SOURCES

To describe the market developments, we start with an overview of negotiaties provided by Van de Voort (Reference Van de Voort1973), with names of bankers, size of the fund, date of initiation of the fund, and coupon. We adjust and correct this data using prospectuses and other deeds from the Amsterdam notary records and information from MBS auction records.Footnote ² We complement this overview with deed records of plantation mortgages from Surinam and Amsterdam notary records.Footnote ³ These deeds provide the date of the mortgage, amount, maturity, interest rate, appraised value of the plantation, and whether it produced sugar or coffee. We calculate the loan-to-value ratio as the mortgage amount over appraised value.

In addition, we collect all protests of dishonored bills of exchange recorded at five Amsterdam notaries who were active in business in Surinam over the period in which we expect the crisis to start, that is, 1770–1774. There are 3,175 such protests, and we consider the amount of the bill and the period between drawing the bill and its presentation at the bank.Footnote ⁴

Our analysis of the development of prices of negotiaties is based on auction prices between 1768 and 1796. These cover all voluntary sales of securities administered by the City of Amsterdam, with information on the date, security characteristics, and price. Our sample has 5,759 transactions for 137 unique securities. We use the periodical Maandelijkse Nederlandse Mercurius for transaction overviews and the original auction records for detailed characteristics to link each security to the negotiatie and the attributes provided by Van de Voort (Reference Van de Voort1973).Footnote ⁵ From 1773 on, around 200 bonds were traded each year. Van Bochove, Boerner, and Quint (Reference Van Bochove, Boerner and Quint2017, p. 4) argue that the auctions were sufficiently liquid and transparent for prices to be informative about the underlying values.Footnote ⁶

For a detailed analysis of developments, we turn to Ferrand Whaley Hudig. Based in Rotterdam, Hudig started his business in 1756, at the age of 21. Hudig was a descendant of a family of merchants and was by blood related to both partners of the prominent Rotterdam merchant house Coopstad and Rochussen, which was involved in the Caribbean and slave trade and the financing of plantations. His family fulfilled important political positions (Hudig Reference Hudig1933; De Groot-Teunissen Reference De Groot-Teunissen2005). In 1759, Hudig started to finance plantations with MBS, initially for two plantations, later expanding to 12. He had minor interests in Surinam shipping and insurance.Footnote ⁷

Hudig is the only merchant-banker active in the plantation negotiatie business whose business archives have completely survived, including his correspondence, contracts, transactions, payments, and detailed overviews of the development of specific plantations and negotiaties.Footnote ⁸ The key sources for our analysis are: (i) the incoming and outgoing correspondence with, among others, his agents and plantation owners; (ii) financial accounts and payments of each plantation; and (iii) legal documents, such as negotiatie contracts, appraisal reports, mortgage contracts, and plantation sale transactions.

In terms of total negotiatie volume, Hudig was in the 19th percentile of about 90 banks active in plantation financing before 1780. There were a limited number of merchant banks that set up large negotiatie funds, such that Hudig’s total negotiatie position amounts to 2 percent of the market. Hudig is more representative of the majority of (smaller) merchant-banks that set up negotiatie funds backed by one or two plantations (Van de Voort Reference Van de Voort1973, pp. 269–323). We discuss Hudig’s representativeness later.Footnote ⁹

FINANCING AND INTERMEDIATION IN THE DUTCH ATLANTIC TRADE

Dutch slave-plantations were mainly located in Surinam, but also in present-day Guyana, primarily producing sugar and coffee. Van Stipriaan (Reference Alex1993) provides a detailed account of the Surinam plantation colony and estimates that Surinam produced about 3 percent of all Caribbean sugar and 14 percent of all coffee in 1775.

There were differences in investments and risks between growing sugar and coffee. To produce sugar, planters needed milling and cooking houses, which meant the minimum feasible scale was larger than for coffee. Growing sugar cane itself was relatively cheap, because among other reasons, canes can already be harvested 16 months after being first planted and normally last for 10 to 15 years. In contrast, coffee trees only start to yield beans after five to six years, and the grounds are exhausted much later. Thus, because of the production costs, the longer time horizon, and the volatility in prices, coffee was a riskier crop than sugar. Nevertheless, the minimum feasible scale was smaller, and entry was easier.

Figure 1 provides an overview of prices and production in Surinam. Coffee prices rose substantially in the second half of the 1760s, which appears to have stimulated investment in coffee plantations (Van Stipriaan Reference Alex1993). Coffee prices fell again after 1770. Sugar prices were more stable. Both commodities saw price increases after 1776 with the American War of Independence and subsequent wars between England, France, and the Dutch Republic. After falling back to peacetime levels in 1783, both commodities saw price increases, especially during the French revolutionary wars of the 1790s. Production in Surinam was stationary during the period, with substantial temporary drops due to war hostilities in 1781 and 1793.

Figure 1 COMMODITY PRODUCTION AND PRICES

Note: Data is available up to 1794.

Source: Van Stipriaan (Reference Alex1993).

The main production factor on plantations were enslaved people, transported to Surinam from Africa (Van Stipriaan Reference Alex1993, p. 100). Enslaved labor was not only responsible for the production of the crops, but also for the necessary infrastructure, such as embankments, irrigation, and drainage systems. The planters’ legal property rights over the enslaved turned them into financial collateral, which helped to fund the expansion of the plantation.

The plantations were largely funded with mortgages provided by bankers from cities such as Amsterdam, Middelburg, and Rotterdam. Mortgages were securitized and sold on to investors in the form of negotiatie funds in certificates of 1,000 guilders. The first negotiatie was initiated by Willem Gideon Deutz in 1754, who initially raised more than one million guilders (Van de Voort Reference Van de Voort1973). Later funds followed this structure closely, leading to highly standardized financial instruments in the 1760s. Investors could get a diversified portfolio of plantation MBS. Moreover, many negotiaties involved a diversified pool of multiple mortgages. Investors were promised a fixed annual interest rate of 5 to 6 percent (substantially above the government yield at the time of about 2.5 percent), while the bankers held an account for each plantation-owner, which allowed payments to investors to be smoothed over time. The maturity period was normally 20 years. Interest payments were due annually from the start of the contract, but amortization only started after ten years. In practice, amortization was rare, and many mortgages were rolled over. A board of commissioners represented investors in an MBS. These appear to have only started playing a role when the negotiatie got into trouble.

