National Accounts – Deutschland in Daten

Thomas Rahlf

English translation of: Thomas Rahlf, Volkswirtschaftliche Gesamtrechnung, in: Thomas Rahlf (Ed.), Deutschland in Daten. Zeitreihen zur Historischen Statistik, 2nd ed., Bonn: Bundeszentrale für politische Bildung 2022, pp. 192-205.

Citation
Thomas Rahlf, National Accounts, in: www.deutschland-in-daten.de, 20.12.2022
< http://www.deutschland-in-daten.de/en/national-accounts >.

Copyright (c) 2022 by Deutschland in Daten, and the author, all rights reserved. This work may be copied and redistributed for non-commercial, educational purposes, if permission is granted by the author and usage right holders. For permission please contact kontakt@deutschland-in-daten.de.

Introduction

National accounts are among the most complicated statistics devised.¹ Unlike many other statistics, which are simply „counted“ (births, fruit trees, votes cast, etc.), here extensive and complex account systems have to be formed and converted into units of quantity that can be added up into units of value in order to arrive at the end at the „one number“², the national or social product, which, according to theory, depicts the performance of an economy. This attempt has meanwhile itself become the subject of historical investigations, on the part of the specialized science³ as well as official statistics⁴. However, organizational developments and discourses have been in the foreground and less the numbers per se. To date, there has been no systematic analysis of the national accounts as historical statistics. Within the framework of its current system, the national accounts can be presented since 1950, taking into account some limitations. Since then, official statistics have provided official data for the Federal Republic of Germany. For the period and the system of the GDR as well as for the period of the German Reich, corresponding data have to be reconstructed.

Terms and definitions

So what exactly are national accounts? In its current form, the quantity expressed in a social product is viewed from three sides: an origin side, a usage side and a distribution side.⁵ On the origin, or output side, it is calculated how the national product was produced. Several sectors are distinguished, ranging from agriculture, forestry and fishing to public and private service providers. If taxes on products are added to this gross value added and subsidies on products are deducted, the gross domestic product is obtained. From the usage or expenditure side, GDP is the sum of consumption expenditure by households and nonprofit institutions serving households, government consumption expenditure, investment and net exports (exports minus imports). From the distribution side, GDP is composed of national income, which in turn consists of the two components compensation of employees and corporate and property income, taxes on production and imports less subsidies, consumption of fixed capital and, finally, the balance of primary income from the rest of the world. Without this last component, gross domestic product is called gross national product (or gross national income); without depreciation, it is called net national product (or net national income) at market prices. Finally, national income is then referred to as net national product (or net national income) at factor cost.
Diagram 1 illustrates these components schematically. Overall, the calculation of the individual variables is very complex and is based on a large number of different ongoing surveys, supplemented by administrative data and other information. In terms of its compilation and dissemination, the national accounts are among the most expensive statistics produced in Germany.

National Accounts as Official Statistics

Today’s system of national accounts has a long history. In Germany, it began shortly after the end of the First World War. Germany had to make reparation payments as part of the „Dawes Plan“, the amount of which was to be determined by a so-called „prosperity index“. Its construction was highly controversial. The Imperial Statistical Office therefore sought alternative calculations. In 1926, an „Enquete Committee“ was set up, whose task was to determine the level of national income in the pre-war period. In 1932, the commission was able to present an initial result, an estimate of national income from 1891 to 1913 and from 1925 to 1931 based on the states of Prussia, Saxony, Baden, Württemberg, and Bavaria, Hesse and Hamburg. A calculation according to the generation or expenditure side was not possible with the available material. Only a calculation according to the income/distribution side could be carried out by taking income tax statistics as a basis.⁶ After World War II, a consistent calculation of output, expenditure and distribution was introduced for the first time in 1957, and three years later an overall accounting system was published for the first time. Since then, the system has been continuously expanded, refined, consolidated and repeatedly adjusted in the course of revisions. Since the 1990s, the changes have been oriented to a large extent to the needs of international standardization.⁷ On the one hand, these revisions in particular make it difficult to compare the national accounts over time. On the other hand, they are necessary when the data situation improves, e.g. because large-scale censuses conducted over several years are now available, because the methods of data collection have changed or improved, or because conceptual changes have occurred in the large and complex system of national accounts that make it necessary to reorganize individual components or add new ones. In the course of such revisions, extensive corrections to past years and decades are also made on a more or less regular basis. Between 1953 and 2019, there have been 16 such revisions in Germany alone, known as general revisions,⁸ which have also found their way into the daily press under headlines such as „Why you will be 1000 euros richer tomorrow morning“ or „Statistically, Germany is now 42 billion euros poorer.“⁹

Main lines of the development

A long-term view since the mid-19th century is only possible to a limited extent due to the numerous conceptual changes and data problems. This is true even for the „one figure.“ Until the early 1990s, gross national product was considered the central figure in national accounts in Germany, then it was replaced by gross domestic product, as has been the international practice since then. Neither the one nor the other has been consistently available for the past 170 years. The gross domestic product prevailing today can be calculated back to 1925, and for the period up to 1901 to 1913, in the absence of data on the balance of primary income from the rest of the world, only the gross national product. Since neither direct data nor reliable estimates of depreciation exist for the 19th century, net national product at market prices, for which estimates are available from 1851 to 1913, must be used as a substitute for this period. On the other hand, net national product cannot be used as a yardstick throughout the period up to the present, as the share of depreciation has changed considerably over the past decades.

Despite these limitations, long-term trends can be identified. In order to exclude territorial changes, population growth and currency devaluations, the „real“ value per capita is used, i.e. adjusted for inflation or price changes. According to current research and official statistics (Fig. 1 ), per capita domestic/social product has increased in real terms

approximately doubled in the second half of the 19th century – with large uncertainty factors,
was already back to the 1936 level in 1950, has
in the Federal Republic of Germany more than quadrupled in the forty years from 1950 to 1989,
in the GDR increased over the same period from just under fifty to about 55 percent of the level in the Federal Republic,
has fallen by over ten percent as a result of reunification,
in Germany as a whole has fallen by 40 percent in the last thirty years since reunification, as well as
increased about sixfold from the beginning to the end of the 20th century.

All this in the context of a significant overall reduction in working hours, rising shares of part-time employment and a substantial drop in the share of persons of working age. This secular development masks cyclical upswings and downswings, which become visible when looking at the annual growth rates (Fig. 2).

For the second half of the 19th century, an average growth rate of 1.6 percent reveals boom phases in which growth was considerably above this average for several years, but also a good dozen „recession years“ in which it was clearly negative. It must be borne in mind that large shares of production first became „marketable“ from the mid-19th century onward, i.e., increasingly moved from the non-measurable domestic sector to the measurable commercial sector. This transformation, which cannot be practically measured statistically, leads to a certain overestimation of growth in this period because the early phase is underestimated in terms of its output, which is measured in too small a proportion in official statistics. As industrialization progresses, the agricultural sector not only loses its quantitative importance, but also the dynamics of its ups and downs become increasingly independent of agricultural crises and business cycles.¹⁰ The more rapid growth from the 1880s onward, which matched that of the 1980s in terms of size, also led to a noticeable improvement in living standards, albeit accompanied by extreme income and wealth inequality.¹¹
The first half of the 20th century, on the other hand, is characterized by extraordinarily sharp swings. The sharp declines that began with the First World War and lasted several years were transformed in 1920 into equally sharp increases, which came to an abrupt end in 1923, the year of inflation, with the sharpest slump ever recorded. A further growth phase is followed by the most severe crisis of the past 170 years from 1929 to 1932. In these years, the national product as a whole declines by about 30 percent, but rises again all the more sharply in the following years. With the same average growth rate as in the second half of the 19th century, this resulted in a completely different dynamic.
The old Federal Republic, whose economy grew by an average of around four percent per year over its four decades, experienced only four – comparatively minor – years of recession after the economic miracle of the 1950s. While annual growth rates averaged over 7 percent in the 1950s, they fell to about 3.5 percent in the following decade, the period of full employment, to less than three percent in the 1970s, the decade of the oil crisis and the first structural problems, and to less than 1.5 percent in the 1980s. The individual factors of growth can only be touched on here: a high labor force potential, high capital stock, strong demand from home and abroad, and a substantial increase in labor productivity were the main, partly mutually beneficial factors that led to the „boom“. The first dip came in 1967, when the national product stagnated for the first time since the war, prompting the Act to Promote the Stability and Growth of the Economy, which is still in force today. The law committed the public budgets to the four goals of price level stability, a high level of employment and external balance with steady and adequate economic growth. Looking at the statistics, it can be said that these objectives were well met in the 1960s, with the exception of the year of the law’s introduction, unlike in subsequent decades. As early as the 1970s, not only did the first year show a declining national product in both nominal and price-adjusted terms, but prices and unemployment figures climbed to heights unheard of since the war. In 1975, the national product was lower than in the previous year for the first time, and in 1982 for a second time; in both cases, this recession was triggered by an „oil price shock“. All in all, the economy of the old Federal Republic underwent a profound structural change. Ten years later, with reunification, 16 million Germans were integrated into the Federal Republic, whose gross domestic product per capita was, optimistically estimated, half that of West Germany. The challenge associated with the reunification and reintegration of East Germany has left long-lasting traces in the national accounts statistics. It was not until the beginning of the 21st century that real GDP per capita again exceeded the value it had reached in the Federal Republic in the year before reunification – although it is currently still only around 75 percent of the western value in the new federal states. However, this period also saw a financial crisis that was unprecedented in post-war history, with a drop in social product of more than five percent. Average growth in reunified Germany is back to the levels seen between 1850 and 1950, although the decline in the hitherto strongest crisis years of 2009 (financial crisis) and 2020 (pandemic) of 5.4 and 4.8 percent respectively was rather moderate in a long-term comparison.

