International Trade and Shipping

In the model of Van Zanden and Van Leeuwen, international trade and shipping contributed fl.42.6 million and fl.8.5 million respectively to Holland’s GDP in 1770, or 24.1 and 4.8 per cent. The overall size of all imports and exports of the Dutch Republic in the 1770s amounted to around fl.250 million, of which around eighty per cent or fl.200 million would have accrued to the province of Holland., The combined value added in international shipping and trade according to the model of Van Zanden and Van Leeuwen would thus contribute to GDP at approximately one quarter of the total value of all commodities coming into and going out of the Dutch Republic, measured in Dutch prices. Given the importance of Atlantic commerce in the total trade of Holland in the second half of the eighteenth century, it can be expected that the main contribution of slave-based activities to the GDP will be found under these two posts.2
1 J.L. van Zanden and B. van Leeuwen, ‘Persistent but not consistent. The growth of national income in Holland 1347-1807’, Explorations in Economic History 49 (2012) 119-130, and online annexes.
2 Patick O’Brien estimated that British inter-continental trade contributed 15 per cent of the total value of the trade in profits alone for merchant houses, shipping companies and insurers (the latter are subsumed under

Van Zanden and Van Leeuwen estimated the total value of trade with different regions of Europe, Asia and the Atlantic, and have derived benchmarks for value added in shipping and trade with each of these regions on the basis of generalized assumptions on the value of inputs before the commodities traded came into the hands of the merchants and the costs of shipping, wages and profits, and Holland’s share in total Dutch trade. While we would prefer to employ the reverse method of estimation, starting from the value added in shipping and trade for specific goods such as coffee and sugar, we will here of necessity start from our findings for the total value of slave-based commodity trade on different routes to calculate value added in shipping and trade following the procedure that Van Zanden and Van Leeuwen developed.
The first difficulty is to determine the total value (in current prices) of the trade on the different trading routes that was slave-based. All extant figures for the value of eighteenth- century trade are rough approximations on the basis of unreliable records for customs and excises, complemented with contemporary observations.3 This is true for the trade in all commodities, not only those that were slave-based. Our starting point thus must be to carefully compare slave-based trade streams with the existing estimates for the total trade of the Dutch Republic.
The slave-based trade activities can be divided in two parts.

a) Direct imports of slave-produced goods from the Atlantic world, both Dutch and non- Dutch, and direct exports from Holland to the Atlantic world for the maintenance of slave plantations in Dutch and non-Dutch slave-colonies;
b) Imports and re-exports of slave-produced goods within Europe, and imports of European goods for re-export for the maintenance of slave plantations in Dutch and non-Dutch slave-colonies;
Both parts contain a component of imports, exports and re-exports for the slave trade, which in some trade-estimates are at least partially treated as a separate stream, and in some are commercial profits in the model of Van Zanden and Van Leeuwen). Guillaume Daudin made a similar estimation for France and England on the basis of slightly different assumptions, and came at 17.6 per cent for France and 17.3 per cent for Britain. All three figures cited in G. Daudin, ‘Profits du commerce intercontinental et croissance dans la France du XVIIIe siècle’, Revue économique 57:3 (2006) 605-613, 606. The GDP of course does not only measure profits, but also other forms of income (wages and rent).
3 These problems are discussed at length in J. de Vries and A. van der Woude, The first modern economy. Success, failure, and perseverance of the Dutch economy, 1500-1815 (Cambridge 1997). FULL BIBLIOGRAPHICAL REFERENCES NEEDED THROUGHOUT

subsumed under a general heading such as ‘Atlantic trade’. Given the importance of this particular branch for debates on the economic contribution of slavery to the Dutch economy so far, we will present activities related to the slave trade within a and b under separate headings marked α and β. The total size of a and b (which thus includes α and β of the slave trade) can then be used to calculate value added in shipping and trade per trade route using the same methods employed by Van Zanden and Van Leeuwen.
It must be noted that when calculating value added on the basis of the total size of trade flows, there is always a danger of counting inputs on the first leg of a voyage (e.g. importing guns for the slave trade into Middelburg) as part of the outputs of the second leg (the value of Dutch exports to the West-African coast). Another perennial problem in such calculations on the basis of trade flows is that the total value of goods traded is based on observations on the value of these goods when they entered the Dutch Republic, not on the prices at the point of acquisition or sale.4 The robustness of the model of Van Zanden and Van Leeuwen thus in large part relies on the way in which they reconstruct net trade flows, that do not count the value of imports as part of the value of re-exports, and are calculated on the basis of local prices at every point.5 The difference between the two is a potential source of considerable confusion, since our most solid overall data remain those for gross trade flows, while the size of net trade flows can only be based on inferences from small amounts of observations. Luckily, the records of the Middelburgse Commercie Compagnie (MCC) allow us to test whether the conclusions of Van Zanden and Van Leeuwen hold for the most important segment of our calculations: the Atlantic trade. Before going into this, however, it is necessary to provide an overview of the (gross) size of slave-based trade, on the basis of the extant estimates for the size of the Dutch Republic’s international trade in the final decades of the eighteenth century.

