The insufficiency is not in the data. Oosterlinck's archival work provides the syndicate composition to the franc, the bond pricing to the basis point, the GDP series to the gold franc. The NYT database provides 53 years of payment data verified against the archives of the Caisse des Depots. Henochsberg provides a continuous GDP series from 1760. The data are richer than what economists typically have for any nineteenth-century case study. The insufficiency is in the absence of a framework that can take these data and produce a disaggregated account of who captured what.

This paper supplies that framework by applying the Elite Extraction with Differential Targeting Model (EEDTM) to the Oosterlinck dataset. EEDTM is a mathematical system built around five core parameters, each given a plain-English name before its Greek symbol:

1. The extraction rate (tau, or τ): the proportion of a population's wealth or income captured per period
2. The elite capture rate (Theta, or Θ): the share of total extracted value that accrues to identifiable elites rather than being destroyed
3. The upstream constant (Phi, or Φ): the financier's guaranteed share of total elite capture
4. The differential targeting coefficient (Gamma, or Γ): the ratio of extraction rates applied to different populations
5. The destruction-capture ratio (DCR): the ratio of value destroyed to value captured

These parameters have been documented across twenty extraction episodes spanning the period 1625 to 2025, including colonial labor systems, industrial monopoly, mortgage securitization, private incarceration, and disaster-debt cycles (Bertil, 2026). The central empirical finding is that Theta clusters around 0.80 and Phi around 0.40 with remarkable consistency across time, geography, and mechanism. Elites capture approximately 80 percent of extracted value, and financiers take approximately 40 percent of what elites capture, regardless of whether the extraction occurs through enslavement in seventeenth-century Saint-Domingue or mortgage fraud in twenty-first-century Cleveland.

The contribution of this paper is threefold.

First, we compute Phi for the Haiti case using Oosterlinck's archival data and find Phi = 0.451. The banking syndicate led by Laffitte and Rothschild captured 45.1 percent of total extraction as the upstream intermediary... nearly equaling the colonial indemnity claimants they were nominally serving. This result falls within one standard deviation of the cross-case mean (0.40 ± 0.05) and represents an independent derivation using primary archival sources that predate the EEDTM framework by two centuries.

Second, we quantify Gamma... the differential targeting coefficient... using two natural comparators that Oosterlinck's data make available. The French Milliard des Emigres (1825) yields Gamma = 10: Haitians were charged ten times per capita what French citizens were charged for an equivalent indemnity program. The Irish Famine Loan (1847) yields Gamma = 22.6 to 173.8 depending on whether Ireland-only GDP (Mokyr, 1985; Andersson & Lennard, 2019) or combined British-Irish GDP is used as the denominator. These values fall within the range documented across other EEDTM cases (Gamma = 3.2 for US subprime lending to Gamma = 6,500 for the Haiti indemnity measured against pre-revolutionary wealth).

A note on positioning. This paper does not claim to replace the existing literature on Haiti's independence debt. The historical scholarship (Blancpain, 2001; Beauvois, 2009; Joachim, 1971; Briere, 2004, 2006), the archival work (Oosterlinck et al., 2021, 2022), the GDP reconstruction (Henochsberg, 2016), and the forensic journalism (New York Times, 2022) are the empirical foundation on which the present analysis rests. Nor does it claim that EEDTM is the only viable framework for analyzing colonial extraction. The odious debt doctrine (Sack, 1927; Oosterlinck et al., 2022), the debt-and-growth literature (Patillo, Poirson & Ricci, 2011; Panizza & Presbitero, 2013; Eberhardt & Presbitero, 2015), and the colonial bond pricing literature (Accominotti et al., 2010, 2011; Flandreau & Flores, 2009) each illuminate different facets of the problem. EEDTM's contribution is to provide a unified mathematical framework that connects these facets: to show that the financier's share (documented by Oosterlinck), the racial premium (documented by Accominotti), and the growth impact (documented by Patillo et al.) are parameters of a single system rather than independent phenomena.

The paper proceeds as follows. Section 2 describes the data: Oosterlinck's archival findings, the EEDTM framework, and new data introduced for the Ireland comparison. Section 3 presents the formula mapping and results. Section 4 develops the Ireland natural experiment. Section 5 reconciles the damages estimates. Section 6 addresses limitations and identifies remaining variable holes as research opportunities. Section 7 discusses implications for reparations, litigation, and the study of sovereign debt.

