Pakistan: Border Changes and National Income Statistics

Takashi Kurosaki


Newsletter of the Asian Historical Statistics Project. No.6, July 1997: 8-10.


The year 1997 marks the 50th year of independence for Pakistan. The basic economic statistics for analyzing the country's economic performance during this period are national income statistics. To analyze the performance in the independent era or earlier periods, it is critically important to be aware that the territory of present-day Pakistan was decided by two border changes, in 1947 and again in 1971.(footnote 1) The purpose of this article is to discuss how to adjust for the territorial changes in Pakistan, but the discussion may also provide useful insights for analysis of other areas, including Malaysia, Russia, and Central Asia, for example, whose borders have also changed significantly.

1. "Partition" in 1947

Pakistan became independent of Great Britain in August 14, 1947. Since the independence was achieved by partitioning the Indian Empire, it is usually called the "Partition" in the Indian Subcontinent. The predominantly Hindu regions of the Subcontinent came to form India while the predominantly Muslim eastern and western regions became Pakistan. The original state of Pakistan thus took on a very strange configuration, being composed of an eastern wing (present-day Bangladesh) and a western wing (today's Pakistan) separated by the vast expanse of India.

The Indian Empire under the British rule was subdivided into provinces of British India and a large number of Princely States, or native states. In the directly ruled provinces, integrated economic statistics on population, prices, and production in agriculture and manufacturing were gathered from early periods. In contrast, the nature and quality of statistics gathering varied greatly in the Princely States. Some compiled figures of the same quality or of even greater details than those of the British provinces, but on the whole, data for the Princely States are poor in quality and coverage. For this reason, the few existing studies on the pre-Partition national income (e.g., [1], [5], [10]) have estimated those statistics for British India only, or have guesstimated very rough figures for the whole empire by extrapolating the estimates for British India to the Princely States.

Upon independence in 1947, West Pakistan received the British provinces of Sindh, the North-West Frontier Province, and a western half of Punjab, as well as some Princely States including a large one of Bahawalpur. This change in boundaries has caused major problems for economic statistical analysis of the regions of present-day Pakistan for periods prior to 1947. First, there are few statistics available for the Princely States which became part of the country. Second, Punjab, the most economically important region in Pakistan, was divided between the two countries. The Partition line even divided two districts (the basic administrative unit within a province), making it impossible to compile from existing data statistics which are consistent between pre- and post-Partition periods. A third problem, which is more fundamental than technical, is that the economies of most regions of present-day Pakistan under the British rule were closely tied to regions of present-day India and did not themselves form an independent, coherent regional economy. This means that attempting to estimate national income prior to 1947 for these regions would be a task of little economic significance. On the other hand, however, compiling pre-1947 statistics of regional economic indicators such as agricultural and manufacturing production or prices for the regions of contemporary Pakistan would have great importance for research on long-term economic development. We hope to assemble such data as part of this project.


2. Independence of Bangladesh in 1971

The phenomenal economic growth of Pakistan during the 1960s widened the economic disparity between East and West Pakistan. For example, most of the rapidly growing manufacturing industries were based in the western wing. In Pakistan's first general election held in December 1970, concerns over the disparity issue led to a landslide victory of the Awami League in East Pakistan, who demanded greater autonomy and economic equality. In March 1971, when it became clear that the government would not convene the newly elected parliament, civil war broke out in East Pakistan led by the Awami League and the independence of Bangladesh was proclaimed. The third Indo-Pakistani war erupted in October 1971 when India intervened in the civil war, giving a decisive advantage to the independence movement. In December, the Indian army captured Dhaka. Independence became reality, and the government of Bangladesh was established in January 1972.

The statistics system of the reconstituted Pakistan, now comprising only the western wing, was soon reorganized. National income figures for the regions of present-day Pakistan were estimated and published, going back to the period of the united Pakistan (see Section 3 below). In this case, unlike the Partition in 1947, two coherent regional economies were officially divided into two nation states, giving sufficient economic significance to constructing retrospective statistics of national income. Because separate statistics had been compiled for East and West Pakistan, which included production statistics in agriculture, mining, and manufacturing, or population and labor statistics, it was not particularly difficult to construct national income statistics for West Pakistan.(footnote 2) However, they were not compiled before 1971, because rulers in West Pakistan feared that publishing information on economic discrepancies would encourage the East's autonomy movement.

