Corporations in the US and Europe 1790–1860

Abstract

Sylla and Wright's statistics of new US special incorporations in 1790–1860 show that they exceeded those in France, Prussia and the UK, but the aggregate paid-up share capitals of extant companies were not so far apart in 1860. The UK continued to lead corporatisation, as measured by the ratio of corporate share capital to GDP. The distinctive features of US corporations were that they were small, diverse and numerous, while UK corporations were larger, more capital-intensive, less prone to disappear and had more dispersed ownership.

Keywords::

• corporations
• commandites
• companies
• US
• Europe

Notes

^1. Harris, Industrializing.

^2. Such terms universally mean, in their widest connotation, ‘any group of people’. I use ‘corporation’ (in the casual American sense to include only business corporations) or ‘corporation proper’ for emphasis, and ‘company’ equivalently in the similarly casual modern British sense, except where the context makes clear that I am using the broader business connotation of ‘company’ to include partnerships.

^3. For example, 15% of the special charters in the Sylla–Wright database were for manufacturers, compared with only 5% in the UK (Shannon, “First,” 420; Levi, “On Joint Stock,” 24–5) and 8% in Prussia (Engel, Die erwerbstätigen, 10). Although more capital-intensive than US equivalents (Field, “Land Abundance”), many UK manufacturers preferred other forms until the later nineteenth century.

^4.Handlins, Commonwealth.

^5. There were some colonial legislatures, but they were subject to the imperial parliament, and separate Scottish and Irish parliaments had been merged with Westminster in 1707 and 1800 respectively.

^6. Although technically some UK municipalities had chartering powers, in practice they no longer used them in the nineteenth century (unlike some German free cities).

^7. The Companies Clauses Consolidation Act of 1845 and the related clauses acts for railways and compulsory purchase were arguably more important than the 1844 act, because they clearly conferred limited liability and applied to most UK quoted capital until near the end of the century.

^8.Glaeser and Goldin, eds., Corruption.

^9.Hessen, Defense; Anderson and Tollison, “Myth.” For the contrary view of corporations as necessarily politicised, see Alborn, Conceiving Companies.

^10.Angell and Ames, Treatise; Dodd, American Business Corporations, 196. Handlin (“Development,” 7, 11) favours lawyer ignorance in a traditional, agrarian society as the explanation. ‘They had had no experience of how to draw up charters, with what went on within the corporation, or with how to resolve the various problems relating to the contract of the corporation with the state’, contrasting this with ‘England, where they knew how to do it’.

^11.Ostergaard and Smith, “Corporate Governance.” They make no mention of liability rules and the literature on Norwegian nineteenth century enterprise (at least in English) is also largely silent on the matter. Dag Michalsen (email to the author of January 23, 2013), professor of law at the University of Oslo, notes that, although the 1814 constitution envisaged the drafting of a corporate law, none was implemented, so article 5–1–2 of the 1687 law on freedom of contract (similar to Denmark's 1683 law) prevailed and the supreme court accepted legal constructions, unfavourable to third parties, which would not have been accepted elsewhere. Professor Knut Sogner of the Norwegian Business School (BI) points to the example in 1895 of the formation of And. H. Kiær & Co Ltd., presumably expecting their self-limited liability to be respected (email to the author, January 21, 2013).

^12.Dupuichault, Loi norvégienne.

^13.Hannah, “Global Census.”

^14. I say ‘near-absolute’ because one should resist a-historically portraying Norway as a doctrinaire libertarian paradise. Export sawmills needed a government concession to operate before 1860 and the Norges Bank from 1816 had a note issuing monopoly. ‘Discounting commissions’ – effectively state-owned commercial banks – dominated the banking market until the later nineteenth century.

^15. I say ‘near-absolute’ because some American unincorporated ‘business associations’ were similar to English deed-of-settlement companies. Livermore (Early American, 215–42; and see also Burns, “Joint Stock”) provides examples in real estate, banking and manufacturing and, perhaps with too much scholarly contortion and not enough quantification, elevates them above the special incorporations studied by Sylla–Wright as the true fons et origo of America's later general chartering statutes. That case can be more convincingly made for the UK: such associations in the US do not appear to have been as numerous (or as large) as English equivalents, nor as creative in limiting shareholder liability, nor to have provided so direct a blueprint for general legislation (as in UK banking in 1826 and, more generally, in 1844). Sylla–Wright consider the 1720 ‘Bubble Act’ damaged the UK, yet its principle – that only the legislature could authorise the corporate form – was actually more consistently enforced in the early nineteenth-century US than in the UK, for example by judges causing difficulties for unchartered associations and legislatures banning unincorporated banks from note issue.

^16.Lastig, Accomendatio.

^17.Kessler, “Limited.”

^18. Guinnane et al., “Pouvoir.” The Kommanditgesellschaft and the stille Gesellschaft in Germany offered partially limited liability in different ways. As Sylla–Wright note, Louisiana recognised its inherited French limited partnership tradition in 1808, but the first state of the Anglosphere (after Ireland) to do so was New York in 1822. The UK was a longer holdout, though limited partnerships were permitted in Ireland from 1782.

^19.Davis, English Shipping, 103.