Let us now take a look at the individual pages of the national accounts. Up to the middle of the 20th century, we have to accept gaps and, in part, take into account large uncertainty factors.

Origin of domestic product

In terms of value added, a fundamental structural change in the economic sectors is noticeable (Fig. 3). The decline of the agricultural sector is most obvious. Its share, which certainly accounted for half of value added in the mid-19th century, was only about ten percent after World War II and fell steadily to one percent by the early 1990s. It has remained at this level ever since. The shares of the other three sectors cannot be described quite so clearly. Here, there are strong differences in real versus nominal development. In both nominal and price-adjusted terms, the service sector overtook the manufacturing sector in the early 1980s, has had the largest share of value added since then and has accounted for more than half of total value added since the mid-1990s. The manufacturing sector remains at a high level until the early 1970s. Although a high level of supply to the population had already been reached before then, the incipient export orientation of the Federal Republic is already evident here.

Between 1950 and 1980, the price-adjusted development is different: in nominal terms, both the service and trade sectors had a share of 20 percent, while manufacturing accounted for half of value added.¹² In 1991 prices, by contrast, the service and manufacturing sectors were on a par in 1950, each with a share of about 40 percent. Until about 1970, however, the share of the service sector then falls initially, while the share of trade has risen steadily since 1950 but has declined in nominal terms. The reason lies in the different price developments of the individual sectors: the service sector was able to benefit far less than the manufacturing sectors from the drop in production costs caused in particular by technical progress. In comparison, therefore, service prices rose disproportionately. Since the turn of the millennium, the shares of the individual sectors have remained largely stable. This does not take into account the accompanying and in some cases even more serious shift in the labor force. For example, the share of employees in the manufacturing sector has declined much more sharply than the corresponding share of value added: A reflection of the equally dynamic and varied development of labor productivity and the impact of technological progress.

Use of domestic product

Structural shifts in use of the domestic product are smaller (Fig. 4), but a fundamental change is also evident here for individual areas. The share of private consumption was around 75 percent in the 1920s, a share that can probably also be assumed for the second half of the 19th century and the early 20th century. Over the past seventy years, too, private consumption formed by far the largest share of the national product, although over the decades it fell from over 60 percent to initially about 56 percent and since 2010 again to now about 52 percent. Government consumption accounted for a much smaller share over the entire period. While its share doubled from around 10 to 20 percent within a few years up to 1938 – a development repeated in the Federal Republic between 1950 and 1975 – it has remained constant at this level since then. Here, too, the price-adjusted side shows a different order of magnitude: private consumption rose from 50 to 55 percent, while the government’s share has been at the level of 20 percent since as early as 1950 and has thus been above the shares of imports and exports since the early years of the Federal Republic. The reason again lies in a change in the price structure: During the first three decades of the Federal Republic, prices in the area of private consumption rose by an average of 3.4 percent annually, while those of government consumption – essentially determined by the wages and salaries of government employees – rose by 5.7 percent.¹³

With regard to the external contribution, the first „globalization shock“ occurred as early as the 19th century.¹⁴ Thus, from the mid-1860s onward, both exports and imports experience an unprecedented surge within a few years. However, the balance of both variables remains largely constant for many years. After World War II, both imports and exports rise steadily, and after a slump in the early 1990s, all the more rapidly, so that the value of both imports and exports now each account for over 40 percent of gross domestic product.¹⁵ This illustrates the extraordinarily strong dependence of the German economy on foreign trade. Finally, investment, about which we do not know too much for the 19th century, was at a fairly high level in a temporal comparison until the early 1970s. This shows the large-scale investments in education and infrastructure, for which there was still a need for a long time after the end of World War II. In the further course one sees declines, in the 1980s long years of the lowest level, to the reunification and for the next ten years again a clearly higher level and finally since the beginning of the 21st century a relapse to the low level of the eighties. But here, too, the pitfalls of statistics become apparent once again. The pre-1991 figures do not take into account the change in national accounts as part of the 2010 revision of the ESA, according to which expenditure on research and development and military weapons systems was still allocated to government consumption and not, as is currently the case, to investment.¹⁶ According to this new definition, investment prior to 1991 is therefore likely to have been higher than reported here, and in the recent past it is lower than ever before compared with past decades.¹⁷

Distribution

Finally, let us turn to the distribution side (Fig. 5). Based on the available data, we can give a) from 1851, shares of compensation of employees, business and property income, and taxes on production and imports less subsidies in net national product, b) from 1901, shares of these three variables plus depreciation in gross national product, and c) from 1925, shares of these four variables plus the balance of primary income in gross domestic product. In the second half of the 19th century, the share of gross income from employment in net national product was consistently between 60 and just under 70 percent, and the corresponding share from entrepreneurial activity and wealth was between 30 and 40 percent. Between 1950 and 1980, the shares of gross income from employment increased significantly, and the corresponding shares from entrepreneurial activity and assets decreased just as significantly. Since then, their shares remained largely constant until around 2004, when the share of income from salaried work subsequently fell to its 1970 level until 2007 and has since risen again significantly. This 2004 slump, which was accompanied by a corresponding increase on the corporate income and wealth side, followed a period in which Germany’s gross domestic product showed the strongest stagnation in the postwar period (2001 to 2005).¹⁸ Overall, Germany can boast a comparatively stable wage share in an international comparison – which, however, says nothing about the distribution of income.

Depreciation, the amount of which is unknown in the 19th century, amounted to about seven to eight percent of gross national product in the first half of the 20th century, but rose continuously after World War II and is now almost 20 percent. Currently, the difference between gross and net national product is thus considerable. The share of taxes remained at a low level, rising from two to about five percent of net national product at the turn of the 20th century. In terms of gross domestic product, the share rises to nearly fifteen percent in the 1930s and levels off at about ten percent in recent decades. Finally, the balance of primary incomes has largely leveled out in recent decades. While the difference between gross national product and gross domestic product has thus been rather small, the past twenty years have seen a comparatively strong increase in the balance, particularly as a result of foreign investment by German investors, so that gross national product is now around two percent higher than gross domestic product, i.e. it has been rising continuously at a faster rate.¹⁹

SBZ/DDR (1945-1990)