Bangladesh began to compile official statistics of national income soon after independence. Data beginning in fiscal 1972-73 are available now [2]. As far as the author knows, however, no one has attempted to delve into the East Pakistan era to compile long-term national income statistics. Almost all the studies on economic disparity between East and West Pakistan are based on production and other related statistics.

A digression on the definition of a fiscal year might be useful here. A fiscal year in Pakistan and Bangladesh runs from July through the following June. For example, "fiscal 1972-73" in this paper indicates the period running from July 1972 through June 1973. In government documents, fiscal 1972-73 is often abbreviated to "fiscal 1972". Confusion is caused because some institutes, which are heavily influenced by practices in the US, might abbreviate the same fiscal year to "fiscal 1973". By using the latter abbreviation, fiscal 1973 in Pakistan or Bangladesh shares nine months with the US fiscal 1973 (October 1972 to September 1973); by using the former abbreviation, fiscal 1973 in Pakistan shares only three months with the US fiscal 1973. It is therefore essential that researchers know which fiscal year system is used when they use statistics of Pakistan and Bangladesh.

Returning to the subject of Bangladesh's independence, the civil war in East Pakistan broke out in fiscal 1970-71, but the subsequent chain of disruptive events -- the intensification of the war, the outbreak of the Indo-Pakistani War, the fall of Dhaka, and the establishment of the Bangladesh government -- occurred during fiscal 1971-72. This is why Bangladesh's official national income statistics commence in fiscal 1972-73 [2], and national income statistics for the original, united Pakistan were compiled only until fiscal 1969-70 or 1970-71. Naturally, the statistics covering East Pakistan/Bangladesh are in disorder during this period, especially 1971-72. For national income statistics for the regions in present-day Pakistan, there are official estimates that link the periods before and after the independence of Bangladesh (see the next section). But one should remember that fiscal 1970-71 and 1971-72 are unusual period, necessitating a special care in conducting econometric or statistical analysis using these data.


3. National Income Statistics for the Regions of Contemporary Pakistan

Given the above conditions, what kind of national income statistics for the regions in present-day Pakistan should we seek to gather from the existing data? A summary of Pakistan's national income statistics is tabulated in the government's Economic Survey [6]. Detailed information on estimation procedures in each sector and data of intermediary estimates are published every year by the Federal Bureau of Statistics (FBS) [7]. Four time series are available from these and other sources.

(1) Current Series

The current series of Pakistan's national income statistics are based on the "new methodology" (1968 SNA base) adopted in fiscal 1988-89. Compared to the previous "old methodology" (1968 SNA base), the new methodology has widened the coverage of direct estimation as well as improved estimation procedures, whose details are given in [7].

It is important to note that the Pakistani national income statistics, whether based on the old or new methodology, are consistently estimated in production accounts only. GDP at factor cost is the sum of value added from eleven sectors -- agriculture, fishing, and forestry; mining and quarrying; manufacturing; construction; electricity and gas; transportation, storage, and communications; wholesale and retail trade; banking and insurance; ownership of dwellings; public administration and defense; and other services. Value added in each sector is estimated primarily through the product approach supplemented by the income approach. The estimates of expenditure on GNP/GDP are published, but the researcher must be aware that the private consumption expenditure is derived as the residual. Only the government consumption expenditure, gross fixed capital formation, and change in stocks are estimated separately from production accounts. Income accounts have not yet been estimated.

In the new methodology also, not all sectors adopt direct estimation procedures. Annual data on actual prices and quantities are gathered through scientific surveys for major crops (wheat, rice, cotton, and sugarcane), large-scale manufacturing, banking and insurance, etc. These sectors account only approximately 30% of GDP. Data for the remaining sectors are estimated indirectly through benchmark and other methods. In order to enhance the accuracy of the national income statistics, the FBS has compiled input-output tables since fiscal 1984-85 [8].

There are four time series in the current series -- production accounts in 1980-81 constant prices, expenditure accounts in 1980-81 constant prices, production accounts in current prices, and expenditure accounts in current prices. All the four series run from fiscal 1980-81 to the present.


(2) Old Methodology Series Covering Present-day Pakistan

This is a series of national income for the regions in present-day Pakistan, estimated through the old methodology. There are four time series, whose summary is conveniently tabulated in the 1988-89 edition of Economic Survey [6]. Production accounts in 1959-60 constant prices are available for the period from 1949-50 to 1987-88, expenditure accounts in 1959-60 constant prices from 1959-60 to 1987-88, and production and expenditure accounts in current prices from 1959-60 to 1987-88. Since these series ceased to be compiled by fiscal 1987-88, necessary revisions to the provisional estimates were not properly implemented during the last few years. Researchers must be warned that the data in this series in the last few years are deviated significantly from those in series (1).

