Jump to ContentJump to Main Navigation
Landownership in Eastern Germany Before the Great WarA Quantitative Analysis$

Scott M. Eddie

Print publication date: 2008

Print ISBN-13: 9780198201663

Published to Oxford Scholarship Online: May 2008

DOI: 10.1093/acprof:oso/9780198201663.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2016. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 08 December 2016

(p.250) Appendix Coding the Handbook Data

(p.250) Appendix Coding the Handbook Data

Landownership in Eastern Germany Before the Great War
Oxford University Press

A.1. Location Coding

To facilitate retrieval, checking, and combining of different groups of properties or owners, each property received special codes for the following: source from which it was taken, geographic location, and type of owner and condition of ownership. Additional codes identified whether it was a principal property of 100 hectares or more, whether it was usable for cluster analysis, and special ‘unfit’ codes which identified whether it could be used in various kinds of regressions or in the cluster analysis. Source coding is explained in Section A.3.2, ‘Reference Variables and Fields’, as are other codes in their respective sections.

The geographic location code identified the province, the administrative region (Regierungsbezirk) and the district (Kreis) of that administrative region in which the property was located. Since the Handbooks listed properties by district, further narrowing down of location was not possible (nor necessary). The location code was a five‐digit number: the first digit identified the province, the second and third the administrative region, and the fourth and fifth the district within that region. I chose to use the administrative divisions of Prussia as they existed in 1895, and used the standard numbering used by the Prussian Statistical Bureau in its yearbooks and other statistical publications for the provinces and administrative regions, as shown in Table A.1.

As one can see from the numbering of the administrative regions, the provincial numbering is redundant, but I retained it anyway because it makes searching for provincial data in the databases considerably easier. To give an example of geographic location coding, all properties located in the district of Ruppin (number 16 within the Potsdam region) received the code 40616: 4 for Brandenburg, 06 for Potsdam, and 16 for Ruppin. Every property in the district of Frankenstein received a location code of 71317: 7 for Silesia, 13 for Breslau, and 17 for Frankenstein district within Breslau administrative region.

Berlin, although included in the coding scheme for completeness, in fact does not figure in the description and analysis of this study because it contained no properties of at least 100 hectares in size.

A.2. Owner Codes

Each property also received an owner code that made it possible to group properties by the social or economic position of their owners. While this code is actually a four‐digit code, only three digits are relevant for this study (in some owner categories only two), and the fourth has remained unused.


Table A.1 Geographic Location Numbering Scheme


Admin. region within province






East Prussia






West Prussia













































(a) Berlin data included in Brandenburg.

(b) Berlin data included in Potsdam.

A.2.1. Physical Persons

The finest divisions of the owner code are found in the category of physical persons, as shown in Table A.2. According to this scheme, a property owned by a baron and his wife would be coded 1320 (the last digit will always be zero), one owned by commoner siblings would be coded 2030, and one owned by the heirs of a count would receive a code of 1440. If a prince were the single owner of the property, its owner code would be 1510, just as it would be 2010 if owned by one person of non‐noble status. In Silesia in 1892 the property of Porschütz (264 ha) was owned by the Count Campani Foundation for Girls, so it received the owner code 1450, whereas the entailed estate of Klein Steegen in East Prussia (1792 ha), owned by Captain of Cavalry (ret.) Oscar von Steegen, received an owner code of 1090.

Royalty were coded separately. Any member of any branch of the Hohenzollern family was considered to be Prussian royalty. Thus the estate of Krojanke in West Prussia (2734 ha in 1894), owned by Friedrich Leopold Hohenzollern‐Sigmaringen, Prince of Prussia, received an owner code of 1710 (aristocrat, Prussian royalty, single ownership), whereas the estate of Kraussnick (490 ha in Brandenburg in 1896), a royal entail of King Wilhelm II, had an owner code of 1770 (aristocrat, Prussian royalty, royal entail). In contrast to Kraussnick, Norkitten in East Prussia (4804 ha in 1907), received a code of 1670 because it was an entailed estate of the Duke of Anhalt‐Dessau, one of the 22 recognized (p.252)

