The five republics of Central Asia-Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan and Uzbekistan-emerged on the map as independent states after the dissolution of the Soviet Union in 1991. These nation-states are situated east of the Caspian Sea between Russia, China, Afghanistan and Iran, covering a territory about twice the size of Western Europe with total population of more than 55 million people (by January 1, 1999).
In any attempt at a comprehensive analysis of the population of Central Asia, the historical backdrop must be taken into account. The rich history of the region can be divided into the following four periods:
1. Before Russian Rule.
2. Under the Russian Empire.
3. During the Soviet Period.
4. Since Independence.
Historical quantitative analysis of Central Asian populations is a complicated undertaking. However, it is a crucial area of research if we hope to attain a better understanding of the historical origins of the urgent problems currently facing people of various ethnic backgrounds in the region.
The main problem in conducting this research is a lack of comprehensive and accurate historical data of population.
Untortunatry, statistical methods, such as invers projection, are not sufficient to reconstruct the Cential Asian population even within modern history. This is primarily due to factor irregularities (wars, famines, revolutions, repression, etc.) which affect the dynamics of population, and the massive yet ill-documented ethnic migration within the region. Further research and data collection is needed to achieve compatibility, especially in light of the multiple changes in state and administrative boundaries, and possible distortions of Census and other statistical data. Given the improved access to previously closed Russian archives, as well as those of other countries, and the establishment of wider international academic cooperation, the time seems ripe for a discussion of the problems and issues related to quantitative demographic analysis for each specific period in Central Asian history.
Unfortunately, there are no consistent demographic and statistical data sets on the peoples of Central Asia up to the second half of 19th century. (Although some fragments of a population census from the 8th century A. D. were unearthed among the documents of the ancient Central Asian state Soghdiana-whose capital was Samarkand-.
Thus, to discover new direct demographic data on the peoples of the region (who played such a significant role in the Great Silk Road trade in ancient and medieval times) is rather difficult. But it is possible to partially cover for this lacuna through the use of additional indirect data (for example, from information contained in old Central Asian as well as Persian, Greek, Arabic, Chinese, Uygur [Orkhon], and Mongolian manuscripts). The collection of all direct and indirect evidence from these sources should help provide a better understanding of the ethnic origins of the contemporary peoples of Central Asia and the demographic processes that were at work in the region. One of the possible sources for tracing and identifying the scale of mass migrations of Turkic people from their geographic place of origin (an area which in modern terms would be identified as Altay, and the northern parts of Mongolia and Manchuria) to Central Asia are old Chinese and Orkhon Mongolian records. The levels of expertise in these scripts are insufficient in the Central Asian republics to fully utilize such sources, but such levels of expertise are available in Japan. In this respect, it is apparent that further cooperation between Central Asian and Japanese scholars on research projects related to Inner Asian studies would be beneficial in expediting the search for new demographic data on population movement from North East Asia to Central Asia.
For accurate quantitative analysis of the Central Asian population within the Russian Empire, there are some data but also obvious difficulties in generating a comprehensive time series for the whole period.
The colonization of the area took place over nearly two centuries (from the early 18th to the late 19th). Two big administrative units-the Steppe and Turkestan Government Generals-were established. The former entity included Uralsk, Turgaisk, Akmolinsk, and Semipalatinsk regions and covered about two thirds of what is now Kazakhstan. The latter was founded in 1867 and initially included Syrdar'ya and Semirech'e, eventually expanding to absorb the Samarkand, Fergana, and Zakaspiysk regions. By the year 1896, the territorial area covered by the two colonial administrations, combined with the states of Bukhara and Khiva (which had become Russian protectorates), reached the size of contemporary Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan and Uzbekistan.
