Hostname: page-component-78c5997874-v9fdk Total loading time: 0 Render date: 2024-11-03T08:59:55.045Z Has data issue: false hasContentIssue false
  • English
  • Français

On the Profitability of Russian Serfdom

Published online by Cambridge University Press:  03 March 2009

Evsey D. Domar  and
Mark J. Machina
Get access Rights & Permissions [Opens in a new window]

Abstract

The paper examines the thesis, popular among Russian Marxists, that Russian serfdom had become unprofitable for the serfowners before the emancipation of the serfs in 1861. Four theoretical models are constructed in order to determine the effects on serfdom of population growth, rise in grain prices, certain restrictions on the serfs' labor obligations, and the replacement of labor services with money payments. Prices of serfs for the several regions and provinces are estimated by regression.With the exception of Lithuania, neither the theoreticalnor the empirical results confirm the Marxist hypothesis.

Type
Articles
Information
The Journal of Economic History , Volume 44 , Issue 4 , December 1984 , pp. 919 - 955
Copyright
Copyright © The Economic History Association 1984

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Gerschenkron, Alexander, “Agrarian Policies and Industrialization: Russia 1861–1917,” The Cambridge Economic History of Europe (Cambridge, 1965), vol. 6, part 2, pp. 706–800.Google ScholarBlum, Jerome, Lord and Peasant in Russia from the Ninth to the Nineteenth Century (Princeton, 1961), pp. 612–18 (Blum also lists fear of serf revolt among the causes [pp. 552, 616–17]).Google ScholarField, Daniel, The End of Serfdom: Nobility and Bureaucracy in Russia, 1855–1861 (Cambridge, Massachusetts, 1976), pp. 96–101.Google ScholarRieber, Alfred J., The Politics of Autocracy (Paris, 1966), pp. 15–58.Google Scholar The general crisis of serfdom is a well-accepted Marxist doctrine. See, for instance, Liashchenko, P. I., Istoriia narodnogo khoziaistva SSSR [History of the peoples' economy of the USSR] (Moscow, 1956), vol. 1, pp. 467–510;Google Scholarchenko, I. D. Koval', Russkoe krepostnoe krest'ianstvo v pervoi polovine XIX veka [Russian serf peasantry in the first half of the nineteenth century] (Moscow, 1967), particularly pp. 378–85;Google ScholarMaslov, P., Agrarnyi vopros v Rossii [The agrarian problem in Russia], 4th ed. (St. Petersburg, 1908), vol. 1, pp. 389–91;Google ScholarFedorov, V. A., Pomeshchich'i kresrt'iane tsentral'nogo promyshlennogo raiona Rossii kontsa XVIII—pervoi poloviny XIX veka [Landowners' peasants of the Central Industrial Region of Russia, end of the eighteenth-first half of the nineteenth centuries] (Moscow, 1974), pp. 256–57;Google Scholar and Pokrovskii, M. N., Russkaia isrorlia s drevneishikh vremen [Russian history from ancient times] (Moscow, 1934), vol. 4, pp. 40–84. Pokrovskii was not the only one to claim that serfdom had become unprofitable for the masters. This view was shared by the other Marxists cited here and by a number of non-Marxist writers. See Blum, Lord, pp. 563–64.Google Scholar

2 In actual fact, a master wishing to free his serfs had to comply with a number of complicated government regulations.Google Scholar

3 The assumption that serfs did not work harder on their own land might appear strange. But data collected by Koval' chenko, Russkoe [Russian], p. 57,Google Scholar for six provinces between 1842 and 1860 show practically equal harvest/seed ratios on both parts of estates. Perhaps the masters appropriated the more productive or more accessible land. Tolstoi's description of how Nikolai Rostov, a model landowner, supervised his serfs on both parts of his estate does not suggest that the serfs displayed any particular ardor when they worked for themselves (War and Peace, translated by Wiener, Leo [New York, 1968], vol. 4, pp. 370–74). If the serfs had really worked much harder on their own land, it would have paid their masters to use the obrok system more frequently than they did (see the Fourth Model). Koval'chenko, Russkoe [Russian], p. 75, confirms that no technological progress was taking place. No distinction is made in this paper between arabic, meadows, waste, or forest.Google Scholar

