1
Feasibility Study on a European
Foundation Statute
2008
Annex A
Methods of the survey
EUF Study 2008 Annex A - Methods of the survey page 2 of 16
2
I. General facts
1. Development of the sample
The sample is divided in two different parts. The first part implies the data that was collected
in spring 2008, whereas the second part includes the data gained by the European Foundation
Centre (EFC).
The questionnaires were sent to 630 different European Foundation whereof 134 responded
and participated in the survey. Therefore there was obtained a response rate of 21 percent. The
result is the first field study in this area and consists of 134 cases of foundations in 24
different countries of the European Union.
The primary data was set as a postal survey, using an address sample. The selection comprises
the largest and most known foundations.
Information about the foundations was taken from different data sources, for instance:
­ Selection made by Donors' and Funders' Networks in Europe (DAFNE) contact:
* Finland
* Poland
­ Taken from Research Task Force (RTF) data:
* Estonia
* Slovakia
* Sweden
All members of DAFNE were contacted, as well as other partners (e.g. EFC members, or
other legal experts) to get the names and addresses of foundations. Data was also taken from
the Research Task Force (RTF)
Besides the help of DAFNE and RTF data, the selection was made by choosing the largest
and most well known foundations.
EUF Study 2008 Annex A - Methods of the survey page 3 of 16
3
This scheme represents the responded questionnaires:
Table A 1: Responded questionnaires
Country received sent percentage additional
Hungary 5 11 45.5 %
Germany 17 42 40.5 %
Czech republic 6 19 31.6 %
Estonia 6 20 30.0 %
Finland 6 20 30.0 %
Netherlands 7 27 25.9 %
Spain 6 26 23.1 % 2
UK 8 35 22.9 %
Italy 13 59 22.0 %
Slovakia 8 37 21.6 %
Belgium 4 19 21.1 %
Malta 4 20 20.0 %
Ireland 3 17 17.6 %
Greece 4 23 17.4 %
Poland 7 41 17.1 %
France 12 74 16.2 %
Cyprus 3 19 15.8 %
Sweden 2 15 13.3 %
Portugal 2 16 12.5 %
Romania 2 16 12.5 %
Denmark 3 29 10.3 %
Lithuania 2 21 9.5 %
Slovenia 1 11 9.1 %
Latvia 1 13 7.7 %
Gesamt 132 630 21.0 % 2
Zusätzlich 2 2 Reports
SUMME 134
The countries Lithuania, Slovenia and Latvia are noticeable at their low return rate, which
ranges around 7,7 and 9,5 percent, whereas the countries Hungary, Germany and Czech
Republic are represented by a high returning rate around 31,6 and 45,5 percent. The overall
result represents an average rate of 21 percent.
EUF Study 2008 Annex A - Methods of the survey page 4 of 16
4
2. Representativeness
The survey serves more as an indicator as a representative study, because of its small case
number and of the intended selection of simply large and well known foundations. Therefore
the selection was not made stochastically as the standard required, but in addition with the
secondary data of the EFC it is scientifically usable to demonstrate an impression of the
sector.
3. Secondary data
There were prior foundation sector studies about foundations we partly draw on. Those are:
- EFC- Data
- American Foundations: Contributions and Promise1
- OECD2
- Comparative Non-profit Sector Project/ The Johns Hopkins Center for Civil Society
Studies3
- The UN Non-profit Handbook4
For a secondary analysis of the European foundation sector we mainly refer to the EFC
provided data.
The second survey coordinated by the Research Task Force was carried out with the
assistance of the Research Task Force members in the participating EU Member States. There
has been one national coordinator for each country who was supporting the project in
collecting and compiling the data of the foundations on the basis of the questionnaire
developed by the group. The survey compiled 20 questions about the foundation sector.
After major efforts to extend the survey to a wider number of EU member states and
consequently confronting the difficulties with the collection of data at national level, the RTF
took a step forward to simplify the survey to the missing countries to six key questions on the
foundation sector (number of public benefit foundations, total assets, total expenditure,
employment and volunteering weight of the sector and fields of activity). After this
simplification two other countries joined the RTF efforts to portrait the dimensions of the
foundation sector in the EU and Hungary.
