Explorations in Economic History 72 (2019) 23–37

Contents lists available at ScienceDirect

Explorations in Economic History

journal homepage: www.elsevier.com/locate/eeh

Russian real wages before and after 1917 ✩

Robert C. Allen a,∗, Ekaterina Khaustova b
a
Faculty of Social Science, New York University Abu Dhabi, United Arab Emirates
b
Arizona State University, United States

a b s t r a c t

The paper measures real wages in St Petersburg, Moscow, and Kursk over 1853-1937. Workers in construction and large scale industry are studied.
For the imperial period and the NEP, new series of prices are collected from archival and printed sources, and these radically revise previous
measures of inflation. Russian living standards grew little between 1853 and 1913, but doubled between 1913 and 1928 due to the exchange rate,
price, and employment policies followed by the regime. Real wages dropped to their pre-War level between 1928 and 1937, as the social surplus
was mobilized for the industrialization drive.

1. Introduction

The working class is the protagonist of the Russian Revolution. In Marxist accounts, the exploitation of the proletariat increased as
the Russian economy expanded before the First World War, and that increasing exploitation was the major background factor leading
to the overthrow of Russian capitalism. Is there, in fact, evidence that real wages declined or that exploitation in any other sense
increased prior to 1913? There have been surprisingly few attempts to measure changes in real wages in this period.1 As socialism
developed after 1917, capitalist exploitation, presumably, ended, so the standard of living of workers should have risen.2 A substantial
literature has found that living standards fell after 1928,3 but what about the years of the NEP? This paper uses newly assembled
data bases to answer these questions.
Much of this paper considers these questions in the traditional way–that is, in a purely Russian context. However, economic history
has become increasingly a global subject in which international comparisons play a leading role. The history of living standards is an
important example of this development. In this paper, we view Russian real wages from an international perspective that modifies
the traditional Russian perspective in important ways.
We are concerned with three periods–the pre-war Imperial, the early Soviet and NEP, and the first two Five Year Plans (1928–
37). The wage and price information covering the three periods we are concerned with is limited in quantity, uneven in quality,
and difficult to compare. There are no price indices covering the whole period. Price indices have been computed for the various
subperiods, but the details of their construction are usually scanty, and one important index is wrong. An important contribution of
this paper is to assemble data sets of prices for many commodities that cover the entire period 1853–1937 and use these to compute
new price indices that cover the whole period in a consistent manner.

✩
We thank New York University Abu Dhabi and the Russian Presidential Academy of National Economy and Public Administration for research
support.
∗
Corresponding author.
E-mail addresses: bob.allen@nyu.edu (R.C. Allen), ekhaustova@mail.ru (E. Khaustova).
1
Mironov (2010) is the most recent. See also Allen (2003, pp. 37–46) and Borodkin et al. (2008). Gregory (1980) studies the growth of consumption
per head, which covers the whole society and not just workers.
2
Most research is concerned with either what happened before the First World War or what happened after the start of the Five Year Plans in
1928. Zaleski (1971, p. 390) did link real wage indices to span the period 1900–1927/8 but never discussed the results!
3
Chapman (1954, 1963), Bergson (1961), Hunter and Szyrmer (1992). Allen (2003) reaches more optimistic conclusions.

https://doi.org/10.1016/j.eeh.2018.12.001
Received 6 August 2017; Received in revised form 6 December 2018; Accepted 7 December 2018
Available online 9 December 2018
0014-4983/© 2018 Elsevier Inc. All rights reserved.
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

Each period we study presents different challenges in view of the presently available data and the historiography. For the pre-war
Imperial period (1853–1913), the only retail prices available to scholars are some time series for bread and flour in St Petersburg
(Borodkin, 2001). We have checked those series against original sources, and - more importantly–expanded coverage to include dozens
of items in the capital as well as in Moscow and Kursk. This allows us to compute new consumer price indices. This is important since
there is only one retail price index available in the literature that covers the entire period.4 It was created by the Soviet economist
Dalmatov (Mironov 2010, p. 52) and has become the standard measure of inflation in this period: Allen (2003, pp. 37–46) and
Mironov (2010, pp. 52–3), for instance, both used it to deflate wages in their studies of the standard of living. We have only a sketchy
idea of the methodology used by Dalmatov. We have not been able to replicate his index, and, indeed, it looks wildly unrealistic
for the years 1853–1880. As a result, all existing real wage indices for Russia in this period are in error. We propose new indices to
replace them.
For the early Soviet period, which ends in 1928 (but for which we use 1913 as the base of our indices), we have made less extensive
revisions to existing information. We have not collected prices for the period of the Civil War or the subsequent hyperinflation due
to the instability of those times. We have assembled new information on prices at the end of the New Economic Policy that augments
the work of Chapman (1963, pp. 190–5), and we use that information to compute price and real wage indices for 1928.
The Conjuncture Institute in Moscow produced a consumer price index, and it has been used to measure real wage changes over
this period (Kingsbury and Fairchild, 1931, p. 78, Zaleski, 1971, pp. 390, 396). Deflating average earnings in large scale industry by
this consumer price index indicates that real wages were about 35% greater in 1928 than they had been in 1913. This substantial
increase has remarkably received very little attention from historians of the period. The coverage and weighting of the Conjuncture
Institutes’ index were changed repeatedly in the period (Zaleski, 1955), and this raises questions about its adequacy. Our price indices
indicate lower rates of inflation than the Conjuncture index and correspondingly higher rates of increase of real wages during the
NEP.
For the first two Five Year Plans (1928–1937) we rely on the prices collected by Chapman (1963) and Zaleski (1955), but integrated
them with our earlier data to construct uniform indices covering the entire period. We do not report price or real wage indices for
years between 1928 and 1937 since most prices were too low to clear markets, so goods were not available at recorded prices. In
1928, some prices were market clearing in some markets but not in others, and we have made adjustments to deal with those issues, as
will be explained. Following the price reform of 1936, when prices were raised to market clearing levels, price and real wage indices
again become meaningful and we compute them for 1937 using the extensive database compiled by Chapman (1963, pp. 190–5).
The data we have assembled points to striking conclusions. In general, with the possible exception of St. Petersburg, there was
little growth in real wages by any measure or in any of our cities between 1853 and 1913. Given the increase in per capita GDP and
output per worker in manufacturing, it is clear that the benefits of growth did not ‘trickle down’ to the working class during this
period. A great change occurred between 1913 and 1928, however, for real wages increased by as much as 100% or more depending
on the measure, while output per worker grew very little. It looks like workers gained a lot through redistribution of income. Their
good fortune did not last long, however. The period 1928–37 witnessed extreme inflation, wages lagged behind prices, and real wages
dropped accordingly. In 1937, working class living standards were back to the level of 1880.

2. Data sources

Our investigation is based on a wide ranging collection of new data for the Imperial period and the NEP. The data include wage
rates and the prices of consumer goods.5 Data were derived from official sources, either in print or manuscript.
The official collection of price data began in the eighteenth century. Administrative practices were codified in 1802 during the
reign of Alexander I when a decree "On drawing of reference prices for provisions in the provinces" was adopted.6 This decree reviewed
the 18th century practice of collecting data and stipulated that prices be approved on a monthly basis in all counties of provinces.
Local authorities were required to determine the reference prices for food and forage to be used as guidelines for the provisioning
of troops. Police collected prices under the authority of provincial governors from merchants and markets. Multiple sources were
surveyed, and prices collected at weekly and monthly intervals. Monthly averages were computed from these data. The reference
prices were reviewed and approved by meetings of leaders of the local nobility, police chiefs, magistrates, mayors, and representatives
of guilds and merchants (Mironov, 1985, p. 20). The prices were issued in the format of statistical surveys or /handbooks.7
The earliest price handbook we found was the Moscow survey for 1801, although our database begins later in 1824 when the
surviving records become fuller. The early archival sources for Saint Petersburg, Moscow and provincial centres followed a consistent
format in line with the 1802 decree. These early sources were hand written.
From the mid-nineteenth century, the data were issued annually in printed form by Tsentralya gorodskaya yprava (Central city
Government), first by St. Petersburg in 1853 and then by Moscow in 1871. Other cities followed. These publications presented reports
of local administrative bodies on the state of weekly, monthly, seasonal and annual prices in the provinces. The purpose of these

4
For the period 1885–1913, there are half a dozen retail price indices for Moscow and St Petersburg and one pseudo retail price index (averaging
wholesale prices with retail weights). (Strumilin, 1966, pp. 89–90, 1967, pp. 431–2). These all show similar rates of inflation and agree our own
calculations.
5
Full references to the data are listed in the Appendix. These data are available online at Allen and Khaustova (2018).
6
Polnoe Sobranie Zakonov Rossiiskoi Imperii-1. - T. XXVII. – Number 20224. – page. 101.
7
While collecting the data, prices recorded in the archives were compared to those in the published volumes to confirm the accuracy of the
published material.

