maghribi Traders’Coalition.”American Economic Review, 83 (3): 525-548.

36

[39] Grossman, Sanford J. and Oliver Hart. 1986. “The Costs and Bene…ts of Ownership: A Theory of Vertical and Lateral Integration.” Journal of Political Economy, 94(4): 691–719. [40] Hart, Jonhathan. Comparing Empires –European Colonialism from Portuguese Expansion to the Spanish-American War. New York : Palgrave Macmillan, 2003. [41] Hart, Oliver and John Moore. 1990. “Property Rights and the Nature of the Firm.” Journal of Political Economy, 98: 1119–158. [42] Hermansson, Robert. 2004. The Great East India Adventure –the Story of the Swedish East India Company. Göteborg: Breakwater Publishing. [43] Hunter, William Wilson. 1919. A History of British India. London: Longmans, Green, and co. [44] Irwin, Douglas A. 1991. “Mercantilism as Strategic Trade Policy: The Anglo-Dutch Rivalry for the East India Trade.”Journal of Political Economy, 99: 1296-1314. [45] Jacobs, Els M. 1991. In Pursuit of Pepper and Tea: the Story of the Dutch East India Company. Amsterdam: Netherlands Maritime Museum; Zuatphen: Walburg Pers. [46] Koninckx, C. 1999. “Sweden and India in the Eighteenth Century: Sweden’s di¢ culty in gaining access to a crowded market.”Merchants, Companies, and Trade: Europe and Asia in the Early Modern Era, ed. Sushil Chaudhury and Michael Morineau. London, New York: Cambridge University Press. [47] Lanciani, Giulia. 1990. "Une Histoire Tragico-Maritime." In Lisbonne Hors des Murs 1415-1580 - L’Invention du Monde par les Navigateurs Portuguais, 89-117. Paris: Autrement. [48] Legros, Patrick and Andrew F. Newman. 2008. "Competing for Ownership." Journal of the European Economic Association, 6 (6): 1279-1308. 37

[49] Lequin, Frank. 1982. Het personeel van de Vereengide Oostindische Compagnie in Azie. Leiden: F. Lequin. [50] Letwin, William. 1965. Law and Economic Policy in America: The Evolution of the Sherman Antitrust Act. New York: Random House. [51] Luz, Francisco Paulo Mendes da. 1960. Livro das Cidades e Fortalezas, que a Coroa de Portugal tem nas partes da India e das Capitanias, e mais cargos que nelas ha, e da importância delles. Lisboa: Centro de Estudos Históricos Ultramarinos, (Orig. pub. ca1582). [52] Manning, Catherine. 1996. Fortunes a Faire – The French in Asian Trade, 1719-48. Brook…eld, VT, USA: Ashgate Pub. [53] Marques, Oliveira A. H. 1982. História de Portugal. Vols. 1 and 2. Lisboa: Palas Editores. [54] Marsden, Peter. 2003. Sealed by Time –The Loss and Recovery of the Mary Rose. Vol. 1. The Mary Rose Trust. [55] Martinez-Hidalgo, José Maria. 1966. Columbus’Ships. Barre Mass.: Barre Publishers. [56] Munro, John. 1994. Handbook of European History 1400-1600 – Late Middle Ages, Renaissance and Reformation, ed. Thomas A. Brady, Jr., Heiko A. Oberman, and James D. Tracy. Leiden; New York: E.J. Brill. [57] Newman, Andrew F. 2007. “Risk-bearing and entrepreneurship.”Journal or Economic Theory, 137 (1):11-26. [58] North, Douglass C. 1991. "Institutions, Transaction Costs, and the Rise of Merchant Empires." In The Political Economy of Merchant Empires – State power and World trade 1350-1750, ed. James D. Tracy. New York: Cambridge University Press.

