1088 Budapest, Rákóczi út 5.; Tel: (36 1) 381 23 47; E-mail: Ez az e-mail-cím a szpemrobotok elleni védelem alatt áll. Megtekintéséhez engedélyeznie kell a JavaScript használatát.
Begegnungen
Schriftenreihe des Europa Institutes Budapest, Band 22:123–136.

SCOTT M. EDDIE

What Size and Kind of Agricultural Units?

Reflections on the Study of Prussian Landownership Before the Great War

 

Introduction

Why are scholars interested in the size of agricultural units? Generally, for two reasons: Their studies arise either from a concern about the distribution of income and/or wealth (and the social or political power that may arise therefrom), or because they wish to investigate production or the efficiency of production. The first reason can be classified as socially-oriented, while the second focuses on more technical issues, and of course the two can overlap: How income is produced and who gets it are obviously a combination that includes both social and technical considerations.

While even the great pioneer of study of the distribution of landed wealth in Prussia, Johannes Conrad, began his series of studies “Agrarian-Statistical Investigations”1 with four technical agronomic articles, his primary concern was with who owned the land. Thus the rest of the articles in the series dealt only with that issue. The popular perception of the Junker landlord dominating Prussia – especially its eastern provinces – led Conrad, and leads us, to concentrate on who owned the land.

The purpose of this essay is to demonstrate that, in studies of the ownership of landed wealth, the choice of what data to use and analyze is critically important to the outcome of those studies. In particular, conclusions about how concentrated (or how unequal) the distribution of land ownership was depend far more on the choice of which data to analyze than on the method used to analyze those data. Using inappropriate – but readily available, easy to use, and therefore tempting – data can introduce errors of serious magnitude. Therefore we consider, in what follows, three fundamental questions:

 

1. Which of several possible concepts of the unit of analysis is appropriate to the given research question(s)?

2. In what units should that concept be measured?

3. How close do available statistics come to satisfying the criteria established by the first two questions?

Because we are concerned with studies of Prussian land ownership, we will confine our attention to the concepts used in that literature, in each case providing the German term for the translations here used.

Several potential concepts offer themselves for our consideration for use in the analysis of land ownership:

 

1. The farm, or operating unit of agriculture (Betrieb).

2. The actual legal unit of land holding, the property (Besitz, or – but only for larger properties – Gut).

3. The manor (Herrschaft), usually made up of several properties.

4. The total amount of land owned by a single owner (Gesamteigentum).

 

In the discussion which follows we will examine the available data for the period from the establishment of the Prussian land tax cadastre in the early 1860s until the First World War, and consider how these have, or could be, used to delineate the structure of land ownership. In the process, we will also consider the general and situational appropriateness of the four concepts listed, and how well the available data fit these concepts.

Available data

That we have to worry about these questions at all is the result of a fundamental lacuna in the Prussian (and German) statistics of the time: Although there were many calls for such data, the government never compiled any statistics of land ownership. In a sense this is surprising, since – as one scholar observed – everything technically needed had been in place for decades, so there were no important hindrances.2 But, since such statistics were never collected, scholars interested in land ownership must direct their attention elsewhere.

For Prussia during the period of our concern, there are four sources of comprehensive data that deal with the size of agricultural units3:

1. The censuses of German agriculture taken in 1882, 1895, and 19074;

2. The reports on private land ownership in Prussia based on the building tax revisions in the years 1878 and 18935;

3. The report of distribution of land6 based on a data from the income and “supplemental”7 tax returns concerning rural indebtedness in Prussia in 1902; and

4. Various directories of land ownership that describe large properties in several provinces.

The choice of which of these sources to use depends first and foremost on the questions to be asked, but also secondarily on the ease of accessibility and breadth of coverage of the statistics contained in these sources. I have listed the sources in decreasing order of their ease of use and breadth of coverage:

 

1. The censuses of agriculture contain all sorts of tables already calculated on the basis of size categories expressed in hectares of agricultural land and include all units that had any land devoted to agriculture, in three different years;

2. The reports on private landownership include some statistics on private landownership – which excluded properties of the State; provinces and lesser government units; communities, churches, and schools; railways; charitable organizations; and economic cooperatives8 – arranged according to size categories for two years;

3. The debt-based statistics show the number of owners, the total number of hectares they owned, and the total amount of land tax net yield for various size categories of that net yield (60-90 Marks, 90-150 Marks, ..., 3000 Marks and up) for the year 1902.