1
Anheier, H. and D. Hammack (2007) American Foundations: contributions and promise. Aspen Project
Volume, 51
2
Main Science and Technology Indicators, Print + PDF Edition (ISSN 1011-792X) - PDF Edition (ISSN 1609-
7327), Issue: Volume 2007/2
3
Lester M. Salamon and S. W. S. a. associates (2004). Global Civil Society: Dimensions of the Nonprofit Sector.
http://www.jhu.edu/~cnp/research/country.html
4
The UN Non-profit Handbook is a project approved by the UN Statistical Commission and developed by the
Johns Hopkins Center for Civil Society Studies, led by Dr. Lester Salamon
EUF Study 2008 Annex A - Methods of the survey page 5 of 16
5
II. Description of survey
The data of the survey was collected by post and is based on standardized questionnaires
existing in different language versions. The questionnaires were sent in English, but there has
been the option to download versions in German or French from the website.
1. Typology of questions
The questionnaires contain open, closed and half standardized questions that apply several
different topic categories.
There were open questions such as:
"Can you briefly describe how you solved one of the most pressing issues?"
There were closed questions such as:
"Do you conduct international activities never, rarely, occasionally, or more or less
regularly?"
There were half standardized questions such as:
"What is the geographic scope of your international activities? (Please check primary
scope)"
2. Procedure of data collection
Table A 2: Procedure of activity
Posting of survey 29. February 2008
1. Deadline 01. April 2008
1. Reminder 08. April 2008
2. Reminder 18. April 2008
2. Deadline 25. April 2008
3. Deadline 09. May 2008
The table demonstrates the procedure of activity, both the posting and the returning of the
questionnaires. The questionnaires were sent in February 2008. A low response after the first
set deadline caused a new setting of the time management. There was also initiated a reminder
to recall the survey. The 9th
of May was set as the ultimate deadline.
EUF Study 2008 Annex A - Methods of the survey page 6 of 16
6
3. The constitution of content of the questionnaire
The survey is divided in seven parts.
Part 1 and part 2 provide general information about the analysed foundations, as well as their
financial structure. Those parts present a short overview about the involved foundations and
classify those that are operating international. The representation of the financial structure is
important to calculate the distribution of their assets and expenditures and build the basis for
the weighing of the data.
Part 3, 4 and 5 provide in general the main analysis and the cooperation of international
operating foundations. Part 3 particularly asks for obstacles and difficulties within
international activities. Part 4 asks especially for the expansion structure and/or for the
involvement in international activities. Part 5 seeks to figure out if international activities
have been reduced. The last two parts (6 and 7) are dedicated to the foundation's attitude
toward the European Foundation Statute and represent reflected advantages or disadvantages
about the European Foundation Statute.
4. The objectives of the study
The study aims to cover the following specific objectives:
Table A 3: Objectives and variables
Subject area Data sheet
Overview of the main types of foundations (or trusts
where appropriate)
V_1.1, V_1.1a, V_1.2, V_1.3,
V_1.4, V_1.5,
V_2,
Screening question 3, V_4
Cross border activities ­ barriers and their economic
relevance
V_3
Estimation of the importance and cost of these barriers
Analysis of possible modalities of elimination of these
barriers (including introduction of a European Foundation
Statute)
V_3.3, V_3.4,
V_4N
Assessment of the possible effects of a European
Foundation Statute.
V_6,
V_7,
Comparison with the United States regarding the
importance of the foundation sector in the economy.
new data
EUF Study 2008 Annex A - Methods of the survey page 7 of 16
7
The specific objectives can be answered by the different variables that are listed on the
column at the right side of the table.
An overview of the main types of the foundation can be given with different variables. In
particular it is asked about the main objectives of the foundation, activity areas, the modality
how to achieve the objectives, the geographic scope of international activities, the conduction
of activities, modality of foundation, as well about financial aspects like assets and
expenditures. To classify the foundations it is analysed if they are international active.
Another objective of the study is to identify the dimension of barriers and their economic
relevance of cross border activities by means of information of experienced significant
barriers, experienced difficulties and the solution of these problems.
Furthermore the possible modalities of the described barriers can be analysed by asking about
different types of solutions to overcome these. By identifying these problems it is next asked
why foundation does not intend to increase international activities, especially to find out what
barriers are most important.
By using a likert scale it is tried to explore the diverse assessments regarding the possible
effects of a European Foundation Statute.
III. Information about procedural method of survey
The data used for the following calculation is drawn mainly from our field study. The data set
contains 134 cases of (public benefit?) foundations in 24 countries of the European Union.
Concerning size and weight, these foundations differ considerably, so that the sample covers
the diversity of the foundation sector at least satisfactory. But due to the quota sample used
for the selection of cases, it is not to be expected that distributions of attributes among the
foundation observed will equal the distribution in the ground population. Therefore a
weighing factor had been developed, to get at least an impression of the sector.
1. Weighing of Data
The task of calculating on the basic population of foundations from the survey is one of the
major methodological tasks of the study. Since there is very little reliable information on the
ground population, we have to weight the observed cases by some very rough indicators.