4 The fact that the serfs worked much harder in summer than they did in winter is disregarded.Google Scholar

5 Actually, as Blum describes, there were considerable wealth and income differentials among serfs. See Lord, pp. 469–74.Google Scholar

6 The assumption of income maximization by the masters probably implies a greater degree of rationality than actually existed. From all accounts, it is clear that Russian serfowners were much less efficient than were American slaveowners.Google Scholar

7 Assuming a given number of serfs on the estate and a constant subsistence level, the maximization of the master's income from the demesne is equivalent to the maximization of total production on the estate. In other words, the estate can be treated like a slave plantation where the slaves work for the master full time and are provided with their subsistence. Although analytically simple, this approach is less satisfactory as a means toward understanding the nature of serfdom and would prevent the use of this model as an introduction to the following ones.Google Scholar

8 Let the number of serfs on a given estate increase by 1 percent. If all of the land for the new arrival is taken from other serfs, the allotment of each will decline by (approximately) 1 percent and their free time will have to be increased accordingly in order to maintain their subsistence. The labor/land ratio on the serf land will then increase by more than 1 percent, but on the demesne it will increase by less than 1 percent, violating the efficiency condition of equal land/labor ratios on both parts of the estate. To restore this equality, the master will have to contribute some of his own land to the new serf's allotment. I. F. Samarin, a liberal landowner highly respected as an authority on serfdom, states that an increase in the number of households (tiagla) on the estate always caused the master to contribute some of his own land to their allotments. He adds that a master who refused to do so and yet expected to receive more labor from his serfs would be condemned by public opinion as a violator of the serfs' rights.Google Scholar See his Sochineniia [Collected works] (Moscow, 1878), vol. 3, p. 205.Google Scholar

9 Blum, Jerome, The End of the Old Order in Rural Europe (Princeton, 1978), pp. 50–59. There was much local variation, however.CrossRefGoogle Scholar

10 See Koval' chenko, Russkoe [Russian], pp. 263–81; Fedorov, Pomeshchich'i [Landowners], pp. 19–25;Google Scholar and Semevskii, V. I., Krest'iane v tsarstvovanie imperatritsy Ekateriny 11 [Peasants during the reign of the Empress Catherine II] (St. Petersburg, 1903), pp. 1–100. Unrelated division of land, labor, and output between the masters and serfs in the eighteenth century reported by these authors did not induce them to deepen their investigations.Google Scholar

11 However, there is some evidence that in the eighteen century the labor services demanded from serfs were very heavy indeed. According to Kliuchevskii, General Peter Panin, having described the intensive exploitation of serfs to Catherine II, suggested to her that their labor obligations be limited to four days a week. See Kliuchevskii, V., Kurs russkoi istorii [A course in Russian history] (Moscow, 1937), vol. 5, pp. 128–29, 146–47.Google Scholar

12 Skrebitskii, Aleksandr, Krest'ianskoe delo v tsarsivovanie Aleksandra II: materialy dlia istorii osvobozhdeniia krest'ian [Peasant affairs during the reign of Alexander II: materials on the history of the emancipation of the peasants] (Bonn, 1865/1966), vol. 2, part 2, pp. 1491–551, and vol. 3, pp. 1227–93. Actually, additional payments in kind were often required.Google Scholar

13 Ibid.

14 If the Cobb-Douglas production function is used, then TsRt, being that part of the serf's income attributed to land, remains constant because in this production function the relative shares of the factors remain constant and the serf's income is assumed to be constant.Google Scholar

15 Long before that happens, an old serfowner, observing that the productivity of an hour of serf labor had declined while its implicit cost to him had risen as the serf population increased, might forget about the increasing supply of labor available to him and bemoan the falling profitability of serfdom. Similarly, in a well-organized slave market, the master, operating on the margin, will profess his indifference between free and slave labor long before the marginal productivity of slave labor comes down to subsistence.Google Scholar See Domar, E. D., “The Causes of Slavery and Serfdom: A Hypothesis,” this JOURNAL, 30 (03 1970), 13–32.Google Scholar

16 Serfdom may also persist if the master's status is determined by the number of serfs he owns. Many such cases can be found in the Russian pre-Emancipation literature. It is also possible that unfree workers (slave or serf) may be more reliable than free workers.Google Scholar

