DOI: 10.1111/ehr.13246
ORIG INAL ARTICLE
Social inequalities in famine mortality in the
manorial system of the tsarist Russian province
of Livland in the mid-1840s
Kersti Lust1,2 Martin Klesment1 Hannaliis Jaadla1,3
1Estonian Institute for Population Studies,
Tallinn University
2School of Economics and Management,
Lund University
3Cambridge Group for the History of
Population and Social Structure,
Department of Geography, University of
Cambridge
Correspondence
Kersti Lust.
Email: klust@tlu.ee
Funding information
Eesti Teadusagentuur, Grant/Award
Numbers: PSG669, PRG71
Abstract
By relying on longitudinal data on two rural parishes in the
Russian Baltic province of Livland, the article analyses two
questions concerning famine’s short-run effects onmortal-
ity in a manorial system: (1) whether there is evidence of a
social gradient inmortality during the famine of 1844–6 and
(2) whether the manors could protect the peasants against
the hardships. The analysis reveals that neither the status
of a farmer peasant nor the landlord saved the local inhab-
itants from an increased risk of dying during the famine of
1844–6. The conventional assumptions about the protective
effect of the higher socio-economic status or type of manor
against subsistence crisis found very little support in the
study.
KEYWORDS
famine, manorial economy, mortality, socio-economic status
Severe food shortages were a recurrent phenomenon in pre-industrial Europe, but conditions
rarely deteriorated to famine levels. Literature on pre-industrial Europe suggests that the link
between famines and production shortfalls (harvest losses) and population pressure weakened
from the eighteenth century onwards, and the role of human agency in ‘making’ the famines,
that is, failing to prevent devastating food shortages andmitigate their immediate effects on demo-
graphic outcomes, increased considerably thereafter.1 In situations of extreme hardship, nutrition
of individuals was affected by individual and household coping strategies as well as reactions from
public actors.2
1 Alfani and Ó Gráda, ‘Famine and disease’, p. 134.
2 Vanhaute, ‘Famines’, pp. 231–2.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and
reproduction in any medium, provided the original work is properly cited.
© 2023 The Authors. The Economic History Review published by John Wiley & Sons Ltd on behalf of Economic History Society.
Econ. Hist. Rev. 2023;1–24. wileyonlinelibrary.com/journal/ehr 1
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 LUST et al.
Famine as an extreme deprivation of food produces significant excess mortality due to either
starvation or hunger-induced diseases.3 In crisis times, malnutrition and infectious disease were
inevitably intertwined, and in most cases it is very difficult, if not altogether impossible, to dis-
tinguish between these two factors of mortality.4 Mortality in the past was sensitive to severe
food shortfalls and in some settings also to smaller yearly fluctuations in food prices or harvest
outcomes.5 Mortality upswings, nevertheless, mostly resulted from non-nutritional factors.6
Being the first of its kind based on Eastern and Central Europe, this article contributes to
two ongoing debates concerning famine’s short-run effects on mortality. First, socio-economic
inequalities and poverty clearly heighten the risk of famine,7 but there is little research based
on individual-level data exploring social differentials in famine mortality.8 While the first part of
the article explores whether the effects of famine were moderated by socio-economic status, the
second part examines the protective effect of an institutional framework – whether the mano-
rial system or a certain type of it provided a safety net. It is commonly suggested that levels of
mortality indicate the ability of a society’s institutions to offer protection to inhabitants’ lives in
the face of pre-existing vulnerabilities.9 An examination of the possible protection of manors for
its peasants against economic hardship and hunger-induced mortality shocks constitutes part of
a greater debate over landlord–peasant relationships in the areas dominated by the East Elbian
manorial system,10 as well as contributes to the strands in famine research that emphasize the role
of local institutions and elites in famines.11 Until the mid-nineteenth century, large noble estates
and manorial economy dominated the economic landscape of the Russian Baltic provinces of
Estland, Livland, and Kurland.
Individual-level data on vital events, socio-economic status, and migration extending back to
times when famines were present exist only for a limited number of rural areas in pre-industrial
Europe. By relying onwell-preserved andhigh-quality longitudinal data on twoEstonian parishes,
Helme and Paistu (figure 1), we address two research questions: (1) whether there is evidence
of a social gradient in mortality during the famine of 1844–6 and (2) whether the manors could
protect the peasants against the hardships. Empirical analysis of specific famines has proved to
be a fruitful way of exploring the inner workings of historical economies.12
Our analysis reveals that neither the social status of a farmer peasant nor the landlord safe-
guarded the local inhabitants from an increased risk of dying during the famine of 1844–6. In
circumstances where hunger promoted the outbreak of devastating epidemics (dysentery and
typhus), landed peasants were not better equipped to shield themselves against infection than the
3 Alfani and Ó Gráda, ‘Famines in Europe’, p. 6; Dyson and Ó Gráda, ‘Introduction’, p. 1; Vanhaute, Paping, and Ó Gráda,
‘The European subsistence crisis’, pp. 25–8; Howe and Devereux, ‘Famine intensity’.
4 Post, ‘Famine, mortality’, p. 26; Ó Gráda, Famine, pp. 109, 116; Alfani and Ó Gráda, ‘Famines in Europe’, p. 6; Dyson and
Ó Gráda, ‘Introduction’, pp. 1, 10, 14; Nefedov and Ellman, ‘Excess mortality’, pp. 174–5; Voutilainen, Poverty, p. 22.
5 Bengtsson et al., Life under pressure, pp. 38, 43–4.
6Wrigley and Schofield, The population history, p. 354.
7 Voutilainen, ‘Income inequality’, p. 6.
8 Such studies mostly analyse mortality response to short-term price fluctuations over longer periods.
9 Van Bavel et al., Disasters, p. 124.
10 Hagen, Ordinary Prussians; Peters, ed., Gutsherrschaft.
11 Vanhaute, ‘From famine’, pp. 47–65; Vanhaute and Lambrecht, ‘Famine’, pp. 156–7, 181; Curran, Luciuk, and Newby,
‘Famines’, p. 9.
12 Alfani and Ó Gráda, ‘Famine and disease’, p. 135.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 3
F IGURE 1 Map of Estonia. The two parishes included in the study are highlighted in grey, Paistu (1), in
which Holstre is located, and Helme (2). Mapped by Ago Tominga.
landless. The protective effect of a ‘good’ or ‘paternalist’ lord, however, was detected at early ages.
Peaks in infant (0–1 years) and childhood (1–4 years) mortality were reduced on the state estate,
compared with privately owned manors. Infants living on the manors owned by ‘good’ lords sim-
ilarly experienced a lower death risk. There were no substantial differences in mortality risks for
older children and adults between the different manor types. Thus, the conventional assump-
tions about the protective effect of the higher socio-economic status or type of manor against
subsistence crisis found very little support in our study.
The rest of this article is structured as follows. Section I summarizes the relevant discussions
and empirical findings of earlier studies. Section II characterizes the famine of the mid-1840s in
Livland. The empirical hypotheses concerning the differences in mortality risks in the short run
are outlined in section III. Section IV introduces our data and methods. In section V, we present
results of relative risk models to study short-term mortality differences by socio-economic status
andmanor type during the famine of 1844–6. The article concludeswith a discussion in sectionVI.
I STATE OF THE ART
Literature suggests that, during famines, mortality has an inverse relationship with socio-
economic status and income.13 Vulnerability to famine is affected by underlying chronic
13 Carter and Kelly, ‘Social inequalities’; Ó Gráda, Famine, p. 90; Dyson and Ó Gráda, ‘Introduction’, p. 11; Häkkinen and
Forsberg, ‘Finland’s famine years’, pp. 106, 108; Voutilainen, Poverty, p. 185; Pitkänen, ‘The patterns of mortality’.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4 LUST et al.
undernutrition, susceptibility (e.g. dependency on a single livelihood), and limited assets at the
disposal of individuals and households.14 Unequal access to assets (in a pre-industrial rural soci-
ety, most importantly to land), credit and social networks and support, made lower social groups
more vulnerable to harvest failures and resulting price shocks.15 In the countryside, landless
labourers were more vulnerable than farmers, and substantial farmers were more likely to sur-
vive than smallholders.16 However, if infectious disease wreaked havoc, even the better-off groups
did not escape increased mortality.17 Socio-economic mortality differences could also vary, either
widen or diminish, over the course of the famine, depending on the amount of social disruptions
and government policy measures. Some twentieth-century famines were socially highly selective.