Table A.2. Owner Codes for Properties Owned by Physical Persons

First digit

Second digit

Third digit

1 Aristocrat

0 Untitled noble

0 unused

1 unused

1 single owner

2 Knight, Baronet

2 with spouse

3 Baron, Freiherr, Edler

3 several owners

4 Count, Reichsgraf, Burggraf, Markgraf

4 heirs

5 Prince, duke, archduke

5 personal foundation

6 Other royalty

6 royal domaine

7 Prussian royal family

7 royal entail

8 unused

8 unused

9 unused

9 entail of other nobility

2 Non‐noble person (‘bourgeois’)

0 in all cases

0 unused

1 single owner

2 with spouse

3 several owners

4 heirs

5 personal foundation

6 unused

7 unused

8 bankrupt estate

9 entail of non‐noble person

sovereigns in Germany. Czeszewo in Silesia (2687 ha in 1896), which belonged jointly to Princess Marie and Hereditary Prince Bernhard of Saxony‐Meiningen, from another of the 22 recognized sovereign houses of Germany, received an owner code of 1630.

A.2.2. Non‐Physical Legal Persons

Non‐physical legal persons presented a different kind of variety. Let us consider first property owned by a religious body or a state (Table A.3). Some examples: Nadziejewo (449 ha in Posen in 1896) was owned by the Catholic Priests’ Seminar, so was coded 3140. The Cathedral in Frauenberg in East Prussia, assumed to be Lutheran, owned the 205‐hectare property of Grundhof, which therefore received an owner code of 3220. The local (Lutheran) church in Guztkow, Pomerania owned a property called Strellin (223 ha) in 1893, so Strellin's owner code became 3230.

State properties are denoted by a first digit of 4. A second digit of 0, 1, or 2 denotes the Prussian State, another German state, or a foreign state, respectively. Finer divisions into different state entities such as forestry, salt production, railways, stud farms, the military, etc. in the third digit were of no consequence for any of the purposes of this study. Thus most state properties received an owner code of 4000 (4010 if owned by the Settlement Commission for Posen and West Prussia), 4100, or 4200 to denote their ownership by Prussia, another German state, or a foreign state.

To illustrate: the estate of Winnigen in Saxony, 1126 hectares in 1885, owned by the Duchy of Brunswick (Herzoglicher Braunschweiger Fiscus), received an owner code of 4100. The Swedish crown property of Grünhufe in Pomerania, 332 hectares in 1884, (p.253)

Table A.3. Owner Codes for Religious Bodies and States

First digit

Second digit

Third digit

3 Religious bodies

1 Roman Catholic

1 Bishopric, archdiocese

2 Reform (protestant)

2 Church district, abbey

3 Greek Orthodox

3 Local church (parish)

4 Greek Catholic

4 Religious schools, foundations, monastic orders

5 Other

4 States

0 Prussian State

0 General fiscus

1 Other German state

1 Prussian settlement commission

2 Foreign state

2 unused

3 Joint ownership with another owner

4 Forest fiscus

5 Domains and stud farms

6 Railway fiscus

7 Military fiscus

8 Royal salt works

was coded 4200. Of properties of the Prussian State in West Prussia in 1894, Freudenthal (154 ha) received an owner code of 4040 because it was owned by the forest fiscus, but Stellinen (814 ha) was coded 4030 because it was jointly owned by the State and Julius Klaasen. The Royal Settlement Commission had purchased, but not yet subdivided, the estate of Zietlonkowo in Posen (559 ha in 1910), so that property received an owner code of 4010. The Prussian military's 5500‐hectare estate of Neuhammer in Silesia in 1909, coded therefore as 4070, was a military exercise ground (Truppenübungsplatz), whereas the property of Ulrichsfelde in West Prussia (137 ha in 1885), received a code of 4060 because owned by the Prussian state railways.

Various levels of communities were combined in most tables and other analyses in the study. The scheme for owner coding of community‐owned properties was as shown in Table A.4 (the third digit was zero in all cases).

