intergenerational mobility: core model of social fluidity 2371

2 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 2 1 1 1 1
1 1 1 2 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
The final inheritance matrix, IN3, reserves a
single distinction for farmers (IVc), among
whom class inheritance is particularly strong:
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 2 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
The third aspect of the core model is a single
matrix, SE, that defines barriers to movement
between the agricultural and non agricultural
sectors. Thus, all cells with agricultural origins
(classes IVc and VIIb) and non agricultural
destinations, and vice versa, are assigned the
value 2:
1 1 1 2 1 1 2
1 1 1 2 1 1 2
1 1 1 2 1 1 2
2 2 2 1 2 2 1
1 1 1 2 1 1 2
1 1 1 2 1 1 2
2 2 2 1 2 2 1
The final component of the model is two
affinity matrices. The first, AF1, identifies
movement between classes III and VIIb and
assigns the value 2 to these cells:
1 1 1 1 1 1 2
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
2 1 1 1 1 1 1
while AF2 identifies reciprocal movement
between III and III; between IVab and IVc;
and between VVI and VIIa; and non reciprocal
movement from IVc to VIIa and from VIIb
to VIIa:
1 2 2 1 1 1 1
2 1 1 1 1 1 1
2 1 1 2 1 1 1
1 1 2 1 1 2 1
1 1 1 1 1 2 1
1 1 1 1 2 1 1
1 1 1 1 1 2 1
In the case of AF1, movement between III
and VIIb is thought to be particularly unlikely
(even over and above the hierarchy, inheri
tance, and sectoral effects already included),
whereas the movements captured in AF2 are
thought to be especially likely.
The model is derived from a number of
propositions about the factors shaping inequal
ity in mobility chances (for a full explanation,
see Erikson & Goldthorpe 1987a). Because
hierarchical movements are assumed to be more
difficult than non hierarchical ones, both HI1
and HI2 should have negative values. The
inheritance effects, by contrast, should all be
positive, while the sector effect, since it captures
the difficulty of moving between the two sec
tors, should be negative. Of the affinity effects,
AF1 should be negative (it captures a disaf
finity) and AF2 positive. Some of the para
meters of the core model are to be interpreted
incrementally. For instance, the relative likeli
hood of long distance mobility is captured by
the sum of the coefficients for HI1 and HI2,
rather than by HI2 alone. HI2, in fact, tells us
the degree to which long range mobility is more
difficult than short range. In the same way, the
propensity to inheritance of class IVc is given by
IN1 plus IN2 plus IN3. So IN3 itself measures
the extra inheritance observed among farmers
compared with classes III and IVab.
Erikson and Goldthorpe (1992: 173) argue
that the FJH thesis requires, as the first criterion
for its support, that a model of fluidity (in this
case the core model of fluidity) should provide
a reasonable fit to the data. But they find that the
core model with the same parameter values for
all countries in their sample does not provide an
adequate fit to the data: in other words, there is
statistically significant variation between coun
tries in the parameters of the model. Allowing
such variation, the model fits the data much
better, and Erikson and Goldthorpe take the
success of this as supporting the FJH hypothesis: