A multistate demographic model for firms in the province of Gelderland

1992 and 1995-1997 were good birth cohorts. 1987 and 1998-2000 were bad start-up
years.

Figures 11, 12 and 13 show combined parameter estimates for age, period, and cohort.
In order to obtain these combined estimates, the original estimates were incremented
with one, and subsequently we multiplied the concerned combinations of parameter
estimates. The “average” mortality level has shifted upwards with one. The combined
effect of cohort and age is demonstrated in figure 11. The three dimensional figure
shows the product of cohort and age effects for each cohort and age. The other
parameter estimates (period, employed, sector and EMS) are excluded from the chart.
The figure reveals that cohort effects become especially pronounced in 1989-1991 and
1995-1997 and strongly suppress the age effects. The socio-economic environment in
these years worked in favour of firms starting in those years. For the yougest ages the
cohort effects are most pronounced.

Figure 12 is a very similar figure, but now the two dimensions age and period are
combined. Especially in the most recent years the effect of period on the young ages is
very strong (high mortality). Also the effect of early years on high ages is very large
(very low mortality). As is evident from the product values, the effect of period on age
is larger than the effect of cohorts.

Figure 13, finally, shows the combined effect of the dimensions year and cohort. As was
said for the cohort-age combination, the cohort effects strongly suppress the period
effects for cohorts 1989-1991 and 1995-1997. And also especially period 1999-2001
strongly reinforces the cohort effects.

16

<<   13   14   15   16   17   18   19   >>

More intriguing information