A grant from the Center for International and Area Studies at Yale helped fund the research. I am particularly indebted to James C. Riley for advice and for permission to cite unpublished work. Helpful comments from George Alter, Peter Lindert, Larry Neal, Richard Sutch, and Andrew Trout as well as from two referees and the editor are also gratefully acknowledged.

¹ See Carter, Alice Clare, The English Public Debt in the Eighteenth Century (London, 1968), pp. 5–7,Google Scholar for a discussion of the Revolution as a prerequisite. Dickson, P. G. M., The Financial Revolution in England (London, 1967), pp. 3–14, emphasizes that Dutch society was the model for English economic reformers even before William of Orange was offered the throne in 1688.Google Scholar

² The War of the Grand Alliance (League of Augsburg), 1689–1697; The war of the Spanish Succession, 1702–1713; The War of the Austrian Succession, 1740–1748; The Seven Years' War (French and Indian War), 1756–1763; The American War of Independence, 1776–1783; The French Revolutionary War, 1793–1801; The Napoleonic Wars, 1803–1815. Each country also fought other battles with less direct involvement of the other.

³ Kennedy, Paul, The Rise and Fall of the Great Powers (New York, 1987), pp. 76–86, credits the financial revolution in Britain as the determining factor in Britain's emerging military supremacy over France.Google Scholar

⁴ Kindleberger, Charles, The Financial History of Western Europe (London, 1984), pp. 158–59.Google Scholar

⁵ The French government resolved in 1763 never again to raise its own funds directly by means of a tontine. During the Revolution the government nationalized a private tontine organized for the duc d' Orléans in 1785. In 1790 the Constituent Assembly rejected a complex plan by Lafarge, a mathematician, who established it anyway. It, too, was eventually nationalized. The Convention Nationale in 1795 established a plan for a tontine narionale (accepting assignats as principal) but revoked it before collecting funds. See Vührer, A., Histoire de Ia deue publique en France (Paris, 1886), pp. 304–10, 376–77.Google Scholar

⁶ The nineteenth-century private life insurance plans based on the “tontine” principle redistributed funds forfeited by nonrenewals of premiums as well as death. Subscribers paid in to private companies on an installment basis, receiving a lump-sum payment at the end of a term (if they survived), or a payment to their heirs (if they died before the end of term), or nothing (if they failed to keep up the payments). On the nineteenth century, see Sutch, Richard and Ransom, Roger, “Tontine Insurance and the Armstrong Investigation: A Case of Stifled Innovation, 1868–1905,” this JOURNAL, 47 (06 1987), pp. 379–90.Google Scholar

⁷ Bosher, J. F., French Finances, 1770–1795: From Business to Bureaucracy (Cambridge, 1970), is especially informative on the revolutionary desire to reform the process of public finance.Google Scholar

⁸ Even the most virulent critic of the monarchy's indebtedness, Marion, Marcel, Histoire financière de la France depuis 1715 (Paris, 1914), acknowledges that Britain was worse in purely quantitative terms (vol. 1., pp. 460–61).Google Scholar

⁹ See Harris, Robert D., Necker: Reform Statesman of the Ancien Régime (Berkeley, 1979), for a recent rehabilitation of Necker's position that the deficit was not of his making.Google Scholar

¹⁰ The figure for French debt service in Table I differs from the total of 318.3 million given by Braesch, F., Finances et monnaies révolutionnaires (Paris, 1936), vol. 2, p. 202, because I excluded 27.2 million in pensions, 7.8 million in overdue expenses of the maison du roi, and 3.1 million in operating expenses, but added 12 million in rentes viagères (annual life annuity payments) from the loan of November, 1787. 1 have retained another 11.4 million in acquisitions and liquidations (payments to members of the aristocracy) that might be interpreted as something other than debt service since the capital received by the king in exchange is not always evident.Google Scholar

¹¹ Mathias, Peter and O'Brien, Patrick, “Taxation in Britain and France, 1715–1810: A Comparison of the Social and Economic Incidence of Taxes Collected for the Central Governments,” Journal of European Economic History, 5 (Winter 1976), pp. 601–50.Google Scholar

