N ELSEVIER Journal of Economic Psychology 18 (1997) 179-199 [610618]

JOURNAL OF
N ELSEVIER Journal of Economic Psychology 18 (1997) 179-199
Values, beliefs and regional variations in new firm
formation rates
Per Davidsson *, Johan Wiklund
J~nki~ping International Business School (JIBS), J~nkiiping University, Box 1026, 551 11 Ji~nk~ping, Sweden
Received 16 January 1996; accepted 15 September 1996
Abstract
There are recurrent claims that culture, understood as prevailing values and beliefs, is an
important determinant of the level of entrepreneurship in a society. In spite of this, relatively few
empirical studies with this focus seem to have been carried out. While some studies of national
culture and the rate of economic development have been published and received wide readership,
empirical studies of values and beliefs in relation to variations in new firm formation rates are
lacking. In a previous study in that vein by one of the authors, it was found that possible cultural
and economic-structural determinants of the new firm formation rate were positively correlated, so
that the unique contribution of each type of explanation could not be determined. In the present
follow-up study, three matched pairs of regions are investigated. While the regions in each pair are
similar on economic-structural dimensions, one region in each pair has shown a higher and the
other region a lower rate of new firm formation than predicted by carefully developed regression
models that use economic-structural factors as explanatory variables. To determine whether
cultural differences can explain the deviations from the predictions based on economic-structural
variables, large samples of 35-40 years old inhabitants in each region were surveyed for cultural
values and beliefs data. The results of this study suggest that both values and beliefs of the kind
investigated do have an effect on regional new firm formation rates. The cultural variation is
small, however, and for contemporary Sweden it appears to be a relatively less important
determinant of new firm formation rates than are variations in economic-structural conditions.
* Corresponding author. E-mail: [anonimizat], Fax: +46 36 165069, Tel.: +46 36 156430.
0167-4870/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved.
PII SO 167 -4870(97 )00004-4

180 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
PsvclNFO classification: 2930; 3040
JEL classification: AI3; M13; RI2
Keywords: Values; Beliefs; Culture; Entrepreneurship; New firms
1. Introduction
Large-scale studies concerning regional environment determinants of new
firm formation rates have recently been carried out in several countries. These
studies, many of which were reported in a special issue of the Regional Studies
journal, have established that about 70 percent of the regional variation in
business start-up rates can be explained – at least in the statistical sense – by
differences in certain economic and socio-demographic characteristics of the
regions (Audretsch and Fritsch, 1994; Davidsson et al., 1994; Fritsch, 1992;
Garofoli, 1992, 1994; Guesnier, 1994; Hart and Gudgin, 1994; Keeble and
Walker, 1994; Reynolds, 1994). As regards the major determinants there is a
high degree of agreement in the results from various countries. The most
recurrent structural determinants of high new firm formation rates appear to be
(a) a high presence of small firms in the regional economy, (b) a large and/or
growing population, and (c) high or increasing unemployment (Reynolds et al.,
1994). Studies that have carried the analysis one step further have found support
for the notion that high rates of new firm formation are positively associated
with regional economic development (Davidsson et al., 1994; Reynolds, 1996;
Reynolds and Maki, 1990).
There is a widespread belief that, in addition to structural factors, cultural
variation is a powerful determinant of cross-regional or cross-national variation
in the 'supply' of entrepreneurship. There are different conceptions of culture.
One way to define culture is to associate it with various behaviours that are
prevalent within an area and passed on from one generation to the next. With
such a conception of culture the above mentioned studies do give some support
for an effect of culture on new firm formation rates, in that the perhaps most
important determinant is the current presence of small, owner-managed firms.
In this article, however, we employ another common conception of culture,
viz. that culture has to do with the mentality, e.g., values and beliefs, that prevail
in a society. The use of concepts like values and beliefs in psychological theory
and research is not entirely consistent. However, there seems to be agreement
that values are more abstract and global psychological evaluations that are
relatively distal to specific behaviours, whereas beliefs are more concrete
perceptions of attributes of objects or other phenomena (Eagly and Chaiken,

P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199 181
1993). This is also the kind of distinction we make here. On the operational
level, our measures of beliefs are based on items that specifically mention small
and new firms, whereas the value measures are not domain-specific.
Systematic empirical research on this conception of culture on the one hand,
and rates of new firm formation on the other, is scarce. With a broader
perspective on entrepreneurial behaviour in a society we have, of course, the
classical and debated attempts of Weber (1930) and McClelland (1961) at
large-scale sociological or macro-psychological explanations for economic de-
velopment at large. In short, Weber held that certain puritan aspects of the
Protestant (in particular Calvinist) moral code led to the striving for profit with
no other goal than re-investment, and hence capital accumulation, on the part of
the employers. At the same time, the same moral code led to high work
productivity on the part of the employees. McClelland, whose work builds on
Weber's (cf. McClelland, 1961, p. 47), held that the diffusion among the
population of a psychological complex of motives, the Need for Achievement,
determined variations in the pace of economic development across countries and
within a country over time. McClelland and his collaborators regularly found
empirical support for their hypothesis (McCleUand, 1965; McClelland and
Winter, 1969; Miron and McClelland, 1979). Across various vocational groups,
Bellu (1987) found differences in the expected direction between the North and
the South of Italy. For various reasons, others have been more reluctant to
accept McClelland's theory (cf. Brockhaus, 1982; Carsrud and Johnson, 1989;
Finison, 1976; Hofstede, 1991).
Hofstede (1991) suggested other but similar, i.e., mentality based, explana-
tions for the post-W.W.II development in East Asia. Shane (1992) investigated
the relationship between Hofstede's measures of national culture and inventive-
ness and found that individualistic and non-hierarchical societies were more
inventive than others. In an empirical comparison of more than forty countries
Lynn (1991) provided additional support for a relationship between certain
aspects of national culture (competitiveness and valuation of money) and
economic growth.
However, none of the above studies specifically related cultural variation to
new firm formation rates. The same is true for the Jackson and Brophy (1986)
study of presumably entrepreneurship-related attitudes in different regions in the
US. Other 'culture and entrepreneurship' studies more explicitly concern new
firm formation, but are delimited to the study of attitudes and motivations of
entrepreneurs in different geographical areas rather than among the general
population (Bellu et al., 1990; Guo, 1991; McGrath and MacMillan, 1992;
McGrath et al., 1992; Scheinberg and MacMillan, 1988; Shane et al., 1991).

