A new estimate of Lithuanian GDP for 1937: How does interwar Lithuania compare?

Abstract

Utilizing archival data from the Lithuanian Central Bureau of Statistics, we piece together the most complete gross domestic product (GDP) figure during the interwar period for Lithuania in accordance with the System of National Accounts (SNA 2008). Lithuania and the other Baltic states have largely been ignored by economic historians due to the short time spans of their independence in the interwar period (1918–1940), their relatively small sizes, and the scarcity of data—we attempt to fill this gap in the historical record. As a practical application of our new GDP estimate for 1937, we convert our estimate into 1990 Geary–Khamis dollars so that we can test the hypothesis that Lithuanian economic growth under communism was below average when compared to its European neighbors using data from the Maddison Project. We find that the economic growth rate in Lithuania was above average from 1937 to 1973, even when compared to other communist economies. Alternatively, from 1973 to 1990, growth collapsed and was below average. Using a difference-in-difference strategy we also find a significant GDP per capita penalty for communist countries in Europe over the studied time periods.

Appendices

Appendix 1—Breakdown of Leonardas Dargis NNI and GNI figures

See Table 6.

Table 6 Breakdown of Dargis’s 1937 NNI and GNI estimates into sources of income.

Appendix 2—Statistical details of 1937 Lithuanian GDP calculation

In Table 7 the calculations of total agricultural land are presented. Agricultural land consists of the land used for crops, animal husbandry, and private forests.

Table 7 Total agricultural land in Lithuania in 1937, ha.

In Table 8 the calculations of GVA from agriculture are presented. The total amount of agricultural land was calculated and is given in Table 7; GVO and IC per hectare were taken from the Agricultural accounts and the subsidies for agricultural products taken from “Ūkininko patarėjas”. It was assumed that there were no taxes on products as there is no separate indication of them. Final GVO and IC values were calculated by multiplying GVO and IC per hectare by the amount of agricultural land.

Table 8 Calculations of GVA from agriculture, LT.

In Table 9 the calculations of GVA from forestry are presented. The GVO at basic prices was calculated by multiplying physical output of logs and firewood by the corresponding basic price quoted by Dargis (1975). The IC of forestry was taken from the Forestry department yearbook; it includes office, inventory expenses, forest growing and forest cleaning expenses. The Forestry department figures only account for Lithuania Major. (Lithuania Minor is the Klaipėda region, and Lithuania Major is the rest of the country.) Therefore, the values for the Klaipėda region and private forests under state supervision are calculated by multiplying the IC per hectare ratio by the area of the Klaipėda region and private forests.

Table 9 Calculations of GVA from forestry, LT.

In Table 10 the calculations of GVA from fishing are presented. The GVO at basic prices was taken from Statistical Yearbook of Lithuania. To calculate IC and GVA, average value-added-to-output ratio from sector A was used. It was obtained by dividing GVA from agriculture and forestry by corresponding GVO. This ratio was then multiplied by GVO from fishing to obtain GVA. Intermediate consumption was then calculated by subtracting value-added from output.

Table 10 Calculations of GVA from fishing, LT.

In Table 11 the calculations of GVA from unaccounted large industrial enterprises are presented. The GVO was taken from Dargis’s article. To calculate IC and GVA, average value-added-to-output ratio from accounted large industrial enterprises was used. It was obtained by dividing GVA from large industrial enterprises by corresponding GVO. This ratio was then multiplied by GVO from Dargis’s article to obtain GVA. Intermediate consumption was then calculated by subtracting value-added from output.

Table 11 Calculations of GVA from unaccounted large industrial enterprises, LT.