Dutch Atlantic trade

Direct imports: Various estimates are available for the size of Atlantic imports and exports around the period investigated by us. All of those are ultimately derived from combining

4 For the difficulty that variations in regional, national and international prices present to economic historians, see: M. Morineau, Pour une histoire économique vraie (Lille 1985).
5 It must also be noted that Van Zanden and Van Leeuwen do not include so-called ‘voorbijlandvaart’, trade by Dutch merchants that never entered Dutch ports, as do all other available estimates for Dutch trade in this period.

incomplete trade records derived from tax records (customs and excises and paalgeld) with the famous 1785 memorandum of VOC director Cornelis van der Oudermeulen, who gave aggregate figures for an unspecified year ‘around 1780’ (in all likelihood a composite of observations from the period immediately before and after the Fourth Anglo-Dutch War) of the value of trade on all important trading routes.6 Various authors have used different methods to adjust these figures, leading to not always entirely compatible results. These are summarized in Table 1 below.

6 C. van der Oudermeulen, ‘Iets dat tot voordeel der deelgenooten van de Oost-Indische Compagnie en tot nut van ieder ingezeten van dit gemeenebest kan strekken’, in: G.K. van Hogendorp, Stukken raakende den tegenwoordigen toestand der Bataafsche beziṄngen in Oost-Indië (The Hague/ Delft 1801).

Table 1 Different estimates for the total size of Dutch trade, ca. 1770

VDO dV&vdW  ENT
c.1780 Neth.    1770s Neth. c.1780 Neth.

Region Value
(f mln) % Value
(f mln) % Value
(f mln) %
Europe:
Sont
Import 27 22
Export 17 17
Total 44 15.4 39 15.7
Europe:
Rest
Import 73
Export 55
Total 143.3 50.1 128 51.7 135.0b 44.2
River
trade
Import 10
Export 20
Total 34 11.9 30 12.1 62.7c 20.5
East
Indies
Import 20.5 20
Export 14.5 2
Total 35 12.2 22 8.9 37.9 12.4
Atlantic
Import 22 22.4
Export 7.5 6
Total 29.5 10.3 28.4a 11.4 70d 22.9
Total 285.8 100 247.4a 100 305.6 100

a De Vries made an upward readjustment of the earlier estimates for Atlantic trade made by him and Van der Woude, which is included here.
b Includes the Sont trade.
c Marked as ‘not specified’.
d The reason why Victor Enthoven’s figures for colonial trade are such outliers, is that he counts European imports and re-exports of colonial goods as part of the totals of Asiatic and Atlantic trade.
Sources: C. van der Oudermeulen, ‘Iets dat tot voordeel der deelgenooten van de Oost-Indische Compagnie en tot nut van ieder ingezeten van dit gemeenebest kan strekken’, in: G.K. van Hogendorp, Stukken raakende den tegenwoordigen toestand der Bataafsche bezittingen in Oost-Indië (The Hague/ Delft 1801) (excluding his figures for gold and coin exported to settle trade imbalances; J. de Vries and A. van der Woude, The first modern economy. Success, failure, and perseverance of the Dutch economy, 1500-1815 (Cambridge 1997) 499;
J. de Vries, ‘The Dutch Atlantic Economies’, in: P.A. Coclanis, The Atlantic economy during the seventeenth and eighteenth centuries. Organization, operation, practice, and personnel (Columbia 2005) 1-29, 28n63; V. Enthoven, ‘An assessment of Dutch Transatlantic commerce, 1585-1817’, in Johannes Postma and Victor Enthoven, Riches from Atlantic commerce. Dutch transatlantic trade and shipping, 1585-1817 (Leiden/ Boston

2003) 385-445, 444.

Table 2 presents various estimates for the size of direct Atlantic imports around 1770 and 1780.
Table 2 Direct imports from the Atlantic world around 1770 and around 1780 in fl. mln., different estimates

vdOudermeulen c. 1780   Van Stipriaan/ Klooster c. 1770 (partial estimates) De Vries, adjusted estimates c. 1770    Our estimate 1770

Suriname 8 7.9 12.6 8.3
Essequibo and
Demerara 4.5 1.7
Berbice 1
St. Eustatius 5.5 4.2 9.8 4.2
Curaçao 2.8 2.8
West Indies (non-
Dutch direct) 0.75 0.75
West Africa 0.3
North America 2.25 0.6
Total 22 14.9 22.4 18.65
Sources: Van der Oudermeulen, ‘Iets dat tot voordeel’; A. van Stipriaan, Surinaams contrast. Roofbouw en overleven in een Caraïbische plantagekolonie 1750-1863 (Leiden 1993) 437; W. Klooster, Illicit riches. Dutch trade in the Caribbean, 1648-1795 (Leiden 1998) 176.