There is some fragmentary statistical data available for the period. According to archives, there was a census organized independently in Khiva between 1856 and 1864 under the rule of Saidmukhammad-Khan . However, there was no equivalent Census type collection of demographic data for the Turkestan, Bukhara state and the Steppe reagion. Even for the latter, which was colonized more than a century before Khiva, population data, especially with respect to the indigenous population that led a largely nomadic lifestyle, is insufficient. The earliest population statistics collected for the Steppe region is dated 1858 .
For 1897-1913, there are more data: accurate information is available from the Census (1897) and other official materials. Although this period has the best statistical base for the pre-Soviet period, two problems still remain. The first is the fact that Russian official data continued to exclude the territories and populations of the former Bukhara Emirate and Khiva Khanate (which were later divided between Uzbekistan, Turkmenistan and Tajikistan). Second, according to extant sources, the figures did not account very well for the natural growth of the population in the region. There was no equivalent of church registrars of births and deaths for local people in the area. Even for the 1909 and 1910 population surveys, the best source for the period, "Asiatic Russia," emphasized, "for the Muslim population of Turkestan, there are no satisfactory statistics of births and deaths, and there are no registration documents (metric notes) for them." The data for the whole region's population was estimated only on the basis of the natural growth of indigenous people of the Zacaspiyskaya and Semirechenskaya oblasts . The problems from using this method were most noticeable in the most populous Syrdaria and Samarkand oblasts . The former was divided subsequently between Uzbekistan and Kazakhstan, and the latter between Uzbekistan and Tajikistan.
Accordingly, the lack of comparable annual data for the protectorates, and the absence of consistent vital statistics for indigenous populations in both Steppe and Turkestan regions makes it difficult to produce comprehensive demographic calculations. In short, there does not exist a complete and accurate statistical data base covering all of Central Asia for the Russian period. The situation is also aggravated by the incompatibility of data for the Russian and Soviet period statistics stemming from differences in administrative units and different coverage areas. However, it is possible to compensate for this incompatibility issue by using maps to compare the composition of different administrative units, and by generating aggregated indices for individual states. In official publications one can find population data for 1897 and 1913: these were the results of recalculations conducted during Soviet period which were then incorporated into official statistics of the respective republics. There are also estimates for the total populations of Uzbekistan, Kyrgyzstan, Tajikistan and Turkmenistan made by Uzbek demographers on the basis of the archives of the Turkestan Government General, in addition to reverse projections down to 1865 [4 and 7]. But estimates by some other researchers for the population dynamics of Central Asian for the period 1897-1913 differed considerably from the Uzbek team's estimates. This would seem to indicate that more detailed analysis is required for this period [4, 5, 6].
The next decade, 1914-1925, is even more difficult to study. The systems of data collection and statistical publications were partially or fully destroyed beginning in 1914. The latter-day researcher is faced with the challenge of generating more accurate population estimates for the period covering World War I, the February and October Revolutions, and the subsequent civil war. In particular, the annual data for both the indigenous populations and the increasing numbers of unregistered migrants who entered Central Asia during these volatile years is very poor. Due to the absence of vital statistics for the period in the area under consideration, even a relatively comprehensive 1926 Census could not fully fill this gap [2a].
The outer frontiers and boundaries of the Russian Empire and its protectorates in 1897, when the first census was conducted, coincide with contemporary ones. However, it is far more complicated a task to generate accurate calculations of time series data for the Central Asian states than for the Russian Federation itself.
The direct origins of all five Central Asian states can be traced back to the creation of the Soviet Union. The formation of Uzbek, Turkmen, Tajik, Kazakh, and Kyrgyz Republics from 1924 to 1936 gave different ethnic groups opportunities to consolidate themselves as modern nations within their respective titular states. Nevertheless, the reconstruction of population data by ethnic background for each of the existing Central Asian states remains a very difficult task, due not only to changes in state borders but also to changes in the self-identification terms and concepts employed by the people themselves. It also requires a very careful study of both Russian and Soviet sources, as well as the archives of the newly independent states (NIS) of Central Asia, especially for the period prior to the delineation of the outlines of the respective nation-states.