Finally we come to the ‘other’ category of owners, where all those owners who do not fit into any of the earlier categories are assigned. The first digit is 9, and the third is 0 in all cases. The second digit denotes the kind of owner, as follows:

  1. 1. Association

  2. 2. Non‐religious foundation

  3. 3. Other legal person or society

  4. 4. Hospital

  5. 5. School

  6. 6. Orphanage

  7. 7. University or academy

  8. 8. Unknown owner

  9. 9. ‘Divided or parcelled’—this designation was for large properties listed in the Handbooks as being, or having been, subdivided into smaller units.1


Table A.4. Owner Codes for Communities or Business Firms

First digit

Second digit

Third digit

5 Community (Gemeinde)

0 all cases

0 all cases

6 Communal owners

0 unused

0 all cases

1 Former serfs

2 Other communal group (e.g. pasturage society)

7 City or larger administrative unit

0 City

0 Sole ownership

1 District (Kreis)

1 unused

2 Administrative region (Regierungsbezirk)

2 unused

3 Province

3 Joint ownership with another body

8 Banks and businesses

0 unused

0 All cases

1 Banks, financial institutions

2 Indl. & commercial companies

3 Savings associations

The 9900 code indicates that such properties should be left out of our working database, and not entered into any of the calculations for large properties performed in this study. Thus a property with the owner code of 9900 plays no part in any of the descriptions or analyses of this book, but has been kept in the underlying database in the interests of retaining potentially useful information.

Some examples of the coding of ‘other’ owners are the following: the 222‐hectare property Friedrichswille in Brandenburg in 1885, owned by the Provincial Society for Combating Vagabondage, received the owner code of 9100; in Saxony in 1885, the estate Nagelstedt (268 ha), owned by the Griefstädter Foundation Fund in Erfurt, received a code of 9200, while the 237‐ha property Schönwiese in East Prussia was coded to 9700 because it was owned by the Academic Senate in Königsberg. All three Silesian properties of the Central Hospital in Görlitz received an owner code of 9400, just as the six Silesian properties of the Royal Land School in Pforta were coded as 9500.

A.3. Other Codes

A.3.1. Additional Characteristics of Owner or Property

Separate variables have been set up in the databases to indicate other characteristics of the owner or the property. (p.255)

  1. 1. For example, the variable OF indicates whether the owner was a military officer or wife of an officer (including active service, reserve, and retired officers, in all branches of the military including the Home Guard (Landwehr)). A value of 1= ‘Yes’ for this variable, 0 = ‘No’.

  2. 2. For the chapter on knight's estates, any property that was a knight's estate received a value of 1 for the variable ‘RG’ (for Rittergut), otherwise 0. If a property listed as a subunit of a manor was a knight's estate, where the manor was recorded as the principal property, then the variable ‘ManorRG’ recorded the number of knight's estates contained within the manor as subunits.

  3. 3. A separate variable, called ‘Unfit’ indicates whether and why a particular property should not be included in any regression analysis or cluster analysis, according to the schema shown in Table A.5.

Depending on the purpose of the statistical analysis, different categories of ‘unfit’ observations had to be eliminated. For example, in the cluster analyses of Chapter 7, using proportions of arable, forest, etc. land as variables, only observations unfit for reasons 1 through 5 had to be eliminated. Numbers 6, 7, and 9 touched only on the land tax assessment, not used in the cluster analyses mentioned, and number 8 could be included because land data was taken from elsewhere only in cases where (a) totals for a given property were exactly the same in two different years, but the land details were missing in one, or (b) land details were the same in two different years, but the totals did not match (this last most likely because of typographical or transcription error).

In regressions such as those of Chapter 9, where the dependent variable was the land tax net yield (GSRE), and independent variables included the areas of ploughland, meadow, forest, etc., not only did observations unfit for reasons 1 through 5 have to be eliminated, but also those unfit for reasons 6 and 9.