2. Data

This section describes three data sources: the archival dataset assembled by Oosterlinck and co-authors (Section 2.1), the EEDTM mathematical framework (Section 2.2), and new data introduced by this paper (Section 2.3).

2.1 Oosterlinck Archival Data

Oosterlinck, Panizza, Weidemaier, and Gulati produced two companion studies of Haiti's independence debt. The first, "The Odious Haitian Independence Debt" (2021), is a working paper containing the economic analysis: GDP reconstruction, debt-to-GDP time series, counterfactual growth models, and damage estimates. The second, "A Debt of Dishonor" (2022), published in the Boston University Law Review, provides the legal analysis and draws on previously unexamined archival documents from the Rothschild Archive in Roubaix, France (Box 132 AQ 73).

Four documents from this archive are central to the present paper:

Table 1. Archival Documents from Box 132 AQ 73, Rothschild Archive, Roubaix

These documents establish the composition and economics of the underwriting syndicate with a precision unusual in the colonial finance literature. The syndicate that purchased the 1825 loan at 80 percent of par comprised:

Table 2. Syndicate Composition, November 4, 1825

The three anchor institutions... Receveurs Generaux, Laffitte, and Rothschild... each held equal 32-million-franc shares, collectively controlling 65.8 percent of the distribution. This concentration matters for the Phi calculation: the upstream intermediary layer was dominated by three entities, all deeply embedded in the French political establishment.

Jacques Laffitte's position is particularly revealing. He was simultaneously a member of parliament (1816-1824), former Governor of the Banque de France, lead underwriter of the Haitian loan, and the party Haiti later asked to negotiate a moratorium on payments it could not meet (Oosterlinck et al., 2022, pp. 1262-1266). Laurent (1842, p. 8) described Laffitte as having "acted in a fraudulent way." Whether or not that characterization is legally precise, it captures a structural feature that EEDTM formalizes through Phi: the financier occupied every node of the transaction.

The pricing structure of the loan further illuminates the extraction mechanics:

Table 3. Bond Pricing and Spread, 1825

Haiti issued bonds with a face value of 30 million francs but received only 24 million in proceeds (80 percent of par). The syndicate planned to float the bonds to the public at 83.5 percent, pocketing a 3.5-percent spread on the face value... approximately 1.05 million francs in underwriting profit. The London Stock Exchange crash of November 11, 1825 eliminated this spread, but the 6-million-franc discount to par remained a permanent extraction from Haiti.

The resulting debt-to-GDP ratio of approximately 270 percent at inception is among the highest documented in the sovereign debt literature. For context, Greece's debt-to-GDP ratio at the onset of its 2010 crisis was approximately 146 percent (Reinhart and Rogoff, 2011). Haiti's initial extraction burden was nearly twice that level. Piketty (2020) independently estimates the ratio at 300 percent, a figure that falls within the margin of error given the uncertainty in the GDP denominator.

The debt-to-GDP trajectory that followed is itself a dataset of considerable value. From the initial 270 percent, the ratio declined to approximately 160 percent after the 1838 renegotiation (which canceled 60 million of the remaining 120 million francs owed), then to approximately 40 percent by 1875 as GDP grew at roughly 2 percent per year. But the ratio then reversed, climbing back to 52 percent by 1915 as new debts accumulated and National City Bank of New York acquired control of Haiti's national bank and government-guaranteed railway debt. Under the US occupation (1915-1934), debt service consumed more than 80 percent of government revenues at its peak, more than 30 percent from 1925 to 1936, and more than 15 percent from 1936 to 1946. The debt was finally retired through an internal dollar-denominated loan in 1947, bringing the ratio below 6 percent by 1950.

The post-1950 trajectory adds a further dimension. Under Francois Duvalier (1957-1971), debt-to-GDP rose by 10 percentage points to approximately 15 percent. Under Jean-Claude Duvalier (1971-1986), it accelerated, peaking at 45 percent in 1987. The HIPC Initiative brought it back to 5 percent by 2011, before it rose again to 15 percent by 2019. The pattern... debt imposed, partially reduced, re-accumulated under kleptocratic governance, partially forgiven, re-accumulated again... is consistent with the EEDTM concept of the "resistance ratchet," in which blocking one extraction mechanism (through debt cancellation, revolution, or international intervention) leads to substitution rather than elimination.