We come to our estimates for the various branches of Atlantic trade in the following way:

• Suriname: Van Stipriaan’s figures for the value of Suriname export goods on the Amsterdam market are usually used as the basis for the overall development of Dutch imports from this region. However, these figures only include the value of sugar, coffee and cotton. While certainly by far the most important trade goods, there were some smaller trade goods that need to be taken into account. Most significantly, Van Stipriaan enlists yearly totals of 100,000 to almost 400,000 kilograms of cacao produced on Surinamese plantations in the 1770s. Including this and other minor trade goods, we find it justified to add another fl.400,000 to Van Stipriaan’s Suriname import figure, bringing the total to fl.8.3 million.

• Essequibo, Demerara and Berbice: De Vries and Van der Woude estimate that output from this region rose from 10 to 20 per cent of Surinamese output in 1700-1739 to over 50 per cent after 1780. According to Van der Oudermeulen’s estimation, the proportion in the latter year even was almost 70 per cent. Contemporary records and later estimates for the amounts of sugar and coffee imported from these Guyana Coast colonies, presented in the main text of the article in Table 4 and 5, suggest that coffee and sugar output only started to approximate the 50 per cent of Suriname output estimated by De Vries and Van der Woude in the second half of the 1770s, and stood much lower before that time. Preferring to err on the side of caution, we estimate the size of direct imports from Essequibo, Demerary and Berbice for our year at fl.1.7 million, 20 per cent of Suriname imports.
• St. Eustatius and Curaçao: We have no reason to diverge from the yearly estimate of Klooster of fl.4.2 million and fl.2.8 million respectively.
• West Indies (non-Dutch direct): this includes direct trade from St. Domingue and the British and Danish West Indies. We assumed that this trade was relatively stable, maintaining Van der Oudermeulen’s estimate of fl.750,000.
• West Africa: Van der Oudermeulen only gives the estimation that a total of fl.2 million was tied up in the slave trade and West African trade, without giving further specifications. . It must be assumed that the largest part of this consisted of goods exported for the slave trade plus profits made by Dutch merchants on the sale of slaves in the Americas (which Van der Oudermeulen might have overestimated, but we will get back to this later in this section). Direct exports from West-Africa to the Dutch Republic by the 1770s were not large. Henk den Heijer shows that the WIC earned fl.131,920 in the sale of African goods in the years 1768-1770, and fl.257,450 in 1771-1773.7 Generously accounting for the increased role of private trade, we set the total of direct imports from West Africa at fl.300,000, with ivory and slave- produced Brazilian gold the most valuable components of this trade.
• North America: Van der Oudermeulen estimated imports in 1784 at fl.2.25 million, with 52 ships clearing in Amsterdam and 13 in Rotterdam.8 Based on the paalgeld records, Welling estimates the value of North-American imports in Amsterdam amounted to fl.323,000 in 1771, and fl.1,191,025 in 1784.9 Given the increase in Dutch-North American trade as a result of the War of Independence, we find it likely that the North American trade indeed was considerably smaller in 1770, putting the total direct imports from the North-American colonies at this period at fl.600,000. It must be noted that for 1771, Welling shows that over 80 per cent of North-American imports came from Charleston (Virginia), making it likely that most North-American imports were slave-produced.
• Of the fl.18.65 million in direct Dutch imports from the Atlantic world, only the imports from West Africa and North America may have contained meaningful proportions of goods that were not slave-produced. Even when assuming that half of the trade with West Africa and North America was not related to slavery or the slave trade (almost certainly a gross over- estimation), we can set the total value of slave-based direct imports from the Atlantic world at fl.18.2 million.