24
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

collections was to provide state institutions with reference prices to guide their purchases and provide the state with a statistical
record.8
By the middle of the nineteenth century, the motives for collecting wage and price data became broader as they intersected with
fundamental questions of economic development. On the one hand, industrial development was an objective in Russia as elsewhere.
On the other hand, the concentration of large numbers of poor workers in dense settlements posed familiar problems of sanitation,
housing, poverty–and potential insurrection. Russian officials wanted the development without the revolution and so had to address
questions of well-being like their western counterparts. This led to similar initiatives and to the collection of similar data to monitor
the situation. It was this context that led states throughout North America and Europe to collect wage and price data from the
mid-nineteenth century onwards (Brower 1990; Kingston-Mann 2005; Lincoln 1986; Smith-Peter 2007).
Statisticians and historians have assessed these price handbooks to determine their reliability. Lositskii (1899, p. 218) investigated
the Saint Petersburg “Vedomosti spravochnikh tsen” for the second half of 19th Century. Rykachev (1911) discussed the strengths
and weaknesses of “Vedomosti Spravochnikh tsen” for Moscow and Saint Petersburg in his work “Tseni na khleb i tryd v Saint
Petersburgе za 58 let.” Mironov (1985, p. 23–5) compared bread prices from ‘Vedomosti spravochnikh tsen’ to publications such
as “Vestnik Finansov Promishlennosti i Torgovli”, Zyrnali Zemskikh sobraniy, and archival sources like hospital account books, and
private collections. The general conclusion of these investigations is that the Vedomosti spravochnikh tsen are a reliable and accurate
source.
While previous work like Mironov’s has concentrated on St. Petersburg, we collected data for Moscow and Kursk as well as the
capital. Mironov collected new data for St Petersburg for 1703–1853 but relied on an outmoded Soviet index to measure inflation
between 1853 and 1913. We have collected original data for St. Petersburg to compute better weighted indices for that period.
Moscow and St. Petersburg were the two largest industrial centres - each with a population over one million in the census of 1897 -
so it was essential to include both of them in view of our focus on the industrial proletariat.
We added Kursk precisely because it was a small, provincial manufacturing city. In 1897, its population was 76 thousand making
it the 19th largest city (out of 54) in European Russia. Flour milling, sugar refining, brewing, and candle making were its major
industries. By the end of the nineteenth century, machine shops were also established. While we do not claim that it represents all of
Russia outside of the two leading cities, we do think it is a weather vane that may indicate trends in outlying districts that warrant
further study.
Our information for Moscow and St Petersburg before the First World War came largely from municipal statistical handbooks:
Vedomosti spravochnikh tsen v Saint Petersburge na pripasi, materiali, platy rabochim i prochee, izdavaemie Saint Petersburg gorodskoi
ypravoi. (Saint Petersburg 1854–1917) and Vedomosti o spravochnikh tsenax na pripasi I materiali v Moskve (Moscow 1871–1917),
Ezemesyachnii statisticheskii bylleten po gorody Moskve (1892–1917). The 1899 volume for St. Petersburg and the 1883 volume for
Moscow were missing from the collection in the Russian National Library, so we interpolated the gaps.
These sources include monthly and sometimes weekly data. We collected monthly data for bread and for the wages of carpen-
ters and labourers. For other items, we collected quarterly data. The data were then averaged to create annual series that were
representative of conditions throughout the year.
The Moscow yearbooks began in 1871. For earlier years, we collected prices and wages from Tsentralniy Gosydarstvennii Arkhiv
Goroda Moskvi. The major data source was Raporti torgovikh starost o tsenax, Spravochnie tseni vedomstva Moskovskoi dvortsovoi konori
a tak ze ministerstva Imperatorskogo Dvora, O dostavlenii spravochnikh tsen i materialov, Delo Moskovskoi rasporiaditelnoi dymi o dostavlenii
spravochnikh tsen, Delo Moskovskoi gorodskoi dymi. With this information, we could trace the same types and qualities of goods and
labour back to the 1820s.
All information for Kursk was obtained from records in the archive Gosydarstvennii Arkhiv Kyrskoi Oblasti. The main information
source was the same kind of statistical hand book (Vedomosti spravochnikh tsen) that we used for Moscow and St Petersburg, and we
treated the information in the same way.
How reliable are these prices? We have considered their provenance above, and it is reassuring. Another approach is statistical.
We can ask if the price structure makes sense economically: if not, the prices are unreliable. This inquiry is best pursued for bread
and flour since the prices of the raw materials–rye and wheat–show cycles, which provide a basis for investigation. We expect that
more processed goods like bread should cost more than less processed goods like rye, that the fluctuations in the series should be
correlated, and the degree of correlation should decline at higher stages of fabrication since the share in cost of the raw agricultural
product declines as the degree of processing increases. We have investigated this for wheat and rye and their products in St Petersburg
and Moscow, and our expectations are confirmed.9 The farm price of wheat was the lowest; the wholesale price of wheat in the cities
was somewhat higher, wholesale flour prices were higher still, and the retail flour and bread prices highest of all. The bread price, in
particular, is more stable than the others but is still affected by large shifts in the price of less fabricated products. It is important for
these comparisons that the prices come from different sources - this is not an exercise in self - validation. The statistical behaviour of
these prices - and they receive a large share of the weight in all of our price indices - is what one would expect if the data are reliable.
To see how living standards after 1917 compared to the Imperial period, we extended our wage and price indices to 1937. We do
not have data for the First World War nor the first years of Soviet rule, so we can say nothing about them. We have not included the
hyper inflation of 1922–3 since prices changed too rapidly to be able to match wages with the corresponding prices. We collected

8
Svininyh (2013) and http://www.gks.ru/wps/wcm/connect/rosstat_main/rosstat/ru/about/history/de354a804bdd96568fa2bfc25c10f730.
9
Farm prices from Gregory (1982, p. 234), wholesale prices from Russia,Ministerstvo torgovli i promyshelnnosti, Svod tovarnykh tsen na glavnykh
russkikh i inostrannikh rynkakh za 1915 god. Sources of retail prices are details in our data appendix.

25
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

new data for 1924–9 covering the New Economic Policy and two years of the first Five Year Plan and we include data for 1937, the
end of the second plan.
Conventional real wage indices like ours are a ratio of wages or earnings to a price index. Such an index is a meaningful indicator
of living standards only if goods are available at the prices in the price index. This condition was frequently violated in the Soviet
Union where official prices were often set below market clearing prices. How serious a problem is this for our indices?
In studying prices in the Soviet period, it is important to distinguish between state and cooperative shops, on the one hand, and
private markets, on the other. In the former, prices were set administratively, so shortages appeared when prices were too low. On
the other hand, prices on private markets where peasants sold food and handicrafts were market clearing prices. Equality of prices
across all the marketing channels is an indicator that prices in state and cooperative shops were, in fact, market clearing prices. When
private market prices exceeded official prices, then there were shortages in state shops at the official prices. Price indices aggregating
official prices when goods are unavailable at those prices would understate the cost of living.
For most years in the 1930s, private market prices greatly exceeded prices in state and cooperative shops. Consequently, we have
not computed real wages for these years. A price reform in 1936, however, raised prices to match those on the free market. The
conclusion of many students of the Soviet economy is that official prices in 1937 were market clearing prices, and that is one reason
it has become a bench market in Soviet economic studies (Chapman 1963, pp. 19–22; Bergson 1961, pp. 157–61; Karcz 1957). We
calculate real wage indices for 1937.
To study wages at the height of the New Economic Policy (1924–9), we collected data from Ezemesyachnii statisticheskii bylleten
po gorody Moskve I Moskovskoi Gybernii where we found prices for many goods sold on the private market as well as in state and
cooperative shops. There were some differences across the marketing channels. In 1928, beef was 3% more expensive in private
markets than in cooperative shops, while rye bread was 5% more expensive. The biggest gap that affects our indices concerned
potatoes where private prices exceeded cooperative prices by 40%. The private market accounted for 11%, 5%, and 76% of sales of
these items. We calculated weighted averages of the prices of major foods where the weights reflected sales in 1928.10 While the price
gap was substantial for potatoes and private market sales accounted for most transactions, the overall impact of price divergences
across marketing channels was minor in our price indices.11

