Sage Journals: Discover world-class research

Abstract

National pride predicts a wide range of politico-economic outcomes, yet what makes individuals proud of their nation is not completely understood. We propose and test a theory that an important but thus far unexplored determinant of contemporary national pride is the longevity of national identity. To measure the longevity of national identity, we construct an index based on responses from an original expert survey designed to trace the emergence of national identity across the polities of Europe and the former Soviet Union. We find that our National Identity Longevity Index is statistically significantly positively associated with the extent of national pride. The implied effect is robust and noteworthy in magnitude. Our results suggest that contemporary national pride inter alia reflects deep, historically rooted societal conventions which take time to emerge.

Keywords

National pride national identity longevity

Introduction

National pride is regarded by social scientists as an important phenomenon, both in its own right and because of its implications for a range of salient issues in comparative politics and economics. National pride has been shown, for example, to predict government effectiveness (Ahlerup and Hansson, 2011), preferences for and extent of redistribution (Qari et al., 2012; Shayo, 2009), tax compliance (Konrad and Qari, 2012), support for protectionism (Mayda and Rodrik, 2005), sentiment toward the Euro (Müller-Peters, 1998), and attitudes toward immigrants and foreign populations (Wagner et al., 2012). Moreover, scholars of nationalism and national identity have long noted the important role that a sense of national belonging may play in nation-building processes (e.g. Bendix, 1980). Yet the determinants of national pride, despite recent increasing interest in the topic (e.g. Bekhuis et al., 2014; Han, 2013; Lan and Li, 2015; Solt, 2011), remain incompletely understood.

In this paper, we contribute to the developing literature on the causes of national pride by proposing and empirically assessing the theory that one fundamental but thus far unexplored factor impacting national pride is the longevity of national identity. The basis of our argument is the notion that a well-defined and entrenched sense of a common national identity, through the process of ongoing diffusion, fosters national pride. Thus, all else equal, national pride should be greater in countries with longer histories of national identity.

To assess the validity of our hypothesis we construct a novel index of the longevity of national identity for polities of Europe and the former Soviet Union. While there exists no comprehensive set of comparable sources that would enable the systematic documentation of the historic emergence of national identities across these polities, some scholars have looked for tangible evidence of the timing of national identity and nationalism. Centeno (2002) examines the number of national monuments as an indirect proxy for national identity. Wimmer and Feinstein (2010) use the date of the first ‘nationalist organization’ to capture the presence of nationalism. Neither of these indirect measures, though, is meant to offer a reliable proxy of the diffusion of national identity in a population. A tiny minority of nationalists with control over resources may be able to construct monuments even if the majority of the population has no such identity. Likewise, we would not expect all cases of the diffusion of national identity to involve nationalist organizations or for the presence of such organizations to consistently reflect the diffusion of national identity in a given polity.

We construct our index by drawing on an original expert survey that we conducted with the aim of obtaining a first comprehensive quantitative picture of the polity-wide emergence of modern-day national identities in the wider Europe.¹ We combine our country-level index of the longevity of national identity with individual-level data from the European Values Survey (EVS). Using an encompassing range of controls, we find that the longevity of national identity is indeed positively associated with national pride. The uncovered effect is sizable in magnitude and robust to a range of alternative model specifications and robustness checks.

Existing research has demonstrated that national pride is a function of a variety of individual-specific characteristics (e.g. Evans and Kelley, 2002; Jones and Smith, 2001; Smith and Jarkko, 1998), market conditions (e.g. Lan and Li, 2015) and even contemporary events (e.g. Kavetsos, 2012). Our analysis sheds novel light on this ongoing debate by demonstrating that contemporary national pride is also a reflection of deep, historically rooted societal conventions that take time to emerge and are slow to change.

National pride and the longevity of national identity: A conceptual framework

Scholars of nationalism recognize that the notion of national identity, like the institutional form of the nation-state, is a ‘modular’ phenomenon (Anderson 1991: 4) that has undergone a process of global diffusion.² Within a given society, national identity particular to that society then spreads internally (Greenfeld 1992: 22). Since there is often a considerable lag between the time that a national identity enters a society and the time when it comes to be shared by a majority of the society’s members—a process that may take multiple generations (see e.g. Weber, 1976)—the first mechanism through which we would expect the longevity of national identity to impact national pride is through scope of diffusion. The longer national identity has been in circulation in a polity’s networks, the more extensive and deep we would expect its diffusion to be.