^20.Passow, Wirtschaftliche Bedeutung, 3; Jantzen, Freiwillige Veräusserung.

^21.Bartlett, British Mining, 21–37. After 1844 the opening of a registration office in Truro encouraged English companies using the cost-book system subject to the Stannaries Court to convert to the standard corporate form. The Prussian state mines administration also loosened its controls over all mining enterprises (both AGs and Gewerkschaften) from 1851.

^22.Christie, “Scottish;” Campbell, “Law.”

^23. Sir William Clay in the 1843 Select Committee, quoted in Taylor, Creating, 141.

^24. Freeman et al., Shareholder Democracy? Harris (Industrialising) takes a more sceptical view.

^25. See also note 15, above, on the Bubble Act. Sylla–Wright cite the upswing in company formations after its 1825 repeal as an indicator of earlier baneful persecution under the Act, but that upswing arguably (as in the US) stemmed from other causes: new railways, general industrial and urban development, freer banking laws and the growing public taste for tradable corporate securities, with the growth of stock exchanges. This is not to deny that both countries' politicians might have done well to rein in anti-corporation sentiments earlier.

^26. Author's calculation from the Freeman et al. book: available from UK Data Archive, University of Essex at http://discover.ukdataservice.ac.uk/catalogue/?sn = 5622&type = Data%20catalogue, reference SN 5622: R Pearson, “Constructing the Company: Governance and Procedures in British and Irish Joint Stock Companies, 1720–1844.”

^27. There is a preponderance of post-1825 banks and insurance companies (with high nominal capitals only partially paid-up) among deed-of-settlement companies, and four large statutory/chartered companies formed before 1720 are omitted.

^28. As can be seen by their continued presence (e.g. among insurance companies) in Burdett's Official Intelligence.

^29. See Getzler and Macnair, “The Firm as an Entity,” for similar British case law.

^30. Hansmann et al., “Law.”

^31. Livermore, Early American, 258–71, 282–94.

^32. From 1834, deed-of-settlement companies could also apply to the Board of Trade for letters patent indisputably conferring limited liability, but few bothered or succeeded.

^33. For example, some statutory lighthouse companies in the UK were corporations sole. One-man companies were not usually allowed for registered companies (national laws typically specified a minimum between two and ten shareholders) but they could be achieved by registering ‘dummy’ holders. See also pp. 19, above.

^34. Blair, “‘Locking In.”

^35. An early sceptic of the importance of limited liability was Heckscher (Mercantilism, I, 367). See also Acheson et al., “Does Limited Liability Matter?”

^36. By 1860 each of these engineering firms had made more than 1000 locomotives (Lever, Railway, 86). Hawthorns incorporated in 1886, Stephensons in 1899, while the other three remained partnerships into the twentieth century. Probably the largest corporate locomotive manufacturer in 1860 was the LNWR, which had integrated backwards into locomotive manufacture at its enormous Crewe works.

^37. The notion that the UK ‘lagged the international frontier’ (Sylla–Wright, “Corporation Formation,” 10) is simplistic.

^38. Previously, in order to limit liability, insurers had required special acts, letters patent, registration as friendly societies or, in the case of deed-of-settlement insurers, specially devised contractual terms.

^39. Wright, Corporation Nation, 243.

^40. Berle and Means, Modern Corporation, 137.

^41.Thieme, “Statistische Materialien,” 292.

^42.Engel, Die erwerbstätigen juristischen, 10–11, 63, 83. See also Bösselmann, Entwicklung, 201.

^43. The lower figure presumably eliminating merger duplication and depreciation. Fremdling (Eisenbahnen, 28) has higher figures, but this is cumulative capital invested including some financed by bonds.

^44. Bösselmann, Entwicklung, 179, n. 1.

^45. Ibid., 179–82. Thieme (“Statistische Materialien,” 285, n. 3), the source used by Sylla–Wright, seems unaware of Bösselman’s earlier correction of Engel's printing error.

^46. For example the Bank für Handel und Industrie (popularly the Darmstädter Bank) was chartered in the latter city in 1853, having failed to get a Berlin or Frankfurt charter, and opened branches in Prussia and elsewhere using the Kommandit form (Riesser, Die deutschen Grossbanken, 1910, 40, 52). It was listed on the Frankfurt exchange from 1853 and Berlin from 1855 (Gommel et al., Deutsche Börsengeschichte, 146).

^47. In some cases there was de facto recognition, though in the Zollverein and under the Franco-Belgian agreement of 1857 and the Franco-British agreement of 1862 mutual recognition was guaranteed by treaty. Later some states began circumscribing the freedom by requiring foreign corporations to register their existence locally before operating.

^48. Engel, Die erwerbstätigen juristischen, 4. The oldest survivor was probably the Mansfeld'sche Kupferschieferbauende Gewerkschaft in Saxony: one of the larger firms quoted on German stock exchanges, founded around 1300.

^49. There were, for example, in 1860, 471 savings bank branches in Prussia (Bösselmann, Entwicklung, 32, n. 3). Pargendler and Hansmann (“New view”) argue that many early European and American corporations were, in effect, consumer cooperatives, with voting structures nearer those of modern co-operatives (one shareholder one vote) than modern corporations (one share one vote).