3. The history of nominal wages

We have collected data for four types of workers–building craftsmen (carpenters, bricklayer, or masons), building labourers, and
employees in large scale manufacturing, and employees in cotton textile mills.
Building craftsmen and labourers figure in virtually all studies of historical real wages since every country had a construction
industry, and building workers were hired by the institutions whose records are the main source of historical wages and prices
(Beveridge 1939, pp. xxi-xxvi, xlix-lii). While payment in building was often purely monetary, there is a question of how continuously
the workers were employed. In our comparisons, we assumed they worked a full year. This is controversial, but there is evidence in
its favour. First, when our sources for the Imperial period give monthly wage rates, they report rates for every month (although they
do not indicate employment levels by month). Second, in the Soviet period, official statistics report average annual earnings in the
construction industry. Dividing by the daily wage of building craftsmen indicates that the average work year was about 260 days.
Third, comparisons with wages in manufacturing provide some reassurance.
Our wages for workers in large scale industry and cotton mill workers are national averages. Moscow and St Petersburg were
centres of industrial production, and this improves comparability between the manufacturing and construction wages. The wage data
for manufacturing are also average annual earnings for employees. The annual averages turn out to be similar to our estimates of the
full, time-full year earnings of building labourers.
We begin with building wages for Russian cities before the First World War. The earnings of labourers were similar in all three
cities and grew little from about 1879 to 1910. Carpenters, of course, earned more than labourers.12 The wages of carpenters were
similar in Moscow and St Petersburg and more than doubled over the period. The wages of carpenters in Kursk were generally the
lowest and grew less than those in the big cities (Fig. 1). By the early twentieth century, carpenters in Kursk earned little more than
labourers. Skilled building workers in Kursk were not keeping up with their counterparts in the big cities in this period. The similarity
of wages in Moscow and St Petersburg raises the possibility that their labour markets were integrated, while the lower carpenter’s
wages in Kursk suggests that integration did not extend to all cities.

4. The cost of subsistence

To assess changes in the purchasing power of wages, the nominal wages just reported must be deflated by an index of prices of
consumer goods and services. There are many approaches to constructing such indices, and we explore several to assess the robustness
of our results.

10
“Comparison of Real Wages in Various Cities,” International Labour Review, 1928, pp. 659.
11
There was a possible mitigating factor. When food was in short supply, there was probably more home grown food either raised in urban areas
or acquired from relatives in the country. There are no comprehensive statistics on this subject. Were it important, then welfare would have been
higher in the Soviet period than our calculations indicate.
12
Plots of these series are available in Allen and Khaustova (2017).

26
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

2
1.8
1.6

rubles per day

1.4
1.2
1
0.8
0.6
0.4
1853 1863 1873 1883 1893 1903 1913

Moscow Kursk
Saint Petersburg

Fig. 1. Russian carpenters’ wages before World War I (rubles per day).
Source: see text and data appendix.

The first approach uses conventional price indices. These require two sorts of data. The first are time series of prices paid by
consumers. These are converted to price relatives by divided all prices in each series by the price in the base year–in our case
1913. The second source of information is one (or more) surveys of consumer spending to establish the shares of spending devoted
to each item.13 Mironov (2010, p. 54) reports expenditure by broad category derived from Prokopovich’s survey of 1907–814 and
Chapman (1963, p. 70) reports similar results for 1937. Our index numbers cover 85–88% of the reported spending as we must omit
some minor categories for which there are no price series.15 Food comprises a large share of spending in both surveys (55% in 1905
and 60% in 1937), while other categories differed substantially between the two dates. Rent drops from 17% in 1905 to 2% in 1937,
while heating and lighting fell from 5% to 2%. Conversely, alcohol and tobacco rose from 6% to 12%, while clothing and shoes
nudged up from 14% to 17%. These broad categories must in some cases be subdivided into categories defined in terms of actual
items consumed like rye bread and rye flour. Chapman reports these subdivisions for 1937, which we use, and we also apply them
to Prokopovich’s survey for which we lack this detail. Despite the differences in category weights, the movement of the indices is
similar, as we will see.
The second approach is the now widely used subsistence basket approach (Allen, 2001). This is a methodology that is specifically
intended for international comparisons as well as measurement over time, so the results derived here for Russia are directly comparable
to those for other countries. We take advantage of this possibility later in the paper to assess Russian real wages from an international
point of view.
Urban households in Russia as well as in many countries averaged about four people,16 so the cost of subsistence for the house-
hold is set at four times the cost of subsistence for the average person. Subsistence baskets are defined that meet nutritional needs
inexpensively and that reflect the food habits of poor people around the world. These are ‘poverty line’ budgets that do not capture
spending patterns in all particulars. Details of the budgets vary between countries to reflect local circumstances, but the overall
structure is the same, and they are intended to represent equivalent levels of well-being (Table 1). All of the budgets contain four
food types–carbohydrates, vegetables, meat, and fat. The carbohydrate is chosen to reflect the predominant food of the country and
the form in which it was usually purchased. Table 1 shows two budgets for Russia, one defined in terms of rye flour and the other
in terms of rye bread. The bread or flour is the main source of calories. The diet also includes 50 kg of potatoes, 5 kg of the cheapest
grade of meat (in this case mutton) and 3 kg of butter or oil. The quantity of the carbohydrate is set at the level that gives a total
dietary calorie content of 2100 calories per day. This represents the US Department of Agriculture’s (2010, p. 44) food security line. It
is intended to be a society-wide average providing many more calories for men doing heavy work and many fewer calories for young
children. On average, everyone receives enough to grow or to work, as appropriate. A quasi-vegetarian diet with no alcohol is the
typical fare of the world’s poor, and, as it happens, was barely affordable by Russia’s low wage workers in the nineteenth century.

13
We note here one issue that we do not pursue, namely, the impact of new goods on the cost of living. In the period under consideration, these
included railway and tram rides, electricity, the cinema, and (possibly) books since, while not new, they became accessible to people who could not
previously read. We do not have good time series of prices of these items. According to Chapman (1963, pp. 70, 314–7), they accounted for 2.2%
of consumer spending in 1928. The small share means that ignoring new goods cannot have a large effect on our conclusions.
14
We use the results of the survey for workers with families.
15
Omitted categories include ‘cultural-education necessities’, ‘socio-political necessities,’ ‘religious necessities,’ ‘remittances to relatives,’ and
‘other.’ Chapman (1963, pp. 312–7) provides weights for individual items in 1937, and these have been adopted. The 1937 weights have also
been applied to Prokopovich’s broad categories to compute weights for individual goods like rye flour, rye bread, and so forth.
16
For India, for instance, Shirras (1923, p. 23–5) shows most working class families in Bombay occupied a single room with an average of 4.03
occupants in 1921. Bombay Labour Office (1928, p. 19) shows an average of 3. 74 people per room in Ahmedabad. Propkopovich (1909, p. 10) reports
an average of 3.78 people per family.

27
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

Table 1
Subsistence and Basic Needs Baskets of Goods (Kilograms per per-
son per year unless otherwise stated).

Subsistence Basic Basic

Bread flour Needs 1937 Needs 1913

Food (kg)
Rye bread 267 257
Rye flour 213 175
Lamb 5 5 12 16
Potatoes 50 50 76 71
Butter 3 3
Oil 3 9
Nonfood
Soap (kg) 1.3 1.3 1.3 1.3
Cloth (metre) 4 4 53 53
Kerosene(litres) 28 28 28 28
Fuel(mBTU) 3 3 12 12
Housing (sq m) 5% 5% 3 3

Source: see text

Notes:
all food quantities are kilograms per person per year.
cloth -metres of cloth per person per year. Cloth is cheap cotton
lighting millions of BTUs per person per year.
heating is from coal.
Housing is 5% of the other costs in the subsistence baskets but it
equals the rental cost of 3 sq m in the basic needs baskets.
Units
kg = kilogram
mBTU = million British Thermal Units
metre = metre
sq m = square metre.

