Social Indicators Research

, Volume 114, Issue 2, pp 441–451

The Pursuit of Happiness in China: Individualism, Collectivism, and Subjective Well-Being During China’s Economic and Social Transformation


    • Department of SociologyPrinceton University
  • Scott M. Lynch
    • Department of SociologyPrinceton University

DOI: 10.1007/s11205-012-0154-1

Cite this article as:
Steele, L.G. & Lynch, S.M. Soc Indic Res (2013) 114: 441. doi:10.1007/s11205-012-0154-1


This paper examines the consequences of China’s dramatic socioeconomic and political transformations for individual subjective well-being (SWB) from 1990 to 2007. Although many still consider China to be a collectivist country, and some scholars have argued that collectivist factors would be important predictors of individual well-being in such a context, our analysis demonstrates that the Chinese are increasingly prioritizing individualist factors in assessments of their own happiness and life satisfaction thus substantiating descriptions of their society as increasingly individualistic. While the vast majority of quality of life studies have focused on Westerners, this study contributes findings from the unique cultural context of China. Moreover, concentration on this particular period in Chinese history offers insight into the relationship between SWB and rapid socioeconomic and political change.


Subjective well-beingHappinessLife satisfactionChinaIndividualismCollectivismCountries in transition

1 Introduction

Since the early 1990s, China has undergone a dramatic economic transformation and has averaged 10 % GDP growth per year. This state-led economic shift has been accompanied by changes in the government’s ideological rhetoric. According to Ong (1997), the market reforms launched in 1991 (“socialism with Chinese characteristics”) represent an attempt to establish greater state control over freewheeling capitalism and to promote the idea that socialism can use capitalism to increase the power of the Chinese nation.

What are the consequences of such economic and political transformations to well-being? Since the vast majority of quality of life studies have focused on Westerners, the insights offered by the unique cultural context of China along with its changing socioeconomic and political circumstances have been largely neglected (Shek 2010). Cheung and Leung (2007) suggest that collectivism may be more important for Chinese subjective well-being (SWB) than individualism. Yet, in his landmark cross-cultural study, Hofstede (1980) examined variation in individualism and collectivism and concluded that economic development encourages individualism. Indeed, Moore (2005) contends that the increasing freedoms that began in the 1980s have encouraged the Chinese to embrace a form of individualism “that stands emphatically opposed to the collectivist spirit promoted during the Cultural Revolution” (p. 357). Nonetheless, China is still generally considered to be a collectivist country. For example, Michailova and Hutchings (2006) describe the Chinese as expecting to subordinate their individual needs, goals, and aspirations to the requirements of the collective.

As China has experienced rapid social and economic shifts, have individualist or collectivist sentiments prevailed, and what have been the consequences in terms of SWB, among its people? We address this question by examining whether individualist or collectivist orientations best predict individual-level SWB and whether the relationship between these larger beliefs and well-being has changed over time during China’s transformation toward a market economy.

2 Background and Literature Review

A key feature of capitalistic ideology is the promotion of individualism, that is, the principle that individual effort is rewarded and benefits the individual. In contrast, a key feature of socialist ideology is the promotion of the good of the collective over the individual. China has been a socialist nation since the latter half of the twentieth century. Yet, the last 20 years have brought extensive economic reforms that have transformed the country into a pseudo-capitalist market economy. Has this economic transformation been mirrored by an ideological one? Wang (2002) claims that as long as the Chinese Communist Party (CCP) continues to govern in the name of socialism, it must continue to avert the emergence of values such as individualism that are clearly identifiable as central to (Western) capitalism. Indeed, some recent research suggests that China is still a relatively collectivist country (e.g., Michailova and Hutchings 2006; Lu and Gilmour 2004).

However, Wang (2002) argues that although the CCP cannot risk its legitimacy by promoting individualism, a moral code that “contain[s] some form of individualism and some corresponding conception of justice is necessary if a society that allows and encourages the pursuit of individual interests, as post-Mao China does, is not to degenerate” (p. 16). Other scholars (e.g., Moore 2005; Yan 2010) have described how social and economic changes in China have already transformed the country toward this end into a much more individualistic one. Yan (2010) described how the Maoist path to modernity was gradually negated and replaced by the language of market economics and privatization. He views the increasing abundance of choices for the individual as the most obvious change in Chinese society. By the turn of the twenty-first century, Yan argues that a twofold social transformation had taken place, encompassing both the rise of the individual and the individualization of the social structure. He contends that even those in socially disadvantaged positions now accept an ethic of personal responsibility for their successes and failures.

