2.1 The social investment paradigm

The concept of a social investment welfare state developed out of acknowledgement that existing welfare structures, developed for the post-war industrial era, were insufficient for the new post-industrial economy of the late twentieth century. The increased flexibility of a post-industrial economy and a change in social values and norms has led to new social risks in contemporary societies that the Keynesian welfare state has been unable to meet (Bonoli, Reference Bonoli2005). The post-war welfare state was built around a single male-breadwinner-based industrial economy and its policies were geared towards income maintenance when faced with ill-health, old age and unemployment (Sainsbury, Reference Sainsbury1996). However, the social investment welfare state is based around not only income maintenance, but also increasing labour market participation and opportunities for workers and citizens with the goal of simultaneously increasing economic growth.Footnote ¹ This represents a clear break from the neoliberal welfare perspective of the 1970s and 1980s that saw welfare as a hindrance to the goal of economic efficiency (Morel et al., Reference Morel, Palier, Palme, Morel, Palier and Palme2012a).

The social investment strategy aims to shift the economy towards a high productivity knowledge-based economy through more inclusive utilisation of human capital, with emphasis on increased female labour market participation. In the short term, investing in the working-age population via active labour market policies that promote re-training and employment services can create better employment opportunities for low-skilled workers, who are at greater risk of social exclusion and poverty due to de-industrialisation (Bonoli, Reference Bonoli, Morel, Palier and Palme2012). In the longer term, the goal is to improve the life-chances and opportunities available to future generations. This is achieved by greater investment into education, especially in early childhood, which is argued to limit the risks of poverty more effectively than later investment in adults (Heckman and Lochner, Reference Heckman, Lochner, Danziger and Waldvogel2000). Investment in tertiary education – both in teaching and research – is also emphasised in recognition of the importance of a high-skilled workforce and technology-led growth in the post-industrial economy (Lundvall and Lorenz, Reference Lundvall, Lorenz, Morel, Palier and Palme2012).

Given the strong push towards adopting social investment packages by national governments, international organisations – e.g. the EU and OECD – and social policy professionals (Nolan, Reference Nolan and Hemerijck2017), there has been considerable research into the positive outcomes of social investment. Increasing women's labour market participation is a crucial goal for the social investment approach due to its positive impact on not just economic productivity but also poverty reduction. There is consistent evidence that households with dual-incomes have a reduced risk of poverty and social exclusion (Nolan and Marx, Reference Nolan, Marx, Salverda, Nolan and Smeeding2009). Consequently, policies that look to remove barriers to female employment, such as public provision of childcare and parental leave, form a key part of social investment packages (Esping-Andersen et al., Reference Esping-Andersen, Gallie, Hemerijck and Myles2002; Pavolini and van Lancker, Reference Pavolini and van Lancker2018). Furthermore, increases in early childhood education and human capital spending have also been shown to have positive effects on employment levels overall, but especially so in knowledge-intensive employment fields (Nelson and Stephens, Reference Nelson, Stephens, Morel, Palier and Palme2012).

2.2 Social investment in Japan

As with other areas of welfare research, research into social investment suffers from Eurocentrism, with literature predominantly focused on policies of European governments and the policy focus of the EU. Furthermore, Japan's place within the broader literature on comparative welfare is also somewhat ambiguous. As the earliest developed East Asian economy, Japan was included in many pieces of early research into welfare states (e.g., Esping-Andersen, Reference Esping-Andersen1990), but most early comparative approaches to welfare policy saw Japan as an exception and it was rarely discussed in depth. Similarly, as research into East Asian welfare states has become more common, the Japanese case has again found itself somewhat outside the comparative literature which has been more focused on China, the ‘tiger economies’ – South Korea, Taiwan, Singapore and Hong Kong – and the newly industrialising countries of South East Asia – Thailand, Malaysia and Indonesia (e.g., Kwon, Reference Kwon2009; Furuto, Reference Furuto2013).