It is common knowledge that the per capita GDP in the GDR was lower than in the Federal Republic. But it would be all too easy to attribute this merely to a superior capitalist economic system of the West. In historical reality, things are often more complex. Before looking for the causes in a system comparison, it must first be considered that regional structural differences also came into play in the two states. These are, on the one hand, simply geographical conditions and, on the other hand, the – partly resulting – long-term developments that have shaped such structures over long periods of time. Even in the old federal states, the gross national product was and still is different. For example, the level in Schleswig-Holstein is currently about 80 percent of the level in Bavaria or Baden Württemberg.²⁰ Just as Schleswig-Holstein and the areas of what is now Mecklenburg-Western Pomerania were and are agricultural, the Ruhr region and Saxony were industrial core zones for many decades and system boundaries. This was also true during the time of the GDR. Looking at their starting conditions, the per capita gross domestic product in the territory of what is now the GDR was higher in 1938 than that in the territory of what is now the Federal Republic.²¹ However, the GDR suffered far more than the Federal Republic from dismantling and reparations, so that per capita GDP around 1950, before the establishment of a socialist economic system, was only about half the level of the West.²²

A look at the following decades shows that the level subsequently rose to around 55 percent (table 6, column 2). According to these figures, the GDR was thus even able to improve its relative position by the end. One possible reason among several for this initially surprising development is certainly the far more dynamic involvement of female workers in the economy. Their labor force participation rate far exceeded that of the Federal Republic (chap. 24). However, the development of value-added shares (columns 3-6) shows that the structural development was also completely different from that in the Federal Republic. While the service sector took the lead there in terms of share from the 1970s onward, the manufacturing sector dominated in the GDR from 1955 onward and until the end, statistically accounting for half of value added from around 1970 onward (column 4).²³ Although purely arithmetical figures show an increasing investment ratio over time (column 9), the real economic reality behind these figures is clear: the controlled focus on the industrial sector, combined with a low level of integration into the world market, led to increasing technological backwardness and path dependency, in which a communist industrialization model was stubbornly adhered to with ever fewer and ever larger enterprises. The result was an increasingly sluggish economic process in which the lack of dynamism of a medium-sized enterprise structure became more and more noticeable with each passing year.²⁴

After the fall of communism, the per capita gross domestic product of the new states collapsed to about 43 percent of the value of the old Federal Republic. In the following four years, it was able to „catch up“ strongly, to about 67 percent, but since then only very slowly to currently about 73 percent.²⁵ How intensively legacies from the GDR era continue to have an effect, how much economic policy mistakes since reunification, system-independent regional structural factors or other reasons play a role, and how these factors should be weighted, is a matter of dispute among researchers.

Concluding Remarks

National accounts can provide a comprehensive picture of a country’s economic development. One of its major advantages is that it provides data for central variables of economic theory in a closed system and thus enables very far-reaching statements to be made, also in intertemporal and international comparisons. At the same time, however, it is also an extremely difficult construction. Some basic conceptual problems are only briefly touched upon here. Large parts of economic activity, such as housework, are performed „outside the market“. On the other hand, „value added“ also includes things that, according to the general understanding, tend to have a negative impact on a country’s economic performance, such as cleaning up environmental pollution or, for some years now, drug trafficking.²⁶ Some of these aspects are taken into account in the national accounts by means of external supplements, for example by additionally compiling an environmental economic account or by recording the services provided free of charge in private households by means of socioeconomic reporting. So far, these supplements remain only so-called satellite systems. But even the national accounts themselves are subject to constant expansion in an evolving world, which makes long-term historical comparisons difficult.
Finally, there is extensive debate about the extent to which even the „one number,“ gross or net social or domestic product, is suitable as a yardstick for assessing prosperity and social progress. For long-term historical comparisons, net social or domestic product offers a number of advantages.²⁷ But even this figure is merely an average value that says nothing about how incomes are distributed. An alternative that is also feasible from a historical perspective is the Human Development Index used by the United Nations, which, in its simplest version, measures life expectancy and access to education in addition to gross domestic product. Finally, this statistic, more than most others, must not be assumed to be scientifically accurate. Decades ago, Oskar Morgenstern, in his famous book „On the Accuracy of Economic Observations,“ demonstrated margins of error in the national accounts in particular that far exceed those of other reporting systems due to their complex construction processes.²⁸ Undoubtedly, measurement techniques are now more accurate, but when comparing over 170 years, the conclusions drawn from such figures should be well considered.

Data basis

The latest data on the national accounts are published by the Federal Statistical Office in a comprehensive form and several times a year in a so-called „national publication program“. Due to the problem of numerous general revisions, which is described in detail in the documentation accompanying this text volume, the „long series“ published by the Federal Statistical Office should be used as far as possible for historical comparisons, as these represent consistent series at least in sections. Data for the GDR were compiled after reunification in a project of the Federal Statistical Office together with the German Institute for Economic Research in Berlin and the Halle Institute for Economic Research. Based on this, Gerhard Heske compiled a national accounts for the GDR for 1950 to 1989, which is the basis for table 6. For long series on the first half of the 20th century, we can refer to the reconstruction by A. Ritschl and M. Spoerer. For the year 1936, an input-output table prepared by R, Fremdling and R. Staeglin is now available, based on newly evaluated archival material. For the second half of the 19th century, C. Burhop. and G. Wolff and U. Pfister presented selected recalculations in critical discussion with W. Hoffmann.²⁹

Appendix: Technical Documentation

Short Version

According to current research, it is not possible to calculate a reliable gross national product, or gross domestic product for Germany for the 19th century. Instead of Hoffmann’s calculations, the net national product calculated by Pfister (2020) should be used.
For the first half of the 20th century, the calculations of Ritschl / Spoerer (1997) should still be chosen.
For the period from 1950 to 1970, the data in Spoerer / Ritschl (1997) are identical with the more comprehensive publication Federal Statistical Office (2021), which contains unrevised results before the introduction of the European System of Accounts in 1995.
For 1970 to 1990, the data published by Ritschl / Spoerer (1997) can be used if comparability of this period with the period before is more important than with the period after.
If a comparison with the period after is more important, the data from Federal Statistical Office (2021), should be used. These are recalculated after the 1991 and 2002 revisions, but not after the 2011, 2014, or 2018 revisions.
The data from 1991 onward fit the previous ones only to a limited extent because they have been recalculated based on the 2011, 2014, and 2018 revisions.

Dataset for the short version: https://doi.org/10.6084/m9.figshare.22704619.v1

Not So Short Version

The following text is an English translation of: Dokumentation zum Zeitreihendatensatz für Deutschland, 1834-2018, Version 01 (= Historical Social Research Transition 32v01), Cologne 2023, pp 1027-1042.

Compilations of the Federal Statistical Office 1950 to 2019

The current data on national accounts are published by the Federal Statistical Office in a comprehensive form and several times a year in a so-called „national publication program“. The frequency of publication depends in particular on the delivery obligations to the Statistical Office of the European Union. In addition to the Genesis-Online database, which offers deeply nested data, albeit only from 1991 onward, the main publication venue is the Fachserie 18 „National Accounts,“ which in turn is subdivided into several subseries. So-called „long series“ from 1970 onward have been published in subseries 1.5 since 2005, exclusively in digital form. Initially planned in two editions per year, further editions have frequently been published as a result of subsequent corrections, with a total of 37 between 2005 and 2019. As part of the general revisions, the series reproduced there are revised back to 1991 in each case, but the data from 1970 to 1990 are left at the level of the 2005 revision. Deflated series are found here only in the form of dimensionless chain indices on a prior-year basis, which the Federal Statistical Office switched to calculating exclusively in 2005. Prior to the introduction of subseries 1.5, long series were published in a special series N, series 3 „Special contributions“ (only one issue in 1972 with data from 1950 to 1970), later in the special series 18 as one-off special volumes: in 1985 as S7 with data for the period from 1950 to 1984, in 1991 as S15 with data for the years 1950 to 1990. The last of these special volumes was published in 2006 and contained revised calculations for the years 1970 to 1991 following the 2005 revision of the national accounts, which were then incorporated into subseries 1.5 at more aggregated levels and have remained unchanged since then. Detailed general notes as well as interpretations of the effects of the revisions are published in parallel in the Federal Statistical Office’s bimonthly periodical Economy and Statistics. The compilation here of national accounts from 1950 onward is based on two publications:

The first four tables in this chapter have been compiled in such a way that the left-hand 5 or 6 columns contain nominal values and the right-hand 5 or 6 columns contain the corresponding price-adjusted values. In table 13.5, only nominal values were included, since net national product at factor cost is not deflated by the Federal Statistical Office. The following diagram illustrates the structure of the first four tables in the colors chosen above:

A largely complete overview of the general revisions, the subsequently revised period in each case, the corresponding publications of the data, and the associated explanations in the journal Economics and Statistics is provided in the following overview. For better orientation, the publications used here have also been color-coded.