In addition, national pride is likely greater in older than in younger nations because older nations have, all else equal, had more time to accumulate experiences and events (related to, for example, literary, scientific and art work, as well as political events and wars) that can be memorialized as national achievements. Accordingly, for a given scope and depth of social networks, the diffusion of national identity is likely greater in older nations due to a greater accumulation of such potential sources of national identity and pride.³

The longevity of national identity might impact contemporary national pride via several additional mechanisms. For example, nationalism may shape economic policy in ways likely to affect a society’s economic performance (Abdelal, 2001) which, in turn, often serves as a basis for national pride (Greenfeld, 2001). Since nationalism emphasizes popular sovereignty as the basis of the state’s legitimacy (Anderson, 1991: 7; Greenfeld, 1992), the longevity of national identity might impact national pride via affecting the responsiveness of state institutions. In other words, because states govern in the nation’s name, an entrenched sense of national identity may underpin civic demand for effective governance; effective governance, in turn, may be conducive to national pride. Finally, given that national identity emphasizes homogeneity and the reduction of intra-national differences (Gellner, 1983), the longevity of national identity might affect national pride through its effect on societal fractionalization. Below, we empirically test for the relevance of some of the above-mentioned potential channels of influence of the longevity of national identity on national pride.

Data and variables

Outcome variable: National pride

Our outcome variable is an individual’s response to the EVS question ‘How proud are you to be a…[country citizen]?’, with answers ranging from ‘not at all proud’, ‘not very proud’, ‘quite proud’, and ‘very proud’. We code these responses using an integer scale ranging from 1 (‘not at all proud’) to 4 (‘very proud’). In our sample of 28,178 citizens of 36 European countries, the mean score on our integer scale equals 3.3. In our sample 44.7% percent of individuals respond that they are very proud of their nation, 42.3% are quite proud, 10.4% are not very proud, and 2.6% are not at all proud.

Focal explanatory variable: Index of the longevity of national identity

Our focal explanatory variable is an index of the longevity of national identity. The basis for the construction of the index was an original expert survey administered between June 2014 and September 2015, in which we asked questions regarding the emergence of national identity in European and former Soviet countries (see Appendix A). Upon completion of the survey, we secured responses from 119 experts on the 36 countries included in our dataset, or on average 3.2 responses per country.

To construct our index of the longevity of national identity we draw on two questions from the expert survey. The first question is ‘Please identify the one-third century during which you believe a majority of the residents of your polity of expertise first adopted national identity’. For each expert’s response, we assigned one point to each one-third century since the passing of that threshold. We discounted the influence of the past at every one-third century. Following Bockstette, et al. (2002), in our main empirical analysis we utilize the index constructed using a 5% discount rate. For example, if an expert response stated that the majority of residents first adopted national identity in the first one-third of the twentieth century, then the discounted value of the expert-specific sub-index equals 1+1×(1+0.05)⁻¹+1×(1+0.05)⁻²+1×(1+0.05)⁻³=3.72.

Our experts do not always agree on the timing of emergence of national identity. Indeed, the within-polity divergence of experts’ responses can be considerable (with experts’ estimates sometimes more than a century apart; see Appendix A), which is hardly surprising given the complexity of the topic at hand. We thus used the experts’ responses to the question ‘How much confidence do you have in the accuracy of your estimate?’ to compute for each polity a weighted average of the above-described discounted expert-specific sub-indices with experts’ confidence levels as weights. The response ‘no confidence’ was assigned the weight of 1, ‘reasonable confidence’ the weight of 2, and ‘high level of confidence’ the weight of 3.