The non-food items in the budgets include clothing, lighting, fuel for cooking and heating, and housing, so requirements for them
must also be set. With the rye flour and bread subsistence budgets, we follow the standard practice in the subsistence basket literature
and set the quantities shown in Table 1. In this approach, rent is assumed to have been 5% of the total spent on food and non-food
goods.
The third approach elaborates the second and is the Basic Needs Poverty Line proposed in Allen (2017). All elements of the
calculation are revised. First, the diet is determined more systematically by using linear programming. Items in the diet and their
quantities are chosen to minimize the cost of achieving a nutritional status defined in terms of calories, protein, fat, iron, thiamine,
folate, niacin, vitamins B12 and C. The vitamins and minerals are those need to prevent anaemia, beri-beri, pellagra, and scurvy. Diets
defined in this way explain much of the food behaviour of the poorest people in the world today. The approach is operationalized for
Russia by solving the linear program diet problem using 1913 prices for a long list of foods17 from Zaleski (1955, p. 376) and again
with 1937 prices for a similar list of foods18 from Chapman (1963, pp. 190–5).
Second, Prokopovich’s budget survey for industrial workers in St. Petersburg was used to define the consumption of clothing, foot
ware, bedding, fuel for heating and cooking, and kerosene for lighting. Expenditures for clothing, foot ware, and bedding for workers
earning 300–400 rubles per year - near the bottom of the earnings distribution - were divided by the retail price of cheap cotton cloth
to convert the expenditures into ‘metre equivalents’ that we could price over time. Expenditures for fuel and lighting were converted
to quantities by dividing by the retail prices of coal and kerosene. Expenditure data for fuel and lighting were available only for the
average worker. To estimate consumption of workers earning 300–400 rubles per year, quantities were reduced in proportion to the
corresponding ratio for clothing, foot ware, and bedding. In the case of fuel, which is much the greater expense, this calculation was
supported by engineering calculations on the fuel required to heat a poorly insulated room with the temperature pattern over the
year in St Petersburg (Allen, 2017 pp. 3710–3). These calculations imply much larger consumption of these items than is assumed in
the subsistence baskets.
Third, housing expense is estimated on the assumption that minimal accommodation is 3 square metres of floor space per person.
Three square metres per person is typical of slums in many third world countries.19 This space is valued at prevailing rents.

17 Rye flour, rye bread, wheat flour, wheat bread, rice, dried peas, beef, mutton, chicken, herring, butter, sugar, sunflower oil, margarine, lard,

milk, cheese, eggs, potatoes, cabbage, sauerkraut, onions, beets, apples, beer, vodka
18
The 1937 list was the same as the 1913 list except that 1937 also included sausage and fresh water fish but omitted beer.
19
In Bombay in 1921, for instance, cotton mill workers lived one family to a room of 13.3 square metres giving each person 2.3 sq metres of space.
In Ahmedabad in the 1920s, the average was 3.6 square metres per head, in Shanghai in 1952, the average resident had 3.4 square metres, and in
the slums of Nairobi today, the rate is 3 sq metres per person. UN Habitat (2003b) proffered a definition of overcrowding as more than 2 people

28
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
1853 1863 1873 1883 1893 1903 1913
flour subsist Dalmatov bread subsist
cpi 1907/8 LP 1937 cpi 1937
LP 1913

Fig. 2. Comparison of consumer price indices for St. Petersburg.

Source: see text and data appendix.
Note: ‘flour subsist’ is the cost of the subsistence basket using flour, ‘bread subsist’ is the cost of the subsistence basket using bread. ‘Cpi 1907/8’ is
a linear cost of living index using weights from Prokopovich’s budget survey of 1907/8, and ‘cpi 1937’ is the same construction using 1937 weights
from Chapman. ‘LP 1913’ and ‘LP 1937’ are the basic needs poverty lines with diets from linear programs run on 1913 and 1937 samples of prices.

We thus have six price indices to measure retail prices in Russia - linear weighted averages of price relatives using 1907–8 and
1937 weights, standard subsistence budgets whose diets are based on rye flour and rye bread, and more sophisticated basic needs
budgets whose linear programming diets reflect 1913 and 1937 prices. We have calculated these for St. Petersburg and Moscow. For
Kursk, we have less extensive price information and calculated the cost of the subsistence budgets and the basic needs budgets.
We begin with the calculations for St. Petersburg in order to compare our results to the Dalmatov index, which was calculated
with prices from that city. Fig. 2 plots our six St Petersburg indices and the Dalmatov index. The graph points to two important
conclusions. First, although there is some disagreement on a year to year basis among our index numbers–the indices based on flour
prices are more volatile than those based on bread prices–they concur in measuring the long run rate of inflation. Second, all of our
indices disagree with the Dalmatov index, especially for the years before 1883. All we know about the Dalmatov index is that it is an
unweighted average of the prices of 26 goods20 ; however, an unweighted average of our retail prices does not show a similar trend.
We have not found any way to replicate Dalmatov’s work, and so we regard it as fundamentally unsound.
It has become standard in cost of living measurement to use ‘superlative’ index numbers that are based on two sets of expenditure
weights since these index numbers track the true cost of living index more closely. We have incorporated that principle in our work
by averaging the price indices in each of our approaches–the consumer price indices, the subsistence baskets and the Basic Needs
Poverty Lines. The averaged indices are not overly dependent on a single set of weights. The resulting indices are shown in Fig. 3 for
St Petersburg. The inflation rate over the whole period was similar Moscow, although the divergence among the indices along the
way was greater in Moscow. Over this period, Russian commodity markets were becoming more integrated as the railway network
was built (Metzer, 1974). In both cities the subsistence baskets were the most volatile, as was the consumer price index in Moscow.
In both cities, the Basic Needs Poverty Line traces out a middle way, so we will use it as our real wage deflator.
We can extend all of these indices to 1928 (the end of the New Economic Policy and the start of the First Five Year Plan) and 1937
(the end of the Second Five Year Plan). The biggest divergence among the indices occurs in the period 1913–28. In St Petersburg,
for instance, the Basic Needs index indicates a 29% rise in the price level, the average subsistence basket shows a 43% rise, while
the linear price indices show at 79% rise. The latter gets closest to the inflation shown by the Conjuncture Institute index of 100%
(Kingsbury and Fairchild, 1931, p. 78). Evidently, the choice of index has a bearing on the measured changes in real wages in this
period, which we will indicate.
There is considerable agreement among our price indices with respect to the price level in 1937 vis-a-vis 1913 since the index
that showed the greatest inflation between 1913 and 1928 shows the least between 1928 and 1937. In St Petersburg, the average of
the conventional cpi shows that the price level increased by a factor of 14.7, the average subsistence basket pointed to an inflation
of 13.7 fold, while the average of the Bask Needs Poverty Lines grew 12.2 fold. In Moscow, the corresponding increases were 15.4,
12.9, and 11.2. We will consider the implications of this variation for real wage measurement shortly.

per room or less than 5 sq metres each. In 2010, however, this standard was revised to 3 people per room, implying less than 3.3 square metres per
person. Shirras (1923, p. 25), Bombay Labour Office (1928, p. 19), Sun and Linyun (2011, p. 57), UN-Habitat (2003a, p. 237), UN-Habitat (2003b,
p.12), UN-Habitat (2010, p. 16)
20
Mironov (2010, pp. 52–3) provides a good discussion of the Dalmatov index.

29
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

1.1

0.9

0.8

0.7

0.6
1853 1863 1873 1883 1893 1903 1913

subsistence cpi LP

Fig. 3. Multi-share price indices, St. Petersburg.

Source: see text and data appendix.

3.5
3
multiples of subsistence

2.5
2
1.5
1
0.5
0
1853 1873 1893 1913 1933

carpenter cotton mills labourer

Fig. 4. Real wages in Moscow, 1853–1937.

These series use the average Basic Needs Poverty Line (LP) as the deflator.
Source: see text and data appendix.

5. Trends in real wages

Our approach to measuring the standard of living uses subsistence baskets and welfare ratios. The welfare ratio equals the annual
earnings of the worker concerned divided by the cost of maintaining a family at subsistence (Blackorby and Donaldson, 1987; Allen,
2001). If the welfare ratio equals one, then the worker earns just enough to keep his family at subsistence. If the ratio is less than
one, then earnings are inadequate even for that low standard and either painful decreases in spending are required or more income
has to be generated by increasing the family’s scale of paid work. If the ratio exceeds one, then there is a surplus over subsistence.
This is often realised by increasing the quality of food consumed as well as by purchasing more goods and housing.
Calculating the welfare ratio requires annual earnings. In the case of cotton mill operatives and employees in large scale industry,
the wage data are average annual earnings, so no adjustment is required. In the case of building workers, however, pre-1917 data
are day wages, which must be multiplied by the number of days worked in a year. We assume ‘full time-full year’ and take that to
have been 250 days, which is commonly used in international comparisons. The calculations with cotton operatives are a check on
this assumption.
The cost of living that we have calculated is the annual cost of an individual. On the assumption that there are four people in a
household–more precisely, that the income earned by the worker supports or herself and three dependents–we calculate the annual
subsistence cost of a family. The ratio of annual income to annual subsistence cost is our measure of the real wage.
We begin by examining the trend in these ratios for Russian workers (Figs. 4–6). Our three cities show three different patterns.
After an increase from very low levels in the late 1850s, real wages in Moscow were flat until the First World War. Labourers earned
bare subsistence, while carpenters took home about 50% more. The real earnings of workers in cotton factories in large scale industry

30
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

2.5

multiples of subsistence
2

1.5

0.5

0
1853 1873 1893 1913 1933

carpenter labourer

Fig. 5. Real wages in St. Petersburg, 1853–1913.