Revisions and back calculations

The first calculation of longer series was made in 1972 for the period beginning in 1950, and was considered necessary because the major revision in 1970 required comparable figures for the 1950s (Stat.BA-FS N-3 1972, p. 8). Subsequently, after major revisions, recalculations were made at different rates and for different lengths of time. Whereas in the first revisions, the main aggregates were revised back to 1950 and further differentiated to 1960, in the later revisions, back calculations were only made to 1970, and from 2006 only to 1990 (Brümmerhof / Grömling, 2011, p. 226).

If the current results of the calculation change due to a revision, the question naturally arises how these then fit to the previous values. To answer this question, we first need to understand the purpose of the calculations in the first place. The Federal Statistical Office defines the main purpose of national accounts as the description and analysis of the economic situation, in particular as a basis for economic policy decisions such as tax estimates and pension adjustments, as a basis for studies on economic structure and interdependence, or as a basis for economic forecasts. The main users are the federal government and ministries, the European Commission, supranational organizations or statistical institutions, business associations and social and economic research institutes. ³⁰ It follows from these intended uses that, if the calculations are changed, comparability with past years must also be possible. However, there are no specifications in this context as to how long back into the past this comparability must be established. The Federal Statistical Office initially endeavored to perform such backward calculations from the beginning, i.e., since 1950. Since the 1990s, it has restricted itself to recalculating the data only back to 1970, and since 2006 only for the period from 1991 onward.
For economic history, this naturally raises the question of comparability over longer periods. As described in the text, the 1999 revision increases average GDP growth from 1991-1998 from an average of 3.7 to 4 percent per year. The 2005 revision brings a further increase in the level for all years back to 1991: 1.4 to 2.6 percent higher than before. Once again, the 2014 revision raises the levels significantly, increasing them by no less than 3-3.5 percent from 1991 to 2013. Taking all the changes together, there are considerable changes even for this highest level of aggregation. In the individual components, one has to reckon with much larger corrections. For the shares of individual economic sectors in value added, such as agriculture, gross investment, net exports, government and household final consumption expenditure, or compensation of employees and corporate income, there are sometimes even significantly larger deviations in the double-digit percentage range. For economic history, this poses a challenge because individual years designated as „crisis years“ on the basis of the data turn out to be different in retrospect: for example, the decline in (real) GDP in 1976 from an original -3.6 percent shrank more and more from revision to revision to -0.9 percent in 2006.³¹ Revisions can have an impact on the level, the business cycle as well as the structure of the economy. To address this problem adequately, one should therefore have to look at time not only in one but in two dimensions. It would be neccessary to create for each (!) time series a table in which the year numbers form not only the rows but also the columns. Such a concept, called real time data, is not new, but for the revisions of the national accounts corresponding compilations are only available in rudimentary form. ³²

Deflation

For a long time, the usual procedure for presenting the real development was to refer to a specific base year. All values are then expressed, for example, „in 1991 prices. Particularly for longer-term comparisons, the problem arises here that not only the general price level changes over time, but also the price structure, i.e. the relationship of the prices of individual goods to one another. For this reason, the Federal Statistical Office has repeatedly shifted the base year to a later year in order to be able to use more up-to-date price structures as a basis. This was accompanied by a number of problems. For example, the serious change in the decline in real GDP in 1974 was largely due to respective changes in the base year.³³ It was mainly for this reason that the Federal Statistical Office switched in 2005 to deflating not in relation to a fixed base year but in relation to the respective previous-year prices. This approach has led to intense controversy³⁴ and has occasionally been called „nonsense“ in academic circles, especially when using longer time series rather than just previous-year comparisons.³⁵ The biggest problem is seen as the fact that with this calculation method the individual components no longer add up to the total aggregate and thus neither volume shares nor growth contributions can be meaningfully calculated. Consequently, the Federal Statistical Office therefore publishes these deflated values only dimensionless in the form of indices with two different „base years“ (1971-1991: 1991=100, from 1991: currently 2015=100). Such values are of course contrary to the need for „actual“ real quantities or absolute values.

In their publications, the German economic research institutes participating in the so-called Joint Economic Forecast have therefore switched to calculating absolute figures from the index values of the Federal Statistical Office. These undoubtedly have the advantage of greater clarity, but also make the above-mentioned problem more visible.³⁶ In any case, it is to be feared that the „real“ values calculated in this way easily lead to false conclusions and interpretations. The bigger problem, however, is that long- and medium-term economic developments on the basis of price-adjusted series are „noticeably different“ from those in current prices, and a consideration of these differences on the basis of chained prior-year price series is at least considerably more difficult.³⁷

A second, related problem is how to account for changes in the price structure. The Federal Statistical Office follows an approach according to which the purchasing power of money is not calculated as a uniform value, since prices for different goods develop differently (=> ch. 14). For this reason, it does not calculate a generally valid price index for deflating national accounts, but several for specific groups of goods.³⁸For gross domestic product and gross value added, a so-called double deflation is applied, since these two quantities do not result directly but as the difference between two others (gross value added: difference between production value and intermediate consumption, gross domestic product: difference between final demand and imports). This procedure of using price-adjusted series is also not uncontroversial among experts. On the one hand, in extreme cases, not only quantitative but also qualitative differences can arise. For example, the so-called net exports, i.e., the difference between exports and imports in the use account, increased sharply in nominal terms in 1986 compared with the previous year, but declined in constant 1991 prices. In 1980, 1991 and 1992, it was negative in nominal terms, but positive at constant prices. In its 1984/85 annual report, the German Council of Economic Experts (Sachverständigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung) commented at length on the question „What is the real national product?“³⁹ In its view, misunderstandings and errors of interpretation frequently occur here. In particular, the approach of the Federal Statistical Office is not helpful for many analyses, since the „real“ social product is not about the construction of a constant purchasing power, but about the elimination of a general depreciation of money. It would therefore make more sense to use a uniform price index for all nominal variables, thus inflation-adjusted series or a constant price level.⁴⁰ Here, too, it is not just a matter of playing with ideas: for example, cyclical upswings and downswings sometimes show up in a different form, depending on whether one uses „real values“ or constant prices as a basis. The Federal Statistical Office has not followed this line of reasoning and has stuck to calculating differentiated deflations.

Effects of the choice of base year

Instead of stating the real values in prices of the respective previous year as a chain index, as practiced by the Federal Statistical Office since 2005, real series to the base year 1991 are given here for the period from 1970 to 2019, i.e. the respective series of the Federal Statistical Office given as a chain index were given here in Tds. or billion euros to the base of the year 1991.

Between 1950 and 2019, prices have increased more than sixfold in the context of the overall economy. In the process, the structure has also changed: While, in terms of the four aggregates of gross value added shown here, prices in the agricultural and trade sectors both increased about fourfold, prices in the manufacturing sector increased almost fivefold, and prices in the services sector increased more than twelvefold. A similar spread can be seen on the use side: here, prices of the import sector increased by about double, those of the export sector by three times, those of the private sector by five times, those of capital goods by almost eight times and those of government consumption by more than eleven times over the past seventy years.

However, the growth rates of the increases also changed differently. While price developments in the agricultural sector were largely linear, price increases in manufacturing and trade slowed down from the 1990s onward. The same applies to the services sector, which until then had risen more strongly than the other three sectors. The situation is somewhat different for price developments on the use side. Here, too, the price developments of the individual aggregates have flattened out since the 1990s, but continued to diverge thereafter.

This divergence has implications for the interpretation of the structural development of the real side of the economy if one wishes to express it in prices of a base year. The following charts illustrate these effects. For the aggregates of the output and expenditure side, they show for three selected base years (1961, 1991 and 2015) in each case the nominal development of the aggregates, the price developments related to the base year, the resulting real developments and finally the resulting shares in or ratios to total gross value added or gross domestic product. If 1961 is chosen as the base year, the real / price-adjusted increase on the creation side has been much stronger for production since 1950 than for trade or services. These two have been diverging only since the 1990s, and the value of services is increasingly greater than that of trade. If, on the other hand, 1991 is chosen as the base year, production and services show roughly the same value up to about 1990, while trade shows a much smaller development in value. A broadly similar picture emerges if 2015 is chosen as the base year.