182 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
A relationship between culture and new firm formation may in principle
occur for two reasons. Firstly, prevailing values and beliefs among others in an
individual's environment may make that individual more or less inclined to go
into business for him- or herself. This is the societal legitimation or supportive
environment perspective put forward by Etzioni (1987). This does not require a
relationship on the individual level, i.e., a relationship on the macro level may
occur whether or not it is those individuals who hold the most 'entrepreneurial'
values and beliefs who also found firms. Secondly, a relationship may occur
because some regions have a larger pool of potential entrepreneurs, mentality-
wise. This latter explanation, which seems to be in line with McClelland's
(1961) reasoning, assumes that it is specifically those individuals who hold more
'entrepreneurial' values and beliefs who also start firms.
While this research is not designed to be able to distinguish between the two
types of explanation, they are still important to our work. Firstly, they have a
bearing upon how the results should be interpreted. We will revert to this issue
in Section 5. Secondly, the latter kind of explanation suggests that we should
also consider value and belief differences that have been found in individual
level comparisons between business founders and other groups. Arguably, the
most recurrent psychological variables in this line of research are a higher need
for Autonomy and the above-mentioned Need for Achievement (cf. Brockhaus,
1982; Brockhaus and Horwitz, 1986; Johnson, 1990). From more general
psychological research, we may conclude that self-efficacy beliefs are likely to
be of importance (Bandura, 1986).
Except for a previous study conducted by one of the authors (Davidsson,
1993, 1995a; Davidsson and Delmar, 1992) we have not been able to track one
single study of the relationship between cultural variation and entrepreneurship,
i.e., a study which specifically compares the values and beliefs among the
general population in different spatial entities and relates the established differ-
ences to spatial variations in new firm formation rates. This is somewhat
surprising, given the interest in regional studies and the emphasis some promi-
nent scholars have given the cultural dimension in theoretical accounts of
environmental influences on entrepreneurship (Shapero and Sokol, 1982; Et-
zioni, 1987; Carsrud and Johnson, 1989).
A probable reason for the relative lack of studies in this area is the
methodological problems involved in such research. Data on culture are difficult
and costly to obtain and if countries were to be compared, language differences
make the validity of established differences uncertain. Structural conditions are
certain to vary considerably and may be difficult to control for. As regards the
dependent variable, data collection methods for new firm formation vary to such

P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199 183
extent, that valid comparison cannot be made. Many of these problems are
eliminated if cultural influences on new firm formation rates instead are studied
across regions within a country.
It could, of course, be argued that such a study would be superfluous given
the relative success of explanatory models based on economic and socio-demo-
graphic variables. However, the results of our previous study (Davidsson, 1993,
1995a; Davidsson and Delmar, 1992) strongly suggest that cultural determinants
of regional variations in entrepreneurship are worthy of further attention. In
summary, it was found that (a) regional variation in business start-up rates did
appear to be related to cultural variation, (b) however, across the (rather large)
spatial units used the cultural variation was small, (c) while regional differences
in more generic values apparently influenced regional start-up rates, differences
as concerns specific beliefs about entrepreneurs and owner-managed firms are
not influential, and (d) structural and cultural factors that appear to influence
regional levels of new fu-m formation were positively correlated, so that no solid
conclusion could be drawn regarding the unique impact of either type of factor.
The finding that values are more important than beliefs is somewhat surprising,
since it is generally concluded in psychological research that more proximal
variables, such as domain-specific beliefs, should have higher explanatory
power with regard to specific behaviours than have distal variables such as
values. However, it is not a given that this generally holds true on an aggregate
level. The result is also of practical significance since values are likely to be
more difficult to affect than are beliefs. The last finding, that cultural and
structural determinants are correlated, is a very important one. It suggests that to
the extent that cultural variation is the real cause for variations in new firm
formation rates, studies that use only structural (i.e., economic and socio-demo-
graphic) explanatory variables may exaggerate the influence of the latter. If so,
the theoretical interpretations, as well as the policy prescriptions that are derived
from such studies, may be at least partly wrong.
2. Research questions
Based on the theoretical considerations reviewed above, and the results of our
previous study, we will focus on the following research questions:
1. Can we find evidence in support of the notion that cultural regional variation
is associated with regional rates of new firm formation, after structural
factors have been controlled for?
2. Can it be confirmed that values are more important than differences concern-
ing beliefs about entrepreneurs and owner-managed firms?

184 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
3. Method
3.1. Selection of regions
In the previous study the cultural variation was found to be small. An
artifactual reason for this might be that some of the regional variations cancelled
out due to the employed definition of region types. Wc therefore decided to use
Labour Market Areas (LMAs), rather than region types as our spatial unit of
analysis in this follow-up study. LMAs are smaller and likely to be internally
more homogeneous than the region types we used in the previous study.
In order to answer the research questions we need region level data on
structural characteristics, new firm formation rates, and culture. The first two
types of data were obtained from a related research project (Davidsson et al.,
1994). In that project, the country was divided into 80 LMAs and regression
analysis was employed which identified seven structural characteristics as
influencing the start-up rate. The explanatory variables were (sign of relation-
ship within parentheses): current small firm density (+), absolute population
( + ), population density ( + ), previous period population growth ( + ), current
unemployment level ( + ), previous period unemployment trend ( – ), and public
business support expenditure per capita ( + ). The model explained 68 percent of
the variation in new firm formation rates, computed as average annual number
of new firms per 1000 working age population.
For the present study, the 80 regions were grouped into eight more homoge-
Table 1
The selected regions and their rates of new firm formation
Matched region pair Annual new firm formation rate a Regression residual b
Kristianstad LMA 10.0 + 0.8
vs.
Karlskrona LMA 7.8 – 1.0
Katrineholm LMA 9.4 + 1.5
vs.
NykiSping LMA 7.5 – 0.9
K/Sping LMA 8.6 + 0.5
vs.
Karlskoga LMA 6.3 – 1.6
a Average annual no. of new single-establishment firms per 1000 inhabitants in the 16-64 years age bracket
during the 1985-9 period. Across the country's 80 regions, this variable has a minimum of 5.7, a maximum of
14.8, mean and median both 8.8, and a standard deviation of 1.4.
b This figure shows how much higher or lower the region's new fu'm formation rate is compared to what is
'normal' for a region with the given characteristics, according to Davidsson et al.'s (1994) regression results.