The bankers kept complete oversight of all cash and information flows. As in Diamond’s (1984) classical banking model, they played a key role in reducing information and agency issues. They marketed the plantations’ products, supplied building materials, insurance, etc., and provided short-term credit so the planters could purchase local goods. The choice to securitize through negotiatie funds, rather than keeping mortgages on their balance sheets, was presumably driven by the small size of banks and the lack of deposit banking. The securitization allowed merchant-bankers to reduce their on-balance sheet exposure. As in DeMarzo’s (2005) seminal model of security securitization, this allowed them to maximize the profits from their expertise without being capital-constrained. However, securitization likely weakened intermediaries’ incentives to properly screen and monitor (Keys et al. Reference Keys, Mukherjee, Seru and Vig2010). Bank reputation functioned as a substitute for skin-in-the-game, which, as we will show for Hudig, was limited (for a theoretical treatment, see Hartman-Glaser, Piskorski, and Tchistyi (Reference Hartman-Glaser, Piskorski and Tchistyi2012)).

Starting in the early 1750s, the Amsterdam market was flooded with capital as the Dutch government started to repay substantial debts, and interest rates were low. The economically most important Dutch province, Holland, reduced its debt from 360 to 320 million guilders between 1750 and 1780 (De Vries and Van der Woude 1997). Dutch financial markets had been familiar with secured bonds since the introduction of so-called quicksilver bonds in 1695, which were backed by the revenues of an Austrian quicksilver mine (Elias Reference Elias1905, part 2, pp. 1047–49).Footnote ¹⁰

Figure 2 illustrates a negotiatie’s structure. Mortgages originated through the bankers’ agents. In Surinam, many of these so-called agendarissen were active, often in associations. These men would screen suitable candidates and then send paperwork to the bankers in the Dutch Republic. They played a crucial role in limiting the agency problems induced by the long distance between the investors and bankers in the Dutch Republic, and the plantations in the colonies. The agents would typically receive a fixed annual payment and a percentage of the new mortgages, which gave them an incentive to overextend credit. Many agents were plantation owners themselves and often worked for multiple banks.

Figure 2 FLOW CHART NEGOTIATIE PROCESS

Source: Authors’ illustration.

Bankers relied on appraisers (priseurs) appointed by the local government to determine the value of a plantation based on market prices of land, enslaved workers, etc. The local government set rules for the appraisal process. Mortgages could cover up to a maximum of 5/8th of the appraised value. In theory, the equity stake of the planter (at least 3/8th of the appraised value) provided a sufficient buffer if the plantation was liquidated. Also, the standardization of lending standards might have been a way for bankers to signal quality, although an over-reliance on such signals could have stimulated fraud. At multiple points in time, usually in response to a credit boom, rules were tightened after financiers complained about lax appraisal standards (Wolbers 1861, pp. 234, 261–2, 303–4; Hudig Reference Hudig1933, p. 36).

Bills of exchange were important, but unlike the English system, they were primarily a means of payment rather than a credit instrument. Plantations commonly used bills to access their negotiatie mortgage credit, or the “overdraft facility” attached to that mortgage. Most bills would be sent to the Dutch Republic immediately to be presented for payment at the merchant-bank. Typically, the banker’s local agent had to approve the bill for the bank to accept it. In case of refusal, the holder of the bill would file a formal protest with a notary and send the bill back to the colony. Refusal was costly, and bankers tried to avoid it. The plantation would, apart from reputation damage, have to pay a fine of 25 percent of the bill’s value, which hurt the plantation, the banker, and the negotiatie investors.

We collect information on 3,175 dishonored bills from 1770 onward. We use this data to investigate the maturity distribution of bills of exchange (see Figure A.1 of the Online Appendix).Footnote ¹¹ The distribution of the time between the issuance of the bill in Surinam and the moment it was first presented at the merchant-bank in the Dutch Republic shows a spike at three to four months, corresponding to the usual travel time between Surinam and the Dutch Republic. The median, average, and 75th percentile are five, seven, and eight months, respectively. This means that the majority of bills were paid well within 12 months, which is considerably faster than was common in the British West Indies (Radburn Reference Radburn2015).Footnote ¹²

The management and supervision of the plantations were in the hands of administrators if the owners of the plantations did not reside in the colony. By custom, they received 10 percent of the plantations’ revenues. Administrators and bankers’ agents had separate roles. However, the two were often the same if the plantation had a mortgage. At least in theory, this improved oversight and minimized information and incentive problems. Administrators often oversaw multiple plantations, often in associations. They resided in Paramaribo, the colony’s administrative center, and visited plantations if needed. Day-to-day affairs on the plantations were supervised by plantation directors, who typically received a flat salary.

The long development times of plantations required a long-term commitment of funds, either in the form of long-term debt or outside equity funding.Footnote ¹³ Theory suggests that due to agency costs, the latter would have been costly (Jensen and Meckling Reference Jensen and Meckling1976). If outside equity owners are unable to (perfectly) monitor, insiders always have an incentive to report that affairs are in a bad state and thereby embezzle plantation revenues. In this scenario, debt contracts can be optimal, even if they limit flexibility (Townsend Reference Townsend1979). In the eighteenth-century Atlantic trade, information problems were severe, and financiers could not verify the true state a capital recipient was in. The fixed interest schemes of debt contracts (under the threat of liquidation) help enforce the truthful verification of the plantation’s state of affairs.

Table 1 and Figure 3 present the development of the negotiatie market using a new combination of sources. Panel A of Table 1 provides information about the different negotiaties. In the early years until 1765, there were 15 negotiaties, raising a sum of 8.9 million guilders. The median size is 106,000 guilders, and the mean coupon for investors is 5.4 percent (with a median of 6 percent). Panel B provides information about the underlying mortgages. On average, mortgage size was 34,721 guilders, while the appraised values were much higher. The median loan-to-value (LTV) ratio is 44 percent, which is well below the maximum in the funds’ prospectuses of 5/8th (i.e., 62.5 percent). The median interest paid is 6 percent (with an average of 5.78 percent). The markup between the MBS coupon and mortgage interest rates is small. The bankers must have derived most of their profits from origination fees and commissions on commodities. Panel C shows that there were a total of 12 bankers involved in the early years.