38

[59] North, Douglass and Barry Weingast. 1989. "Constitutions and Commitment: The Evolution of Institutional Governing Public Choice in Seventeenth-Century England." Journal of Economic History, 49(4): 803-832. [60] Outhwaite, R. Brian. 1971. "Royal Borrowing in the Reign of Elizabeth I: The Aftermath of Antwerp." English Historical Review, 86(339): 251-263. [61] Phillips, Carla Rahn. 1994. "The Caravel and the Galleon –The Galleon." In Cogs, Caravels and Galleons: The Sailing Ship 1000-1650, ed Robert Gardiner, 98-114. London: Conway Maritime Press. [62] Pollard, Abert Frederick 1910. “Henry VIII.” In The Encyclopædia Britannica. Cambridge, England: at the University Press. [63] Renault, Gilbert. 1959. The Caravels of Christ. London: Allen & Unwin. [64] Sanceau, Elaine. 1967. Good Hope – the voyage of Vasco da Gama. Lisboa: Academia Internacional da Cultura Portuguesa. [65] Senhadji, Abdelhak. 2000. "Sources of Economic Growth: An Extensive Growth Accounting Exercise." International Monetary Fund Sta¤ Papers, 47(1): 129-57. [66] Serrão, Joaquim Veríssimo. 1980. História de Portugal. Vol. 2. A Formação do Estado Moderno (1415-1495). 3rd revised ed. Lisboa: Editorial Verbo. [67] Serrão, Joel. 1977. Cronologia Geral da História de Portugal. 3rd ed. Lisboa: Livros Horizonte. [68] Steensgaard, Niels. 1965. "Freight Costs in the English East India Trade 1601-1657." Scandinavian Economic History Review, 13(2): 143-62. [69] Steensgaard, Neils. 1970. “European Shipping to Asia 1497-1700.” Scandinavian Economic History Review, 18(1): 1-11. 39

[70] Steensgaard, Niels. The Asian Trade Revolution of the Seventeenth Century –The East India Companies and the decline of the Caravan Trade. Chicago: The University of Chicago Press, 1974. [71] Subrahmanyam, Sanjay. 1989. “The Coromandel Trade of the Danish East India Company, 1618-1649.”Scandinavian Economic History Review, XXXVII(1): 41-56. [72] Unger, Richard W. 1978. Dutch Shipbuilding Before 1800. Amsterdam, Assen: Van Gorcum. [73] Vogt, John. 1979. Portuguese Rule on the Gold Coast 1469-1682. Athens: The University of Georgia Press. [74] Wee, Herman van der. 1963. The Growth of the Antwerp Market and the European Economy (fourteenth-sixteenth centuries). The Hague: Nijho¤. [75] Williamson, Oliver E. 1975. Markets and Hierarchies, Analysis and Antitrust Implications: a Study in the Economics of Internal Organization. New York: Free Press.

40

Figure 1: European Shipping to Asia 1.0

% of total tonnage

0.8 PT

0.6

EN NL

0.4

FR DN

0.2

16 91 -1 70 0

16 71 -1 68 0

16 51 -1 66 0

16 31 -1 64 0

16 11 -1 62 0

15 91 -1 60 0

15 71 -1 58 0

15 51 -1 56 0

15 31 -1 54 0

15 11 -1 52 0

14 91 -1 50 0

0.0

Source: Steensgaard (1970).

Figure 2: Outcome on the Pareto Frontier

v

KC

•A

(0, λc) MC

(λc, 0)

u

41

Table 1: Portuguese discoveries and conquests in the …fteenth century Year 1415 1419 1421 1427 1434 1436 1442 1443 1446 1446 1458 1460 1471 1471 1474 1481 1482 1487 1489 1498

Location Ceuta Madeira Islands Cape Não Azores Islands Cape Bojador Gold River Cape Branco Cape Verde –Senegal Grande River Guinea El-Qsar-es-Seghir Cape Verde Islands and Sierra Leone Asilah and Tangier Gulf of Guinea Cape Saint Catherine Mina –Ghana Congo River Cape of Good Hope Sofala –Mozambique Calicut –India

Region Northern Africa Atlantic Islands West African coast Atlantic Islands West African coast West African coast West African coast West African coast West African coast West African coast Northern Africa West African coast Northern Africa West African coast West African coast West African coast West African coast Southern African coast East African coast East

Source: Serrão (1980:175).

Table 2: Frequency of assemby of cortes from 1385 to 1580 1385-1414 1417-1495 1498-1580

Average interval (years) between meetings 1.5 3.0 9.1

Source: Serrao (1977), lists all years in which meetings were held. Note: I calculated the average interval between meetings in the three sub-periods.