P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199 185
neous categories with respect to these seven variables, by means of cluster
analysis. Through this procedure, three structurally matched pairs were distilled
for further investigation, i.e., the regions in each pair belong to the same cluster.
In each structurally matched pair one region has a higher, and the other a lower
new firm formation rate than predicted by the regression model. While perfect
matching across seven variables is not possible to attain, the regions' positive
and negative residuals need another explanation than these variables, under the
assumption that the linear model specification is correct. Cultural variation,
which was not included in the regression model, is a plausible candidate for
explanation of the regions' higher- and lower-than-predicted rates of new firm
formation. In addition to the requirement that the regions in each pair belong to
the same cluster they are located in the same part of the country. Hence, the
control of potential, unwanted influences was further increased. Details on the
matching are given in Appendix A. The regions and data on their new firm
formation rates are specified in Table 1.
3.2. Data on culture
In order to obtain data on cultural variables a mail survey was directed to
random samples of 35-40 years old inhabitants in each region. This age group
was considered to be the most relevant with respect to new firm formation. A
narrow age span was chosen in order to avoid that sample differences in the age
distribution distort the results. Descriptive data on the samples are given in
Table 2. The response rates are high and reasonably even, while the gender and
entrepreneur representations vary. We will revert to the latter issue later in
Section 3.2. The survey was carried out in February-March, 1994.
Table 2
Sample characteristics
Region (LMA) Sample No. of useful Response rate Percent males Percent with
size responses entrepreneurial
experience a
Kristianstad 300 230 76.7% 49.6% 21.3%
Karlskrona 300 216 72.0% 51.4% 9.7%
Katrineholm 300 220 73.3% 51.4% 21.3%
NykiSping 300 219 73.0% 43.8% 17.4%
KiSping 300 212 70.7% 55.5 % 19.3%
Karlskoga 300 216 72.0% 50.0% 13.4%
Total 1800 1313 72.9% 50.2% 17.3%
a Includes full-time and part-time business owner-managers as well as ex-entrepmneurs.

186 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
Apart from personal background questions, the questionnaires included ques-
tions concerning values and beliefs that could have an effect on entrepreneurial
behaviour. The values part of the questionnaire comprised items reflecting
change-orientation, achievement motivation, need for autonomy, Jante-mentality
(a Scandinavian breed of cultural repression – "you shouldn't think you are
someone"), acceptance of capitalism, competitiveness, and valuation of money.
This selection of value dimensions was based on theories and empirical results
concerning national culture and economic development, in-depth studies of
communities considered to be more and less entrepreneurial, and research results
concerning psychological characteristics of entrepreneurs (cf. Davidsson, 1993,
ch. 2 and pp. 38-40). Each value dimension was measured by an index based on
four or five 4-point agree-disagree statements. The former five indices were
developed in our previous study, while the latter two were the strongest
predictors of economic growth rate in Lynn's (1991) study. They were translated
and validated by Lynn's Swedish collaborators (Ekehammar and Niemenmaa,
1992).
As mentioned already, the items used for measuring values make no mention
of entrepreneurs, business start-ups, owner-managed f'Lrrns, and the like. Such
items were used instead for the beliefs part, which concerned the societal
contribution of entrepreneurs as well as beliefs concerning entrepreneurs' gains
and sacrifices in terms of money, workload, risk-taking, and social status. The
risk-taking index was developed for this study while the others were used also in
the previous study. A further addition in the present study are two more
self-oriented beliefs concerning personal know-how about where to turn for
finance and advice if one had an idea for a business start-up, and whether the
respondent perceives the idea of going into business for oneself as something
'for me'. Both of these can be regarded as aspects of self-efficacy (Bandura,
1986). Otherwise, the selection of specific beliefs to include was guided by
common sense reasoning rather than established theories or previous research
findings.
In order to make maximum use of the available data, mean substitution was
employed for internal non-response for individual items in the values and beliefs
indices (1 to 28 cases out of 1300 + ). The values and beliefs indices are further
described in Table 3.
As may be noted we have been fairly permissive, accepting indices with
Cronbach's alpha values down to 0.50. This accords with the Nunnally (1967)
recommended lower limit for exploratory research. It may be debated if this
research is to be called exploratory. However, the risk we run by using indices
with questionable reliability is to underestimate, not exaggerate, real relation-

P. Davidsson, J. Wiklund/Journal of Economic Psychology 18 (1997) 179-199 187
E
E
r~
o • – Z.~
.N
[.. :~:~:~:~:~:=~~ :~:~:=~:=~ ~ ~ ~ ~ = ~ :~
.=. J::
"~o'~ ~o =
.N
o ,. =.=
~ ,~ Z ~-~
~<~> "o
~ 0 0 "I
Q _=
E >
t- t-
m 0
N "
~.-~
K t-
O C ,.~ .~,
~ m
°I
E~ ~-~
._= .~
~2 *" 0
~._=
~-~