In Vietnam (1944–5) the landless and those dependent on wage labour, lacking entitlement to rice
rations and losingmost of their income in themidst of soaring food prices, were by far more likely
to die of famine than other social groups.18 Furthermore, even a small parcel of land afforded a
valuable safety net for its owners and dramatically reduced household death rates. In one of the
very few papers dealing with famine mortality and social class in a pre-industrial society, Pitkä-
nen confirms that, during the Finnish famine of 1866–8, the risk of death increased less for farmer
peasants than for crofters and labourers.19
Recent literature suggests that pre-industrial societies did not exhibit social class differences
in mortality in normal years, except in some cases in early childhood.20 However, several studies
exploring the influence of economic fluctuations on individual mortality risks have been able to
identify sizeable differentials in the mortality response to changes in food prices before the onset
of modern economic growth.21 Socio-economic differences in mortality varied by age. For exam-
ple, social gradients in mortality response to changes in grain prices, which were often negligible
or missing for infants, became increasingly more apparent in early childhood.22 Studies on mor-
tality differences by socio-economic status amongworking-age adults have produced inconsistent
results.23 In very few historical settings, the parish death records have been digitized to allow for a
more detailed analysis of famine mortality on an individual level, which makes the current study
an important contribution to this field.
In a manorial system, the question of mortality response to economic hardships was intrin-
sically related to the issue of the protective role of manorial lords. In international research,
there is a widespread view that manors insured their farmer peasants against uncertainties.24 It
is believed that the relation between tenant and landlord was contractual, in which the tenant
worked for the landlord in exchange for various forms of protection. The large size of estates
14 Howe, ‘Famine systems’, p. 146.
15 Carter and Kelly, ‘Social inequalities’.
16 Ó Gráda, Famine, pp. 90–1.
17 Bengtsson et al., Life under pressure, pp. 63–4.
18 Huff, ‘The Great Vietnam famine’, pp. 287, 298–300.
19 Pitkänen, ‘The patterns of mortality’; see also Häkkinen and Forsberg, ‘Finland’s famine years’, p. 108.
20 Bengtsson, Dribe, and Helgertz, ‘When did the health gradient’; Jaadla et al., ‘Infant’.
21 Bengtsson, Dribe, and Helgertz, ‘When did the health gradient’, p. 956; Bengtsson et al., Life under pressure; Willführ
and Störmer, ‘Social strata’; Jennings, Quaranta, and Bengtsson, ‘Inequality’.
22 Van Poppel, Jonker, and Mandemaker, ‘Differential infant’; Breschi, Manfredini,and Pozzi, ‘Mortality’; Willführ and
Störmer, ‘Social strata’.
23 Bengtsson and vanPoppel, ‘Socioeconomic inequalities’; Bengtsson,Dribe, andHelgertz, ‘Whendid the health gradient’.
24 For a review, see Dribe, Olsson, and Svensson, ‘Was the manorial system’, pp. 296–7.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 5
offered opportunities to insure its subjects against extreme events, such as harvest failures, which
would otherwise be impossible to achieve without extensive public intervention. Recent literature
emphasizes the protective effect of manors, particularly in the context of serfdom, claiming that
it was in the interests of the nobles and the state to make sure that serfs were guaranteed subsis-
tence.25 Research has revealed that during the crisis of the 1840s, the response of landlords was
mostly proactive, for example, in Scotland, butmuch less supportive in Ireland and East Prussia.26
Despite the claim that manorial lords often undertook action to help, historians have seldom
presented empirical evidence showing that manors’ support to the peasants facing subsistence
problems saved the latter from increasedmortality in times of severe food shortages.27 In a micro-
study of Borshevka parish in nineteenth century Russia, Hoch detected little connection between
mortality and harvests or grain availability, but explained it with household patterns and mecha-
nisms of land distribution that kept the portion of the rural populationwithout access to a sizeable
plot of land small, and not with the support from themanors.28 Similarly, for the large Voshchazh-
nikovo estate in Iaroslavl’ province, Dennison andNafziger found no references to grain shortages
or to mortality crises even in the years of poor harvests.29 Their research, furthermore, did not
reveal special instructions to bailiffs regarding efforts to cope with harvest failures or petitions
from local peasants requesting famine relief.
Several famine researchers have established that it was rather the actions of public institutions
and relief measures that affected the outcomes and characteristics of famines.30 In the context of
the Hungry Forties, however, the effectiveness of measures taken by the central authorities has
been questioned, whereas the role of locally based institutions like municipalities and voluntary
organizations in coping with the subsistence crisis has been highlighted.31
Recently, Dribe, Olsson, and Svensson tried tomeasure the efficiency of themanorial system in
protecting its inhabitants against economic stress in southern Sweden 1749–1859 by studying the
impact of grain prices onmortality outcomes on the basis of parish-level data on vital events.32 For
children living in parishes characterized by a high degree of manorialism, there appears to have
been a protective effect in the short run, as there was nomortality response to current prices; how-
ever, this protective effect disappeared after 1 year. For adults, the pattern of mortality response
was similar to that of children. The authors concluded that themanorial estate was able to protect
its inhabitants against the risks of economic stress, but the protective effect was highly imperfect
and only functioned in the short term. Categorizing parishes according to the percentage of land
25 Klussmann, ‘Leibeigenschaft’, pp. 238–40; Prange, ‘Das Adlige Gut’, 67–8; Behrens, Das Adlige Gut Ahrensburg, 46–7;
Moon, ‘Reassessing Russian serfdom’, pp. 503–6; Moon, The Russian peasantry, p. 88.
26 Devine, ‘Why the Highlands’; Vanhaute, Paping, and ÓGráda, ‘The European subsistence crisis’, p. 31; Bass, ‘The crisis’,
pp. 194–5, 199.
27 For example,Hagen, one of themost prominent authors in the camp of ‘revisionists’, has failed to produce such evidence.
In Stawenov, severe harvest failures led to high mortality. Farmers borrowed grain from landlords at high prices, but the
landlords at least charged no interest and awaited with repayment. Farmers also suspended seigneurial rent in hard years.
Hagen, Ordinary Prussians, pp. 193–4.
28 Hoch, ‘Famine’.
29 Like in Livland, in 1844 excessive moisture ruined part of the harvest, and in 1845, due to the drought, grains turned out
poorly: Dennison, Institutional framework, pp. 33, 36; Dennison and Nafziger, ‘Living standards’, pp. 429–33.
30 Post, ‘Famine, mortality’, p. 22; Alfani and Ó Gráda, ‘Famines in Europe’, pp. 3, 19–24; Nefedov and Ellman, ‘Excess
mortality’, pp. 183–4.
31 Vanhaute, Paping, and Ó Gráda, ‘The European subsistence crisis’, pp. 30–5.
32 Dribe, Olsson, and Svensson, ‘Was the manorial system’.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6 LUST et al.
owned by manorial lords made it possible to measure the insurance effect or its absence in a
manorial system as a whole, but did not make it possible to distinguish between different types
of manorial lords. In manorial economies, however, there was considerable variation in the way
landlords managed their estates and made use of the powers at their disposal, and the degree of
variation appears to have been higher east of the Elbe. In addition, the analysis of parish-level
demographic response to short-term fluctuations in rye prices prevents the exploration of how
the mortality response at the time of economic stress influenced social differentials within estate
populations.
II FAMINE OF THEMID-1840S IN LIVLAND
Famine is often defined through its effects onmortality. The 2017 volume Famine in European his-
tory explains that it is ‘a shortage of food or purchasing power that leads directly to excessmortality
from starvation or hunger-induced diseases’.33 The editors of the volume also define a common
methodology to be applied to regional reconstructions relying on time series of deaths/burials.
There was crisis if the number of deaths exceeded the ‘normal’ (calculated by different methods)
by 50 per cent or more.
The subsistence crises of the 1840s in Europe, known as the Hungry Forties, were regional
phenomena, and the reasons for these regional differences may be attributed to the organization
of the regional and local economy and society.34 The latest comparative volume on famine in
European history assures that this famine was confined to a ‘fairly limited part of the continent’.35
This statement may need to be reconsidered in the light of data on eastern and central Europe. In
Europe, the crises weremostly triggered by the failure of the potato crop due to blight. By contrast,
in the three Baltic provinces of Russia, two successive grain failures in 1844–5 led to a spike in
grain prices in 1845–6, producing a strong mortality response, most notably in Livland.36 Famines
lastingmore than 1 year, so-called back-to-back famines, were rare inworld-historical perspective,
and when they occurred, they often turned out to be lethal.37 The case of Livland illustrates the
point. In 1845–6, even population growth was negative in the Estonian part of Livland (figure 1),
around −2 per cent.38 The rural population suffered much more than the urban population.