Finally, we computed the normalized value of the country-level index of the longevity of national identity. To this end, we divided the country-specific index (weighted average of the above-described discounted expert-specific sub-indices) by the maximum possible value of the country-level discounted longevity index (in our sample, the value for Portugal). Hence, our normalized country-level index of the longevity of national identity ranges between 0 (no history of national identity) and 1 (longevity of national identity for Portugal, our maximum). Table 1 provides the values of our (normalized) National Identity Longevity Index for the 36 countries in our sample.

Table 1.

Values of the National Identity Longevity Index.

Country	National Identity Longevity Index
Albania	0.30
Austria	0.38
Azerbaijan	0.30
Belarus	0.22
Belgium	0.32
Bosnia and Herzegovina	0.08
Bulgaria	0.53
Croatia	0.47
Czech Republic	0.42
Denmark	0.52
Estonia	0.38
Finland	0.47
France	0.48
Georgia	0.40
Germany	0.57
Great Britain	0.83
Greece	0.60
Hungary	0.50
Iceland	0.21
Italy	0.48
Latvia	0.41
Lithuania	0.38
FYR of Macedonia	0.34
Moldova	0.28
Netherlands	0.64
Norway	0.65
Poland	0.45
Portugal	1.00
Romania	0.48
Russia	0.29
Slovakia	0.37
Slovenia	0.65
Spain	0.59
Sweden	0.42
Switzerland	0.47
Ukraine	0.34

FYR: Former Yugoslav Republic.

Note: The table presents the values for the (normalized) National Identity Longevity Index for countries included in the estimating sample for the results reported in Tables 2 and 3. The section headed Focal explanatory variable provides an account of the construction of the index. Appendix A provides a detailed description of the underlying expert survey and experts’ responses.

Given the lack of reliable comparable historical sources on the timing of emergence of national identity in the polities of our interest, a survey of experts, who have the requisite knowledge as well as the possibility to access and process diverse sources of information, is an appropriate data collection method for our purposes (Hooghe et al., 2010: 689, 692). Expert surveys, however, also raise concerns about the validity and reliability of responses (see e.g. Budge, 2000; Hooghe et al., 2010; Steenbergen and Marks, 2007). Our expert survey is not immune to these concerns. On the issue of validity, we have verified that the estimates of the longevity of national identity based on our National Identity Longevity Index are broadly consistent with an illustrative sample of historical narratives, drawn from a secondary source, about the timing of emergence of national identity (see Appendix A). To probe the consequences of variability in expert responses, we examined if our empirical results are robust to using versions of the index that utilize only extreme expert responses for each polity. Indeed, we find that they are (see below). This suggests that, despite uncertainty in expert estimates of the longevity of national identity, the average from expert responses is a sensible measure of the timing of polity-wide emergence of national identity and, furthermore, the underlying relationship of our interest (between the longevity of national identity and contemporary national pride) is robust to the choice of extreme expert estimates.

Controls

The timing of the historic emergence of national identity in a polity is plausibly exogenous to an individual’s contemporary national pride. To minimize omitted variable bias, however, we nevertheless use both individual and country-level controls. Our individual-level controls are drawn from the EVS and capture an encompassing range of demographic, socio-economic, and politico-economic factors that have been shown to be correlated with an individual’s national pride (see e.g.; Bekhuis et al., 2014; Coenders and Scheepers, 2003; Han, 2013; Jones and Smith, 2001; Jung, 2008; Kavetsos, 2012; Smith and Jarkko, 1998).⁴ To minimize the scope for any specification bias due to the potential non-linearity of the effects, we capture the effect of each of the corresponding categorical controls through inclusion of a full range of dummy variables.

As noted above, one mechanism through which the longevity of national identity might affect national pride is through the record of historical events about which a nation’s members might feel pride. But any such effect might also be due to the antiquity of the state. To ensure that any seeming effect of the longevity of national identity does not reflect an underlying effect of the antiquity of the state, we therefore include the index of state antiquity developed by Bockstette et al. (2002). We further include a series of country-level covariates which plausibly capture some of the channels, other than the historic diffusion of national identity per se, through which the longevity of national identity may influence an individual’s national pride.