¹² Williamson, Jeffrey G., “Why Was British Growth So Slow During the Industrial Revolution,” this JOURNAL, 44 (09. 1984),Google Scholar table I, used an 1801 nominal GNP estimate by Deane, Phyllis and Cole, W. A., British Economic Growth, 1688–1959 (Cambridge, 1962)Google Scholar, table 37, back-extrapolated at ten-year intervals using the real output growth rates from the same source (table 19), and adjusted to current prices using the price index in Mathias, and O'Brien, , “Taxation” (table 2, p.605).Google Scholar The figure of 107.1 million for 1781–90 is an average of estimates for 1780 and 1790. Two objections can be raised. The year 1801 is a poor choice of benchmark because it was a period of rapid change in both relative and absolute prices, when Britain was off the gold standard. It is difficult to know what level of prices corresponds to the nominal income data, and the results of back-extrapolation are highly sensitive to the choice. Mathias and O'Brien's price index is a decennial average and therefore very unlikely to provide an exact match. Second, Crafts, Nicholas, British Economic Growth During the Industrial Revolution (Oxford, 1985), has compiled a set of important revisions to Deane and Cole's real output growth rates.Google Scholar

I begin instead with Deane and Cole's (table 37) estimate for 1831, separated into an agricultural and a nonagricultural component (79.5 and 260.5 million, respectively). Crafts, , British Economic Growth gives estimates of real output growth for agriculture (p. 42) and total output (p. 45) for 1780–1801 and 1801–1831, and the share of agriculture in each period (p.45),Google Scholar from which a nonagricultural growth rate can be calculated. Assuming the growth rate constant within the first period, I obtain estimates of real output in 1780 and 1790 relative to the 1831 level for both sectors. Mitchell, B. R. and Deane, Phyllis, Abstract of British Historical Statistics (Cambridge, 1962), pp. 468–71, provide the necessary price data. The Schumpeter-Gilboy series (1696–1823) was spliced at 1815–19 to the Rousseaux series (1800–1913). For agriculture, the Schumpeter-Gilboy “Consumers' Goods(a)” series was spliced to the Rousseaux “Total Agricultural Products” series. For nonagricultural prices, the Schumpeter-Gilboy “Consumers' Goods other than Cereals(b)” series was spliced to the Rousseaux “Overall Index.” Prices in 1780, 1790, and 1831 were measured as three-year centered averages.Google Scholar

The 1780 level of nominal agricultural output is then calculated as the product of 1831 nominal agricultural output times the ratio of real output in 1780 to that in 1831, times the ratio of agricultural prices in 1780 to 1831, or 79.5 times 0.6012 times 0.7977 equals 38.1 million pounds. The real output and price level multipliers for 1790 agriculture are (0.6479, 0.8597); for nonagricultural output in 1780: (0.3767, 0.85 14), and in 1790: (0.4442,0.8955). Total nominal output in 1780 and 1790 is estimated at 121.7 and 147.9 million pounds, and the mid-decade average at 134.8.

¹³ I take 24 French livres to the pound as a reasonable estimate of currency exchange rates in the 1780s. Both countries adhered to fixed specie content of their currency. Following the French monetary stabilization of 1726, par of exchange was 29.2d of English money per French ecu (24.66 French livres tournois per pound sterling), according to McCusker, John J., Money and Exchange in Europe and America, 1660–1775: A Handbook (Chapel Hill, 1978), who finds market rates in London between 30 and 32d in most years up to 1775.CrossRefGoogle ScholarBouchary, Jean, Les Marchés de change de Paris a la findu XVIIIe siécle (Paris, 1937), finds the rate lower, around 29d per écu, in the mid–1780s (pp. 107–8).Google Scholar The two alternative purchasing power parity exchange rates for the 1780s calculated by O'Brien, Patrick and Keyder, Caglar, Economic Growth in Britain and France, 1780–1914 (London, 1978), p. 47, were 20.2 and 24.2 livres per pound. This is close to the currency market rate, although their own estimate of the market rate (29 livres to the pound) is inconsistent with Bouchary and McCusker.Google Scholar