This period witnessed a mass exodus of indigenous Central Asians during the civil war which followed the revolutions in February and October 1917. The collectivization policies of the early 1930s generated unjustifiably great losses of nomadic settlers of the Central Asian steppes, deserts and mountains, while local political and intellectual elites were decimated under Stalin's repression. On the other hand, during the famines of 1921 and 1933, there were massive influxes of people from the European areas, especially from the regions around the Volga River, to Central Asia through state-organized and private ethnic migration. Further reassessment of population dynamics in this period can be pursued through the use of data from the recently declassified 1937 Census , which had been inaccessible of the public from right after its completion.
Central Asia was the site of unprecedented ethnic migration by Slavs for more than a hundred years, and also the recipient of mass deportations of Germans, Koreans, Chechens, Karachays, Meskheti Turks and Crimean Tatars immediately before, during, and after World War II. The consequences of deportations policies were revealed under Gorbachev's democratization in the second half of the 1980s. Many of these people, however, were unable to reach a suitable solution for their situations, and in some cases merely augmented the instability of the newly formed independent states.
Immediately after World War II, unlike in Japan or many European countries which suffered from the war, in the USSR, no censuses were conducted. As a consequence, it is very difficult to estimate the overall population losses and movements for a period of over ten years. Recent estimates of population losses from the war placed the total at 25.3 million Soviet lives in contrast to the previous estimate of about 20 million . This revised estimate still requires further clarification that includes population data for the period from the Central Asian republics as well.
There are more or less comprehensive data for the period starting with the end of 1950s covered by the four censuses (1959, 1970, 1979 and 1989) conducted at regular intervals with annual publications of current demographic statistics in between. However, most analysis of these data have been conducted within the framework of the dominant ideology, with the primary goal of proving the advantages of socialist, non-capitalist laws of population movement. The majority of publications under the Soviet period were devoted to maintaining unity within the Soviet Union, and thus, emphasized the discovery of evidence indicating regional similarities rather than disparities .
Meanwhile, during this period, most of the Central Asian states experienced an enormous population explosion, an explosion whose momentum remains strong to this day. From the first Russian census in 1897 to the 1959 census, the population of Central Asia increased from 10.5 million to 24.7 million people. By 1989, the year of the last Soviet census, the population for the Central Asia region had doubled, reaching a total of 49.5 million. In particular, the states of Tajikistan, Turkmenistan, Kyrgyzstan and Uzbekistan doubled their populations within a period of less than 30 years.
Since independence, the population of the Central Asian countries has increased in total by about ten per cent, despite the large outflow of Russian-speaking people from the region. Dissolution of the former Soviet Union had the largest effect on the population growth of Kazakhstan, as this country had the largest share of non-titular inhabitants in the former Soviet Union. However, the disparity between dynamics of population in the Slavic states (Russia, Ukraine, Belarus) and other Central Asian nations remains. The former countries have recorded rather high negative growth rates of population owing to sharp increases in death rates. In contrast, the latter states have continued to register high birth rates and positive population growth. The interconnections and linkages between systemic social-economic change and demographic transition is the most complex and most important issue to tackle, especially in the Central Asian states, where indigenous populations have been growing along more or less traditional patterns.
The threat of ethnic conflicts, the mass outflow of Russian speaking populations from newly independent states of Central Asia, and the inflow of forced migrants from neighboring areas of civil and economic unrest, have been among the most urgent concerns for the new states since independence, and will obviously remain major concerns for the next few years. The thorough analysis of current migration patterns, therefore, is among the most urgent tasks for any statistical analysis of the area.