Figure 3 GROWTH AND DECLINE OF NEGOTIATIE MARKET

Sources: Van der Voort (1973). We have adjusted the figures of Van der Voort using information from two main sources: (1) Prospectuses and other deeds of Amsterdam notaries, City Archives Amsterdam, Collection of the Amsterdam Notaries 1578–1915 (5075). (2) Negotiatie details from the MBS auction records, Archive of the Burgomasters, City Archives Amsterdam.

Table 1 OVERVIEW OF NEGOTIATIES MARKET

Notes: Pooled negotiaties are based on multiple mortgages. The loan-to-value ratio is the mortgage sum divided by the appraised value. The number of banks indicates the number of banks issuing a new negotiatie. Amounts are in guilders. This table covers the period 1750–1773. Beyond 1773 the volume of newly issued negotiatie is very small, as illustrated in Figure 3.

Sources: Van der Voort (1973). City Archives Amsterdam, Collection of the Amsterdam Notaries 1578–1915 (5075). Dutch National Archives, Notarial Archives of Surinam 1699–1845 (1.05.11.14). City Archives Amsterdam, Archive of the Burgomasters (5068), with authors’ calculations.

The years 1766–1768 show a steep increase in the number of new negotiaties to 27. The average mortgage size doubles to 64,921 guilders in this period. Strikingly, the median LTV ratio is exactly the maximum of 5/8. In these years, the mean MBS coupon increases to 5.8 percent, and the markup is compressed to virtually zero. The number of bankers increases to 20. The years 1769–1770 can be considered the boom period for plantation MBS. Negotiatie volume and mortgage size roughly doubled compared to three years before. Although the same number of bankers are active, the size of their negotiaties increases dramatically. Appraisals also increase, and the mean LTV creeps close to the maximum of 5/8. Strikingly, the median coupon and interest rates were 6 percent throughout all three periods, including the boom.

THE BUILD-UP OF A PLANTATION BUSINESS: HUDIG’S NEGOTIATIES

Table 2 provides an overview of all negotiaties issued by Hudig and the underlying plantations (the locations of the plantations are in Figure A.2 of the Online Appendix). Hudig built up his portfolio over the years 1759 to 1778. He set up his first negotiatie in 1759, backed by a 100,000-guilder mortgage on sugar plantation Roosenburg and coffee plantation Mon Bijoux. The mortgage had a maturity of 20 years, with annual principal repayments of 5,000 guilders and an interest rate of 5 percent that was fully passed on to investors. Hudig received a fee of 0.5 percent of the principal value at origination, and a 2 percent commission on sales of the plantation products over the life of the mortgage (Oostindie Reference Oostindie1989, p. 346). The owner resided in the Dutch Republic, while three administrators in the colony supervised the plantations. Hudig appointed one of the administrators, Dirk van der Mey, as his agent (in partnership with Adriaan Gootenaar). The negotiaties were sold among Rotterdam elites, including Hudig’s relatives.

Table 2 HUDIG’S NEGOTIATIES

Note: This table gives an overview of the different mortgages contracted on the plantations for which Hudig issued and managed Negotiatie funds.

Sources: City Archives Rotterdam, Collection Coopstad and Rochussen, 68, with authors’ calculations.

In 1764/5, five years after Hudig’s first negotiatie, he initiated two more negotiaties on the plantations of Annaszorg and Somerszorg. The mortgage contracts were similar, although with an extended maturity (both 28 years) and lower annual amortization. For the Annaszorg mortgage, Hudig claimed an additional 0.5 percent fee for the arrangement of shipping insurance. The mortgage sums of Hudig’s early negotiaties were all well below 50 percent of the appraised plantation values.

In 1766, Hudig raised a negotiatie collateralized by a mortgage on a coffee plantation, Janslust, and its corresponding woodlands, Blockenbosch, and two more negotiaties in 1768: one with a mortgage on coffee plantation La Confiance, and one with a mortgage on plantation Venetia Nova and its wood grounds, Ma Resource. By then, the negotiatie market was growing rapidly, and the surging number of investors and banks in the market had led to increasingly standardized mortgage contracts (Van de Voort Reference Van de Voort1973). The mortgages of La Confiance, Venetia Nova, and Ma Resource followed the mortgage blueprint with mortgage sums equal to 5/8 of the appraised plantation value, interest rates of 6 percent, and a 20-year maturity starting with a ten-year interest-only period. The mortgage contract for plantation Janslust looked similar, only with an even higher sum of ¾ of the appraised plantation value. Adding these plantations to his portfolio, Hudig displayed a certain bullishness about the negotiatie market. Due to the abundant supply of capital, investors were not compensated for the increased risks in these new mortgage contracts, as they still received the 5 percent interest on their bonds. Hudig, however, charged 6 percent to the plantation owners and generated a 1 percent interest markup.

Plantation revenues were boosted by a surge in coffee prices in the late 1760s. Although there had been considerable price volatility in the first half of the eighteenth century, by 1766, the coffee price had been relatively stable for a decade between 0.70 and 0.82 guilders per kilo and was moving upwards from 1766 onward. The higher commodity prices facilitated additional funding. The plantations under Hudig’s negotiaties that were issued before 1766 renewed and increased their mortgages. Hardly any amortization took place. Planters used new and higher plantation appraisals to justify the additional debts, which Hudig funded by enlarging the negotiaties and issuing more securities to investors.

Somerszorg is an informative example. Initially, Hudig declined to pay the plantation’s outstanding bills of exchange because the plantation had already drawn credit up to the maximum mortgage sum. However, when the planter provided a new appraisal report in 1766, which listed an increased number of enslaved people and a higher appraised value of 157,805 guilders, the mortgage was increased to 98,627 guilders, equal to 5/8th of the newly appraised value of the plantation. Three years later, Hudig received a new appraisal report of Somerszorg valued at 234,755 guilders (a 49 percent increase), although the number of enslaved people had decreased since the previous appraisal report.

The increase in appraised values reflects the boom in plantation prices, but also points to the possibility of overappraisals. Just five days after the new appraisal, Somerszorg was sold for 202,500 guilders, about 15 percent below the appraised value. Such cases should have been a warning signal. However, there was fierce competition between bankers. For example, in 1768, Hudig lost out to a competing Amsterdam banker with regard to the mortgage on plantation Purmerend that Hudig was in the process of securitizing.Footnote ¹⁴ Bankers wishing to impose stricter credit terms would have a hard time finding borrowers.