17 Pokrovskii, Russkaia [Russian], pp. 40–55.Google Scholar The liberal official was A. P. Zablotskii-Desiatovskii, sent by Count P. D. Kiselev to make a firsthand investigation of peasant conditions. This report was included in his major work, Graf Kiselev i ego vremia [Count Kiselev and his time] (St. Petersburg, 1882), vol. 4, pp. 271–345.Google ScholarActually, the behavior of grain prices, as cited by Pokrovskii, hardly supported his contention. But a more serious recent study made by two Soviet historians does indicate a considerable rise in prices of rye and oats in most provinces during the 1846–1855 (or 1847–1856) period. See Pokrovskii, Russkaia [Russian], vol. 4, pp. 44–46, 53;Google Scholar and Koval'chenko, I. D. and Milov, L. V., Vserossiiskii agrarnyi rynok XVIII—nachalo XX veka [The All-Russian agricultural market, eighteenth—beginning of the twentieth century] (Moscow, 1974), pp. 394–97. But did agricultural prices rise relative to other prices?Google Scholar

18 We are concerned here with that part of Pokrovskii's argument that is of greatest theoretical interest. We should add that he expected the rise in grain prices not only to make serfdom unprofitable but also to lead to a general reorganization of the serf estates into capitalist enterprises. The serfowners expected to obtain the capital needed for this purpose from the redemption of their serfs by the government on generous terms (pp. 55–84). That such a redemption was in fact carried out after 1861 makes his argument suspect. Pokrovskii's style is sharp, witty, and partisan. We doubt that his opinion on the profitability of Russian serfdom would have changed had grain prices fallen instead of risen. Gerschenkron, “Agrarian,” p. 726, also spoke about the “sinking profitability of the [serf] estates” without presenting any evidence.Google Scholar

19 The two economists were Stolper, Wolfgang F. and Samuelson, Paul A.. See their “Protection and Real Wages,” Review of Economic Studies, 9 (11 1941), 58–73.CrossRefGoogle Scholar For a more recent discussion of this problem, see Jones, Ronald W., “A Three-Factor Model in Theory, Trade, and History,” in Trade, Balance of Payments and Growth: Papers in International Economics in Honor of Charles P. Kindleberger (Amsterdam, 1971), pp. 3–21;Google Scholar and Samuelson's, “Summing up on the Australian Case for Protection,” Quarterly Journal of Economics, 96 (02 1981), 149–60.CrossRefGoogle Scholar

20 Between 1840 and 1856 urban population in European Russia increased by 21.8 percent, or at an annual rate of 1.23 percent. It grew more slowly than in the 1825–1840 period (2.25 percent per year), but faster than in the period 1856–1863 (1.02 percent per year). (The Crimean War was fought from 1854 to 1856). As a fraction of total population, the urban population increased from 9.27 percent in 1838 to 9.98 percent in 1863. See Rashin, A. G., Naselenie Rossii za 100 let (1811–1913 gg.) [The population of Russia for 100 years (1811–1913)] (Moscow, 1956), pp. 86, 98.Google Scholar Urban population, however, is not a good proxy for employment in crafts because many craftsmen resided in villages. See Tengoborskii, L. V., O proizvodi:el'nykh silakh Rossil [On the productive forces of Russia] (Moscow, 1854), part 2, p. 146.Google Scholar

21 The fact that in some European countries the inheritance of a serf's plot was not automatic and had to be secured by the payment of a fee (such as the surrender of the best animal)—a custom completely unheard of in Russia, where every additional working serf was always welcome— suggests that these masters were willing to part with their serfs provided they retained the land.Google Scholar Early in the nineteenth century, the serfs in the Russian Baltic provinces were given their freedom (more or less) without land. According to Kliuchevskii, V., Kurs [A course], vol. 5, pp. 299–300, their position became worse at once.Google Scholar

22 Blum, Lord, pp. 445–47. According to Samarin, , Sochineniia [Collected], vol. 2, p. 421, the law was very poorly worded. It was not clear to whom this limit of three days per week applied.Google Scholar

