# Research Insider
North Korea’s GDP: Features and Issues
The basis for studying the North Korean economy is to gain an understanding of the size of the overall economy and the trends within it. However, owing to the fact that North Korea does not release data on its GDP and GNI, the task has fallen on the shoulders of outside researchers and such institutions as the Bank of Korea (BOK). Still, with such a lack of official data, the estimates produced have been subject to much controversy, and the question of accuracy and propriety remains unanswered.
Against this backdrop, we at the KDI Office of North Korean Economic Studies invited Dr. Kim, Byung-Yeon, an economics professor at Seoul National University and one of the most prominent experts in this area, for an interview.
♦ Date and Venue: Namhakdang, Friday, June 12, 2020
♦ Interviewer: Lee, Suk (Senior Fellow at KDI)
♦ Interviewee: Kim, Byung-Yeon (Economics Professor at Seoul National University)
Estimating North Korea’s GDP Prior to the 1990s
For this issue, we are privileged to have Dr. Kim, Byung-Yeon, economics professor at Seoul National University, with us to talk about an important economic aggregate of North Korea, the GDP. Dr. Kim is a leading expert on the North Korean economy and GDP, and as such, I am certain that we will be able to learn a lot from today.
Lee, Suk Q.
Since 1990, the Bank of Korea (BOK) has periodically released its growth rate estimates for the North Korean economy in cooperation with other South Korean government agencies. As an important source of information on the status of the North Korean economy, these figures are widely used by researchers, and hold considerable sway over estimations on economic growth. For the sake of convenience, let us divide the period into pre-1990 and post-1990. But first, I would like to ask you about concepts of an economic aggregate in a socialist country. North Korea adopted the socialist economic system, and as far as we know, the terms used in such an economy differ from those used in an market economy, for example, GDP and GNI. This contrast would have been much more marked before 1990. So, in relation, could you tell us about the indicators used in socialist economies and the differences with those used in market economies. Also, how would we interpret them?
Kim, Byung-Yeon GDP is the fundamental on which evaluations of a country’s economic conditions are based. However, estimates on North Korea’s GDP and GDP growth rate have been the subject of endless controversy due to the lack of statistical data in particular, and also due to the different approaches used by researchers.
In terms of the concepts for economic aggregates, Marx believed that, in principle, services do not create value-added. This notion laid the foundation for the use of ‘net material product (NMP)’ instead of GDP in socialist economies. It was understood that value-added could only be generated through the input of labor and subsequent production of materials. Hence, buying goods at a low price and selling it at a higher price as is customary in distribution was not deemed to be creation of value. Rather, it was ruled as an act to self-profit, and was punishable on the grounds of speculation.
Thus, NMP can be understood to be GDP minus the value-added created in the service industry. According to CIA estimates, the Soviet Union’s GDP was 25% higher than its NMP. That difference of 25% is the value-added generated through services. The same is true for North Korea where education, distribution, and financing are considered non-value-added sectors. Having said that, we have not even seen data on its NMP for a long time.
You are absolutely right. North Korea has released rough estimates of its NMP a few times in the past, but it has hardly been regular. In regard to the 25% difference, while we can estimate and understand it as you have explained, there is still a lot of confusion. So, are there any other ways to estimate GDP without official data from North Korea. That is, in addition to converting the economic aggregate of socialist economies into GDP, are there ways to measure the GDP of economies like North Korea using other indicators?
North Korea released relatively detailed statistics in the 1950s, but stopped publishing official data from the 1960s. Japanese scholars, Goto and Niwa, tried to estimate North Korea’s GDP using a method similar to the NMP-based method applied by the CIA to estimate the Soviet’s GDP (Goto, 1990; Niwa and Goto, 1989). However, while this worked until 1960, the lack of data made this endeavor no longer feasible. In addition, because there was significantly less data for North Korea than the Soviet Union, various assumptions had been made in the estimation of GDP, making the accuracy of the estimates somewhat suspect.
Even though we cannot directly assess North Korea’s GDP, there are other alternatives. One is to use real indicators which are related to GDP. For example, researchers studying the income of periods for which statistical data are mostly unavailable, such as ancient or medieval times, use people’s height as a proxy for per capita income. Specifically, it is assumed that in come is closely tied to calorie intake, and as such, the higher the calorie intake, the taller a person would be. But, caution is required here. We need to differentiate between flow and stock. For instance, while a person’s height is a stock variable that is measured at a particular point in time, GDP is a flow variable. Therefore, using height, it would be difficult to estimate per capita income that changes annually.