As stated in Section 2, this series includes national income data for regions in present-day Pakistan before the independence of Bangladesh. It is therefore an important data source for analyzing the country's long-term economic development.


(3) Old Methodology Series for the United Pakistan

National income statistics for the united Pakistan (East and West Pakistan combined) were compiled based on the old methodology until fiscal 1970-71. Reflecting the history of the development of national accounts statistics in Pakistan, these series experienced numerous revisions in estimation procedures [3] [7]. In addition, these series came to end during the turmoil of Bangladesh's independence war. These factors have resulted in so many variations in estimates according to the data source. For the time being, the most reliable data are those compiled by the FBS in 1972 [9].

The existing four time series cover different periods -- production accounts in 1959-60 constant prices (1949-50 to 1970-71), production accounts in current prices (1959-60 to 1970-71), and expenditure accounts both in 1959-60 constant prices and in current prices (1959-60 to 1969-70).

By subtracting figures in series (2) from those in series (3) for the period through 1970-71, we can estimate national income statistics for East Pakistan, which are not covered by Bangladesh Bureau of Statistics publications [2]. This procedure should prove useful for studying Bangladesh's long-term economic development.


(4) Extended New Methodology Series

Although the full data have not yet been published, preparatory work is underway on projects intended to extend the new methodology to cover the entire Pakistan period from1947. For example, the collaborative study by the FBS and the Pakistan Institute of Development Economics [4] have produced five national income indices for three sectors. The five indices are the value of output, value added, labor stock, capital stock, and intermediate inputs used in production, all of which are compiled in 1980-81 constant prices; the three sectors are agriculture, manufacturing, and the others. These series are notable for estimating capital stock by accumulating flow time series. As Prof. Ishiwata emphasized in his article in No.5 of this Newsletter (April 1997), estimating capital stock is very important to conducting long-term analysis of economic development. We are presently examining the utility of the stock data in these series as a benchmark for Pakistan.


Footnotes

1. This article does not discuss the issue of coverage within the national border. For example, the extent to which statistics capture the economic activities of the Federally Administered Tribal Areas, bordered by Afghanistan, varies by era. Even today, the existing statistics in Pakistan do not accurately reflect the economies of these regions.

2. As described in Section 3, Pakistan's national income statistics are consistently estimated in production accounts only. Statistics for production accounts for almost all sectors were collected separately for East and West Pakistan. The exceptions were the value added from Pakistan International Airlines, banking and insurance, the federal government, and net factor income from abroad. For these (sub)sectors, the value added was allocated on the basis of the ratios worked out and used in 1970-71 (see [6], 1988-89 edition, Appendix, p.29).

References

[1] Arora, H. C., and K. R. R. Iyengar, "Long Term Growth of National Income in India, 1901-1956," in V. K. R. V. Rao et al. (eds.), Papers on National Income and Allied Topics, Volume I, New York: Asia Publishing House, 1960.

[2] Bangladesh Bureau of Statistics, Twenty Years of National Accounting of Bangladesh (1972-73 to 1991-92), Dhaka, 1993.

[3] Kamanev, Sergei, The Economic Growth of Pakistan, Lahore: Vanguard Books, 1985.

[4] Kemal, A. R., and S. Islam Ahmad, "Final Report of the Sub-Committee on Sources of Growth in Pakistan," unpublished paper prepared for the Eighth Five Year Plan, Islamabad: Pakistan Institute of Development Economics, 1992.

[5] Mukherji, K. "A Note on the Long Term Growth of National Income in India 1900-01 to 1952-53," in V. K. R. V. Rao et al. (eds.), Papers on National Income and Allied Topics, Volume II, New York: Asia Publishing House, 1962.

[6] Government of Pakistan, Economic Adviser's Wing, Economic Survey, various issues, Islamabad.

[7] Government of Pakistan, Federal Bureau of Statistics, National Accounts of Pakistan, various issues, Karachi.

[8] Government of Pakistan, Federal Bureau of Statistics, Input-Output Table for 1984/85, Karachi, June 1993.

[9] Government of Pakistan, Central Statistics Office, 25 Years of Pakistan in Statistics 1947-1972, Karachi, 1972.

[10] Rao, V. K. R. V., The National Income of British India 1931-1932, London: Macmillan, 1940.