6 Planes de gestión
6.5 Plan de gestión de recursos humanos
As discussed in subsection 3.5.2.2 a reason why immigrants are less likely than natives to suffer a Crime by Acquaintances could just be that immigrants are also more likely, particularly the most recent ones, to have a smaller network of acquaintances. However, this cannot be the case for domestic crime because, as we saw in subsection 3.5.2.1, immigrants’ households consists of more members (even for the most recent immigrants). Unfortunately, the BCS does not provide any information on the number of acquaintances the respondent has. Nevertheless, in this subsection, I examine the “network effect” hypothesis by assuming that immigrants expand their networks of acquaintances as they stay longer in the country. Therefore, a linear trend that measures the number of years of an immigrant in the country is used, once I control for immigration status and basic demographic characteristics. If the aforementioned hypothesis holds, we expect the linear trend to have a positive significant effect.
At a first glance, the results which are presented in Table 3.18 provide some support on the above hypothesis. We can see from specification 1 that the linear trend has a positive and significant at 10% effect. Thus, more recent immigrants are much less victimized than
187 natives, but immigrants’ victimization probability converges to natives’ one as time spent in the host country increases, perhaps because of network effects. However, we can also see that this assimilation is very slow as it takes around 70 years for immigrants to reach natives’ victimization probability. Moreover, the results in specification 3, where I use four assimila- tion dummies, also show that more recent immigrants are less victimized by acquaintances. However, they also indicate that time spent in the country does not affect the victimization probability linearly but a quadratic trend would be more appropriate. This is evident in specification 2 as well, where a quadratic term is also included. It is clear that starting with a very large difference, the victimization differential between immigrants and natives closes but it never becomes zero. The gap reaches its minimum at around 30 years in the country and then starts increasing.
If the effect of the trend was purely due to networks effects, we would expect it to have a linear effect. Therefore, we must be cautious with the interpretation of these results as there might be some other unobserved factors involved that give rise to the observed relationship. From specification 4 we can see that there is a linear assimilation trend for Domestic Crime as well, even though, as we saw before, immigrants’ families are larger.1 Moreover, specification 5 indicates that a weak quadratic assimilation pattern exists for Crime by Strangers too, even though networks should make no difference in crime by strangers.
Given the evidence from all crime types, it seems that if networks effects exist in Crime by Acquaintances, they are quite weak. According to these results the following story could be more appropriate. More recent immigrants, perhaps because they consider themselves more vulnerable, set strategies associated with lower victimization. As time spent in the country increases, immigrants assimilate in natives’ lifestyle, or increase their networks of familiar people, resulting in a smaller victimization difference for all crime types. However, for earlier immigrants, the picture looks different for Crime by Acquaintances and Crime by Strangers. Earlier immigrants seem to suffer fewer violent incidents even though we control for differences in the age distribution.2 Hence, earlier immigrants, due to some unobserved
1The quadratic trend does not fit well in domestic crime.
2It is essential to control for Age because earlier immigrants are older and therefore, they have a lower
victimization probability. Moreover, it is important to stress that controlling for Age by including a quadratic or a cubic term instead of dummies makes no difference in the assimilation patterns found for domestic and crime by acquaintances (but slightly weakens the assimilation relationship for crime by strangers). Therefore,
188 factors (perhaps cultural), follow social lifestyles associated with lower victimization than natives with the same basic demographic characteristics.
Summing up, immigrants face a lower probability of violent victimization and we have argued that this might be because immigrants follow social lifestyles associated with lower victimization. However, further analysis showed that this difference exists only for crime by familiar people, as immigrants face the same victimization probability for crime by strangers if we do not control for racially motivated crime. This should be considered as unexpected if immigrants follow a lifestyle that draws them away from crime activities, given that violent crime depends a lot on interactions between potential victims and potential offenders. How- ever, we provided evidence that this difference is not because of more under-reporting by immigrants. Moreover, some of this difference can be explained by racially motivated crime, and perhaps in a small degree by “network effects”. In addition, we should not ignore that crime by strangers is more “random”. Finally, the fact that the proportion of immigrants in the “strangers group” is probably smaller than the proportion of immigrants in the “family and acquaintances group”, as natives account for the 90.5% of the population (at least in the BCS data of 2007-08), could be another reason to explain the aforementioned difference, given that as found by Papadopoulos (2010b), immigrants are slightly less violent as offender than natives.
An interesting question emerging from our analysis is the following: if immigrants set the aforementioned lifestyle strategies, why do we still observe the positive (although in- significant) association for personal thefts? First of all, as has been stressed throughout this paper, personal behaviour is a highly more important determinant for violent crime than for personal thefts. This is closely related to the “randomness” that I have discussed through- out this study. Therefore, the aforementioned lifestyles of immigrants would have a much stronger effect on violent crime than on personal thefts, which has as a result to overbalance the positive victimization-immigration association because of higher proximity for violent crime, but not for personal thefts. In a cost-benefit setting, the above can be explained by the fact that it is much more costly (in the sense that it needs much higher effort) to reduce
we can argue that it is not the case that we observe the negative relationship for earlier immigrants because we were not able to capture the age distribution properly.
189 the uncertainty of suffering a personal theft than to reduce the uncertainty of suffering a violent crime.