During the height of the boom, Hudig added the plantations of Duringen, Driesveld, and Bijgeleegen to his portfolio. Duringen was an established plantation and had the same owner as Somerszorg. Its interest coverage ratio was higher than that of La Confiance and Venetia Nova (Table A.1, Online Appendix), indicating that Hudig became slightly more conservative. The coffee plantations of Driesveld and Bijgeleegen were both owned by Bernard Texier, the future governor of the colony. Texier agreed with Coopstad and Rochussen in 1770 that they would accept a mortgage on his plantation, Driesveld (and later the adjoining grounds Bijgeleegen), if his current bankers, Hermael and Van den Bosch, declined to do so.Footnote ¹⁵ When Hermael and Van den Bosch declined, Coopstad and Rochussen made a deal with Hudig to establish a negotiatie collateralized by Texier’s extended mortgage on plantation Driesveld in Hudig’s name. Hudig was bound by this agreement, even though he was no longer keen to provide new plantation mortgages after 1770.Footnote ¹⁶ The mortgage terms were standard. The Duringen and Bijgeleegen mortgage sums both equaled 5/8th of the appraised plantation values, while Driesveld obtained the credit of 192,500 guilders on an appraised value of 271.230 guilders, equaling as much as 71 percent of the appraised plantation value.

How did Hudig manage the cash flows and debts of his portfolio over time? Table 3 provides a summary of the financial accounts, aggregated over all negotiaties and plantations from 1766 to 1780. We start with the plantation owners’ debts to Hudig’s negotiatie funds. The total negotiatie funds increased from 226,000 to 868,000 guilders between 1766 and 1772, when Hudig’s business grew from two to eight funds. Next, we present the gross revenues net of Hudig’s administration fees and shipping and insurance costs, scaled by total negotiatie debts. For example, in 1768, the revenues were 73,852 guilders, or about 16 percent of the negotiatie debts. Interest to bondholders held the primary claim on revenues. Until 1775, plantation revenues were sufficient to cover interest payments—Table 3 shows that the actual interest payments were close to the contractual obligations. Remaining revenues were used for plantation expenses and investment. Plantation owners were the residual claimants.

Table 3 HUDIG’S FINANCIAL ACCOUNTS

Notes: The Number of negotiaties is the total number of negotiaties outstanding and Negotiatie debt is the total value of negotiaties under direction of Hudig outstanding, both at the end of the book year. Plantation revenues are the revenues from commodity sales, net of Hudig’s administrator fees and shipping costs. The Interest – payable and Interest – actual payments are the contractual interest obligations and the actual payments, respectively. Total Income Hudig consists of the Interest markup arising from the negotiaties, the Interest from debt on balance sheet the planters owed Hudig directly, and Transport and insurance fees. Total debt Hudig is the sum of Long-term debt held by Hudig, the Current account debt the planters owed Hudig on short-term advances, and the Negotiatie debt on balance sheet, all measured at the end of the book year. Amounts are in guilders. The detailed analysis of Hudig’s accounts covers the period 1766–1780. This period includes the build-up of Hudig’s negotiatie portfolio during the second half of the 1760s, and the growth of plantation debts during the 1770s.

Sources: City Archives Rotterdam, Collection Coopstad and Rochussen, 68, with authors’ calculations.

As Hudig issued more funds between 1766 and 1773, his gross earnings increased steadily. About half came from the interest markup, that is, the 1 percent difference between the 6 percent mortgage interest rate and the 5 percent paid on negotiate bonds. The other half came from interest on planation credit on Hudig’s balance sheet and transport and insurance fees. Over 1766–1771, Hudig’s total gross earnings amounted to 36,655 guilders.Footnote ¹⁷

Finally, we present the total debts on Hudig’s balance sheet. These consist of the current account debts of the plantation owners with Hudig, direct (long-term) positions in plantation mortgages, and Hudig’s own position in the negotiatie funds. The current account debt reflects the junior (unsecured) short-term debt that the planters had with Hudig. Before 1771, these reflected relatively small amounts (between 1766 and 1770, 0.75 percent of the total negotiatie sums). From 1771 onward, this position started to increase as Hudig allowed the planters to draw bills of exchange on him. In 1772, this amounted to 3.5 percent of the total negotiatie sums. Hudig’s direct (long-term) positions in plantations originated from converting existing unsecured short-term debt into secure, long-term mortgage debt. This was a way for Hudig to secure the advances to planters with a formal mortgage. Later, this would become a source of contention with investors, who now saw the emergence of a claim on equal standing with theirs. Finally, Hudig’s own position in the negotiatie funds, at least up to 1775/1776, mainly reflected the warehousing of negotiatie bonds between formal issuance and sale to investors. In 1773, when Hudig stopped issuing new MBS, virtually all negotiatie debt had been placed (see row “Negotiatie debt on balance sheet”). Hudig only kept a MBS position of 6,000 guilders.

In summary, the development of the overall market and Hudig’s experience demonstrate how the negotiatie system reduced plantation owners’ financial constraints. In combination with a well-informed intermediary and the availability of collateral in the form of a plantation, debt contracts appear to have been an effective solution to the information problems inherent to the plantation business. The opportunities for investors to diversify over several negotiaties, and the institutional “plumbing” involving bankers, local agents, and appraisers allowed Dutch investors to fund many colonial enterprises. The market developed as a pooling equilibrium, where the instruments were perceived to be homogenous and, as a consequence, saw limited price discrimination.

However, as the boom progressed and more capital became available, bankers like Hudig, relying on inflated appraisal reports, provided too much debt to the planters. The underlying dynamics are consistent with the model of Povel, Singh, and Winton (Reference Povel, Singh and Winton2007). As in their model, toward the end of the boom, bankers appear to have primarily based their lending decisions on public signals and information provided by the borrowers, in particular coffee prices, aggregate credit flows, and (overly optimistic) appraisal reports, rather than due diligence. Furthermore, there was strong competitive pressure at the height of the boom, which forced bankers to play along or drop out of the game entirely (Ruckes Reference Ruckes2004).

