Mortality rates
In demography (gross) mortality rates are calculated as a ratio between the number of
deaths and the population. We used the same procedure for calculating mortality rates
for the population of firms. Since we want to perform APC analysis on the data, we
selected only those firms that started in 1986 (the 1986 birth cohort) or later. A firm
born in 1986, did not exist in the beginning of 1986, but appears for the first time in the
database at the beginning of 1987. A firm closing down cannot be observed until one
year later. Our period dimension therefore starts in 1988. This selection reduces the
number of firms and the number of deaths available to our analysis, substantially. We
now have information on 9,615 existing firms in 1988 to 118,077 firms in 2001 and 397
to 7,753 firms closing down respectively. For each of the three time dimensions (age,
period and cohort) mortality rates are plotted in figures 3 to 5. Figure 3 shows the
mortality rates for the period dimension.
Figure 3 Mortality rates for firms in Gelderland by period, 1988-2001
In general the mortality rates show an increasing trend over time, with two peaks around
1988 and 1999. Not surprisingly, the mortality rates show a similar pattern as in figure
2, absolute numbers of deaths.
In figure 4 the mortality rates are shown by birth cohort. The picture is very clear: older
cohorts have lower mortality rates than the younger cohorts do. Especially the most
recent years (from 1998 onwards), were relatively speaking bad startup years. For firms
that started in these years, mortality rates are the highest.
The last time dimension, age, is plotted in figure 5. As expected from the literature,
younger firms show higher mortality rates than older firms do. Mortality rates very
nicely decrease with age.
11