The study area was among the less affected areas in the Estonian part of Livland. The territories
east and south of the study area were particularly badly hit, while those to the west were the
least afflicted by the crisis in the Estonian part of mainland Livland.39 In the two parishes, Paistu
and Helme, mortality estimates in 1845 were 150 and 120 per cent higher than average in the
preceding 10 years (1834–43); in 1846, crisis-related mortality increase was very light – 18 and 10
per cent, respectively.40 Figure 2 presents yearly death occurrence rates for four age groups in the
33 Alfani and Ó Gráda, ‘Famines in Europe’, p. 2.
34 Vanhaute, Paping, and Ó Gráda, ‘The European subsistence crisis’.
35 Alfani and Ó Gráda, ‘Famines in Europe’, p. 16.
36 Klesment and Lust, ‘Short-term economic stress’, p. 29.
37 Ó Gráda, Famine, p. 8; Ó Gráda, ‘Making famine history’, pp. 7–9
38 According to the annual reports of the Governor: National Archives of Estonia [hereafter RA, EAA. f. 296, n. 4, s. 1179
(hereafter 296/4/1179]; Latvian State Historical Archives [hereafter LVVA, 1/ 4/317]).
39 Comparative figures are provided for a number of areas in Lust, ‘The question’, pp. 54–5, table 1, and figure 2. The reasons
for regional variation in Livland mainland have not been studied.
40 The figures pertain to the whole parishes and not only the selected manors.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 7
F IGURE 2 Yearly death occurrence rate by parish, 1834–84. Note: The years highlighted in grey had the
worst harvest failures, 1844–6.
two parishes separately. Throughout the period of 1834–84, the years of 1844 and/or 1845 stand out
as one of the years of highest death rates for almost all age groups. During the years of 1845–6, the
hunger-related diseases dysentery, typhus, or ‘typhus-like’ diseases41 accounted for a large share
of the deaths among individuals above the age of 1 year. Reduced fertility is also a common effect
of famine. The decline in births is presented in figure 3. During those years, the numbers of births
declined equally for all social groups.
As several other famine-prone societies, the Baltic provinces were low-productivity agricultural
areas, characterized by high socio-economic inequality.42 Virtually all agricultural land belonged
either to nobles or to the Russian crown. Noble manors as well as state estates were large-scale
agricultural enterprises based on corvée (labour rent). The main sources of income for manors
were distilling liquor fromgrain andpotatoes, and fattening oxenwith the draff for sale on theRus-
sian market.43 Sheep farming was of minor importance. The commercial manor economy existed
side-by-side with the highly self-sufficient peasant economy, characterized by limitedmarket pro-
duction and a heavy dependence on rye and barley for subsistence. Due to the high concentration
of land in the hands of Baltic German nobles, and low level of urbanization and industrialization,
the share of landless labourers and semi-landless cottagers was large: They formed roughly two-
thirds of the rural population in the study area, and for most being a farm or manor worker was a
lifelong occupation.
41 Vanhaute, Paping, and Ó Gráda, ‘The European subsistence crisis’, p. 26; Voutilainen, Poverty, p. 16.
42 Voutilainen, Poverty, p. 17.
43 Kahk and Tarvel, An economic history, p. 83.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 LUST et al.
F IGURE 3 The size of birth cohorts in Helme and Holstre parishes by father’s socio-economic status,
1834–52. Note: The years highlighted between two dashed lines had the worst harvest failures, 1844–6.
Natural conditions within the boundaries of a parish did not vary much, and the possibilities of
earning supplemental income from non-farming activities were fairly similar. The nearest town
lay a half-day’s walk from Holstre (Viljandi), and roughly a day’s walk from the Helme manors
(Valga) (figure 1). Large and middle-sized farms dominated the area (table 1). The tenant farmers
paid rent, normally in the form of labour dues and dues in kind, and the rates were determined
either by the manor owner (on noble manors) or the state (on state estates). On noble manors,
tenancies were short term and unstable.
To have sufficient labour to discharge labour duties to the estate and to work the large holdings,
a farmstead population normally included servants in addition to the family of the farmstead’s
head. The farmstead servants were both male and female, and included married and unmarried
people. Multiple-family farmsteads were typical of the area, with the average number of conjugal
units per farmstead ranging from2.7 to 3.6.44 Besides the families of farmsteadheads, servants, cot-
tagers, widows and widowers, incapacitated veterans, wives of conscripts, paupers, and orphans
all showed up as members of the farmstead. Some of themwere also related by blood or marriage.
Archival sources do not distinguish between those who dwelt in separate cottages and those who
livedwith a farmer, but customarily cottagers andmarried servants lived separately, while unmar-
ried servants shared the house and the same bowl and table at meals with the farm head’s family.
Married servants were given a small patch of land, foodstuffs, clothes, lodging, heating, and the
right to keep cattle, a pig, etc., while unmarried servants received fixed payments in kind and
cash. In general, the non-monetary remuneration and home production made up most of their
income. The heads of farmsteads who were responsible for operating the farm and fulfilling the
44 Lust, ‘A not so undesirable status’, pp. 79–80.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 9
TABLE 1 The size-structure of farms and land:population ratio in the study area, during the first half of the
nineteenth century.
Number of farms
Farm size by
thaler
valueb Total
Manor Large Medium Small Total Mean SE Thalers Peoplea
Land:
population
ratio
Holstre
state
estate
35 68 – 103 18.3 5.2 1929.25 2528 0.76
Helme
church
estate
1 5 – 6 15.7 4.3 94.38 146 0.65
Jõgeveste 18 15 2 35 18.9 4.4 661.21 663 0.99
Riidaja – 35 – 35 17.8 1.6 650.33 654 0.99
Lõve 21 18 – 39 20.8 4.6 853.25 756 1.13
Helme 37 9 – 46 24.1 5.4 1108.71 1029 1.1
Patküla 39 5 – 44 21.6 2.8 942.23 795 1.19
Leebiku-
Vanamõisa
12 21 1 34 19.2 7.5 652.57 566 1.15
Total 163 176 3 342 19.8 5.3 6891.93 7137 0.96
aThe population size derives from the 1834 poll tax list.
bThe size of farms was measured in thalers. Medium farms were 10–20 thalers and large farms were over 20 thalers. One thaler
of land amounted roughly to 2.2 hectares of agricultural land. Later, the size in thalers served as a basis for calculating the rates
of money rent and land purchase prices. As the farms’ size in thalers derives from different years from the period 1817–41 (the
value increased with time as a result of the increase of arable in size and land improvement), the differences between estates were
probably smaller. Furthermore, it might be that the higher value of thalers for the noble estates rather reflects the incentives of
their owners to request higher dues from the tenants, as the dues were traditionally assessed on the basis of farm’s size in thalers
(even if formally, the dues were subject to the arbitrary will of the manor owner) than real differences in land size and soil fertility.
For example, using the size of arable land as a basis for classification of farm sizes, large farmswith arable land over twelve hectares
constituted two-thirds of the farms in Holstre.
Notes: SE, standard error.
Sources: RA, EAA 567/1/370; EAA 932/1/6; EAA 934/1/2e; EAA 934/1/2 h EAA 934/1/2m; EAA.1296/1/326; EAA 2419/1/381.
duties and obligations to the owner of the manor and to the crown divided the duties (e.g. corvée
days) among the able-bodied inhabitants of the farmstead, apportioned the land, and set the con-
ditions according to which each family unit in the farmstead would till it. Joint residence and
eating habits, kinship relations, short-term tenancies, and the farmstead head’s duty to maintain
his labourers all reduced the social gap between the landed peasants and the landless.
Hunger struck first the southeast of the Estonian-speaking part of Livland in 1840–1.45 In this
province, the general food availability started to decline further after the harvest failure in 1844.
In the worst afflicted parishes, the crisis was still raging as late as in spring 1847. In 1844, frequent
rainfall and storms damaged the potato, grain crops, and hay. The resulting lack of forage and
the use of rotten hay led to a massive loss of horses and cows.46 A shortage of good-quality seed
grain heightened the food crisis into the next harvest. Contemporary observers described the years
45 Kruus, Talurahva käärimine; Kahk,Murrangulised, pp. 45–8.
46 RA, EAA 296/4/1177, annual report of Livonian Governor Georg von Fölkersahm to the Emperor, 1845.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10 LUST et al.
1844–5 as ‘some of the hardest andmostmiserable times Livland has ever seen’.47 Due to a drought
in 1845, yields fell to about half. In Viljandi County (including the parishes of Paistu and Helme),
rye yielded roughly three grains per seed sown in 1844 and 2.5 grains in 1845; barley yielded two
grains in 1845.48 As far as the estate administrations’ reports tell, the yields were probably slightly
lower in Holstre than in the manors of Helme.
