POPULATION
The Demography of Growing European Identity
Wolfgang Lutz,1-2* Sylvia Kritzinger,3 Vegard Skirbekk1
T:
\he process of European integration appears to be in disarray. After rejection of the new European constitution by referenda in France and the Netherlands and serious quarrels over the future budget of the European Union (EU), observers have warned that the EU is entering a period of stagnation or even disintegration (1, 2). But observers should not be overly impressed by short-term events and need to study the important underlying forces. One such force is the slowly evolving feeling of identity in the national and European context. Here we study the trends in identity and project them into the future.
Easton (3) suggested that development of identity is crucial for the legitimacy of a political system. Eurobarometer (EB) surveys provide a consistent series of accessible data with individual answers to the following question: "In the near future, do you see yourself as [Nationality] only, as [Nationality] and European, as European and [Nationality] or European only?" We combined the three categories that have at least some European element, and called this category "multiple identities"^).
In the EB survey of2004, 42% of the population above age 18 said that they felt themselves to be solely nationals of their own country, whereas 58% gave an answer that reflected at least some European identification. This implies that 130 million adult citizens of the EU-15 consider themselves only as nationals and 177 million as having multiple identities. But there are differences by country of residence (see table) and by age (see chart). The
'World Population Program, International Institute for Applied Systems Analysis (NASA), Laxenburg, A-2361, Austria. 2Vienna Institute of Demography (Austrian Academy of Sciences), Vienna, Austria, institute for Advanced Studies, Vienna, Austria.
'Author for correspondence. E-mail: lutz@iiasa.ac.at
PREVALENCE OF MULTIPLE IDENTITY
older the respondents, the higher is the chance that they feel only a national identity.
Do these data allow us to make projections? No, because this empirical pattern at only one point in time could be due to (i) a cohort effect, i.e., the current younger generations having been socialized in such a way that they will maintain their multiple identities throughout their lives, or (ii) an age effect, which would assume that peoples' identities change over their life course. Age profiles at different points in time (which the EB data provide) allow us to distinguish between these possibilities.
When the data for 1996 (the first year after the EU expanded to 15 member states) are com-
EU-15 POPULATION WITH MULTIPLE IDENTITIES
Country Percent*	
Luxembourg	78
Italy	72
France	68
Spain	64
Belgium	59
Netherlands	59
Germany	56
Denmark	54
Ireland	53
Austria	51
Portugal	50
Greece	46
Sweden	45
Finland	43
UK 40	
•Average of 1996-2004	
,<y 70-
'1 60-1 S 50
-a. 40
1 30-1
E 20
¥ -P ¥ -¥ &
<í?" & A
Age (in years)
pared to those for 2004, the proportion with multiple identities was lower at each age in 1996, although the general shape of the curve was maintained. This upward shift of the profile from 1996 to 2004 indicates that the pattern is not primarily due to an age effect, but rather is dominated by cohort effects. Visual analysis also shows that the curve is not merely shifted upwards, but that the humps and valleys are also shifted to the right, i.e., along cohort lines.
This visual pattern was confirmed analytically by a demographic age-period-cohort model (5). The model shows a strong and highly significant positive cohort effect. The coefficient we calculated of 0.48 means that for cohorts born 1 year later, the proportion with some European identity is on average half a percentage point higher. An age effect
Younger Europeans are more likely than older groups to consider themselves to have a European identity in addition to their national one.
also mattered, but only to a secondary degree, with the tendency to multiple identities reaching a peak around age 50 to 60 and then starting to decline around an age that seems to coincide roughly with retirement age.
This quantification allows us to forecast future trends by assuming that the estimated effects will continue to prevail over the coming 25 years. In 2030, under the stated assumptions, there will be only 104 million adult EU-15 citizens who have strictly national identities and 226 million with multiple identities. Age-specific proportions with multiple identities in 2030 show a marked upward shift. In the age group 30 to 44, those who have some identity as Europeans will outnumber those with strictly national identities by more than three to one.
To test the sensitivity of our results to political events at the European level, we ran an alternative model that included dummy variables for 3 years, reflecting the negotiations of the Amsterdam and Nice Treaties, as well as the introduction of the Euro (6). This did not change the results. Multivariate models, including education, urban versus rural place of residence, and occupation, showed that those segments of the population that are likely to increase in size have more multiple identities (6). In addition, changing socialization processes such as expanding European-level media impact, increasing mobility of students and tourists, as well as labor migration within the EU, may also enhance the prevalence of European identity.
Our conclusion is that as older, more nationally oriented cohorts die, there are likely to be significant changes in the pattern of European identity. Although the politics of European integration remain volatile and unpredictable, these long-term tectonic shifts in identity are likely to have major and enduring consequences for the future of Europe.
References and Notes
1. L. Cohen-Tanugi, Foreign Affairs 84, 6 (2005).
2. "Crisis, what crisis?," Economist, 8 September 2005, p. 37
3. D. Easton, A Systems Analysis of Political Life (Wiley, New York, 1965).
4. This question does not cover subnational identities.
5. R. V. Robinson, E. F. Jackson, Soc. Sci. Res. 30,117 (2001).
6. For detailed information and analyses, see supporting online material.
Supporting Online Material
www.sciencemag.org/cgi/content/full/314/5798/425/DCl
10.1126/science.ll28313
www.sciencemag.org   SCIENCE   VOL 314   20 OCTOBER 2006
Published by AAAS
425