Note: these series use the average Basic Needs Poverty Line (LP) as the deflator.
Source: see text and data appendix.

2.5
multiples of subsistence

1.5

0.5

0
1853 1873 1893 1913 1933

carpenter labourer

Fig. 6. Real Wages in Kursk, 1853–1937.

Note: For Kursk, wages were deflated by the bread version of the Basic Needs Poverty line since we do not have a flour price series for Kursk.
Source: see text and data appendix.

- a Russia-wide index - are also shown on this graph, and it is reassuring that their real incomes are closely in line with those of
building labourers in Moscow. This provides some assurance that the real wages of the building workers, which were calculated on
the assumption of full year employment, are not seriously misleading.
St Petersburg presents a mildly more optimistic picture than Moscow. In the capital, the real wages of unskilled workers fluctuated
around subsistence for most of the period, as they did in Moscow. While the real wages of carpenters were 50% higher in the 1880s,
they gradually trended upwards averaging around 1.75 times subsistence until the First World War. In the 1860s real wages were
lower in St. Petersburg than in Moscow but ended up considerably higher.
The pattern for real wage changes that we compute is very different from that pattern reported by Mironov (2010, pp. 56–7) for
carpenters in St. Petersburg. He concluded that real wages fell from the 1850s to the 1880s when they began to rise, reaching the
same level at the outbreak of the First World War as they had achieved in the middle of the nineteenth century. In contrast, we find
rising real wages across the period. The difference reflects the difference in consumer price indices, previously discussed.
In contrast to St. Petersburg or even Moscow, the labour market in Kursk was much less favourable. In Kursk, there was scarcely
any indication of long run improvement in real wages, labourers often earned less than the cost of subsistence, and craftsmen often
earned no more than unskilled labourers. Anyone with skill would have had a strong incentive to move to either of the major cities.
What standard should we use to judge the history of real wages in Russia? One possibility is to compare the change in real wages to
the change in output per worker, for that shows whether workers were maintaining their share of the economic pie. Building workers
in Moscow and Kursk did badly by this criterion since their real wages did not rise even though GDP per head was increasing.
The growth in the real wage in St. Petersburg, however, came close to matching the growth in GDP per capita. Not only are these

31
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

1200

1000

1913 rubles/year
800

600

400

200

0
1885 1895 1905 1915 1925 1935
real Value Added in Industry/worker
real wage in industry

Fig. 7. Real Value Added per Worker and the Real Wage in Industry.
Real value added in industry in 1913 equals value added in large scale industry on Soviet interwar territory according to Falkus (1968, p. 62). Real
value added for earlier years was computed by carrying this 1913 figure backwards with Kafengauz’s ‘expanded index’ (Gregory, 1997, p. 198).
Later years were extrapolated forward with Markevich and Harrison (2013, Table 5).
Employment in large scale industry from Crisp (1978, pp. 348–9) and Nutter (1962, p. 504).
Source: Real wage in industry was average annual earnings in large scale industry (before 1917 without Poland and Finland) deflated with our
average basic needs price index for Moscow.

comparisons muddled by the different experience of workers in different cities, but GDP includes the very large agricultural sector,
and changes in its circumstances may swamp other factors influencing distribution.
We can eliminate agriculture and get a comparison that focuses on income distribution in industry by comparing the growth in
real value added per worker in industry to real annual earnings per worker in industry (Fig. 7). The comparison in this case is more
exact since the work force is the same and value added equals wages plus profits, so any shortfall in wages was a gain for capitalists.
And there was such a shortfall. The average real wage was flat from 1885 to World War I, while value added per worker doubled.
The share of value added going to industrial workers dropped from 54% to 25%. The gains from growth were going to capitalists
rather than workers. Similar patterns have been observed in recent decades as well (Piketty, 2014).
While most of the gains from economic growth were going to groups other than workers in the late nineteenth and early twentieth
centuries, it is a very long way from that finding to the conclusion that rising inequality caused revolution in 1917. Nevertheless, the
Bolshevik Revolution was made in the name of the working class, and we ask, therefore, if it served their interests better than the
pattern of growth achieved under the Tsars.
The short answer is ‘yes.’ The greatest gains realized by Russian workers between 1860 and the Second World War occurred in
the 1920s. We can compare the earnings of carpenters and building labourers in 1928 with their counterparts in 1913, and we find
that the real wages of building workers in Moscow rose by about 70%, while the incomes of workers in large scale industry jumped
by a factor of 2.7 with cotton mill operatives going up by a factor of 2.8. These increases were greater than those realized during the
expansion of the late Imperial economy. It looks like Russian workers really were gainers from the 1917 Revolution.
This progress was the result of policy decisions, some of which allowed market forces to operating and some of which interfered
with them. The real wage is the ratio of the nominal wage to a price index, and policies affected both wages and prices. On the price
side, there were three important developments. The first was the monetary reform that ended the hyperinflation of the early 1920s.
This reform pegged the ruble to gold at the pre-revolution parity. Since the price level in the world at large had almost doubled,
world prices for Russia’s exports and imports, when converted to rubles, were close to twice what they had been in 1913, and, indeed,
Russian price indices were 80–100% greater than they had been before the First World War (Kingsbury and Fairchild (1931, pp. 78,
117). Soviet policy makers recognized that Russian prices could not be allowed to diverge from international prices at the exchange
rate if the Soviet Union hoped to be competitive in international markets (Carr and Davies, 1969, p. 707). Second, the macroeconomic
environment was further modified by specific policies that tended to raise real wages. The nationalization of the housing stock led
in some cases to the abolition of rents and in others to very low levels (Carr, 1958, I, pp. 404–5). This cut the rate of inflation and
led to a drop in the share of workers’ spending that went on housing from more than 20% to less than 10%. Food policy made a
further contribution. Prior to the War, the prices received by Russian farmers for wheat and rye were similar to the prices received
by American farmers. In 1927, the prices paid by the Soviet government for wheat and rye were approximately three quarters of
the price received by American farmers.21 While low procurement prices made peasants less willing to sell grain, they did allow the
production of flour and bread at lower prices than would otherwise have prevailed. In the United States wheat bread was 60% more

21
See Gregory (1982, p. 234) for pre-War Russian prices, Kontrolnye tsifry narodnofo khoziaistva na 1928/29 god, pp. 504–5, for Russian procurement
prices in the 1920s, and Historical Statistics of the United States, millennium edition online, series Da 719 and Da742 for USA prices. Exchange rate
calculations in Imperial period done via the rate on sterling in Denzel (2010, pp. 367–72).

32
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

Table 2
Real wages by Industry, 1913, 1927/8.