188 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
ships. Two planned indices were excluded from the analysis because of Cron-
bach's alpha's lower than 0.50. A factor analysis was performed which largely
confh-rned the presumed dimensionality of the pre-specified indices. An addi-
tional reason for being permissive with the Cronbach's alpha criterion is that
since the analysis is conducted on the regional level individual level correlations
may not be relevant. For instance, Hofstede's widely used indices are based on
country level, not individual level inter-correlations (Hofstede, 1991, pp. 142-
143). Nevertheless, the indices yield statistically significant, and logically
reasonable differences when males and females are contrasted, or when en-
trepreneurs are compared with non-entrepreneurs. Analyses on the individual
level also disclose reasonably strong relationship between values and beliefs on
one hand, and entrepreneurial intentions on the other (Davidsson, 1995b).
3.3. Data analysis
The presented analyses are straightforward. Within each matched region pair,
arithmetic mean scores are compared for one index at a time. Multivariate
analysis would of course be preferable, but with six regions and fourteen
explanatory variables this is not possible. As a weak substitute, we perform sign
tests for the total number of region differences that run in the expected direction
compared with the number of differences in the opposite direction plus ties.
The absolute values of the index scores are arbitrary in the sense that they are
affected by the scoring system and the number of items included. For ease of
comparison and interpretation, differences concerning standardised versions of
the indices will be reported (i.e., zero mean; unity standard deviation). For
standardised variables, mean differences of 1.0 can, as a rule-of-thumb, be
considered large. Differences of 0.50 are still substantial, while differences
around 0.25 are small but non-negligible (Cohen, 1962). We regard these
rules-of-thumb concerning effect size as more important than significance levels,
since we have no basis for weighing the possibility of type I errors against
possible type II errors. In the absence of other information on the cultural
differences between the regions, we regard our point estimates (rather than a
zero difference) as our best guess regardless of the level of significance. The
certainty of these point estimates are, of course, contingent on the sample sizes
and the indices' variances. To give an indication of statistical certainty we report
exact associated probabilities rather than a particular cut-off level for statistical
significance.
Throughout, the indices are computed so that a higher value is expected for
the region with high new firm formation. The hypotheses thus are uni-direc-
tional, and the reported associated probabilities have been computed accord-

P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199 189
ingly. In order to avoid effects of sampling differences in gender representation,
the analyses are weighted for equal gender representation. As entrepreneurs are
part of the general population, their responses should be included. However, if
random sampling error makes the proportion of individuals with entrepreneurial
experience in the samples non-representative for the regions, the results may be
distorted. Therefore, analyses including and excluding such individuals are
reported.
It should be noted that while culture data were collected in 1994, the data on
new fhm formation rates concern the 1985-9 period. This means that any
interpretation involving that cultural variation causes variations in new firm
formation require that regional differences in culture and new firm formation be
fairly stable over medium range periods of time. This is a plausible assumption.
We know, for example, that on the county level (24 counties), the correlation
between 1988 and 1994 start-up rates in relation to the stock of fLrrns is
reasonably high, 0.64 (SCB, 1990, 1995). As an additional check that the
regions' deviating patterns in terms of new firm formation during the 1985-9
period were not random or temporary phenomena, we have compared the
regions in each pair concerning 1994 questionnaire data on the respondents
intentions to go into business for themselves. In all three pairs the respondents
from the region assumed to be more entrepreneurial show more entrepreneurial
intentions. We can therefore consider the results as support for the assumption
that regional differences in new firm formation rates are fairly stable over
medium range periods of time. After this research was designed and after the
first version of this manuscript was prepared, follow-up data on the regions' new
firm formation rates during the 1990-93 period have become available. In order
to avoid confusion, we postpone the discussion of these results to Section 5.
4. Results
4.1. Results concerning values
The results concerning values are displayed in Table 4. The matched pairs are
presented with the region with high new firm formation rate appearing first, i.e.,
Kristianstad, Katrineholm, and KiSping, respectively. The first row of figures for
each value index refers to the analysis including all respondents, while all
italicised figures refer to the analysis excluding respondents with entrepreneurial
experience. The first column in each match displays the standardised mean
difference between the two regions (cf. Section 3). When the difference is in the
expected direction, a positive value appears. The second column displays the

190 P. Davidsson, J. Wiklund/Journal of Economic Psychology 18 (1997) 179-199
Table 4
Matched pair comparisons of value index means
Values I st Match 2nd Match 3rd Match
Kristianstad, n = 230/172 Katrineholm, n = 220/173 K/Sping, n = 212/170
VS. VS. VS.
Karlskrona, n = 216/193 Nyktiping, n = 219/180 Karlskoga, n = 216/184
Standardised Assoc. Standardised Assoc, Standardised Assoc.
difference probabil, difference probabil, difference probabil.
Change- 0.16 0.05 0.13 0.08 – 0.04 –
orientation 0.19 0.04 0.21 0.02 – 0.02 –
Need for 0.08 0.18 – 0.11 – 0.00 –
achievem. 0.04 0.34 – 0.02 – 0.09 0.22
Need for 0.19 0.02 0.13 0.10 0.05 0.29
autonomy 0.21 0.02 0.14 0.11 0.05 0.33
Jante- – 0.07 – 0.01 0.44 – 0.07 –
mentality 0.05 0.32 – 0.01 – – 0.12 –
Acceptance 0.17 0.04 – 0.20 – 0.00 –
of capitalism 0.20 0.03 – 0.30 – 0.04 0.36
Competitiveness 0.16 0.05 0.07 0.22 0.02 0.43
0.14 0.09 0.11 0.15 0.04 0.34
Valuation 0.03 0.37 0.00 – 0.10 0.16
money 0.04 0.37 0.03 0.36 0.17 0.07
Italicised figures refer to the analysis excluding respondents with entrepreneurial experience. The associale, d
probabilities refer to uni-directional tests.
associated probability, or 'absolute significance level'. That is, values of 0.05 or
smaller translate to the conventional 'significant at 5 percent risk level.' The
associated probabilities refer to uni-directional tests. Hence, they indicate the
probability that a difference of this size or larger in this direction be obtained, if
no difference exists between the underlying populations.
A f'n-st observation is that all differences are small (cf. the rule-of-thumb
given in the Section 3). Few of the differences are statistically significant at
conventional levels. Further, for only two of the seven indices do all tests
uniformly yield positive differences and in neither of these cases are all
differences of a satisfactory magnitude. It can thus hardly be argued that the
analysis points out a particular set of values as consistently influential with
regard to regional rates of new firm formation. As regards Need for Achieve-
ment and Jante-mentality the analysis does not provide any support at all for the
notion that they, regarded as cultural variables, have the expected effects.
However, the overall results support the notion that values of the investigated