4. The directories of large landed properties typically cover all properties of 100 or more hectares (some later dropped the lower limit to 50 hectares) or having a land tax assessment of 1500 marks or more, for a particular year in a given province; these directories appeared in editions updated typically every 4–7 years,9 and show who owned each property as well as its makeup (how much ploughland, meadow, pasture, woodland, etc.), its tax assessment, whether it had any industrial or mining establishments located on its territory, or whether it specialized in some kind of arable agriculture or animal husbandry. The directories even gave some data on the administration and leasing of the property’s land as well as about its location, transport connections, and even phone numbers.

 

Therefore a scholar wishing to deal comprehensively with land holding or land use in Prussia faces two problems: Which sources to select as most appropriate for the issues he or she is concerned with, and among those, which does he or she have the resources to use? The more difficult it is to extract and analyze the data from the sources, the more time and money it takes to work them up. Hence scholars – especially in the past and especially before the advent of personal computers – often had to compromise: They had to use what they could afford to work up.

Johannes Conrad was nevertheless able, perhaps because he was the editor of the journal where he published his studies but more likely because the university where he taught put considerable resources at his disposal,10 to use the most difficult sources, the directories of land ownership, to find answers to questions that could not be effectively addressed using the data of the other sources. We will return to this point after a discussion of appropriate concepts.

Farm vs. Property

From the various German statistics, we are able to use different kinds of units of agricultural land, measured on different scales:

 

1. The agricultural censuses present data organized according on the basis of the farm (Betrieb), the operating unit of agriculture. One farm could be

a.) one part of a larger estate which is let to and/or operated by different farmers,

b.) congruent with a particular ownership unit (e.g., the proverbial family farm), or

c.) be made up of several properties owned by or leased from the same or different owners (e.g., the big tenant farms of England).

2. The statistics of private ownership categorize and summarize the units according to the amount of land tax net yield (Grundsteuerreinertrag), which was a measure of the amount of net income a given piece of land could be expected to yield in its given location under ordinary management and with local farm-gate prices.11 At the time of the original land tax assessment (even much later, as I have shown in an earlier study12), this net yield could therefore be expected to be a nearly perfect proxy for the market value of the land. There are also some tables of the number of properties (but not their area or tax value) falling into each size class according to total area of the property in hectares. The principal tables are cross-tabulations of the number of properties according to their size in hectares and their assessed total land tax net yield in Thaler13. These surveys counted properties by land tax district (typically the same as the riding [Kreis], but often smaller, sometimes larger), so an owner with land in more than one tax district would be counted more than once.

3. The debt-based statistics, on the other hand, classified ownership parcels according to their total land tax net yield in marks and their total area in hectares. This survey omitted all land holdings with less than 60 marks of land tax net yield, on the ground that 60 Marks represented the minimum economic size of farm that could provide its owner with a living without by-employment. The top category of “3000 Marks and more” does not permit much differentiation among larger land holdings, since in most provinces a property of 200 hectares could easily have this much land tax net yield or more. This survey counted each owner only once within the borders of the entire Kingdom of Prussia,14 so that it comes much closer to being a dataset of total ownership (by private individuals only), whereas the previous statistics of private ownership came much closer to being a dataset of properties, rather than owners. This difference stemmed from the fact that the 1902 survey based itself on the income and earnings tax returns, while the 1878 and 1893 surveys were based on the data of the land tax and the building tax.