^50.Rauchberg, Berufs- und Gewerbezählung, 310 for 1895 data.

^51. I have followed Sylla–Wright in converting £s sterling at a standard $5, thalers at $0.75 and francs at $0.2, ignoring temporary exchange rate fluctuations.

^52.Riesser, Deutschen Grossbanken, 56, 63, 599. The Prussian state retained control over note issue and new entry into Prussian banking was only grudgingly conceded. The average US corporate bank in the Sylla–Wright database had an authorised capital of only $194,122, that in Bodenhorn's database (“Voting Rights,” 47) $179,280; and only a few of these had more than $5 m authorised or paid-in capital (van Fenstermaker, Development). In 1860 the Bank of England had £14.553 m ($73 m) paid-up share capital, the Bank of Ireland I£3 m (about $14 m), the Royal Bank of Scotland £2 m ($10 m), the Oriental Bank £1.26 m ($6 m) and eight other London and Edinburgh banks £1 m ($5 m) paid-up capital each (Evans, Banking, 136ff).

^53. Although Thieme (“Statistische Materialien,” the Sylla–Wright source) lists 16 AG banks formed by 1867, 13 of which survived, these are mainly small, local banks. Compare Schauer, Preussische Bank and Radtke, Preussische Seehandlung. The distinctly-non-defunct Preussische Bank resembled the thrice-defunct Bank of the United States (which only appears in the Sylla–Wright database once on its third – Pennsylvania – re-incarnation: they omit the earlier federal charters).

^54. Schauer, Preussische Bank, 32, 63–4.

^55. Gommel et al., Deutsche Börsengeschichte, 144–6.

^56. Ibid., 147. There were also 115 railway bonds listed in Berlin.

^57. Another issue is whether such prices at par reflect true values. Because of stricter oversight of capitalisation and discouragement of new entry, German share prices often exceeded par (for example, at the 1856 peak 293% above par, Gommel et al, Deutsche Börsengeschichte, 148); US stock prices were often below par.

^58. Thieme (Statistische Materialen, 288) suggests only 62 (21%) of Prussian AGs had abortive formations or failed by 1867, significantly lower than Sylla–Wright's estimate of 50–67% for US corporate failure. However, some other German quasi-corporate forms (particularly KGs) were smaller and probably also had lower survival rates.

^59.Jobert, Entreprises, 106.

^60. A problem in achieving greater precision on this dimension is that an unknown proportion of early US and UK incorporations did not clearly confer limited liability, while most SAs (and all commandites) appear to be limited.

^61. A government survey of extant companies in 1898 (Anon, “Les sociétés”) suggests higher SA survival rates over the period 1808–98 than Sylla–Wright suggest for the US in a shorter period.

^62. The cost for a French SA was £750, with on-going costs of £400–500 annually (Freedeman, Joint Stock, 139). Levi (“Joint Stock,” 14) estimated the cost of a simple special incorporation in the UK at £402, but some turnpikes paid as little as £51, while some banks and railways paid more. In the USA fees, taxes and other imposts were more varied: some were massively higher, many as trivially low as official fees for UK company registrations from 1844 (or lawyers' fees for deed-of-settlement companies earlier).

^63.Neymarck, Aperçus, 172–5; Worms, Sociétés, 136–43 for companies listed on the Bourse in 1856.

^64.Broderick, First Toll Roads.

^65.Return of Partnerships(Ireland), 1863 (BPP, LXVIII, 1863)

^66. On the non-separation of powers, see May, Treatise, 385. This meant that the legislature did not suffer the inconveniences that their supreme courts inflicted on US legislatures (e.g. the Dartmouth College judgment of 1819). Yet US corporations had more frequently suffered from the revocation of their privileges by states than British ones (Lamoreaux, “Scylla,” 17–18).

^67. They would surely hesitate before concluding that the number of stocks quoted in New York (69 in 1825 and 112 in 1840, see Banner, Anglo-American Securities Regulation, 255) was the number of American companies in existence. The 51 local companies traded even in Edinburgh's share market from the mid-1820s had perhaps twice the capital of the 67 New York companies then traded in Wall Street (Michie, Money, Mania and Markets, 39, 44–6; Hilt, “When did,” 675).

^68. They are in high-class company: see North et al, Violence, 219. If one traces the three lists – 156 in 1824 (England's list), 717 in 1843 (Spackman's list, covering more non-London securities than England's) and 947 in 1844 (the Registrar's list) – back to their sources and compares the names of the firms in each and in the new list available in the 1720–1844 database for the Freeman et al. book, it is striking how little overlap there is. It is likely that all four lists underestimate the number of companies by between two-thirds and nine-tenths. The lack of overlap suggests that they are all samples of a population of firms in existence sometime between 1720 and 1844 that is even larger than the total of at least 1500 estimated by Freeman et al (Shareholder Democracies?, 15–16). Moreover, their sample explicitly excludes companies formed before 1720, those with fewer than 13 shareholders or non-transferable shares, and turnpikes, in which four categories alone there were a greater number than 1500. Their count, being based on surviving records, is also likely to underestimate the many short-lived and stillborn companies included in the US data: for example, their sample covers only 50 (8%) of the 624 they note from another (itself only partially complete) source as formed in 1824–25 (p. 31), most of which were abortive or soon failed. If these lists were all independently drawn random samples of the same population, we could estimate the total population reliably from the degree of overlap, but as three of the four (i.e. all except Freeman et al.) were not even nearly representative samples, that approach would yield a seriously downward-biased estimate. In the (exceptional) case of turnpikes we know that there were only seven named in any of these lists (the few quoted on the London Stock Exchange), while parliament had authorised well over 1000.