If we look at the percentage development,⁴¹ so we see that, with a base year of 1961, the share of services does not exceed that of production until the turn of the millennium, while the shares of trade and services roughly balance each other out until around 1990. By contrast, the development is quite different if 1991 is chosen as the base year. Here, the services sector already exceeds that of production in 1980. Moreover, its share is significantly higher over the entire period than when 1961 is chosen as the base year. The reason for this is that the prices of the services sector rise disproportionately until 1990. By contrast, if 2015 is chosen as the base year, there are hardly any differences compared with the base year 1991, which is due to the fact that price developments have largely converged since 1990.

Effects of different base years (production side)

The use side shows different effects. In nominal terms, private consumption rose earlier and more strongly than the other aggregates, investment and government consumption developed largely at the same level, as did imports and exports until the early 1990s; only then did these two rise disproportionately in the wake of globalization. Unlike the value added components of the output side, however, the prices of the aggregates of the output side continue to diverge after 1990. As a result, if 1961 is chosen as the base year, imports and exports develop in practically the same way in value terms over the entire period after 1950 up to the present – in 1961 prices -, diverge from the value of investment and government consumption as early as around 1970 and exceed the value of private consumption shortly after the turn of the millennium. In 1991 prices, on the other hand, this „increase in value“ does not occur until twenty years later and the surpassing of private consumption is also delayed by about ten years. If, on the other hand, 2015 is chosen as the base year, it is not until around the turn of the millennium that the value of imports and exports starts to rise above that of investment and government consumption, and the value of private consumption has not been reached until today.

Effects of different base years (consumption side)

Effects of the individual revisions

We are not aware of a systematic assessment of the impact of individual revisions for economic history since 1950.⁴²
In the following, the main quantitative effects in particular are listed in key words, first for the last five respective current revisions, then for the two revisions in each case, which concern the period 1970 to 1990 and before 1970, respectively. For more detailed discussion, please refer to the corresponding articles in Economics and Statistics listed in the table above. It should be noted in the data that all impacts are always described only in comparison with the immediately preceding revision. Cumulatively, therefore, the effects can (often) be considerably larger, or (less often) cancel each other out.

Revisions for the period from 1991 (in each case up to the current margin)

Revision 1999

Conversion to EVST 95
significant conceptual enhancements in the investment calculations
Input ratios of value added of economic sectors change considerably
serious changes in the calculation of gross domestic product due to the introduction of the ESA 1995 in combination with the new classification of economic activities 1993
a number of changes with both positive and negative effects on the level of gross domestic product as a whole
Flatter time course but at a higher level
Average growth 1991-1998 3.7 instead of 4 percent
Household consumption expenditure increases by 0.4 (1998) to 2.1 percent (1991) per year
Government consumption expenditure increases by 1.2 to 1.7 percent per year
Gross capital expenditures increase 6 to 7 percent per year from 1991 to 1996
in roughly the same amount, imports and exports also
national income is 0.5 percent higher as a result of the revision in 1998, and 2.5 percent higher in 1991

Revision 2005

Conversion to chain indices
Level of gross domestic product increased by the revision-related changes in all years since 1991; GDP in current prices proven to be about 30 to 47 billion euros higher than before (+1.4 and +2.6 percent higher than before)
Change in the calculation methods of deflators for certain groups of goods or sectors of the economy.
Only minor changes due to new classification of economic activities (WZ 2003)
in the „Agriculture, forestry, fishing“ sector, gross value added in 2000 was 1.60 billion euros higher at current prices; price-adjusted development of gross value added in this economic sector much flatter: increase between 1991 and 2004 revised by 4.9 percent, previously 23.3 percent
In the manufacturing sector excluding construction, data-related changes in 2000 resulted in an increase in gross value added of 8.16 billion euros at current prices; price-adjusted trend much flatter, increase between 1991 and 2004 revised by 7.2 percent, previously 13.1 percent
in trade, hotels and restaurants, and transport, gross value added in 2000 was 7.97 billion euros higher at current prices; price-adjusted trend much flatter, increase between 1991 and 2004 revised by 24.8 percent, previously 33.2 percent
for public and private service providers, a 19.03 billion euro increase in gross value added at current prices in 2000; price-adjusted trend much steeper; increase in 1991 and 2004 revised by 26.1 percent, previously 19.1 percent
Final consumption expenditure by households and nonprofit institutions serving households (NPISHs) in current prices is reported to be between 1.5 percent and 4.6 percent higher in 1991 and 2004 than previously

Revision 2011

Conversion to WZ 2008
Agriculture, forestry, fisheries for 2008: instead of 19.96 billion euros, 20.87 billion euros (change of +4.6 percent), in the other years under
Gross capital expenditures will increase by 4.4 percent in 2008 as a result of the revision.
net exports decrease by 3.3 percent
Otherwise effects at the level considered here rather low

Revision 2014

Implementation ESA 2010
Gross domestic product higher by about 3 to 3.5 percent per year from 1991 to 2013
gross value added at current prices decreases by 3.6 percent for agriculture, forestry and fishing and increases by 2.9 percent for manufacturing and 4.1 percent for services in 2010 due to revisions.
private consumption expenditure increases by 0.7 percent, government consumption expenditure by 1.4 percent, gross capital formation by 16.3 percent and net exports by 4.7 percent in 2010. Roughly the same orders of magnitude apply to the remaining years from 1991 onward.
government revenues as well as expenditures are shown to be higher for all years from 1991 onward, but this increase was lower than the increase in GDP, so the government-to-GDP ratio became lower than previously reported (but only by 0.1 to 0.9 percentage points).

Revision 2019

Predominantly data-related changes
Previous cyclical trend largely maintained; level of GDP in current prices on average 0.4 percentage points lower (1.2 lower in 2018, 0.4 percent higher in 1991)
lower private consumption expenditure (-25.8 billion euros), downward revision of 1.5 percentage points compared with the previous result. Government consumption expenditure 0.7 percentage points (4.5 billion euros) higher than before the revision.
serious changes in the components of the external contribution: 6.7 percentage points (16.7 billion euros) lower, due to
lower exports of – 0.5 percentage points (- 7.8 billion euros) and
Higher reported imports of + 0.7 percentage points (+ 8.9 billion euros)
Revision of deflators from input-output accounting: real growth rates of price-adjusted results (unlike in previous years of the time series) were revised significantly upward.

Changes 2024

The next major revision is currently scheduled for 2024. This will primarily involve adjustments resulting from the planned update of the ESA 2010, which will in turn follow a planned revision of the United Nations System of National Accounts. It may also be possible to take into account an update of economic classifications at the international level.

Revisions for periods further back in time

Revision 1972 (1960-1970)

Gross national product at current prices 1960 to 1968 between 1.6 and 2.1 greater than previously shown
for 1962 growth of 8.3 instead of 8.7 percent, for 1966 of 6.6 instead of 6.2 percent
Value added: considerable increase in the contribution of goods-producing industries, correction by 3.3 percent (1962) to 6.2 percent (1967, 1968)
Contribution of service companies reduced by 1.9 percent (1961) to 7.5 percent (1968)
In the case of income from entrepreneurial activity and property, the deviations from the new to the old calculation are the largest in percentage terms of all the variables in the calculation of social product and amount to between 2.0 percent (1962) and 6.4 percent (1966)
Household income increases 3.5 percent (1962) to 5.0 percent (1966) 1960 to 1968

Data for the period from 1950 to 1970 were last revised in 1985 in the course of the major revision (as of 2009):

Revision 1985 (1960-1984)

relatively small corrections for the gross national product as a whole.
in the gross value added at market prices of agriculture and forestry changes from +1.8 percent (1976) to -5.8 percent (1984)

Revision 1985 (1950-1960)

Total social product only minor changes
Variations in gross value added in services between +12.4 percent (1950) and +19.4 percent (1960)
on the expenditure side: government consumption deviations between 1950 and 1955 between -21.2 and -28.6 percent per year
Exports 1950 to 1955 Variations between +10 and +33 percent

Revision 1991 (1970-1990)

Only very slight deviations overall
for gross value added in trade and transport between 1980 and 1990 annual deviations due to revision ds. about 5-6 percent
There are hardly any deviations in the use calculation
Values 1950 to 1969 identical to publication 1985