4. The directories of large properties permit, but do not themselves present, the classification of land units according to size of holding (Besitz) or size of ownership by an given owner (Gesamteigentum). The measurement units which could be chosen, using these data, are total area in hectares, area of land devoted to agriculture (as in the censuses), or amount of land tax net yield (as in the surveys of private ownership). Thus the directories provide the most flexible source of data, but with two serious limitations: (1) They cover properties (and therefore NOT farms), and only if they exceed a certain minimum area, and (2) they contain no summary data. All results therefore depend on first transferring the raw data into a computer or other device in order to do any summaries, cross-tabulations, or similar calculations. Because each of these directories contains anywhere from about 2500 to over 7500 individual entries with information on several variables, their use has in most cases proved to be cost-prohibitive.

 

Further problems arise with the practical definition of terms. What determines what an operating unit (Betrieb) is? For example, the Bolko von Hochberg estate of Neuschloss in Silesia (6877 hectares in 1909) had a general agricultural administrator, a chief forester, and an administrator for ponds and water resources (each of whom had his own staff), but only one accountant. Is this one, two, or three economic units?15 Moreover, the same owner had another estate elsewhere in Silesia (Rohnstock: 2163 hectares), with no accountant listed for it. If the same accountant did the books there as well, were both estates together just one big economic unit?

What, for that matter, is a property? Is it the manor (Herrschaft) or the individually- named properties that make it up – and which could have been acquired at different times? If the manor was an entail (Fideikommiß) that could not be alienated in part but had to be passed on intact to the next heir when the current land holder died, does that make the definition of what is a property different from the case of a manor which was not an entail, and whose parts could be sold off, inherited, or given away piecemeal?

The least problematic concept is the concept of total ownership (Gesamteigentum). This is the concept which Conrad employed in his pathbreaking series. Whether each individual part of a manor is counted as a property, or whether the manor itself is considered to be a single property, when we add up all the properties owned by a single owner, it should come to the same total. The adding-up process is unaffected by the definition of what is a property. Here there are only marginal problems, save one: Does the current holder of an entailed estate (Fideikommiß) really “own” it? He or she (or in some rare cases, they) would have lifetime use of the land and its income, but the estate must be passed on intact to the designated heir.16 If no such designated heir exists, the terms of the entail usually provide that the land revert to the crown. Moreover, the actual legal owner of the land, as shown in the land registry, is the “foundation” that establishes the entail, not the current beneficiary of that foundation. I do not know of any case in which this fine distinction ever came into play: All studies which I have seen treat the beneficiary of an entail as a real owner in compiling their statistics.

Compared to the problem of entail, the other conceptual problems of total ownership are minor. Foremost among these is how to deal with joint ownership (husband and wife, siblings, parent and child, unrelated individuals in some sort of partnership, etc.) where one of the joint owners is also the single owner of one or more other properties. My own personal solution to this problem is to consider each group of two or more owners as a separate owner, which itself can lead to anomalies: If a husband and wife own property together, and each also owns property separately, then in my schema these two individuals constitute three separate owners. Assigning the rightful share to each joint owner to overcome this anomaly would be a task for which the cost would vastly exceed the benefit, at least among the Prussian data where joint owners are relatively rare. Moreover, how could one assign shares, and to whom, when the number and names of owners are unknown? How many people owned the estate of Granowo in the province of Posen in 1910, whose owners are simply listed as “the heirs of Prince Zdzislaus Czartoryski?” And who, exactly, were they?

Given all these difficulties, it is unsurprising that the first place many scholars turn, when they wish to examine the distribution of ownership of land, is to the very handy summary tables of the agricultural censuses. Without considering any of the definitional problems, many of them make the mistake of confusing “farm” (Betrieb, the operating unit) with “property” (Besitz, the ownership unit), or – what amounts to the same thing – assuming that the farm and the property are one and the same unit. They then use the data of the agricultural censuses as if they were data on properties.