23 But, according to Semevskii, Emperor Paul extended serfdom to several new provinces and raised labor obligations in Little Russia from two to three days a week. See his Krest'ianskii vopros v Rossii v XVIII i pervoi polovine Xix veka [The peasant question in Russia in the eighteenth and the first half of the nineteenth centuries] (St. Petersburg, 1888), pp. xiv–xv.Google Scholar

24 We have no information on the length of time the masters might have taken to move their serfs from point G to N. In the country as a whole, the serf population grew slowly early in the nineteenth century and became almost stationary after about 1830. It is possible that many serf estates remained in a state of transition between G and N until the very end in 1861. See Troinitskii, A., Krepostnoe naselenie v Rossii po 10-i narodnoi perepisi [The serf population of Russia according to the 10th population census] (St. Petersburg, 1861), pp. 54–56 (English translation by Elaine Herman [Newtonvile, Massachusetts, 1982], pp. 68–71);Google Scholar and Hoch, S. L. and Augustine, W. R., “The Tax Censuses and the Decline of the Serf Population in Imperial Russia, 1833–1858”, The Slavic Review, 38 (09 1979), 403–25.CrossRefGoogle Scholar

25 See the sources cited in footnote 10. A rise in grain prices, discussed in the First Model, would have little effect on serfdom under Paul's Law. It would merely result in smaller land allotments.Google Scholar

26 The law, enacted in 1814 and 1827, forbade the alienation of estate land below a minimum of 4.5 desiatinas per soul (1 desiatina = 1.09 hectares) (Blum, Lord, p. 532). It is not clear to us whether this minimum applied to the land allotment of each serf or to the land/labor ratio of the whole estate. Skrebitskii, Krest'ianskoe [Peasant], vol. 2, part 2, pp. 1491–539, cites many cases when the land allotment for serfs was less than 4.5 desiaginas per soul. In any case, the law merely forbade the alienation of land; it did not compel the masters to acquire additional land to restore the minimum. It was not likely to have been important because the restriction only mattered in densely populated areas where Pau's Law must have been ineffective.Google Scholar

27 See, for instance, Fedorov, Pomeshchich'i [Landowners'], pp. 42–49;Google ScholarIndova, E. I., Kreposinoe khoziaisivo v nachale XIX veka pa materialam votchinnogo arkhiva Vorontsovykh [The serf economy at the beginning of the nineteenth century according to the materials in the patrimonial archives of the Vorontsov family] (Moscow, 1955), pp. 178–82; and Koval'chenko, Russkoe [Russian], pp. 128–57, 177.Google Scholar

28 According to Blum, Lord, pp. 460–62, the Inventories were definitely imposed on the three Southwest provinces (Kiev, Volyniia, and Podoliia) in 1848. But in Lithuania and white Russia they were postponed because of the opposition of the serfowners. However, Skrebitskii, Krest'ianskoe [Peasant], vol. 3, pp. 1266–73, describes serfs' obligations in Lithuania as being subject to the Inventories. It seems that the same held true in the Minsk province, but not in the rest of White Russia. He warns that the Lithuanian Inventories need not reflect the actual state of affairs (p. 23). The Russian government did want to improve the welfare of these serfs, but not to the point of arousing the envy of others who lived elsewhere under Orthodox masters.Google Scholar

29 This is a great simplification. Actually, the serfs' duties were quite complex.Google Scholar

30 It existed implicitly and was also constant in the Second Model because Ls was determined by Paul's Law and Ts by the subsisteice requirement.Google Scholar

31 The intersection of Z* = Ts[H - Ls] with E = Lsα T sß (where E is the subsistence level) results in Ls = H - (Ts/Z*), where Ts is obtained from the equation H - (Ts/Z*) = [E/Tsβ](1/∞), which can be solved by trial and error.Google Scholar

32 From the definition of Z* = Ts/[H − Ls], we obtain Ts = Z*[H − Ls]. Hence |δTs| = Z*δLs, or Z* = |δTs/δLs|.The master will not move his serfs down the Z* ray so long as δLsRL > |δTs|RT, that is, RL/RT > |δTs/δLs|, or, RL/RT Z*.+|δTs|RT,+that+is,+RL/RT+>+|δTs/δLs|,+or,+RL/RT+Z*.>Google Scholar