Methods and results

Empirical approach

Our data are hierarchical in that individuals are nested in 36 countries. Our individual-level observations for any given country are thus not statistically independent, but rather (positively) correlated. To take such intra-cluster correlation into account, we include individual and country-level covariates in a single pooled model and base statistical inference on heteroscedasticity-robust standard errors clustered at the country-level. Since our interest lies in estimating the average marginal effect of the longevity of national identity on national pride, as opposed to quantifying, for example, the extent to which differences in national pride across individuals reflect country-level differences versus individual-level characteristics, such an approach is fully appropriate (see e.g. Wooldridge, 2002: 331). However, we also show that our results are robust to alternative empirical approaches such as the estimation of a multilevel (mixed) model.

Our outcome variable is an ordered response variable. To explore the relationship between an individual’s national pride and the longevity of national identity, we therefore estimate an ordered probit model. The marginal effects in an ordered probit model vary across individual observations and are nonlinear functions of the covariates (see e.g. Wooldridge, 2002). Furthermore, each category of responses (ranging from ‘very proud’ to ‘not at all proud’) has its own set of effects. We focus on the response probability that the individual is ‘very proud’ and compute the associated average marginal effects, i.e. the average of the marginal effects for all sample observations.

Main results

Table 2 reports the average marginal effects for our primary explanatory variable—the National Identity Longevity Index—as well as for the country-level covariates.⁵ Based on the baseline specification without any controls (column (1)), a one-unit increase in the value of the index of the longevity of national identity (the difference between Portugal’s longevity of national identity and no history of national identity) is associated with an average increase of 0.24 in the probability of a response that an individual is ‘very proud’ of his or her country. The addition of a full set of individual-level controls (column (2)) only marginally reduces the magnitude of this effect to 0.23. Additionally controlling for the state antiquity index (column (3)) increases the magnitude of the average marginal effect of the longevity of national identity on the probability that an individual is ‘very proud’ of their nation to 0.29.

Table 2.

Average marginal effects for probability that the answer to the national pride question is ‘Very Proud’.

Explanatory Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
National Identity Longevity Index	0.2397* (0.1341)	0.2292* (0.1264)	0.2859*** (0.1131)	0.2797** (0.1170)	0.2778** (0.1196)	0.2829** (0.1143)	0.2672** (0.1078)	0.2085* (0.1134)	0.1527⁺ (0.1017)	0.2731*** (0.1028)	0.1952* (0.1014)
Individual-Level Controls	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country-Level Controls	No	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
State Antiquity Index			−0.1520 (0.0933)	−0.1566 (0.0973)	−0.1578 (0.0984)	−0.1517(0.0950)	−0.2169** (0.0931)	−0.1614** (0.0795)	−0.1538** (0.0712)	−0.0853(0.0722)	−0.1686*** (0.0605)
GDP Per Capita (in USD1,000)				0.0004 (0.0012)							− 0.0019 (0.0020)
Government Effectiveness					0.0073 (0.0195)						0.0058 (0.0339)
Gini						0.0009(0.0032)					0.0006 0.0042
Socialist							−0.0813** (0.0351)				−0.0906** (0.0448)
Ethnic Fractionalization								−0.1726** (0.0817)			0.1491 (0.1153)
Linguistic Fractionalization									−0.2517*** (0.0781)		−0.2822** (0.1217)
Religious Fractionalization										−0.3034*** (0.0878)	−0.1540 (0.1188)
No. Obs.	28,178	28,178	28,178	28,178	28,178	28,178	28,178	28,178	28,178	28,178	28,178
Pseudo R²	0.0033	0.0390	0.0403	0.0403	0.0403	0.0403	0.0418	0.0419	0.0452	0.0466	0.0492

GDP: Gross Domestic Product; Gini: The Gini coefficient of income inequality.

Note: The table reports the average marginal effects based on the ordered probit regressions. Heteroscedasticity-robust standard errors clustered at the country level in parentheses. *, **, and *** denote statistical significance at the 10%, 5%, and 1% level, respectively, for a two-sided test of significance. ⁺ denotes statistical significance at 10% level for a one-sided test of significance.