Haiti's revenue data further contextualize the extraction burden. Beauvois (2009) estimates average government revenues at 2,581,210 francs per year for the period 1818-1824. At this level, the 150-million-franc indemnity represented approximately 58 years of total government revenue. Even using Ternaux's (1825) grossly inflated claim of 37 million francs in annual revenue... a figure Oosterlinck et al. describe as "well above reality" and "divorced from reality"... the indemnity represented more than four years of revenue. The collapse of coffee prices from 290 francs per 100 pounds in 1821 to 140 francs in 1825 and 85 francs by 1830 (Blancpain, 2001) further eroded the revenue base. Haiti's capacity to pay was deteriorating at the precise moment the payment demand was imposed.

The Henochsberg (2016) GDP series, available through the Paris School of Economics (piketty.pse.ens.fr), is the foundational data source for the long-run analysis. Henochsberg estimates Haiti's GDP from 1760 to 1915 using gold franc denominators, producing the only continuous economic series available for the country during this period. The methodology involves triangulation from export data, customs records, and fiscal accounts. While any single-country GDP estimate for the early nineteenth century carries uncertainty, the Henochsberg series has been adopted by multiple research teams (Oosterlinck et al., 2021; Piketty, 2020) and is consistent with the available fiscal and trade data.

2.2 The EEDTM Framework

The Elite Extraction with Differential Targeting Model provides a mathematical system for tracing value flows in extraction events (Bertil, 2026). The framework was developed through empirical analysis of twenty extraction episodes spanning five centuries and four continents, ranging from the Atlantic slave trade to twenty-first-century mortgage securitization.

The core insight is that extraction follows predictable mathematical patterns regardless of time, geography, or mechanism. Six formulas are used in the main text of this paper; the complete set of forty-one equations is provided in Appendix A.

The extraction rate captures the intensity of extraction relative to a population's economic capacity:

The elite capture rate measures what proportion of total extracted value accrues to identifiable elites rather than being destroyed through institutional damage:

Across twenty validated cases, Theta clusters at approximately 0.80, with a dual regime: direct extraction (Theta_d = 0.85 ± 0.07) and crisis extraction (Theta_c = 0.45 ± 0.15). The direct regime characterizes mechanisms where value is efficiently transferred... colonial tribute, labor exploitation, monopoly pricing. The crisis regime characterizes mechanisms where most value is destroyed rather than captured... foreclosure cascades, famine-driven asset stripping, disaster-debt cycles.

The upstream constant disaggregates elite capture into its financier (upstream) and operator (downstream) components:

Phi measures the financier's guaranteed share of total extraction. Across the validation set, Phi clusters at approximately 0.40: financiers capture 40 percent of what elites capture, regardless of the specific extraction mechanism. This is the "upstream constant" of the paper's title.

The differential targeting coefficient quantifies how extraction rates differ across populations:

The present-value damages formula compounds historical extraction to its current value:

The destruction-capture ratio measures the relationship between value destroyed and value captured:

The validation context is important for interpreting the results. EEDTM's twenty-case validation set includes colonial labor systems (convict leasing, 1865-1928, Theta = 0.85), industrial monopoly (Gary, Indiana steel, 1906-1970, Theta = 0.87), mortgage securitization (US subprime, 2004-2008, Theta = 0.80), private incarceration (US private prisons, 1983-2025, Theta = 0.92), disaster-debt cycles (Haiti TPS termination, 2017-2026, Theta = 0.45), and state-level comprehensive analyses (Washington State, 1850-2025, Theta = 0.952). The mean Theta across all validated cases is 0.80 with a median of 0.86 and a 95 percent confidence interval of 0.73 to 0.87 (Bertil, 2026).

The dual Theta regime was identified in December 2025 through cross-case analysis. Cases where extraction mechanisms transfer value efficiently (colonial tribute, labor exploitation, monopoly pricing, financial intermediation) cluster at Theta_d = 0.85 ± 0.07. Cases where extraction mechanisms destroy most value in the process (foreclosure cascades, famine-driven asset stripping, forced migration, disaster-debt cycles) cluster at Theta_c = 0.45 ± 0.15. The distinction matters because it determines what proportion of total harm is attributable to identifiable elite capture versus diffuse institutional destruction.

The Phi constant was identified through decomposition of elite capture across cases where the archival record distinguishes between financier (upstream) and operator (downstream) shares. The Atlantic slave trade produces Phi values of 0.38 to 0.42 (the banking share of total slaving profits). US subprime securitization produces Phi = 0.41 (the securitizer share of total extraction from homeowners). Convict leasing produces Phi = 0.39 (the lessee's financial share versus the state's operational share). The convergence at approximately 0.40 across these disparate contexts suggests a structural constant rather than a coincidence.