In criticizing the approach of these scholars, it is important to re-emphasize that the German agricultural censuses explicitly counted farms, not properties, and indeed took care to point out that there was a difference between the two concepts. From the introduction to the 1882 census volume17 we can read that

One should never lose sight of [the following]: first, that areas of operation18 are treated here, not areas of ownership; second, that only the agricultural areas of operation, not the total area of the farms, come into the presentation...

Conrad did not make this mistake: He used the agricultural census of 1882 for background data regarding agriculture, but was careful never to mix any of these data on farms in with his data on properties.

 

Empirical issues

The distribution of agrarian land can therefore be considered from at least two fundamentally different standpoints:

 

1. If we are concerned with production, or the economic function of rural land, it makes sense to construct a distribution of land in farms (Betriebe), since the farm is the operating unit for production. Such a distribution might show, for example, that average farm size in North America has increased sharply in the last several decades, with a consequent loss of farm population and concomitant hard times for small towns whose main economic function had been to serve local agriculture.

2. If we are concerned with land as an asset, a component of wealth, then it makes sense to construct a distribution according to ownership19 of the asset. Such a distribution can take two basic forms:

a) A distribution by property (Besitz, or often Gut, in the case of larger properties); this is the legal unit of ownership, for which customarily a deed is issued, and this is the unit that can be most readily sold, mortgaged, exchanged, given or gambled away, etc.

b)A distribution of total ownership (Gesamteigentum) is appropriate when one’s concern is the wealth (or status or influence) that the ownership of land confers upon the owner, and the concentration of that wealth.

Problems arise when data do not exist in the form necessary to make whatever distribution one’s research goal requires. For example, we noted above that the distributions of ownership based on the data in address books of landed properties usually are truncated on the lower end, for these address books normally have a minimum size criterion for inclusion, typically 100 hectares. The researcher must then either be content to present a truncated distribution, or else make some rather heroic assumptions to use other data, such as those from agricultural censuses, to fill in the lower part of the distribution.

Problems also arise, again as noted above, when researchers fail to recognize the difference between a farm and a property, and mistakenly use the data from the German agricultural censuses as distributions of ownership.20 There is, however, no necessary identity between a farm and a property: One property may be let in parcels to several renters, each of whom operates a farm. In such a case one property comprises several farms, or parts of several farms. Or, alternatively, one farmer may rent land from several different owners, so that one farm comprises several properties or parts of properties. There is really no good excuse for confusing farms with properties, since the census data always show how many farms in a given size category had exclusively owned land, how many had a combination of owned and rented land, or how many operated with exclusively leased land. They do not show, however, whether the farms with exclusively leased land were leased from more than one owner; but clearly for any farm that consisted partially of owned land and partially of rented land, the farm and the property could not possibly be one and the same thing.21 As Klaus Hess correctly pointed out: “Because of the method of their collection, the agricultural operational statistics [NB: agricultural census (SME)] in no way convey the character of ownership statistics, contrary to the common opinion.”22

A particularly egregious case of the confusion of farm with property can be found in a recent study by Ilona Buchsteiner,23 who compounded the confusion by inventing a term, Gutsbetrieb (literally “property-farm” or “estate-farm”) which, as a general concept of ownership, is nonsense, for the reasons given in the previous paragraph. The Gut (large individual property) and the Betrieb (farm) can be one and the same thing, as in the case of the family farm which neither rents in nor rents out land, but the concept of Gutsbetrieb is worse than meaningless for any farm that is either part of a larger property, or which is made up of two or more smaller property units. In modern England and North America, for example, the large farming operation consisting of several or even many properties – some owned, some rented – has become the norm for agriculture in many regions. Even in the 19th century there already were large tenants (Grosspächter) whose farm operations comprised land from several different properties, often owned by different persons.