Aggravating to the unsatisfying data situation is the fact that some information are only
available for some countries, so that we have to deal with averages rather than exact figures.
The way we chose for the weighing of the cases was to calculate the distribution of assets and
expenditure since these are the most validated numbers especially from the EFC surveys. It
seems appropriate to divert the cases in Top15 and Non-Top15, since the concentration of
EUF Study 2008 Annex A - Methods of the survey page 8 of 16
8
capital on the 15 biggest foundations is so high that this diversion seems to depict the sector
quite well.
The weighing of the data followed 5 steps, which are described in detail below
Step1: Extraction of breaks for expenditures and assets according to EFC Data
From the EFC survey, we can learn about the Top 15 share of assets and expenditures for 10
out of 24 countries.
Table A 4: Breaks for Top 15 Assets/ Expenditures
Country Break Assets Break Expenditures
Belgium 7,000,000  2,000,000 
Cyprus X X
Czech Republic X X
Denmark X X
Estonia 2,000,000  2,000,000 
Finland 56,000,000  8,000,000 
France 103,000,000  60,231,000 
Germany 35,000,000  15,000,000 
Greece X X
Hungary 13,000,000  7,000,000 
Ireland X X
Italy 691,000,000  18,000,000 
Latvia X X
Lithuania X X
Malta X X
Netherlands X X
Poland X X
Portugal X X
Romania X X
Slovakia X X
Slovenia X X
Spain 93,000,000  57,000,000 
Sweden 98,000,000  X
United Kingdom 350,000,000  37,500,000 
EUF Study 2008 Annex A - Methods of the survey page 9 of 16
9
Step 2: Calculation of missing values.
To fill the gaps in theTable A 4 above, we calculated averages for the cut-off values in
clusters of countries. These clusters are drawn from the welfare state typology by EspingAndersen
and from Anheier (2005)5
Table A 5 depicts the clustering of countries. Within each cluster, we calculated the average
for the dividing value between Top15 and Non-Top15 foundations. Than we related these
averages as `new' dividing value (Grenzen) to those countries, we have no information about
Top15 foundations from the EFC data.
Table A 6 shows the same values than Table A 5, including the calculated averages. One
special case is Ireland. Here both variables (Assets and Expenditures) are in all three cases
observed that low, that it makes no sense whatsoever to implement a cut-off value here. In the
following, we classified all three cases as Non-Top15.
Table A 5: Cluster of Countries
Countries Classification
Belgium
Netherlands Continental Europe /
France conservative welfare states
Germany
Finland
Sweden Scandinavia /
Denmark social democratic welfare states
Cyprus
Greece Southern Europe
Italy
Malta
Portugal
Spain
Ireland Liberal welfare states
United Kingdom
Czech Republic
Estonia
Hungary
Latvia Eastern Europe
Lithuania
Poland
Romania
Slovakia
Slovenia
Including the averages, Table A 4 becomes completed to Table A 6.
5
{Anheier, 2005 #8}
EUF Study 2008 Annex A - Methods of the survey page 10 of
16
10
Table A 6: Breaks for Top 15 Assets/ Expenditures (including averages)
Country Break Assets Break Expenditures
Belgium 7,000,000  2,000,000 
Cyprus 392,000,000  37,500,000 
Czech Republic 7,500,000  4,500,000 
Denmark 77,000,000  8,000,000 
Estonia 2,000,000  2,000,000 
Finland 56,000,000  8,000,000 
France 103,000,000  60,231,000 
Germany 35,000,000  15,000,000 
Greece 392,000,000  37,500,000 
Hungary 13,000,000  7,000,000 
Ireland X X
Italy 691,000,000  18,000,000 
Latvia 7,500,000  4,500,000 
Lithuania 7,500,000  4,500,000 
Malta 392,000,000  37,500,000 
Netherlands 48,333,333  25,743,667 
Poland 7,500,000  4,500,000 
Portugal 392,000,000  37,500,000 
Romania 7,500,000  4,500,000 
Slovakia 7,500,000  4,500,000 
Slovenia 7,500,000  4,500,000 
Spain 93,000,000  57,000,000 
Sweden 98,000,000  X
United Kingdom 350,000,000  37,500,000 
Note: Averages are highlighted bold.
EUF Study 2008 Annex A - Methods of the survey page 11 of
16
11
Step 3: Implementation of variable "Top15ASS" and "Top15Ex"
Using the break values from step2, we created two new variables, named "Top15ASS" and
"Top15Ex". With this step, we achieved the first classification of our cases.