42

Table 3: Estimated returns of the Portuguese expansion (Kg of Gold) 1415a 1419 1421 1427 1434 1436 1442 1443 1446 1458a 1460 1471a 1471 1474 1481 1482 1487 1488 1489 1494 1495 1496 1497 1498 1499 1500

wL 97.3 1.4 1.5 2.2 2.4 2.6 3.5 4.7 6.2 75.1 3.0 139.5 4.2 12.7 7.3 14.6 21.7 28.7 29.7 -

Cost rK – α1615 21.5 0.3 0.3 0.5 0.5 0.6 0.8 1.0 1.4 16.6 0.7 30.9 0.9 2.8 1.6 3.2 4.8 6.3 6.6 -

rK – α1859 116.1 1.8 1.8 2.6 2.9 3.1 4.2 5.6 7.4 89.6 3.6 166.5 5.0 15.2 8.7 17.4 25.9 34.2 35.4 -

Revenue 3,620b 226.3c 226.3c 226.3c 649.7c 649.7c 649.7c 372.6c 372.6c 372.6c 372.6c

Revenue wL + rK 11.8-21.2 4.8-8.5 3.6-6.5 5.7-10.3 -

Notes: aMilitary expeditions; bLoot from Asilah (Renault 1959:71); cEntries of gold in the Lisbon mint from the Gold Coast only (Vogt 1979:217-9). Vogt's series goes from 1481 to 1572 and is more precise for later years. Early years calculated by averaging the total quantity of gold that the head of the Lisbon mint received over his mandate (3-4 years).

Table 4: Forced loans by the late Tudors and early Stuarts Amount £20,000 £23,200 £111,891 £116,381 £96,466 £60,000

1598 1600 1604-5 1611-2 1617 1625

Interest rate ? ? 10 10 10 8

Sources: North and Weingast (1989:820); Dietz (1964:64).

Table 5: Cost of the East India Company voyages in the …rst 15 years Voyage # 1601 1604 1607 1608 1609 1610 1611 1612 1612 1613-16* Totals

1 2 3 4 5 6 7 8 9 10-13 14

Ships sent out 4 4 3 2 1 3 4 4 1 29 55

Cost of Ships and Victuals £39,771 £48,150 £28,620 £14,000 £6,000 £32,200 £42,500 £48,700 £5,300 £272,544 £538,385

Source: Hunter (1919:291, 307) Note:*First joint stock, as opposed to the prior system of separate voyages.

43

Discounted Cost (1601) £39,771 £36,176 £16,155 £7,492 £2,799 £13,656 £16,386 £17,069 £1,858 £75,700 £227,062

Table 6: Descriptive statistics: Bonus Shares Average Median St. Dev. Min. Max. N

Portugal .600 .765 .396 .000 1.00 198

Netherlands .363 .466 .209 .000 .895 115

Total .511 .500 .358 .000 1.00 313

Sources: Lequin (1982), Luz (1960).

Table 7: Ships in the Carreira da India 1497-1612 1497-1579 # % 620 100.0 325 52.4 6 1.0 31 5.0 0 0.0 2 .3 256 41.3

Sent Returned Aborted voyage Lost Taken by enemies Burned Stayed in the East

1580-1612 # % 186 100.0 100 53.7 14 7.5 35 18.8 4 2.2 4 2.2 29 15.6

1497-1612 # % 806 100.0 425 52.7 20 2.5 66 8.2 4 .5 6 .7 285 35.4

Source: Falcão (1859) lists the absolute numbers.

Table 8: Ships Built in VOC Shipyards 500-800t 68

800-1,000t 14

>1,000t 14

66%

24%

5%

5%

1650-99

230

131

29

39

54%

31%

6%

9%

1600-99

412

199

42

53

58%

28%

6%

8%

1600-49

<500t 182

TOTAL 278 428 706

Source: Bruijn, Gaastra, Schöffer (1987:52).