33 The minimum number of serfs that allows the master to obtain this maximum income is obtained by solving the equation (HZ*S − aT)α/S = E(Z*)Y(αT)β for S. The maximization of YM = LMα;TMβ subject to Z* = T5/[H − Ls] = [T – TM]/LM gives TM = βT and LM = αT/Z*, both expressions being independent of S.Google Scholar

34 The collective responsibility of peasant communities for tax collection created a similar situation in old Muscovy, where the old-timers objected to the departure of any members. See Domar, “The Causes,” p. 25.Google Scholar

35 The number of serfs that maximizes the income of each serf is αT/βHZ*.Google Scholar

36 Since the Inventories have the effect of reducing the optimal number of serfs on the estate, they make serf agriculture less labor-intensive in the long run. These enactments, or similar ones in other countries, can thus create a class of landless peasants that, strangely enough, can coexist with serfdom.Google Scholar

37 See Skrebitskii, Krest'ianskoe [Peasant], vol. 3, pp. 1228–65.Google Scholar

38 See Koval'chenko, Russkoe [Russian], pp. 62–63; Semevskii, Krest'iane [Peasants], pp. 48–51, 591–92; and Blum, Lord, pp. 394–401. Although there were regional variations between the two periods, the totals for the whole country remained stable.Google Scholar

39 The difficulty of collection is stressed by Samarin. He also asserts that it is dangerous for peasants used to living and working under their master's supervision to be transferred to the obrok system. See his Sochineniia [Collected], vol. 3, pp. 44, pp. 233–48. On the master' choice between labor services and rent there exists a considerable literature. See, for instance, Koval' chenko, Russkoe [Russian], pp. 163, 207, 212–13, 222; Fedorov, Pomeshchich'i [Landowners’], p. 29;Google ScholarKatsenelinboigen, Aron, “Disguised Inflation in the Soviet Union: The Relationship Between Soviet Income Growth and Price Increases in the Postwar Period,” in Abouchar, Alan, ed., The Socialist Price Mechanism (Durham, North Carolina, 1977), p. 174, note 2; Folke Dovring, “Bondage, Tenure, and Progress,” Comparative Studies in Society and History (April 1965), 309–23;Google Scholar and Fenoaltea, Stefano, “Authority, Efficiency, and Agricultural Organization in Medieval England and Beyond: A Hypothesis,” this JOURNAL, 35 (12 1975), 693–718.Google Scholar

40 See the sources cited in footnote 1.Google Scholar

41 It was shown in the First Model that rising land values could destroy the profitability of fixed arrangements under serfdom.Google Scholar

42 See Blum, Lord, pp. 449–51, and Koval' chenko, Russkoe [Russian], pp. 131, 295–97. The later claims that in the nineteenth century the magnitude of obrok was increasing in real terms.Google Scholar

43 See Indova, Krepostnoe [The serf,] pp. 178–82; Fedorov, Pomeshchich'si [Landowners's], pp. 42–50; and Koval'chenko, Russkoe [Russian], p. 151.Google Scholar

44 Some serfs, particularly those who belonged to wealthy masters, did get some benefits from this arrangement, such as support during famines and protection against other noblemen and government officials. In Turgenev's famous story, “Khor' and Kalinych,” Khor', a wealthy obrokpaying serf, does not want to acquire his freedom because the “every beardless person would be [his] boss” (noblemen and government officials wore no beards). (Zapiski Okhotnika [Sportsman's sketches] (Moscow, 1961), p. 13).Google Scholar

45 See the discussion at the end of the First Model.Google Scholar

46 “Svedeniia o prodhnykh tsenakh tsenakh na zemli” [“Data on selling prices of land”], Zhurnal ministerstva vnutrennikh del [Journal of the Ministry of Interior Affairs] (1859), book 7, pp. 1–46, book 8, pp. 95–118.Google Scholar Early in this century, Maslov, , Agrarnyi [The agrarian], vol. 1, pp. 463–64, used these data to show that serfdom had become unprofitable. But he did not calculate serf prices separately from land prices. Blum did. See Lord. p. 372.Google Scholar