Furthermore, recent volumes with the stated goal of expanding the discussion of social investment beyond Europe have also not included Japan (Midgeley et al., Reference Midgeley, Dahl and Wright2017; Deeming and Smyth, Reference Deeming and Smyth2019). One reason for this is that Japan has lacked ‘explicit public policy discourse on social investment’ (Peng, Reference Peng2011: 44) and has comparatively little implementation of social investment policy (Fleckenstein and Lee, Reference Fleckenstein, Lee and Hemerijck2017b; Miura, Reference Miura2018). However, this is not to say that social investment policies do not exist in Japan. In fact, given that economic growth in Japan has been stagnant since the 1990s and there is a strong insider–outsider division in the labour market between regular and non-regular workers (Thelen and Kume, Reference Thelen and Kume2006; Song, Reference Song2012), social investment policies would appear to target important social issues, especially since the relative poverty rate in Japan has rapidly increased since the 1980s (Miura, Reference Miura2012). Policies promoting better work–family balance would also tackle the issues of a rapidly ageing society and conservative attitudes towards gender roles: Japan's fertility rate in 2018 was 30th out of 37 OECD countries (OECD, 2020) and between 2005 and 2009, 62% of women quit their jobs upon becoming pregnant (Gender Equality Bureau, 2015).

However, the implementation of social investment policies has been lacklustre and often failed to deliver necessary changes. The provision of childcare facilities was noted as key to tackling the issue of declining birth rate in the 1990s (Ministry of Health and Welfare, 1996). However, although strides were made to improve the public provision of childcare services in the 1990s and early 2000s, this progress has been arrested and policy has reverted to focus on more traditional child benefits (Fleckenstein and Lee, Reference Fleckenstein and Lee2017a). Consequently, public childcare and early education services could provide places for just 32% of children aged 0–5 in 2009 (An, Reference An2013). Similarly, in 2003 the Japanese government set a target of having 30% of leadership roles be filled by women in both public and private sector employment by 2020, but the target date was recently pushed back to 2030 after the government found it was well short of meeting its goals – 5.3% in leadership positions in the public sector in 2019 and 8.4% in the private sector in 2018 (Hori, Reference Hori2020).

One important factor to note when discussing social investment policies in the Japanese context is that the term ‘social investment’ is rarely used and little understood outside of academic circles (Miura, Reference Miura2018). Direct translations of social investment into JapaneseFootnote ² are associated with either crowd-sourcing investment for start-ups or the corporate social responsibility schemes of large enterprises. As a result, social investment policies are not presented as forming part of a complete policy package as is common in the EU and its member countries (Astor et al., Reference Astor, Fransen, Vothknecht and Hemerijck2017). Instead, they are individual policies that tackle separate unrelated issues, e.g. the expansion of public childcare services in Japan is geared towards improving child welfare and is not conceptually linked by bureaucrats and politicians to increasing women's labour market participation (Boling, Reference Boling2015).

3.1 Who are the supporters of social investment?

The key finding from previous research into voter attitudes towards welfare is that attitudes are highly dependent on the characteristics of voters. Overwhelmingly, it has been shown that voters from economically disadvantaged groups – low income, low educational attainment – are the strongest supporters of welfare (Blekesaune, Reference Blekesaune2007; Pfeifer, Reference Pfeifer2009). However, the social investment literature expects there to be a difference between the supporters of traditional social transfers and the supporters of social investment, with wealthier middle-class voters being the core supporters of the latter (Gingrich and Häusermann, Reference Gingrich and Häusermann2015). This is due to their more secure socio-economic standing, which allows them to make long-term wagers on policy returns that the working-class cannot (Beramendi et al., Reference Beramendi, Häusermann, Kitschelt and Kriesi2015).

Voters' normative beliefs have also been shown to be correlated with attitudes towards welfare. Voters who believe in egalitarianism are more likely to hold positive views on social transfers, whereas voters who believe that welfare leads to state dependency hold more critical attitudes, which reflects known left- and right-wing attitudes towards the welfare state (van Oorschot, Reference van Oorschot2006). However, this left-right economic dimension should not to be conflated with attitudes on the social libertarian-authoritarian dimension, which consistently finds authoritarian voters to be more supportive of traditional transfer policies (Staerklé et al., Reference Staerklé, Likki, Scheidegger and Svallfors2012). By contrast, the middle-class supporters of social investment are expected to hold social libertarian values (Kitschelt, Reference Kitschelt1993; Inglehart, Reference Inglehart1997).

Using a survey of voters in eight European countries, Garritzmann et al. (Reference Garritzmann, Busemeyer and Neimanns2018) provide information on the support for different social policy packages – social investment, passive transfers and workfare – by asking respondents for their opinion on hypothetical welfare reforms. Overall, social investment is found to be the most popular package followed by workfare and lastly passive transfers. Supporters of the social investment package were more likely to be highly-educated and hold economically left-wing and socially liberal views, whereas voters favourable to passive transfers were more likely to have lower levels of income and education and hold more authoritarian views.