In Table 12 the calculations of GVA from small industrial enterprises and crafts are presented. The procedure of calculating output was used from Fridbergas’s book, and gross value-added and intermediate consumption were both calculated using GVA-to-GVO ratio obtained in Table 9. According to the Statistical yearbook and magazine called “Amatininkas”, there were 5207 small industrial enterprises (VIII patent category), 7789 urban craftsmen and 19,661 rural craftsmen in Lithuania Major (Central Bureau of Statistics 1938) (Tamulevičius 1939). Numbers for rural craftsmen are for the year 1938, but it is assumed that the number was the same in 1937. In Klaipėda region, according to Fridbergas, there were 742 small industrial enterprises and 906 urban craftsmen (Fridbergas 1988). Number of rural craftsmen in Klaipėda region was obtained by calculating the ratio of urban craftsmen in Klaipėda region to urban craftsmen in Lithuania Major. Adding the numbers together, the total amounts of enterprises and craftsmen were obtained. According to Fridbergas’s procedure, then these numbers should be multiplied by average number of workers in these enterprises to obtain the final numbers of workers. Fridbergas used the average number of workers without including the owner as a productive worker. In small craft companies the owner worked no less than his assistants (Dargis 1975). Therefore, the average number of workers used by Fridbergas for small industrial enterprises was raised by one (Fridbergas 1988). Then the figures are multiplied by assumed output per worker ratio to obtain final GVO value (Fridbergas 1988). Output per worker for rural craftsmen was assumed to be the same as for urban craftsmen.

Table 12 Calculations of GVA from small industrial enterprises and crafts, LT.

In Table 13 the calculations of GVO from private construction are presented. As repair expenses were not mentioned anywhere in the yearbook, Dargis suggests adding 10% for repair expenses (Dargis 1976).

Table 13 Calculations of GVO from private construction, LT.

In Table 14 the calculations of GVO from public construction are presented. These construction activities include railway construction works, land drying, etc. As repair expenses for public residential buildings were not mentioned anywhere in the yearbook, Dargis suggests adding 10% for repair expenses (Dargis 1976).

Table 14 Calculations of GVO from public construction, LT.

Table 15 represents the separation of enterprises that were considered “Trade in broad meaning” during the interwar years into economic sectors of origin ruled by ISIC.

Table 15 Lithuanian trade enterprises in 1937, distributed among the ISIC economic sectors.

In Table 16 the calculations of other taxes less subsidies paid by enterprises of the “Trade in broad meaning” are presented. The separation of these enterprises into modern economic sectors is presented in Table 15. Other taxes less subsidies on production are the taxes levied on firms regardless of the amounts of production sold or produced. These taxes include license fees, labor taxes, etc. These taxes and subsidies were taken from the Budget execution account. However, job income tax paid by the enterprises was not separated on the basis of sectors the firms operated in. This was overcome by taking the share of population working in “Trade in broad meaning” as a percentage of all working population as a proxy indicator for the share of taxes payable. Such use of working population was extensively used by authors of Czechoslovak aggregate production in the interwar period (Pryor et al. 1971) and was discussed in the review of literature.

Table 16 Calculations of taxes paid by Lithuanian “Trade in broad meaning” enterprises, LT.

In Table 17 the calculations of gross value-added from sector G are provided. The percentage of value that comes from sector G in the “Trade in broad meaning” is taken from Table 15. Information on the number of enterprises, number of employees, average yearly salary, and net operating surplus (NOS) was taken from Dargis’s article. Compensation of employees (CE) was calculated by multiplying number of workers by average salary. It is assumed that profits indicated by Dargis are equal to NOS. Other taxes less subsidies on production (other production (t − s)) figure are taken from Table 16. Values for sector G only were obtained by multiplying values for “Trade in broad meaning” by 79.7% from Table 15. Consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis.

Table 17 Calculations of GVA from sector G, LT.

In Table 18 the calculations of gross value-added from railway and land transportation are provided. It was assumed that revenues of transportation services in the sources were presented at producer’s prices. GVO at basic prices was obtained by subtracting taxes less subsidies on products. Figures of GVA were obtained by subtracting IC from GVO at basic prices. For private railways and land transportation, only output figures were available; therefore, a GVA-to-GVO ratio and taxes (less subsidies) on products ratio from public railways were calculated and used.

Table 18 Calculations of GVA from railways and land transportation, LT.