These six formulas, applied to the Oosterlinck dataset, produce the results reported in Section 3.

2.3 New Data

This paper introduces three data sources not present in the Oosterlinck studies.

The NYT Haiti Double Debt Database. In May 2022, the New York Times published a year-by-year payment database for Haiti's "double debt" (1825-1888), compiled from the archives of the Caisse des Depots et Consignations, the French Foreign Ministry at La Courneuve, and the French National Overseas Archives at Aix-en-Provence (New York Times, 2022). The database contains 197 rows covering indemnity payments, 1825 loan principal and interest, and late fees, with CPI adjustment and 2021 USD conversion. The verified total is 112,323,476 francs across 53 payment years (1826-1888), comprising 66 million in indemnity, 41.1 million in loan principal and interest, and 5.2 million in late fees introduced after 1870.

This dataset fills a critical gap. Oosterlinck et al. work with a net figure of 90 million gold francs (the 150-million-franc indemnity minus the 60-million-franc reduction negotiated in 1838). The NYT database provides the total actually paid, which is the appropriate denominator for Theta and Phi calculations: Haiti's losses are measured by what Haiti actually paid, not by what was contractually owed net of renegotiation.

The NYT methodology is rigorous. The payment data were tabulated from the archives of the Caisse des Depots et Consignations (the French public bank that collected and distributed the payments), cross-referenced with the French Foreign Ministry archives at La Courneuve, captain's logs in the Landes archives (documenting the physical transport of coffers in 1826), and parliamentary records of the 1838 treaty in the French National Archives. Historical exchange rates use Victor Bulmer-Thomas's unpublished series, reprinted with permission. GDP estimates use the Henochsberg (2016) series. The 2021 USD conversion follows a standard methodology: payment in francs divided by the franc-to-USD exchange rate, multiplied by the ratio of 2021 CPI to the historical CPI.

Several methodological choices in the NYT database are worth noting. The 24 million francs paid from the 1825 loan proceeds (which Haiti immediately surrendered to pay the first installment of the indemnity) are counted under loan payments to avoid double-counting. Bonds retired between 1827 and 1838 are valued at an average of 500 francs per bond (between the market price and par). Interest payments for 1839-1842 are estimated at 3 percent of outstanding bonds, reflecting the post-renegotiation interest rate reduction from 6 to 3 percent. These choices are conservative: they tend to produce a lower total than alternative methodologies, which is appropriate for a newspaper of record but also means that the 112.3-million-franc total may slightly understate actual payments.

The database also reveals a feature not discussed in the Oosterlinck studies: France introduced late fees on arrears beginning in 1870, totaling 5.2 million francs. This is double extraction within double extraction: Haiti was charged penalties for late payment of an obligation imposed under military duress to compensate enslavers for the loss of the people they enslaved. Each layer of the extraction generated its own revenue stream for French creditors.

Table 12. DCR Across Selected Cases

Haiti's DCR of 5.25 places it between the standard crisis range (DCR approximately 1.22) and the annihilation range (DCR greater than 10, exemplified by Tulsa). The interpretation is consistent with the historical record: the indemnity did not merely transfer Haitian wealth to France. It destroyed Haiti's institutional capacity for two centuries... the fiscal system, the education system, the infrastructure, the governance capacity... in ways that far exceeded the financial value received by French elites.

The 1949 IMF staff memorandum confirms the mechanism: "Fiscal policy in Haiti in general has not been development promoting, because it has been primarily motivated by considerations of immediate revenue yields and has generally failed to give active encouragement to economic development. The business community complains of a lack of sympathy on the part of the fiscal authorities, sudden arbitrary new taxes, and the like. Instances have been reported to show that promising new industries have been thwarted to the point of being forced out of business by the application of new taxes that made the venture unprofitable" (IMF, 1949).

This passage is remarkable for two reasons. First, the IMF itself... the institution most closely associated with fiscal austerity in developing countries... recognized in 1949 that Haiti's fiscal policy was pathological. Second, the mechanism it describes (taxing promising industries to death in pursuit of immediate revenue) is precisely the institutional damage channel that DCR captures. The debt imposed a fiscal imperative... service the debt... that precluded the developmental spending that might have produced catch-up growth. The destruction was not incidental. It was structural.