What is the consequence of the ideological transition from a more collectivist orientation to a more individualistic one to individual-level subjective well-being in the country? In a multi-country, macro-level study, Diener et al. (1995) found that, across four data sets, multiple measures of the concept of individualism—defined by personal freedom and the ability to pursue one’s own interests—were more frequently positively correlated with higher SWB than any other predictor across 55 countries. Thus, we might expect SWB to increase in China as it transforms. However, Brockmann et al. (2009) found that one of the characteristics of rapid transitions—that increasing inequality causes the relative financial position of many people to deteriorate despite absolute gains in income—depressed Chinese life satisfaction from 1990 to 2000. Moreover, at the micro level, transitions from a collectivist orientation to an individualistic one are almost certainly heterogeneous during the larger, macro level transition, and so, heterogeneity in SWB is also likely. For example, Di Tella and MacCulloch (2005) found that individuals declared themselves to be happier when the party they supported was in power. And Napier and Jost (2008) found conservatives to be happier than liberals throughout the world, with the differences being most pronounced in countries where the overall quality of life was relatively low. While data on left-wing/right-wing political orientation is not available for China, those who would be considered conservative would be the supporters of the CCP, which is also the only political party to have held power in China since 1949.

During the period we examine, expressing support for the CCP would have entailed embracing the ideology it was promoting most actively in practice—national pride. The 1989 anti-government demonstrations in Tiananmen Square highlighted the fact that the shift away from Maoist collectivism (Moore 2005) and toward capitalism required the CCP to further justify its legitimacy. Beijing identified the most unifying ideology to be national pride, the only important value shared by the regime and its domestic critics (including the Tiananmen demonstrators, who equated promoting democracy with patriotism). Hughes (2006) points out that the CCP leadership had to ensure that the version of national pride that was disseminated was compatible with market reforms. Thus, the CCP defined its role in this process as the guardian of national pride, promising to enhance the collective image of Chinese society while providing political stability and economic prosperity (Zhao 1998).

Having adopted the advancement of nationalism as the new strategy for holding the country together during what was expected to be a rapid and turbulent transformation, Beijing launched an extensive propaganda campaign in the early 1990s to educate the people in patriotism. Zhao (2005) describes how students at high schools and universities were taught to be proud of being Chinese by concentrating on the “great achievements” of the Chinese people. Indeed, Gries (2004) defines national identity in China in very collectivist terms, as “that aspect of individuals’ self-image that is tied to their nation, together with the value and emotional significance they attach to membership in the national community” (p. 19). Given the established links between national politics and individual well-being, the importance of collectivist ideologies in China’s recent history, and the even more recent attempts by the government to integrate specific nationalist goals with individual self-interest, we expect to find a relationship between individual SWB and contemporary manifestations of collectivism.

Recent research has painted a more nuanced picture of the relationship between collectivism (nationalism) and SWB. Jagodzinski (2010) demonstrated that at the individual level, respondents’ national pride was an important predictor of happiness in both Europe and Asia. At the country level, in Asia the decline in national pride was associated with dissatisfaction. Moreover, Inglehart et al. (2008) found, for example, that happiness is related to collectivistic orientations—in particular in-group solidarity, religiosity, and national pride—more strongly in less-developed countries than in more developed ones. In more developed countries, the importance of free choice surpasses that of solidarity in predicting happiness, which is consistent with Diener et al.’s (1995) finding about the importance of personal freedom.


Individuals who are more collectivist, specifically those with greater national pride and those who support collectivist economic policies, will have higher SWB scores. However, we also expect that the importance of collectivism will have declined as the transformation toward capitalism has transpired.

Decades of quality of life research have documented the relationship between individual characteristics and SWB. In Van Hoorn’s (2007) synthesis of this research, he found that the variables pertaining to individual status that had most consistently corresponded to higher levels of SWB included having better health, being married (as opposed to being single, divorced, widowed, etc.), and being female. He also determined that the relationship between age and SWB displayed a U-shape; it was higher among young people, declined among the middle aged, and increased again at older age.