In any event, the consistency of the higher British nominal per capita GNP estimate with the nominal wage data is independent of the exchange rate chosen. Phelps-Brown, E. H. and Hopkins, Sheila, “Seven Centuries of Building Wages,” Economica (1955)Google Scholar, reprinted in CarusWilson, E. M., ed. Essays in Economic History (London, 1962), vol. 2, p. 178,Google Scholar give daily wages of 19 pence per day for laborers in the building trades around 1785. Durand, Yves, “Recherches sur les salaires des macons a Paris au XVIIIe siécle,” Revue d'histoire économique et sociale, 44 (1966), pp. 468–80, shows summer daily wages of 28 sous for laborers. At 240 pence to the pound, 20 sous to the livre, and 24 French livres to the pound, the British wages were equivalent to 38 sous. Laborers' wages were therefore 36 percent higher in Britain than in France at currency exchange rates.Google Scholar

¹⁴ Crudely, at 5 dollars to the pound and 24 livres to the pound, American per capita GNP is 283 livres versus 262 in France and 345 in Britain.

¹⁵ Of the 292.2 million livres in French debt service, 65.8 million were explicitly for debt redemption (remboursement), mostly for short-term loans contracted during the American War. See Braesch, , Finances, vol. 2, pp. 192–203. That is about 36.5 percent of the non-life annuity portion of the debt service. In addition, the 102.3 million in annual payments on life annuities and tontines, mostly bought at 8 to 10 percent interest, contain a substantial fraction of amortization in addition to a high rate of interest. It is difficult to separate the two. If 20 percent is taken as a probable lower bound on the share of amortization, the life annuity interest rate is between 6 and 8 percent and amortization is not less than 30 percent of total debt service.Google Scholar

¹⁶ The nominal consol yield was 5.15 percent in 1785, 4.26 percent in 1786, and around 4 percent to 1790 (Heim, Carol and Mirowski, Philip, “Interest Rates and Crowding-Out During Britain's Industrial Revolution,” this JOURNAL, 47 [03. 1987], table 1). The total annual interest of 9.229 million pounds would therefore have been evaluated at around 217 million pounds in 1786.Google Scholar

¹⁷ My estimates agree with Braesch, Finances, vol. 2, with regard to the floating and short-term debt plus the acquisitions and liquidations, that is, everything except perpetual rents, life annuities, and tontines. Annual interest charges were assessed by the government at 5 percent, so the corresponding capital is estimated here at 20 times the annual interest, or 1,420.66 million livres. Braesch inappropriately used the same multiplier for perpetuals and life annuity rents after taxes. For each life annuity loan (including tontines), I calculated the fraction of original rents extinguished as of 1789 from the summary table in Marion, Histoirefinancière, applied that fraction to the original capital raised, and subtracted it from the original capital to get the surviving capital. Summed over all loans, that left 1,117.694 million livres in capital. An analogous procedure for the perpetual rents (some of which had been redeemed) suggests 2,042.054 million livres in capital by 1789, for a total debt of 4,580 million. Over half the perpetual debt (1,190.214 million) was attributable to two huge loans in 1720 associated with the liquidation of Law's system, at interest rates of 1 percent and 2.5 percent, for which much of the capital provided was depreciated paper. Reevaluating those two loans at 5 percent interest, the remaining perpetual capital would be 1,339.408 million and the total debt 3,877.8 million.

¹⁸ Riley, James C., The Seven Years War and the Old Regime in France: The Economic and Financial Toll (Princeton, 1986),Google Scholar and developed further in Riley, James C., “The Seven Years War and the French Revolution,” unpublished manuscript, Indiana University, 1983.Google Scholar

¹⁹ Williamson, “British Growth,” has been challenged by Heim and Mirowski, “Interest Rates and Crowding-Out,” and reassessed by Mokyr, Joel, “Has the Industrial Revolution Been Crowded Out?,” Explorations in Economic History, 24 (07 1987), pp. 293–319.CrossRefGoogle Scholar

²⁰ Interest charges as defined here exclude remboursemengs, but include the total of life annuity charges, which include some amortization. As a general rule, interest on offices is also excluded where possible.