Real indicators include production indicators such as cement and steel output, and nutrition indicators such as daily calorie intake. And while we can also use health indicators, such as infant mortality and life expectancy, to estimate per capita GDP, there are limitations as health indicators which also possess properties of a stock variable. When there is no available data on the GDP of a certain country, the real indicator method will first examine the relationship between the real indicators and GDP of other countries, and subsequently use this relationship to make estimations for GDP of the country under investigation. Specifically, after the standard parameters are obtained using data from other countries, estimates for the GDP-a dependent variable-are computed by multiplying these with the amount of indicators (cement and steel output, nutrition, health, etc.) for a data-lacking country. Another real indicator that is commonly used is night lights. In this case, we evaluate the relationship between the night lights and per capita income of other countries, and then apply this relationship to a country that has data for night lights but not for per capita income to obtain the latter.
However, because there is considerably less data for North Korea than for the Soviet Union and other socialist states, the problem is much more serious. If the availability of official data for the Soviet Union is 100, then, I believe, that for North Korea was only 10 up to the early 1960s, and not even 5 thereafter. Therefore, the CIA method does not work for North Korea. Instead, researchers more often use the short-cut method. Using real indicators can be considered as one of the short-cut method.
"If the availability of official data for the Soviet Union is 100, then, I believe, that for North Korea was only 10 up to the early 1960s, and not even 5 thereafter."
Although not a real indicator, North Korea’s figures on budget and settlement of accounts is the only official data that is regularly released by the regime. Some scholars assume that there is a constant and stable relationship between budget and GDP, and use these data to estimate national income. In the case of the Soviet Union, government expenditure was as much as 60-65% of GDP. This high share of government expenditure in GDP is because public finance was used to finance most investments in enterprises in socialist economies. If we assume that the ratio of government expenditure to GDP has been stable in North Korea, one could think that the changes in government expenditure could be used to estimate GDP. However, data on the budget are a nominal variable that includes changes in prices, and thus, this short-cut method would mismeasure or overestimate GDP by confusing the rate of change in the nominal variable with the real economic growth rate. Although, in principle, there should be no inflation in a socialist economy, even the official data from the Soviet Union show an inflation rate of 1% per annum. If the informal sector is included in the estimation, overall inflation would be higher. Moreover, the Soviet government included its central bank loans in the budget income. All of this could be the same in North Korea. Therefore, we have to bear in mind that considering the rate of change in the budget to be the equivalent to GDP growth could be extremely problematic.
Because there is such a shortage of data on the North Korean economy, researchers sometimes produce implausible figures based on some of the assumptions they make. In one case, using the above method of equating the rate of change in fiscal expenditure with GDP growth, it was reported that North Korea posted an annual economic growth rate of 20-30% until the 1990s.
Recent discussions on the North Korean growth rate also show potential for such errors. In a recent paper, the author assessed the economic conditions of North Korea to be reasonable after believing that the 4.2% growth in the 2020 budget plan released in April was the economic growth rate. But, as I pointed out earlier, the budget and settlement of accounts is a nominal variable, and in the Soviet’s case, it included its central bank loans in the budget income. There is no guarantee that North Korea is any different. And, unless specific items of the budget account are clearly stated, doubt will only escalate.
Whether it be using budgetary growth or other nominal variables, or using other data to estimate inflation and deducting it from the budgetary growth rate, it seems to me that there have been diverse and risky attempts made to measure North Korea’s GDP; although not always accurate. Given that, I would like to ask you a more realistic question. How should we view the actual trends in North Korea’s GDP and growth rate? Could you elaborate on your research and explain how these trends have changed since the 1990s?