Germany never published a directory of farms based on her agricultural censuses. For an illustration of how dangerous it is to assume that farm = property we can, however, turn to data from that other quintessential land of large estates in Europe, Hungary,24 which did produce a directory of farm operators (Gazdacímtár) from her 1895 agricultural census.25 In the following table, I have extracted a single page from the county of Moson in western Hungary26 which shows the names of farm operators and the names of owners of the land that they farmed.

 

Table 1: Farms and owners in town of Szolnok, Moson county

Farm
Nr.

Name of farm operator

Names of landowners

110

Unger József

Unger József, Ung Mihály, Zwickl András, Czellei János

111

Unger Mária

Unger Mária, Unger Mátyás, Czellei János etc.

112

Unger Mátyás (House Nr. 152/153)

Unger Mátyás, Unger Mátyás dr., Lang József etc.

113

Unger Mátyás (House Nr. 24)

Unger Mátyás, Waldherr József, Czellei János, etc.

114

Unger Mihály (House Nr.13)

Unger Mihály, Zechmeister Mihályné, Czellei János etc.

115

Unger Mihály (House Nr.1)

Unger Mihály, János and Mátyas; Czellei János

116

Unger Pál

Unger Pál, Zwickl Mihály, Czellei János, Town of Szolnok

117

Wachtler János (House Nr.7)

Wachtler János, Town of Szolnok, Grazi Mária

118

Wachtler János (House Nr.154)

Wachtler János, Haiden János, Weisz József etc.

119

Waldherr Mihály

Waldherr Mihály, Haiden János, Town of Szolnok

120

Weinhandl János

Weinhandl János, Boldizsár Erzsébet, Reiter János etc.

121

Weisz János (House Nr.46)

Weisz János, Reiter János, Bobert István, Town of Szolnok

122

Weisz János (House Nr.111)

Weisz János and Márton, Czellei János, Trummer Pál

123

Weisz János (House Nr.21)

Weisz János and József, Czellei József, Trummer Pál

124

Weisz József (House Nr.12)

Weisz József, Trummer Pál, Town of Szolnok

125

Weisz József (House Nr.23)

Weisz József and Anna, Town of Szolnok

126

Weisz József (House Nr.41)

Weisz József, Town of Szolnok

127

Weisz József (House Nr.37)

Weisz József, Town of Szolnok

128

Weisz József (House Nr.142)

Weisz József and János, Town of Szolnok, Bobert István

129

Weisz József (House Nr.103)

Weisz József, Neuberger György, Town of Szolnok

130

Weisz Mátyás (House Nr.105)

Weisz Mátyás, Czellei János, Town of Szolnok

131

Weisz Mátyás (House Nr.9)

Weisz Mátyás, Huber Ferencz, Czellei János etc.

132

Weisz Mátyás (House Nr.42)

Weisz Mátyás, Zechmeister Mihályné, Town of Szolnok

133

Weisz Mátyás (House Nr.47)

Weisz Mátyás, Town of Szolnok

134

Weisz Mátyás (House Nr.11)

Weisz Mátyás, Zwickl András, Weisz József

135

Weisz Mátyás (House Nr.3)

Weisz Mátyás and József, Town of Szolnok

136

Weisz Mihály (House Nr.122)

Weisz Mihály, Hutflesz András, Grazi Mária etc.

137

Weisz Mihály (House Nr.25)

Weisz Mihály, Zwickl András, Rom. Cath. Church, etc.

138

Weisz Pál (House Nr.112)

Weisz Pál, Czellei János, Reiter János, Town of Szolnok

139

Weisz Pál(House Nr.57)

Weisz Pál, Czellei János, Reiter János, Grazi Mária

140

Weisz Teréz

Weisz András, Zwickl András, Czellei János

141

Zechmeister János (House Nr.98)

Zechmeister János, Bobert István, Czellei János etc.

