transform of population size, leading to a more symmetrical distribution was used
(LGPOPUL).
2- Urbanization/agglomeration. With this variable, it is intended to define if there is a
relationship between change in firm population and urban areas due to their advantages such
as ease of access to customers and inputs, supply of a developed infrastructure etc. The
variables used are the square root of population density in order to reduce the overwhelming
effect of the very high densities recorded by Istanbul, Kocaeli etc. (SQDENS), average
household size (HOUSEH), in-migration rate (MIGR) and the proportion of young
population (25-44 years old) (YOUNGPOP) in 1985.
3- Unemployment: Reynolds et. al. (1994) claim that “when people lose their jobs with
existing organizations, they may well turn to self-employment or attempt to start a new firm”
(p.446). However, as Audretsch and Fritsch (1994) claims, "high regional unemployment
rates would indicate slack growth, thereby dampening incentives for new firms to locate
within the region" (p.360) at the same time. To measure unemployment’s impact, the
unemployment rate for active population in 1985 has been used (UNEMPLOY).
4- Employment structure. To identify the relationship between labor composition, two
variables are included: the rate of illiterate population (NONQUAL), and that of graduates in
the active population (GRAD).
5- Social structure. It is assumed that areas characterized by social mobility and high
proportion of individuals in self employment will have higher new firm formation rates. The
two variables used are both the proportion of workers (WORKERS) and the proportion of
employers (EMPLOYER) in manufacturing industry in 1985. The other variables are the rate
of manufacturing (MANUF) and agricultural (AGRIC) employment in total employment.
6- Household wealth. Capital is one of the main requirements for business start-up. It is
assumed that the greater wealth in an area would bring “greater availability of startup capital
and/or borrowing costs” (Reynolds 1994, p.446). The variable used is the amount of bank
deposits per capita (DEPCAP).
7- Size structure. With regard to the size, the hypothesis of small firm incubator, “suggesting
that areas dominated by large firms are likely to generate relatively few new firms” (Garofoli
1994, p.389) is tested. However, to eliminate the bias coming from the use of number of
11