Official reports and correspondence speak almost univocally about ‘general misery’ and do not
highlight social strata differences as a factor in coping. Theymention differences betweenmanors
rather than between various social groups in the villages.49 Famished individuals ate unhealthy
food substitutes for sustenance. Eating chaff, half-rotten potatoes, sparrowgrass, sainfoin, nettle,
burdock, thistles, and other indigestible substitutes made the peasants, according to the governor
of Livland, susceptible to typhus and dysentery.50 The needy might have tried to augment their
odds of survival by migrating, but the crisis generated little migration from the area under study
to towns or less affected parishes. In Livland, out-migration from the stricken areas was possi-
ble only with permission from the local noble estate-owners or peasant community authorities.
Any person who left the community without authorization and was caught begging for alms was
forcibly returned home by local police courts.51 The law required that the community ‘feeds all its
poor’. Temporary migration was also limited because the areas lying to the south, east, and north
of the study area were affected more severely than the area itself. In the wake of the subsistence
crisis, mass conversions from the Lutheran church to the Russian Orthodox Church took place in
Livland in the hope for purported social and material benefits of taking the ‘tsar’s faith’.52
The serf emancipation laws of 1816–9 had left the provision for the peasantry on noble estates
solely to communal and public institutions (the provincial food supply commission) and had
declared that henceforth peasants would lose their right to demand advance loans frommanorial
lords. Communal granaries containing spring and winter grain supplies gathered from landed
peasants were meant to provide the needy with assistance in the form of loans in times of dis-
tress, and in extraordinary situations, when granary supplies became insufficient, the community
authorities were able to apply for a state relief loan. Due to low productivity, a portion of the
region’s tenant farmers needed to take advance loans from the communal granary stocks almost
every spring, and not only in years of crop failure. In 1845–7, the imperial government directed
674 506 rubles of aid in money, and tens of thousands of chetverti of grain and flour to Livland.53
Overall, the state assistance loan to Livland in both forms amounted to about 1.2 million rubles.54
The first shipments of food supply from Russia began to arrive in summer 1845. Monetary aid was
mostly used to purchase grain and redistribute it within the province.
General poverty and collective responsibility for repaying the relief loans made the community
authorities unwilling to include the landless sufferers in their appeals for relief loans, regardless of
47 RA, EAA 1267/1/36, f. 16; Kahk,Murrangulised, p. 106.
48 Klesment and Lust, ‘Short-term economic stress’, p. 25.
49 Lust, ‘Feeding the landless’, p. 95.
50 Kahk,Murrangulised neljakümnendad, p. 45; RA, EAA 296/4/1177, annual report of Livonian Governor Fölkersahm to
the Emperor, 1845.
51 Lust, ‘Feeding the landless,’ p. 95.
52 Kruus, Talurahva käärimine.
53 1 chetvert of rye = 3 bushels = 360 pounds = 147 kg, 1 chetvert of barley = 300 pounds = 123 kg.
54 Lust, ‘Feeding the landless’, p. 91.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 11
the official regulations and sanctions against the delinquents.55 The rate of relief loans for peasants
– a maximum of a pound of grain per day per adult – was similar for all peasant strata, but grain
rations were cut for everybody when relief reserves became scarce. Tenant farmers could also
borrow grain for seed and consumption from the communal granary stocks.
The manorial lords commanded ample resources that could have been used for helping the
needy.56 Its grain stocks far exceeded the needs of the noble family and their servants and other
hired personnel; manorial lords had access to bank credit and could create jobs. Bailiffs and over-
seerswerewell informed about the economic capacity of the farms and the villagers’ need for help.
From manorial reserves, grain could be delivered without much delay. The Livonian provincial
authorities explicitly obligedmanorial lords to grant subsistence loans to needy peasants,57 which
was primarily a practical necessity, because, under corvée labour, the welfare of manor farming
depended directly on how the farms fared, and howmanyworkmen and draught horses the farms
could provide. Official reports, as well as correspondence, contain many references to providing
peasants with both seed grain and baking grain from the manors.58 In addition to advance loans
to tenant farmers, the manorial lords also sold large quantities of grain to the communities that
paid the ‘market price’, that is, almost double the price compared with the non-famine years, for
it. Communities purchased the grain from the manors with the sums they had received from the
state as a relief loan. For manorial lords, it was a profitable transaction. Furthermore, the pro-
duction of vodka was not officially limited, although it swallowed a great share of the grain and
potatoes produced on the demesne. Neither the nobility nor the governor favoured restrictions
on distilling, as it would disrupt the functioning of the manor economy and the fulfillment of
the crown’s orders for vodka and, therefore, diminish the profits of manor renters and state tax
revenues from selling vodka.59
To sum up, the crisis in the study area was relatively short and weak within the context of the
province. In the areas under analysis, in 1844 and 1845, harvests failed twice, causing grain prices
to soar between 1845 and 1846. This influenced demographic patterns in the short term – therewas
a distinctive mortality crisis in 1845 but also a decline in births in 1845–6. In regions characterized
by export-oriented manor economy and a large number of landless agricultural labourers, as in
East Prussia and Livland, economic and social access to food and the relief regime affected the
scale of the crisis considerably more than the aggregate grain availability.60
III RESEARCH QUESTIONS
Our paper addresses two main research questions. First, we are interested in whether mortality
risks for some social groupswere exacerbated during the period of extremehardship and famine in
local rural society. Based on previous research we expect to see indications of social differences in
55 Ibid., pp. 96–7.
56 In southern Estonia, manors over 1000 ha constituted 90 per cent of all noble manors. On noble manors, the ratio of
demesne to peasants’ land was 1:2; on the state estate of Holstre, it was roughly 1:10.
57 Lust, ‘Feeding the landless’, p. 100; RA, EAA 291/1/8270, fo-s 424–8v: Governor Fölkersahm to Governor-General
Golovin, 22.11.1845.
58 Lust, ‘Die Rolle’, pp. 233–43.
59 Ibid., p. 228; minutes of the Livonian food supply commission, 5.10.1845: LVVA 29/1/1451, fo-s 117–22.
60 For East Prussia, see Bass, ‘The crisis’.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
12 LUST et al.
mortality, namely that the landless strata weremore vulnerable to harvest failures and subsequent
food scarcity than tenant farmers, and thus experienced a higher risk of mortality. However, as
mentioned earlier, social gaps in local farmsteads were not very large among landed and landless
peasants.
In rural areas, access to land largely determined the socio-economic status of an individual and
a family. Our study sample is divided into three broad categories: tenant farmers with large and
medium-sized farms, skilled workers, and the semi-landless (cottagers) and landless. Cottagers
had only a small parcel of land or no land at all and normally worked for wages. Those without
land circulated among the farmsteads and manors as hired labour.
Parish registers normally do not differentiate between different segments of landless and semi-
landless strata, or do so inconsistently. The greater vulnerability of cottagers relative to farmhands
and manorial workers is not made evident by our data. Different households within the same
socio-economic group experienced food shortages to a different extent.61 As these social groups
exhibited within group variation, income level would be a more precise measurement of socio-
economic inequality. In the absence of an income tax (instead, therewas a flat poll tax) and income
data, the available sources allow differentiation only between the three main social categories.
This might have a levelling effect on relative mortality risks.
Second, the paper addresses whether the type of manor or manorial lord played a role in reduc-
ing the vulnerability of villagers in times of harvest failures. Practices in manorial societies varied
across manor types and from landlord to landlord. Based on previous research, we hypothesize
that state estates and/or paternalistic landlords could protect peasants relatively more effectively
against increased mortality risks when the crops failed and subsistence problems occurred.
Our study encompasses six privately owned manors and one church estate in the parish of
Helme. For comparison, we use a large state-owned estate of Holstre in the parish of Paistu. The
legal as well as economic situation of the state peasants was better than that of their counterparts
living on noble manors and church estates.62 Unlike on noble manors, the tenant farmers living
on state estates had the prerogative to renew the farm contract if they had fulfilled their duties
and obligations to the estate and crown. The rent level was fixed and lower than on noble manors.
On noble land, by contrast, the labour dues were unlimited, and landlords could extract as much
labour rent and rent in kind from the tenant farmers as could be obtained without compromising
the viability of farm households.
