Volume 52, May/June 2017, Number 3 · pp. 159-164
Ten Lessons from a Decade of Debating an EUBS: Robust Findings, Popular Myths and Remaining Dilemmas
Sebastian Dullien, HTW Berlin – University of Applied Sciences, Germany; and European Council on Foreign Relations, Berlin, Germany.
This article is part of the Forum A Fiscal Stabilisation Function for the Eurozone.
The concept of a European unemployment insurance or a European unemployment benefit scheme (EUBS) has seen a number of ups and downs. The first mention of the idea dates back to the 1970s, when the Marjolin report proposed a "community unemployment benefit fund" (among other measures) to guarantee a smooth working of a potential common currency.1 Roughly 20 years later, in the preparatory technical documents for the European Economic and Monetary Union (EMU), a scheme similar to those discussed today under the term "equivalent EUBS" was analysed (see below for the distinction between "genuine" and "equivalent" EUBS).2 However, the ideas failed to get political traction. As a consequence, throughout most of the preparation for and run-up to EMU, as well as for most of the first decade of the euro's existence, there was not much mention of European unemployment insurance.
The modern debate on a potential EUBS began in earnest around a decade ago, just prior to the onset of the global financial crisis. In the mid-2000s, current account imbalances and diverging unit labour costs triggered a new debate about transfer systems between euro area countries and, particularly, a European unemployment insurance.3 However, the general perception was that no transfer system for the euro area was necessary, as large current account deficits and surpluses were mostly seen as benign, and countries were expected to be able to borrow in financial markets should a macroeconomic shock occur.4
The idea of an EUBS received a new boost when the euro area stumbled from the global financial and economic crisis into the euro crisis, with the European Commissioner for Employment at the time, László Andor, pushing the debate forward. Since then, a large number of academic and policy-related contributions on the EUBS have been published, the most recent and comprehensive one prepared for the European Commission by a consortium led by CEPS.5
This contribution attempts to extract the most important ten lessons from this body of literature. To this end, it will point out five findings which most of the contributions on an EUBS agree upon, three commonly perceived myths about the necessity and the potential impact of an EUBS and, lastly, two remaining dilemmas which stem from the interaction of the findings and the current political environment in which euro area governance reform is debated.
Among the robust findings of the research on an EUBS is first that safeguards against macroeconomic shocks in the euro area did not work as well as had been hoped for. Modern textbooks on international capital flows propose two channels through which financial markets can help countries deal with shocks: First, with open capital markets, households can diversify their portfolios by investing abroad. If an asymmetric shock then hits their home economy, income flows from abroad can substitute for the drop in domestic income. Second, if a temporary shock hits, households can borrow from abroad to sustain consumption despite a drop in domestic income and production.
Unfortunately, however plausible this story appears in the textbooks, these mechanisms did not play out in the euro area in the years after 2008. First, households had kept most of their wealth in domestic assets, including – but not limited to – housing, so that when the severe recessions in some countries set in, households had almost no income from abroad to rely upon. Second, while households had borrowed through their banking systems from abroad prior to the crisis in order to increase consumption, banks cut back on lending once the crisis hit. Hence, economies in recession could only draw from international financial markets to a limited extent.
The second important finding is that European fiscal rules in place prior to the crisis were not sufficient to make sure that governments had ample fiscal room to counteract a downturn. One problem was that governments in general had not cut deficits strongly enough in good times. Another reason was that for governments already running surpluses – such as Ireland or Spain – prior to 2007, it proved politically very difficult to sustain large surpluses. Instead, the governments concerned used the ample public funds to increase public sector spending (including public sector wages) or to cut taxes, even though their economies were already booming. This behaviour not only limited the room for manoeuvre in the following crisis, but it arguably made the macroeconomic imbalances in these countries worse by further contributing to an inflating real estate bubble and overshooting wage increases.
This behaviour is linked to the third lesson: even for countries which adhered to the fiscal rules, in some cases the downturn proved too large to be offset by the borrowing of national governments. Beyond the question of shock absorption through financial markets by the private sector discussed above, it is usually assumed that governments with sound public finances can borrow in international markets during recessions to sustain spending even if revenues tank. During the eurocrisis, this hope turned out to be false. Even though countries like Ireland and Spain had been running budget surpluses prior to 2007 and had debt-to-GDP levels of less than 60%, during the crisis they were effectively cut off from financial markets when problems in the banking sector required large bailouts. Hence, the self-insurance of governments through financial markets also did not work as expected.