Table A.5. Values for the Variable ‘Unfit’


What the value indicates

0 or blank

Fit to be included in any kind of regression or cluster analysis


Arbitrary allocation of land


No details of land, only sum, in source


Data taken from elsewhere, but anomalies remain


Land data estimated on basis of GSREa, or empty rowb


Sum of details differs from total by 5% or more


GSREa allocated arbitrarily between or among properties


GSRE data taken from elsewhere, OK


Land data taken from elsewhere, OK


GSRE allocated according to Kreis averages

Note: 1–9 indicate why variable should not be used in statistical analysis.

(a) GSRE = Grundsteuerreinertrag, or ‘land tax net yield’, the basis for calculating the land tax payable for a property.

(b) ‘Empty row’ indicates that only the property name, and perhaps the name of the owner, was given, but there was no other information about the property.

(p.256) A.3.2. Reference Variables and Fields

To facilitate easy reference to any single observation, and for later checks on the accuracy of data entry, several reference variables were added to the database. The most important of these are the following:

  1. 1. Province: The Handbooks had separate editions for each province. This identifies the individual province editions from which the data were taken.

  2. 2. Year: This is the year of publication of the particular edition of the Handbook used for the province in question.

  3. 3. Source: The ‘Source’ code identifies the page and line number of the volume where the given property is to be found: the first three digits are the page number, the last two the line number. Thus 45307 means page 453, line 7, and 215 (really 00215, but most spreadsheet and database software does not display initial zeros) means page 2, line 15.

  4. 4. SchlPropNr: The 1909 Silesian address book used gave each property a unique sequence number. For the observations from this province only, and only in the c.1907 database, ‘SchlPropNr’ recorded these unique sequence numbers to facilitate even more accurate reference to where the property could be found in the source book. Additionally, each of these properties also received a five‐digit ‘Source’ number as in number 3 above.

  5. 5. Notes: This is perhaps the most important of the reference fields. It contains information about the state of the property and any changes made to the data in the course of ‘cleaning’ and checking those data.

    1. (a) Sometimes there was a note about the property in the source book itself. The single most common one of these was ‘Subdivided or in the process of subdivision’, which led to the assigning of owner code 9900 and the exclusion of the property in question from any of the data used in this book. Other notes appeared, such as ‘see appendix’, in which there was usually some modification to the information about the property, or ‘data unavailable’, in the case where a property (and usually its owner as well) was listed, but no further details were shown. All such notes were included.

    2. (b) When one or more industrial establishments listed in the source book for a given property were assigned to the industrial category ‘other’, they were named in the Notes field.

    3. (c) Any changes to the data of the property also appear in the Notes. If, for example, the areas of ploughland and meadow were combined in the source, but assigned separate values later, the assignment and reasons for it appear in the Notes. Perhaps the single most common note would be ‘GSRE allocated according to Kreis averages’ (leading to a value of 9 for the ‘Unfit’ variable, see above) for those properties for which no land tax net yield was reported in the source, leading to an attribution based on the average land tax net yield per hectare of each of the land types in the district in which the property was located. All other modifications and attributions, with their reasons, would also appear in the Notes.

    4. (p.257)
    5. (d) Even if no modification were made, a Note might still be added to indicate an unsuccessful attempt at supplementing or modifying the data. For example, for the estate Gross Tromp in the district of Braunsberg in East Prussia, 482 hectares and home to a brickworks and a dairy, owned by Gustav Fuchs in 1907, no land details were given, only the total hectares and the land tax net yield. An unsuccessful attempt to find details in the East Prussian Handbooks of 1884, 1895, and 1929 led to the following Note for this property: ‘No details. Does not exist in 1929. Smaller in 1884 and 1895.’ Or, if the reported details of ploughland, meadow, pasture, etc., did not add up to the total area reported for the property, and the anomaly could not be resolved, a Note would point this out. If the difference were more than ± 5 per cent, this would also lead to a value of 5 for the ‘Unfit’ variable, as noted above.


(1) Some of these properties were still technically owned by the Royal Settlement Commission for Posen and West Prussia, but nevertheless received an owner code of 9900 because they had already been subdivided, or were in the process of subdivision. Only those properties of the Settlement Commission that were still intact, or those unsubdivided parts retained as a Restgut (remainder property) by the Settlement Commission received the owner code of 4010.