Columns (4) through (10) report the results based on specifications where we sequentially introduce a series of further country-level covariates. The inclusion of GDP per capita (column (4)), measures of government effectiveness and economic inequality (columns (5) and (6)), and an indicator for socialist past (column (7)) only marginally change the magnitude of the effect of the National Identity Longevity Index. The estimated average marginal effect of the longevity of national identity remains stable across specifications (3) to (7) and equal to between 0.27 and 0.29. This suggests that the effect of the longevity of national identity on national pride does not operate through the channels of economic development, quality of governance and economic inequality. Furthermore, among country-level controls included in columns (4) through (7), only the average marginal effect of having a socialist past is statistically significant. All else equal, a history of socialist rule decreases the probability that an individual is ‘very proud’ by about 0.08.

In columns (8) through (10) we further include a set of controls that capture various dimensions of country-level societal heterogeneity. The inclusion of measures of ethnic and linguistic fractionalization (but not of religious fractionalization) notably reduces the magnitude of the average marginal effect of the longevity of national identity to 0.21 (in the case of ethnic fractionalization; column (8)) or even to 0.15 (in the case of linguistic fractionalization; column (9)). The effect of the longevity of national identity, however, remains statistically significant. All three dimensions of country-level heterogeneity exhibit a statistically significant and quantitatively noteworthy negative average marginal effect on national pride. As we might expect, given the close relationship between language and ‘national membership’ (e.g. Laitin 2007: 31), linguistic fractionalization seems to be particularly important.

The above analysis suggests, first, that ethnic and linguistic fractionalization provide empirically salient channels through which the longevity of national identity affects national pride. Intuitively, the longevity of national identity is negatively correlated with ethnic and linguistic fractionalization (with correlation coefficients equal to −0.4951 and −0.4326, respectively), and less ethnic and linguistic fractionalization in turn increases national pride. Second, the effect of the National Identity Longevity Index on the probability that an individual is ‘very proud’ remains positive, statistically significant, and non-trivial in magnitude even upon simultaneous inclusion of a full set of individual and country-level controls (column (11)). Thus, the longevity of national identity exhibits an effect on national pride that is independent of the fractionalization channel.⁶ These empirical results are consistent with our hypothesis that the extent of diffusion of national identity represents an important channel through which the longevity of national identity impacts contemporary national pride.

Robustness checks and alternative model specifications

We subjected our results to a series of robustness checks. Table 3 reports the results. We constructed and used alternative indexes of the longevity of national identity by varying the discount rate between 0% and 10% (columns (1) through (4)). We estimated the model while dropping the observations for Portugal (column (5)) as an outlier with regard to the values of our index, and for the five countries (column (6)) in our sample for which the value of our index, reported in Table 1, is based on a single expert response (see Appendix A). We estimated the model using different re-coded versions of the index: a version that incorporates the responses of two experts who did not elaborate on their definition of national identity (column (7)); a version that is based only on the set of those expert responses, one for each polity, that imply the smallest longevity of national identity (column (8)); and a version that is based only on the set of those expert responses, one for each polity, that imply the greatest longevity of national identity provided that the expert’s level of confidence in such a response was ‘high’ (column (9)).⁷ None of our findings concerning the positive effect of the longevity of national identity change as a result.

Table 3.

Results for robustness checks.

Explanatory Variables	Alternative Discount Rates				Dropping Outliers and Alternative Coding					Further Controls
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
National Identity Longevity Index	0.2022** (0.0974)	0.1983** (0.0976)	0.1987** (0.0984)	0.1921* (0.1042)	0.1945⁺ (0.1326)	0.2532* (0.1341)	0.2895** (0.1214)	0.2115*** (0.0733)	0.6578*** (0.1397)	0.2497** (0.1011)
Individual-Level Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country-Level Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
No. Obs.	28,178	28,178	28,178	28,178	27,708	24,609	25,766	28,178	12,790	25,054
Pseudo R²	0.0492	0.0492	0.0492	0.0493	0.0475	0.0505	0.0520	0.0505	0.0715	0.0561