^69. These are based on reasonably accurate counts of private acts (arbitrarily discounted for duplications) and company registrations, with an estimate for other forms based on the lists in note 68, above.

^70. Levi, “On Joint Stock,” 11.

^71. Ibid.; Burdett, Official Intelligence 1882, ix.

^72. The UK distinction between statutory and registered companies parallels what Sylla–Wright describe as special versus general acts in the US: statutory companies required individual parliamentary approval (or, latterly, a provisional order from the executive, which only took effect if no member of the legislature objected). In 1845–60 2300 companies were registered under UK general acts (Levi, “On Joint Stock”) and there were 2861 private acts (Williams, Historical, 59, 125–6), not all of which resulted in new corporations. The numbers chartered individually by letters patent etc. were smaller.

^73. See Burdett (Official Intelligence 1884, cxxiv) for the relationship, disaggregated by sector, of authorised and paid-up capitals in 1853, 1863, 1873 and 1883 for officially listed companies (most statutory not registered).

^74.Shannon, “Coming;” idem, “First Five Thousand;” idem, “Limited Companies.” Shannon's statistics are on numbers of failures; since large firms were less likely to fail than small ones, his loss ratios exaggerate the capital losses from failures.

^75. When the IRS in 1909 examined state registers in order to administer the new federal corporate excise tax, they found tens of thousands of non-existent companies registered in some states.

^76. Levi (“On joint stock”) for registered companies after 1844. There is some information earlier for individual sectors. For example, Wright (2014) notes 314 insurance companies chartered in America in 1790–1830, probably a larger number than in the UK (compare Walford, “On Fires,” 394–6; Andras, Historical Review, 101ff), but there appear to have been more large and surviving UK offices (see Stalson, Marketing, 717–9, 748–53 for life offices).

^77. Author's estimates based on the flow of new companies and the stock existing in the early twentieth century. Of course, bankruptcy did not lead to the social loss of all capital: some was recycled to corporate or other businesses. Countries differed in their tolerance of such churn rates: following liberalisation in 1870, Germany briefly experienced US corporate failure rates during its Gründerboom, resulting in a moral panic and progressive re-regulation of the company formation process.

^78. This is less likely to be a work of fiction, though is not entirely devoid of problems. Some American ‘paid-up’ common stock was issued below par or even without any payment being made, as a ‘bonus’ for subscribers to preferred stock. The monitoring of shares issued in payment for existing assets (rather than for cash) appears to have been tighter in France and Germany than in the UK or US. On the other hand, paid-up capital in some cases understates the resources available to firms: for example, some US banks had double liability and for some UK banks and insurers with subscribed capital only partly paid-up, the capital authorised may better represent their capital backing, typically at this time quadruple the amount paid-up (Burdett, Official Intelligence for1884, cxxiv).

^79. Gallman's estimates of domestic capital formation in 1860 prices, from Historical Statistics of the United States. This is not defined in the same way as corporate capital (which may, for example, additionally include promoter's profits, the purchase of land or simple ‘watering’ of stock), but there is a good deal of overlap.

^80.Gatty, Portrait, 241–3.

^81. It is a reasonable simplifying assumption that all railway investment was corporate (and accounted for 15% of US investment outlays in 1849–58, Fishlow, American Railroads, 101f) and none in agriculture and housing. We know the portion in banking by mid-century and Schwartz (“Gross Dividend,” 428) estimates that nearly a third of the manufacturing capital stock ($311 m), 50% of mining capital ($32 m), and all gas-light ($27 m), canal ($42 m) and street railway ($13 m) capital was in corporations in 1859.

^82. Common and preferred stocks only, excluding bonds (in order to match the Sylla–Wright database which is for stocks alone).

^83.HSUS gives Poor's figure for total railway capital as $1,149 m (which is 26% of GDP). For an exhaustive discussion of the capital invested in railroads see Fishlow (American Railroads, 341–401): he increases Poor's figure for the cumulative cost of all capital invested to the end of 1860 by only $2 m (p. 358). If the proportion in bonds was 37.3% (the average of that known for 1855 and 1867), Poor's figures suggest $720.4 m in corporate stock. Taylor and Neu (American Railroad) count 210 extant railroads in 1860, but exclude short lines; I estimate there were 90 of the latter, making 300 in all.

^84. Robert Wright tells me the later rail figures are distorted by a number of very large, but abortive, trans-continental railroads. Robert Wright, email to the author 29 February 2012.

^85.Weber (“Early State Banks”) reports 1345 incorporated banks in existence at the end of 1860. The slightly higher figures in HSUS (1562 with $422 m capital) presumably include some private banks. The figure I have used for capital is given for 1396 banks in March 1861 in Secretary, Banks, but Jaremski and Rousseau (“Banks,” 37), relying on contemporary directories, report a higher capital figure (without defining whether that includes reserves as well as share capital) of $436 m for 1144 extant chartered banks in 1860 and $101 m more for 512 ‘free’ banks (i.e. those not individually chartered but allowed under general (‘free banking’) legislation and sometimes referred to as ‘private’ banks).