COLLAPSE OF THE PLANTATION CREDIT MARKET, 1770–1773

In early 1770, Jan Nepveu, Surinam’s governor, expressed his concerns to his principals back in the Dutch Republic about the “exorbitant” amounts of credit leading to surging plantation prices. According to the governor, the price of some plantations had risen by more than 100,000 guilders in six months.Footnote ¹⁸ He also expressed concerns about “speculative” plantation buyers’ limited wealth and their inability to pay for plantation maintenance in a downturn.Footnote ¹⁹

In 1770, the problems in the negotiatie system started to become apparent. First, there was a spike in the refusal of bills of exchange. Figure 4 shows the guilder amount of dishonored bills per quarter for our sample of 3,175 refused bills from 1770 onward. In the third and fourth quarters of 1770, around one million guilders in bills were refused, indicating that, on many occasions, bankers (and their investors) refused to accept a mortgage and were unwilling to transfer the promised sums. Van der Meulen (Reference Van der Meulen1904, p. 517) provides anecdotal evidence that the Surinam agent La Croix had offered mortgages to fund a plantation’s entire purchase price. When prospective buyers tried to access the negotiatie credit in Amsterdam, their bills of exchange were refused. This shocked the market, triggering a herding reaction with more investors pulling out.

Figure 4 BILL OF EXCHANGE REFUSALS

Notaries: De Wilde 14319–14333; Pool 12724–12742; Van den Brink 10582–10618; Van Heel, 12880–12884; Van Homrich, 12390–12421.

Source: City Archives Amsterdam, Collection of the Amsterdam Notaries 1578–1915.

The spike in bill refusals was followed by severe problems for the planters. On his arrival in Paramaribo on 22 October 1770 from a trip inland, Governor Nepveu found “very big perplexity” in the town. The ships that had just arrived from Holland had returned a great number of dishonored bills of exchange.Footnote ²⁰ Nepveu now feared that the majority of the bills that still circulated (estimated at 1.5 million guilders) would also be refused. The Surinam planters had already suffered from a drought in 1769, reducing revenues in 1770. A stop in credit would lead to a further deterioration of the plantation economy.

Afterward, Surinam was hit by a number of other shocks, aggravating the crisis. In 1771, there were renewed attacks from maroons (Van der Meulen Reference Van der Meulen1904, pp. 506–507). Coffee prices declined as competition from the French colonies increased (Combrink Reference Combrink2021). Further, the financial crisis of 1772/3 in London and Amsterdam likely reduced the availability of credit when a number of prominent Amsterdam merchant-banks, such as Clifford and Sons and Pels and Sons, failed. Moreover, two houses heavily involved in the negotiatie business, Abraham ter Borch and Sons and Clifford and Chevalier, did not survive the crisis either (Kosmetatos Reference Kosmetatos2014, 2018a, 2018b; Goodspeed Reference Goodspeed2016; Koudijs and Voth Reference Koudijs and Voth2016). Finally, in 1780, war broke out between the Dutch and English, further aggravating Surinam’s problems. The English invaded the adjoining Dutch colonies of Demerara and Essequibo, and there was a substantial risk the same would happen to Surinam (Wolbers 1861). Moreover, the war restricted Dutch transatlantic trade, and colonial products had trouble reaching the Dutch ports. Slave imports were also affected (Van de Voort Reference Van de Voort1973, p. 5; Oostindie Reference Oostindie1989, pp. 366–7).

Figure 3 shows that the origination of new negotiaties declined substantially after 1770, presumably due to a combination of tightening credit in Amsterdam and worse plantation fundamentals. The system did not implode from one day to the next. Table 1 shows that there was still a non-negligible amount of new negotiaties in 1771–1773 (predominantly in 1771 and 1772). The average loan-to-value ratio was lower than during the 1766–1770 period, indicating that only (somewhat) safer mortgages found their way to the market. In 1773–1774, there were still 2 million in negotiaties placed with investors. A closer inspection reveals that many of these were not newly placed issues with investors, but restructurings of existing negotiaties under new management.

In Figure 5, we present an index of negotiatie prices from Amsterdam auctions. We report the average yearly price of all securities traded (full sample) as a dashed line. While prices remain around par value until 1771 and even into 1772, they spiral downward afterward. In 1780, the average bond traded at 30 percent of its original value. At this point, the negotiatie system had virtually collapsed.

Figure 5 NEGOTIATIE PRICES

Notes: The dashed (solid) black line shows an (entropy balance-corrected) index price of all negotiaties sold in auctions. We report the average yearly price of all securities traded. Auction prices of Hudig’s negotiaties are in red. The gray bars are the number of sales per year.

Sources: Maandelykse Nederlandische Mercurius, Auction records, period 1768–1795, Archive of the Burgomasters, 70134–12198, City Archives Amsterdam.

HUDIG: INTERMEDIARY IN A PROBLEMATIC POSITION

To better understand the developments after 1770, we turn to Hudig, who also started to experience problems. First, we show that Hudig’s experience was broadly representative of market-wide developments using an entropy balancing approach (Hainmueller Reference Hainmueller2012). We construct a sample of negotiaties for which the average characteristics resemble Hudig’s portfolio (details are in Table A.2 of the Online Appendix). We start with 101 non-Hudig negotiaties for which we have complete data on their characteristics. Comparing Hudig to this larger sample shows that he had a somewhat smaller negotiatie portfolio than the average banker. Further, Hudig was not involved in any activities other than the negotiatie business, had relatively more exposure to sugar plantations, and was relatively early in its negotiatie issues, with most being issued between 1766 and 1768. Next, we give each of the 101 non-Hudig negotiaties a weight such that, on average, the weighted group resembles Hudig’s portfolio.Footnote ²¹ The average characteristics of this synthetic portfolio are presented in the column balanced control. Using the balancing weights, we construct a weighted price index that gives average prices for the synthetic portfolio. This is presented in Figure 5, together with actual transaction prices for Hudig’s negotiaties reported in his archives. The figure shows that the negotiatie price indices for the whole market and the synthetic Hudig portfolio are strikingly similar. In addition, the actual transaction prices for Hudig’s negotiaties are generally in line with the two indices.Footnote ²² In other words, Hudig’s experience is similar to the overall market.