Note: The table reports the average marginal effects based on the ordered probit regressions. Columns (1) through (4) report results when the National Identity Longevity Index is constructed using alternative discount rates equal to 0% (column (1)), 0.1% (column (2)), 1% (column (3)), and 10% (column (4)). Column (5) reports results when dropping observations for Portugal. Column (6) reports results when dropping observations for Albania, Georgia, Iceland, Norway and Slovenia, as the polities for which the value of the National Identity Longevity Index reported in Table 1 is based on a single expert response (see Appendix A). Column (7) reports results when dropping observations for Albania, Georgia and Norway and using an alternative coding for the National Identity Longevity Index when responses of two experts (for Iceland and Slovenia) who did not elaborate on their definition of national identity are included in the calculation of the value of the index. Column (8) reports results when the National Identity Longevity Index is based only on the set of those expert responses that, for each polity, imply the smallest longevity of national identity. Column (9) reports results when the National Identity Longevity Index is based only on the set of those expert responses that, for each polity, imply the greatest longevity of national identity provided that expert’s level of confidence in such a response was ‘high’. The set of individual and country-level controls in columns (1) through (9) is as featured in column (11) of Table 2. Column (10) reports results using the following additional country-level controls (see Tables B1 and B7 in Appendix B): World War II Axis Power, Democracy, Accumulated Democracy, and In War 1946–2008. Heteroscedasticity-robust standard errors clustered at the country level in parentheses. *, **, and *** denote statistical significance at the 10%, 5%, and 1% level, respectively, for a two-sided test of significance. ⁺ denotes statistical significance at 10% level for a one-sided test of significance.

We re-estimated the model after including additional country-level covariates (column (10)). Specifically, following Elkins and Sides (2007), we control for the level of democracy, a measure of accumulated democracy, and World War II Axis power status. We further control for whether a country was involved in an interstate armed conflict between 1946 and 2008. The effect of the longevity of national identity remains positive and statistically significant. Among the additional controls included in the specification reported in column (10), only country’s status as World War II Axis power is statistically significant (see Table B7 in Appendix B); as anticipated, history as an Axis power exerts a negative effect on national pride.

We also examined several alternative model specifications. Table 4 summarizes the results (in the form of a coefficient, odds ratio, or average marginal effect). Column (1) reports the estimated average marginal effect and the odds ratio based on the proportional-odds ordered logit model as another commonly utilized ordered response model. Column (2) reports the estimated odds ratio based on the multilevel, random-intercept ordered logit model (see e.g. Rabe-Hesketh and Skrondal, 2012) as an alternative approach to addressing the issues of clustering and contextual effects in hierarchical data (see e.g. DiPrete and Forristal, 1994; Steenbergen and Jones, 2002).⁸ Column (3) reports the OLS coefficient from a model where the outcome variable national pride is viewed as cardinal rather than ordinal. Finally, column (4) reports the estimated coefficient from applying the Hausman and Taylor (1981) instrumental variable approach (see e.g. Wooldridge, 2002: 325-328; Rabe-Hesketh and Skrondal, 2012: 253-257) when we treat the longevity of national identity as the only endogenous regressor and model the outcome variable as cardinal. In sum, the results in Table 4 show that our key finding concerning the effect of the longevity of national identity on national pride is robust to alternative model specifications.

Table 4.

Results for alternative model specifications.

	(1) Ordered Logit		(2) Random-Intercept Ordered Logit	(3) OLS	(4) Hausman-Taylor
Explanatory Variables	Avg. ME	Odds Ratio	Odds Ratio	Coeff.	Coeff.
National Identity Longevity Index	0.2010* (0.1091)	2.4419* (1.1880)	2.3185* (1.1406)	0.3108* (0.1647)	0.2696⁺ (0.2008)
Individual-Level Controls	Yes		Yes	Yes	Yes
Country-Level Controls	Yes		Yes	Yes	Yes
Country-Level Variance			0.1711*** (0.0501)
No. Obs.	28,178		28,178	28,178	28,178

ME: marginal effects.