142

Zechmeister János (House Nr.147)

Zechmeister János, Bobert István, Town of Szolnok

143

Zechmeister József (House Nr.52)

Zechmeister József, Town of Szolnok

144

Zechmeister József (House Nr.60)

Zechmeister József, Bobert István, Town of Szolnok

145

Zechmeister Pál (House Nr.15)

Zechmeister Pál, Czellei János, Grazi Mária etc.

146

Zechmeister Mihály

Zechmeister Mátyás, Trummer Pál, Czellei János etc.

147

Zechmeister Pál (House Nr.45)

Zechmeister Pál, Brasch József, Czellei János etc.

148

Zwickl András

Zwickl András, Laszl Mátyás, Town of Szolnok

149

Zwickl János

Zwickl János, Bobert István, Town of Szolnok

150

Zwickl József

Zwickl J., Zwickl J. Jnr., Hutflesz K., Schay B.

151

Zwickl Márton

Zwickl M., Haiden J., Reiter J., Town of Szolnok

152

Zwickl Mátyás (House Nr.5)

Zwickl Mátyás, Town of Szolnok

153

Zwickl Mátyás (House Nr.51)

Zwickl Mátyás, Town of Szolnok, Czellei János

154

Zwickl Mátyás (House Nr.140)

Zwickl M., Thullner M., Trummer Pál, Town of Szolnok

155

Zwickl Mihály

Zwickl M., Czellei J., Grazi Mária, Town of Szolnok

156

Zwickl Pál (House Nr.121)

Zwickl Pál, Zwickl Mihály, Czellei J., Town of Szolnok

157

Zwickl Pál (House Nr.130)

Zwickl Pál, Reiter János, Weisz A., Town of Szolnok

 

The original table only listed a maximum of four owners per farm; the 13 cases in which “etc.” appears at the end of the list of owners indicate farms comprising land belonging to least 5 different properties. This means that the 48 farms in the table, on average, contained land from at least 3.6 different properties. Moreover, certain names that appear frequently (Trummer Pál, Czellei János, Town of Szolnok, etc.) show that property of one owner was often part of several different farms. While this may be an extreme instance – there is not a single farm that did not use some land that did not belong to the farmer – it does pointedly illustrate that farm and property are clearly not equivalent concepts.

For Prussia we have to use more aggregate data, but the implications are the same: Table 2 below compares the distribution of land by farms from the 1882 agricultural census with the Handbook data for properties (1884) for the province of Pomerania, which was the subject of Buchsteiner’s book. This table shows clearly that the size distribution from farm data and the size distribution from property data are markedly different. In general, for the smaller size categories (the area of agricultural land, as used by the census) there are fewer properties with less area than for farms, but that in the larger size categories exactly the opposite is true. If these data are consistent with each other, that can only mean that, on balance, the larger properties tended to comprise more than one farm (i.e., that the operating unit was smaller than the ownership unit). Overall, the census counted 400 more farms (16%) than the Handbook did properties, yet both total agricultural area and total area overall were nearly the same (slightly larger for the properties in both cases). Taking the distribution of large farms from the agricultural census to represent the distribution of large properties thus markedly understates both the average size of large properties and the dispersion of property sizes.

The principal point having been made, further comparisons can most easily be summarized in charts. Charts 1, 2, and 3 which follow use exactly the same data sources as Table 2 to compare farms, properties, and total ownership using the size categories of the agricultural censuses, measured in hectares of agricultural land. These charts show, respectively, the number of units, the agricultural area of these units, and their total area. The pattern in all three charts is the same: As we move up the size categories, the properties grow in importance relative to farms, and ownership grows in importance relative to both, exactly as we would have predicted. The differences among categories, moreover, are very striking.

