How the CIA estimated the Soviet Economy

Last year I took a class on the Soviet and modern Russian economy. Part of that class involved a term paper. I choose to do mine on how the CIA estimated Soviet economic statistics. The class was also thought by a leading economist in Soviet economics so I was pretty hype when he called it “a very nice little paper!”

Soviet GNP and American Estimations

Many economists have noted that national income estimation is closer to an art than a science and they are not being entirely hyperbolic. Between the variety of choices that have to be made in the estimation process, it is far from a clean process. Yet, national income estimation was one of the most important parts of the cold war and arguably economics most significant contribution to it. By estimating the size of the Union of Soviet Socialist Republics, USSR, the United States and her allies were better able to understand the threat posed by the socialist regime and how to counter it effectively.

Related: Debunking Progressive Cliches, Entry 5: “Socialism can be democratic”

National income is “the total market value of production in a country’s economy during a year” (Ott). While modern economics looks at Gross Domestic Product (GDP), the USSR collapsed before the switch from Gross National Product (GNP) to GDP so we will focus on the former. GNP is specifically defined as “an estimate of total value of all the final products and services turned out in a given period by the means of production owned by a country’s residents” (Chappelow). In other words, GNP represents the final value, at prevailing prices, of currently produced goods and services with no deduction for durable capital depreciation (Bergson, 15).

The main difference between the two is that GDP includes foreign owned production in the country and excludes domestically owned foreign production. The Soviet Union had almost no foreign investment in the country or international economic investments causing GDP and GNP to be essentially the same.

Towards the end of World War Two, the Office of Strategic Services (OSS) began measuring Soviet National Income for lend-lease purposes. While Paul R. Gregory described Tsarist Russia as having “…comparable [national income series] to the historic series of most of the major industrialized countries…” (Gregory, 9) but valuable data dried up following the Russian Revolution and was “largely suppressed” (Bergson, 3). Outside a wave of data in the 1920’s, the OSS had little information to go off besides information in percentages which may or may not have been reliable. This lead the US government to need to reliably know how large the USSR was prior to the war in order to provide them with an adequate amount of support in World War 2. This need continued after the war in Europe and the Pacific ended and the Cold War took its place.     

In his 1961 book The Real National Income of Soviet Russia since 1928, Abram Bergson laid out five principles in estimating national income. First, accurate measurements of national income are required to learn production potential and welfare. The proper price valuations on goods depend on which application is used. This distinction is particularly important in the case of the Soviet Union because the preferences of planners often did not align with those of Soviet consumers. For example, planners focused on military projects over consumer goods (Bergson, 26).

Related: Debunking Progressive Cliches, Entry 2: “Cuba is actually really good”

Second, in comparing two periods, outputs in both periods are valued either using period one prices or period two prices meaning that if an economist wanted to measure the change in GNP between 1950 and 1960 they would either need to use only 1950 prices or only 1960 prices (Bergson, 26).

Third, the magnitude of “equivalent variation” will depend on the nature of the composition. This means that for a certain combination of goods, the change in output needed to achieve the same effect will increase or decrease depending on its share of total output (Bergson, 28).

Fourth, when measuring national income over multiple periods choosing between a base year and a given year offers significantly different results. If you choose the base year then each period is evaluated using the same price but if the given year is used then the given price varies between periods (Bergson, 29).

Finally, the schedule of production possibilities depends on the period length. This means, for example, how resources can be reallocated is different in the short-run than it is in the long-run (Bergson 30).

There are two methods used for providing specific estimations of national income. They are:

1. End use estimations and

2. Sector of origin estimations

End use estimations are your traditional method of estimating national income by aggregating consumption, investment, government spending and net-exports. Sector of origin calculates national income by aggregating economic activity in industry, construction, agriculture, and a handful of other economic activities (Reuss, iv).

End use estimations and sector of origin estimations were both developed by the United States Department of Commerce and the OECD but the latter method was the method preferred by the CIA to measure total GNP growth. Even though the methods used were based off Department of Commerce methodologies, the “…organization of the Soviet economy and the limited amount of data…” caused the CIA to modify and simplify the process. For example, defense expenditures were not identified separately in the Soviet GNP accounts because so much of it was included in other GNP components like investment (Reuss, vi).

Related: 10 Hard Questions for Progressives

To estimate GNP growth with the sector of origin approach, each sectors real growth rate was multiplied by its weight (Boretsky, 518). Income flow estimations were used to construct base year weights (Reuss 31). Since the CIA did not have access to data for every sector, it used a sample of 312 products based on partially released Soviet data. Of the 312 products, 68 were different automobile mobiles, 9 were different excavator models, and 19 were different locomotive models. When it came to non-trade services, the CIA used a man hour index on which to base its estimates (Boretsky, 518).

Another important aspect of national income accounting to understand are the accounting units and how finances flow between them. The four accounting units were budgets institutions, Khozraschet enterprises, collective farms and private households (Reuss, 27-28).

An important nuance to remember is when discussing government spending in national income calculations we are talking about their budget. For example, in the Soviet Union the government controlled essentially the entire economy and all economic activity went through it but only its budget was counted as government spending. The government current expenditures were the economic output of the government. This relationship was the same as in the United States (Reuss, 29).

Similarly, Khozraschet enterprises are state owned enterprises that engage in “buying and selling” but their financial flows are not accounted for under the budget, even though they are state owned because they are not part of the budget (Reuss, 27).

There was also double accounting that occasionally occur. For example, an enterprise would be budgeted via the state budget, but would also sell a service. Museums and theaters were examples of this phenomenon (Reuss, 29).