We categorize the noble manors in the parish of Helme according to the characteristics of the
manorial lord and the relationship between the lord and the peasants.63 First, we consider the
manorial lord paternalistic if the farmswere permanent and seldomchanged hands, if labour dues
and monetary payments were lower than on nearby manors, and if in- and out-migration levels
were low. Second, the rate of conversion to the Orthodox Church in the mid-1840s gives insight
into the nature of lord–peasant relationships. The converts were seeking freedom from mano-
rial authority, tax relief, monetary aid, and so on. If the conversion rate remained low (less than
5 per cent, according to Riidaja and Jõgeveste manor), the landlord–peasant relationships were
likely better. Third, the manorial lord is considered to be paternalistic if folklore, oral tradition
collected in the 1920s and 1930s, and published memoirs support it. According to these three cri-
teria, Riidaja manor, Jõgeveste manor, and the Helme church estate were owned by paternalistic
61 Van den Berg, Pinger, and Schoch, ‘Instrumental variable estimation’, p. 466.
62 Klesment and Lust, ‘Short-term economic stress’, pp. 25–6.
63 Ibid., p. 26.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 13
landlords. These three were grouped together and coded as ‘paternalistic’. The other four manors
in Helme – Helme, Lõve, Patküla, and Leebiku-Vanamõisa – were owned by conventional land-
lords. This resulted in three different manor groups: one state estate, one paternalistic group, and
one conventional group divided into two sub-groups.
IV DATA ANDMETHODS
In this study,wemakeuse of a dataset that draws on individual-level data fromparish registers and
listings ofmigrants. Famines rarely have clear-cut beginnings or endings, and the excessmortality
only refers to the terminal phase of the process.64As a crisis period, we define the 2 years that
witnessed serious harvest failures and food scarcity – from harvest 1844 (1 October) to harvest
1846 (1 October). As shown above, these 2 years featured extraordinarily low harvests as well as
extraordinary relief measures by local and state institutions, in particular.
For the analysis of short-runmortality effects, our dataset includes, besides vital and migratory
events, also data from the 1834 poll tax list, which provides us with an enumeration of the year’s
population, comprising 4737 persons on the sevenmanors inHelme, and 2537 onHolstre in Paistu.
The poll tax list includes data on sex, age, marital status, and socio-economic status (SES). To this
‘initial’ dataset, all following births in the area until 1856 are added from the Lutheran andRussian
Orthodox parish registers. In addition to births, people who migrated to the manor are recovered
frommigrant listings, parishmember registers, and church certificates (Parochial Schein). Annual
listings ofmigrants include time of arrival or departure. Similarly, parishmember registers provide
the date of migration. Church certificates, issued by the pastor, also include the date of birth of
the migrant. These sources were used to determine the time of arrival or departure of people
moving in and out of the community. Linking migration records to poll tax list and birth records,
we obtained a combined dataset, which includes people who lived in the area under observation
between 1834 and 1856. To this combined dataset, the information on time of deathwas linked from
parish registers. As a result, for each individual, we have information about the time of entering
under observation (birth, in-migration) and leaving it (death, out-migration).
From the observed population, we excluded those without information on sex and date of birth,
as well as stillbirths. In total, we have 11 616 observations for the analysis. From this we had to also
exclude 381 people who lacked information on SES, thus the final study sample includes 11 235
individuals (table 2).
Data on SES is derived from birth and death registers and poll tax lists. In the regression
modelling, we focus on the three main groups – farmers, skilled workers, and landless and
semi-landless labourers – by grouping together certain History of Work Classification System
(HISCLASS) categories.65 The first group, corresponding to HISCLASS 8, is comprised of ten-
ant farmers. The second group includes ‘skilled workers’ (Classes 6–7) – families of individuals
who did not make their living from agricultural production (e.g. forest wardens, peasant school
teachers, storekeepers, bailiffs, foremen, gardeners, innkeepers, and millers). The third category,
‘landless and semi-landless’, consists of low-skilled (Classes 9–10) and unskilled workers (11–
12), and includes families whose land holdings were below subsistence level (i.e. semi-landless
cottagers, landless labourers, and artisans).
64 Voutilainen, Poverty, p. 24; Howe and Devereux, ‘Famine intensity’, pp. 356–64; Carter and Kelly, ‘Social inequalities’, p.
7.
65 Van Leeuwen and Maas, HISCLASS.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
14 LUST et al.
TABLE 2 The size of the study sample by manors and socio-economic status, 1834–56.
Socio-economic status
Farmers Skilled labourers Labourers Total
(1) (2) (1) (2) (1) (2) (1) (2)
Manor type
Paternalista 815 210 138 26 1416 404 2369 640
Conventionalb 1737 558 288 91 3335 1042 5360 1691
State estatec 1003 293 110 31 2393 701 3506 1025
Total 3555 1061 536 148 7144 2147 11 235 3356
Notes: Number of observations and deaths: (1) Number of observations; (2) Number of deaths.
aJõgeveste, Riidaja, and Helme church manor.
bHelme, Leebiku-Vanamõisa, Lõve, and Patküla.
cHolstre.
Our empirical strategy is to apply Cox proportional hazard model. As not all individuals in the
dataset are followed from birth to death, accounting for censoring is necessary. To compute sur-
vival times, being at risk of death starts at birth and ends when an individual reaches the age of 55
years (or the end of observation window in November 1856), after which they are considered right
censored. Right censoring also occurs in the event of out-migration: The individual is considered
right censored as of the month of departure. For those born before 1834 or who moved into the
parish after their birth, hazard rate is conditional on their survival to 1834, or to the month of
in-migration.
Addressing the research questions requires accounting for calendar time. To include calendar
time in the regression models, we categorize the relevant timespan into two longer (1834–44 and
1846–56) and one shorter period (1844–6), each beginning on 1 October. The inclusion of the 10-
year periods before and after the crisis facilitates comparison of relative risks of dying. In the
models, the first period (1834–44) is the reference category, that is, the baseline against which the
two following periods are compared.
Since our focus is on mortality up to the age of 55 years, this broad age range still accommo-
dates substantial differences in risk of death by age but also considerable differences in risk factors
associated with mortality across the age range. For this reason, we estimate regression models
separately for four age groups: infants up to 12months of age, children aged 1–4 years (early child-
hood), older children aged 5–14 years, and adults aged 15–55 years. The mortality of those above
working age is not considered since, in older ages, the socio-economic status differences between
retired tenants and cottagers were fuzzy before the introduction of peasant land ownership.
Our models include socio-economic status and type of manor, plus control variables for the
calendar period and sex of the individual. To gain further insight into the role of socio-economic
status and type of manor over time, we also added interaction terms of these two variables to
the models. For easier interpretation, the hazard ratios obtained from the interaction models are
presented in graph form.
V RESULTS
We fitted four Cox proportional hazards models – one for each age group – as shown in table 3.
The estimate for the period of the subsistence crisis (1844–6) reveals a rise in infant mortality of
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 15
TABLE 3 Hazard ratios for different age groups in Helme and Holstre, 1834–56.
Infant Age 1–4 years Age 5–14 years Age 15–54 years
Time period
1834–44 1 1 1 1
1844–6 1.284* 2.010*** 3.558*** 2.272***
1846–56 0.966 1.247** 1.105 0.919
Sex
Male 1 1 1 1
Female 0.858** 1.129† 0.981 1.008
Manor
Paternalist 1.050 1.007 0.566*** 0.654***
Conventional 1.376*** 1.142† 0.761* 0.767***
State estate 1 1 1 1
SES
Farmers 1 1 1 1
Skilled workers 1.084 0.889 0.599† 0.615†
Labourers 1.040 1.031 1.206† 1.105
No. of subjects 6298 5588 5763 6062
No. of deaths 1234 914 424 781
†p < 0.10.
*p < 0.05.
**p < 0.01.
***p < 0.001.
approximately 28 per cent over the previous 10-year period. Regarding sex differences, girls had
an approximately 14 per cent lower risk of dying during their first year than boys. We also find
differences by manor group, as infants in conventional manors exhibit a significantly higher risk
of dying in the first year of life than infants in the state estate and the paternalistic estates. There
are no clear SES differences in infancy.
The next model in table 3 applies to children aged 1–4 years. Unlike the infant mortality model,
in which the period effects were only moderately influenced by the subsistence crisis, the risk
of dying in the early childhood model doubles during the crisis period. The differences between
manor groups follow the same pattern as that of infant mortality, with the other groups exhibiting
a higher risk than the state estate.