Direct exports: De Vries and Van der Woude estimate the total value of exports to the Atlantic world at fl.6 million. They themselves acknowledge that this is a very rough figure, based only on estimated exports ‘to the colonies’ (presumably the Dutch colonies on the Guyana coast) plus Van der Oudermeulen’s estimate of total exports of fl.3.5 million to to St. Eustatius and Curaçao.10 A much older estimate by De Hullu put the export to the Dutch Atlantic colonies around 1800 at fl.7.4 million.11 Excluding exports to North-America and non-Dutch Atlantic colonies, but including the slave trade, Van der Oudermeulen put the direct exports from the Netherlands to the Atlantic world at fl.7.5 million, an amount that seems more or less consistent with De Hullu’s figure. Considering the figures on which De Vries and Van der Woude based their conservative estimate, it seems fair to say that their export figure actually does not reflect Atlantic exports in total, but only those exports that directly supported the Dutch Atlantic colonies, and through St. Eustatius and Curaçao to the other slave colonies in the West-Indies. We therefore take fl.6 million as the total value of slave-related exports to the Atlantic world, except for the direct export for the slave-trade, which we will now treat separately.
α. Slave trade: As noted above under direct exports, De Vries and Van der Woude’s estimates for the Atlantic trade are lower than Van der Oudermeulen’s, primarily because they seem to not include the slave trade. Indeed, Van der Oudermeulen estimated the export to the Dutch Atlantic colonies at fl.6 million and the African slave trade at fl.1.5 million.12 How accurate was Van der Oudermeulen’s estimate? According to the most recent reconstruction of revenues from slave-sales by Dutch traders, the total value of the slaves imported to the Dutch Guyanas and Caribbean in 1780 amounted to fl.1,226,234, the composition of which was:

• 1,329 slaves disembarked in the Dutch Caribbean, sold at an average price of fl.312;
• 2,267 slaves disembarked in the Dutch Guyanas, sold at an average price of fl.358.13

Including the direct exports of trade goods from Africa to the Dutch Republic, which Van der Oudermeulen included under the same bracket as the slave trade, indeed gives a total only slightly above fl.1.5 million. However, the value of exports related to the slave trade of course should be measured not by looking at the prices at which the slaves were sold in Suriname, but on the basis of the prices that Dutch merchants paid for their human cargoes on the African coast. For 1780, this amounted to a sum of fl.580,608 (accounting for 4,032 embarked slaves and average slave prices of fl.144). In 1770, it amounted to the much larger sum of fl.917,595 (6,797 embarked slaves at an average price of fl.135). These exports for the African slave trade should be added to the fl.6 million of direct exports to the Americas under ‘direct Atlantic exports’, bringing the total of direct Atlantic exports to fl.6.9 million.

Indirect slave-based imports and (re-)exports

Indirect Atlantic imports: The Dutch not only imported slave-produced commodities from their own colonies, but also imported large quantities of coffee and sugar from rival Atlantic empires, especially the French. Part of this trade came in through legal and illegal channels via the West-Indies, in particular through Curaçao and St. Eustatius. In addition, the Dutch
12 It is possible that De Vries and Van der Woude assumed that Van der Oudermeulen included the importation of slaves in the Dutch colonies in his figure of 6 million, as part of the triangular trade. This, however, is not the case. Van der Oudermeulen, ‘Iets dat tot voordeel’, 333.
13 Figures courtesy of Matthias van Rossum and Karwan Fatah-Black.

imported a substantial amount of colonial goods from French ports. The two routes were in inverse proportion, so that in years in which much coffee and sugar from St. Domingue was imported through St. Eustatius, less was imported directly from Bordeaux, and vice versa. In 1773, Dutch ships represented 29 per cent of the total foreign tonnage in the trade from Bordeaux.14 On the basis of the French Trade Records Database TOFLIT, it can be estimated that the Dutch imported fl.5 million worth of sugar and fl.3 million worth of coffee through French ports.15 Comparing to the immense European imports from French slave-plantations, imports through other European countries in 1770 (a down year for this side of the trade) was relatively small and throughout this period remained extremely volatile. For 1770 the Portuguese, English and Danish trade in these plantation goods added only fl.0.3 million for sugar, and fl.0.1 million for coffee. Sugar and coffee formed the most valuable goods produced by slaves of other Atlantic Empires imported through European ports. Other goods included gold produced in Brazilian mines that was imported through Lisbon, indigo produced in the Caribbean imported through French, British and Danish ports, and tobacco that was partly produced by slaves in North America and was imported in large quantities from Britain. The value of these imports was far from negligible. Using British export data, Roessingh has estimated that between 1770 and 1774, the Dutch imported an annual average of 86 million pounds of Virginia and Maryland tobacco from England and Scotland, at a price of around fl.0.21 per pound, totalling over fl.18 million in just five years.16 Dutch custom records give much lower import figures for tobacco, perhaps indicating large scale smuggling of this commodity. According to Welling’s paalgeld database, Amsterdam imported well over fl.400,000 worth of tobacco from the British North-American colonies and Brazil in the 1770s.17 However, the most important port of entry for tobacco was Rotterdam, not Amsterdam. French trade records show that Dutch merchants imported almost fl.500,000 in
14 S. Marzagalli, ‘The French Atlantic and the Dutch, late seventeenth-late eighteenth century’, in: G. Oostindie and J.V. Roitman (eds.), Dutch Atlantic connections, 1680-1800. Linking empires, bridging borders (Leiden/ Boston 2014) 103-118, 113.
15 French trade records database, TOFLIT 18, http://toflit18.medialab.sciences-po.fr/#/home (accessed 31-05- 2019). Further estimates on the basis of these data courtesy of Tamira Combrink.