P. Davidsson, J, Wiklund / Journal of Economic Psychology 18 (1997) 179-199 191
kind do have an effect. The distribution is clearly biased towards positive
differences rather than negative ones. In the analysis including entrepreneurs we
find 13 positive differences, three ties and five negative differences. In the
analysis excluding entrepreneurs these figures become 16, 0, and 5, respectively.
According to simple sign-tests, such distributions are unlikely to be randomly
generated (p < 0.05 in both cases). The most reasonable interpretation of the
results therefore is that regional differences concerning values of the studied
kind among the general population in relevant age categories do have an effect
on regional rates of new firm formation. It is also reasonable, however, to
assume that these effects are marginal and that the cultural differences estab-
lished here may not be the sole explanation for the positive and negative
deviations from the expected levels of new firm formation.
4.2. Results concerning beliefs
The results concerning beliefs are very similar to those concerning values (see
Table 5). Again, while most differences are in the expected direction, no
differences are large and few reach statistical significance at conventional levels.
For three belief indices, Social status, Workload, and 'For-me', the differences
are uniformly positive across all comparisons. However, a difference of satisfac-
tory magnitude and statistical certainty does not appear consistently for any
variable, even if the 'For-me' dimension comes close to that. For one index,
Financial pay-off, there is no support for an influence of the belief on new firm
formation. This is in line with the recurrent result in Scandinavian studies of
motivations of business founders that expectations of monetary gain generally
rank low among stated reasons for founding one's own firm (cf. Scheinberg and
MacMillan, 1988).
Despite the relatively weak separate differences, the overall results in terms
of the sign of the differences clearly support the notion that regional differences
regarding the kind of beliefs investigated are associated with differences in
regional rates of new firrn formation. In the analysis of all respondents, the
region with high new fn'm formation has a higher value for entrepreneurial
beliefs in 16 cases out of 21. Four differences run in the opposite direction while
one comparison yields a tie. In the analysis excluding entrepreneurs we find
eighteen positive differences, one tie, and two negative ones. Again, such
distributions are unlikely to be randomly generated (p < 0.01 in both cases;
sign-test). Therefore, unlike the previous study (Davidsson, 1993, 1995a;
Davidsson and Delmar, 1992) these results are in line with the assumption that
the relative prevalence of this kind of beliefs do affect regional levels of new
firm formation, although this influence may be marginal.

192 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
Table 5
Matched pair comparisons of belief index means
Beliefs 1st Match 2nd Match 3rd Match
Kristianstad, n = 230/172 Katrineholm, n = 220/173 KiSping, n = 212/170
VS. VS, VS.
Karlskrona, n = 216/193 Nyk6ping, n = 219/180 Karlskoga, n = 216/184
Standardised Associated Standardised Assoc. Standardised Assoc.
difference probability difference probabil, difference probabil.
Societal 0.24 0.01 0.07 0.22 0.00 –
contribution 0.26 0.01 – 0.03 – 0.05 0.33
Financial – 0.02 – – 0.07 – – 0.03 –
pay-off 0.10 0.19 0.03 0.36 – 0.05 –
Perceived – 0.03 – 0.07 0.23 0.06 0.25
risk 0.03 0.40 O. 15 0.08 O. 13 O. 10
Social status 0.04 0.35 0.08 0.21 0.14 0.08
0.09 0.21 0.03 0.37 0.13 0.11
Workload 0.08 0.20 0.01 0.44 0.11 0.11
0.08 0.23 0.08 0.23 O. 15 0.07
Know-how 0.16 0.05 0.06 0.28 0.08 0.20
0.14 0.10 0.09 0.20 0.00 –
'For-me' 0.26 0.003 0.13 0.08 0.14 0.08
0.15 0.07 0.17 0.06 0.16 0.06
Italicised figures refer to the analysis excluding respondents with entrepreneurial experience. The associated
probabilities refer to uni-directional tests.
4.3. Summary of results
One rationale for the design of this study was the suspicion that with
relatively small spatial units, we would be able to establish larger cultural
variation than was the case in the previous study. This was not borne out. The
differences that were established were all small in magnitude. For those indices
that were used in both studies, the range of variation across all six regions in this
study was narrower for each index, than was the corresponding range of
variation across the six region types in our previous study (Davidsson, 1993,
1995a; Davidsson and Delmar, 1992). Indirectly this supports the finding from
the first study, that cultural and (some) structural determinants of new firm
formation rates are positively correlated. In this study we have deliberately
restricted the structural variation. Consequently, we find less cultural variation.
Our first research question concerned whether we can find evidence support-
ing the notion that culture has its own influence on new firm formation rates.