The third model in table 3 reveals children aged 5–14 years as the most vulnerable group in the
face of distress: their risk of dying more than triples in the crisis years. The results also demon-
strate a clear social gradient in mortality. As expected, being a skilled worker is associated with
considerably lower risk of dying (–40 per cent) and being landless with slightly higher risk of
dying (+20 per cent) compared with farmers. In contrast to the first two models, children aged 5–
14 years living in both the conventional and paternalist manors exhibit lower mortality risks than
those in the state estate. The relative advantage of living on paternalisticmanors accords well with
our initial hypothesis about the better conditions in estates owned by ‘good lords’. However, the
beneficial effect of manor type on survival at times of hardship appears to be age dependent.
For adults aged 15–55 years, the years of the subsistence crisis are associated with a mortality
risk 2.3 times higher than the preceding decade. There also appears to be a social gradient that
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
16 LUST et al.
F IGURE 4 Socio-economic status and calendar period interaction effects, 1834–56. Notes: 90 per cent
confidence intervals of predicted values. Interaction models include controls for the sex of the child and the
parish of birth. Farmer in 1834–44 as a reference group.
puts skilled workers at a lower risk (–38 per cent) compared with farmers, while landless labour-
ers show 10 per cent higher (though statistically not significant) risk of dying. The model again
predicts pronounced differences between manor groups. Similar to children aged 5–14 years, the
mortality risk of adults was lower in noble estates, regardless of their owner, than in a state estate.
In conclusion, the models on infant and early childhoodmortality support our hypothesis with
respect to the protective effect of a state-owned manor, compared with those that belonged to
noble lords. The type of lord also proves important, as the mortality risks were higher on manors
owned by conventional as opposed to paternalistic lords. However, after the age of 5 years, the
picture appears to be reversed, with the risk of dying being highest in the state estate. Regarding
socio-economic status, we find the strongest social gradient inmortality between the ages of 5 and
14 years. Themortality analysis of age group 15–55 years does not support our hypothesis regarding
the disadvantaged position of the landless and semi-landless in comparison to landed peasants.
As expected, the skilled workers demonstrate lower mortality risks compared with farmers. The
positive effect of higher status is markedly stronger for young adults and adults.
To disentangle the social gradients in mortality during times of extreme hardship and famine
conditions, we estimated another set of models with interaction effects of socio-economic status,
type of manor, and calendar period. We used fitted models to predict hazard ratios with respect
to the reference category of the interacted variables. The hazard ratios resulting from each model
are shown in figures 4–6.
First, to explore further if landless groups were more vulnerable during the famine of 1844–6,
the models in table 3 were extended by adding SES and calendar period interaction terms. Haz-
ard ratios were then predicted using farmers in 1834–44 as a reference group, and the results are
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 17
F IGURE 5 Manor group and calendar period interaction effects, 1834–56. Notes: 90 per cent confidence
intervals of predicted values. Interaction models include controls for the sex of the child and SES of the father at
birth. State estate in 1834–44 as a reference group.
shown in figure 4. We find a short-term increase in mortality during the crisis years for all four
age categories. The increase in infant mortality during the period of 1844–6 relative to the risk of
dying in infancy in a farmer’s family in earlier period (1834–44) is relatively small (over 1.5 times
only for ‘skilled workers’) and statistically significant only for the landless group. In contrast, the
subsistence crisis elevates the risk of death for all SES categories in early childhood (1–4 years),
compared with the base level. However, the increase predicted by the model is statistically signif-
icant for farmers and landless. In the two older age categories, the hazard ratio during the crisis
years is 2.5–3.8 times higher compared with the base level, except for skilled workers, who again
do not show a statistically significant difference from the comparison level. As mortality rise is
similarly present in both major socio-economic groups, the results suggest that farmers (landed
peasants) did not have an advantage for survival compared with landless groups during this crisis
period.
Second, to test whether different types of manors could protect their peasants against hardship,
we estimated another set ofmodelswith interaction terms between the calendar period andmanor
group variables (figure 5). Again, relative risks were predicted with respect to the reference group,
which in this case is state manor in the period of 1834–44. In the following, each age group is
examined separately with regard to the developments that occur within the manor groups.
Regarding the overall short-term effect of the famine, the mortality rate in all age groups rose
during the years of the crisis. This temporary increase is smallest for infant mortality, but still
exhibits a gradient bymanor group. For state-owned andpaternalisticmanor groups, the predicted
hazard ratio during the crisis is not significantly different from the level of the state manor before
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
18 LUST et al.
F IGURE 6 Manor group and calendar period interaction effects by socio-economic status, 1834–56. Notes:
90 per cent confidence intervals of predicted values. Model controls for sex. State estate in 1834–44 as a reference
group.
the crisis. By contrast, for conventional manors, the infant mortality rate is significantly higher
compared with the state estate before, during, and after the crisis period 1844–6.
In the models fitted for the three older age groups, all manor groups show statistically signifi-
cant highermortality during the crisis over the reference level, with the exception of children aged
1–4 years on the state estate. For the latter, confidence intervals of prediction overlap with the ref-
erence level. This finding pertaining children aged 1–4 years further supports our hypothesis with
respect to the protective effect of the state estate, as already shown in table 3. The protective effect
of paternalistic lords comparedwith conventional lords, by contrast, remainsmodest and is clearly
discernible only for infant mortality.
The state estate appears less favourable in the 5–14 years and 15–55 years age categories. Outside
the crisis period, mortality in paternalistic, and conventional estate group is estimated to be lower
than for the state estate. During the crisis period, all estate groups show a statistically significant
increase in mortality over the level during the previous period. Given the large confidence inter-
vals of prediction for 1844–6, it is not likely that any of the manor groups would be substantially
less affected during the crisis, although the estimates in comparison with the reference category
are biggest for the state estate.
Third, as a sensitivity test of our results above, we further investigated the vulnerability of farm-
ers and the landless peasants during times of crisis across different manor types. For this we fitted
separate models for the subsets of farmers and the landless by age group, applying interaction
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 19
between calendar period and type of manor. In each of the eight groups by age and SES, the
reference category is living on a state estate in 1834–44 (figure 6).
First, during the subsistence crisis there is a marked survival disadvantage for farmers beyond
infancy in conventional estates. On a state estate this disadvantage applies only to farmers who
are adults. Importantly, farmers on paternalistic estates do not show statistically significant higher
mortality in relation to the state estate during 1844–6 in any age group. Before the crisis period,
children aged 5–14 years on conventional and paternalistic manors exhibit even lower mortality
compared with the state estate.
The landless are somewhat more disadvantaged during the crisis period if they live on a con-
ventional manor. Infant mortality of the landless is consistently higher on conventional manors
and is about twice as high during the crisis time compared with the landless on the state estate.
Early childhood mortality during the crisis years is significantly higher both in paternalistic and
conventional manors in relation to the state estate, whereas in the latter there is no significant
rise in risk of dying. With regard to the children aged 5–14 years during the crisis years, we infer
that their death risk increases more in paternalistic and conventional manors, since, relative to
the state estates, they show a lower level of mortality before the crisis years. For the adult landless,
there is an increase in mortality across all manor types in times of hardship.
The comparison of interaction models of farmers and the landless suggests that both groups
shared some common features. Regarding early childhood survival, they were both more vul-
nerable when living on conventional estates compared with the state estate. Also, in terms of
adult mortality, on conventional and state estate they both experienced higher mortality during
the crisis years. Differences between the landless and the farmers appear in infant mortality –
the landless labourers’ survival was worse on conventional estates. Landless young children also
show significantmortality increase during the crisis years on paternalistic and state estates. Along
the same lines, their older children and adults were exposed to higher crisis period mortality in
all estates, whereas for farmers this happened only in some types of estates. Overall, mortality
response of the landless is more likely stronger across the estates.
VI CONCLUSION AND DISCUSSION
This article has considered two contested, yet little studied, aspects of famine mortality: mortal-
ity differentials by socio-economic status and the protective role of manor type. We tested two
hypotheses. First, we expected to find socio-economic differences in mortality between farmer
peasants and landless and semi-landless labourers during famine years. The second hypothesis
suggests that manorial lords protected their peasants from economic hardship in times of dis-
tress, and we expected to find that the type of manor – state- or privately-owned – played a role
in its inhabitants’ mortality risk. Using the individual-level longitudinal data from two Estonian
rural parishes, Helme and Paistu, we addressed these questions by applying survival regression
techniques.
Our interaction models suggest that the response to short-term economic stress was generally
strongest in the 5–14 years age group and weakest among infants. Among the latter, only the land-
less exhibit statistically significant higher mortality during the stress years, while for the other
groups the increase remains within the margin of error. Comparing the mortality response of the
landless with the farmers, point estimates suggest that the rise in children’s mortality rates dur-
ing the crisis years was bigger for landless labourers. However, the difference in how much the
risk of death increases is not statistically significant between the two groups. In the same way, the
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
20 LUST et al.
mortality response of adults exhibits a considerable overlap between the landless and the farmers,
although the landless fare slightly better in this age group.