The fourth lesson learned concerns the indicators to which transfer payments could be linked. While in the early years of the modern discussion of a euro area fiscal capacity, some proposals were brought forward that would have made payments dependent on a country's output gap,6 these ideas are now all but off the table. A specific problem is the growing unease with the reliability of the output gap estimates. The European Commission repeatedly revises and refines its methodology, and the current documentation of the methodology runs to more than 100 pages,7 but the output gap estimates are still vulnerable to strong revisions and large uncertainties. For example, during the years of the global financial crisis (2008-09) and the subsequent eurocrisis (from 2010 onwards), output gap estimates for some European countries were not only revised by several percentage points, but in many cases, negative output gaps for past years were revised into strongly positive values. Hence, under a system of payments linked to the output gap, payments would have been made to countries which (according to current data) were already producing above their potential output.8
The fifth lesson is that a so-called "equivalent" EUBS is easier to implement than "genuine" forms. The literature usually classifies proposals for an EUBS as genuine if individuals have an actual claim against some kind of benefits from a European unemployment system (much like they have today against their national systems).9 Systems are considered equivalent if, instead of providing individual claims, funds are funnelled from a European system to national systems – depending on unemployment levels in the country concerned. "Equivalent" here means basically the same as "reinsurance".10
Complications with genuine systems are of a technical, legal and political nature. On the technical side, the question is how payments from a European unemployment insurance would interact with the national tax, social security and welfare systems, as well as with other active and passive labour market institutions, including activation policies. This problem mainly arises because national systems are still very different, and it is difficult to design a genuine EUBS which does not increase coverage compared to existing national insurance in at least some countries, raising the question of interaction with other national benefit systems.
From the legal side, the problem is that a genuine EUBS might interfere with the rights of EU member states to define social protection in their jurisdiction. While a transfer of certain competences to a European institution might be possible under European law, it would need constitutional amendments in some countries. In contrast, simple transfers among European countries, even if tied to an indicator such as the unemployment rate, can be relatively easily agreed among countries that wish to participate.11
From the political side, the problem is that in some countries, social partners are involved in the administration of the national unemployment insurance. In these cases, unions and employers fear that they might lose their influence if a significant part of the system is Europeanised. As these fears are not relevant for equivalent systems, where money is distributed only between national systems, equivalent systems are much less politically controversial.
In addition to these five conclusions, there are a number of popular myths about an EUBS. I use the term "myth" here if a certain claim is regularly made in the debate, but the support for it in the research is either limited or upon closer inspection rather flimsy.
The first such popular myth is that even in other federal systems, most of the stabilisation in individual states happens through cross-border private capital flows and that government transfers only play a very limited role. The most widely cited contribution in this tradition is by Asdrubali et al., and it is usually quoted as demonstrating that only about 13% of shocks in the US are bolstered by government transfers, while 23% are smoothed by credit markets and 39% by capital markets, with the rest remaining unsmoothed.12
While similar results have been published for other federations using this methodology, the measuring approach is actually highly questionable. First, by employing time-fixed effects in their estimation, the inquiry limits itself to asymmetric shocks. Large transfers in a nationwide recession (such as those happening in the US through extended and emergency unemployment benefits are actually taken out of consideration. While the Asdrubali et al. approach might be justifiable if one wants to examine concurrent inter-state risk-sharing of asymmetric shocks in a federation, it is inappropriate if the question at hand is how much of a single state's recession is actually stabilised by the federal government. In determining whether a federation needs fiscal transfers to stabilise business cycles at the level of member states, the latter question is more relevant, especially if member states are constrained in their ability to borrow.
Second, because of the absence of data on capital flows across state borders within the US, all smoothing of dividend payouts from the corporate sector to the private sector is measured as an "inter-state risk sharing through capital markets", even if both the companies paying out dividends and the households receiving them are located in the same state. Third, a reduction of savings in the wake of a fall in income is taken as "inter-state risk sharing through credit markets", even if a similar reduction could have taken place in a completely closed economy.13
As this methodology systematically understates the role of the federal government in shock absorption and overstates the role of cross-border capital flows, in reality, federal transfers, taxes and spending can be expected to contribute much more to stabilisation at the state level in federations such as the US. Thus, the widely presented argument – that if the US can live without substanial stabilisation by government transfers, then so too should the eurozone – needs to be qualified.
