1 Introduction

In the 2016 State of the Union, Jean-Claude Juncker, the President of the European Commission, called for increased efforts towards construction of the European Pillar of Social Rights, stressing that Europe was not social enough and it needed a change in that respect (European Commission 2016). Building upon this topic, in January 2017, President Juncker conceptualized his earlier statement specifically within the framework of the minimum wage policy, stressing that while all member states are free to set the minimum wage in line with the local conditions, “There is a level of dignity we have to respect” (Guarascio 2017). Such calls have been reflected also in the demands of the European citizens (Maselli 2016). In this paper, we would like to contribute to the debate started by President Juncker by attributing a monetary value to the notion of the “minimum level of dignity” for the individual EU countries and asses the scope for adjustments needed to achieve this aim.

Societies struggle to find agreement on minimum wage (MW) policies, because of the assorted distributive effects. Economic literature on MW (Cahuc and Zylberberg 2004; Borjas 2015) and empirical research on the topic (Adams and Neumark 2005; Neumark 2014) by and large agree that while raising MW increases wages of low-income workers in particular, it also results in companies laying off workers (with low-income workers again being among the first to be fired) and decline in economic output.Footnote 1 The great ideological, institutional and political differences between the European Union (EU) member states result in the lack of consistence in setting minimum wages in Europe. Some societies are more willing to safeguard growth and employment, while others are more concerned about equality.

The struggle between different ideologies on the topic of MW policies is a healthy manifestation of democratic political order, where different values and arguments compete for the support of the public. Nonetheless, it is important to recognize it produces widely different outcomes. The relative minimum wages in the EU vary anywhere between 37 and 62% of the median wage (Schulten and Müller 2017). While this diversity of outcomes is not a problem by itself, as it represents legitimate outputs of a democratic process, there are robust normative and economic arguments (summarized in the “Literature review” section) which justify placing a base value, which is necessary to safeguard President Juncker’s “minimum level of dignity”, which we instrumentalize in this paper using the concept of the living wage (LW).

Our main intended contribution is to put an amount in Euro to LW for several European countries so that we can contrast the calculated amount with the national minimum wage. Based on this exercise, we present how several of the member states of the EU currently set the MW beneath the LW level as well as amounts of MW increase, which would be necessary in each of these countries to guarantee the LW at least for all Europeans working full-time. Importantly, in line with the LW literature (Anker 2011), the presented calculations should be understood as “absolute minimum amount needed for a dignified life” and nothing more. Under no circumstances should our number be interpreted as “sufficient” level of MW. On the contrary, as argued above, we fully appreciate the current diversity of approaches to MW as an expression of legitimate democratic will of European peoples and do not argue for any harmonization beyond ensuring that MW in no member state falls under the LW.

The following text consists of four main sections. Firstly, the “Literature review” section summarizes the history of thinking about LW, along with the main arguments behind the concept. Following is the “Data and estimation strategy” section, which presents the underlining reasons for using the WageIndicator Cost of Living survey, details how the LW amounts are calculated and how we approach the comparison with MW. The results of our calculation, along with policy implications, are presented and discussed in the “Results and discussion” section. Finally, the “Conclusions” section contains a succinct summary of the results, discussion of limitations of our approach and policy implications.

2 Literature review

The concept of just remuneration for work is as ancient as civilisation itself, being present in the writings of ancient Greeks and Romans, all the world’s great religions (in the modern times most notably in the Social Teaching of the Catholic church), classical economics, and the modern concept of Social Rights born with the establishment of the United Nations (Stabile 2009, 2016). Interestingly, in spite of the rich history, the meaning behind the LW concept has mostly been quite stable and straightforward, typically expressed as a wage necessary for survival based on the real cost of living (May 1982; Wills and Linneker 2014). More recently however, there have been attempts to enlarge the definition to include parameters of “decent life”, rather than mere survival, such as the capacity to support families, maintain self-respect and be free to participate in the civic life of the country concerned (Glickman 1999; Brenner 2002). At the same time, it has been stressed that too comprehensive calculation might result in a LW that is too easy to discredit through competitiveness concerns (Hirsch 2017).