Next, we delve into Hudig’s detailed records to better understand the position bankers found themselves in. On the eve of the crisis, Hudig noticed that the tide was turning. In May 1770, the owner of La Confiance sold the plantation to a, in Hudig’s eyes, “speculative buyer.” To finance the purchase, the buyer requested that the mortgage be increased from around 50,000 guilders to 60,000 guilders. Hudig refused, as this would exceed 5/8 of the appraised value. However, his agents, Van der Mey and Gootenaar, had already signed off on the mortgage request. Hudig was forced to take the plantation into receivership and bring it up for auction. In the summer of 1770, the owner of Venetia Nova and Ma Resource passed away, and Van der Mey and Gootenaar brought the plantations to auction in August 1770. Due to low demand, Van der Mey himself purchased the plantations for 80,916 guilders (appraised for 83,449 guilders two years earlier).Footnote ²³ This example shows that, already in the summer of 1770, prices had fallen below those of 1768. Moreover, the pool of potential buyers was starting to shrink in response to a decline in the availability of mortgage credit. Going forward, it would be harder to liquidate plantations.

Hudig wrote Van der Mey and Gootenaar in October 1770 that the intended auction of plantation La Confiance would need to be canceled—no one would be willing to extend credit to Surinam.Footnote ²⁴ The credit crisis hit before they could sell. An execution price far below fair value would damage Hudig’s reputation and access to credit.Footnote ²⁵ Moreover, Hudig would be forced to incur a substantial loss as his short-term credit was junior to the mortgages backing the negotiatie bonds.Footnote ²⁶ Hudig was critical of his colleagues in Amsterdam. He blamed the credit crisis in 1770 on their indiscriminate acceptance of mortgages before 1770 and their equally indiscriminate refusal of bills of exchange afterward.Footnote ²⁷

After the events of 1770, Hudig became dissatisfied with his agents. He got into direct conflict with Van der Mey over his purchase of Venetia Nova and Ma Resource and subsequent request for an additional mortgage.Footnote ²⁸ In response, he replaced his agents with Jean Rocheteau, who was heading to Surinam from the Republic to administer his family’s plantations, Janslust and Blockenbosch, after the passing of his father-in-law. Hudig also appointed Rocheteau as administrator of most other plantations in his portfolio (except for Annaszorg and Duringen). After his arrival in Surinam, Rocheteau wrote to Hudig that he was not selling Janslust and Blockenbosch just yet—due to the scarcity of credit, he expected the current sale price to be less than half of the mortgage amount.Footnote ²⁹ He added that only selling the plantations’ enslaved workers would not be advisable either, because slave sales would generate little revenue in such depressed times.Footnote ³⁰

In Table 4, we present yearly averages of each plantation’s revenues and expenses, interest coverage ratio (earnings/interest), and interest arrears for five-year periods. For seven plantations, we have data for both 1766–1770 and 1771–1775. For at least three of these plantations, revenues and the interest coverage ratio fell, indicating that the plantations’ economic prospects deteriorated. For two newer plantations, we only have data from 1771–1775 onward. For one of these, Driesveld, the interest coverage ratio was below one between 1771 and 1775, indicating that this plantation was unable to pay its interest payments.

Table 4 HUDIG’S PLANTATIONS: PRODUCTION, REVENUES, AND EXPENSES

Notes: The Revenues are the value of commodity sales net of administration fees and shipping costs. Expenses are the bills of the plantation owner paid by Hudig: they cover local expenses (most importantly slave purchases, taxes, director salaries, building materials, labor hire, and food), and imported goods from the Dutch Republic (most importantly building materials, tools, and clothes). Coffee, cotton, and sugar production are estimates in bales or hogsheads. Amounts are in guilders. We have no information on slave investments over 1766–1775. We have no information on plantations La Confiance, Venetia Nova, and Bijgelegen.

Sources: City Archives Rotterdam, Collection Coopstad and Rochussen, 68, with authors’ calculations.

Hudig kept the plantations afloat by extending short-term credit. Table 3 shows that up to 1775, the plantations’ actual interest payments equaled the interest payable even though revenues fell. Between 1771 and 1775, the plantations’ current account debts with Hudig increased dramatically from about 7,000 to 80,000 guilders, while the total plantation debt on Hudig’s balance sheet (including the mortgage debt and negotiatie debt Hudig kept on the balance sheet) rose to 147,474 guilders, or 17.5 percent of the negotiatie debt. The only alternative to extending credit was liquidating the underlying plantations and using the proceeds to repay investors. Selling at a discount would have damaged Hudig’s reputation. This behavior is consistent with Rajan’s (1994) model of bank credit policies. Toward the end of a boom cycle, banks will keep lending to insolvent borrowers to keep up the pretense that the existing loans are current. Only when the entire market turns (and the bank cannot be singled out for mismanagement) will the bank admit to the problems in its loan portfolio. If all banks follow this policy, this will prolong the boom and deepen the bust.

Hudig tried to limit the amount of short-term credit given to the plantations, but did not always succeed. On several occasions, Hudig refused to honor bills of exchange that his agents had not approved.Footnote ³¹ However, Hudig often felt forced to accept problematic bills. In 1767, he urgently asked his agents in multiple letters to stop drawing bills of exchange on behalf of Somerszorg as the outstanding debt had already greatly exceeded the mortgage amount.Footnote ³² Nevertheless, Hudig kept accepting new bills and only threatened to refuse them next time, which would lead to “excessive damage and shame.”Footnote ³³ Refusing a bill meant a fine for the plantation of 25 percent of the value of the bill, which would aggravate the solvency problems. Hudig also realized that the planters might try to find other means to generate income in the future, possibly by selling products illegally. This gave the planters and Hudig’s agents an incentive to overdraw, hoping Hudig would accept.Footnote ³⁴ As debts increased and plantation values fell, the planters became increasingly levered and thus prone to risk-shifting.

Table 4 has no information about La Confiance or Venetia Nova. The former simply had no revenues. After Hudig took La Confiance into receivership, the plantation remained unproductive in the absence of a new owner. When Hudig finally managed to sell it, the new owner absconded. The plantation fell into disrepair and was sold in 1780 at a fraction of its initial value. In the end, investors received only 20 guilders per bond with an initial value of 1,000 guilders.Footnote ³⁵ Venetia Nova also fell into disrepair, and revenues were small. The number of enslaved workers dropped from 61 in 1773 to 37 in 1779. Debtholders continued to receive interest payments (reduced to 2.5 percent) in about half of the years. In 1791, planter Van der Mey bought out the outstanding debt for 25 percent of its face value.Footnote ³⁶

LONG-RUN RETURNS TO INVESTORS

As the negotiatie system collapsed and bankers such as Hudig experienced severe problems with their plantation portfolios, investors started losing money. Figure 5 shows a drop in plantation MBS prices, but this ignores coupon payments. We use more detailed data to calculate actual long-run returns. We take 1795 as the end-year, when the Dutch Republic became a French client-state and Dutch financial markets started to undergo dramatic changes.