Table A 7: Classification in "Top15ASS" and "Top15EX"
Country Assets Expenditures
Top15Ass Not-Top15Ass Top15EX Not-Top15EX
Belgium 4 0 4 0
Cyprus 0 3 0 3
Czech Republic 2 4 0 6
Denmark 2 1 2 1
Estonia 4 2 3 3
Finland 5 1 4 2
France 2 12 1 13
Germany 10 6 5 11
Greece 0 4 0 4
Hungary 0 5 0 5
Ireland 0 3 0 3
Italy 1 12 2 11
Latvia 0 1 0 1
Lithuania 0 2 0 2
Malta 0 4 0 4
Netherlands 2 5 1 6
Poland 2 5 2 5
Portugal 1 4 1 4
Romania 0 2 0 2
Slovakia 0 7 0 7
Slovenia 0 1 0 1
Spain 2 4 1 5
Sweden 1 0 1 0
United Kingdom 4 4 1 7
Total 42 92 28 106
From Table A 7 we can see that not all categories are represented in each country. This is
mostly due to the small sample but also to the rough classification.
Step 4: Correction of breaks to get all categories filled
We choose a pragmatic way to overcome the unpleasant fact of having not all categories
represented in the sample. Since the break values taken from the EFC data are in fact not that
reliable also (the collection of data was all but uniformly across all countries), we "moved"
these values in a way that we have at least one case in both categories in every country.
EUF Study 2008 Annex A - Methods of the survey page 12 of
16
12
Therefore we looked at the minima and maxima of the respective variables and corrected the
bread value correspondently.
Table A 8: Corrected Breaks for "Top15ASS" and "Top15EX"
Country Break Assets
(uncorrected)
Break Assets
(corrected)
Break
Expenditures
(uncorrected)
Break
Expenditures
(corrected)
Belgium 7,000,000  21,000,000  2,000,000  2,000,000 
Cyprus 392,000,000  30,000,000  37,500,000  17,500,000 
Czech Republic 7,500,000  7,500,000  4,500,000  3,200,000 
Denmark 77,000,000  77,000,000  8,000,000  8,000,000 
Estonia 2,000,000  2,000,000  2,000,000  2,000,000 
Finland 56,000,000  56,000,000  8,000,000  8,000,000 
France 103,000,000  103,000,000  60,231,000  60,231,000 
Germany 35,000,000  35,000,000  15,000,000  15,000,000 
Greece 392,000,000  30,000,000  37,500,000  1,300,000 
Hungary 13,000,000  1,000,000  7,000,000  6,000,000 
Ireland X 10,000,000  X 450,000 
Italy 691,000,000  691,000,000  18,000,000  18,000,000 
Latvia 7,500,000  7,500,000  4,500,000  4,500,000 
Lithuania 7,500,000  7,500,000  4,500,000  70,000 
Malta 392,000,000  392,000,000  37,500,000  80,000 
Netherlands 48,333,333  48,333,333  25,743,667  25,743,667 
Poland 7,500,000  7,500,000  4,500,000  4,500,000 
Portugal 392,000,000  392,000,000  37,500,000  37,500,000 
Romania 7,500,000  7,500,000  4,500,000  300,000 
Slovakia 7,500,000  2,000,000  4,500,000  200,000 
Slovenia 7,500,000  7,500,000  4,500,000  4,500,000 
Spain 93,000,000  93,000,000  57,000,000  57,000,000 
Sweden 98,000,000  98,000,000  X 8,000,000 
United Kingdom 350,000,000  350,000,000  37,500,000  43,000,000 
Note: corrected figures are highlighted bold
EUF Study 2008 Annex A - Methods of the survey page 13 of
16
13
Table A 9: Classification in "Top15ASS" and Top15EX" (corrected)
Country Assets Expenditures
Top15Ass Not-Top15Ass Top15EX Not-Top15EX
Belgium 3 1 2 2
Cyprus 1 2 1 2
Czech Republic 2 4 1 5
Denmark 2 1 2 1
Estonia 4 2 3 3
Finland 5 1 4 2
France 2 12 1 13
Germany 10 6 5 11
Greece 1 3 1 3
Hungary 1 4 1 4
Ireland 1 2 1 2
Italy 1 12 2 11
Latvia 0 1 0 1
Lithuania 0 2 1 1
Malta 0 4 1 3
Netherlands 2 5 1 6
Poland 2 5 2 5
Portugal 1 4 1 4
Romania 0 2 1 1
Slovakia 1 6 1 6
Slovenia 0 1 0 1
Spain 2 4 1 5
Sweden 1 0 1 0
United Kingdom 4 4 1 7
Total 46 88 35 99
Notes:
a) The Top15 Ass category in Lithuania, Malta and Romania is empty, because the cases
observed are much too small to be reasonably categorized as Top15 Foundations.
b) In countries where only one case is observed, obviously one category stays empty (Latvia,
Slovenia and Sweden)
Step 5: Creation of weighing factor according to number of cases in each category
Starting from the classification shown in Table A 9, we constructed two weighing factors, one
for the distribution for assets and one for expenditures. For this task, we used the number of
foundations per country, given in the EFS data and divided this number by the number of
cases observed in each category. This operation leads to the following (artificial) data set of
about 83,000 cases.