Research into Japanese voter attitudes towards welfare also shows similar findings to that found in European contexts (Sumino, Reference Sumino2014), and as a result, we would expect voter attitudes towards social investment in Japan to also follow the patterns found in Europe. Finally, given that many social investment policies – including those we investigate in this paper – aim to increase the economic participation of women, it is likely that gender will also impact on voter attitudes to social investment, especially given Japan's conservative attitudes towards gender roles. Based upon this we can derive the following hypotheses for testing in this paper:

3.2 When do voters support social investment?

The literature on voter attitudes to social welfare above finds a distinction between the supporters of social investment and traditional social security policies. This dichotomy between newer investment policies and traditional security lies at the heart of much research into government social policy choices (Barbier, Reference Barbier and Hemerijck2017). Budgetary pressures mean that governments who seek to introduce social investment reforms to the welfare state, may have to do so at the expense of social security (Cantillon and van Lancker, Reference Cantillon and van Lancker2013), although there is little evidence that social security has been replaced by investment in government spending patterns (Hudson and Kühner, Reference Hudson and Kühner2012).

However, how voters would react to the introduction of social investment policies at the expense of social security policies is an open question. The literature on welfare state reform has provided a convincing argument that welfare states are path dependent and institutionally sticky in part because of the popularity amongst voters for traditional social security entitlements (Pierson, Reference Pierson2001). Research into European voter attitudes has shown that support is high when voters are asked about social investment in isolation (Busemeyer et al., Reference Busemeyer, Garritzmann, Neimanns and Nezi2018), but that once the introduction of social investment is portrayed as part of a trade-off with social security policies, the level of support drops significantly (Busemeyer, Reference Busemeyer and Hemerijck2017). Consequently, we expect there to also be objections amongst Japanese voters to the introduction of social investment policies if they come at the expense of social security benefits:

5.1 Characteristics of social investment supporters

Using the conjoint analysis, we can also see how the characteristics of voters impacts on their support for candidates proposing social investment policies. Voters from higher socio-economic backgrounds are expected to be key supporters of social investment policies (hypothesis 1) and in order to investigate this we examine voter household income levels (Figure 2).Footnote ³ We see evidence that there is a decreased likelihood of middle-income voters supporting candidates promising retrenchment in all three of our social investment policies – although there is no corresponding greater likelihood of support for candidates promising increased investment. We also see decreased likelihood of support by high-income voters for candidates reducing education and childcare spending and also an increase in support for candidates promising childcare investment. Although the findings are slightly mixed, this does provide some evidence for the argument that middle-class voters are supporters of social investment policies – or at least oppose its retrenchment.

Figure 2.

Income differences in conjoint analysis social investment preferences. The AMCEs conditional on income and 95% confidence intervals are displayed. The other attributes are omitted from the figure.

When looking at other characteristics, voter gender (hypothesis 3) appears to be the clearest factor impacting on support for social investment policies (Figure 3). We see candidates who promise increased spending on any of our three social investment policies should expect an increase in support from women. However, the same effect is not found for male respondents who are not expected to support candidates that promise any social investment increases. Looking at the ideological values held by voters (hypotheses 2a,b), we find evidence that candidates promising education and childcare spending increases receive greater support from social libertarian respondents, whereas authoritarian respondents are not significantly likely to increase their support for social investment promoting candidates. However, the economic values of voters do not appear to impact on their support for social investment policies.Footnote ⁴

Figure 3.

Gender differences in conjoint analysis social investment preferences. The AMCEs conditional on gender and 95% confidence intervals are displayed. The other attributes are omitted from the figure.

5.2 Social investment and social security

One of the key findings in the literature is that support for social investment policies declines if it is placed in a trade-off with social security policy (hypothesis 4). By subsetting our conjoint experiment data according to the position of hypothetical candidates on social security spending, we can see the differential impact of this on support for candidates who promise social investment increases.

In Figure 4, we can see that a candidate's position on social security spending is a critical factor for our respondents in determining whether they will support increased social investment spending. The analysis shows that candidates receive greater voter support for social investment increases – specifically in childcare and education – if they also promise an increase in social security spending. However, candidates who promise to make cuts to social security (or even maintain current spending levels) do not receive increased support as a result of promising increased social investment. We do not, however, see a decrease in support for candidates who promise social investment increases while also promising decreases in social security spending. This evidence suggests that instead of a trade-off between support for social investment and security, a prerequisite of winning voter support for social investment spending appears to be also promising increases in traditional social security policies.

Figure 4.

Support for social investment modified by social security spending position. The AMCEs conditional on social security spending and 95% confidence intervals are displayed. The other attributes are omitted from the figure.