Regarding the much-debated relationship between income and SWB (for example, see Easterlin 1995 versus Stevenson and Wolfers 2008), Van Hoorn found that within a given country, findings had most consistently shown that richer people on average reported higher SWB than did poorer people in the same country, though increasing per-capita income within a country over time was not associated with higher SWB. Thus, for the most part, those who generally have the highest status in society—healthy, married, middle-aged individuals with high incomes—have the highest SWB.


Drawing on these findings as well as Moore and Yan’s theories, we hypothesize that factors reflecting individualist concerns, such as freedom of choice and measures of individual status, predict higher SWB and will have become more important predictors of SWB over time as China has transformed toward a market economy.

3 Data and Methods

3.1 Sample

We use data from the World Values Survey (WVS) for our analyses. The WVS is a repeated, multi-country cross-sectional survey that has been conducted in China in 1990, 1995, 2001, and 2007. We exclude the 1995 wave in our analysis because its sampling frame differed dramatically from those of the other three waves. Although there were several design changes across the waves of the study, we were able to restrict the remaining three samples to represent similar subpopulations. Ultimately, our pooled sample is representative of persons ages 18–65 years living in urban China. The original pooled sample size was 4,015. After removing all agricultural (i.e., non-urban) workers and those over age 65, 2,287 observations remained. We use only cases with complete information in the analyses, leaving us with an analytic sample of size n = 1,798. (1990 n = 884; 2001 n = 350; 2007 n = 564). The majority of missing data are scattered across variables in our models; however, income and support for collectivism had considerable missingness. Logistic regressions predicting missingness on income and support for collectivism showed that employed persons were more likely to report their income and views about collectivism than those who were not employed (retired, housewives, students, unemployed, or other) and women were less likely to report their views about collectivism, but no other differences between the missing and nonmissing were observed.1 Ultimately, we opted to use listwise deletion of missing data, because most contemporary methods for handling missing data assume the missing data are missing at random—an assumption that cannot be evaluated—or rely on other assumptions that cannot be tested.

3.2 Measures

Our key dependent variable is an index of two measures of subjective wellbeing: happiness and life satisfaction. Happiness was measured via the question: “Taking all things together, would you say you are [very happy, quite happy, not very happy, or not at all happy]?” Life satisfaction was measured via the question: “All things considered, how satisfied are you with your life as a whole these days,” measured on a scale from 1 to 10, with 1 representing “dissatisfied” and 10 representing “satisfied.” We combined these measures by standardizing them within each survey year and summing them. The Cronbach’s alpha was 0.33. Although this alpha is quite small, analyses using one measure versus the other led to the same substantive conclusions presented here, which were obtained using the combined index.

Our key research question concerns whether, as China has transitioned to become increasingly capitalist, individualistic (vs. collectivistic) factors have become increasingly important in predicting SWB. Thus, our key predictors include “individualist” measures and “collectivist” measures. Our key individualist measures include marital status, self-rated health, income, employment status, and freedom of choice and control over life.2 Marital status was measured with a dummy variable reflecting the marital status of the respondent (married or unmarried) at the time of the interview. Employment status was measured with a dummy variable reflecting employment status of the respondent (full-time, part-time, or self-employed) at the time of the interview. Self-rated health was measured with a five category Likert scale in 1990 and 2001 (very poor, poor, fair, good, and very good) but with a four category scale in 2007 (poor, fair, good, and very good). We collapsed the former two categories in the early waves, yielding a four category measure for each wave. Income was measured in wave-specific deciles. Sense of freedom was measured with the following question: “Some people feel they have completely free choice and control over their lives, while other people feel that what they do has no real effect on what happens to them. Please use this scale where 1 means ‘none at all’ and 10 means ‘a great deal’ to indicate how much freedom of choice and control you feel you have over the way your life turns out.”

Our key collectivist measures include an item measuring one’s feeling of national pride and an index measuring one’s support for collectivist policies. National pride was measured with a single item asking how proud the respondent feels to be Chinese (1 = not at all proud, 2 = not very proud, 3 = quite proud, 4 = very proud). Support for collectivism was measured with an index constructed from three items. One item asked whether the respondent believes that “income should be made more equal versus we need larger income differences as incentives.” The second asked whether “people should take more responsibility to provide for themselves versus the government should take more responsibility to ensure that everyone is provided for.” The third asked about agreement with the statement: “Competition is good. It stimulates people to work hard and develop new ideas verus competition is harmful. It brings out the worst in people.” The Cronbach’s alpha is .30 for the three-item index, which has a range from 1 to 10, with 10 indicating strong support for collectivist policies.3