The Oosterlinck authors' own analysis of the institutional damage channel is revealing. They note that "it is thus possible that the kleptocratic regimes that stunted Haiti's development are a consequence of the independence debt" (2021, p. 20). This is a carefully hedged statement, and the hedge is appropriate given the difficulty of establishing causation over two centuries. But the DCR framework provides a quantitative structure for the argument: if 84 percent of the total cost was institutional destruction rather than direct capture, then the kleptocratic regimes are not merely possible consequences but likely ones. A fiscal system permanently oriented toward debt service creates incentives for governance structures that prioritize revenue extraction over economic development. The Duvalier kleptocracy (1957-1986) is, in this reading, the long-term institutional consequence of fiscal arrangements imposed in 1825.

The debt-to-GDP ratio's trajectory between 1875 and 1915 provides further evidence. After the 1838 renegotiation and the gradual reduction of the French debt, Haiti's ratio fell from 160 percent to approximately 40 percent by 1875. One might expect institutional recovery to follow debt relief. Instead, the ratio rose back to 52 percent by 1915 as new debts accumulated, National City Bank acquired control of the national bank, and the government took on railway-guaranteed debt that consumed more than 80 percent of revenues. The pattern is consistent with the resistance ratchet: reducing one debt created space for new creditors to fill the same fiscal niche. The extraction mechanism changed (from French indemnity to American commercial debt) but the extraction intensity was preserved.

3.7 Life Expectancy (Supplementary Application)

EEDTM includes a heuristic life-expectancy function calibrated to cross-country data:

This application is flagged as the weakest in the paper. The function was calibrated to contemporary cross-country data and is not designed for historical application. We include it only because it illustrates the human cost in accessible terms.

Using the 1825 extraction rate of tau = 0.556:

The model predicts a life expectancy of approximately 73 years under the 1825 extraction rate. Haiti's actual life expectancy in 2024 is 64 years... worse than the model predicts. The gap suggests either that the extraction rate has been sustained or compounded through subsequent extraction layers (the 1875 renegotiation, the National City Bank fiscal capture, the 1915-1934 US occupation, the Duvalier kleptocracy), or that the life-expectancy function does not adequately capture cumulative institutional damage. The result is illustrative, not probative.

Before proceeding to the Ireland natural experiment, we summarize the Section 3 findings:

Table 12.5. Summary of EEDTM Results for Haiti 1825

4. The Ireland Differential

4.1 Natural Experiment Design

The strongest empirical contribution of this paper may not be the Phi calculation but the natural experiment embedded in the Rothschild bank's treatment of two populations in the same era.

In 1825, the Rothschild bank (Paris branch) participated as a co-equal partner in the syndicate that underwrote Haiti's 30-million-franc independence loan. The syndicate purchased the bonds at 80 percent of par, extracting a 6-million-franc discount (20 percent of face value) as the cost of capital access. Annual interest of 6 percent compounded the extraction. Fees and commissions accumulated over the life of the debt.

In 1847, the Rothschild bank (London branch) arranged an £8 million loan for the British government to finance famine relief in Ireland. The Rothschild Archive's institutional history records that Lionel de Rothschild "waived all commission" on this transaction. No comparable waiver has been documented for any other major Rothschild sovereign transaction in the nineteenth century.

The juxtaposition is made more striking by the scale of human suffering in each case. In Haiti, the indemnity was imposed under the threat of re-enslavement, on a population that had liberated itself through a revolution that claimed approximately 100,000 lives (of a pre-war population of roughly 500,000). The 14 warships and 528 cannons that arrived at Port-au-Prince on July 3, 1825 represented not merely a diplomatic show of force but a concrete threat to reverse the only successful large-scale slave revolution in history. In Ireland, the famine killed approximately one million people and displaced another million through emigration between 1845 and 1852, reducing the island's population by roughly 25 percent. The British government's response... widely criticized by contemporaries and subsequent historians as inadequate... included the £8 million loan as one component of a larger relief effort that also involved public works programs and soup kitchens.

In both cases, the Rothschild bank occupied the same structural position: intermediary between a demanding sovereign power and a suffering population. The bank's treatment of the two populations diverged completely.

The two events share five features that make them suitable for comparison:

Table 13. Haiti-Ireland Comparison Matrix

The populations differ in one salient dimension: Haiti was a Black, formerly enslaved population that had won independence through armed revolution. Ireland was a white, European, British-subject population suffering under colonial misgovernance. The question is whether this difference... race, colonial status, or both... explains the differential treatment.