^86. ‘free’ banks (those not individually chartered) had proliferated in the 1850s, though, as the previous footnote suggests, were generally smaller and less clearly corporate.

^87.Wright (Corporation Nation, 241–2) notes $191 m minimum authorised capital in general charters by 1860, with data for some states missing; however, for present purposes banks (and perhaps a few railways) need to be deducted and, as the average size was smaller than for special charters, they were probably more likely to fail.

^88.Secretary, Report, 53.

^89. $600 m is 32% of our estimate for total corporate capital. The sectors we have fully covered (rail and banks) accounted for 48.3% of corporate net earnings in New York state in 1867 and the ones partially covered by our $150 m figure (insurance, canals and gas) another 42.7%, leaving as omitted only 0.7% of earnings in express companies, waterworks, turnpikes and bridges and 8.3% in miscellaneous (including manufacturing) (Schwartz, “Gross Dividend,” 412, n. 17).

^90. Ibid., 417.

^91. 1859 is more representative of normal expectations than the (higher) 1860 rate, which was affected by international investor perceptions that America was about to descend into civil war.

^92. Schwert's index (HSUS, Cj808) shows the dividend yield on US traded corporate stocks as 5.2% in 1859.

^93.Hawke and Reed, “Railway capital,” 271.

^94. As in the US, we omit railway bond and loan capital, which would add £81.889 m ($409 m), bringing total rail capital to £348.130 m (43% of GDP, compared to 26% including rail bonds in the US).

^95. Field, “Land Abundance,” 409.

^96.Casson, World's first railway system; Foreman-Peck and Millward, Public and Private Ownership, 13–28. On the other hand, Smith (“Longest”) reviews contemporary concerns of collateral damage from central rail planning mistakes in France.

^97.Poor, History, 226, 421, 580. More than half of all railway construction expenditure to 1845 in America, Belgium, France, Germany and the UK combined was in the UK alone (Hobson, Export, 116–17).

^98. The 1861 Almanac shows £43 m paid-up capital (excluding large accumulated reserves) in 104 UK joint-stock banks, plus £4 m in London's 15 colonial banks with sterling capital in November 1860 (author's calculation). Thirteen more joint-stock banks are listed, with capital unknown, but were small, with perhaps no more than £1 m capital in total. Over a third of this £48 m capital (in the central bank and some others taking advantage of special charters or post-1858 registration) had fully limited liability, while others had fixed liabilities on capital subscribed but not paid-up, specified reserve liabilities, or completely unlimited liability. All these figures exclude the capital of private bankers.

^99.Collins (Money, 102) for rapidly declining bank capital-asset ratios in the UK.

^100.Economist, August 4, 1860, 842.

^101. Secretary, Banks, 168, 306.

^102. Specialist bill brokers are also excluded from the UK commercial bank statistics. Both they and merchant banks in 1860 were largely unincorporated (Overend Gurney – infamously – only became a limited company in 1865).

^103. Burdett reports £60.3 m of paid-up capital at par in non-rail, non-bank listed securities in January 1853, rising to £67.5 m in January 1863. This includes some foreign companies and all corporate bonds (though at this time the totals for both were small outside the rail sector) and preference shares. Hickson et al. (“Rate of Return”) estimate the market value of domestic non-rail equities (excluding preferences, debentures and infrequently-traded shares) on the LSE as £62.6 m in 1853, £55.3 m in 1860 and £71.1 m in 1863, figures which – by comparison with Burdett – suggest non-rail equities were generally above par. They also report 322 equity securities (200 of them non-railway and non-bank) of 266 companies (banks and railways not separately enumerated) officially listed on the LSE in 1860. The number of such companies listed would be lower than 200 if some issued two types of equity and higher if some only issued preference shares or debentures. Such securities, though common in railways, were still relatively rare elsewhere.

^104.Burdett (Official intelligence 1884) later listed more than 2000 UK companies with traded securities, a substantial majority not officially listed in London and there were many more infrequently-traded companies. Of course, provincial finance of canals and turnpikes had been the norm even before local stock exchanges existed: formal provincial exchanges followed the much larger sums involved in railway finance.

^105. In 1864 the par value of the paid-up capital of the 2,479 limited liability companies formed under the 1856 act by 1862 was £31.31 m, of which 16.5% was then in companies dissolved or being wound up (Shannon, “Coming,” 379) and 512 companies formed in 1861–62 need to be deducted from this Figure (Shannon, “First Five Thousand,” 421), while 966 need to be added for companies registered under the 1844 Act, which were probably larger. Levi (“On Joint Stock”) gives a figure of £96 m for the ‘nominal’ (authorised) capital of the 1655 new companies formed under the 1856 Act in 1856–60, but paid-up capital would have been less.

^106. As public capital markets were probably less centralised and developed in America a larger multiplier there might be appropriate.