47 Additional details of these data are given in the Statistical Appendix.Google Scholar

48 The name of the editor is not given, but, according to D. I. Rikhter, he was the well-known statistician A. G. Troiniskii. See D. I. Rikhter, “Zabytyi material po statistike prodazhnykh tsen na zemliu” [“Forgotten materials on the statistics of selling prices of land], Trudy Imperatorskogo Vol'nogo Ekonomicheskogo Obshchestva [Works of the Imperial Free Economic Society] (1897), vol. 2, book 4, pp. 1–28.Google Scholar

49 According to Blum, Lord, p. 81, the redemption period was 40 years before 1830 and three years between 1830 and 1917.Google Scholar

50 “Svedeniia” [“Data”], p. 3.Google Scholar

51 This implies that on the average some. 7 percent of all serfs were sold per year. A small fraction should be added to this number to include serfs (servants, craftsmen, and so on) sold without land.Google Scholar According to R. W. Fogel and S. L. Engerman, 1.92 percent of the slave population of Maryland was sold each year in the period 1830–1840. They accepted this figure as the national average. See their Time on the Cross: The Economics of America Negro Slavery (Boston, 1974), vol. 1, p. 53.Google Scholar

52 In addition to Maslov and Blum (see footnote 46 above), this assumption was made by Gerschenkron, “Agrarian,” p. 738; Liashchenko, Istoriia [History], vol. 1, p. 584; Pokrovskii, Russkaia [Russian], vol. 4, p. 93;Google ScholarRobinson, G. T., Rural Russia under the Old Regime (Berkeley, 1932, 1960), p. 88; and others who were intersted in the fairness of prices charged to former serfs for land allotted to them by the Emancipation.Google Scholar

53 Skrebitslii Krest'ianskoe [Pesant], vol. 3, p. 17. In some regions, like Lithuania, White Russia, and Little Russia, these percentages were even smaller: 4.4,2.7, and 5.3 respectively. For the country as a whole, the quantity of unpopulated land sold was some 24 percent of total land sold; in White Russia it was quantity of unpopulated land sold was some 24 percent of total land sold; in White Russia it was only 8.6 percent, and in the Southwest–a tiny 2.8 percent.Google Scholar

54 Indicating the first and second methods by superscripts, we can express the regional serf prices obtained under each method byGoogle Scholar

and

where Vp is the value of populated estates; and Tp are the number of serfs and the amount of land, respectively, on such estates; Tuis the amount of unpopulated land sold; and PTu is the price of unpopulated land. It is obvious from (In) that in the first method the average price of populated land in each region is the weighted average of prices of unpopulated land in each province weighted by the provincial quanties of unpopulated land. In the second method (2n), it is the provincial quantities of populated land that are used as weights, a procedure that seems to us more justofiable than the first. From (1n) and (2n) it can easily be deduced that if and only if

A negative relationship between quantities and prices has been found by Gerschenkron and others. See his A Dollar Index of Soviet Machinery Output, 1927–28 to 1937 (Santa Monica, California, 1952).Google Scholar

For a mathematical treatment of the “Gerschenkron Effect” see Ames, E. and Carolson, J. A., “Production Index Bias a Measure of Economic Development,” Oxford Economic Papers, 20 (03 1968), 12–24. If this effect holds for quantities and prices of land as well, than a negative relationship between the prices and quantities of unpopulated land is more probable than between the prices of the former and the quantities of populated land; hence, this inequality is likely to be true more often than not.CrossRefGoogle Scholar

55 The negative Moscow price is obviously wrong. In the adjacent province of Vladimir, the serf price was a suspiciously high 211.82 rubles—the highest of all shown in Appendix Table 1 in the Statistical Appendix. Sixty-five percent of all serfs in the Moscow province were on obrok, paying an average annual sum of 10.84 rubles. Capitalized at 8 percent per year, this would amount to a price of 135.50. Assuming that non-obrok serfs were only two-thirds as profitable as the ones on obrok, we obtain 120 rubles as the average serf price. Even if the non-obrok serfs were completely useless to their masters, the average serf would still be worth 88 rubles. For the Saratov and Orenburg provinces, where the proportions of obrok serfs were only 30.8 and 8.6 percent respectively, such an exercise is less meaningful, but, for whatever it is worth, the assumption of two-thirds would yield prices of 82 and 84 rubles, respectively. Of course the 8 percent capitalization rate and the assumption of two-thirds are arbitrary, but it is clear that any reasonable change in these magnitudes would also fail to reduce the serf prices in these three provinces to zero. (Sources are given in Table 2.) The land redemption bonds given to the former serfowners after the Emancipation carried an interest rate of 6 percent per year. Hence the 8 percent capitalization rate used here looks reasonable. Although every serf was mortal, he or she was expected to leave offspring.Google Scholar