The absence of significant disparities between mortality responses of the two major socio-
economic groups can be partly explained by factors that counteracted the gains from better
resources. Even though tenant farmers had (better) access to land, credit (especially, from the
manor), and support networks in the years of crisis, the persistence of the habit of living and
eating together, and of old rural hygiene and sanitary customs, made all household members
similarly vulnerable to epidemics.66 Crop failures also turned tenant farmers into ‘deficit
producers’ who were forced to apply for loans to meet their household needs. As the peasant
household cared for the maintenance of orphans, the elderly, the sick, and the poor who lived (as
‘lodgers’) in peasant households, this may have further diluted the effects of social differences in
mortality in the mid-1840s, when households took subsistence loans to provide for its landless
members. Tenant farmers as employers had to feed, clothe, and accommodate landless rural
workers – corvée workers as well as farmhands. Furthermore, in some localities such as the
state estate of Holstre, farmhands and farmers were often relatives, diminishing effects of status
differences in survival. As mentioned above (in section III), the category of unskilled labourers
may also be more heterogeneous than could be determined on the basis of our sources, including
subgroups with higher risk of mortality. There were clear nutritional differences between
manorial servants on the one hand (fed by the manor) and cottagers on the other hand, who
could not expect sufficient support from the tenant farmers unless the cottagers were their close
relatives. Initially, cottagers were not entitled to aid from communal granaries, but as the food
scarcity considerably increased in 1845, they could also apply for relief. The fixed norm of aid,
one pound per person, however, was very low. In the words of one parish magistrate, it was
barely enough to avoid death by starvation.67 The aim of our study, however, was not to delineate
the exact hierarchies of suffering among various social segments of rural people but to reveal
whether the social class mortality differences appeared at all during the crisis. Our results reveal
that there were no statistically significant differences between the landed and landless groups in
mortality.
Our second hypothesis regarding the protective role of the type of manor or manorial lord
found more support. During the crisis years of 1844–6, infants on a state estate and estates owned
by a paternalistic lord had lower mortality than on conventional manors. As better conditions
in these estates probably resulted in more favourable work schedules for mothers, the positive
impact on infant mortality is expected. Early childhood mortality on the conventional and
paternalistic manors more than doubled during the period of economic stress, but on the state
estate, the increase was more moderate – approximately 50 per cent. Such a clear difference
probably reflects less exploitative conditions on a state estate. The protective effect of manor type,
however, did not extend to older ages. In the 5–14 years age group, the risk of death during the
crisis years multiplied in all estate groups. In the oldest age group, the state estate exhibited an
approximately threefold increase in mortality risk, while for the others the risk was doubled.
Our findings indicate that, in post-emancipation Livland, the protective effect of the manors
against hunger-induced mortality increases was probably even more constrained than it has
been demonstrated in southern Sweden. In the latter case, the protective effect for children
and adults was limited to 1 year.68 In our study, we found no protection for older children and
66 Pärdi, Eesti argielu, pp. 45–210.
67 Lust, ‘Die Rolle’, p. 243.
68 Dribe, Olsson, and Svensson, ‘Was the manorial system’, p. 307.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 21
adults. Moreover, large differences apparent in mortality risks between manor types in age
groups above 5 years over the entire study period (1834–56) become statistically insignificant in
the years of distress. To sum up, neither the manorial system nor the ‘paternalist’ landlord could
shield its inhabitants from increased mortality risk in times of severe food shortage, although the
well-developed manorial economy commanded resources that could have been used for feeding
the rural population in times of distress.
In general, the absence of consistent and significant differences in mortality between the two
major socio-economic groups in the villages in times of distressmight be partly associatedwith the
exploitative manorial system in the Baltic provinces, which kept peasants poor through burden-
some taxes in labour and kind, and restricted theirmobility. If surplus product did not accumulate
in the hands of tenants but was appropriated by the manorial lords by raising the rent, the farm
economy withered. Short-term tenants on noble land lacked the security of tenure and the con-
trol over farm succession that were vital economic incentives for pre-industrial tenant farmers.
In Russia proper, agricultural productivity (e.g. seed-yield ratios) was also very low, and much of
the surplus was similarly confiscated by the landlord or the state, but there the manorial system
was combined with communal land tenure and looser restrictions on (temporary) migration. As
long as there is no up-to-date research on subsistence crises in pre-emancipation rural Russia, we
might acquiesce to Hoch’s suggestion that communes and multiple-family households, sizably
reducing the harvest-sensitive portion of the population, really afforded protection from hunger-
inducedmortality shocks in amanorial system.69 Greater flexibility and diversity in the economic
activities of the Russian peasants might have mattered, as well.70
The negative effects of calamitywere difficult to absorb ‘locally’ if epidemics broke out. The case
of Livland rather suggests that the functioning of the relief systemat largewas of great importance.
Government aid, even if it arrived with delays and in the form of interest-bearing loans, mitigated
the food shortage and mortality response to harvest failures as well as soaring grain prices for
those who had survived by the time of the aid’s arrival (Supporting Information).
ACKNOWLEDGEMENTS
Weare grateful to the referees andTommyBengtsson for their valuable comments and suggestions
on the earlier version of the paper.
FUNDING
This work was supported by the Eesti Teadusagentuur (Estonian Research Council) PSG669 and
PRG71.
ORCID
Kersti Lust https://orcid.org/0000-0003-0384-5722
MartinKlesment https://orcid.org/0000-0001-6460-0839
Hannaliis Jaadla https://orcid.org/0000-0002-8569-347X
69 Hoch, ‘Famine’, p. 363.
70 The two micro-studies denying the connection between mortality increases and grain shortages in Russia in the mid-
nineteenth century give much reason to assume that actually there might have been several factors at play. Migratory
work, engagement in trade and rural industry, and payment of feudal dues in cash and kind rather than in labour services
are some of them.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
22 LUST et al.
REFERENCES
Alfani, G. andÓGráda, C., ‘Famines inEurope: An overview’, inG.Alfani andC.ÓGráda, eds.,Famine inEuropean
history (Cambridge, 2017), pp. 1–24.
Alfani, G. andÓGráda, C., ‘Famine and disease’, inM. Blum andC. L. Colvin, eds.,An economist’s guide to economic
history (Cham, 2018), pp. 133–41.
Bass, H. H., ‘The crisis in Prussia’, in R. Paping, E. Vanhaute, and C. Ó Grada, eds.,When the potato failed. Causes
and effects of the last European subsistence crisis, 1845–1850 (Turnhout, 2007), pp. 185–208.
Behrens, A., Das Adlige Gut Ahrensburg von 1715 bis 1867: Gutsherrschaft und Agrarreformen, Stormarner Hefte, 23
(Neumünster, 2006).
Bengtsson, T., Campbell, C., Lee, J., et al., eds. Life under pressure: mortality and living standards in Europe and
Asia, 1700–1900 (Cambridge, 2004).
Bengtsson, T., Dribe, M., and Helgertz, J., ‘When did the health gradient emerge? Social class and adult mortality
in southern Sweden, 1813–2015’, Demography, 57 (2020), pp. 953–77.
Bengtsson, T. and van Poppel, F., ‘Socioeconomic inequalities in death from past to present: an introduction’,
Explorations in Economic History, 48 (2011), pp. 343–56.
Breschi, M., Manfredini, M., and Pozzi, L., ‘Mortality in the first years of life: socio-economic determinants in an
Italian nineteenth century population’, in M. Breschi and L. Pozzi, eds., The determinants of infant and child
mortality in past European populations (Udine, 2004), pp. 123–38.
Carter, B. and Kelly, L., ‘Social inequalities and famine and severe food insecurity risk’ (2021),
https://opendocs.ids.ac.uk/opendocs/bitstream/handle/20.500.12413/16735/954_Social_inequalities_and_
famine_and_severe_food_insecurity_risk%20.pdf?sequence=3&isAllowed=y
Curran, D., Luciuk, L., and Newby, A., ‘Famines in European economic history: introduction’, in D. Curran, L.
Luciuk and A. Newby, eds., Famines in European economic history. The last great European famines reconsidered
(London and New York, 2015), pp. 1–16.
Dennison, T., The institutional framework of Russian serfdom (Cambridge, 2011).
Dennison, T. and Nafziger, S., ‘Living standards in nineteenth-century Russia’, Journal of Interdisciplinary History,
43 (2013), pp. 397–441.