^107. Harris, Industrializing, 195, following Harris's convention that England's GDP was 80% of the UK's. Sylla–Wright cite Harris's data, not noting how close his total is to their figures for the US flow in all special incorporations in 1790–1843, which surely overstates the extant US stock at the latter date. They suggest deducting from Harris's figure the capital of three ‘old, moneyed companies’, presumably the Bank of England ($54.6 m), South Sea Company ($18.3 m) and East India Company ($30 m) – without explaining what they have against old money – but this would only reduce the ratio slightly, to 51%.

^108. Privately funded turnpike trusts had taken over 17% of the road network from parishes by the 1830s, though there was then a gradual drift back to local government control (Bogart, “Did turnpike trusts,” 440). In Ireland, where the system had peaked at 1,300 miles, there were only 325 miles left in 1858 when turnpike trusts were abolished and the roads reverted to grand jury control (Broderick, First Toll Roads, 240, 272).

^109. This was when the registrar began publishing annual figures for the surviving stock of registered companies and we have data on the accumulated capital stock of statutory companies (Clifford, History, 266, 492), leaving only royally chartered and remaining unauthorised companies to be estimated. My estimate for UK corporate share capital for that year is 110% of GDP (148% including bonds).

^110. Hannah (“Global census”) suggests that, though at par US corporate share capital (173% of GDP) had overtaken the UK‘s ratio (162%), at market prices the UK remained ahead. All our calculations for earlier years are in par values. Casual inspection of 1860 stock prices suggests that they were a little below par in the UK and further below par in the US, so it is possible that our analysis at par flatters the US. Fishlow (American Railroads, 354) notes that a lot of railway stock had been sold at less than par in the 1850s. Hilt (“When did”) also shows NY market prices below book values in 1826.

^111.Frère (Etudes, I: 128) suggests a Belgian level in the 312 extant SAs at the time of general incorporation in 1873 of F1,271 m (21.3% of GDP), though this excludes commandites and many (though not, before the 1870s, the majority of) Belgian railways which were state-owned.

^112. Competition in transport and capital markets (and some regulation of tolls and fares) led to quite low investor returns on capital invested in UK transport: around 4% in turnpikes, 3.5–4.5% in railways and 6% in canals (Bogart, “Transport Revolution,” 23).

^113.Cottrell, Finance, 75, 105; Jefferys, Business, 57–8, 75, 105, 118–9, 142; Hannah, Rise, 19; Alborn, Conceiving Companies, 129, 203–4; Forbes, “Limited Liability”; Smart, “On Limited Liability”; Johnson, Making, 122–6.

^114. Davis and North, Institutional Change, 138n. They were presumably misled by the 1855 legislation making limited liability available by simple registration, not realising that other means of limiting liability were widely used earlier.

^115.Ferguson, Great Degeneration, 93.

^116. Harris, Industrializing, 195, 222. GDP figures from www.measuringworth.com, following Harris’s convention that England was 80% of Britain.

^117.Kocka, “Entrepreneurs,” 538–9.

^118. Email to the author February 14, 2013.

^119. This suggests my conjectural Prussian estimates for an approximately 1200 m thaler (M3600 m) flow of new incorporations before 1870 in Prussia (which had slightly more than half Germany's population) may perhaps be too high, or perhaps a large portion was lost or absorbed in later AGs.

^120.Van der Borght, Statistische Studien, 217. I have added M120 m to his (and other) totals to allow for the Reichsbank (an investor-owned utility, which had replaced the Preussische Bank in 1875). Wagon, Finanzielle Entwicklung,. 4 gives a figure of 1311 AGs with M3918.7 m capital for 1883, which suggests Borght omitted many small AGs not publishing balance sheets, plus the effect of further rail nationalisation; GDP from the Groningen Growth Project website, assuming 1880 prices were 84% of 1913's.

^121. Hannah, “Global Census.”

^122. Notably the Hanse cities. For experience of Napoleonic occupation driving faster German trade development and encouraging investment in corporate (rather than – less trade-promoting – government) railways, see Keller and Shiue, “Link”.

^123. Author's calculation of 45% (par) and 49% (market) of GDP from Anon, “Les Sociétés,” with the addition of the Banque de France (Théry, Valeurs, 97). Bonds and loans are excluded (to facilitate comparison with the US and UK figures for share capital alone), though French companies (especially railways) had unusually high leverage: including bonds would more than double the French 1898 ratios. No figures are given for commandites simples.

^124. Securities quoted on the Paris bourse in 1859 totalled only F8980 m (Hautcoeur and Gallais-Hamonno, Marché, 227) and around three-fifths of these were corporate and government bonds, leaving about F3592 m for corporate equities (20% of 1859 GDP). There were also provincial bourses and unquoted shares so this would be a minimum.

^125. However, because the French state guaranteed railway bonds, French railways were heavily leveraged, so their share capital was quite small.

^126. The best data are for physical measures (though it should be borne in mind that railway building was generally more expensive in Europe; compare Table 3): in 1860 the US had 49,292 km of railway track (and more sparsely populated Canada a further modest 3359 km) and Europe 51,862 km (of which the UK had 16,787 km, and more sparsely populated Germany 11,633 km, France 9528 km and Austria-Hungary 4543 km and the rest 9371 km), with only small networks on other continents (their longest national systems in 1860 being India's with 1350 km and Cuba's with 604 km), see Woytinsky, Welt, V: 34–7. By physical measures, the UK also led continental Europe in road and waterway investments (Bogart et al., “State,” 87–94; Bogart, “Transport,” 11).