The foregoing comparisons dealt with number and area of agricultural units. If we should use the value of the land (or its proxy, the land tax assessment) as a measure of size instead of area, the results of our inquiry could be very different indeed. As Klaus Hess found in his study of landownership in all of Prussia, both West and East:

The most important and in fact also most surprising result of the examination of the land tax net yield statistics lies in the re-evaluation of large properties and large farms in western Prussia ... Relative to the respective areas of the western and eastern provinces, the ownership of large properties in the West – according to the attribute of land tax net yield – was not only much more numerous than in East Elbia, but even of equal significance in regard to its economic importance.27

Unfortunately, the agricultural censuses do not report any data about the land tax net yield in the various farm size categories, so a comparison to the data of the property directories cannot be made. Needless to say, we should expect it to follow roughly the same pattern of increasing relative importance of properties compared to farms as we move up through the size categories, but to a lesser extent than in the area comparisons: The larger properties had relatively higher shares of those land types assessed at lower rates (forest, pasture, moorland, water, untaxed land), so that average tax assessment per hectare tended to fall as property size increases.

Thus, when discussing land ownership in Prussia – or anywhere else, for that matter – it is imperative that one understands the concepts relevant and appropriate for the questions being asked, as well as being familiar with the data available, what they measure, and their limits and non-compatibilities. Armed with this knowledge, a scholar would run much less risk of coming to wrong or misleading conclusions.

 

Notes

1

Jahrbücher für Nationalökonomie und Statistik, 1876 - 1898.

2

Eduard Müller, Der Grossgrundbesitz in der Provinz Sachsen: Eine agrarstatistische Untersuchung, “Sammlung nationalökonomischer und statistischer Abhandlungen des staatsw. Seminars zu Halle a.d.S.,” vol. 67 (Jena: Gustav Fischer, 1912), p. 1.

3

Another source, the volumes concerning the land tax assessment of the 1860s, present extensive data regarding the area and quality of land, without any indication of size of units. Prussia. Ministerium der Finanzen, Grundsteuerveranlagung (25 vols., Berlin: Staatdruckerei, 1864ff). The most important of these data are reprinted in August Meitzen, Der Boden und die landwirtschaftlichen Verhältnisse des Preussischen Staates, vol 4 (Berlin: Paul Parey, 1869), Appendix.

4

Germany. Kaiserliches Statistisches Amt, “Statistik des Deutschen Reiches,” various volumes: 1882: N.F. vol. 5 (1885); 1895: N.F. vol. 112 (1898); 1907: N. F. vol. 212 (1909).

5

The data for both 1878 and 1893 are conveniently summarized in the 1893 volume: Prussia. Königliches Statistisches Bureau, Grundeigenthum und Gebaeude im Pr. Staate, auf Grund der Materialien der Gebaeudesteuerrevision vom Jahre 1893, “Preussische Statistik,” No. 146 (Berlin: Verlag des Königl. Stat. Bureaus, 1898).

6

Preussisches Statistisches Landesamt, Grundbesitzverteilung in Preussen nach den Ergebnissen der ländlichen Verschuldungsstatistik für 1902, Zeitschrift des Preußischen Statistischen Landesamts, Ergänzungsheft 42 (Berlin: Verlag des PSL, 1921).

7

This was in fact a general tax on wealth, including landed wealth. For further details, see for example my “Estimating the value of land from Prussian wealth tax data,” paper presented to the First Conference on German Cliometrics, University of Toronto, September 1999.

8

These amounted to only about 5 per cent of the total land of the country but over 10 per cent in Province of Saxony, Silesia, Hannover, and the Rhineland. (Grundeigenthum, pp. XV - XVI).

9

The space between editions was, for the province of Silesia in the most important of these series of directories (Handbuch des Grundbesitzes im Deutschen Reiche from Paul Parey Verlag in Berlin), considerably longer: First 12, then a full 30 years (1880 - 1892 - 1922). This was very unusual, and almost certainly the result of competition from a Silesia-based series that was long-established, concentrated on and dealt very comprehensively with that province only, and was revised quite frequently (Schlesisches Güteradreßbuch from Verlag Wilhelm Gottlob Korn in Breslau). The main general competition for the Parey series was Niekammers Landwirtschaftliche Adreßbücher, from the Niekammer publishing house in Stettin, but this series started much later and covered fewer provinces.

