Garskova I., Moscow State University and
Akhanchi P., Azerbaijan Institute of History

Discrimination in the Labour Market
in the Baku Oil Industry (Late Nineteenth to Early Twentieth Century)

   The persistence of discrimination in manufacturing in the southern United States as a consequence of historical factors has been described inter alia by Gavin Wright.1 When northern manufacturers moved to the south in the twentieth century, with few exceptions they followed local segregation practice until strong political pressures forced a change. As a consequence the labour market became segmented by race, with factories hiring only black workers coexisting in a single region with factories hiring only whites.

   We find for the city of Baku, one of the centres of oil production in late imperial Russia, a similar situation with regard to the hiring of Russians and Tatars (Moslems from Iran, the Volga region and the Caucasus). Foreign firms such as the one we select, Nobel Brothers, located in Baku and, rather than introducing a non-discriminatory hiring practice, as would have been operative in its home state, Sweden, adjusted to local conditions and tended to hire mainly Russians, Armenians and Europeans as workers (that is, Christian or Jewish, or in other words non-Moslem) for jobs requiring literacy and skill and Moslem workers for lower paying, less skilled jobs.

   The labor market was in a 'path dependency' toward certain hiring practices. 2 Whether or not this hiring practice produced efficiencies, however, is a difficult to test. Obviously, the transactions costs of breaking down hiring traditions would have been significant, and may have been sufficient to prevent non-discriminatory practices. However, as in the US South, we find that other firms appeared in Baku where only Moslem workers were hired. In those firms Moslem workers could rise to positions of responsibility and were hired for skilled jobs.

   Lastly, we find that Nobel Brothers also did more than perpetuate a previously existing situation. The firm enhanced discriminatory practice, all the while declaring that no discrimination existed in hiring. Since the flow of Russians to Baku, however, was substantial, continuous and voluntary, these Baku oil producers did not have to pay transport costs (that is, pay unusual costs) in finding skilled workers. Indeed the efficiencies gained may have been in identifying by ethnicity the skilled worker in the absence of other satisfactory means of skill determination. This led, perhaps, to a form of 'statistical discrimination'.

   I. Historical Background

   The important role of Baku, a half-eastern, half-European city in the economy not only of Azerbaijan but also of Russia, was expressed at the time by Larin: 'the well-being and earnings of millions of people in our country depends upon oil in Baku'. 3 In the thirty years between 1872 and 1903 the production of oil in Baku region increased by more than 170 times; in 1897 oil production in Baku was equal to that in the United States but by 1898 it was greater, the highest in the world. 5 A multi-ethnic migration from the Russian provinces, but especially from the Caucasus and neighbouring Iran, fuelled Baku industries and assisted the rapid development of the oil industry and the transformation of Baku into one of the world's most concentrated industrial centres.

   Firms such as Nobel, Rothschild, Mantashev, the Baku Oil Association, the Moscow-Caucasian Society and others were the great oil companies in Baku. Among firms, the Nobel Association was distinguished by a highly organised work force, a higher pay and a bureaucratic administration. The personnel files of 2000 workers of the association 4 were the main source for this paper: the files include information about every worker for the entire time of employment by the firm, covering about 10 per cent of all workers, and contain information about nationality, literacy, skill-level, age, place of birth and, above all, promotions, pay, and other benefits (including those in kind), fines, accidents and so forth.

   On the one hand, there was no official discrimination according to ethnicity; non-discriminatory practices were established as a policy. Religious holidays were paid; the administration was tolerant of absence for religious practices. Thus, in the correspondence of the administration of the Nobel Association with the Baku sector in September 1911, the main reason cited for the fall in the production of oil was the observation by Moslem workers of religious holidays during 'Uraza'. 6

   Provision of housing, board and wages by Nobel Brothers was in principle non-discriminatory with regard to ethnicity. Instructions for the hiring, maintenance and dismissal of skilled and unskilled workers, both in the oil fields and the machine shops, stated that the 'wages of Russians and Tatars are the same'. But practice was not entirely consistent with this. A wage increase of 15 September 1916, for example, gave 1.30 rubles for Tatars and 1.50 rubles for Russian workers. 7

   In practice, Nobel Brothers gave preference to Russians in the hiring process. The skilled workers were mainly Russians, Armenians and other Non-moslems, and their wage rate was higher. Moslems tended to be unskilled and semi-skilled, and they were therefore less well paid. So, despite declarations of non-discriminatory practice, there was wage differentiation between different national groups.