Devine, T., ‘Why the Highlands did not starve. Ireland and Highland Scotland during the potato famine’, in R. Pap-
ing, E. Vanhaute and C. Ó Gráda, eds.,When the potato failed. Causes and effects of the last European subsistence
crisis, 1845–1850 (Turnhout, 2007), pp. 111–22.
Dribe, M., Olsson, M., and Svensson, P., ‘Was the manorial system an efficient insurance institution? Economic
stress and demographic response in Sweden, 1749–1859’, European Review of Economic History, 16 (2012), pp.
292–310.
Dyson, T. and Ó Gráda, C., ‘Introduction’, in C. Ó Gráda and T. Dyson, eds., Famine demography: perspectives from
the past and the present (Oxford, 2002), pp. 1–18.
Hagen, W. W., Ordinary Prussians: brandenburg junkers and villagers, 1500–1840 (Cambridge, 2002).
Häkkinen, A. and Forsberg, H., ‘Finland’s famine years of the 1860s: a nineteenth-century perspective’, in D. Cur-
ran, L. Luciuk and A. Newby, eds., Famines in European economic history. The last great European famines
reconsidered (London and New York, 2015), pp. 99–123.
Hoch, S., ‘Famine, disease andmortality patterns in the Parish of Borshevka, Russia, 1830–1912’, Population Studies,
52 (1998), pp. 357–68.
Howe, P., ‘Famine systems: a new model for understanding the development of famines’,World Development, 105
(2018), pp. 144–55.
Howe P. and Devereux S., ‘Famine intensity and magnitude scales: a proposal for an instrumental definition of
famine’, Disasters, 28 (2004), pp. 353–72.
Huff, G. ‘Causes and consequences of the Great Vietnam Famine, 1944–5’, Economic History Review, 72 (2019), pp.
286–316.
Jaadla, H., Potter, E., Keibeks, S., and Davenport, R., ‘Infant and child mortality by socio-economic status in early
nineteenth-century England’, Economic History Review, 73 (2020), pp. 991–1022.
Jennings, J. A., Quaranta, L., and Bengtsson, T., ‘Inequality and demographic response to short-term economic
stress in North Orkney, Scotland, 1855–1910: sector differences’, Population Studies, 71 (2017), pp. 313–28.
Kahk, J.,Murrangulised neljakümnendad (Tallinn, 1978).
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SOCIAL INEQUALITIES IN FAMINE MORTALITY 23
Kahk, J. and Tarvel, E., An economic history of the Baltic countries (Stockholm, 1997).
Klesment, M. and Lust, K., ‘Short-term economic stress and mortality differentials in rural Estonia, 1834–1884’,
Scandinavian Economic History Review, 69 (2021), pp. 22–40.
Klussmann, J., ‘Leibeigenschaft im frühneuzeitlichen Schleswig-Holstein: Rechtliche Entwicklung, öffentlicher
Diskurs und bäuerliche Perspektive’, in J. Klussmann, ed., Leibeigenschaft: Bäuerliche Unfreiheit in der frühen
Neuzeit (Köln, 2003), pp. 213–40.
Kruus, H., Talurahva käärimine Lõuna-Eestis XIX sajandi 40-ndail aastail (Tartu, 1930).
Lust, K., ‘Die Rolle der Gutshöfe bei der Hungerhilfe für die livländischen Bauern in den Krisenjahren 1841−1847’,
Zeitschrift für Ostmitteleuropa-Forschung, 61 (2012), pp. 219–46.
Lust, K., ‘Feeding the landless in post-emancipation Livland during times of famine (1840s–1860s)’, Slavonic and
East European Review, 92 (2014), pp. 81–106.
Lust, K., ‘The question of moral economy and famine relief in the Russian Baltic provinces of Estland and Livland,
1841–69’, Collegium: Studies across Disciplines in the Humanities and Social Sciences, 22 (2017), pp. 46–66.
Lust, K., ‘A not so undesirable status? Widowhood options and widows’ living conditions in post-emancipation
rural Estonia’, The History of the Family, 26 (2021), pp. 74–99.
Moon, D., ‘Reassessing Russian serfdom’, European History Quarterly, 26 (1996), pp. 483–526.
Moon, D. The Russian peasantry 1600–1930. The world the peasants made (London and New York, 1999).
Nefedov, S. and Ellman, M., ‘Excess mortality in Russia in 1868–1912 and its historiographic implications’, Social
Evolution & History, 20 (2021), pp. 173–97.
Ó Gráda, C., ‘Making famine history’, Journal of Economic Literature, 45 (2007), pp. 5–38.
Ó Gráda, C., Famine. A short history (Princeton, NJ, 2009).
Pärdi, H., Eesti argielu. Teekond moodsasse maailma (Tallinn, 2017).
Peters, J. ed., Gutsherrschaft als soziales Modell. Vergleichende Betrachtungen zur Funktionsweise frühneuzeitlicher
Agrargesellschaften (Historische Zeitschrift, Beihefte 18) (München, 1995).
Pitkänen, K., ‘The patterns of mortality during the Great Finnish Famine in the 1860s’, Acta Demographica (1992),
pp. 81–102.
Post, J. D., ‘Famine, mortality, and epidemic disease in the process of modernization’, Economic History Review, 2nd
ser., XXIX (1976), pp. 14–37.
Prange, W., ‘Das Adlige Gut in Schleswig-Holstein im 18. Jahrhundert’, in C. Degn and D. Lohmeier, eds.,
Staatsdienst und Menschlichkeit. Studien zur Adelskultur des späten 18. Jahrhunderts in Schleswig-Holstein und
Dänemark, Kieler Studien zur deutschen Literaturgeschichte, 14 (Neumünster, 1980), pp. 57–75.
vanBavel, B., Curtis, D. R., Dijkman, J., Hannaford,M., DeKeyzer,M., vanOnacker, E., and Soens, T., eds.,Disasters
and history: the vulnerability and resilience of past societies (New York, 2020).
Van den Berg, G. J., Pinger, P. R., and Schoch, J., ‘Instrumental variable estimation of the causal effect of hunger
early in life on health later in life’, Economic Journal, 126 (2016), pp. 465–506.
Van Leeuwen, M. and Maas, I., HISCLASS: a historical international social class scheme (Leuven, 2011).
Van Poppel, F., Jonker, M., and Mandemakers, K., ‘Differential infant and child mortality in three Dutch regions,
1812–1909, Economic History Review, LVIII (2005), pp. 272–309.
Vanhaute, E., ‘From famine to food crisis: what history can teach us about local and global subsistence crises’,
Journal of Peasant Studies, 38 (2011), pp. 47–65.
Vanhaute, E., ‘Famines: causes and impact,’ in B. Pritchard, R. Ortiz, and M. Shekar, eds., Routledge handbook on
food and nutrition security (2016), pp. 227–38.
Vanhaute, E. and Lambrecht, T., ‘Famine, exchange networks and the village community. A comparative anal-
ysis of the subsistence crises of the 1740s and the 1840s in Flanders’, Continuity and Change, 26 (2011),
pp. 155–86.
Vanhaute, E., Paping, R., and Ó Gráda, C., ‘The European subsistence crisis of 1845–1850: a comparative perspec-
tive’, in R. Paping, E. Vanhaute, and C. Ó Gráda, eds., When the potato failed. Causes and effects of the last
European dubsistence crisis, 1845–1850 (Turnhout, 2007), pp. 15–40.
Voutilainen, M., Poverty, inequality and the Finnish 1860s famine (Jyväskylä, 2016).
Voutilainen,M., ‘Income inequality and faminemortality: evidence from theFinnish famine of the 1860s’,Economic
History Review, 75 (2022), pp. 503–29.
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
24 LUST et al.
Willführ, K. P. and Störmer, C., ‘Social strata differentials in reproductive behavior among agricultural families in
the Krummhörn region (East Frisia, 1720–1874)’, Historical Life Course Studies, 2 (2015), pp. 58–85.
Wrigley, E. A. and Schofield, R., The population history of England 1541–1871 (Cambridge, 2002).
SUPPORT ING INFORMATION
Additional supporting information can be found online in the Supporting Information section at
the end of this article.
How to cite this article: Lust, K., Klesment, M., and Jaadla, H., ‘Social inequalities in
famine mortality in the manorial system of the tsarist Russian province of Livland in the
mid-1840s’, Economic History Review, (2023), pp. 1–24. https://doi.org/10.1111/ehr.13246
 14680289, 0, D
ow
nloaded from
 https://onlinelibrary.w
iley.com
/doi/10.1111/ehr.13246 by Test, W
iley O
nline Library on [24/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License