16 H.K.Roessingh, Inlandse tabak. Expansie en contractie van een handelsgewas in de 17e en 18e eeuw in Nederland (Zutphen 1976) 339. For the 1770s, it is not controversial to assume that the tobacco imported from the British colonies in North-America was produced by slaves. After 1700, slave labour rapidly replaced indentured servants up to the level of the supervisory and skilled jobs on the tobacco plantations of the Chesapeake. See: R.M. Menard, Migrants, servants and slaves. Unfree labor in colonial British America (Aldershot 2001) IV, 51. Tobacco and slavery were inexorably linked in Virginia already in the early eighteenth century the slave populations were concentrated in the tobacco belts. See: idem, III, 383-384. Maryland may have followed a little bit later. A.L. Kulikoff, Tobacco and slaves. Population, economy and society in
eighteenth-century prince George’s county, Maryland, dissertation Brandeis University 1976, 169, 171.
17 Welling, Prize of neutrality.

indigo in 1770.18 To be on the safe side, we estimate that taken together, these ‘lesser slave- based trade goods’ valued fl.2 million. As the British export figures for tobacco mentioned above make clear, this must be a considerable underestimation.
In addition, part of the fl.6 million worth of goods exported by the Dutch Republic to its Atlantic colonies for the upkeep of plantations (excluding goods for the slave trade, that will be treated separately below under β)were in turn imported from other European countries. De Vries and Van der Woude estimate that half of these exports were re-exports of goods procured in other parts of Europe. However, basing ourselves on estimates for the proportion of imported goods in the cargoes of Middelburg slave traders (see under subsection 3), we think it is more likely that the part of goods produced domestically formed only about one quarter of the total. Of course, the three quarters of the re-exported goods were not bought at the same prices at which they were re-exported. Allowing for a very generous discounting of 33 per cent, we estimate a value of fl.3 million in goods for export to the Atlantic colonies was imported from elsewhere in Europe.
Re-export of Atlantic slave-produced goods: Most slave-based imports into the Dutch Republic were not intended for domestic consumption, but were re-exported to Northern and Central Europe. Tamira Combrink estimates the total value of sugar exports from the Dutch Republic at fl.10.1 million for the year 1770. Practically all of this (98 per cent) was sugar from Atlantic slave plantations that was either re-exported as raw sugar or refined in the Dutch Republic first. Clearly, with these proportions and no alternative source of sugar at hand, dependency on Atlantic plantation production was almost complete, leading us to count fl.9.9 million as slave-based.19 The total value of coffee exports was fl.9.3 million. As in sugar, by far the largest proportion came from the Atlantic colonies, leaving a total of fl.7.5 million in slave-based coffee exports.20 Once again, the re-export of slave-based commodities of lesser importance in the overall trade needs to be taken into account, especially tobacco and indigo. Unfortunately, figures to base this estimate on at this time are not available, but
18 French trade records database, TOFLIT 18, http://toflit18.medialab.sciences-po.fr/#/home (accessed 31-05- 2019).
19 The estimate is based on custom records for 1753 and 1790, and then interpolated on the basis of the trade volumes, with a correction factor based on the development of the Rhine trade. The corresponding values are estimated on the basis of Posthumus prices in Amsterdam. These prices are specific to kind and quality, making it necessary to estimate the mixture of types of sugar exported, which has been done by Tamira Combrink on the basis of qualitative and quantitative sources available and calculations of weight loss in sugar refining for the different types, the estimates of which will be presented in more detail in her dissertation.
20 Most imported coffee that was not produced by slaves in the Atlantic region was produced on Java, and might actually have involved significant proportions of slave labour. However, we do not examine revenues from forced labour in the East Indies as part of our total in this article.

estimating the total value of this trade at fl.1 million will put us once again on the conservative side. On this basis, we set the total value of slave-based European re-exports at fl.18.4 million.
β. Indirect imports and exports for the slave trade: Apart from European imports of slave- produced commodities, re-exports from the Netherlands of slave-produced commodities, and European imports aimed for re-export to the Dutch Atlantic colonies, the Dutch Republic also imported goods from various parts of Europe destined for its own slave trade (Swedish iron, Silesian linen, fire arms from the Southern Netherlands, wine from France, etc.), and exported domestic products used in the slave trade of other nations (cheese, gin, etc.). The former is relatively easy to estimate. We have already seen that the export of goods to West- Africa bound up with the Dutch slave trade amounted to fl.900,000, and have estimated that approximately three quarter of these cargoes were brought into the Dutch Republic from other parts of Europe, which has to be discounted by one third to account for lower prices of acquisition, making a total of fl.445,500. It is impossible to reconstruct at this point how much of the cheese the Dutch exported to France, or the gin exported to England was subsequently used in the slave trade, but it seems reasonable to estimate this at no more than fl.500,000.