Revision 2002 (1970-1991)

Conversion to EVST 95
Total gross domestic product 1970 to 1990: divergent trends both upward and downward, consistently at higher levels (+1.9 percent 1970 to 2.8 percent 1990)
In the revised usage aggregates, in some cases significantly different time patterns than previously
Private consumer spending consistently higher than previously evidenced
over the entire revision period, government final consumption expenditure increases by 12 percentage points more than before in current prices; 9 percentage points in constant prices
Level of gross investment (one of the focal points of the revision by ESA 1995) increases by 5.6 percent in 1970, by 6.7 percent in 1990.
in net exports: +13.8 percent in 1970 due to data, -5.8 percent due to concept, 8 percent in total, and 37.6 percent in 1980.
Depreciation increased substantially (1990: 15.7 percent), mainly due to conceptual changes in ESA 1995
1980-1990 Compensation of employees increases more strongly and corporate income less strongly than before

Revision 2006 (1970-1991)

Level increase in nominal GDP through allocation of financial intermediation services (FISIM) among users.
partly noticeable changes in the overall trend, with differences between the rates of change ranging from – 0.6 to +0.8 percentage points.
Value added: Agriculture 11 percent lower in 1990 due to revision
Usage: private consumer spending consistently higher than previously recorded, up by 24.88 billion euros or 3.3 percent in 1991
From 1970 onward, the deviations of government final consumption expenditure in current prices from the pre-revision level fluctuate significantly due to FISIM.
Gross fixed capital formation only marginally revised
Compensation of employees consistently slightly higher than before
Corporate and property income also increase

About the reconstructions for the GDR

The national accounts of the Federal Republic and the GDR corresponded to completely different systems. The latter was based on the „Material Product System“ (MPS), which was oriented on K. Marx and did not correspond to the SNA concept of the Federal Republic in many areas. For example, in the calculation of output, intermediate consumption and depreciation were determined in a completely different way than in the West. In particular, the output of government institutions and social organizations, the credit, insurance, housing, education and health care systems, and direct consumer services, i.e., large parts of public and private services, were not included.⁴³ If corresponding figures have nevertheless been included here, then with the caveat that this missing share, which is certainly not constant in its development over time, must be taken into account when making comparisons. A back-calculation of social product data in the national accounts system for the years 1970, 1972 and 1975 as well as 1978 to 1989 was attempted within the framework of a DFG research project by the Halle Institute for Economic Research and the German Institute for Economic Research on the basis of archive material. Neither the price problem nor the exchange rate problem could be solved, so all data are given in current prices in marks (East).

– Federal Statistical Office (ed.), Sonderreihe mit Beiträgen für das Gebiet der ehemaligen DDR. Heft 33: Entstehung und Verwendung des Bruttoinlandsprodukts 1970 bis 1989. Ergebnis eines von der Deutschen Forschungsgemeinschaft (DFG) geförderten Forschungsvorhabens, Wiesbaden 2000.

In addition, Gerhard Heske continued the work of Udo Ludwig and Rainer Stäglin and calculated the development of East German gross domestic product in deutschmarks for the 1970s and 1980s on the basis of internal documents of the GDR statistics:

– Heske, Gerhard, Gross Domestic Product, Consumption and Employment in East Germany 1970-2000. New Results from National Accounts (= Historical Social Research Supplement 17), Cologne 2005.

– Heske, Gerhard, Volkswirtschaftliche Gesamtrechnung DDR 1950-1989: Daten, Methoden, Vergleiche (= Historical Social Research Supplement 21), Cologne 2009.

Reconstructions for the period before 1950

No national accounts in the modern sense are available for the period before the Second World War. If one nevertheless wants to look further back, corresponding data must be reconstructed. On the part of official statistics, no retrospective reconstructions have been carried out apart from the earmarked calculations corresponding to the concepts of the time and an early and limited exception ⁴⁴. Instead, since the mid-1950s, the economist Walther G. Hoffmann and several collaborators undertook the effort and, in many years of work, compiled a comprehensive national accounts in long series, which was published in 1965.⁴⁵ Hofmann not only reconstructed the social product via national income back to the mid-19th century on the basis of tax statistics, but also attempted, with great effort, an estimate of the use side, combining numerous consumption statistics with further data. He also attempted an extrapolation of the generation side. This was not possible in a direct way due to the lack of differentiated production statistics. To do this, he had to take a roundabout route: he started from the level of national income and calculated the shares of the individual sectors via wage totals in various industries, but the agricultural sector in particular led to implausibly high values. In the end, however, both the calculation via the use side and the generation side could only make approximate assumptions for important components such as investments or depreciation. A systematic recalculation since the beginning of the 20th century, largely based on the official calculations of the Statistisches Reichamt on national income, came to the conclusion that the level of gross national product before and between the two world wars had been overestimated by Hoffmann.⁴⁶ For the second half of the 19th century, new estimates of industrial production, investment and domestic and foreign capital income have meanwhile been produced, which show higher economic activity and thus lower average growth rates for this period.⁴⁷ Complementing this, more recent calculations on land rents have meanwhile been able to provide more reliable projections of the income side of net national product, which also show a higher level since the mid-19th century and thus lower growth rates.⁴⁸