Notes: The table reports the average marginal effects, odds ratios, and coefficients for the effect of National Identity Longevity Index based on different model specifications (see section Robustness checks and alternative model specifications). The set of individual and country-level controls is as featured in column (11) of Table 2. Heteroscedasticity-robust standard errors clustered at the country level in parentheses. *, **, and *** denote statistical significance at the 10%, 5%, and 1% level, respectively, for a two-sided test of significance. ⁺ denotes statistical significance at the 10% level for a one-sided test of significance.

Conclusion

We contribute to the ongoing scholarly debate on the determinants of national pride by proposing a novel explanatory factor: the longevity of national identity. We draw on an expert survey to construct a very first index of the longevity of national identity for the polities of wider Europe. We then demonstrate empirically that an individual’s pride in his or her nation indeed reflects deep societal roots stemming from the historic emergence and diffusion of a common national identity.

Our measure of the longevity of national identity is certainly not meant to be final. To confront the uncertainty in our expert responses, future work should seek ways to re-assess our index by surveying other experts and expanding the number of experts per polity. Future research could also aim to explore the external validity of our finding by extending our analysis to other regions in the world. Such an endeavor will require global collection and evaluation of data on the historic emergence of national identity. Given the importance of nationalism with respect to numerous key questions in contemporary social science and scant comparative empirical research on the factors explaining national pride, the study of the emergence of national identity and its repercussions for contemporary national pride represents a worthwhile avenue for further research.

Footnotes

For helpful comments and suggestions we thank John Breuilly,John L. Campbell,Liah Greenfeld,John Hall,Sara Helms,Nicolas Prevelakis,John Stone,Andreas Wimmer,participants at the annual meeting of the Southern Economic Association and workshop of the Department of Sociology and Anthropology at Washington and Lee University,the referees for the American Sociological Association,Fund for the Advancement of the Discipline grant competition,two anonymous referees of this journal and the editor Kristian Skrede Gleditsch.

Declaration of conflicting interests

The authors declare no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

This research was funded by the American Sociological Association’s Fund for the Advancement of the Discipline Award supported by the American Sociological Association and the National Science Foundation.

Supplementary material

The online appendices are available at:

Carnegie Corporation of New York Grant

The open access article processing charge (APC) for this article was waived due to a grant awarded to Research & Politics from Carnegie Corporation of New York under its ‘Bridging the Gap’ initiative. The statements made and views expressed are solely the responsibility of the author.

References

Abdelal

(2001) National Purpose in the World Economy: Post-Soviet States in Comparative Perspective. Ithaca, NY: Cornell University Press.

Ahlerup

Hansson

(2011) Nationalism and government effectiveness. Journal of Comparative Economics 39(3): 431–451.

Anderson

(1991) Imagined Communities: Reflections on the Origin and Spread of Nationalism. New York, NY: Verso.

Bekhuis

Lubbers

Verkuyten

(2014) How education moderates the relation between globalization and nationalist attitudes. International Journal of Public Opinion Research 26(4): 487–500.

Bendix

(1980) Kings or People: Power and the Mandate to Rule. Berkeley, CA: University of California Press.

Bockstette

Chanda

Putterman

(2002) States and markets: The advantage of an early start. Journal of Economic Growth 7(4): 347–369.

Budge

(2000) Expert judgements of party policy positions: Uses and limitations in political research. European Journal of Political Research 37(1): 103–113.

Centeno

(2002) Blood and Debt: War and the Nation-State in Latin America. University Park, PA: Penn State University Press.

Coenders

Scheepers

(2003) The effect of education on nationalism and ethnic exclusionism: An international comparison. Political Psychology 24(2): 313–343.

10.

DiPrete

Forristal

(1994) Multilevel models: Methods and substance. Annual Review of Sociology 20: 331–357.

11.

Elkins

(2010) Diffusion and the constitutionalization of Europe. Comparative Political Studies 43(8/9): 969–999.

12.

Elkins

Sides

(2007) Can institutions build unity in multiethnic states? American Political Science Review 101(4): 693–708.

13.

Evans

MDR

Kelley

(2002) National pride in the developed world: Survey data from 24 nations. International Journal of Public Opinion Research 14(3): 302–338.

14.

Gellner

(1983) Nations and Nationalism. Ithaca, NY: Cornell University Press

15.