I think one of the most important studies on the long-term growth of the North Korean economy is that of Kim ByungYeon, Kim Suk-Jin, and Lee Keun (2007).3 The study was published in the Journal of Comparative Economics, and estimates North Korea’s economic growth in 1954-1989. To prevent the growth rates from being distorted by the aforementioned problem of prices, the authors used a methodology that weight averages the output growth rate by industry (agriculture, mining and manufacturing, and services). Specifically, the hidden inflation, which was estimated using the data from the Soviet Union, was deducted to obtain the output growth rate for the mining and manufacturing industry. This adjustment was made because, in a socialist economy, the hidden inflation is included in the industrial growth rate, which could, in turn, lead to an overestimation of the overall growth rate. This approach applies both the CIA’s method and data. That is, the discrepancies between the Soviet’s official growth rate and the CIA’s estimate for GDP growth were equated to be the differences in inflation, and such discrepancies were deducted from the output growth rate of North Korea’s mining and manufacturing industry. In addition, the industrial ratio from North Korea’s report to the UN in 1992 was used to calculate the weighted value of each industry. The structure of each industry was assumed to be the same in 1989 as in 1992, and a reverse estimation was done consecutively until 1954.
Simply put, the authors used data on output rather than figures including price changes in accordance with the CIA method, and deducted the hidden inflation from the growth rate, which includes price changes, to obtain this data. Maddison also used output data to estimate China’s economic growth rate (Maddison, 1995; 1998). This is because, as noted in many studies, there is a high level of uncertainty when it comes to the reliability of China’s official growth statistics. Maddison subdivided China’s industries, and used the output growth rate and weighted value of each industry to assess the economic growth rate. He found that the official growth rate had been overestimated. The method used by Maddison is a form of the short-cut method while our’s (Kim, Byung-Yeon, Kim, Suk-Jin, and Lee, Keun) is a combination of the CIA method and Maddison’s.
I would like to elaborate on the CIA method.4 Estimating the GDP of a socialist economy requires (1) obtaining output data, (2) obtaining price data, and (3) resolving the exchange rate problem to denominate the national income in dollars. However, in terms of output data, even the CIA used the Soviet’s official statistics while making its own estimations for the price data. This is because prices reflect the scarcity value in a market economy while they are arbitrarily set by the government in a socialist economy, and as such, it is difficult to accurately estimate the Soviet’s GDP with the official prices. As for the exchange rate, the agency resolved the issue by estimating the GDP value of the Soviet Union and the United States using the purchasing power parity (PPP) approach, both in ruble and dollars terms, and by obtaining the geometric mean of these two estimates. But in North Korea’s case, obtaining the output data in itself represents a challenge; which is why the short-cut method is so often used.
From the studies published so far, I see that the short-cut method is the most favored means for many researchers to obtain North Korea’s economic growth rate for before the 1990s. So, what about the subsequent economic growth trends?
Initially, growth rates were high, but they have declined gradually. This is a typical pattern of socialist economies such as the Soviet Union and others in Eastern Europe. However, there is a difference. For North Korea, it exhibited a much sharper decline in its growth rate compared to other socialist states from the early 1960s, just as socialism was taking root. In 1954-1989, the average annual growth rate of GDP posted 4.4% while that of per capita GDP marked 1.9%. Still, it should be noted that even these figure may have been slightly overestimated. This is because, although the hidden inflation was deducted for the mining and manufacturing industry, it was not adjusted for the agricultural industry owing to the lack of data.
"South Korea’s per capita GDP already exceeded that of North Korea in 1968."
These findings indicate that South Korea’s per capita GDP already exceeded that of North Korea in 1968. The common notion, however, is that this reversal took place in the 1970s. It also suggests that even the UN’s data can be misleading which reveals that the overtake happened in 1974. According to the data, the per capita GDP of North and South Korea were $480 and $416, respectively, in 1973, and $518 and $575 in 1974.
Led by Bruce Cumings, numerous studies believe that South Korea pulled ahead of North Korea in the mid-1970s. However, UN statistics are usually based on data from its member states. Therefore, if there is fault with North Korea’s statistics, there is also fault with the UN’s. The long-held belief that North Korea was affluent until the mid-1970s is a myth, and the time has come to discard such fallacies. As I said earlier, our (Kim et al.) estimates show that the tables had already turned on the Peninsula at the end of the 1960s. Thereafter, North Korea’s growth rate contracted even further. According to recent BOK data, North Korea’s average annual growth rate in 1956- 1989 was 4.7% (Jo, Se-Hyung and Minjung Kim, 2020). This represents only a 0.3%p difference with our estimates. Although different methodologies were used, similar patterns and result were drawn, leading me to conclude that this reinforces the reliability of our results.
For more, please refer to the attached file.