The most obvious weakness in the estimation process was lack of value information from the Soviet Union. The Soviet Union provided valuable economic information in waves. The first wave was during the New Economic Policy of the 1920’s, again during Khrushchev’s reign, and once again when Gorbachev ruled (class notes). From the 1930’s until 1956, “not a single volume of statistics was published”. This applied to basic economic information like population, wages, and output. The only data that was published was in percentages.

Even once the Soviet Union began to publish data it was notoriously limited. Industrial output was overaggreated, goods were omitted while others would appear and disappear. Wages did not include actual earnings. Social statistics included some information for one nationality but had different information for another nationality. The data that was published also did not have clear definitions (Nove, 365-366). All that is to say, Soviet economic data was “notoriously imperfect and incomplete” (Bergson, 2).

Another weakness in the estimation procedure is the phenomenon known as Index Number Relativity (INR). INR is a statistical phenomenon whereby more rapid national income growth rates are obtained when early year valuation weights are used. This is explained by the inverse relationship between output growth and price growth (Gregory, 26). This phenomenon is especially abundant in developing countries. Any estimate had to choose between overestimating Soviet growth by choosing an early base year or underestimating Soviet growth by choosing a later year.

After all this, how reliable were the results generated by the US government? Abram Bergson estimated Soviet industry to be about sixty perfect of US industry while their overall economy to be just over forty percent of ours in 1960. Similarly, Gertrude Schroeder estimated Soviet industry to be sixty percent of ours in 1979 (class notes, lecture 15). But just over a decade later, Soviet economists estimated their economy to be fourteen to twenty-eight percent the size of the United States (Los Angeles Times).

A 1991 Government Accountability Office report supported the idea the Soviet economy was overestimated saying “these methods are unlikely to produce accurate results, primarily due to data limitations and problems inherent in estimating a nonmarket economy’s GNP.” This report specifically focused on the Central Intelligence Agency’s estimate of the Soviet economy but it noted “…[their] methods for estimating Soviet GNP are, where possible, consistent with western national income concepts.” The report went on the clearly state the CIA “…probably overstated the relative size of the Soviet economy…” (GAO, 3).

Similarly, by comparing UN GDP estimates and CIA GNP estimates we can see any significant divergences. While these statistics are not always identical, they are similar enough that a significant difference would not be expected. In the Tyranny of Numbers, Nicholas Eberstadt estimated that for western European countries the ratio of the CIA to Un estimate was approximately 1– meaning that they were nearly the same– but when he analyzed eastern European states he found ratios of 1.48 for Hungary, 1.23 for Yugoslavia, and 1.64 for Poland. All of which suggest significant overestimations by the CIA. While the UN did not produce a figure for the Soviet Union, it would be fair to assume given all the other evidence that the systematic overestimation for socialist countries would apply to them (Eberstadt,145).

It should be noted that the CIA has responded to these criticisms. Bruce Berkowitz noted in the International Journal of Intelligence and Counterintelligence that the CIA “… detected the slowdown in the Soviet economy; it noted that the Soviet leadership was running out of options to save the country…”. While this is true it is also not the same as correctly estimating the size of the Soviet economy. At best, the CIA only was able to accurately estimate parts of the Soviet economies condition (Berkowitz, 237).

While estimating Soviet national income had obvious a priori weaknesses and the estimates themselves were deeply inaccurate, it was not necessarily the fault of western economists. Soviet economic data was, as mentioned above, was incomplete, inconsistent, and even outright impossible to find. The country’s economy was essentially a black box economists had to study but could not access information on. Economists like Abram Bergson were able to pioneer a methodology of studying one of the world’s largest economies that was irrationally planned and produce a result that was in the ballpark of being accurate. That, in itself, is worth something.

One of the most significant economic questions of the 20th century was the conflict between American capitalism and Soviet socialism and which system was superior. The question had both practical implications for which system was superior and for the Cold War. In this question, discovering the real Soviet national income was one of the most significant queries. In answering the question, American economists battled incomplete, inconsistent, and often missing information, yet they were able to develop a result which allowed for a ballpark accurate estimation of Soviet national income.

Works Cited

 Boretsky, Michael. “The Tenability of the CIA Estimates of Soviet Economic Growth’ .” Sciencedirect, Journal of Comparative Economics, 31 May 1985,

Becker, Abraham S. Soviet National Income, 1958-1964. University of California Press, 1969.

Bergson, Abram. The Real National Income of Soviet Russia since 1928. Harvard U.P.;Oxford U.P, 1961.

Chappelow, Jim. “Gross National Product (GNP) Definition.” Investopedia, Investopedia, 15 Apr. 2019,

Berkowitz, Bruce D. “U.S. Intelligence Estimates of the Soviet Collapse: Reality and Perception.”, The Brookings Institution Press, 2008,

Eberstadt, Nicholas. The Tyranny of Numbers: Mismeasurement and Misrule. AEI Press, 1995.

GAO, US. “Soviet Economy: Assessment of How Well the CIA Has Estimated the Size of the Economy.” U.S. Government Accountability Office (U.S. GAO), 30 Sept. 1991,

Gregory, Paul R. Russian National Income, 1885-1913. Cambridge Univ. Pr., 1982.

Nove, Alec. Soviet Economic System. Allen & Unwin, 1986.

Ott, Mack. “National Income Accounts.” Econlib, 5 Feb. 2018,

Pine, Art. “Soviets Say CIA Overestimates Size of Their Economy : Foreign Affairs: The U.S. View of Gross National Product Is Rosy Compared to Reality, Panel Asserts.” Los Angeles Times, Los Angeles Times, 24 Apr. 1990,

Reuss, Henry S. “USSR: Measures of Economic Growth and Development, 1950-80.”, Joint Economic Committee , 8 Dec. 1982,

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s