^127.Picard, Chemins de fer, 16. This appears to be cumulative expenditure, including that financed by bonds. So – if we include railway bonds – the French level of 22% of GDP is a little below the US level of 26% and well below the British level of 43%.

^128.von Schreiber, Die preussischen, 87 gives lower figure of 352 m thalers; I use Fremdling's (Eisenbahnen, 28, 184–5) upward adjustment, which includes some state-controlled railways (which mainly issued bonds). US and UK GDPs at current prices from www.measuringworth.com. France F20,684 m GDP in 1860 at current prices from the Groningen website. It is possible that German railways were more highly valued by stock exchanges (see e.g. the figures in Engel, Die erwerbstätige) than US and UK ones and correcting for this would bring the Prussian figures nearer British levels, though it is a moot point how to interpret this (the par value may be nearer to actual cost and the high stock market values might show the lower levels of competition in Germany which followed from its lower investment in competing railways than the US and UK).

^129. The Sylla–Wright estimates for the flow of authorised capital refer to shares, so I focused on these for comparative purposes in the text. The US's somewhat higher leverage derives from foreign investors' requirements, given the difficulty of absentees influencing railway governance and dividend policy and the clarity of the bond interest contract. Following the French commercial panic of 1857, the government agreed to new guarantees of the interest on railway bonds in 1859 and within a decade French railways (which had previously funded two-thirds of their capital with shares), overwhelmingly relied for their expansion on bonds which were almost as highly rated as government rentes (in 1856 only 32% of rail capital was in bonds; by December 1869, the capital of French railways on the Paris bourse was 72% in bonds at par, 75% at market, Congrès International, Rapport, vol. 2, 9).

^130. Bogart, “Transport,” 13, 22.

^131. Mitchell (International Historical Statistics) gives the share of GDP generated in agriculture around 1860 as 45% in Germany, 40% in France and 18% in the UK; for the US his earliest figure is 21% for 1869. The share of the US labour force in agriculture was much the same at both dates (over 50%), so the US was probably nearer to the UK in 1860 than it was to the continental European agrarian states. It is debateable whether such an adjustment is appropriate, for the UK did not consume less agricultural produce, but rather procured its beaver skins from Hudson's Bay, tea from Bombay, wheat from Stettin and Odessa, wine from Bordeaux, cotton from New Orleans, rice from Charleston, sugar from Jamaica, and so on, through (substantially British-owned) supply chains (including trading companies, commodity exchanges, banks, insurers, ships, railways and ports) that were more capital-intensive and more corporation-intensive.

^132. Casual inspection of newspapers and share indexes suggests that 1860 was generally a low point for share prices, with American stock prices below par and continental European ones above and the UK somewhere in between.

^133. This is contested terrain. I follow Lindert and Williamson (“American Incomes”) rather than Maddison on US living standards.

^134.Maddison (World Economy, 436–7) puts the French GDP per head in real terms in 1860 at 67% of the UK's, Prussia's at 58%.

^135. Based on their stated capital values for 22,419 companies in the US, 717 in the UK, 265 in Prussia and 642 in France.

^136. As is suggested by the first figure for the extant stock of German AGs in 1883, which gave an average of M2,989,069 ($747,267) paid-up capital. Before free incorporation in 1870 and the nationalisation of major railways (the largest contemporary companies) in the previous five years, the mean would presumably have been larger, especially if we exclude turnpikes.

^137. Sylla–Wright kindly facilitated this comparison by providing this series omitting turnpikes (which are excluded from the British data and were generally small) and for appropriate dates (email from Robert Wright, February 29, 2012).

^138. Author's calculation from the Freeman et al. database. The non-statutory companies in the database were actually larger than statutory ones, especially after 1825, while some US general acts (such as New York's for manufacturing) limited incorporation by registration to those below $100,000, so, to this extent, the British lead in corporate size will be understated. However, this is still not conclusive because Sylla–Wright is a full population, while Freeman et al. is a sample, and the survival of archival and published references from which they draw it may be biased to larger companies.

^139. Author's calculation from the data in HSUS and Hunt, Development, 146, based on 469 companies in Massachusetts, 52 in New Jersey and 3503 in England, though the capital actually paid-up was less than that nominally registered (see n. 73, above). Most of the UK companies were not offered to the public: for the 876 that were, the mean size of their capital was higher: £426,062 authorised and £306,115 offered. In Massachusetts the mean capital of new charters was only $51,225 in 1851, peaked in 1864 at $252,172, but by 1889 had fallen to $45,705 (Falkner, “Statistics,” 60–61). Some UK registered companies were also surprisingly small: the average size of the 2479 UK companies registered between 1856 and 1862 was only £12,630 paid-up (Shannon, “Coming,” 379).

^140. English, Complete View. Excluding insurance companies (with more than £25 m capital) whose numbers are not given by Spackman (Morgan and Thomas, Stock Exchange, 279).

^141.English, Complete View, 31; Hilt, “When did,” 663. Both are biased to the major commercial centre and so may overstate the national average, but English's data are more seriously biased by including only companies with traded or publicly offered shares.