In Table 19 the calculations of gross value-added from postal, telegraph, telephone, and radio services are provided. It was assumed that revenues in the sources were presented at producer’s prices. GVO at basic prices was obtained by subtracting taxes less subsidies on products. Figures of GVA were obtained by subtracting IC from GVO at basic prices. IC components of postal services, which are attributable to sector H in Ministry of Transport and Communications yearbook, were presented in yearbooks together with telephone, telegraph, and radio services which are attributable to sector J. The separation was done by taking the percentage of output from postal services to total output and using this ratio to split the intermediate consumption between postal services and telegraph and telephone services. Then the obtained IC was subtracted from GVO and separate GVA values thus calculated.

Table 19 Calculations of GVA from postal, telegraph, telephone, and radio services, LT.

In Table 20 the calculations of gross value-added from distribution of trade goods and water transportation are provided. The calculations were done using the component of GVA approach, used in sector G calculations. For distribution of trade goods, the percentage of value that comes from sector H in the “Trade in broad meaning” is taken from Table 15. Information on the number of enterprises, number of employees, average yearly salary, and net operating surplus (NOS) was taken from Dargis’s article. Compensation of employees (CE) was calculated by multiplying number of workers by average salary. It is assumed that profits indicated by Dargis are equal to NOS. Other taxes less subsidies on production (other production (t − s)) figure are taken from Table 16. Values for sector H only were obtained by multiplying values for “Trade in broad meaning” by 8.8% from Table 15. Consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis. For water transportation, the CE figures indicated by Dargis were used. It was assumed that there were no profits and no taxes in water transportation as there is no information on this.

Table 20 Calculations of GVA distribution of trade goods and water transportation, LT.

In Table 21 the calculations of gross value-added from sector I (accommodation and food service activities) are provided. The calculations were done using the component of GVA approach, used in sector G calculations. The percentage of value that comes from sector I in the “Trade in broad meaning” is taken from Table 15. Information on the number of enterprises, number of employees, average yearly salary, and net operating surplus (NOS) was taken from Dargis’s article. Compensation of employees (CE) was calculated by multiplying number of workers by average salary. It is assumed that profits indicated by Dargis are equal to NOS. Other taxes less subsidies on production (other production (t − s)) figure are taken from Table 16. Values for sector I only were obtained by multiplying values for “Trade in broad meaning” by 9.4% from Table 15. Consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis.

Table 21 Calculations of GVA from sector I, LT.

In Table 22 the calculations of gross value-added from financial (banking) activities are provided. The GVO at basic prices of banking is calculated by taking the difference between the total interest revenues and total interest expenses in Lithuania in 1937 which was done by Dargis in 1978. The intermediate consumption (IC) and gross value-added (GVA) were estimated by calculating the GVO, IC, and GVA of largest Lithuanian banks in 1936 from their published income statements and using the same GVA-to-GVO ratio for the whole Lithuanian banking sector in 1937. The ratio was then multiplied by the total GVO of Lithuanian credit sector estimated by Dargis to obtain GVA. The difference between these two figures is IC.

Table 22 Calculations of GVA from financial activities (banking), LT.

In Table 23 the calculations of gross value-added from insurance activities are provided. The GVO was calculated by taking the difference between the total premiums received and total benefits paid in 1937. IC and GVA were calculated using the GVA-to-GVO ratio obtained in Table 22 as there are no separate indications of IC of insurance activities. Data on activities by non-Lithuanian insurance companies were only available for the year 1936, so it was assumed that the figures were the same in 1937.

Table 23 Calculations of GVA from insurance activities, LT.

In Table 24 the calculations of gross value-added from all real estate activities in Lithuania in 1937 are provided. The total real estate taxes amount was obtained from the Budget execution account for treasury taxes and from Statistical Yearbook of Lithuania for municipal taxes. Revenues from government real estate activities were also taken from the budget. GVO at producer’s prices was obtained by dividing the total tax amounts by the real estate tax rate provided by Dargis and adding the governmental real estate revenues. GVO at basic prices was obtained by subtracting real estate tax amounts (taxes on products) from the GVO at producer’s prices. Ratio of GVA to GVO was taken from the European commission’s database. Multiplying the ratio by GVO at basic prices the GVA was obtained and the difference between the two produced the IC figure.