   Table 1 shows that most of the Nobel Association workers were Russian; Persians comprised half that number. Lezgins, Kazan Tatars, Iranian Azeris, Armenians and Baku Azerbaijanians were also hired in significant numbers. In addition there were Georgian, Osetin, German, Polish, one or two Jewish, Finnish, Tadzhik and some simply 'Moslem' workers. These groups are small in number and excluded from Table 1. The most surprising feature of this table (which details literacy, skills, marital status, age, place of work, job tenure and wages by ethno-religious groups) is the low proportion of local workers, that is natives of Baku and its province. Such workers included Moslems other than Azeris, as well as Armenians, Russians and other Christians, but even taking this into consideration, we find only 131 (7 per cent) were local. The proportion of Baku Azeris working for Nobel Brothers was much lower than their share in the province 8 and somewhat lower than their share in the oil-producing region. 9

Table 1. Nobel Association, Baku: Ethno-religious groups of workers, 1878-1921

Group	Number of workers	Literacy, %	Degree of skill, %	Marital status, %	Age	Shop, %	Lenght of service, (months)	Wages/aver.wages, %
Azeri	38	18	5	66	32	55	30	95,3
Persian	462	5	1	41	27	13	17	84,7
Iranian Azeri	96	7	6	39	28	15	48	97,2
Lezghin	234	5	4	44	27	6	34	88,6
Kazan' tatar	101	19	8	59	29	9	47	94,5
All Moslems	931	7	3	44	28	13	28	88
Russians	951	74	25	74	30	38	48	105,7
Armenian	51	51	29	69	32	22	47	114,1
Other christans	41	83	43	63	28	68	71	147,8
All christians	1043	73	26	73	29	39	49	107,8

   Russians formed the largest group of Nobel Brothers workers. The proportion of Russians (46.5 per cent) was much higher than in the general population of Baku province (23.5 per cent) ;10 and of the oil-producing region as a whole (24.7 per cent). ;11 The Russian immigrants to Baku as a rule came from the north: the central industrial region (13 per cent), the central black-earth region (26 per cent) and middle- and lower-Volga regions (49 per cent). Kazan Tatars were also originally from the Volga region (88 per cent), the Lezgins (87 per cent) came from Dagestan (Caucasus), and Persians (94 per cent) and Azeris (96 per cent) came from Iran. 'Beginning in the second half of August, the oil fields are quickly filled with workers from the Caucasus and Dagestan as well as from the Volga. . .This influx of workers creates for oil production in Baku the situation that summer daily wages of unskilled workers increases up to one ruble and even higher, but in the other seasons wages decrease to 0.80 rubles for Russians and to 0.50-0.60 rubles for workers from Persia and the Caucasus. . . As for skilled workers, there is no shortage in any season'. ;12 The other significant uninterrupted influx of workers into Baku was from Iranian. 'A poor person, having crossed the frontier, travels by foot to the object of his dream (namely wage employment), hungry, ragged, hardly earning a livelihood on the way'. 13

   II. An Analysis of Discrimination

   In general, supervisors perpetuated ethnic stereotypes of work-based segregation. For example the overwhelming majority of workers in Nobel Brothers plants were Moslems from Baku and southern Azerbaijan. A regional mechanical engineer wrote, 'they are most often appointed to jobs which demand care, for instance oil distillation and kerosine cleaning, or unskilled jobs out of doors. Armenian workers, especially from Shusha district of Elisavetpol province carry out duties of steam engine stokers and even replace machinists and metal-workers in small plants. In larger plants almost all jobs demanding knowledge of metal-working and boilers and other machines are carried out by Russian workers. Other nationalities that can be found, but as the rare exception, include Jews, Germans, and Poles'. 14

   Analysing the dynamics of employment for different ethnic groups from 1878 to 1921 one notices the surprising stability of the Russian share, while other groups change. After 1914, with some decline in the proportion of Russian workers (connected with mobilization for the battlefront of Russian but not Moslem or foreign workers) one can see a sharp rise in the Persian share; the share of Iranian Azeris, which began high, steadily declined, providing evidence of their substitution by Kazan Tatars and Lezgins. In general however, the share of all Moslem workers is quite stable. Along with the low share of workers from Baku (Table 1), we also find a steadily decreasing share of native local workers from Baku (both Azeris and Armenians), from 10 per cent to 3 per cent. We suggest that the high share of Russian (and generally speaking non-Moslem) workers and extremely low portion of Azeris gives evidence of some ethnic discrimination.

   The form discrimination took was indirect, for different wages for identical jobs performed by workers of different nationalities, direct discrimination is difficult to pinpoint. Indirect discrimination is linked an inheritance of segregation among jobs and associated levels of skill, education and so forth. The existence of a competitive labour market does not exclude, but can supports and reinforce racial discrimination in the form of professional segregation. Racial job segregation became firmly established and institutionalised, we argue, because there were some efficiencies involved. In practice one must analyse wage discrimination by comparing wages of two groups of workers with the same education, skill level, occupation, marital status, age and so forth, but belonging to different ethnic groups. 15 In our case we cannot isolate 'pure' groups, because of the existence in this period of rigid occupational segregation in Baku oil production. Moreover, the mean wage of Moslem workers was lower than than the average for all workers whereas that of non-moslem workers were higher. Although the standard deviation of wage distribution declined during the period we cannot absolutely conclude that this gap between ethnic groups was declining.

   For the next step we aggregated 8 original ethno-religious groups into two: Russians and Moslems. Table 1 shows that almost all variables (except age) of Moslems were similar and were sharply distinguished from the same variables for Christians (usually Russians, 96 per cent), and that the total number of Moslem workers approximately equalled the total number of Russians.