Trade to different regions in Europe

One final preparatory step before we can apply the model of Van Zanden and Van Leeuwen to these overall estimates for the size of the different components of slave-based trade, is to differentiate within the European side of this trade between three large streams: trade through the Sound with Northern and North-Eastern Europe; the river trade, mainly directed to the German hinterland via the Rhine; and the trade with the rest of Europe over sea. The reason why we apply these very large container categories, is that Van Zanden and Van Leeuwen calculated the contribution of trade and shipping to GDP on the basis of weighted measures for these specific routes. We have subdivided the different European streams of goods under the following, slightly simplifying assumptions:
a) Imports of goods produced in other European Atlantic Empires through European ports have all been headed under the rubric ‘rest of Europe’. Only a tiny proportion of goods imported from Danish ports might be mislabeled by doing so.

b) Half the goods imported from Europe for export to the Dutch Atlantic colonies have been assigned to the Sound, the other half to the rubric ‘rest of Europe’. The same has been done for European goods used in the slave trade.
c) For the division of re-exports over the three different branches of European trade, we could use export data compiled by Gerhard de Kok on the basis of custom records for Amsterdam and Rotterdam in the years 1753 and 1790. Taking the average between the two to represent exports in 1770, we assume the following proportions:
Sugar: Rhine 71%, Sound 27%, Rest 2%

Coffee: Rhine 61%, Sound 37%, Rest 2%

Tobacco: Rhine 24%, Sound 52%, Rest 24%

Other lesser trade goods have been subdivided following the same proportions as tobacco.

d) All exports of domestic goods for the slave trade of other nations have been estimated to have gone to the ‘rest of Europe’.

Table 3 summarizes the results of our calculations thus far.

Table 3 Total value of slave-based commodity trade Dutch Republic in millions of guilders, 1770

Import  Export  Total Import +

Export
Total direct
Atlantic 18.2 6.9 25.1
Sound 1.7 6.0 7.7
Rhine – 11.8 11.8
Rest of Europe 12.1 0.6 12.7
Total indirect
European 13.8 18.4 32.2
Total import +
export 32.0 25.3 57.3

From total trade flows to value added for shipping and international trade

One way to calculate the contribution made by this large sum of fl.57.3 million to Holland’s GDP would be to estimate the proportion of this trade that was based in Holland as opposed to the rest of the Dutch Republic, calculate value added on the basis of a value-chain approach for all the different trade goods involved subtracting inputs and outputs to avoid double counting or counting non-domestic activities, and relate the total that is thus acquired to the total GDP of Holland of fl.177 million estimated by Van Zanden and Van Leeuwen.
However, as explained extensively in the main text of this article, in our eyes this procedure would be a partial retreat back into a naïve approach. Since the method of Van Zanden and Van Leeuwen relies on using benchmark figures to adjust sectoral proportions, related to a total figure for the GDP that is interpolated from known figures for 1504-1510 and 1807, it would create serious distortions to arbitrarily use different figures for one segment of Holland’s trade without recalculating all other parts. Therefore, we believe that following the methodological assumptions of Van Zanden and Van Leeuwen and relating our own figures to their benchmarks following their methodological assumptions will produce the most reliable estimates. To relate our own 1770 estimates for the total size of Dutch slave-based trade flows to Van Zanden and Van Leeuwen’s total value added in the trading sector, we have used the following procedure:

Step 1

Based on our calculations on the size of slave-based trade to and from different regions (summarized in Table 3), we calculated the proportion of slave-based trade of the total trade with the different trading areas distinguished in Van Zanden and Van Leeuwen’s model (Atlantic, Sound, Rhine, Rest of Europe). To calculate which proportions of the total trade for these routes were slave-based, we have used De Vries and Van der Woude’s trade estimates for the 1770s shown in Table 1, except for the Atlantic trade which we have re-estimated above for the year 1770 specifically. Doing so gives us the following proportions (p) of slave- based imports and exports on the four relevant routes: pATL = 0.984, pSND = 0.197, pREST = 0.099, pRIV = 0.393.