P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199 193
Our conclusion is that the most reasonable interpretation of the results is that
cultural variation does affect the new firm formation rates. With such small
cultural variation, cultural factors of the kind investigated would, however, be
likely to explain less of the total variation in new firm formation across a larger
number of regions than the share that previous studies have shown can be
accounted for by structural factors such as current small firm density or
population growth.
While the tentative conclusion from the previous study was that values
influence business start-ups while beliefs do not, the results of the present study
do not confirm this pattern. Our answer to the second research question thus is
that the results of this study suggest that both sets of factors have small effects
in the expected direction. With regard to particular values or beliefs, we are
unable to point out specific aspects of regional culture which consistently appear
as a determinant of the new firm formation rate.
5. Discussion
5.1. An overall interpretation of the results
Is the supply of entrepreneurship in a society contingent on the values and
beliefs that are prevailing? Of course it is! Nevertheless, our results give only
limited support for this idea. While other research has demonstrated that
demographic and economic-structural factors can explain a major share of the
regional variations in new firm formation rates, cultural differences appear at
best to be another contributing, although marginal factor. Then, why could no
strong relationships between culture and entrepreneurship be established in this
study? Let us examine some possible explanations.
The first explanation would be that the measurement of cultural variations is
inadequate, even if the limitation to 'mental software' is accepted. This objec-
tion can be raised at several different levels. Firstly, it could be claimed that it is
generally impossible to measure aspects of culture by means of questionnaires.
The frequent use of similar approaches in social science research and – not least
– commercial activities is but one of several possible rebuttals to such an
objection. It should be admitted, however, that frequent does not equal valid and
that no survey study could give more than a crude and over-simplified image of
a culture. If the approach as such be accepted it can still be claimed that the
study did not capture the right aspects of culture. However, a comprehensive
search for presumably relevant values and beliefs was conducted prior to the
first study, and taken together the two studies do cover a fairly broad set of

194 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
values and beliefs. If this is accepted it can still be argued that some of the
theoretical variables included in the study have an influence on new firm
formation, but this does not come through in the results because of measurement
error. This point is relevant in the sense that several of the indices have lower
reliability coefficients than could be wished for. However, as was mentioned in
Section 3, the indices yield meaningful relationships in individual level analyses,
which gives reason to believe that they do capture psychologic-cultural realities.
In summary, measurement problems do not seem to be the main reason why
strong relationships between culture and entrepreneurship could not be estab-
lished in this study.
A second reason for the relatively weak relationships has to do with the
selection of regions. We selected pairs of regions that during one period
deviated in opposite directions from their predicted levels of new firm forma-
tion. The reason for their deviations may in part have been a result of stochastic
variation. If so, the difference between the regions in the dependent variable
may be exaggerated and likely to be smaller in the following period. This would
be a case of the 'regression to the mean effect' that occurs for any follow-up of
members of extreme categories on a scale. After this study was carded out, and
after the first version of this manuscript was written, Davidsson et al. (1996)
conducted a follow-up study of new firm formation rates in Swedish LMAs
during the 1990-93 period. Their data show that the firm start-up rate difference
between the 'high' and 'low' region in each pair was indeed smaller during the
latter period. This supports the suspicion that the region differences in the
previous period were in part stochastic. However, for each pair the difference is
still in the same direction, and the relation between their regression residuals
remain the same as those reported in Table 1. Therefore some unexplained
variation that is potentially attributable to cultural differences also remains.
The third explanation for the relatively weak relationships would be that a
region's rate of new firm formation is too crude an indicator of that region's
degree of entrepreneurship. We have refrained from using the terms 'new firm
formation' and 'entrepreneurship' interchangeably. However, our choice of
cultural explanatory variables has been guided by an implicit assumption that
firms are founded for 'entrepreneurial' reasons. Such is not always the case. In
recessions and in deprived areas, going into business for oneself is often a
choice of last resort. In many other cases the founder has grown up in a family
of business owners, and founding his/her own business has therefore become
the most natural career choice this individual could possibly make. It can be
argued that both of those alternatives are 'less entrepreneurial' than when the
prime motivation is the wish to exploit an attractive market opportunity. Quite

P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199 195
certainly, the distribution of different start-up motivations, and the sectoral split
of start-ups between infant and mature industries, is not the same across regions.
The overall start-up rate may therefore not satisfactorily reflect the theoretical
dependent variable we are trying to capture.
It is unlikely, however, that any other measure of the regions' level of
entrepreneurship would have yielded distinctly different results. The reason for
this is also the fourth explanation for the relatively weak relationships that were
obtained, viz. that the cultural variation is very limited. The country appears to
be culturally homogeneous 'on average.' A constant cannot explain variations.
Therefore, the main reason for the relatively weak relationship between culture
and new firm formation seems to be the relative lack of cultural variation.
But then again, how strong relationships could we reasonably expect? And
how much cultural variations would it take to produce the deviating patterns of
new firm formation that the regions have shown? These are important questions.
Let us focus on one particular difference, viz. the 0.19 higher standardised Need
for autonomy index value that Kristianstad has compared with Karlskrona. The
original variable is an index were the answers to four 4-point agree-disagree
statements have been summed. The range of the scale thus is 4-16, and it turns
out that Kristianstad's mean is 0.38 higher than is Karlskrona's. All it takes to
produce it is that we can pick 20 people (out of 200 + ) from each sample in
such a way that the 20 'Kristianstaders' express strong agreement (4) with each
statement while the 20 'Karlskroners' only express moderate agreement (3),
while the remainder of the samples have identical mean values.
Now, could this apparently tiny difference have any effect on the regions'
new firm formation rates? The answer to that question is partly contingent on
the nature of the culture ~ entrepreneurship relationship. As was mentioned in
Section 1, we may assume that everybody in a region would be a little more
inclined towards entrepreneurship (regardless of his/her own values and beliefs)
because the favourable values and beliefs of others in the region create a
supportive environment for entrepreneurship. We would argue that for this
mechanism to have a measurable effect, larger cultural differences than the one
described above would be needed. Only then could they become noticeable for
individuals as a characteristic of their general environment. Alternatively, the
higher value on the cultural variable may be interpreted as showing that the
region has a larger pool of potential entrepreneurs. We then assume that it is
specifically those individuals who have the more entrepreneurial values and
beliefs who found firms, i.e., an aggregated version of the 'psychological
characteristics' explanation for individuals' entrepreneurial behaviour. Remem-
ber that what we are trying to explain is that Kristianstad has had about one too

