Page 163 - Contributed Paper Session (CPS) - Volume 3

CPS1982 Dmitri J. et al.
2. Inter-Censal Estimates and Migration
Contrary to common beliefs, international migration has been low for
decades. In 2015, about 3.3% of the world's population, or 244 million people,
lived in a country other than the country of birth (Willekens et al., 2016).
However, the distribution of migration is often highly uneven. First, it
affects working ages and have significant influence on fertility estimates. For
countries with a long history of erroneous statistics of migrants it also has
serious consequences for estimation of mortality at old ages because the
proportion of underestimated or overestimated number of migrants becomes
higher with age. Second, uneven distribution across countries might be a
serious problem for countries with substantial in- or out-migration and small
populations.
Population censuses provide the most reliable basis for retrospective
estimation of both population estimates and international migration. In many
cases, current annual migration statistics are unreliable due to a lack of
accurate flow data. In general, arrival data are more reliable than information
on out-migration. The last round of censuses in 2010-2012 allowed to produce
reliable (at least in most of the cases) population estimates around the time of
the census. In many countries the new census-based population estimates
significantly differ from the post-censal estimates based on the updating the
previous census . The next step would be to recalculate annual population
estimates back using the last census data. Unfortunately, not all countries do
it. Moreover, a number of countries showing disruption in annual population
estimates has increased in comparison with the previous round of censuses in
2000.
The standard HMD methodology (Wilmoth et al., 2007) for the cases when
population estimates between two neighboring censuses are either not
available or unreliable is based on the assumption of uniform distribution
across the entire inter-censal period. This assumption works well in many
conventional situations, but may be violated in the case of special events. For
example, the collapse of the USSR and abrupt social-economic changes in
Eastern Europe produced several migration waves at the end of the 1980s and
over the 1990s. Huge and irregular migration waves followed the EU
enlargement in 2004 and the financial crisis in 2008-2009. In such cases direct
application of the basic HMD approach would not yield satisfactory results.
For example, the official population estimates for Bulgaria show a sudden
drop in the total population count for the census years 1985, 1992, and 2001
(Figure 1). According to the official data, the total number of males decreased
by about 252,000 between 1991 and 1992 (the census year). A similar notable
discrepancy is found when one compares the official post-censal estimates as
of December 31st, 2010 to the 2011 census counts (as of February 1st, 2011).

152 | I S I W S C 2 0 1 9