Page 13 - IJPS-9-1

International Journal of
Population Studies Local population changes as a spatial varying multiscale process

2020; Raymer et al., 2019; Voss, 2007; Weeks, 2016) and
the need to use appropriate spatial methodologies in
population-based studies, i.e., considering space in the
analysis (Chi & Zu, 2008; Matthews, 2019; Matthews &
Parker, 2014; Weeks, 2004). In this general framework, a
crucial variable is the scale of analysis (Burillo et al. 2020;
Oshan et al. 2022).

In this study, we showed that this is particularly true
for Italy and its local demographic dynamics but with
two major additions: the spatial varying relationships and
multiscale nature of these relationships. In our view, this
proves the spatial complexity of demographic changes in
Italy and the need for measuring demographic processes
without a constant scale approach.
Figure 3. Specificity for ITAPGR, NATPGR, and MIGPGR (a) Indeed, it can be misleading if modeling the spatial
Note: (a)In brackets the number of municipalities. ITAPGR: Yearly demographic process is without considering the spatial
average of Italian population growth rate, NATPGR: Yearly average dimension (classic OLS model), without considering local
natural population growth rate, and MIGPGR: Yearly average internal dimensions – such as, spatial global regression models like
migratory population growth rate.
Source: Authors’ elaboration on Istat data. spatial lag model, spatial error model, and spatial Durbin
model – or without a multiscale framework (classic GWR
of municipalities that the focal variable has the strongest model).
significant impact on the dependent variable, TOTPGR At least, the case for Italy for the period 2011 – 2019
(Yang et al., 2022a; 2022b). as this paper clearly proves it. We argue that the results
Multiscale results with the three dimensions are presented achieved provide new insights into the importance of
in Table 2. They are quite interesting because they prove the treating the population process as spatial phenomena and
relevance to modeling population growth not only as a spatial in particular as local and multiscale (spatial) phenomena.
process but, most of all, as local spatial varying process. In The achieved results also have relevance in terms of policy
particular, we can see – column (a) – that each independent implications. In Italy, as in other parts of Europe, there are
variable plays primary level of influence on the dependent vast areas of land in systematic depopulation (shrinking
variable (TOTPGR). Therefore, the local importance of each regions) (Klingholz, 2009), a real challenge for territorial
covariate is high. Moreover, all the independent variables prove planners and policy makers. Adopting this type of model
to be local determinants in terms of scalability so that their (MGWR) allows the depopulation phenomenon to be
effects have to be detected at local level. In terms of specificity, modeled locally by identifying the radius of influence of
we can appreciate a quite high heterogeneity between the the different explanatory variables and thus enabling the
dependent variables. ITAPGR records the highest specificity territorial calibration of policies to counter it.
while INTPGR and FORPGR presents no specificity.
Acknowledgments
The map of specificity in Figure 3 reveals different
spatial patterns for the three variables that prove to have The authors would like to thank the two anonymous
a specificity effect, namely, ITAPGR, NATPGR, and reviewers and the editor for their suggestions, which
MIGPGR. In particular, we can observe how the effect of helped to improve the paper significantly.
ITAPGR involves much more municipalities than the other Funding
two. Most of them are located in the southern Italy but also
in the north-east area. The NATPGR specificity cover the None.
central part of Italy and the north-west too. Finally, the Conflict of interest
MIGPGR local specificity distribution covers few areas
that are almost located in the northern part of Italy. The authors declare that they have no competing interests.
4. Concluding Remarks Author contributions

In recent years, many papers have underlined the intrinsic Conceptualization: Federico Benassi, Gerardo Gallo
spatial nature of demography (De Castro, 2007; Gu et al, Investigation: Federico Benassi

Volume 9 Issue 1 (2023) 7 https://doi.org/10.36922/ijps.393