²¹ Among other problems, the sources do not provide a consistent accounting of amortization, or interest on floating debt, or of changes in the gages paid to the private financiers of the government in exchange for their loans (a mechanism favored by Louis XIV and virtually abandoned by his successors). Annual gross tax receipts for 1690 to 1715 are reported by Guéry, Alain, “Les Finances de Ia monarchic francaise sous l'Ancien Régime,” Annales: Economies, Sociéiés, Civilisations, 33 (03.-04. 1978), pp. 216–39, p. 237.CrossRefGoogle Scholar Debt charge observations for 1712 (65.4 million in perpetual rents only) from Riley, , Seven Years War, p. 166.Google ScholarMarion, , Histoirefinancière, vol. 1, p. 63, gives 45 million in rents and 40 million in other charges in 1715.Google Scholar Vührer, Dette publique, cites 11.7 million in rents in 1689. Claageran, J. J., Histoire de l'impôt (Paris, 1876), vol. 3, gives 24 million in rents in 1699 (p. III), 68 million in 1734 (p. 279), and 71 million in rents and charges in 1725 against taxes of 204 million (p. 232).Google Scholar

²² Grellier, J.J., The National Debt (London, 1810), pp. 343–44, reports details of the funded and non-funded debt, showing a total of 77,097 pounds per year in payments on all types of life annuities in 1786, out of a total debt charge of 9.5 million, that is, less than one percent. Fixed-term annuities amounted to 1.26 million, and short-term debt interest 208,749.Google Scholar

²³ See Coudy, Julien, “La Tontine royal sous Ie régne de Louis XIV,” Revue historique de droit françcais er élranger, 4ème série, 35 (1957), pp. 128–33.Google Scholar

²⁴ Placing administration in the hands of the subscribers is not unrelated to the verification issue. Subscribers had a strong interest in preventing fraudulent receipts because they would reduce the payments to true survivors. The government's payout did not depend on number of survivors, so it had no incentive for verification until the very end when costs were low.

²⁵ An excellent summary work is Jennings, Robert M. and Trout, Andrew P., The Tonline: From the Reign of Louis XIV to the French Revolutionary Era (Homewood, IL. 1982).Google Scholar

²⁶ Assuming the exchange rate to have been 13 livres to the pound in 1689 and 1696, and 23 after 1726 (McCusker, , Money and Exchange, pp. 93–97), the 108 million livres raised in the ten tontines was equivalent to 5 million pounds. The English raised 2,548,000. The Irish tontine was designed and implemented by the Irish Parliament and raised the equivalent of 928,000 pounds sterling.Google Scholar

²⁷ The Bibliothèque Nationale has the published annual reports for some years between 1741 and 1769: Listes des rentes viagéres dites Tontines (Paris, various years).Google Scholar These form the basis of the work by Wyler, Julius, Die Tontinen in Frankreich (Munich, 1916), and Jennings and Trout, The Tontine. The number of original subscribers in each active division of each existing tontine is reported each year, along with the number of survivors and deaths of the past year, and the value of that year's payout. A life table based on this data will be reported at some later date.Google Scholar

²⁸ Archives Nationales de Ia France, Série P. Chambre des Comples, P5875-P5932, appears to be complete for the fourth through tenth tontines.

²⁹ The exact rates, for the fourth through tenth tontines were: 2.289, 2.877, 3.035, 3.101, 3.022, 2.177, and 3.231. The rates were higher in older age groups than in younger, indicating more multiple purchases at higher ages, except in the lottery-tontines of 1743, where they were nearly equal. Note that these rates indicate number of divisions per nominee. Subscribers could invest on several lives other than their own. In practice, this seems to have been infrequent and generally confined to other family members, who would then inherit the shares on their own lives. It seems reasonable to consider these eventual owners as creditors of the government.