1913 1927/8 Relative increase

Metal 425 559 1.32

Coal 413 413 1.00
Petroleum 366 540 1.47
Leather 300 600 2.00
Wood working 264 436 1.65
Chemical 240 534 2.23
Paper 216 429 1.98
Textile 204 388 1.90
Alimentary-food 192 550 2.86
Average 300 464 1.55
CV 0.31 0.16

Source: nominal wages from Kingsbury and Fairchild

(1931, p.77).

expensive in the 1920s as it had been before the War.22 In the Soviet Union, increases in these prices were modest. This difference
between the countries may also have been due to direct subsidies in the Soviet manufacturing sector, but such policy is hard to detect.
These policies moderated inflation and contributed to rising real wages. They were particularly marked in our Basic Needs Poverty
Lines and subsistence baskets since bread and flour loomed so large in them.
Soviet labour policies also led to rising real wages. Workers were organized in trade unions. These were nominally independent
of the state, but were led by Bolsheviks. Under the NEP, they were charged with the inconsistent tasks of protecting the interests
of their members vis-a-vis their employers and with advancing state objectives (Carr, 1952, II, pp. 323–7). Trade Union congresses
met throughout the period and debated many issues of labour organization and wage policy. Should wage increases be linked to
productivity growth or should wage increases be directed to low income workers? How the balance of interests played out cannot be
easily traced in administrative histories, and historical discussions are often inconclusive (Carr, 1952, pp. 317–59, Carr and Davies,
1969, pp. 520–44, Dewar, 1956, pp.119–157). Bergson (1944) took a statistical approach and compared wage distributions in 1914–
1928 and concluded that wage equalization was substantial.
We come to a similar conclusion using industry data. Table 2 compares real wages in large scale industries in 1913 and 1927/8.
This table uses our average linear price index as the deflator, which generates the highest rate of inflation of any of our indices in
this period. The average real wage went up 55%, but advances were highly unequal. The industries with the highest earnings in 1913
were coal mining, metallurgy, and petroleum. Real wages rose less in these industries than in the lower paid industries in 1913. With
wage growth fastest in the poorly paid industries, wage inequality declined. This is shown statistically by the drop in the coefficient of
variation from 0.31 in 1913 to 0.16 in 1927/8. One reason for the differential experience was that the lower paid industries were the
most profitable during the NEP since they supplied consumer goods which experienced buoyant demand as the economy rebounded.
Another factor was the egalitarian wage policy. This stance was reversed, however, after 1928 during the industrialization drive.
The real wage gains of the NEP also proved to be short lived. During the first Five Year Plans, there was rapid inflation, as we
have seen. Between 1928 and 1937, consumer prices rose much more rapidly than urban wages, and real wages sagged, as most
historians have observed (Chapman, 1954, 1963; Zaleski, 1955; Bergson, 1961; Hunter and Szyrmer, 1992). The effect was to push
Russian real wages back to where they had been around 1880–at the start of the Imperial boom. Using our basic needs poverty line,
the welfare ratio of the average employee in large scale industry plunged to 0.97 in 1937 implying that earnings were 3% less than
required to purchase the very minimal standard we have set for a family’s subsistence. Carpenters earned 8% more than subsistence,
while labourers only realized 57% of that cost. Either other family members had to work to make ends meet or labourers could not
have been supporting a family at all. The subsistence budget is so abstemious that there was little scope for reducing spending.
Confirmation of the real wage calculations comes from the budget surveys of Prokopovich and Chapman. Their food shares were
55% and 60% in 1907/8 and 1937, respectively. According to Chapman (2963, p. 70), the food share in 1928 was only 50%. A lower
food share is consistent with a higher standard of living. Furthermore, the average worker spent 10% of his or her income on meat
and poultry in 1928 but only 6% in 1937. The share of spending on manufactured goods likewise dropped from 34% to 27% over
the same period. These shifts in spending shares are consistent with a rise in real income between 1907/8 and 1928 and a drop in
real income in the next decade.
The evolution of real wages after 1913 had important implications for the overall distribution of income. Fig. 7 shows that real
output per worker in large scale industry increased marginally between 1913 and 1928 and then dropped in 1937 to a level little
above its 1913 value. The rise in real wages between 1913 and 1928 translated into an increase in labour’s share of industrial value
added from 25% to 59%. Confirmation for such a massive change comes from Novokmet et al. (2017, Fig. 8c)’s finding (using very
different data sources) that the share of income going to the top 10% of Russians fell from about 47% to approximately 23% between
their benchmark dates of 1905 and 1928. By 1937, labour’s share had dropped back to 29%, as real wages fell. The upshot is that
the share of value added going to workers dropped from 59% to 29%. In 1913 most of the profits in the industrial sector had gone

22
US Census Bureau, Statistical Abstract of the United States: 1924, p. 309.

33
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

0
1860 1875 1890 1905 1920 1935

Cairo Bombay Boston

Manchester Russia

Fig. 8. Real wages of building labourers.

Other cities–see Allen (2014).
Source: Russia–see text and data appendix.

to private capitalists. In the Soviet period, they went to the state. Between 1913 and 1928 industrial profits were redistribution from
capitalists to workers; between 1928 and 1937 they were again transferred–this time from workers to the state. The increased income
of the state was used both to finance industrial investment and to expand and re-equip the military.
The shift from positive to negative real wage growth after the start of the Five Year Plans in 1928 was the result of changes in
economic policy. The best economic policy for the USSR was fiercely disputed by leading Bolsheviks in the famous Industrialization
Debate of the 1920s (Ehrlich 1960). Starting in 1928, Stalin implemented policies like those advocated by the left wing in the
Industrialization Debate. Had the right wing prevailed, it is likely that the human costs of rapid industrialization would have been
less. Those costs included the famine following collectivization as well as the fall in real wages. It is arguable that rapid growth could
still have been accomplished within the institutional frame work of the New Economic Policy and the real wage gains, in particular,
could have been maintained (Allen, 2003). The Stalinist turn thwarted the potential of the 1917 Revolution. That is a longer answer to
the question of whether the Bolshevik Revolution served the interests of the working class better than the pattern of growth achieved
under the Tsars.

6. International comparison

When we view Russian wage history from an international perspective, we get different insights. Fig. 8 tracks the subsistence ratios
of building labourers in Moscow, Manchester, Boston, Cairo, and Bombay. The values for the non-Russian cities are from Allen (2014).
The indices, which are ratios of full-time annual earnings to the cost of maintaining a family at subsistence, are designed to make
international comparisons of well-being possible.
The most important insight conveyed by Fig. 8 is that the rises and falls in Russian real wages become almost indiscernible in the
context of the large differences between rich and poor countries. The trend lines in Fig. 8 divide into two groups. At the top are the
USA and UK. There was very little difference between real wages in the two countries before 1900 when both were growing rapidly.
The real earnings of American building labourers continued to grow rapidly through the First World War and even the 1930s. British
real wages stagnated after 1900 in the British climacteric and remained depressed through the 1920s as the British pound returned
to the gold standard at the pre-war parity. At the bottom were India, Egypt, and Russia. Real wage growth looks almost nonexistent
compared to the USA throughout and the UK pre-1900. The jump in Russian real wages in the 1920s is dimly perceptible but is
dwarfed both by the difference in levels between the poor and the rich countries and by the growth in the USA.
The patterns are similar for skilled building workers and cotton mill operatives.23 In both cases, the wage trajectories divide into
the two groups of rich and poor countries. With respect to building craftsmen, the main difference is that the slowdown in wage
growth in the UK started decades earlier than it did with building labourers. The American industrialization boom saw a dramatic
rise in skilled wages relative to unskilled wages. The growth in the real wage of skilled craftsmen in Russia is difficult to observe and
looks like a catching up to Indian levels. The rise in real wages in the 1920s in Russia is matched by a similar rise in India in the
1930s. Thus Russian experience does not look very different from that of any other poor country.
The real wage curves of cotton operatives also divide into rich and poor groups. Among the rich countries, there was little difference
between the real earnings in the USA and UK. American workers generally earned a premium, but it was small, and real wages grew
from the 1860s to the 1930s in both countries. Real wages were lower in both Russia and India. There was little difference in real
earnings between the two countries before the First World War. The rise in real earnings following the Russian revolution is apparent
in this graph. Its significance, however, is called into question both by the fall in the 1930s and by the pronounced rise in real earnings
that took place in India in that decade.

23
Graphs of these series are shown in Allen and Khaustova (2017).

34
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

7. Conclusion

The history of real wages in Russia went through three phases between the 1850s and the 1930s. In the late nineteenth century,
particularly between the 1880s and the First World War, real wages increased in St Petersburg but stagnated in Moscow and Kursk.
The share of wages in industrial value added declined. This was a period in which all of the ‘gains from growth’ accrued to capitalists
rather than workers. This distributional story provided a fertile field for the spread of socialist ideas.
The second phase in Russian wage history was defined by the War, the first years of Communist rule, and the New Economic
Policy. We have little information about the war and the first years of Soviet rule. After the end of the hyperinflation in 1924, the
situation was calm enough for real wage calculations to make sense. This period reveals a new pattern of higher real wages. In part
this was due to particular price policies like the reduction of rent charges in apartments and agrarian and possibly manufacturing
policies that resulted in low prices for flour and bread. In addition, the industrial work force was highly organized and engaged with
the state. Wage equalization was a motif of state policy. Our data indicate this complicated set of political interactions was resolved
by a substantial rise in the average industrial wage. Gains were much bigger for workers in poorly paid industries in 1913 than in the
more highly paid. The result was a decline in wage inequality across industries. In addition, the share of wages in industrial value
added rose, as workers received more of the economic pie.
The third phase in Russian wage history was delimited by the First Two Five Years. Most years between 1928 and 1937 were too
chaotic for accurate wage and price measurement, but in the two end years were at or close enough to equilibrium levels for plausible
real wage measurements to be made. They indicate a dramatic fall in the purchasing power of industrial wages and a return to the
standards of living prevailing in the 1880s. The share of wages in industrial value added also dropped back to earlier levels, as the
share of profits, which accrued to the state, increased. This forced savings contributed to the rapid industrialization of the period.
The import of these conclusions is tempered by the international comparisons, for changes which loom large in a purely national
context look less significant in an international one. This is not to say that the doubling or halving of the real wage is not unimportant
but rather to emphasize that the differences between rich and poor countries were so large even a century ago, that the scale of change
in the real wage that was possible by redistribution was not enough to catch up to the rich countries. Only sustained economic growth
could do that.