EUF Study 2008 Annex A - Methods of the survey page 14 of
16
14
It is important to note that these are two separate calculations. It is not possible to calculate on
both factors at the same time. So we either look at foundations that are different in size in
terms of assets or in terms of expenditure.
Table A 10: Weighed data set for assets and expenditures
Country Assets Expenditures
Top15Ass Not-Top15Ass Top15EX Not-Top15EX
Belgium 15 385 15 385
Cyprus 15 40 15 20
Czech Republic 15 1190 15 1240
Denmark 15 13985 15 13985
Estonia 15 168 15 168
Finland 15 2585 15 2585
France 15 1211 15 1211
Germany 15 11985 15 11985
Greece 15 474 15 474
Hungary 15 16692 15 16692
Ireland 15 92 15 92
Italy 15 4705 15 4705
Latvia 0 130 0 130
Lithuania 0 350 15 130
Malta 0 400 15 300
Netherlands 15 1385 15 1155
Poland 15 5985 15 5985
Portugal 15 470 15 470
Romania 0 500 15 250
Slovakia 15 388 15 388
Slovenia 0 128 0 128
Spain 15 10820 15 10820
Sweden 15 0 15 0
United Kingdom 15 8785 15 8785
Total 285 82853 330 82083
For the following calculations, we will give three numbers each, where possible. One for the
raw data and one weighed for the two criteria described above.
To check the quality of the weighing factors, we compare correlations between the variables
,,Value of assets" (V_2.1.3) and "Operating expenditure" (V_2.3.3). In all cases, we see a
highly significant result, which is 0.532 for the unweighed data set, 0.679 for the data
weighed for expenditures and 0.150 if we weigh the data for assets. Because we have no
indication how the correlation in the ground population might be, the high value for the
weighing factor for expenditures does not necessarily mean that this is the most representative
model.
EUF Study 2008 Annex A - Methods of the survey page 15 of
16
15
Table A 11: Correlation of assets and expenditures (weighed for unweighed)
Correlations
Value of assets
(total)
Operating
expenditures
(total)
Pearson Correlation 1,000 ,532**
Sig. (2-tailed) ,000
Value of assets (total)
N 124 117
Pearson Correlation ,532**
1,000
Sig. (2-tailed) ,000
Operating expenditures (total)
N 117 119
**. Correlation is significant at the 0.01 level (2-tailed).
Table A 12: Correlation of assets and expenditures (weighed for expenditures)
Correlations
Value of assets
(total)
Operating
expenditures
(total)
Pearson Correlation 1,000 ,679**
Sig. (2-tailed) ,000
Value of assets (total)
N 71199 67541
Pearson Correlation ,679**
1,000
Sig. (2-tailed) ,000
Operating expenditures (total)
N 67541 75887
**. Correlation is significant at the 0.01 level (2-tailed).
EUF Study 2008 Annex A - Methods of the survey page 16 of
16
16
Table A 13: Correlation of assets and expenditures (weighed for assets)
Correlations
Value of assets
(total)
Operating
expenditures
(total)
Pearson Correlation 1,000 ,150**
Sig. (2-tailed) ,000
Value of assets (total)
N 70929 68430
Pearson Correlation ,150**
1,000
Sig. (2-tailed) ,000
Operating expenditures (total)
N 68430 76776
**. Correlation is significant at the 0.01 level (2-tailed).
The two weighing factors open the opportunity to estimate a range for the calculated figures.
None of them is capable to determine the "true" value of a variable in the ground population
but the estimation of a possible range is probably more useful than to only look at the
aggregation of cases which are not representative.
The greatest advantage of the weighing factors is that the results get sensible for the size of
the respective foundation and therefore somehow compensates the lack of representativity of
the sample. Especially the weight of smaller foundations is heightened, despite that they are
systematically underrepresented in the sample. Due to the small number of foundations that
took part in the survey, there is a chance that below a certain level of (economic) weight,
foundations refused to answer completely. That would lead to an additional overestimation of
the sector, which in turn is unavoidable since we cannot introduce ghost-cases in the data set.