³⁰ Dickson speculates that William Paterson, the author of the Bank of England charter, was also the architect of the tontine plan (Financial Revolution, p. 52).Google Scholar

³¹ Finlaison, Alexander len, Report and Observations on the Mortality of the Government Life Annuities. Ordered, by the House of Commons, to be Printed, 1860, p. 10.Google Scholar

³² Rotman-Zelizer, Vivian, Morals and Markets: The Development of Life Insurance in the United States (New York, 1979).Google Scholar

³³ CaIdwell, John C., The Theoty of Fertility Decline (New York, 1982).Google Scholar

³⁴ Encyclopédie, ou Dicrionnaire Raisonné des Sciences, des Arts, et des Métiers (Neuchâtel, 1765), vol. 16.Google Scholar

³⁵ See Alter, George and Riley, James C., “How to Bet on Lives: A Guide to Life Contingent Contracts in Early Modern Europe,” Research in Economic History, 10 (1986), pp. 1–53, for comparisons of many forms of contracts, including tontines, using the present discounted-value approach.Google Scholar

³⁶ In 1751 Britain consolidated various 3 percent perpetual annuities into one general stock. The term consol is an abbreviation of the Three per cent. Consolidated Annuities created in that year. To avoid confusion between a historically specific government security and a general type of asset, the term “consol” will be reserved for British 3 percents after 1751.

³⁷ See Vührer, , Dette publique, pp. 191–96, 263–69, for descriptions.Google Scholar They were especially popular in the 1740s and again under Necker. Lotteries were often used in both countries to introduce variance into an average rate of return on any form of loan. In Britain they appear to have been the rule rather than the exception; see Grellier, J. J., The Terms of All the Loans (London, 1805). Although we now think of higher variance as requiring a premium on the expected return, they evidently thought that the lure of a gamble would draw in funds at a lower average rate.Google Scholar

³⁸ Riley, James C., Seven Years War, pp. 174–75, shows that in at least one case the government actually offered to borrow at zero interest while listing the (initial) interest rate at 3 percent.Google Scholar

³⁹ They were known as douceurs in English financial discussions. To my knowledge, no one has speculated on why English financiers chose a French word to describe subsidized interest payments. The British government began issuing life annuities more widely after 1808.

⁴⁰ See Grellier, J. J., The Terms of All the Loans.Google Scholar

⁴¹ The mortality data are reported in Deparcieux, Antoine, Essai sur les probabilités de Ia durée de Ia vie humaine (Paris, 1746).Google Scholar

⁴² Five-year classes were used in the French tontines of 1689, 1696, 1709, 1734, 1743, 1744, and 1745. In 1733 and 1759 the classes were grouped by ten-year intervals, but each class had many subdivisions within which the tontine principle applied. I have assumed that the subdivisions sorted by age. Thus, all the French tontine internal rates of return assume that life table effects cancel out.

⁴³ All the English and Irish tontines were in this category. For the cancelled tontine of 1757, the age distribution of the (very similar) tontine of 1789 was used.

⁴⁴ In 1693 the alternative was a 14 percent life annuity for any age. In 1757 the tontine plan offered an alternative fixed-term annuity without survivorship benefits. The entire plan was then displaced with a mixed offering of life annuities and consols. In 1789 the alternatives was a 4.25 percent annuity for 69 years.

⁴⁵ In 1694 the English government chartered the Bank of England in exchange for a 1.2 million pound loan at 8 percent. The New East India Company charter went for the same terms in 1698.

⁴⁶ Statutes at Large, 32 George II, c. 19, establishes the replacement bill and notes the history of the attempted tontine. Since a 15 percent deposit was all that was required initially, it is unclear whether the 12.5 percent raised by May was a nearly complete deposit or a woefully short total contribution.

⁴⁷ The 1765 tontine was an optional douceur on a larger loan. Again the tontine was priced to pay the same returns as the alternative annuities. Evidently only a fraction of those eligible for the tontine actually subscribed. English financial expertise does not seem to have influenced the Irish Parliament in constructing its three tontines in 1773, 1775, and 1777, nor did the successful French plans based on many age classes. The Irish tontines had a uniform interest rate of 7.5 percent in three age classes: 0–20, 20–40, and 40 and above. Not surprisingly, enrollment was heavily skewed to the youngest class and to the younger ages within classes.