4.2 Three Explanations

Three hypotheses might explain why Rothschild charged commission on Haiti but waived it on Ireland.

Hypothesis 1: Credit risk. Haiti was a new sovereign with no credit history and an uncertain capacity to pay. Ireland was part of the British Empire, whose credit was the gold standard of the era. The commission differential might reflect a rational pricing of default risk.

This hypothesis has surface plausibility but does not survive scrutiny. Flandreau and Flores (2009) show that "the Rothschilds chased good securities and good securities only" and that "securities underwritten and issued by the House of Rothschild clearly outperformed the rest" (quoted in Oosterlinck et al., 2022, p. 1264). If the Haiti loan was too risky for Rothschild's standards, the bank should not have participated. Having chosen to participate, the commission should be understood as compensation for lending the Rothschild brand to a transaction the market would otherwise have priced lower... precisely the Phi mechanism.

Moreover, the Irish Famine Loan was not riskless. Ireland's economy was in catastrophic collapse. One million people died and another million emigrated between 1845 and 1852. The borrower was the British government, but the economic purpose... famine relief in a devastated colony... was at least as uncertain in its returns as Haiti's sovereign capacity. The credit-risk hypothesis cannot fully explain the differential.

Hypothesis 2: Market conditions. The 1825 London stock exchange crash occurred one week after the Haiti bonds began trading. Market conditions in 1825 were generally volatile. By contrast, 1847 was a period of relative stability in British government credit. The commission differential might reflect differences in market appetite.

This hypothesis explains the pricing of the Haiti bonds on the secondary market (the crash reduced the syndicate's planned spread) but does not explain the syndicate's ex ante decision to purchase at 80 percent of par. The 20-percent discount was agreed before the crash, based on Haiti's negotiating position (it wanted 90 percent), the competitive bidding process (first bids came in at 76 percent), and the syndicate's desired spread. Market conditions were a factor in the realized return, not in the commission structure.

Hypothesis 3: Race and colonial status (the Gamma residual). After controlling for credit risk and market conditions, a residual differential remains. Rothschild waived commission for a white European population and charged it for a Black Caribbean population. This residual is what EEDTM formalizes as Gamma.

To be precise: Hypothesis 1 explains some portion of the commission differential (Haiti was indeed a riskier credit than the British Empire). Hypothesis 2 explains some portion of the bond pricing dynamics (the 1825 crash was real). But neither hypothesis, alone or in combination, explains a complete commission waiver for Ireland alongside a full commission charge for Haiti. The residual... the gap between what credit risk and market conditions predict and what was actually charged... must be attributed to other factors. The most parsimonious explanation for that residual, given that the two populations differ most saliently in race and colonial status, is that these characteristics functioned as pricing variables.

This interpretation is not an imputation of individual racial animus. Lionel de Rothschild, who arranged the Irish loan, was himself a target of discrimination: he was the first Jewish member of the British Parliament, and his admission was delayed for over a decade by religious tests. The commission waiver may have been motivated by political calculation (cultivating British government goodwill), humanitarian sentiment (genuine sympathy for Irish suffering), or reputational strategy (the British Relief Association, which Rothschild founded, enhanced the bank's public standing). None of these motivations were extended to Haiti, and the question is why.

The residual interpretation is strengthened by the broader pattern of Rothschild lending in the period. The bank participated in the British government's 1833 Slave Compensation loans... £20 million to compensate enslavers for the loss of their "property"... and later managed the London Gold Fix for 85 years (1919-2004). In each case, the bank's commission structure tracked the racial and colonial hierarchy of its clients.

We do not claim that Lionel de Rothschild made a conscious decision framed in racial terms. Institutional racism operates through structures, not necessarily through individual intention. What the data show is that the same institution, facing comparable transactions in the same era, produced systematically different outcomes for populations that differed in race and colonial status. The differential is quantifiable:

Formally, Gamma is undefined when the comparator extraction is zero. Substantively, the result says that any finite commission on Haiti, when Ireland's was waived, produces an infinite differential for the commission channel specifically. This is an extreme case of Gamma... more extreme than can be captured by the per capita burden comparison (Gamma = 10 to 174) discussed in Section 3.4.