There are three negotiatie funds with sufficiently detailed information to calculate long-run returns. This includes the very first negotiatie initially set up by Willem Gideon Deutz in 1753, later taken over by Jan and Theodoor van Marselis, and two negotiaties (Letters “A” and “C”) issued by Harmanus van de Poll & Co in 1765 and 1769, respectively (see De Jong, Kooijmans, and Koudijs (2023) for information about these funds). All three bankers had a high reputation; Elias (Reference Elias1905, table VII) lists Marselis and Van de Poll as two of Amsterdam’s “most prominent merchant houses,” and returns likely provide an upper bound for the plantation MBS market as a whole. In terms of size, the three negotiaties were the first, third, and fourth largest ones in the market, accounting for 22 percent of total negotiatie volume up to 1773. The Deutz negotiatie initially had a coupon of 6 percent, which was lowered to 5 percent when Marselis took over. Van de Poll’s Letters “A” and “C” negotiaties had coupons of 5 and 5.5 percent, respectively. We use Van der Meulen (Reference Van der Meulen1904) and Van Stipriaan (Reference Alex1993) to reconstruct actual annual interest and amortization payments and the Maandelijkse Nederlandsche Mercurius for secondary market prices from auctions.

Results are in Figure 6, Panel A. We calculate the annualized return for investors between the issuance of the MBS and a sale in the secondary market, taking all intermediate coupon payments into account.Footnote ³⁷ Since we observe multiple transactions for each negotiatie, this gives us a set of returns over time, allowing us to look at short- and long-run returns. Results show that, initially, the Deutz/Marselis negotiatie offered investors returns of around 5 percent. This started to drop after 1780. Marselis lowered the coupon to 3 percent but continued to pay interest every year. In 1796, long-run returns fell to around 4 percent. The two Van de Poll negotiaties saw a much earlier and stronger decline in investor returns. Van de Poll reduced coupon payments much more aggressively than Marselis, leading to large price declines. Around 1780, investors faced an annualized loss of 2.5 percent on Letter “A” and 6 percent on Letter “C.” Afterward, as Van de Poll stabilized its annual coupon payments, prices recovered somewhat and investor returns improved, converging to about 1 percent in 1795.

Figure 6 LONG-TERM RETURNS

Notes: We calculate the annualized effective compounded returns for each year in which we can observe the closing of a position. If we observe multiple closings in the same year we take the average rate of return.

Sources: Van der Meulen (Reference Van der Meulen1904), Van Stipriaan (Reference Alex1993), Maandelykse Nederlandische Mercurius, Auction records, period 1768–1795, Archive of the Burgomasters, 70134–12198, City Archives Amsterdam. City Archives Rotterdam, Collection Coopstad and Rochussen, 68, with author’s calculations.

These results help us understand developments from investors’ perspective. Coupons were 5 to 6 percent. Given the risks inherent to the plantation business, investors likely expected to make less than that, but still earn a “risk premium” over the 2.5 percent coupon offered on Holland government bonds. Ex post, however, returns hardly exceeded 2.5 percent. Only the Deutz/Marselis negotiatie offered an ex post risk-premium of about 1.5 percent. However, this was the oldest negotiatie with most mortgages extended before the start of the boom and subsequent increase in plantation valuations. Moreover, its plantations had a comparatively long period to generate high returns before the system imploded. The Van de Poll negotiaties were placed later and performed substantially worse. Results, therefore, suggest that investors substantially overvalued plantation MBS during the boom of the late 1760s.

Figure 6, Panel B, presents the long-term returns on Hudig’s negotiaties using the same approach. Secondary market prices come from Hudig’s records and are sparser than the negotiatie prices discussed previously. Similar to Panel A, older negotiaties performed better. In fact, returns for Annaszorg (issued in 1765) are close to the coupon of 5 percent until at least 1780. Roosenburg & Mon Bijoux (issued in 1759) yielded around 3.75 percent for investors holding the bonds until 1788. Other, newer negotiaties returned substantially less than 2.5 percent. The lower bound was –9 percent (Driesveld) in 1777. In sum, the returns on Hudig’s negotiaties are in line with what we find for the Deutz/Marselis and Van de Poll negotiaties and point to generally poor returns, though older negotiaties, issued before the boom, performed better.

DEBT OVERHANG, RESTRUCTURING, AND THE LONG-RUN EFFECTS OF THE CREDIT CRISIS

The poor investor returns went hand in hand with economic stagnation in Surinam. Figure 1 shows that the credit boom of the late 1760s was initially associated with an increase in commodity production. Sugar and coffee production peaked in 1771 and 1775/7, respectively. By 1780, a number of shocks, including the credit crisis of 1770/1, had brought the negotiatie system to its knees. Figure 1 shows that total sugar and coffee production in Surinam stopped growing or even declined. Sugar production only reached its earlier peak in 1790. Coffee production, which had relied more on credit expansion, never reached its earlier peak (at least not before 1794, when the data ends). In other words, the plantation economy in Surinam recovered only partially and slowly.

One potential reason for this slow recovery is debt overhang (Myers Reference Myers1977). In a situation of distress, the claims of existing debtholders on future profits deter equity and debt infusions for new investments. This means that plantations with large debt burdens, lacking new investments in slave labor and the maintenance of buildings, sugar fields, and coffee trees, may have recovered more slowly or not at all. To test this hypothesis, we turn to Hudig’s experience. We first describe his attempts after 1775 to keep the plantations afloat and restructure their debt positions. Next, we study whether negotiaties with higher relative debt burdens saw a worse economic performance.

Hudig’s cash flow position was greatly impaired in 1776, and he was unable to provide more funding to the plantations. On paper, Hudig had greatly benefited from the negotiatie boom. Table 3 shows that the negotiatie business provided Hudig with gross earnings totaling 89,585 guilders over 1766–1775. However, these earnings took the form of an increase in planters’ current account debts, not actual cash flows. The revenues from selling the plantations’ coffee and sugar were hardly enough to cover interest payments to the negotiatie investors, let alone Hudig’s fees. Further, Hudig had paid some of the investors’ interest payments out of his own pocket, again through an increase in the planters’ current account debts. To improve his cash flow position, Hudig had to borrow 26,399 guilders from his brother in July 1774.Footnote ³⁸ In the following year, Hudig stopped granting additional credit to the plantation owners and refused several bills of exchange the plantations had drawn on him.

