The study of migration in a comparative perspective implies that we find a common solution to three kinds of problems.
This problem should probably not be over-estimated. Migration is not mobility. Mobility could be any movement but migration is an official change of domicile. In China, Japan, Sweden, and Belgium as well as in Venice, migration could simply be a change that the social system of this place and this time considered as important enough to be noted in an official document. Classically, some movements are missing but they normally produce an administrative reaction ("inscription by order" or "radiation" in Belgium; indication of a desertion in the Chinese case; and so on). The exception is rural Italy where a direct registration of migration did not exist, and where the statistical unit could be either a change of household or a change of parish in the year between two status animarum. If we decide that it could be useful it will be easy to "simulate" the rural Italian situation in the other databases.
Basically, we will all stay with two types of migration:
For the study of the first type, the accuracy of some data can be questioned in the Belgian case and perhaps elsewhere (Venice ?), as well as the capacity of a household to be a coherent statistical unit for a longitudinal analysis (see point 2). For the second type, we must keep in mind that, other things equal, geographically larger parishes will have lower external migration rates than smaller ones.
In continuation of the preceding discussion, we have to distinguish between different types of migration, considering that they respond to different causes and that different models of event history analysis will be needed to be explain them– among other things.
Suggestions are to consider separately:
The best records for migration, the Japanese ones, cover also movements like trade or military service that are interesting and could be a topic for several specific studies, but this is not migration in the sense defined above (i.e. a change of domicile).
It is important to isolate migrations of servants if they are not kin of the household head because in many countries, the registration of their migration is very bad. However, we encourage those who have good data to investigate this case because it is essential for the examination of theories about family systems, like life-cycle servanthood and its real nature. As Martin pointed out in his Beijing paper, there is a potential problem with servant migration when it comes to certain more specific issues. Due to the high general mobility of servants (a majority moving every year) it can be difficult to connect servant migration to such issues as economic stress since the main response might be to not taking in a new servant, rather than letting a servant move. In this way, the response to economic stress will manifest itself as a change in household composition (the number of servants) rather than in migration of servants per se. This is probably more a matter of definition, but might nevertheless be worth considering.
This discussion proofs that we need another discussion on the choice of our statistical unit. This is especially the case when we work on family migration. Indeed in three pilot studies, we observe three different solutions. Catherine has chosen to work on the individuals married and aged 25-54 (both males and females) using the composition of their household as variables, while Marco and Matteo have used the household as the unit of analysis. In the Swedish case, individuals (for leaving home and servant migration) or nuclear families (for family migration) leaving a household (and/or the parish) are considered as the units of analysis. In any case, it is important not to include the same event more than one time, i.e. when analyzing family migration the event of both husband and wife should not be included, since they can not be regarded as independent events.
It could be quite easy to reach a consensus on this question of statistical unit, after some discussions, but it is important to keep in mind also the problem of how to treat the cases of splitting (when a nucleus leaves another one, a complex household becoming a simple; or when an individual, someone other than a single children or a servant, leaves an extended household). These cases, which could be relatively numerous in some Asian cases, could form a fourth type, but as for many other points, this has to be verified by empirical investigations within the different teams.
Finally, in Beijing Tommy suggested an analysis of return-migration, because they are an immigration for which we know the population at risk and for which we are able to compute explanatory variables. This could be extremely interesting on the basis of a distinction between systems of "true" mobility and of "circular" moves. Practically, I think that such a study is not possible in Sart or Madregolo, but it could give us a lot of information on life cycle mobility (servanthood ?) in Japanese villages or in Scania or in Belgian towns.
I will be brief here because the points below are developed in Catherine’s paper, which will be distributed within the group as an EAP working paper. Just to remind you:
1)
2)
While it has been assumed many times that migration was a marginal component of demography in pre-industrial societies, in the last 10 or 20 years it has become more and more clear that this is false, and that a wide variety of behaviours existed throughout the world. However, quantitative evidence and demonstrations remain rare, and even if we are able to make more complex things with our data, we must not forget to provide simple measures. They can be ordered in four sub-chapters, the first three depending on each team, the fourth one being a comparative exercise.
The model below is quite classical for members of the EurAsian Project. We just stress the interest, demonstrated in Marco-Matteo’s first analysis of Madregologo, in making a distinction among the short-term economic stresses between prices and epidemic events.
| Stresses: | Family systems: | Migration: | ||
| Malthusian | Nuclear | Individual:
|
||
Short-term:
|
Stem | Family | ||
| Family: | Joint | Nucleus within household | ||
| [interaction] | ||||
| Socio-economic Status |
For the migration research as well as for all the other EurAsian studies, it is important to continue to discuss the interpretation of price fluctuation in measuring economic stress. This includes more detailed analysis of the impact of prices and harvests on producers vs. consumers, on different groups over time. Last developments (see especially George answer to Tommy questions about rural economy and short-term fluctuations of mortality in East Belgium, on iu_eapcore) stress the importance to, where available, use harvest and climatic data together with prices in the analysis.
This is a set of questions of great importance, but also certainly the one on which we need the most work and progress. At first glance, we can mention
Connected with 1.1.
For an introduction to 4.2. and 4.3. see papers presented by Akira and Martin in Beijing.
From the Belgian case especially, we can see complex connections between mortality and migration. At an aggregate level, out-migration reduces the overall tension between economy and population in a village like Sart. This explains the paradox of a decrease in economic resources going together with a progressive increase of life expectancy. However, at a family level, it is very clear that out-migrants were recruited disproportionately from healthy households (selection). And finally, higher sensitivity to economic stress has been observed among young single adults. Even if it is difficult to provide a statistical demonstration that this is a result of their higher mobility and contacts with the dangerous epidemiological places in the industrial towns, Muriel has recently found qualitative evidence on this point. There are 19th-century discussions in the Belgian parliament about the law on the "domicile for assistance", something like the discussions about the settlement of the poor in the English poor laws. Some representatives of the countryside said explicitly that young rural immigrants fell sick in the towns and went back to their villages to receive care and sometimes die, which imposed expenses on the rural charitable administration due to an infection acquired in a town.
Except for the idea of differential fertility of natives and immigrants, the connections between the study of migration and research on fertility concern essentially methodological problems. For example, how important were the first observed births of immigrant families, which are sometimes excluded from our analysis? Did this bias change over time? representativity…
These connections are mainly derived from two points of view :