Descriptive statistics for all variables used in the analyses are shown in Table 1. The pooled sample is 56 % male with a mean age of 38.6 years and with 78 % of respondents being married. Mean happiness increased over time while mean life satisfaction decreased. However, the patterns are both u-shaped; the difference in the linear trends results from the fact that happiness has a higher mean in 2007 than in 1990 while life satisfaction has a higher mean in 1990 than in 2007. The standardized SWB variable also follows a u-shaped pattern over time but trends slightly upward.
Table 1

Descriptive statistics for variables used in analyses (World Values Survey, 1990, 2001, and 2007








Mean or  % (SD)

Mean or  % (SD)

Mean or  % (SD)

Mean or  % (SD)


[1, 4]

2.97 (0.73)

2.94 (0.81)

2.91 (0.62)

3.05 (0.66)

Life satisfaction

[1, 10]

7.08 (2.22)

7.28 (2.10)

6.46 (2.49)

7.15 (2.17)

Subjective well-being

[−2.59, 1.39]

0.06 (0.79)

0.08 (0.76)

−0.11 (0.83)

0.13 (0.79)



55.90 %

59.95 %

51.71 %

52.13 %


[18, 65]

38.60 (12.48)

38.34 (12.82)

38.63 (11.48)

39.00 (12.55)



78.09 %

78.28 %

82.29 %

75.18 %


[1, 10]

4.35 (2.25)

3.22 (1.70)

6.98 (1.84)

4.49 (1.79)



79.03 %

87.56 %

68.57 %

72.16 %

Self-rated health

[1, 4]

2.89 (0.93)

2.82 (0.94)

2.78 (0.94)

3.06 (0.90)

Freedom of choice/control

[1, 10]

7.15 (2.21)

7.07 (2.12)

6.84 (2.49)

7.46 (2.14)

National pride

[1, 4]

3.15 (0.76)

3.23 (0.78)

3.06 (0.75)

3.07 (0.71)

Support for collectivism

[1, 10]

4.08 (1.74)

3.57 (1.58)

4.44 (1.75)

4.67 (1.72)

Number of respondents






3.3 Analytic Strategy

We estimated a series of linear regression models to address the hypotheses. First, we estimated a linear model for the pooled sample. Next, we estimated separate models by year in order to determine the extent to which the relationships between individualist and collectivist views and SWB varied across time.

After estimating the regressions, we illustrate the results by generating predicted values for SWB across time by four combinations of minimum and maximum values of the key predictors: minimum individually and minimum nationally oriented scores (income = 1; self-rated health = 1; support for collectivism = 1; national pride = 1); maximum individually and minimum nationally oriented scores (income = 10; self-rated health = 4; support for collectivism = 1; national pride = 1); minimum individually and maximum nationally oriented scores (income = 1; self-rated health = 1; support for collectivism = 10; national pride = 4); and, maximum individually and maximum nationally oriented scores (income = 10; self-rated health = 4; support for collectivism = 10; national pride = 4). Other variables were set to their means in these calculations.

4 Results

Table 2 presents estimates of regression coefficients for the pooled sample and by wave. In the pooled regression model, we find that the coefficient for the year 2001 is negative and statistically significant, suggesting that there may be a period effect. The coefficients for marriage and employment were positive and highly significant in the pooled model. Models by wave show that they are much smaller and non-significant in 1990, becoming larger in 2001, then more significant in 2007. Thus, the importance of being married and employed for SWB increased over time.
Table 2

OLS regression models of subjective well-being by individualist and collectivist predictors


Pooled coefficients

1990 Coefficients

2001 Coefficients

2007 Coefficients

Year 2001

−0.240*** (0.057)


Year 2007

−0.005 (0.041)




0.182*** (0.051)

0.028 (0.078)

0.280** (0.128)

0.305*** (0.075)


0.042*** (0.009)

−0.013 (0.014)

0.057*** (0.021)

0.111*** (0.015)


0.119*** (0.045)

0.019 (0.087)

0.078 (0.093)

0.127** (0.064)

 Self-rated health

0.198*** (0.018)

0.138*** (0.026)

0.129*** (0.043)

0.329*** (0.031)

 Freedom of choice/control

0.101*** (0.008)

0.096*** (0.011)

0.121*** (0.017)

0.091*** (0.013)


 National pride

0.179*** (0.022)

0.210*** (0.031)

0.124** (0.053)

0.141*** (0.036)

 Support for collectivism

−0.035*** (0.010)

−0.039*** (0.015)

−0.029 (0.023)

−0.014 (0.016)



−0.077** (0.033)

−0.030 (0.048)

−0.131* (0.078)

−0.067 (0.053)


−0.018* (0.010)

0.020 (0.016)

−0.053** (0.026)

−0.033** (0.015)


0.000* (0.000)

−0.000 (0.000)

0.001** (0.000)

0.000** (0.000)


−1.681*** (0.207)

−1.949*** (0.303)

−1.101** (0.543)

−2.162*** (0.341)











p < .05, ** p < .01, *** p < .001 (two-tailed test)

Standard errors in parentheses

In the pooled sample, higher incomes and self-rated health, greater feelings of freedom of choice and control over life and national pride predict greater SWB, with positive and statistically significant coefficients. Support for collectivism, which has a negative and significant coefficient, predicts lower SWB.

In the wave-specific models, the coefficient for income is small and non-significant in 1990, increases and becomes statistically significant in 2001, and then becomes even larger in 2007. The coefficients for self-rated health are significant in all waves and almost identical in 1990 and 2001, with a marked increase in 2007. Conversely, from 1990 to 2001, the coefficient for national pride decreases in magnitude, although it remains significant in all waves. The coefficients for support for collectivism, which are negative, tend toward zero across waves, becoming nonsignificant by the second wave. The pattern of the coefficients for the variable measuring freedom of choice and control over life, which are positive, can be interpreted in two ways: (1) Assuming the true pattern is linear, the coefficients show that the relationship between freedom of choice and SWB is stable; or (2) assuming the pattern is nonlinear, there is a slight inverted u-shaped pattern in the relationship between freedom of choice and SWB.

In terms of the control variables, the coefficient for gender was negative and statistically significant in the pooled model and in the 2001 model. The coefficients for age and age-squared were statistically significant in the pooled model, the 2001 model, and the 2007 model; in those models, being older has a negative effect on SWB, although the relationship between age and SWB is actually u-shaped.

To summarize, the regression results show that more individualistic explanatory factors—income, marital status, employment status, and health—have become more important for SWB over time; one individualist factor—freedom of choice and control—has remained stable over time; and factors reflecting respondents’ collectivist sentiment—national pride and support for collectivist policies—have become less important. This finding is substantiated through examination of predicted values of SWB at each point in time for high and low values of both individualistic and collectivist factors (see Fig. 1). A 40-year-old employed married male with the highest scores for income (10), self-rated health (4), and freedom of choice and control over life (10) had higher and increasing SWB scores over time regardless of whether scores for support for collectivism or national pride were set at their maximum or minimum values. Conversely, those with low individualism scores had lower and decreasing SWB scores over time, even when collectivism scores were high.
Fig. 1

Predicted values of subjective well-being over time by individualist and collectivist predictors

5 Discussion

Do collectivist or individualist factors better predict SWB in China, and has their relative importance changed over time during China’s transformation toward a market economy? Our first hypothesis (H1) stated that we expected individuals who are more collectivist to have higher SWB scores, although we expected the importance of collectivism to have declined over time. The results reported above support the latter proposition in that the coefficients and statistical significance of both national pride and support for collectivist economic policies decrease over time. While those with higher national pride scores do indeed report greater SWB, albeit less so over time, greater support for collectivist economic policies is associated with lower SWB in 1990 and the pooled model.

In fact, the descriptive statistics reported in Table 1 reveal that support for collectivism has actually increased over time as China’s policies have shifted farther and farther away from that ideology. This increasing support may be one manifestation of the rapid pace of China’s transformation. However, if the Chinese are dissatisfied with the changes that have occurred, and this dissatisfaction has affected their outlook, our results suggest that they no longer relate this outlook to assessments of their own well-being. One possible explanation is that for those who are the most collectivist—those who are most invested in society’s well-being—their own personal SWB may not be an important consideration. Another potential explanation is that this trend reflects confusion. The government promotes collectivism rhetorically while implementing economic and social policies that are antithetical to it. Moreover, the generations that were obligated by political and social forces to reject individualism in their youth remain loath to return to the coercive collectivism of that era (Moore 2005). The Chinese may be more collectivist in theory, but in practice they increasingly tie individualistic factors to their valuations of their own SWB. Indeed, our second hypothesis (H2)—that factors pertaining to individualism would predict higher SWB and would have become more important predictors of SWB over time—is well supported by the findings about the increasing importance of income, self-rated health, and marital and employment status.