56 See Domar, “The Causes,” p. 26, and Blum, Lord, p. 367. According to Pokrovskii, Russkaia [Russian], vol. 4, pp. 9–10, Nicholas I suggested to one of the committees on the peasant question that it should be forbidden to sell estates or to grant mortgages, unless the number of desiatinas was indicated next to the number of souls. In spite of this, published official statistics on mortgages of populated estates indicated the number of (male) serfs but not the quantity of land. See “Bankovye dolgi i polozhenie gubernii v 1856 godu” [“Bank debts and the condition of provinces in 1856”], Zhurnal ministerstva vnutrennikh del [Journal of the Ministry of Internal Affairs] (1856), part 3, book 2, pp. 199–234. Many other examples can be given.Google Scholar

57 For the nature of this restriction, see Romanovich-Slavatinskii, A., Dvorianstvo v Rossii ot nachala XVIII veka do otmeny krepostnogo prava [Nobility in Russia from the beginning of the eighteenth century to the abolition of serfdom] (St. Petersburg, 1870), pp. 272–86.Google ScholarIn Butovskii's, D. note, “Prodazhnye tseny na zemli v Poltavskoi gubernii” [“Selling prices of land in Poltava Province”], Zhurnal ministerstva vnutrennikh del [Journal of the Ministry of Internal Affairs] (1860), book 1, pp. 1–8, the author remarks that Cossack lands in that provinces sold “incomparably” cheaply because only Cossacks were permitted to buy them. He also asserts that land prices in the sales reports (“Svedeniia”[“Data”]) were greatly understated. Samarin, Sochineniia [Collected], vol. 2, p. 121, advocated that personal noblemen, honorary citizens, and merchants of the first two classes be permitted to buy populated estates. As a result of this measure, he expected the prices of these estates to rise. (He also expected the new owners to run their estates more efficiently.)Google Scholar

58 In pre-famine Ireland (1841), village industries were found to be economically important, and the relationship between income and the land/labor ratio was negative. See Almquist, Eric L., “Pre-Famine Ireland and the Theory of European Proto-Industrialization: Evidence from the 1841 Census,” this JOURNAL, 39 (09 1979), 699–718,Google Scholar and Mokyr, Joel, “Malthusian Models and Irish History,” this JOURNAL, 40 (03 1980), 159–66.Google Scholar

59 The ratio of the weighted slave price in the Lower South to that in the Upper South in the United States was 1.82 in 1830–1835 and gradually declined to 1.28 in 1856–1860. See Fogel, and Engerman, , Time, vol. 2, p. 73. The ratio of the highest to the lowest state price, however, would be more relevant here.Google Scholar

60 This positive correlation between regional grain and serf prices adds a bit more evidence against Pokrovskii's hypothesis examined in the First Model.Google Scholar

61 The regional distribution of labor in a country may have several optimums depending on the goals of the decision-makers (such as serfowners or slaveowners, free workers, central planners, and so on).Google Scholar

62 Here are a few straws in the wind: with the exception of the Lake region, the percentage increases (or decreases) in the numbers of serfs and of free males in the period 1835–1859 by regions were positively correlated. The rank correlation between real prices of serfs and of the percentage increases in free males by regions for that period would be reasonably high if not for the low nominal serf prices in New Russia and the Volga regions obtained by our regression. As observed above, these two prices must have been underestimated.Google Scholar

63 See Field, The End, pp. 80–81; and Blum, Lord, pp. 401, 579. The serfowners in the Southwestern region had a reputation for efficiency. Among other things, they cultivated sugar beets on a large scale. It is, of course, impossible to judge the effectiveness of any particular set of Inventories without knowing local prices, wages, etc.Google Scholar