Historically, the concepts of LW and MW were used interchangeably. Nonetheless, after the war, the meanings started to diverge, because workers receiving MW started being reasonably well protected from poverty, which became a problem concentrated among unemployed and underemployed workers (Stigler 1946). Nonetheless, the connection has recently reestablished as MWs started to lose buying power over time and falling under the rational definition of poverty as governments in different countries failed—for ideological or other reasons—to keep the minimum wages in line with living wages (Schenk 2001). Phenomena such as inflation and a rise in labour productivity provide profit opportunities mainly to high-wage workers, while low wages do not profit from increased wealth to a comparable degree (Pollin 2007; Piketty 2014). As a result, a liveable MW became a challenge again.

LW policy can be justified both on normative and economic/social good grounds. From a normative perspective, the argument for living wage is quite straightforward. According to arguably the most influential western concept of justice formulated by John Rawls as “justice as fairness”, inequalities are admissible only as long as they do not hurt the worst off in the society (Rawls 1999). From this it follows that a society generating substantial wealth, a condition EU countries fulfil, should provide decent living standards even for low-income workers. In addition to this normative argument, several economic arguments have been put forward in favour of securing decent income levels for low-income workers, including that it contributes to macroeconomic stability, maintenance of effective private demand and high level of employee’s satisfaction and productivity. Additional arguments are that in combination with welfare state, low wages lead to outsourcing of externalities on taxpayers, for example when the low-skilled workers are unable to save up sufficient resources for retirement (Stigler 1946; Kaufman 2010). Finally, as stressed by Piketty (2014), extreme forms of inequality are dangerous to democratic society. Sufficient MW as a redistributive policy in its core is one way to address this issue (Freeman 1996).

Politically, the efforts to guarantee LW have been connected mainly with the English-speaking world. The first modern LW campaign started in Baltimore, MD, USA, in 1994 through local legislation, following a proposal from a coalition of civil society actors. This campaign was successful and spawned a range of successors across the USA, as well as in the UK, Canada, Australia and New Zealand (Levi et al. 2002; Luce 2004). Interestingly, there has not been much spillover towards the continental Europe. Instead, the concept gained some prominence within the debate about fair incomes in low-income countries, particularly within the context of the Asian garment industry (Luce 2009). Table 1 contains an overview of some examples of LW initiatives.

Table 1 Examples of well-known living wage movements
Full size table

3 Data and estimation strategy

In general, there are two ways of estimating a threshold for minimum income: as some percentage of average income or through LW. This is what is often referred to as standard of living versus the minimum level of resources debate. In general, we have seen a shift in the policy debate from the minimum level of resources measured through a goods basket towards the relative standard of living conceptualization (Atkinson 2003). For example, the social inclusion agenda of the European Union had been based on criteria such as the top/bottom earners ratio or the share of population under the poverty line before and after social transfers (Atkinson et al. 2002). As a result, the shift towards LW heralded by President Juncker represents something new for continental Europe.

A shift towards using the LW as a poverty threshold, however, requires dependable data, which are short in supply outside of the UK, due to embeddedness of the LW debate in the English-speaking world. Consequentially, there are no internationally comparative official calculations of LW available that would cover the EU. The national statistical offices do collect price information to calculating statistical indicators, most prominently inflation. Nonetheless, normally only aggregated price indices are published rather than individual prices necessary to calculate the LW.

Particularly in Central and Eastern Europe, some countries calculate a “subsistence minimum”. For example, in SlovakiaFootnote 2, it is adjusted annually based on average earnings and prices of a basket of items consumed by low-income citizens. Since 2013, it has remained stable at about 200 EUR, about half of the minimum wage amount in the country. The last published amount in neighbouring Hungary, a country with a rather generally lower level of income than Slovakia, valid for 2015, equalled about 290 EUR and was widely considered rather low (Kiss 2015). In another neighbouring country, CzechiaFootnote 3, the amount is just about 120 EUR. Given the big differences between amounts in comparable countries and their low value, it is hard to see how these official figures could have been used in an international comparison.

The lack of official data sources is partly filled by several online tools. Perhaps the most famous of these tools is the MIT Billion Prices Project that collects online prices online to estimate price indices in 50 countries since 2008, which have achieved remarkable similarity with the official statistics (Cavallo and Rigobon 2016; Cavallo 2017). Nonetheless, this dataset is limited to supermarket prices and as such cannot be used to estimate LW, which necessarily contains items like housing, which are not traded in supermarkets.