10

Unfortunately, the Conrad papers in the archive of the University in Halle do not include anything about this series of studies.

11

It is therefore a precise equivalent of, for example, the Hungarian concept of “net cadastral income” (tiszta kataszteri jövedelem), the basis of the land tax in Hungary.

12

See my “The Distribution of Landed Properties by Value and Area: A Methodological Essay based on Prussian Data, 1886 - 1913,” Journal of Income Distribution, vol. 3, no. 1 (Summer 1993), pp. 101 - 140.

13

One Thaler = 3 Marks.

14

See Klaus Hess, Junker und bürgerliche Grossgrundbesitzer im Kaiserreich, “Historische Forschungen,” ed. Karl Erich Born and Harald Zimmerman, vol. 16 (Stuttgart: Franz Steiner Verlag, 1990), p. 61.

15

The 1907 agricultural census does not help us much to determine the answer to this question. In the administrative district of Breslau, where the estate Neuschloss was located, there were 12 farms with 1000 or more hectares of agricultural land, comprising in total 26,214 hectares. By riding, the largest size category is 200 or more hectares. Thus we have too little information from which to deduce how the estate of Neuschloss entered into the 1907 agricultural census.

16

The entail specifies the inheritance principle: primogeniture, seniorat, or majorat in most cases. Other forms were also possible, such as secundogeniture or inheritance by a female.

17

Germany: Kaiserliches Statistisches Amt, Landwirthschaftliche Betriebsstatistik nach der allgemeinen Berufszählung vom 5. Juni 1882, “Statistik des Deutschen Reichs,” N.F. vol. 5 (Berlin: Puttkammer & Mühlbrecht, 1885), p. 9*.

18

Betriebsflächen.

19

There is a legal distinction between possession (Besitz) and ownership (Eigentum). Since this legal distinction has no real relevance for our study, we will not distinguish between these two concepts, and use the term “ownership” to encompass them both.

20

In the German terms, the Betriebsstatistik is taken to be a Besitzstatistik.

21

Except perhaps in a case where, for example, several heirs inherited a property in joint ownership, and one of them farmed it while paying rent to the others.

22

Hess, p. 97. Emphasis added.

23

Ilona Buchsteiner, Großgrundbesitz in Pommern 1871-1914: Ökonomische, soziale und politische Transformation der Großgrundbesitzer (Berlin: Akademie Verlag, 1993). For example, on p. 29 Buchsteiner refers to the properties in the various directories of large properties as Betriebe, and avers that the most important reason that the data of the agricultural censuses cannot be compared with the data of the property directories is that the two sources use different measures of size: the former uses agricultural area, the latter total area. She also goes on to point out (p.30) that different treatment of woodland also contributes to the observed differences, that the directories of large properties left out some “peasant properties,” and that some satellite properties (Vorwerken) might have been treated separately in the census, but included with the main property in the directory. Entirely missing from this discussion is any notion that the concepts themselves are different. Indeed, throughout the entire book she consistently refers to both farms and properties as Betriebe.

24

For a comparison of the structure of land ownership in Hungary and Prussia, see my “Junkers and Magnates: The Social Distribution of Landed Wealth in Prussia and Hungary; A Case Study of Pomerania and Transdanubia, 1893,” Österreichische Osthefte, Nr. 1, 1994, pp. 109 - 131.

25

Földmívelésügyi Magyar Királyi Miniszter [Royal Hungarian Minister of Agriculture], Magyar Korona Országai Mezőgazdasági Statisztikája II. kötet: Gazdacímtár [Agricultural statistics of the Lands of the Holy Hungarian Crown, vol. II: Directory of Farm Operators] (Budapest: Magyar Királyi Állami Nyomda [Royal Hungarian State Press], 1897).

26

After World War I, the Treaty of Trianon awarded a large part of Moson county to Austria, and some villages to Czechoslovakia.

27

Hess, p. 98.