Literature

Adler, Walther / Hauf, Stefan / Schäfer, Dieter, Gross Domestic Product 2018 and Investments in Germany, in: Economy and Statistics 1 / 2019, pp. 89-115.
Barro, Robert J. / Ursúa, José F. (2008) „Macroeconomic Crises since 1870,“ Brookings Papers on Economic Activity, Spring, pp. 255-335.
Federal Ministry of Labor and Social Affairs, Statistical Overviews of Social Policy in Germany since 1945 (volume SBZ/DDR), authored by André Steiner with the assistance of Matthias Judt and Thomas Reichel, Bonn 2006.
Brümmerhoff, Dieter / Grömling, Michael, Are National Accounts Revisions Harmful for Historical Comparisons?, in: World Economics 13/4 (2012), pp. 79-97.
Federal Ministry of Labor and Social Affairs, Statistical Overviews of Social Policy in Germany since 1945 (volume SBZ/DDR), authored by André Steiner with the assistance of Matthias Judt and Thomas Reichel, Bonn 2006.
Burhop, C. / Wolff, G., A compromise estimate of German net national product, 1851-1913 and its implications for growth and business cycles, in: Journal of Economic History 65/3 (2005), pp. 613-657.
Coyle, Diane: GDP: A Brief but Affectionate History. Princeton et al. 2014.
Croushore, D. / Stark, T., A Real-Time Data Set for Macroeconomists, in Journal of Econometrics 105 (2001), pp. 111-130.
German Bundestag, 10th legislative period, printed matter 10/2541 of Nov. 30, 1984.
German Bundestag, 10th legislative period, printed matter 10/4295 of Nov. 22, 1985.
Fürst, Gerhard, Wandlungen im Programm und in den Aufgaben der amtlichen Statistik in den letzten 100 Jahren, in: Statistisches Bundesamt (ed.): Bevölkerung und Wirtschaft 1872-1972, Stuttgart/Mainz 1972, pp. 13 – 83.
Grömling, Michael, Sinn und Unsinn von Quoten auf Basis von preisbereinigten Werten, in: Allgemeines Statistisches Archiv 89 (2005), pp. 451-468.
Grömling, Michael, Entwicklung der makroökonomischen Einkommensverteilung in Deutschland, in: Vierteljahresschrift zur empirischen Wirtschaftsforschung 44/1 (2017), pp. 77-98.
Grömling, Michael, Measuring Modern Business Investment. A Case Study for Germany, in: World Economics 21/1 (2020), pp. 39-64.
Hoffmann, Walter G., Das Wachstum der deutschen Wirtschaft seit der Mitte des 19. Jahrhundert. Berlin/Heidelberg/New York: Springer 1965.
Jordà, Òscar / Schularick, Moritz / Taylor, Alan M., Macrofinancial History and the New Business Cycle Facts, in NBER Macroeconomics Annual 2016, volume 31, edited by Martin Eichenbaum and Jonathan A. Parker. Chicago 2017.
Knetsch, T.A. / Reimers, H.-E., Dealing with benchmark revisions in real-time data: The case of German production and orders statistics, in: Oxford Bulletin of Economics and Statistics 71/2 (2009), pp. 209-235.
Krug, Walter / Nourney, Martin / Schmidt, Jürgen, Wirtschafts- und Sozialstatistik: Gewinnung von Daten, 5th ed.
Leifer, Hans-Albert / Tennagels, Peter, Preisbereinigtes Bruttoinlandsprodukt: Publikationspraxis im In- und Ausland, in: Wirtschaftsdienst 3/2008, pp. 203-207.
Lepenies, Philipp: The Power of One Number. A Political History of the Gross Domestic Product. Berlin 2013.
Lippe, Peter von der, Kritik internationaler Empfehlungen zur Indexformel für Preisindizes in der amtlichen Statistik. Some Remarks on the Economic Theory of Index Numbers and on Chain Indices, in: Jahrbücher für Nationalökonomie und Statistik 218/3+4 (1999), pp. 385-414.
Lippe, Peter von der, Der Unsinn von Kettenindizes, in: Allgemeines Statistisches Archiv 84 (2000), pp. 67-82.
Lippe, Peter von der, Zur Interpretation der ,,Vorjahrespreismethode“ der Deflationierung, in: AStA Wirtschafts- und Sozialstatistisches Archiv 1 (2007) , pp. 61-71.
Ludwig, Udo, Die gesamtwirtschaftliche Entwicklung der SBZ/DDR (1949 bis 1990) – eine Bilanz, in: Günther Heydemann / Karl-Heinz Paqué (eds.), Planwirtschaft – Privatisierung – Marktwirtschaft. Wirtschaftsordnung und -entwicklung in der SBZ/DDR und den neuen Bundesländern 1945-1994 (= Schriften des Hannah-Arendt-Instituts für Totalitarismusforschung 63), Göttingen 2017, pp. 111-147.
Lützel, Heinrich, Entwicklung des Sozialprodukts 1950 bis 1984, in: Wirtschaft und Statistik 6/1985, pp. 433-444.
Maddison, Angus, Dynamic Forces in Capitalist Development. A Long-Run Comparative View, Oxford 1991.
Maddison, Angus, Monitoring the World Economy, 1820-1992, Paris 1995.
Morgenstern, Oskar, Über die Genauigkeit wirtschaftlicher Beobachtungen, 2nd ed.
Mucha, Tanja, Volkswirtschaftliche Gesamtrechnungen, in: Statistisches Bundesamt / Wissenschaftszentrum Berlin, Datenreport 2018. Ein Sozialbericht für die Bundesrepublik Deutschland, Bonn 2018, pp. 129-137.
Neubauer, W., Deflationierung, in: Brümmerhoff, D. / Lützel, H. (Hgg.), Lexikon der Volkswirtschaftlichen Gesamtrechnungen, 3rd ed.
Nierhaus, Wolfgang, Zur gesamtwirtschaftlichen Preisentwicklung, in: ifo Schnelldienst 56/3 (2003), pp. 37-38.
Nierhaus, Wolfgang, Preisbereinigtes Bruttoinlandsprodukt: Zur Veröffentlichungspraxis im Gemeinschaftsgutachten, in: ifo Schnelldienst 61/9 (2008), pp. 15-18.
Pfister, Ulrich, The Crafts-Harley view of German industrialization: an independent estimate of the income side of net national product, 1851-1913, in: European Review of Economic History 24/3 (2020), pp. 502-521.
Pfister, Ulrich, Die 1870er Jahre als Strukturbruch der langfristigen wirtschaftlichen Entwicklung Deutschlands, in: Ulrich Pfister / Jan-Otmar Hesse / Mark Spoerer / Nikolaus Wolf (eds.), Deutschland 1871: Die Nationalstaatsbildung und der Weg in die moderne Wirtschaft, Tübingen 2021 (= The Unity of the Social Sciences in the 21st Century 6), pp. 97-118.
Piketty, Thomas / Zucman, Gabriel, Capital is Back: Wealth-Income Ratios in Rich Countries 1700-2010, in The Quarterly Journal of Economics, Volume 129, Issue 3, August 2014, Pages 1255-1310.
Quaas, Georg, Real variables and price indices in the old and new national accounts systems, MPRA Paper No. 22316, https://mpra.ub.uni-muenchen.de/22316/.
Raabe, Karl-Heinz, Die langfristige Entwicklung des Sozialprodukts im Bundesgebiet, in: Wirtschaft und Statistik 1954/2, pp. 63-66.
Räth, Norbert, Recessions in Historical Perspective, in: Economy and Statistics 3/2009, pp. 203-208.
Räth, Norbert / Struck, Bernd / Voy, Klaus, Zur Geschichte der Deflationierung in den Volkswirtschaftlichen Gesamtrechnungen, in: Voy, Klaus (ed.), Kategorien der Volkswirtschaftlichen Gesamtrechnungen. Vol. 4. On the History of National Accounts after 1945, Marburg 2009, pp. 209-230.
Rinne, Horst, Wirtschafts- und Bevölkerungsstatistik. Erläuterungen – Erhebungen – Ergebnisse, 2nd ed.
Ritschl, Albrecht, Wirtschaftliche Folgen des Ersten Weltkriegs, in: Marcel Boldorf (ed.), Deutsche Wirtschaft im Ersten Weltkrieg (= Handbücher zur Wirtschaftsgeschichte), Berlin 2020, pp. 601-617.
Ritschl, Albrecht/Spoerer, Mark, Das Bruttosozialprodukt Deutschlands nach den amtlichen Volkseinkommens- und Sozialproduktstatistiken 1901-1995, in: Jahrbuch für Wirtschaftsgeschichte 1997, pp. 27-54.
Speich Chassé, Daniel: The Invention of the Gross National Product. Global Inequality in the Knowledge History of Economics (Critical Studies in the Study of History 212). Göttingen et al. 2013.
Stahmer, Carsten, Organizational New Beginnings and First Calculations. Early History of National Accounts in West Germany, in: Economy and Statistics 2/2010, pp. 179-195.
Federal Statistical Office, Strategy and Program Plan. For the years 2017-2021, o.O. 2017.
Federal Statistical Office (ed.) 2010, National Accounts. Results in current prices and in 1991 prices according to ESA, 2nd edition, 1950 to 1970, Wiesbaden (unpublished) (Download).
Federal Statistical Office (ed.), 2021, Fachserie 18, National Accounts. Series 1.5: Domestic Product Calculation – Long Series from 1970 onward, vol. 2020, Stuttgart. (Download)
Federal Statistical Office (ed.), 1991, Fachserie 18, Volkswirtschaftliche Gesamtrechnungen. Series S.15: Revised Results 1950 to 1990, Stuttgart/Mainz: Kohlhammer.
Federal Statistical Office (ed.), 2005, Fachserie 18, Volkswirtschaftliche Gesamtrechnungen. Series S.21: Revised Results 1970 to 2004, Stuttgart/Mainz: Kohlhammer.
Federal Statistical Office, Fachserie 18, Reihe S.32, National Accounts. Methods of price and volume measurement, o.O. 2017.
Federal Statistical Office (ed.), 1985, Fachserie 18, Volkswirtschaftliche Gesamtrechnungen. Series S.7: Long Series 1950 to 1984, Stuttgart/Mainz: Kohlhammer.
Voy, Klaus (ed.), On the History of National Accounts after 1945. Volume 4: Categories of National Accounts. Marburg (Metropolis), 2009.