Since 2014, the Amsterdam-based WageIndicator Foundation (WI) started collecting prices in nearly 90 countries through online surveys with the specific aim of being able to calculate LW globally (Guzi 2014). The project is based on a number of websites, which contain information about wages and working legislation, in each of the individual countries, visited by about 34 million people on an annual basis (Guzi et al. 2016). Every day, the visitors are asked to report the price of an item picked from a list of items through a banner. Those who click on the banner are then asked to report their location and the price of the “item of the day” and potentially some related items (see Fig. 1). Those who provide this information are further given a long list of items and asked to report the prices of as many items on the list as possible. Where relevant, the respondent can choose to provide price information for different package sizes.

Fig. 1
figure 1

The banner and the “item of the day” question on the British website for 7th Match 2017. Source: paywizard.co.uk

Full size image

Using the survey, WI collects price data on a continuous basis since 2014. For all items, which have at least ten observations, WI publishes the 25th percentile and median statistic updated twice a year and publishes them online.Footnote 4 Furthermore, WI calculates LW using a methodology rooted in the ILO approach (Anker 2005, 2011). The LW consists of three basis items necessary for life: food, housing and transportation. Food costs are calculated using approximately 50 food items, weighted by item groups based on their consumption in the country as defined by FAO. For housing, apartment rental outside of the city centre is considered. The transportation item is based on a monthly public transportation pass.Footnote 5 A 10% premium is added to cover other necessities such as clothing, hygienic needs, culture or civic participation. LW is calculated as a range, using the 25th percentile of prices as the lower bound and the median value as the upper bound. Even though the respondents report their precise location, for now, LW is published on the country level. The amounts are published in the national currency and the euro, using the average exchange rate for the 6 months preceding the publication of the data.

WI calculates LW on the individual and family levels. Family is assumed to have two adults (a male and a female) in a productive age and several children depending on a fertility rate in the country (in the case of the EU, countries vary between 1.3 and 1.9). For income, the employment rate for men and women in the country is considered (in the case of the EU countries, 1.6–1.8 income is assumed for a family). For an individual, the LW considers a one-room apartment, while for a family, a three-room apartment is deemed appropriate. For transportation purpose, it is assumed the household is located in an urban environment and there is a public transportation system available.

While WI is a valuable source, it is not a representative one, given that respondents are self-selected and the survey is only accessible through the Internet. There is a wide debate in the literature about the merit of such convenience sample-based surveys (Couper 2000; Dillman and Bowker 2001; Bethlehem 2010). A substantial number of empirical experiments with the WI survey generally show that the survey is appropriate for explorative research but not necessary for inferring relationship between variables (de Pedraza et al. 2010; Tijdens and Steinmetz 2016). As such, the WI datasets are increasingly used to particularly study topics which are hard to address using traditional data sources (Besamusca and Tijdens 2015, Guzi and de Pedraza 2015; Tijdens et al. 2015; van Klaveren et al. 2015). The limits and potential of using the WI data, as well as other web-based data sources, have been recently discussed in the IZA Journals Series as well (Mýtna-Kureková et al. 2015; Lenaerts et al. 2016).

We then compare LW with MW in European countries to see if the MW at least covers the minimum requirements for living. MWs are obtained from Eurostat. Unfortunately, we do not see any way to control for the effect of income taxes and social security deductions, given that they depend on many individual-level variables. It has been argued that this is not a major issue as the amounts tend to be quite modest for the low-income segment of the population (Guzi 2014) due to tax credit/personal allowance/zero tax rate threshold policies. Furthermore, people on low incomes tend to benefit more from welfare transfers. Nonetheless, it is likely that if we could take taxes into account, the calculated LW would be slightly higher.

4 Results and discussion

Looking at the data, we were able to calculate LW for 14 member states. Those that are missing either do not have a MW (Denmark, Italy, Cyprus, Austria, Finland and Sweden) or there is insufficient data collected so far through the WI survey to calculate the LW (Luxembourg, Malta, Lithuania, Estonia, Latvia, Ireland, Slovenia, Croatia).

The data that are available confirm the major centre-periphery divide (Table 2). In general, life in the core countries is quite expansive but matched with sufficient earnings to cover the needs. Meanwhile, in the periphery, it is not the case and MWs are commonly not sufficient to cover basic needs.

Table 2 Comparison of living and minimum wage
Full size table

In the Northwestern EU countries (Benelux, Germany, France, UK), even the upper range of LW rarely overcomes 80% of the MW threshold and the lower threshold can even go below 50%. In other words, in the core countries, MW earners can secure basic living necessities and still have 20–50% for additional expenses or savings. MW earners can afford to live in relative comfort, even though life in those countries is not cheap—the LW tends to amount to about 1000 EUR or more. Nonetheless, with MW starting at over 1300 EUR in the Netherlands and reaching up to over 1800 EUR in Germany, the MW earners are able to cover their needs. Among the peripheral countries, the upper LW threshold is above MW.