Table 9: Losses of Ships in the VOC Voyages 1600-49 1650-99 1600-99

Wrecked outbound inbound 18 11 23 23 41 34

Captured outbound inbound 2 2 12 7 14 9

% Total Voyages outbound inbound 2.7% 3.5% 3.2% 4.5% 3.0% 4.2%

Source: Bruijn, Gaastra, Schöffer (1987:75, 91). Notes: 715 (368) outbound (inbound) voyages from until 1649 and 1,107 (662) after.

7

Appendix

7.1

Model Solutions

In stage (iii) merchants choose optimal e¤ort e = [s c(1

)]1= D, given c, d, and .42 The

choice of d in stage (ii) occurs after the realization of the state of the world , and varies according to the control structure: a) MC: maximization of u results in dM C = , which gives DM C = 1, and e¤ort eM C = [s c(1

)]1= . Utilities for merchants and king are uM C =

42

1

[s c(1

)]1= , and vKC =

Notice that bargaining in stage (i) occurs over the utility possibilities frontier generated in expectation of e which is chosen after d is chosen in stage (ii). Therefore, for a given s, when the king chooses d he is always at least as well o¤ as when the merchant does.

44

1 s [s s(1 )

)]1= +

c(1

;

b) KC: maximization of v results in dKC = ing a distorted value of DKC = 1 2 s2 (1

)]1=

2(1 2 s2

2(1

s)2 [s

s)2 [s

(1 ) c(1 )]1=

)2

c(1

)]1=

2 s2 (1

2(1

s)2 [s

c(1

s(1 ) (1 s)[s c(1 )]1=

+ )2 )]2=

>

, which gives e¤ort eKC = [s c(1

. Utilities for merchants and king are uKC =

, and vKC =

1 s [s s(1 )

c(1

)]1= +

= dM C , generat-

2 s(1

2(1 s)[s c(1

) )]1=

1

+

[s c(1

)]1=

.

Varying s will change e , d , and the utility pair (u ; v ) which result in a concave frontier for each control structure, as shown in Figure 3 for all interior solutions e > 0.43

Figure 3: Utility Possibilities Frontiers v

KC

MC

u

Only the decreasing portions of either frontier are relevant to the analysis as they constitute the set of Pareto dominating sharing rules: increasing the share of any given party increases its utility without raising the other party’s. Also Pareto dominated is the decreasing portion of MC (KC) that lies below KC (MC). Therefore the dashed portions are dropped from the analysis in section 2. The merchants’share s, increases as we move to the right on either frontier, increasing 43

Parameter values:

= 2, c = 1,

= :5,

= :26 and

45

= :7.

e and also u: the rise in share of output more than compensates merchants’ disutility of e¤ort. Moving up on either frontier increases the king’s share (1

s), and despite the lower

e¤ort level e (@e=@s > 0 merchants have little incentive to work hard the higher the share going to the king) the king gets better o¤: the increase in the king’s share compensates the lower output under MC, while under KC the king is getting utility from increased d. It can be shown that u =

1

e under either control structure, so if eM C > eKC then

uM C > uKC , that is, merchants are always better o¤ under MC, which is visible in Figure 2. On the other hand, because dKC > dM C , the king is always better o¤ when in control vKC > vM C , despite the lower level of output under this control structure. The consequences of control allocation have tangible impacts on the venture’s business decisions d (such as the size of ships), on how much the merchants get invested in the venture e (and on the venture’s viability

7.2

as a consequence), and also on the generated social surplus u + v.

Maritime Voyages: Labor Requirement and Duration

The voyage from Lisbon to Ceuta lasted 28 days, but chroniclers are imprecise about other numbers involved in this conquest. Serrão (1980:22) reports 225 ships, and 45,000 men, which would give an average of 200 men per ship – possibly more, as several ships could have been employed in food storage, quite common practice in long voyages and/or large crews. However, two hundred men per ship is likely to be an overstatement for two reasons: (1) caravels, very much used in the 1400s, had modest crews of 20 to 40 men (MartinezHidalgo 1966). This lower crew size, however, may have been rule in discovery voyages, but not necessarily in military ventures; (2) Hunter (1919:159-161) provides numbers for military operations in the Indian Ocean in the sixteenth century. By then ships had increased in size and crews could be as large as 700 men per ship in larger vessels (Marsden 2003). In all battles for which Hunter provides numbers for ships as well as men, the average is 50 to 75 men per ship. Evidence from shipping technology and other battles seems to dismiss the idea of a large army of 45,000 men in the conquest of Ceuta. Taking the upper bound of 46