4.3 Implications for Sovereign Debt History

The Ireland comparison connects to a broader finding in the sovereign debt literature. Accominotti, Flandreau, Rezzik, and Zumer (2010) document a "race premium" in colonial bond yields... the finding that bonds issued by colonies with nonwhite populations traded at systematically higher yields than bonds issued by colonies with white populations, even after controlling for economic fundamentals. Their title, "Black Man's Burden, White Man's Welfare," captures the pattern.

The connection between Phi and Gamma... between the financier's share and the racial premium... is the paper's deepest finding. If the upstream intermediary charges more when the borrower is nonwhite, then the supposedly neutral financial intermediation layer is not neutral. It is a channel through which racial hierarchy is monetized. The Rothschild bank's brand carried a premium (Flandreau & Flores, 2009), and that premium was selectively deployed: offered as a charitable waiver for Ireland, charged at full market rates (or above) for Haiti. The financier's cut is not merely a function of intermediation services rendered. It is also a function of what the market will bear, and what the market will bear is, in part, a function of the borrower's race and colonial status.

Our finding extends Accominotti et al. in two ways. First, we document the race premium operating not in bond yields (the secondary market) but in underwriting commissions (the primary market). Rothschild's decision to charge commission on Haiti and waive it on Ireland is an upstream version of the yield differential that Accominotti et al. document downstream. Second, we provide a natural-experiment design (same bank, same era, comparable scale) that strengthens the causal inference beyond what cross-country yield comparisons can deliver.

The implication for the study of sovereign debt is that Phi... the upstream financier's cut... may itself be a function of Gamma. Financiers may extract at higher rates from populations they can differentially target. If the upstream constant varies systematically with the race and colonial status of the borrower, then the Phi = 0.40 average conceals a distribution in which Black and colonized populations face Phi values above the mean and white and metropolitan populations face Phi values below it. This hypothesis is testable with Flandreau and Flores's (2009) dataset on 1820s sovereign bonds, which includes both underwriting terms and issuer characteristics.

5. Damages Reconciliation

Three damage estimation approaches produce overlapping ranges:

Table 14. Damages Reconciliation

Methods A and B capture the restitutionary component: the present value of what was directly taken from Haiti and received by French elites. Methods C and D capture the compensatory component: the economic cost of institutional damage caused by the extraction.

The methods are not additive... they measure overlapping channels of harm. The direct extraction is a subset of the total economic cost. A reconciled range uses Method B (the direct extraction channel at the Schmelzing rate) as the lower bound and Method C (the conservative growth-effect channel) as the upper bound:

Reconciled damages: 21 to 144 billion

The lower bound (21B) represents the present value of documented financial flows, compounded at the 700-year average real rate. It is a conservative, well-sourced figure suitable for restitutionary claims against successor institutions. The upper bound (144B) adds the conservative growth-effect estimate from Patillo, Poirson, and Ricci (2011), representing the total economic cost including institutional damage. It corresponds to approximately ten times Haiti's 2020 GDP.

The ranges converge around 3 percent for a specific reason. At this rate, the direct extraction PV (8.1B using the NYT base) and the sovereign-wealth-fund counterfactual (123B from Oosterlinck et al., 2021, Table 1) produce a ratio of approximately 1:15. This ratio... one dollar captured for every fifteen dollars of total cost... is consistent with the dual Theta finding: the crisis Theta of 0.16 implies that direct capture represents approximately one-sixth of total loss.

This reconciliation also contextualizes previous estimates. Piketty's (2020) estimate of approximately 30 billion euros (roughly 35 billion) falls within the reconciled range, as does the BARSS EEDTM range of 100 to 170 billion (Bertil, 2026). The Aristide government's 22 billion demand (de Cordoba, 2004) corresponds closely to the direct extraction PV at the Schmelzing rate. Far from being contradictory, these estimates illuminate different channels of the same extraction event: Piketty captures the nominal reflation, the Aristide team captured the direct PV, and the growth-effect literature captures the institutional damage.

The reconciled range has a useful property: it corresponds to observable economic magnitudes. The lower bound (21 billion) is approximately 1.5 times Haiti's 2020 GDP of 13.6 billion. This is the "what was taken" figure... the restitutionary quantum. The upper bound (144 billion) is approximately 10 times Haiti's GDP. This is the "what it cost" figure... the compensatory quantum. The ratio between the two (approximately 7:1) is consistent with the dual Theta finding: the crisis-frame Theta of 0.16 implies that direct capture is approximately one-seventh of total loss.