Notes

The article benefits from a lively exchange with Ulrich Pfister and Mark Spoerer. Any remaining errors are the responsibility of the author. I would also like to thank Rainer Fremdling and Nikolaus Wolf for further valuable comments.
Philipp Lepenies: The Power of One Number. A Political History of the Gross Domestic Product, Berlin 2013.
Lepenies (note 2); Daniel Speich Chassé: Die Erfindung des Bruttosozialprodukts. Globale Ungleichheit in der Wissensgeschichte der Ökonomie (Kritische Studien zur Geschichtswissenschaft 212), Göttingen et al. 2013; Diane Coyle: GDP: A Brief but Affectionate History, Princeton et al. 2014.
For the Federal Republic: Klaus Voy (ed.): Zur Geschichte der Volkswirtschaftlichen Gesamtrechnungen nach 1945. vol. 4: Kategorien der Volkswirtschaftlichen Gesamtrechnungen, Marburg 2009.
As an overview: Tanja Mucha: Volkswirtschaftliche Gesamtrechnungen, in: Statistisches Bundesamt / Wissenschaftszentrum Berlin, Datenreport 2018. Ein Sozialbericht für die Bundesrepublik Deutschland, Bonn 2018, pp. 129-137.
According to this method, smaller incomes and incomes of the agricultural sector could only be recorded inadequately and had to be supplemented by estimates. Cf. Gerhard Fürst: Wandlungen im Programm und in den Aufgaben der amtlichen Statistik in den letzten 100 Jahren, in: Statistisches Bundesamt (ed.): Bevölkerung und Wirtschaft 1872-1972, Stuttgart/Mainz 1972, p. 40.
Thus, the major revision in 1999 is an implementation of the revision of the European System of Accounts (ESA) of 1995, which in turn is closely based on the revision of the United Nations System of National Accounts (SNA) of 1993. Its new edition in 2008, in turn, was adopted as ESA 2010 by the European Union in 2011 and implemented retroactively by the Federal Statistical Office in 2014. The compilation of national accounts according to this system is mandatory for all EU member states.
For a detailed overview cf. the documentation.
DIE WELT, Aug. 13, 2014; Tagesspiegel, Aug. 28, 2019.
Cf. Ulrich Pfister: Die 1870er Jahre als Strukturbruch der langfristigen wirtschaftlichen Entwicklung Deutschlands, in: Ulrich Pfister / Jan-Otmar Hesse / Mark Spoerer / Nikolaus Wolf (eds.), Deutschland 1871: Die Nationalstaatsbildung und der Weg in die moderne Wirtschaft, Tübingen 2021 (= Die Einheit der Gesellschaftswissenschaften im 21. Jahrhundert 6), pp. 97-118.
Cf. Albrecht Ritschl: Wirtschaftliche Folgen des Ersten Weltkriegs, in: Marcel Boldorf (ed.), Deutsche Wirtschaft im Ersten Weltkrieg (= Handbücher zur Wirtschaftsgeschichte), Berlin 2020, p. 612.
The problem of price adjustment is explained in more detail in the documentation.
Cf. Heinrich Lützel: Entwicklung des Sozialprodukts 1950 bis 1984, in: Wirtschaft und Statistik 6/1985, p. 439. Lützel points out, however, that due to the problems of measuring prices in the government sector, the figures are subject to a certain degree of uncertainty.
Pfister (note 11), p. 109, after Lampe/Wolf in this volume.
When calculating, it should be borne in mind that only the difference between exports and imports is to be taken into account as a component for the gross domestic product.
According to Walther Adler / Stefan Hauf / Dieter Schäfer: Gross Domestic Product 2018 and Investment in Germany, in: Economy and Statistics 1 / 2019, p. 98, 45 percent of gross fixed capital formation in 2018 was equipment, including military weapons systems and research and development.
Cf. Adler / Hauf / Schäfer (Note 17), p. 111.
Michael Grömling: Entwicklung der makroökonomischen Einkommensverteilung in Deutschland, in: Vierteljahresschrift zur empirischen Wirtschaftsforschung 44/1 (2017), pp. 77-98. However, this statistic should also be viewed with caution, as it only says something about the sums and not the numbers of individuals. For example, since 1950, many self-employed persons have switched to dependent employment, so that the sum of employee compensation was distributed among correspondingly more heads and that of corporate income among correspondingly fewer heads. Cf. Lützel (note 14), p. 441. It should also be noted that in many studies on the long-term development of capital and wealth, the capital share does not take into account the income of the self-employed but, unlike in the national accounts, accounts for this wholly or partly in labor income, e.g. Erik Bengtsson / Daniel Waldenström: Capital Shares and Income Inequality: Evidence from the Long Run, in: Journal of Economic History 78/3 (2018), pp. 712-743.
Cf. Thieß Petersen: Measuring Welfare: Domestic Product versus National Income, in: Wirtschaftsdienst 99/10 (2019), pp. 725-730.
At the county level, the GDP per capita of the bottom quarter of all counties is only 40 percent of that of the top quarter. Data: Arbeitskreis Volkswirtschaftliche Gesamtrechnung der Länder.
Nikolaus Wolf: Regional economic growth in Germany, in: Joan Ramón Rosés, Nikolaus Wolf (eds.), The Economic Development of Europe’s Regions. A Quantitative History since 1900, London: Routledge 2019, p. 160.
See Jaap Sleifer, Planning Ahead and Falling Behind: The East German Economy in Comparison with West Germany, 1936-2002, Berlin 2006.
In this context, however, the different calculation concepts of the two systems must be taken into account, in which the service sector in particular was viewed differently. Cf. Udo Ludwig: Die gesamtwirtschaftliche Entwicklung der SBZ/DDR (1949 bis 1990) – eine Bilanz, in: Günther Heydemann / Karl-Heinz Paqué (eds.), Planwirtschaft – Privatisierung – Marktwirtschaft. Wirtschaftsordnung und -entwicklung in der SBZ/DDR und den neuen Bundesländern 1945-1994 (= Schriften des Hannah-Arendt-Instituts für Totalitarismusforschung 63), Göttingen 2017, p. 146.
Cf. Ludwig (Note 24), p. 147.
Federal Ministry for Economic Affairs and Energy (ed.): Annual Report of the Federal Government on the State of German Unity 2018, Munich 2018, p. 16.
For example, with the introduction of the „European System of Accounts“ (ESA) in 2010, illegal activities such as domestic drug production were included for the first time in the 2014 German calculation, which meant that the Federal Statistical Office then also had to „value drug quantities produced in Germany at street prices. Cf. Federal Statistical Office, Strategy and Program Plan. For the years 2017-2021, op. cit. 2017, p. 25. However, such illegal activities have of course not only existed since 2014.
Cf. e.g. the arguments of Joseph E. Stiglitz / Amartya Sen / Jean-Paul Fitoussi: Report by the Commission on the Measurement of Economic Performance and Social Progress, op. cit. 2009.
1950, repeated again in 1965 in essence. Oskar Morgenstern: On the Accuracy of Economic Observations, 2nd edition, Vienna / Würzburg 1965, pp. 244-284.
Full details in each case in the data documentation.
Cf. Federal Statistical Office (2017, Strategy and Program Plan).
Cf. Brümmerhoff / Grömling (2012).
In the U.S., such „Real-Time Data Sets for Macroeconomists“ were started to be built up already thirty years ago. Based on the U.S. concept, the Research Center of the Deutsche Bundesbank began in 2004 to establish a corresponding offering for Germany. On the structure of the data and the concept: Croushore, D. / Stark (2001); for Germany Knetsch / Reimers (2009). In the database currently offered on the Internet by the Bundesbank, apart from gross domestic product (from 1995), the series start in 2005 at the earliest.
Cf. Räth (2009, p. 207f).
Räth / Struck / Voy (2009).
Lippe (1999, 2000, 2007), Grömling (2005). Keywords: error propagation, non-additivity, path dependence and no pure price or volume comparison.
Cf. Nierhaus (2008), Leifer / Tennagles (2008, p. 207): „All in all, multiplying the chain index by the reference year values is not very helpful, if not misleading.“ On the other hand, there is also a lack of an alternative.
Grömling (2005, p. 406).
Cf. Krug / Nourney / Schmidt (1999, pp. 388-389), for calculation Nierhaus (2003).
Deutscher Bundestag (1984, pp. 139-143).
Specifically, the „price index for the last domestic use“ is proposed.
due to the non-additivity of the chained indices, the sum of the individual components does not add up to the total series. In fact, however, the deviations are very small (\pm 1-2 percent),
See, however, Brümmerhof / Grömling (2012).
Cf. Federal Ministry (2006, p. XXIIff.)
Raabe (1954).
Hoffmann (1965).
Ritschl/Spoerer (1997). For the period 1970 to 1990, the Ritschl/Spoerer data are not used here because they have been extensively revised in the meantime, mainly due to adjustments to ESA 1995 and ESA 2010 and the new classification of economic activities. For 1950 to 1969, on the other hand, the data used here are identical to Ritschl/Spoerer.
Burhop/Wolff (2005).
Pfister (2020).