A second widely shared myth is that an EUBS in principle is a (potentially inferior) substitute for greater cross-border mobility of the labour force. According to this narrative, if the unemployed moved to other euro area countries in the wake of a macroeconomic shock, a transfer system would not be needed. Proponents of this argument push for lowering barriers to cross-border migration or even for putting pressure on the unemployed to look for jobs outside their home country.
However, this argument is at least somewhat dubious. Languages and customs still differ greatly among euro area countries, and this fact is not going to change anytime soon. While language barriers today play less of a role at the bottom of the qualification scale (e.g. construction workers) and at the top of the qualification scale (e.g. scientists), they do for many of the qualifications in between. Unemployed jobseekers with very language-specific or culture-specific qualifications cannot actually be expected to find a job in another country in which they can make full use of their human capital. A Spanish bank teller who loses her job in Madrid will most likely not be able to work in a bank after moving to Stuttgart in Germany but might end up in a job there for which much lower qualifications are needed. As long as there is a chance of re-employment in a qualified position in her home country, migration by this person would amount to a destruction of part of her human capital, which is economically inefficient.
While nothing is wrong with making migration easier (which might improve the euro area economy by allocating human capital more efficiently), one should be aware of these downsides when pushing people to migrate who might have good economic reasons to stay put. An EUBS could help a country hit by a strong macroeconomic shock to get through a temporary slump and might help to preserve human capital by preventing unnecessary migration. At the same time, making EUBS benefits portable on a voluntary basis would enable the unemployed to engage in beneficial migration while not forcing workers to engage in excessive migration.
The last myth is that stabilisation through an EUBS would be negligible. Such a conclusion can be easily reached when one looks at the average stabilisation across the whole business cycle. Most simulations here indicate a reduction in the average output gap of maybe 0.1 or 0.2% of GDP. Some authors conclude from this result that introducing an EUBS is not worth the effort. However, it seems questionable whether such an analysis is really the correct metric. Small fluctuations in the output gap or in unemployment over the cycle are a problem that does not matter economically or politically, and national governments are usually well positioned to deal with these fluctuations. In contrast, large economic fluctuations are those in which stabilisation by fiscal policy is most needed – first, because they are more painful for the private sector, and second, because in times of deep recessions, the power of national fiscal institutions to counteract the downturn might be more limited.
Hence, it would be more appropriate to measure "marginal stabilisation", that is the stabilisation potential following a severe economic shock. When measuring the potential impact in this way, it becomes clear that an EUBS actually can have a rather large mitigating effect during a recession, at least in some countries. For example, according to the simulation studies conducted by the CEPS-led consortium, depending on the specific variant of EUBS under consideration, Spanish GDP in 2012 could have been 1.6 percentage points higher had a European unemployment insurance been implemented at that point.14 If we compare this hypothetical level of GDP to the actual downturn in 2010-2012, we can see that the EUBS would have absorbed more than 40% of the actual fall in GDP (see Figure 1).
Figure 1 (back to the text)
Shock absorption capacity of an EUBS in the Spanish recession of 2012
Source: Author's depiction.
Roughly 40% is of course far from a perfect stabilisation, but compared to the 13% usually cited from the Asdrubali et al. literature for the entire federal tax and transfer system in the US, this number is quite high. Moreover, this figure alone might even understate the stabilisation benefits. First, Spain at that point was on the verge of losing access to financial markets because of concerns about debt sustainability. This loss of investor confidence is a non-linear process. A small further increase in the debt-to-GDP ratio at such a point (for example, caused by a further downward correction in the growth outlook) can lead to investors' expectations shifting to a new equilibrium, which can then have a very large negative impact on economic growth. Similarly, if a strong rise in unemployment brings populists close to an election victory, this can also trigger capital outflows and will have a non-linear negative effect on the economic outlook.
Thus, despite the perception to the contrary, upon closer inspection, an EUBS might have significant stabilisation properties at least for some countries during times when this is important.
Despite these positive conclusions on the potential merits of an EUBS, there are a number of unresolved issues. The first concerns the degree to which an equivalent EUBS might actually stabilise in the way the models predict.