The plantations faced substantial debt burdens. Table 5 compares each plantation’s total debt position at the end of 1775 (negotiatie debt plus any debt owed directly to Hudig) to its most recent appraised value. All plantations’ debt levels were well above 5/8th of the last appraised value, with the exception of Roosenburg, Mon Bijoux, and Annaszorg. These appraised values presumably exceeded realistic 1775 market values. When Venetia Nova and La Confiance were sold in 1770 and 1773, respectively, transaction prices were 3 percent and 22 percent below the last appraisal values from 1768. Between 1773 and 1775, plantations’ market values fell even further.

Table 5 HUDIG’S PLANTATIONS: DEBT TO APPRAISED VALUE

Note: “Low-debt” plantations in bold.

Sources: City Archives Rotterdam, Collection Coopstad and Rochussen, 68, with authors’ calculations.

This situation was unsustainable. The plantations’ high debt burdens and low revenues led to interest arrears and necessitated debt restructuring. On 8 March 1776, Hudig called a first general meeting with his investors and proposed to appoint six men as directors (commissarissen) to represent the investors in the future co-management of the negotiaties. All directors were prominent Rotterdam citizens; many had large stakes in the negotiaties, and one was a former Surinam official.

The process was slow and fraught with conflict. A key point of contention was the status of planters’ current account debts with Hudig, which had been used to advance interest payments to investors. Depending on the negotiatie, discussions took two to three years. In the end, Hudig’s claims were largely converted into mortgages on the plantations that came on equal footing with the original negotiatie mortgages (see Table 3).

Little progress was made in the restructuring of debts. Hudig and the directors disagreed on how much new credit should become available to the planters to make the necessary investments and who should provide it. Hudig’s capacity to provide new credit was limited. The new directors were usually only willing to forego interest payments to support urgent needs. Investors were typically unwilling to provide new funding. In the absence of additional credit, most of the plantations did not have sufficient funds for maintenance or to replace old and deceased enslaved workers. Pleas from Rocheteau, Hudig’s agent in the colony, for new investment were often to no avail. Even if Hudig proposed to follow Rocheteau’s advice, directors often called investor meetings, where votes often fell in favor of interest payments rather than investments.Footnote ³⁹

The economic effects were detrimental. Table 4 presents the plantations’ production, gross revenues, expenses, and interest arrears for five-year periods. Expenses cover local costs (purchase of new enslaved labor, taxes, director salaries, building materials, and food) and imported goods from the Dutch Republic (building materials, tools, and clothes).Footnote ⁴⁰ Over the decade 1776–1785, low coffee prices, combined with a lack of credit, led to strongly reduced expenses, likely driven by a dramatic fall in investment in enslaved labor. At the same time, production fell and interest arrears increased.

We can differentiate between plantations that, according to Table 5, had relatively high (Somerszorg, Janslust, Duringen, and Driesveld) or low (Roosenburg, Mon Bijoux, and Annaszorg) debt burdens in 1775. Problems were more severe for the high-debt plantations. For example, at Janslust, the number of enslaved fell from 85 in 1771–1775 to 69 in 1781–1785. Mortality exceeded fertility (Van Stipriaan Reference Alex1993, p. 316), and hardly any new enslaved workers were purchased. At the same time, coffee production declined from 159 to 60 bales, a much larger decline than at other (lower-debt) plantations. By 1785, interest arrears had increased to 39,000 guilders, 39 percent of the original mortgage sum. Rocheteau continuously requested investment in new enslaved labor and other necessities. There was a shortage of everything. The enslaved workers likely suffered. Many walked around half-naked, and Rocheteau begged for clothes, in part to avoid fines by the local colonial government.Footnote ⁴¹ These patterns are consistent with two other high-debt plantations, Somerszorg and Duringen. Investors received full interest payments from both plantations until 1778, which left no room for investments in enslaved workers, whose number fell dramatically. Expenses were brought down to a minimum over 1776–1785, while production fell. After 1778, interest in arrears mounted. Again, Rocheteau pleaded for using the plantation’s cash flows to buy enslaved labor instead of paying interest, but to no avail.Footnote ⁴²

For the low-debt plantations of Roosenburg, Mon Bijoux, and Annaszorg, the situation was less problematic. For example, Roosenburg and Mon Bijoux did receive new credit. In 1776, the combined negotiatie ran a current account deficit of 4,500 guilders. To cover the payment of bills of exchange and interest due, and investments in the sugar cooking facility and new enslaved labor, Hudig and the directors decided to raise an additional negotiatie of 20,000 guilders, which would be senior to the old mortgages. The plantations were able to preserve the number of enslaved workers and stabilize production over 1771–1785. This example suggests that plantations that had not dramatically increased their debt levels during the boom retained access to funds and were able to sustain production.

From 1785 onward, the coffee price recovered, and Table 4 shows that all plantations’ revenues increased.Footnote ⁴³ This provided the funds for necessary maintenance expenses. High-debt plantations’ prospects improved. Somerszorg and Duringen used funds to deal with labor shortages first by renting slave labor in 1785–1790 (a relatively rare phenomenon) and then, in 1790–1795, by buying additional enslaved workers that stabilized the size of the workforce. Driesveld made significant purchases of enslaved labor between 1790 and 1795, which brought the size of the workforce back to the same level as in 1779. Still, the high-debt plantations were unable to meet these expenses without the willingness of the debtholders to forgo interest payments, and interest arrears kept mounting. Low-debt plantations also saw improved conditions. The number of enslaved workers on Annaszorg grew after 1785. Roosenburg and Mon Bijoux acquired new enslaved labor, though this was not sufficient to stop the decline in enslaved labor on Roosenburg. Their interest arrears remained minimal.

In summary, the overextension of credit during the late 1760s could not easily be reverted. Hudig’s experience shows that restructuring was problematic. In line with Myers’s (1977) model of debt overhang, this had large negative real consequences. The debt contracts that were efficient at the beginning of the negotiatie system crippled it later on.