196 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
many and Karlskrona one too few start-ups annually per one thousand 16-64-
year-old inhabitants. We have just shown that for the autonomy dimension
alone, Kristianstad has in the 35-40 age bracket an 'excess capacity' of about
one hundred somewhat more entrepreneurially inclined people per one thou-
sand, compared with Karlskrona. Viewed from this perspective, the difference
does not appear marginal.
Our preferred interpretation of the results is that the cultural differences are
minor and that their effects are likely to be small in comparison to the effects of
some structural factors. The above exercise, however, illustrates a problem with
all attempts at explaining minority phenomena such as the number of business
start-ups in relation to the size of the population. Maybe we are often looking
for unrealistically large effects; perhaps do we often disregard 'small' effects
that are important results and of the largest magnitude that could reasonably be
expected.
5.2. Implications for policy-making and research
The major policy implication of this study and the previous study is that
given the limited cultural variation, there seems to be little need in contemporary
Sweden for customising policies regionally because of cultural variation of the
investigated kinds. It must be stressed, however, that the situation may be totally
different in other countries. It should further be noted that the results in no way
rule out the possibility that cultural variation over time may be substantial for
the country as a whole, and that this may have a substantial impact on temporal
variations in entrepreneurial activity.
A research implication is that the general research approach is viable, i.e., it
is possible to get decent response rates and response quality and thus possible to
measure cultural variation by means of questionnaires directed to samples of the
general population. It is difficult, however, to obtain perfect matching for
important structural characteristics of regions. The alternative, to obtain both
structural and cultural information on a large enough sample of regions to make
multivariate analyses possible may be difficult to realise because of the high
costs for collecting data on culture. Given the limited cultural variation in
Sweden the following extensions of this research may be fruitful. Firstly,
cultural variations across countries could be studied. While this most likely
would capture more variability, such an approach would add problems related to
questionnaire translation and different data collection methods for new firm
formation, as has been discussed in the introduction. Secondly, the approach
could be replicated within other countries that are less culturally homogeneous.
Thirdly, the approach could be used longitudinally rather than for cross-section

P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199 197
comparisons. In all of these extensions, it would be beneficial to collect more
detailed data on the composition – in terms of sectors and preferably also in
terms of start-up motivations – of the total number of new In'ms. Of course,
measuring the dependent variable some time after the cultural data have been
collected would be preferable in all the above cases.
Finally, since this research has focused mean differences across regions, it
can be accused of emphasising too much the 'mainstream' or 'average' culture
of the regions. Issues of cultural variability within regions, the prevalence of
certain sub-cultures et cetera may be equally or even more important. Paying
more attention to these issues may prove fruitful in future attempts at isolating
the role of cultural differences for regional variations in entrepreneurship.
Acknowledgements
This research was funded by the Swedish Foundation for Small Business
Research. We are also grateful to Professor Leif Lindmark and Professor
Christer Olofsson for their permission to use the data on new firm formation and
structural characteristics of regions for this study.
Appendix A
Table 6 presents details of the structural matching of the investigated region
pairs (standardised deviation from the mean of all 80 LMAs).
Table 6
Structural variable Pair 1 Pair 2 Pair 3
Kristianstad Kariskrona Katrineholm Nyk~ping K~ping Karlskoga
Small firm density 0.15 -0.88 – 1.07 -0.92 – 1.05 – 1.18
Absolute population 0.65 0.71 0.08 0.32 – 0.08 – 0.03
Population density 0.75 0.66 0.03 0.29 0.47 0.43
Population growth 0.60 0.06 – 0.43 0.43 – 0.84 – 1.05
Unemployment level – 0.26 0.44 – 0.76 – 0.95 0.14 0.51
Unemployment trend 1.00 0.29 -1.10 -0.64 -0.59 -0.10
Support expenditure – 0.53 0.60 – 0.70 0.12 – 0.44 – 0.22
As mentioned in the main body of the text, perfect matching across seven variables is not possible to attain.
While the regions in each pair are members of the same cluster they may differ substantially on individual
variables. However, under the assumption that the linear specification of the regression model is correct these
differences have already been accounted for and cannot explain the regions' deviations from predicted levels
of new firm formation.