Peripheral countries can be divided to southern (Greece, Spain, Portugal) and eastern (Visegrad countries, Bulgaria and Romania). The southern group treats its MW earners relatively better. The MWs tend to be around 600 EUR, which is enough to cover life LW in Greece and is at a lower threshold of LW in Portugal and Spain. Nonetheless, the MW falls short of covering the upper edge of the LW, particularly in Spain, where LW is slightly higher than in Greece or Portugal. Meanwhile, in the eastern countries, the MW falls 10–60% short of covering even the lower threshold of the LW. In this group, we identify two distinct subgroups—the Visegrad countries which are similar to southern periphery in terms of living costs, but MW is only about 400 EUR, and Bulgaria/Romania, where LW as well as MW is lower.

The situations improve when looking at the family LW. Here, the lower threshold lies underneath the combined MW of the two adults in the household, and the upper threshold is above the MW in only three countries: Bulgaria, Hungary and Romania.

Looking at the graphical representation of LW and MW, we can see the four clusters quite clearly (Fig. 2). Two things are of note here—the substantial distance between the core and the periphery and the relatively substantial extent to which peripheral countries (other than Greece) must improve to achieve the minimum standard of dignity called for by president Juncker. Such an increase in minimum wage, of course, could potentially have a significant detrimental impact on the competitiveness of those countries, which have based their business model largely on the low-wage model of “embedded neoliberalism” (Bohle and Greskovits 2007, 2012; Greskovits 2015). As such, the figures presented in this paper relate not only to the minimum wage debate but also the larger debate about “variants of capitalism” in the EU.

Fig. 2
figure 2

Graphical distribution of countries based on individual MW and LW. Source: own calculation based on WI from December 2016. Q1 2017 Eurostat data used for MWs

Full size image

As a final point, when looking at the structure of LW per country, we see a degree of variation (Fig. 3). This is particularly the case when housing, as the single biggest parameter in the LW calculation, is concerned. The housing costs appear to form a disproportional share of LW, over 70% compared to about 65% in average, in the UK, along with three Visegrad countries: Czechia, Slovakia and Poland. In most countries, The UK balances out expensive housing with high incomes; nonetheless, in the Visegrad countries, this factor certainly contributes to MW insufficiently covering the LW.

Fig. 3
figure 3

Relative share of individual items in the lower threshold of LW. Source: own calculation based on WI from December 2016

Full size image

The policy implication of our analysis appears quite straightforward. If building a social pillar of the EU integration means a guarantee of minimum level of dignity for all, the MWs in the peripheral countries of the Union need to increase to cover at least the MW. Because MWs in the core countries of the EU generally significantly surpass the LW, a partial degree of convergence is sufficient. In addition, making housing more accessible to low earners might be an appropriate alternative policy response, particularly in the Visegrad countries. Given that an increased MW will likely cover the costs of basics, such as food and housing, such a policy can be expected to boost domestic demand. Particularly, in the eastern peripheral EU states, with relatively low unemployment rates, the case for MW increase appears to be clear cut.

5 Conclusions

The Social Pillar of the EU brings an impetus to think about MW from the perspective of the LW. Making wages “liveable”, guaranteeing workers the possibility to pursue basic liberties and reducing wealth and income inequality should be the main priority of EU institutions. In this paper, we present data, which support the claim that a Social Pillar of the EU should entail a hike in MW in the peripheral EU countries. Except for Greece, we found MW in all examined peripheral countries is lower than the LW for individuals and in the case of Hungary, Bulgaria and Romania also for families with more than one earner. In the core EU countries, the MW sufficiently covers the LW.

While our findings are clear in their general message, more work is needed to develop robust indicators of LW compliance in Europe. Firstly, the consumer price data collected by the statistical offices should be used to regularly calculate a LW for all EU member states according to a single methodology. The methodology should probably be more comprehensive than one presented in this paper and take other important costs, such as civic participation into account. It should also consider taxation and social transfer. At the same time, the LW should not be too high, leaving it up to member states to strike the right balance between equality and competitiveness. In this effort, the online-based calculations, such as WI, can provide a useful advance indicator, if they can be benchmarked against an official dataset.