64 Between 1835 and 1859 the number of serfs in the Viatka-Perm' region increased by 42.1 percent, as compared with and average increase in European Russia of only. 95 percent. But these figures may not be completely comparable since they come from two separate sources and are subject to different definitions. The 1835 figures are taken from Keppen, P., Deviataia reviziia: izsledovanie O chisle zhitelei v Rossii [The ninth census: an investigation into the number of people in Russia] (St. Petersburg, 1857), pp. 199–200, and the 1859 figure from Troinitskii, Krepostnoe [The serf], p. 49 (English translation, pp. 61–63).Google Scholar

65 See the sources cited in footnote 64.Google Scholar

66 See Troinitskii, Krepostnoe [The serf], pp. 55–56 (English translation, pp. 68–71); and Hoch and Augustine, “The Tax Censuses.”Google Scholar

67 See Blum, Lord, p. 380. The figures are from Troinitskii, Krepostnoe [The serf], p. 65, note 2 (English translation, p. 83, note 2).Google Scholar

68 Calculated from data in “Svedeniia” [“Data”]; Skrebitskii, Krest'ianskii [Peasant], vol. 2, part 2, pp. 1492–1551; and Troinitskii, Krepostnoe [The serf], p. 45 (English translation, pp. 55–57).Google Scholar

69 See Blum, Lord, p. 381. The magnitude of the special debt was taken from Skrebitskii, Krest'ianskii [Peasant], vol. 4 p. 1241.Google Scholar

70 In 1980, the ratio of all liabilities to all nonfinancial assets on American farms was 17.8 percent. See U. S. Department of Agriculture, Agricultural Statistics 1980 (Washington, D.C., 1980), p. 425.Google Scholar

71 The regional distribution of the serfowners' debts (without that special amount and without private debts) was taken from Skrebitskii, , Kerst'ianskii [Peasant], vol. 4, pp. 1246–49.Google ScholarWe found the lowest ratio of debt to the value of serf estates to be in New Russia–6.4 percent. (That region contained fewer than 400,000 serfs.) In Lithuania it was 12.2 percent. The highest percentage was in the Volga region (29.1), followed closely by White Russia, Viatka-Perm' and the Central Industrial region. A comparison of the burdens of landowners' debts over space and time should be very interesting. The percentage of Lithuanian serfs mortgaged was taken from Blum, Lord, p. 381.Google Scholar

72 Besides listing the above values for each country, the report also gives the provincial sums for each of the five series, as well as listing various ratios of these series for each country (Tp/S, Vp/S, Vp/Tp and Vu/Tu). While these additional listings are strictly speaking redundant, they provide a means of cross-checking the original series for typographical errors. Such a check revealed about a dozen clearly identifiable typos—that is, where a particular correction in the listed data value served to make both the sum and the ratio(s) correct. However, this check also revealed inconsistencies between the original series, their sums, and their ratios, which could not be so easily corrected. Since the original calculations were in all likelihood performed on abaci, whenever there was any discrepancy, we adjusted (or did not adjust, accordingly) the listed data value on the assumption that the calculated sum rather than the calculated ratio was correct. The large number of counties for which none of the reported ratios were correct provides a further justification of this procedure.Google Scholar

73 Not all counties reported sales of both populated and unpopulated estates.Google Scholar

74 This equation was run over all counties that reported sales of both populated and unpopulated estates.Google Scholar

75 The above estimation and test were performed correcting for heteroscedasticity in the manner described below.Google Scholar

76 We also experimented with adding S2, T2p, and other variables to the right side of this equation, with no changes in the results.Google Scholar

77 That the variance of the error term is essentially proportional to S is highly plausible, since each soul corresponds to a random number of efficiency units of labor (including any attendant women and children), and we might expect this variation to be independent across the S souls in each observation. It is somewhat surprising, however, that the variance of Ēp evidently does not depend upon the amount of land.Google Scholar

78 All tests were at the 5 percent level. The order of the two tests (equality of serf prices, equality of land prices) was chosen more or less arbitrarily, and upon the assumption that serf prices were more likely to be equalized than land prices (serfs presumably being more homogeneous and more mobile than land). This conjecture is verified by the outcomes of the various tests (Appendix Tables 1 and 2).Google Scholar