Under most of the simulations, it is assumed that the money sent to member states from an equivalent EUBS would be either directly spent on government investment or public sector wages or used for an additional transfer to households. This assumption is necessary to obtain the relatively high multipliers necessary to show a significant degree of stabilisation of the system.
In reality, this assumption might be too optimistic. Most forms of an equivalent EUBS include some kind of trigger mechanism. Funds are only disbursed if the unemployment rate rises above a certain threshold (or, in models which are linked to the employment rate, employment falls below a certain threshold). This means that in the case of an unexpectedly deep downturn or recession, the governments concerned might not know whether they will receive funds while they put together their budgets and coordinate them with Brussels during the European semester (usually around September of the preceding year). For them, it is hence logical (and might actually be required by the budget procedures under the European fiscal rules) to plan a budget without the revenue from the equivalent EUBS. If the decision is made to disburse EUBS funds to national governments, it is not clear whether the governments would manage to actually spend these funds in the year concerned. Instead, it is conceivable that these funds would just be used to lower borrowing requirements. While using the funds to cut public deficits might help with debt sustainability, it would not help much in terms of the stabilisation of aggregate demand.
Note that this problem does not occur in the case of a genuine EUBS. In this case, the unemployed would directly receive funds from the European system as soon as they lost their jobs. These transfers would substitute for lost income and could be expected to be spent directly by the households concerned. This additional spending would be independent of the government's budget process, and the government could just continue its normal budget process and spending plans.
The second issue is that some of the compromise proposals on the table which specifically address some of the concerns of Northern euro members might be so restrictive in terms of when and how many funds are disbursed that the stabilisation impact would be compromised. An example of this can be seen in some of the versions brought forward by the Italian government,15 according to which countries would pay into a European unemployment system but could only withdraw an amount equivalent to two times their original pay-ins and would have to replenish the funds earmarked for their country-specific support later. The problem with such limitations is twofold. First, the stabilisation impact in a deep recession would be limited. Second, such a system of basically nationally earmarked funds does not really improve debt sustainability in the wake of a large shock, as the country in question basically only borrows against its own future contributions (which at low interest rates is almost identical to allowing the country to borrow more under the Stability and Growth Pact).
Given that some more far-reaching proposals have also been discussed by the Italian treasury, it is very likely that these restrictions were introduced as a reaction to German (and other North European) concerns. This is especially regrettable, as some of the often-voiced concerns are broadly unfounded. For example, the vast majority of studies debunk the concern that the introduction of European unemployment insurance would lead to large permanent net transfers from Northern EMU members to the South. Instead, most studies which include a longer period of time find that Germany would have been a net receiver of funds in the early 2000s and that most of the potential variants of an EUBS do not lead to permanent transfers.16
A second, widely voiced concern is that the introduction of an EUBS would lead to moral hazard, keeping countries from making their labour markets more flexible. However, this worry also does not hold up to closer scrutiny. The generally discussed variants of EUBS focus on short-term unemployment. It is well known that short-term unemployment reacts more strongly in a recession the more flexible a labour market is. In contrast, it is usually long-term unemployment (which is not covered by an EUBS) which is linked in the literature to labour market inflexibilities. Hence, if one wants to argue using such political economy considerations of an EUBS influencing labour market reforms, one would have to conclude that an EUBS increases the incentive for making national labour markets more, not less, flexible.
While, of course, one needs to take into consideration the political dynamics of proposals for the eurozone, it would be regrettable if compromises borne out of mostly unfounded political concerns lead to proposals whose impact remains strongly below their potential.
Summarised in Figure 2 are the ten lessons discussed in this article that I have learned from a decade of debating a European unemployment benefit scheme. In conclusion, one can say that the past decade's research on an EUBS has shown that such a system is feasible and would potentially have relevant stabilisation benefits for the euro area. Yet it remains unclear whether the political impediments which originally led similar proposals to be discarded in the euro area architecture have weakened sufficiently to move this project beyond theoretical discussions. Moreover, despite the large progress made in researching an EUBS over the past decade, still some conceptual work needs to be done before a workable, politically acceptable and effective EUBS design that is ready for implementation is on the table.
Figure 2 (back to the text)
Ten lessons from the EUBS debate
Source: Author's depiction.
1 See European Commission: Report of the Study Group "Economic and Monetary Union 1980", 1975, p. 34 ff., available at http://ec.europa.eu/archives/emu_history/documentation/chapter7/19750308en57reportstudygroup.pdf.