The above-mentioned negative coefficient of the year 2001 in the pooled regression model suggests that there is something unique about that particular period. Although scholars such as Brockmann et al. (2009) wondered why Chinese SWB declined so dramatically from 1990 to 2001 as standards of living were increasing, our descriptive results show that there is a u-shaped pattern with SWB rebounding to 1990 levels by 2007. There are a number of reasons why the late 1990 s/early 2000 s may have been a uniquely unsatisfactory period in recent Chinese history. Huang (2008) describes how President Jiang Zemin and Premier Zhu Rongji ushered in the “Shanghai model” of development after the 1989 Tiananmen Square protests; while China’s cities sprouted gleaming skyscrapers, foreign investment exploded and GDP continued to grow, growth in average household income and poverty eradication slowed and income differences and social tensions widened. By the late 1990s, the government had begun to take notice of some of the potential ramifications. The State Council launched a series of income redistribution programs after the Asian financial crisis of 1997/1998 to boost domestic demand (Hung 2008). More comprehensive policy shifts were ushered in with the change in leadership at the 2002 16th Party Congress. Hu Jintao’s regime came to power and began placing more emphasis, at least rhetorically, on social harmony by introducing the notion of a “Xiaokang Society,” which maintains high growth as a goal but also incorporates attention to equity, social harmony, and sustainable development (Gustafsson et al. 2008).

6 Conclusions

Our findings show that while both individualist and collectivist factors predict SWB in China, individualist factors have become more important over time, and collectivist factors have become less important. These results substantiate the link between higher SWB and individualism found by Diener et al. (1995).

Although many still consider China to be a collectivist country, and some scholars have argued that collectivist factors would be important predictors of individual well-being in such a context, our analysis demonstrates that the Chinese are increasingly prioritizing individualist factors in assessments of their own happiness and life satisfaction. This substantiates descriptions of Chinese society as becoming increasingly individualistic (Moore 2005; Yan 2010).

Adherence to the national ideologies promoted by the CCP, which Hughes (2006) describes as having been defined as synonymous with loyalty to the nation, has waned over time, as has this factor’s ability to enhance individual SWB. Concomitantly, improvements in the areas that have become more important for SWB over time, including household income and individual self-rated health, may be unraveled if disparities continue to grow. Although Whyte (2010) finds that, as recently as 2004, increases in market-oriented policies and socioeconomic inequality were broadly accepted by the Chinese, Huang (2008) points out that since the early 2000s, growth in average household income and poverty eradication have slowed while income differences and social tensions have widened. While overall health in China has continued to improve on average (UNDP 2010), inequities in the health care system are increasing in some cases (Xu et al. 2007).

Wang (2002) describes a mismatch in China’s moral culture; the official code continues to invoke collectivistic values while the increasingly capitalistic economic order encourages individuals to pursue their own interests in competition with others. Our findings show that while support for collectivist economic policies has increased in tandem with the transition to a market economy, the link between these attitudes and individuals’ evaluations of their well-being has diminished. Thus, the individualist moral code may be trumping the collectivist one in terms of what is most important to individuals in their everyday lives.


In addition, when income and support for collectivism were excluded from the models, which yielded a much larger analytic sample of size n = 2,071, none of the substantive findings for the other variables in the models differed.


We were unable to include an educational attainment measure due to the absence of a consistent and meaningful measure across waves of the WVS. The only measure of education that was included in all waves of the WVS in China was the age at completion of education, which is not particularly meaningful. Moreover, when this variable, which has considerable missingness, was included in the models presented in this paper, it did not alter any of the substantive conclusions presented.


When these three items are included in the models separately, the results do not differ substantively from those reported in this paper.



The authors would like to thank Martin Ruef and Gilbert Rozman for providing invaluable comments on early drafts of this paper. We are also grateful for the constructive suggestions from the SIR editor and anonymous reviewers. The content is solely the responsibility of the authors.

Copyright information

© Springer Science+Business Media Dordrecht 2012