the caravel crew (40 men), I multiply it by the 225 ships reported in the chronicle (probably also overstated) to get a total of 9,000 men involved in this particular venture, surpassing all but one of the numbers provided by Hunter. All military expeditions are assumed to be of the size of Ceuta’s. In the case of exploration voyages, historical records often report only the numbers and type of ships used –barcas or caravels –but not the number of men. Because barcas were larger than caravels (Elbl 1994:92), I assume a crew size of sixty for the former and forty for the latter. Exploration voyages pre-1471 (until the Gulf of Guinea) were of pure exploration with very uncertain return, for which the monarch would send only one barca, such as in the 1419 voyage to the Madeira islands (Barros 1988). Voyages after 1471, were directly engaged in the search for the Eastern passage, in which case I assume they were of the same size as Da Gama’s voyage in 1498 –170 men and four ships. Da Gama’s expedition returned to Lisbon with only two ships and …fty-…ve (very sick) men on board (Sanceau 1967). My calculations, however, take into account the full original investment, that is, costs associated with men and ships sent. The 1415, 1487 and 1498 voyages to Ceuta, Cape of Good Hope and Calicut are well documented and we know their exact duration. For all other events, I estimate duration based on the distance from Lisbon to destination using either the baseline duration of the 1415 voyage for events before the Gulf of Guinea, or the duration of the 1487 voyage. In addition to the latitude and longitude coordinates of origin and destination, I also use those of intermediate locations in the ocean to approximate the course of the trip to the sea route.

7.3

Labor Share Calculations Portuguese India ship, 1615

1. New ship (11,150$000 reis) in the India trip and associated cost of wages and victuals (2,100$000 reis - four months, or 525$000 reis a month) for a crew of 123 (Falcão 1859:205).

47

2. Assuming each round-trip to India lasted twenty-four months (Steensgaard 1965)44 I then multiply the monthly value of wages and victuals by 24 months and calculate total labor cost wL1615 for the India trip as 12,600$000. 3. Assuming each ship made only four round trips (Phillips 1994:102, Steensgaard 1965) to India I divide the original capital cost by four and calculate the capital cost rK1615 for a ship in the India trip as 2,787$500. 4. Plugging the values for wL1615 and rK1615 in the labor share formula we get 1 2;787$500 1+ 12;600$000

1615

=

= :8188.

5. I omit the value of cargo because I am measuring the cost of …fteenth century exploration voyages, in which ships were not loaded with bullion. Portuguese ships to Brazil, 1633 and 1580-1640 Costa’s detailed study on Portuguese Atlantic transportation to Brazil provides speci…ed costs for the production of a 600 ton ship in 1633 and also estimated daily costs for the same type ship in the periods 1580-1601, 1602-1614, 1615-1623, and 1624-1640 (Costa 2002:866). For the 1633 ship the values are: (1) wages and victuals for 210 days: 994$980 reis = wL1633 ; (2) cost of the ship and artillery: 2,460$000 reis; (3) again assuming a life time for the ship of four round-trips rK1633 = 615$000 reis. The labor share yields

1633

=

1 615$000 1+ 994$980

= :6180.

Using daily costs per 600 ton ship, the estimated costs are 1/31 to 1/9 of the given revenues, perhaps re‡ecting shorter duration of seventeenth century trips. Return ratios would even be more magni…ed had I de‡ated seventeenth century numbers.

44

In theory, the round-trip to India could take less than two years given that the outward trip to Asia lasted 6-8 months and 7-9 months home (Steensgaard 1965). The weather conditions, however, dictated the optimal times of departure (from Europe in April, and from India in December) in order to avoid winters on the Cape and monsoons in the Indian Ocean. The time of absence of the ships was also a¤ected by political issues such as the need of ‡eets to …ght in the Indian Ocean. Steensgaard (1965) notes a decline (below 24 months) in the time of absence of Dutch ships only after 1640.