Table 24 Calculations of GVA from total real estate activities, LT.

In Table 25 the calculations of gross value-added from real estate activities that are part of sector T are provided. The only information on activities of households as real estate lenders is provided in the Agricultural accounts. GVO at producer’s prices was obtained by multiplying the total amounts of agricultural land, presented in Table 7 by the leasing output per hectare as recorded in the Agricultural accounts. GVO at basic prices was obtained by subtracting real estate tax amounts (taxes on products) from the GVO at producer’s prices. Ratio of GVA to GVO was taken from the European commission’s database. Multiplying the ration by GVO at basic prices the GVA was obtained and the difference between the two produced the IC figure.

Table 25 Calculations of GVA from total real estate activities of sector T, LT.

In Table 26 the calculations of gross value-added from sector M (professional, scientific and technical activities) are provided. All activities that were included in this sector in interwar Lithuania were non-market, meaning that the prices charged for the services provided by these activities were insignificant and did not cover the costs of service production. Therefore, the components of value-added approach had to be used. For non-market activities net operating surplus (NOS) is zero. The GVA of sector M is therefore the sum of CE, other taxes less subsidies on production and CFC. The compensation of employees (CE) and other taxes less subsidies on production were taken from the Budget execution account. Consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis.

Table 26 Calculations of GVA from sector M, LT.

In Table 27 the calculations of gross value-added from sector N (administrative and support service activities) are provided. The calculations were done using the component of GVA approach, used in sector G calculations. The percentage of value that comes from sector N in the “Trade in broad meaning” is taken from Table 15. Information on the number of enterprises, number of employees, average yearly salary, and net operating surplus (NOS) was taken from Dargis’s article. Compensation of employees (CE) was calculated by multiplying number of workers by average salary. It is assumed that profits indicated by Dargis are equal to NOS. Other taxes less subsidies on production (other production (t − s)) figure are taken from Table 16. Values for sector N only were obtained by multiplying values for “Trade in broad meaning” by 0.38% from Table 15. Consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis.

Table 27 Calculations of GVA from sector N, LT.

In Table 28 the calculations of gross value-added from sector R (arts, entertainment, and recreation) are provided. The calculations were done using the component of GVA approach, used in sector G. The percentage of value that comes from sector R in “Trade in broad meaning” is taken from Table 15. Information on the number of enterprises, number of employees, average yearly salary, and net operating surplus (NOS) was taken from Dargis (1976). Compensation of employees (CE) was calculated by multiplying the number of workers by the average salary. It is assumed that the profits indicated by Dargis (1976) are equal to NOS. Other taxes less subsidies on production are taken from Table 16. Values for sector R were obtained by multiplying values for “Trade in broad meaning” by 0.43% from Table 15. In addition, some activities in sector R were non-market (meaning zero profit or insignificant prices). These activities were recorded in the Budget execution account and in the Statistical Yearbook of Lithuania. Separate compensation of employees for municipal recreation services was not available. This issue was solved by taking the average municipal salary-to-ordinary expenses ratio (35.7% of 6,519,400.6 LT of ordinary recreation expenses) from the municipal accounts (in the Statistical yearbook) and using it to calculate the CE of municipal activities. The consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis.

Table 28 Calculations of GVA from sector R, LT.

In Table 29 the calculations of gross value-added from sector S (other service activities) are provided. The calculations were done using the component of GVA approach. The percentage of value that comes from sector S in “Trade in broad meaning” is taken from Table 15. Information on the number of enterprises, number of employees, average yearly salary, and net operating surplus (NOS) was taken from Dargis (1976). Compensation of employees (CE) was calculated by multiplying the number of workers by the average salary. It is assumed that the profits indicated by Dargis are equal to NOS. Other taxes less subsidies on production are taken from Table 14. Values for sector R only were obtained by multiplying values for “Trade in broad meaning” by 0.62% from Table 13. In addition, some activities in sector S were non-market (meaning zero profit). These activities were recorded in the Budget execution account and were also calculated here. Consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis.