2 See A. Majocchi, M. Rey: A special financial support scheme in economic and monetary union: Need and nature, in: European Economy – Reports and Studies, No. 5, 1993, pp. 457-480; J. Pisani-Ferry, A. Italianer, R. Lescure: Stabilization properties of budgetary systems: A simulation analysis, in: European Economy – Reports and Studies, No. 5, 1993, pp. 511-538.
3 See S. Dullien, D. Schwarzer: Bringing Macroeconomics into the EU Budget Debate: Why and How?, in: JCMS: Journal of Common Market Studies, Vol. 47, No. 1, 2009, pp. 153-174; S. Dullien: Eine Arbeitslosenversicherung für die Eurozone, SWP-Studien, 2008/S01, 2008, available at https://www.swp-berlin.org/fileadmin/contents/products/studien/2008_S01_dullien_ks.pdf; R. Deinzer: Konvergenz- und Stabilisierungswirkungen einer europäischen Arbeitslosenversicherung, Berlin 2004, Duncker & Humblot.
4 See for example B. Schmitz, J. von Hagen: Current account imbalances and financial integration in the euro area, in: Journal of International Money and Finance, Vol. 30, No. 8, 2011, pp. 1676-1695, who write: "[...] the observed current account imbalances should be regarded as signs of the proper functioning of the euro area rather than a sign of improper macro-economic adjustment."
5 See M. Beblavý, K. Lenaerts: Feasibility and Added Value of a European Unemployment Benefits Scheme, Brussels 2017, CEPS.
6 See for example H. Enderlein, L. Guttenberg, J. Spiess: Blueprint for a Cyclical Shock Insurance in the Euro Area, Jacques Delors Institute Studies & Reports, 2013.
7 See F. D'Auria: The production function methodology for calculating potential growth rates and output gaps, in: European economy. Economic papers, No. 420, 2010.
8 See S. Dullien: A European unemployment benefit scheme: How to provide for more stability in the euro zone, Gütersloh 2014, Verlag Bertelsmann Stiftung, pp. 93 ff.
9 For prime examples of this approach, see ibid.; or S. Dullien: Eine Arbeitslosenversicherung für die Eurozone, op. cit.
10 A typical example for this approach is M. Beblavý, D. Gros, I. Maselli: Reinsurance of National Unemployment Benefit Schemes, CEPS Working Document No. 401, 2015, available at http://www.ceps.eu/system/files/wd401.pdf.
11 For a summary, see M. Beblavý, K. Lenaerts, op. cit.; or for a detailed discussion see A. Repasi: Legal Options and Limits for the Establishment of a European Unemployment Benefit Scheme, European Commission 2017.
12 See P. Asdrubali, B.E. Sorensen, O. Yosha: Channels of Interstate Risk Sharing: United States 1963-1990, in: The Quarterly Journal of Economics, Vol. 111, No. 4, 1996, pp. 1081-1110. To be fair, it needs to be stated that the authors actually suggest caution and indicate certain caveats regarding their estimation, but this careful wording is usually lost when this contribution is quoted.
13 For details on these arguments, see S. Dullien: Risk sharing by financial markets in federal systems: What do we really measure?, Forum for Macroeconomics and Macroeconomic Policies, FMM-Working Papers No. 2, 2017.
14 M. Beblavý, K. Lenaerts, op. cit.
15 See Ministero dell'Economia e delle Finanze: Simulations on the European Unemployment Benefit Scheme for the 1999-2015 period, 2016, available at http://www.mef.gov.it/inevidenza/documenti/EUBS_simulazioni_EN.pdf; Ministero dell'Economia e delle Finanze: European Unemployment Benefit Scheme, 2016, available at http://www.mef.gov.it/inevidenza/documenti/Unemployment_benefit_scheme_rev_2016.pdf; Ministero dell'Economia e delle Finanze: A European Unemployment Benefit Scheme: Nine Clarifications, 2016, available at http://www.mef.gov.it/inevidenza/documenti/EUBS_chiarimenti_EN.pdf.
16 Note, however, that Dolls finds that four countries would have been permanent net receivers or net beneficiaries, but that the likelihood of such permanent transfers can be easily reduced; see M. Dolls: Chances and Risks of a European Unemployment Benefit Scheme, ZEW policy brief No. 7, 2016.