   All parameters except age and length of service in the firm differ by ethnic groups (Russians and Muslems). Russians had greater literacy, were more highly skilled and carried out more complicated jobs in machine shops and plants. Russians were more often married and worked in the firm for longer terms. It is not surprising that their wage rate was higher. Contemporaries understood the rationale for hiring Russians: 'it is both practical and profitable to hire workers who are literate, whether skilled or unskilled. They are more intelligent. They grasp the demands of their job more easily and execute it better. All these circumstances are considered in the labour market, where Russian workers are judged as more intelligent and get higher wages'. 16

   Table 2 provides evidence of ethnic differentials in the frequency of upward occupational mobility from unskilled to skilled jobs, or from unskilled and skilled blue-collar to white-collar professional. Note that those Russians and Moslems who moved up the professional ladder were as a rule married and literate (and undoubtedly skilled). The characteristics of the Russian worker determine the differential between the two left-hand columns and the two right-hand columns in Table 2.

Table 2. Nobel Association, Baku: Annual average mobility of workers, 1878-1921

Group	Russian			Moslem
	Skilled	White-collar		Skilled	White-collar
Skilled		15			2
Learners	14	1		2	-
Semi-skilled	43	10		15	2
Unskilled	18	1		4	1
Others	3	12		1	3
Total	78	39		22	8

   The differentials between two main groups of workers may be demonstrated in the distribution of such variables as length of service in the firm: the share of Moslems is higher for the lower lengths of service, whereas the proportion of Russians is higher for longer service.

   We have also collected dynamic data for different ethnic groups for such parameters as skill and age. In our further work we intend to examine David's contention that in periods of faster industrial growth under conditions of general labour scarcity a rise in wages is accompanied by a reduction in the gap between skilled and unskilled labour. 17 Generally in developing countries in the early stages of industrialisation (as in Russia at the late nineteenth and early twentieth centuries), the gap between skilled and unskilled wages was wide, and the high wages of skilled workers were maintained because of an influx from the old industrial centres. The share of skilled workers changed in line with the demand for labour, whereas skilled workers' wages moved in the opposite direction. In contrast with the situation David described for Chicago, the main feature of the Baku labour market was an absence of any labour shortage.

   During 1906 - 1916 the peak of Moslem wages shifted downward whereas the curve of Russian wages steepened for greater values. This differential confirms the existence of wage discrimination in the predominance of Moslem workers in lower paid jobs.

   In conclusion, it should be noted that these results agree with the theme in Wright and Whatley 18 about a direct link and feedback between low levels of education and skill on the one hand, and occupational segregation on the another. That is, because they do not expect to get a good job, some ethnic groups do not invest enough in human capital, in particular in education.

Notes

---

1 G. Wright, 'Postbellum Southhern Labor Market' in Peter Kilby (ed.), Quantity and Quaddity: Essays in United States Economic History (Middletown, CT: Wesleyan University Press, 1987), pp. 98-134.

2 P. David, 'Path Dependance: Putting the Past into the Future of Economics', Technical Report, no. 533, August, 1988.

3 J. Larin, Rabochie neftjanogo dela (iz byta i dvijenija 1903-1908) (Moscow, 1909).

4 M.A. Ismailov, Sotsilno-ekonomicheskaya struktura Azerbaijana v epokhu imperializma (Baku, 1982).

5 Central State Archive of юzerbaijan (CSAA), F. 798, Op. 3, Dd. 2, 3, 5, 11, ..., 186, 187, 189, etc.

6 CSAA, F. 1458, Op. 1, D. 239, L. 2.

7 CSAA, F. 798, Op. 2, D. 3585, L. 222.

8 D.I. Ismail-zade, Naselenije gorodov Zakavkazskogo kraya v XIX - nachale XX Б. (лoscow, 1991).

9 A.M. Stopani, Neftepromyshlenny rabochij i ego budget (Baku, 1925).

10 Ismail-zade, op. cit.

11 Stopani, op. cit.

12 CSAA, F. 45, Op. 2, D. 351, L. 13.

13 M.A. Ismailov, Kapitalizm v sel'skom khozajstve юzerbaijana na iskhode XIX i nachale XX vv. (Baku, 1964).

14 CSAA, F. 20, Op. 3, D. 2304, L. 21.

15 Paul A. Samuelson and William D. Nordhaus. Economics (New York: McGraw-Hill, 1989).

16 CSAA, F. 1458, Op. 1, D. 239, L. 2.

17 Paul A. David. 'Industrial Labor Market Adjustments in a Region of Recent Settlement: Chicago, 1848-1868', in P. Kilby (ed.), Quatity and Quiddity: Essaysin United States Economic History (Middletown, Ct: Wesleyan /university Press, 1987).

18 W. Whatley and G. Wright, 'Race, Human Capital, and Labor Markets in U.S.History', American Economic Review, vol. 82, no. 2, May 1992, also in G. Grantham and M. MacKinnon (eds.), The Evolution of Labour Markets (London: Routledge); and theit unpublished working paper, Getting Started at the Ford Motor Company, Black Workers in the Automative Industry 1918-1947 (Stanford University Working Paper Series, 1993).