Data appendix

I. Russia: Saint Petersburg

Building wages and prices except rent, 1853–1917

Rossiyskaya Natsionalnaya Biblioteka (Russian National Library) Saint Petersburg. Vedomosti spravochnikh tsen v Saint Peters-
burge na pripasi, materiali, platy rabochim i prochee, izdavaemie Saint Petersburg gorodskoi ypravoi 1853–1917. 3/314. The volume
for 1899 was missing, so we interpolated the index for that year.
Our main series of cotton cloth prices is for a variety called polotno flamskoe. It was about twice as expensive as the cheap cotton
cloth that formed the bulk of Russian mill production (Odell, 1912, p. 28), so we divided its price in half for our calculations. This
adjustment is made for all three Russian cities.
Rent per square metre.
1853–1913 assumed to equal rent in Moscow.

II. Russia: Moscow

Building wages and prices except rent, 1824–1870.

Tsentralnii Gosydarstvennii Arkhiv Goroda Moskvi. Raporti torgovikh starost o tsenax; fond 14, opis 4, No. 141, 153, 154, 166,
179, 193, 194, 206, 208, 223, 235, 237, 249, 262, 263, 292, 704.
Spravochnie tseni vedomstva Moskovskoi dvortsovoi konori a tak ze ministerstva Imperatorskogo Dvora; f 14, opis 4, No. 275.
O dostavlenii spravochnikh tsen i materialov; f 14, opis 4 No. 567, 664, 666.
Delo Moskovskoi rasporiaditelnoi dymi o dostavlenii spravochnikh tsen; f 14, opis 4, No. 630.
Delo Moskovskoi gorodskoi dymi; f 14, opis 4, No. 420.
Building wages and prices except rent, 1871–1917
Rossiyskaya Natsionalnaya Biblioteka (Russian National Library) Saint Petersburg. Vedomosti o spravochnikh tsenax na pripasi I
materiali v Moskve 1871–1917. 3/315.
The volume for 1883 was missing, so we interpolated the index for that year.
Ezemesyachnii statisticheskii bylleten po gorody Moskve (1892–1917) 3/1058.
Prices except rent 1924–1929.
Rossiyskaya Natsionalnaya Biblioteka (Russian National Library) Saint Petersburg. Ezemesyachnii statisticheskii bylleten po
gorody Moskve I Moskovskoi Gybernii 1924–1929. Sostavlen statisticheskim otdelom goroda Moskvi П23/959.
Prices 1937.
Chapman (1963, pp. 190–5).
Rent.
1913: Zaleski, (1955, p. 380).

35
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

Pre-1913: The rent for 1913 was extrapolated backward with the Dalmatov index of the cost of rental accommodation in
Strumilin (1954, pp. 431–2).
1914–17, 1924–7: assumed equal to 1913 and 1928, which were virtually equal.
1928–40: Zaleski (1955, p. 380), Chapman (1963, p. 195).
Wage rates.
Building craftsmen and labourers.
1926–1928 carpenters daily wage in USSR Statisticheskoe obozrenie (Russian National Library Saint Petersburg, Code is П23-1520,
Number 4–6).
1928: carpenters and labourers in Moscow, daily wage in“Comparison of Real Wages in Various Cities,” International Labour Review,
1928, p. 658. The Moscow and USSR wage for carpenters in 1928 are close.
1926–8: labourers daily wage. Moscow wage was adjusted in 1926 and 1927 in proportion to movement of carpenter’s wage.
1937–The 1928 figures were extended forward in proportion to average earnings in construction Zaleski (1971, pp. 346–7, 1980,
pp. 564–5, series 215).
Cotton mill operatives.
These wages are national averages and not city specific.
1885–1913: Strumilin (1926, 1966, pp. 92, 94) reports average factory earnings for the pre-World War I period. In the period
1901–1910, cotton industry employees earned 4% more than the industrial average, and the latter was increased by 4% in the other
years to reflect that.
1924–1928 Ezemesyachnii statisticheskii bylleten po gorody Moskve i Moskovskoi Gybernii 1924–1929 in Russian National Library
Code: П23/959.
1929–1931 Tryd v SSSR. Statisticheskii spravochnik, 1932.
1932–1935 Tryd v SSSR. Statisticheskii spravochnik, 1936.
1937: Zaleski (1980, pp. 362–3) series 205.

III. Russia: Kursk

Building wages and prices except rent, 1850–1917.

Gosudarstvennii Arkhiv Kyrskoi Oblasti. Raporti i svedeniya yezdnikh ispravnikov; f 1, opis 1, No. (2073.30), (2134.25), (1976.31),
(1921.101), (2073.30), (2134.25), (2310.27), (2984.24); f 4, opis 1, No. 117, 108, (2187.703), (1921.100), (1976.31).
Vedomosti spravochnikh tsen na proviant i fyraz; f 39, opis 1, No. 102, 103, 104, 369; f.125, opis 1, No. 128, 198, 200, 202.
Vedomosti spravochnikh tsen po gubernii; f 1, opis 1, No. 431, (2644.4), 2697, (2872.13), (3015.11), (3077.5), (3241.9).
Mesyachnie vedomosti spravochnikh tsen na proviant I fyraz; f 1, opis 1, No. 68, (1841.33), (2113.15), (2147.14), (2205.31),
(2296.1), (2310.27), (2331.4), (167.144); f.33, opis 2, chast 2, No. 5513; f.33, opis 2, chast 4, No. 11,696, 11,696; f 33, opis 2, chast
5, No. 16,121; f.56, opis 1, No. 263, 358.
Svedeniya ob yrozae khlebov I tsenax na rabochyy sily; f.1, opis 1, No. 2389, (2452.85).
Vedomosti o torgovikh tsenax na proviant I fyraz; f. 1, opis 1, No. 1700.122.
Mesyachnie listi tsen na osnovnie prodykti pitaniya; f.33, opis 2, chast 2, No. 3433.
Vedomosti o tsenax; f. 33, opis2, chast 2, No. 4250; f. 56, opis 1, No. 331.
Vedomosti o spravochnikh tsenax; f. 33, opis 2, chast 3, No. 7670, 7837, 8400, 9315, 10231.
Materiali, prislanniye dlya opyblikovaniya v gazete “Kyrskiye Gybernskiye Vedomosti”; f.33, opis 2, chast 3, No. 6514, 7248.
Perepiska s gybernskim prisytstviem o spravochnikh tsenax; f. 33, opis 2, chast 5, No. 14081.
Perepiska s glavnim tyremnim ypravleniem o spravochnikh tsenakh na proviant; f. 33, opis 3, No. 759.
Perepiska s prisytstvennimi mestami; f.33, opis 3, No. 1922, 841.
Protokoli stroitelnogo otdeleniya o proverke vedomostei o spravochnikh tsenax; f. 33, opis 2, chast 4, No. 10268.
Perepiska s Kyrskoi kontrolnoi palatoi; f.33, opis 2, chast 4, No. 11737.
Mesyachnie I polymesyachnie vedomosti tsen na prodovolstvie I fyraz; f.4, opis 1, No. 103, 95; f. 56, opis 1, No. 390.
Tablitsi o raspredelenii zemli i o srednikh tsenakh na prodovolstvie; f.4, opis 1, No. (1963.255).
Statisticheskie svedeniya o chisle gorodov Kyrskoi Gybernii, o tsenakh na proviant i fyraz; f.4, opis 1, No. (2072.586), (139.134).
Vedomosti tsen na selsko-khozyaistvennyu prodyktsiy, milo, tkan, sakhar; f.4, opis 1, No. (2133.645); f.143, opis 1, No. 41.
Vedomosti Gybernskoi Zemskoi Ypravi o tsenax na proviant; f.4, opis 1, No. (2165.672).
Vedomosti ystanovlennikh tsen; f. 56, opis 1, No. 395, 398, 406.
Rent–assumed to equal half of the rent in Moscow.