Step 2

We used these proportions (p) to relate our estimates of slave-based activities per trading route to the total value of trade on the different trading routes that forms the basis for Van Zanden and Van Leeuwen’s GDP calculation. By taking this second step, we make sure that we follow the same assumptions that were followed in this model when isolating the output of the trading and shipping sector from the total value of trade streams calculated on the basis of end prices of goods on the Dutch market (which still contain the value added in the regions in which these goods were produced). The total size of the different net trade-flows TF for the Dutch Republic used in their model, subtracting prior inputs, is as follows:
TFATL (Atlantic trade): fl.20.85 million

TFSND (Sound trade): fl.32.14 million

TFREST (Europe rest): fl.37.41 million

TFRIV (River trade): fl.43.08 million

The size of the slave-based trade flow, when related back to the figures of Van Zanden and Van Leeuwen, is given by the formula p . TF.
We are well aware these totals depend on many assumptions, quite a few of which are based on calculating steps that remain implicit in the published data series and methodological annex that accompany Van Zanden and Van Leeuwen’s original article.
Starting from the various extant estimates of the total size of trade measured in Amsterdam bourse prices presented in Table 1, it is clear that Van Zanden and Van Leeuwen assigned smaller weight to the trade for their category ‘Rest of Europe’ than the gross size of this trade flow would seem to suggest, and must have started from a far larger gross size of the river trade than is usually assumed.21 On the other hand, their adjustment factors to subtract inputs from the gross trade flow for the Sound trade are slightly below what we would expect to see, although admittedly these expectations are largely based on price differentials between wholesale and retail trade within the Netherlands and on import and export prices in Amsterdam, rather than prices at which merchant bought and sold commodities abroad used by Van Zanden and Van Leeuwen. Furthermore, Van Zanden and Van Leeuwen did not
21 Indeed, recent data on the growth of the trade between the Dutch Republic and the German hinterland show such an adjustment to be necessary. U. Pfister, ‘The quantitative development of Germany’s international trade during the eighteenth and early nineteenth centuries’, Revue de l’OFCE 140:4 (2015) 175- 221.

include any contribution to GDP from the substantial shipping through the river Rhine, instead calculating the contribution of international shipping only on the basis of the sea trade only, but the relatively high proportion of value added in international trade on this route might compensate for this. By and large, we accept these assumptions as necessary parts of the model, especially since the potential overestimations and underestimations that flow from it by and large cancel each other out. Under step 4, we examine the question of potential underestimations and overestimations stemming from the application of this model more closely.

Step 3

The net trade flows form the basis on which Van Zanden and Van Leeuwen calculated the value added in international trade and shipping. They did so on the basis of specific selections of trade goods, that of course do not correspond neatly to the mixture of commodities that we have studied. We therefore have chosen to reconstruct composite weights for each route, working backwards from Van Zanden and Van Leeuwen’s totals while making some minor adjustments for irregularities in the data. The resulting weights are presented as proportions of the net trade flows TF. Finally, all calculations so far have been done on the basis of figures for the trade flows to and from the Dutch Republic as a whole. Van Zanden and Van Leeuwen in their model have used the admittedly very rough measure of an unchanging share of Holland in Dutch trade for each of the different routes, that we express here with the Greek letter γ.

Table 4 Weighing factors used by Van Zanden and Van Leeuwen

Value added in trade as proportion of TF, assumptions vZ

and vL (X) Value added in shipping as proportion of TF (Y) Share of Holland in trade of Dutch Republic (γ)
Atlantic 0.372 0.138 0.78
Sound 0.301 0.059 0.95
Europe: Rest 0.301 0.055 0.75
River trade 0.150 0 0.8

For each route, the value added in international trade will thus be given by the simple formula

p . TF . X . γ, and the value added in international shipping by the simple formula p . TF .Y . γ. On the basis of the figures given above, we estimate the contribution of slave-based international trade to the GDP of Holland to be:
Atlantic trade: fl.5.95 million; Sound trade: fl.1.81 million; Rest of Europe: fl.0.83 million; Rhine trade: fl.2.03 million;
Bringing the total value added of slave-based international trade to fl.10,62 million.

Following the same procedure, we estimate the value added from slave-based international shipping to be:
Atlantic trade: fl.2.21 million; Sound trade: fl.0.35 million;
Rhine trade: not included in the model (subsumed under international shipping) Rest of Europe: fl.0.15 million,
Bringing the total contribution of slave-based international shipping to fl.2.71 million.

Step 4

Step 1 to 3 give us the total value added in international shipping and trade following the assumptions of the model of Van Zanden and Van Leeuwen, and using composite adjustment factors to adapt gross trade flows to net trade flows per trade route. However, our starting point for these calculations were source-based estimates for the total value of slave-based imports and exports. Our final step is to go back to these original numbers, and verify whether implementing the various weighting factors actually leads to credible results. In

order to do so, we pretended that the value added in trade and shipping was derived directly from the total value of trade, without the intermediary steps dictated by the model.
Calculating the value added in both branches as proportions of the gross trade flows gives us numbers that can be meaningfully compared to what the sources tell us about the profitability of international trade and shipping in this period. Table 5 gives these proportions (note that in order to compare to the gross value of trade for the Dutch Republic as a whole, the figures given in step 3 have to be divided by γ).