Table 29 Calculations of GVA from sector S, LT.

In Table 30 the calculations of gross value-added from sector O (public administration and defense; compulsory social security) are provided. All activities that were included in this sector in interwar Lithuania were non-market, meaning that the prices charged for the services provided by these activities were insignificant and did not cover the costs of service production. Therefore, the components of value-added approach had to be used. For non-market activities net operating surplus (NOS) is zero. The GVA of sector O is the sum of CE, other taxes less subsidies on production and CFC. The compensation of employees (CE) and other taxes less subsidies on production were taken from the Budget execution account for treasury expenses and from Statistical Yearbook of Lithuania for municipal expenses. Separate compensation of employees for the Klaipėda region and the Lithuanian municipalities was not available. This issue was solved by taking the average salary-to-total ordinary expense ratio from the Budget execution account and using it to calculate CE of Klaipėda region activities, and by taking the average salary-to-total ordinary expense ratio from the municipal accounts (from the yearbook) and using it to calculate CE of municipal activities. Compensation of employees also includes non-monetary salary, like social benefits, etc. As there were not enough data on military activities, an estimation of CE received by soldiers made by Dargis was used. Consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis.

Table 30 Calculations of GVA from sector O, LT.

In Table 31 the calculations of gross value-added from sector P (Education), are provided. Because there were no data on private education, only public education GVA was calculated here. As public education is a non-market activity, the components of value-added approach had to be used. For non-market activities net operating surplus (NOS) is zero. The GVA of sector P is therefore the sum of CE, other taxes less subsidies on production and CFC. The compensation of employees (CE) and other taxes less subsidies on production were taken from the Budget execution account for treasury expenses and from the Statistical Yearbook of Lithuania for municipal expenses.

Table 31 Calculations of GVA from sector P, LT.

In Table 32 the calculations of gross value-added from sector Q (human health and social work activities) are provided. Because there were no monetary data on social work activities and private health institutions, only public hospitals’ GVA was calculated here. As public human health activities are non-market, the components of value-added approach had to be used. For non-market activities net operating surplus (NOS) is zero. The GVA of sector Q is therefore the sum of CE, other taxes less subsidies on production and CFC. The source for hospital expenditures was the Statistical Yearbook of Lithuania. Consumption of fixed capital (CFC) norm for interwar Lithuania was taken from “Makroekonomika” by R. Čiegis.

Table 32 Calculations of GVA from sector Q, LT.

In Table 33 the Lithuanian gross value-added figures of all ISIC sectors A–U for the year 1937 are presented together and aggregated into one GDP at factor cost figure.

Table 33 Sectorial breakdown of GVA figures and calculation of GDP at factor cost, LT.

In Table 34 Lithuanian gross domestic product at market prices figure is obtained by adding taxes less subsidies on products to GDP at factor cost. Taxes less subsidies on products are the ones already presented in Tables 8, 18, 19, and 24 plus total excise taxes from the Budget execution account.

Table 34 Sectorial breakdown of GVA figures and calculation of GDP at factor cost, LT.

Appendix 3—Difference-in-difference regressions

In Table 35 we run a regression to find the difference-in-difference estimator for the effect of communism on economic growth from 1937 to 1990. Regression was run in order to provide estimates of the standard errors and p-values for the coefficients. The diff-in-diff estimator is the interaction term: 1990*Communism.

Table 35 Difference-in-difference regression.

Table 36 Difference-in-difference regression.

We also attempted to find the diff-in-diff estimator for the communist countries in Europe which were not a part of the Soviet Union to use as a comparison. However, the results of the regression are not statistically significant (Table 36).