198 P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199
References
Audretsch, D.B. and M. Fritsch, 1994, The geography of firm births in Germany. Regional Studies 28,
359-366.
Bandura, A., 1986. The Social Foundations of Thought and Action. Englewood Cliffs, NJ: Prentice-Hall.
Bellu, R.R., 1987. A behavioral explanation of the differential economic development of the center-north and
mezzogiorno of Italy. Working Paper, Kingsborough College of the City University of New York.
Betlu, R.R., P. Davidsson and C. Goldfarb, 1990. Toward a theory of entrepreneurial behavior: Empirical
evidence from Israel, Italy and Sweden. Entreprenenrship and Regional Development 2, 195-209.
Brockhaus, R.H. Sr., 1982. 'The psychology of entrepreneurship'. In: C.A. Kent, D.L. Sexton and K.H.
Vesper (Eds.), The Encyclopedia of Entrepreneurship. Englewood Cliffs, NJ: Prentice-Hall.
Brockhaus R.H. Sr. and P.S. Horwitz, 1986. 'The psychology of entrepreneurship'. In: D.L. Sexton and R.W.
Smilor (Eds.), The Art and Science of Entrepreneurship, Cambridge, MA: Ballinger/Englewood Cliffs,
NJ: Prentice-Hall.
Carsrud, A.L. and R.W. Johnson, 1989. Entrepreneurship: A social psychological perspective. Entrepreneur-
ship and Regional Development 1, 21-31.
Cohen, J., 1962. The statistical power of abnormal social psychological research: A review. Journal of
Abnormal and Social Psychology 69, 145-153.
Davidsson, P., 1993. Kultur och entrepren~rskap [Culture and Entrepreneurship]. Stockholm: NUTEK.
Davidsson, P., 1995a. Culture, structure and regional levels of entrepreneurship. Entrepreneurship and
Regional Development 7, 41-62.
Davidsson, P., 1995b. On the formation of entrepreneurial intentions. Paper presented at IAREP's 20th Annual
Colloquium in Economic Psychology, Bergen, Norway, Aug. 2-5.
Davidsson, P. and F. Delmar, 1992. 'Cultural values and entrepreneurship'. In: Frontiers of Entrepreneurship
Research 1992. Wellesley, MA: Babson College.
Davidsson, P., L. Lindmark and C. Olofsson, 1994. New firm formation and regional development in Sweden.
Regional Studies 28, 395-410.
Davidsson, P., L. Lindmark and C. Olofsson, 1996. NS_ringslivsdynamik under 90-talet [Business Dynamics
during the Nineties]. Stockholm: NUTEK.
Eagly, A.H. and S. Chaiken, 1993. The Psychology of Attitudes. New York: Harcourt, Brace Jovanovich.
Ekehammar, B. and P. Niemenmaa, 1992. Attitudes to work and money: Analyses of various scales applied to
a Swedish sample (mimeo).
Etzioni, A., 1987. Entrepreneurship, adaptation and legitimation. Journal of Economic Behavior and Organiza-
tion 8, 175-189.
Finison, L.J., 1976. The application of McClelland's national development to recent data. Journal of Social
Psychology 98, 55-59.
Fritsch, M., 1992. Regional differences in new firm formation: Evidence from West Germany. Regional
Studies 26, 233-241.
Garofoli, G., 1992. New firm formation and local development: the Italian experience. Entrepreneurship and
Regional Development 4, 101-125.
Garofoli, G., 1994. New firm formation and regional development: the Italian case. Regional Studies 28,
381-394.
Guo, M., 1991. Attitudinal and motivational characteristics of Chinese entrepreneurs and a comparison
between Chinese and American entrepreneurs (attitudinal characteristics). Dissertation Abstracts.
Guesnier, B., 1994. Regional variations in new firm formation in France. Regional Studies 28, 347-358.
Hart, M. and G. Gudgin, 1994. Spatial variations in new firm formation in the Republic of Ireland,
1980-1990. Regional Studies 28, 367-380.
Hofstede, G., 1991. Organisationer och kulturer [Organisations and Cultures]. Lund: Studentlitteratur.
Jackson, J.E. and D.J. Brophy, 1986. 'The environment for entrepreneurship'. In: Frontiers of Entrepreneur-
ship Research 1986. Wellesley, MA: Babson College.

P. Davidsson, J. Wiklund / Journal of Economic Psychology 18 (1997) 179-199 199
Johnson, B.R., 1990. Toward a multidimensional model of entrepreneurship: The case of achievement
motivation and the entrepreneur. Entrepreneurship Theory and Practice, Fall, 39-54.
Keeble, D. and S. Walker, 1994. New firms, small firms and dead firms: Spatial patterns and determinants in
the United Kingdom. Regional Studies 28, 411-428.
Lynn, R., 1991. The Secret of the Miracle Economy. Different National Attitudes to Competitiveness and
Money. London: The Social Affairs Unit.
McClelland, D.C., 1961. The Achieving Society. Princeton, NJ: Van Nostrand.
McClelland, D.C., 1965. n Achievement and entrepreneurship: A longitudinal study. Journal of Personality and
Social Psychology 1,389-392.
McClelland, D.C. and D.G. Winter, 1969. Motivating Economic Achievement, New York: The Free Press.
McGrath, R., I.C. MacMillan and S. Scheinberg, 1992. Elitists, risk-takers, and rugged individualists? An
exploratory analysis of cultural differences between entrepreneurs and non-entrepreneurs. Journal of
Business Venturing 7, 115-135.
McGrath, R. and I.C. MacMillan, 1992. More like each other than anyone else? A cross-cultural study of
entrepreneurial perceptions. Journal of Business Venturing 7, 419-429.
Miron, D. and D.C. McClelland, 1979. The impact of achievement motivation training on small business.
California Management Review 21, 13-28.
Nunnally, J.C., 1967. Psychometric Theory. New York: McGraw-Hill.
Reynolds, P.D., 1994. Autonomous fn'm dynamics and economic growth in the United States, 1986-1990.
Regional Studies 28, 429-442.
Reynolds, P.D., 1996. Business volatility: Source or symptom of economic growth. Paper presented at the
workshop Entrepreneurship, SMEs, and the Macro Economy, JtJnkiSping, Sweden, June 13-15.
Reynolds, P.D., D.J. Storey and P. Westhead, 1994. Cross-national comparisons of the variation in new firm
formation rates. Regional Studies 28, 443-456.
Reynolds, P.D. and W.R. Maki, 1990. Business volatility and economic growth. Final project report submitted
to the US Small Business Administration.
SCB, 1990. Nyf'rretagandet i Sverige 1988 och 1989 [New firms in Sweden 1988 and 1989] Statististiska
Meddelanden, F15 SM 9001. Orebro: SCB Publication Service.
SCB, 1995. Nyf'rretagandet i Sverige 1993 och 1994 [New firms in Sweden 1993 and 1994] Statististiska
Meddelanden, F15 SM 9501. Orebro: SCB Publication Service.
Shane, S.A., 1992. Why do some societies invent more than others? Journal of Business Venturing 7, 29-46.
Shane, S.A., L. Kolvereid and P. Westhead, 1991. An exploratory examination of the reasons leading to new
firm formation across country and gender. Journal of Business Venturing 6, 431-446
Shapero, A. and L. Sokol, 1982. 'The social dimension of entrepreneurship'. In: C.A. Kent, D.L. Sexton and
K.H. Vesper (Eds.), The Encyclopedia of Entrepreneurship. Englewood Cliffs, NJ: Prentice-Hall.
Scheinberg, S. and I.C. MacMillan, 1988. 'An 11 country study of motivations to start a business'. In:
Frontiers of Entrepreneurship Research 1988. Wellesley, MA: Babson College.
Weber, M., 1930. The Protestant Ethic and the Spirit of Capitalism. New York: Scribner's.