48

Table 10: Labor share: Portuguese shipping to Brazil 1580-1601 1602-1614 1615-1623 1624-1640

Wages and victualsa 2$106 1$600 2$887 2$671

Shipb 4$210 3$210 5$790 5$420

Ship (1 voyage)

1$052.5 0$802.5 1$447.5 1$355.0

α .6668 .6660 .6661 .6634

Source: Costa (2002:367) reports a and b. Note: Costa does not take into account fixed capital depreciation: labor share calculations with the total value of the ship for the periods of analysis yield the same as the calculations with the total value for the 1633 episode. I therefore divide the value of the ship by four, since the reported cost of ships and victuals is for the 210 day voyage.

English Merchant ship, 169345

Table 11: Labor cost calculation: English merchant ship Boatsmarins Gunners Carpenters Coopers Cooks Cash for beer and water Provisions for 100 men for 24 months Cost of Labor per trip (wL)

35 30 60 10 25 105 1900 £2,165

Table 12: Capital cost calculation: English merchant ship Ship 300 tons New ropes and store of others Rigging and two suits of sails Blockmaker Long boat Sheathing Nails Iron work Anchors and spare 20 guns Iron crows 20 barrels Small arms for 100 men Dry casks Brazer for hearth and bell Other costs River clearing at launch and stock to sea Cost of capital per trip (rK)

1800 560 325 35 30 150 176 45 100 96 20 70 100 35 65 30 200 £984.25

Source: Manuscript: Charles Blunt of London: Account book (1684-1722) available in the British National Archives (Ref. PRO, C114/165). Note: In the cost of capital per trip I assume that each merchant ship makes four round trips only.

Labor share: 45

1693

=

1 1+ $984:25 $2;165

= :6875.

Kindly provided by Joachim Voth.

49

American Whaling ship, 1859

Table 13: Labor cost calculation: American whaling ship Occupation Captain 1st mate 2nd mate 3rd mate Boatsteerer Cooper Carpenter Cook Steward Skilled Seaman Semi-skilled Seaman Unskilled Seaman Totals

Average mo. wage 104.48 70.92 42.52 27.32 16.87 27.29 8.77 10.18 10.31 9.42 9.24 7.49 -

Average crew per ship 1.00 1.07 1.07 1.07 3.70 1.44 0.91 1.00 0.99 2.19 2.55 10.78 27.77

Monthly labor cost 104.48 75.88 45.50 29.23 62.42 39.30 7.98 10.18 10.21 20.63 23.56 80.74 $510.11

Total (37 mo.) 3,865.76 2,807.72 1,683.37 1,081.60 2,309.50 1,454.01 295.29 376.66 377.66 763.30 871.79 2,987.46 $18,874.12

Source: Davis et al (1997) - Average monthly wages on p. 176, average number of crewman by ship on p. 155, average length of a trip on p. 258. Note: $18,874.12 is the expense in wages to which we need to add the cost of food and provisions for the crew in order to get wL.

Table 14: Capital cost calculation: American whaling ship Average outfit expense Cost of vessel (per ton) Average vessel size (tons) Cost of a vessel of 350 tons Average age of a vessel (mo.) Average length of a trip (mo.) Average number of trips per vessel Cost of a vessel per trip Food and provisions man/mo. Food and provisions crew (37 mo.) Cost of capital per trip (rK)

30,500.00 74.08 350 25,928.00 291.60 37 7.88 3,289.90 5.00 5,137.45 $28,652.45

p. 214 p. 244 p. 218 p. 231 p. 258 p. 211

Source: Davis et al (1997) Note: rK per trip = outfit – food and provisions + cost of vessel per trip.

1859

=

1 $28;652:45 1+ $24;011:57

= :4559.

The signi…cantly low

1859

is likely to result from technological improvements in shipping

(over more than two centuries), but also from the high excess capacity K=L of whaling ships which carried very large volumes of whale bone and oil on homebound trips with a relatively small crew.46 46

The average India ship in the early seventeenth century had 550 tons and a crew of 123 men, which gives a ratio of 0.224 men per ton. The average whaling ship in the mid nineteenth century had 350 tons for a crew of 28 men, yielding a ratio of 0.079 men per ton, almost three times lower than the former ratio. Moreover, when compared with other maritime enterprises, whaling ships used an unusually large amount of unskilled labor, which depressed labor relative to capital costs.

50