The real scale and scope of the demographic changes that have occurred in Central Asia since disruption of the former Soviet Union can be traced in a new round of censuses undertaken by the newly independent states. For example, the State Statistical Agency of Kazakhstan has recently announced preliminary figures from the first post-Soviet census in the republic (February 25 - March 4, 1999). They showed, for instance, that the population had fallen by more than a million persons from 16,199,000 in 1989 to 14,951,000, a drop of 7.8 per cent. The decrease was mainly attributed to the outflow of Russians, Ukrainians, Germans, Jews and other Russian-speaking people. Now ethnic Kazakhs account for 53.4 per cent of the population compared to 36 per cent at the time of the 1979 census (when total population was 14,684,000, just marginally less than now). During the ten years since the 1989 census, the urban population of Kazakhstan fell by 8.4 per cent and rural population by 6.8 per cent. Today, 8.3 million people (55.9 per cent of the total) live in cities and the remainder in villages. The largest Central Asian state in area, about the size of western Europe, has become one of the least densely-populated states in the world. The final results of the census will summarized at the end of this year . These statistics will allow for more mature conclusions on demographic trends in Kazakhstan over the past decade, and for more accurate predictions for population trends in the forthcoming years.
Meanwhile, it is clear that further research on demographic change in Central Asian populations throughout 20th century, particularly changes within each specific period (e.g. under the Russian Empire, Soviet Union and post-independence) is vitally important to fill statistical and analytical gaps.
This brief review of the problems and issues related to research on the population of Central Asia has highlighted the limits of the existing information, including officially published sources, for any comprehensive analysis. Further research on historical population data of Central Asian peoples, therefore, requires considerable work in such forms as, collecting additional data from recently opened archives; applying more sophisticated quantitative methods to adjust for information paucity through more accurate estimates; and more analyses which start from non-ideological standpoints. Such work will hopefully produce a better understanding of the different periods in Central Asian demographic history and promote better policy formulation and decision-making in the present and the future.
(Institute of Economic Research, Hitotsubashi University)
 Asiatskaya Rossia, tom pervyi, Ludi I poriadki za Uralom. Izdanie pereselencheskogo upravleniya Glavnogo upravleniya zemleustroistva I zemledelia, S-Peterburg, 1914.
 Andreev E.M., Darskiy L.E., Khar'kova T.L. Naselenie Sovetskogo Soyuza: 1922-1991, Moskva, "Nauka", 1993.
[2a] Andreev E.M., Darskiy L.E., Khar'kova T.L. Demographicheskaya istoriya Rossii: 1927-1959, Moskva, "Informatika", 1998.
 Boyarskiy A. Ya., Shushurin, Demographicheskaya statistika, Gos.stat.izdat, Moskva, 1955.
 Brachnost', rojdaemost', smertnost' v Rossii I v SSSR. Pod redaktsiey Vishnevskogo A.G., Moskva, "Statistika",1977.
 Demographicheskaya statistika, Otv. Redaktor T. V. Ryabushkin, Moskva, "Nauka", 1986
 Jiromskaya V.B., Kiselev I. N., Poliakov Yu.A., Polveka pod grifom "Secretno": Vsesoyuznaya perepis'naseleniya 1937 goda, Moskva, "Nauka", 1996.
 Karakhanov M.K., Necapitalisticheskiy put' razvitiya I problemy narodonaseleniya, Tashkent, "Fan", 1983.
 Naselenie Rossii za 100 let (1897-1997), Stat. sbornik, Goscomstat Rossii, Moskva, 1998.
 Rashin A.G., Naselenie Rossii za 100 let (1811-1913), Pod redaktsiey akademika Strumilina C.G., Gos.stat.izdat, Moskva, 1956.
 Tatimov M.B., Razvitie narodonaseleniya I demographicheskaya politika, "Nauka" Alma-ata, 1978.
 Vodarskiy Ya. E., Naselenie Rossii za 400 let (XVI-nachalo XX vv), Moskva, "Prosvetshenie, 1973
 BBC Monitoring: Inside Central Asia, N 270, April 12-18, 1999; Interfax April 13 and May 15, 1999; Reuters April 14, 1999; Radio Free Europe/Radio Liberty, April 15, 1999; BBC April 15, 1999.