Table 5 Value added in international trade and shipping expressed as percentage of gross trade flows on each trade route

Total value of the trade

(million fl.) Value added in int. trade
(million fl.) % Value added in shipping
(million fl.) %
Atlantic 25.1 7.6 30 2.8 11
Sound 7.7 1.9 25 0.4 5
Europe: Rest 12.7 1.1 9 0.2 2
River trade 11.8 2.5 22 0 0
Total 57.3 13.1 23 3.4 6

It is obvious from Table 5 that the question whether this model gives a realistic perspective on value added in slave-based trade and shipping stands or falls with the proportions for Atlantic trade. Fortunately, the robustness of the model for this branch can be tested with an unusual amount of precision, due to the availability of the very detailed accounts for a very large number of voyages in the MCC records. Gerhard de Kok has reconstructed the accounts for all Atlantic voyages of the MCC, presented in Annex B of his dissertation.22 The completeness of these records, based on over hundred voyages between 1730-1795, allows us to do something that is very rare indeed: test our outcome for the most important component of our GDP calculation on a contemporary source, using exactly the same measures that we applied in our model. The results are presented in Table 6. The value of outgoing goods (exports for the slave trade) and incoming goods (imports from the West-African coast and West-Indian plantation goods) are both measured in current prices paid for those goods in the Dutch Republic itself, just as in our overall estimates for the size of the various trade flows.

Value added per voyage is represented here by net profits, sailors’ wages and bonuses, the contribution paid by the MCC to Dutch West India Company (WIC) (a redistribution of value added between commercial companies), and insurance fees minus damages claimed for the profits drawn from these voyages by insurance companies.23 Since the MCC was both trading company and shipping company at the same time, the value added in trade and shipping is taken together.

23 Van Zanden and Van Leeuwen do not include a separate post for the profits of insurers in their GDP model, thereby assuming this was part of the value added in trade. We are aware that insurers made some domestic costs apart from the sums paid in damages, but on our totals, these sums are negligible.

The period that includes the year 1770 show a percentage of value added of 41.5, slightly above the number inferred by us for the Atlantic trade in Table 5. In fact, the five MCC voyages for the single year 1770 were all highly profitable. Basing ourselves on that year alone, we would get a value added of 56 per cent of the gross trade flow. This result is all the more notable, as the MCC was primarily a slave-trading company, with trade in slave- produced goods always coming second. The Dutch literature, as much of the international literature, generally assumes the slave trade to be one of large windfalls and even larger losses, a reality reflected in the accounts presented above in the volatility of profits and losses. The bilateral trade in Atlantic goods, while contributing less to the value added in shipping, is generally assumed to have been much more stable and profitable. This bilateral trade formed the majority of Dutch Atlantic voyages. About 80 per cent of Dutch ship traffic with Suriname in the 1770s was bilateral trade (an annual average of 59.6 bilateral voyages, against 14.6 triangular voyages).24 Therefore, the proportion of value added by the MCC in shipping and trade can be taken as strong confirmation of the model applied by us.
What about the other trade routes? We have already made a remark under step 2 about the relatively low value added captured in comparison to the size of the gross trade flow in the trade with the category that Van Zanden and Van Leeuwen have named ‘rest of Europe’ (including France, Britain and the Mediterranean), and the relatively high proportions in the Sound trade. Adjusting the proportion of the value added in the Sound trade downward to 15 per cent, while readjusting the value added in the trade with the rest of Europe upward by a similar proportion, also to 15 per cent, would lead to a net result just above the total given by our model (an upward adjustment of fl.150,000). We therefore stick to the conclusion that the potential over- and underestimations for these two trade flows cancel each other out, and decide not to make any adjustments. Finally, for the Rhine trade, we have seen that Van Zanden and Van Leeuwen apply a proportion of 15 per cent value added in trade and shipping for this route. However, since they assume a TFRIV that is substantially larger than the total inferred from the existing estimations, the inferred proportion as calculated in Table 5 comes out higher than this 15 per cent. Since we have started our calculations not from projections of the size of the trade on this route based on a model, but from the actual size of the trade that can be deferred from export figures for sugar, coffee and tobacco, there is in principle no reason for such an upward adjustment. We therefore would deem it more responsible to stick to the proportion of 15 per cent value added in international trade. On the other hand, not calculating any value added for shipping on the Rhine trade presents quite a substantial under-estimation. Once again, we can assume that these two cancel each other out.