⁴⁸ It is unclear whether this expertise extended fully into the pricing of life annuities. When Britain began to sell life annuities in 1808, it adopted the Northampton life table constructed by the renowned Dr. Richard Price in the 1770s. As subsequent government actuaries showed (A. G. Finlaison, Report and Observation on the Mortality of the Government Life Annuities), the Northampton table greatly overstated mortality for annuity purchasers in the early years of the nineteenth century (it is unclear how much of the difference was due to class, region, or time period). That was good for life insurance company profits, but bad for government debt service. National vanity may well have prevented them from using Deparcieux's tables, which would have been much closer to the mark.

⁴⁹ Deparcieux's Essai sur les probabilités, in addition to setting out the mathematics of present- value calculations for various investments, constructed life tables based on 9,000 nominees in the first two French tontines. Dutch writers had constructed life tables on much smaller samples, and the English astronomer Halley built his on death registers from the town of Breslau in Silesia; hardly a sound basis for the bourgeois investors of London or Paris.

⁵⁰ Riley, , Seven Years War, p. 174.Google Scholar

⁵¹ France issued vast quantities of life annuities in 1717, 1720, 1722, 1723, and 1724. All were at a constant interest rate for all ages. The nominal initial interest rates were low, but capital could be furnished in depreciated billets d'Etat. Vührer, A., Derre publique, p. 185–87.Google Scholar

⁵² Vührer thinks Necker must have known how to calculate the costs of life annuities (Dette publique, p. 273).Google Scholar

⁵³ Necker, Jacques, Oeuvres, vol. 5. p. 491.Google Scholar

⁵⁴ Riley, Seven Years War, based on selling prices of bonds on the Compagnie des Indes.

⁵⁵ Fachan, , Hisrorique de la rente française, p. 63.Google Scholar

⁵⁶ Marion, , Histoire financière, vol. 1, pp. 248–51.Google Scholar

⁵⁶ On Terray's reforms, see Vührer, , Dette publique, pp. 241–51,Google Scholar and Marion, , Histoire financière, vol. 1, pp. 247–79.Google Scholar The other main changes were a reduction of perpetual rents by about II million annually and the conversion of floating short-term paper (rescriptions) into long-term debt at 5 percent with an annual lottery for reimbursement.

⁵⁸ Terray converted tontine rents in which the government's obligation is constant until the last death to life annuity rents in which the rents due each individual end with his or her death. For each of the tontine classes under age 95 in 1770, I projected forward for each year to 1850 the fraction of the total rents to be extinguished due to mortality from its age in 1770. I used a single-year-of-age life table derived from Deparcieux's data to project mortality. Had tontines stayed in force the cumulative total would have been 297 million; as life annuities, 135 million.

⁵⁹ Although they were nominally at 8 percent initial interest, Terray proposed to accept depreciated government paper for half the capital, boosting the effective rate above 10 percent.

⁶⁰ In 1776 Terray's “memoirs” were written by Jean-Jacques Coquereau (Mémoires de l'abbé Terrai) as a virulent attack. They included some of the popular jokes about him circulating in Paris.

⁶¹ Lüthy, Herbert, La banque protestante en France de Ia révocation de l'Edit de Nantes á la Révolution, 2 vols. (Paris, 1959–1961), describes the development of the scheme. Thirty young girls was the typical pool. These were primarily used in the French life annuities after 1770 and not in the earlier tontines (vol. 2, pp. 464–591).Google Scholar

⁶² See Furet, François, Interpreting the French Revolution (Cambridge, 1977),Google Scholar and Hunt, Lynn, Politics, Culture, and Class in the French Revolution (Berkeley, 1984).Google Scholar

⁶³ France. Assemblée Nationale Constituante. Cormité des pensions, Etal Nominatif des Pensions sur le Trésor Royal, 1789.Google Scholar

⁶⁴ In his Dénonciarion de l'agiotage (Paris, 1787). It is one of history's ironies that government profligacy should be condemned by Mirabeau, a man who had been imprisoned by his father, the Physiocratic economist, for nonpayment of debts, and who fell from grace after it was revealed posthumously that he had conspired with the king in exchange for relief of debts.Google Scholar

⁶⁵ Lüthy, , La banque proresranre, vol. 2, p. 561.Google Scholar

⁶⁶ J. F. Bosher, French Finances, 1770–1795: From Business to Bureaucracy.