^142.Freedeman, Joint Stock, 82 for France; earlier text figures for Prussia, UK and US, estimating there were 300 US railroads and 100 in the UK in 1860. Sylla–Wright's data for railroad incorporations 1825–60 suggest a lower mean US railway size below $0.9 m. Hickson et al. (“Rate of return”) report the average LSE-traded rail security 1825–70 had an even higher market value of £3.61 m (an underestimate since some railroads issued more than one equity security).

^143. Dobbin, Forging, 25.

^144. Equity market capitalisations from Hickson et al. (“Rate of Return,” 37). In the US Treasury list of 1853 the largest gas company, Manhattan Gas Lighting, had $1.3 m paid-in share capital (of $2 m authorised) and the Chesapeake & Ohio Canal $8.2 m paid-in capital (the Erie was state-owned and financed by bonds). In the 1860s things began to change. Western Union (the 1865 merger of Morse systems, with the equivalent of £8 m share capital and £1 m bonds, see Economist, October 2, 1869, 1164) was bigger than the largest British cable company (Submarine Cable, which had laid the cross-channel cable in 1853, capitalised at £6.6 m in 1870), though the UK government had nationalised all domestic telegraph corporations in 1868–70, so only international traffic was open to the private sector in Britain (and most of Europe).

^145. See n. 85, above.

^146. Freedeman (Joint Stock, 82, 118) shows bank and insurance SAs formed in France in 1847–59 averaged F4,794,340 ($958,868) capital and banks formed in 1860–66 averaged F34,181,1818 ($6,836,364). On other banks around 1860, compare n. 53, above. The average bank traded on the LSE in 1825–70 (Hickson et al., “Rate of Return,” excluding the Bank of England) had an equity market capitalisation of £640,000 ($3.2 m).

^147. Hunt, Development, 150; Engel, Die erwerbstätigen, 10; Freedeman, Joint Stock, 82. The capital actually offered to the public by the UK companies was lower than that authorised at £229,339 ($1,146,694) per company; we do not have equivalent figures for the other countries. Even if we assume that British manufacturers incorporating but not offering shares to the public (i.e. that remained private, close companies) numbered three times as many and each had only £1 capital, the average British manufacturing corporation remains larger. Moreover, the manufacturing firms under US general acts were probably smaller than the Sylla–Wright average: in New York, for example, the general act of 1811 did not permit incorporations with more than $100,000 capital.

^148. Joseph Whitworth in Rosenberg, ed., American System, 338.

^149. For the 1880s see Yonekawa (“Growth”, 4); the US did not catch up until around 1913 (ibid., 10).

^150. Hilt, “When did,” 664. This may overstate shareholder dispersion: shareholder lists could be found only for a minority of companies and it is unclear whether survival is size-related.

^151. Hilt, “When did,” 663–4.

^152. London companies can be traced in the twice-weekly Course of the Exchange and later, more comprehensively, in Burdett's Official Intelligence. US companies can be traced in local newspapers, from which Sylla is assembling an impressive additional database.

^153.Evans, British Corporation Finance, 26; Thomas, Stock Exchanges, 140.

^154.Anon, Return.

^155.Neymarck, Ce qu'on appelle, 8. This understates the totals by excluding bearer shares.

^156. It was probably not until the early twentieth century that the number of stock-exchange traded companies' shareholders first exceeded a million: about 1% of the US and 2% of the UK population (Hannah and Rutterford, “When did”).

^157. Yonekawa, “Growth,” 26: Toms, “Producer co-operatives.” On the other hand, there is evidence of quite widespread shareholding in (mainly unlisted) companies in Pennsylvania, Massachusetts and elsewhere in the early days of the republic, and much of that continued in local US banks and utilities.

^158. E.g. Wright, “Bank Ownership.”

^159.First Annual Report of the Directors of the Pennsylvania Railroad Company, 10. This hardly matched the reported 24,000 subscribers in Lyons alone to a French railway issue in 1844 (Colling, Bourse, 234)

^160. originally municipally owned by the city of Troy, which had sold it to one of the merger negotiators in 1852.

^161.Stevens, Beginnings, 352.

^162.Foreman-Peck and Hannah, “Extreme divorce.”

^163. North, Wallis, and Weingast, Violence; Acemoglu and Robinson, Why; Hoffman. Postel-Vinay, and Rosenthal, Surviving; Wright, Corporation Nation.

^164. Smith, “Longest.”

^165. For a persuasive rebuttal of the persistent strand in the literature projecting back modern relative capital–output and capital–labour ratios onto nineteenth-century US and UK business, see Field, “Land Abundance.”

^166. Canada, Australasia, Singapore and Hong Kong are other candidates.

^167. Johnson, Making; Wright, Corporation; Mayer, Firm.

An earlier version has benefited from suggestions from Kristine Bruland, Carsten Burhop, Eric Hilt, Robin Pearson, Richard Sylla, John Wallis, Robert Wright and two anonymous referees. Responsibility for remaining errors is mine.

Additional information

Notes on contributors

Les Hannah

Leslie Hannah is Professor of Economic History at the London School of Economics and a frequent contributor to this journal.