Greenfeld

(1992) Nationalism: Five Roads to Modernity. Cambridge, MA: Harvard University Press.

16.

Greenfeld

(2001) The Spirit of Capitalism: Nationalism and Economic Growth. Cambridge, MA: Harvard University Press.

17.

Han

(2013) Income inequality, international migration, and national pride: A test of social identification theory. International Journal of Public Opinion Research 25(4): 502–521.

18.

Hausman

Taylor

(1981) Panel data and unobservable individual effects. Econometrica 49(6): 1377–1398.

19.

Hobsbawm

Ranger

(eds) (1983) The Invention of Tradition. New York, NY: Cambridge University Press.

20.

Hooghe

Bakker

Brigevich

. (2010) Reliability and validity of measuring party positions: The Chapel Hill expert surveys of 2002 and 2006. European Journal of Political Research 42(4) 684–703.

21.

Jones

Smith

(2001) Individual and societal bases of national identity: A comparative multi-level analysis. European Sociological Review 17(2): 103–118.

22.

Jung

(2008) Growing supranational identities in a globalizing world? A multilevel analysis of the World Values Surveys. European Journal of Political Research 47(5): 578–609.

23.

Kavetsos

(2012) National pride: War minus the shooting. Social Indicators Research 106(1): 173–185.

24.

Konrad

Qari

(2012) The last refuge of a scoundrel? Patriotism and tax compliance. Economica 79(315): 516–533.

25.

Laitin

(2007) Nations, States, and Violence. New York: Oxford University Press.

26.

Lan

(2015) The economics of nationalism. American Economic Journal: Economic Policy 7(2): 294–325.

27.

Leerssen

(2011) Viral nationalism: Romantic intellectuals on the move in nineteenth-century Europe. Nations and Nationalism 17(2): 257–271.

28.

Mayda

Rodrik

(2005) Why are some people (and countries) more protectionist than others? European Economic Review 49(6): 1393–1430.

29.

Meyer

Boli

Thomas

. (1997) World society and the nation-state. American Journal of Sociology 103(1): 144–181.

30.

Müller-Peters

(1998) The significance of national pride and national identity to the attitude toward the single European currency: A Europe-wide comparison. Journal of Economic Psychology 19(6): 701–719.

31.

Qari

Konrad

Geys

(2012) Patriotism, taxation and international mobility. Public Choice 151(3): 695–717.

32.

Rabe-Hesketh

Skrondal

(2012) Multilevel and Longitudinal Modeling Using Stata, Volume I: Continuous Responses, 3rd Edition. College Station, TX: Stata Press.

33.

Shayo

(2009) A model of social identity with an application to political economy: Nation, class, and redistribution. American Political Science Review 103(2): 147–174.

34.

Smith

(1978) The diffusion of nationalism: Some historical and sociological perspectives. British Journal of Sociology 29(2): 234–248.

35.

Smith

Jarkko

(1998) National pride: A cross-national analysis. General Social Survey Cross-National Report No. 19. National Opinion Research Center, University of Chicago.

36.

Solt

(2011) Diversionary nationalism: Economic inequality and the formation of national pride. Journal of Politics 73(1): 821–830.

37.

Steenbergen

Jones

(2002) Modeling multilevel data structures. American Journal of Political Science 46(1): 218–237.

38.

Steenbergen

Marks

(2007) Evaluating expert judgments. European Journal of Political Research 46(3): 347–366.

39.

Wagner

Becker

Christ

. (2012) A longitudinal test of the relation between German nationalism, patriotism, and outgroup derogation. European Sociological Review 28(3): 319–332.

40.

Weber

(1976) Peasants into Frenchmen: The Modernization of Rural France, 1870–1914. Stanford, CA: Stanford University Press.

41.

Wimmer

(2012) Waves of War: Nationalism, State Formation, and Ethnic Exclusion in the Modern World. New York, NY: Cambridge University Press.

42.

Wimmer

Feinstein

(2010) The rise of the nation-state across the world, 1816–2001. American Sociological Review 75(5): 764–790.

43.

Wooldridge

(2002) Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: The MIT Press.