References

Allen, R.C., 2001. The great divergence in European wages and prices from the middle ages to the first world war. Explor. Econ. Hist. 38 (4), 411–447.
Allen, R.C., 2003. Farm to Factory: A Reassessment of the Soviet Industrial Revolution. Princeton University Press, Princeton.
Allen, R.C., 2014. American exceptionalism as a problem in global history,” EHA presidential address. J. Econ. Hist. 74 (2), 309–350.
Allen, R.C., Khaustova, E., 2017. Russian Real Wages Before and After 1917: in Global Perspective. University Abu Dhabi, Department of Social Science, New York
Working Papers 20170003revised May 2017.
Allen, R.C., Khaustova, E., 2018. Russian Real Wages Before and After 1917. Inter-university Consortium for Political and Social Research, Ann Arbor, MI 2018-12-05
doi:10.3886/E107509V1.
Bergson, A., 1944. The Structure of Soviet Wages: A Study in Socialist Economics. Harvard University Press, Cambridge.

36
R.C. Allen and E. Khaustova Explorations in Economic History 72 (2019) 23–37

Bergson, A., 1961. The Real National Income of Soviet Russia Since 1928. Harvard University Press, Cambridge.
Blackorby, C., Donaldson, D., 1987. Welfare ratios and distributionally sensitive cost-benefit analysis. J. Public Econ. 34, 265–290.
Bombay, Labour Office, 1928. Report on an Enquiry into Working Class Family Budgets in Ahmedabad. Government Central Press, Bombay.
Borodkin, L.I., 2001. Динамкиа заплаты строительных рабочих, а также динамка цен на муку и хлеб б. С. Петербурге (1853-1910 гг.)
http://www.hist.msu.ru/Labour/Database/bor_base.htm.
Borodkin, L., Granville, B., Leonard, C., 2008. The rural/urban wage gap in the industrialisation of Russia, 1884–1910. European Review of Economic History 12,
67–95.
Brower, D.R., 1990. The Russian City Between Tradition and Modernity, 1850-1900. University of California Press, Berkeley, CA.
Carr, E.H., 1952. The Bolshevik Revolution, II. Macmillan and Co Ltd, London.
Carr, E.H., 1958. Socialism in One Country. Macmillan, London.
Carr, E.H., Davies, R.W., 1969. Foundations of a Planned Economy, One-II. Macmillan and Co Ltd, London.
Chapman, J.G., 1954. Real wages in the Soviet Union, 1928-52. Rev. Econ. Stat. 36, 134–156.
Chapman, J.G., 1963. Real wages in Soviet Russia since 1928. In: Comparison of Real Wages in Various Cities,” International Labour Review, 1928. Harvard University
Press, Cambridge, pp. 654–660.
Crisp, O., 1978. Labour and Industrialization in Russia. In: Mathias, P., Postan, M.M. (Eds.). In: The Cambridge Economic History of Europe, VII. Cambridge University
Press, pp. 308–415.
Denzel, M.A., 2010. Handbook of World Exchange Rates, 1590–1914. Ashgate.
Dewar, M., 1956. Labour Policy in the USSR, 1917-1928. Royal Institute of International Affairs, London.
Erhlich, A., 1960. The Soviet Industrialization Debate, 1924-1928. Harvard University Press, Cambridge, MA.
Falkus, M.E., 1968. Russian National Income, 1913. In: Economica, 35, pp. 52–73. new series.
Gregory, P.R., 1980. Russian living standards during the industrialization era, 1885-1913. Rev. Income Wealth 26, 87–103.
Gregory, P.R., 1982. Russian National Income, 1885-1913. Cambridge University Press, Cambridge.
Gregory, P.R., 1997. Searching for consistency in historical data: alternative estimates of Russia’s industrial production, 1887-1913. J. Econ. Hist. 57, 196–202.
Hunter, H., Szyrmer, J.M, 1992. Faulty Foundations: Soviet Economic Policies, 1928-1940. Princeton University Press, Princeton.
Karcz, J.F., 1957. Soviet Agricultural Marketings and Pirces, 1928-1954. The RAND Corporation Research Memorandum RM-1930, Appendix E.
Kingsbury, S.M., Fairchild, M., 1931. Employment and Unemployment in Pre-War and Soviet Russia, Report Submitted to the World Social and Economic Congress,
Amsterdam, August 23-29, 1931. Vice-President, International Industrial Relations Association, New York.
Kingston-Mann, E., 2005. Statistics, social science, and social justice: the Zemstvo statisticians of pre-revolutionary Russia. In: McCaffray, S.P., Melancon, M. (Eds.),
Russia in the European Context, 1789–1914: A Member of the Family. Mcmillan, Palgrave, pp. 113–139.
Lincoln, W.B., 1986. In the Vanguard of Reform: Russia’s Enlightened Bureaucrats, 1825-1861. Northern Illinois University Press.
Lositskii, A.E., 1899. “Torgovoe dvizenie k Saint Petersburgy i iz nego,” Statistical Survey of Saint Peterburg 1896-1897g. SPB.
Markevich, A., Harrison, M., 2013. Pervaia mirovaia voina, Grazhdanskaia voina i vosstanovlenie : natsional’nyi dokhod Rossii v 1913-1928 gg. Moscow, Misl’.
Metzer, J., 1974. Railroad development and market integration: the case of tsarist Russia. J. Econ. Hist. 34 (3), 529–550.
Mironov, B.N., 1985. Khlevnie Tseni v Possii za Dba Stoletia (xviii-xix bb.). Leningrad, Nayka.
Mironov, B.N., 2010. Wages and prices in imperial Russia, 1703-1913. Russ. Rev. 69, 47–72.
Novokmet, F., Piketty, T., Zucman, G., 2017. From Soviets to Oligarchs: Inequality and Property in Russia 1905-2016. WID.world Working Paper Series, No 2017/09.
Nutter, G.W., 1962. Growth of Industrial Production in the Soviet Union. Princeton University Press, Princeton.
Odell, R.M., 1912. Cotton Goods in Russia, US Department of Commerce and Labor, Special Agents Series, No. 51. Government Printing Office, Washington.
Piketty, T., 2014. Capital in the Twenty-First Century. Harvard University Press, Cambirdge, MA.
Prokopovich, S.N., 1909. Biudzhety petersburgskikh rabochik. Shreder, St. Petersburg.
Rykazhev, A., 1911. Tseni na khleb i tryd v S.-Peterbyrge za 58 let 1911.
Smith-Peter, S., 2007. Defining the Russian people: Konstantin Arsen’ev and Russian statistics before 1861. Hist. Sci. 45 (Part 1), 47–64.
Shirras, G.F., 1923. Report On an Enquiry into Working Class Budgets in Bombay. Government Central Press, Bombay.
Strumilin, S.G., 1926. Dinamika oplaty promyshlenogo truda v Rossii za 1900-1924. Planovoe khoziastvo (9) 1926.
Strumilin, S.G., 1966. Ocherki Ekonomicheskoi Istorii Rossii i SSSR. Nauka, Moscow.
Sun, L., 2011. In Search of an Affordable Housing System for Shanghai, China, Thesis for Master of Science in Real Estate Development, MIT.
Svininyh, E.A, 2013. Institute of state purchases of goods, works and services for country’s defence and state security in the pre-revolutionary Russian law. Mil. Law
2, 7–8.
UN-Habitat, 2003a. Rental Housing: An Essential Option for the Urban Poor in Developing Countries. United Nations Human Settlements Program, Nairori 2003.
UN-Habitat, 2003b. The Challenge of Slums: Global Report on Human Settlements, 2003, UN Human Settlements Program. Earthscan Publications, Ltd., London.
United States Department of Agriculture, 2010. Food Security Assessment, 2010-2020. Government Printing Office, Washington.
Zaleski, E., 1955. Les fluctuations des prix de détail en Union Soviétique. Études et Conjunctures 329–384.
Zaleski, E., 1971. Planning For Economic Growth in the Soviet Union, 1918-1932. University of North Carolina Press, Chapel Hill.
Zaleski, E., 1980. Stalinist Planning for Economic Growth, 1933-1953. University of North Caroline Press, Chapel Hill.

37