*7.1. Hypotheses H1 to H3*

The median firm sub-sample reflects the whole sample well. For all campsites, the value of β equal to 1.0 is rejected at the 5% significant level (SYS-GMM estimates) in favour of β < 1.0. Consequently, there is a tendency for Norwegian campsites to revert to some mean, steady state, or natural size. This is in contradiction to previous studies of campsites, but not in contradiction to other studies using the SYS-GMM estimator with moderately long time dimension. The results suggest that smaller campsites are growing significantly faster than larger ones. For the smallest campsites, there is considerable variation in the estimates, and this gives an uncertain result. The picture is different for the medium-sized campsites, in which the smallest firms grow significantly more than the mean company in this segment. This result is not found for large campsites, indicating that after the campsites have reached a certain level (more than 25 employees), the growth for the individual enterprise is independent of its size. We can thus conclude that the growth of a company within the camping sector is also independent of its size after a certain size is reached. Because the campsites belong to a labour-intensive sector in which there are limited localization opportunities (e.g., near the sea or lake), the possibility of economies of scale may be limited, explaining the results for the mean reversion for the Norwegian camping sector as a whole. Consequently, the crux of Gibrat's Law, the random walk of firm size for the sector, is rejected in favour of mean reversion.

Notice also that H2 (ρ = 0) is only rejected for large companies in favour of ρ > 0 with a significance level of 1%. This is an interesting result. If a large campsite does something extraordinary in a year that contributes to higher growth, this will keep going into the following year, which will show above average growth as well. That is, it has a spill-over effect for the next period, and therefore the company achieved in higher-than-average growth also for the subsequent period. The company will also have an advantage in subsequent periods based on the prior year's success. Similarly, a fiasco one year resulting in lower growth will persist from one period to the next. For the camping sector generally, however, there is no significant spill-over effect from year to year.

The volatility of companies in camping services is strongly dependent on size. With a negative significant value for the heteroskedasticity test, this means small campsites have a greater variation in growth than the average company in the sector. This confirms the results of other researchers with data from other sectors (Calvino et al. 2018; Coad 2008). There is noticeably more stability within large firms than among small ones. Smaller firms show greater fluctuation in their growth than the others. This is strongly the case for small- and medium-sized enterprises. The exception is for campsites with more than 25 employees; in this category, volatility does not depend on firm size. Volatility is highly dependent upon size, meaning that H3 is also rejected in line with existing literature.

Analysis from other sectors suggests this difference between large and small may be due to better use of technology, more differentiated activity that reduces risk and longer company history (Begenau et al. 2018). It is reasonable to assume that large campsites are more likely to have several units with different geographical locations and with a focus on multiple segments (year-round operation, winter and summer holidays, cabins, campers) and various types of activities (family activities, hiking, fishing, etc.) aimed at both domestic and foreign visitors. In this way, the risk is spread out and the firm is less vulnerable if there is a decline in any individual field (bad weather, a reduction in foreign visitors, etc.). Smaller companies cannot spread out their activity and risk in the same way and therefore become more vulnerable, which causes greater fluctuations. The larger corporations seem to be working on a longer time scale, a dynamic that is often observed both in economics and in nature.
