Volume 52, May/June 2017, Number 3 · pp. 154-159
Risk Reduction, Risk Sharing and Moral Hazard: A Vaccination
Frank Vandenbroucke, University of Amsterdam, The Netherlands.
Nearly all monetary unions are true "insurance unions": they not only centralise risk management with regard to banks, they also centralise unemployment insurance. The European Economic and Monetary Union (EMU) is the one exception, but it is gradually developing policies driven by the need for mutual insurance, notably in its progress towards a Banking Union.1 This paper focuses on unemployment insurance; it sketches the rationale for a degree of centralisation that would create a lean insurance union, addressing the risk of large economic shocks. Both politically and technically, the debate about the merits and drawbacks of an EU role in unemployment insurance is complex, and we should not expect any definitive resolution soon. However, at this juncture of Europe's political history – with a new political landscape unfolding after the Brexit vote and Macron's victory in France – it is important to pursue this debate and to clarify the arguments. Automatic stabilisers are a basic feature of national welfare states. Hence, there is an intrinsic link between our conception of eurozone automatic stabilisers on one hand and our conception of the EU's role in social policy on the other.
Inspired by one particular strand of "optimum currency area" theories, eurozone stabilisers are often discussed against the backdrop of a trade-off between the "symmetry" across the member states and the "internal flexibility" within the member states. This trade-off can be mitigated by fiscal transfers: when asymmetric shocks occur, fiscal transfers reduce the need for internal flexibility.2 However, one can also approach the eurozone's predicament from the perspective of theories on international risk sharing.3 In this contribution, I apply a vaccination metaphor that is compatible with both approaches. The metaphor highlights the fact that one should avoid a simple contrasting of "risk reduction" versus "risk sharing", and it signals social policy priorities with regard to the joint resilience of national welfare states and the eurozone.
Why centralise unemployment insurance?
Monetary unions either opt for a downright centralisation of unemployment insurance (as in Canada or Germany) or they streamline unemployment insurance and provide a degree of reinsurance and centralisation when the need is especially high (as in the US, which combines centralisation and decentralisation in unemployment insurance). This is rational behaviour for two reasons. First, risk pooling enhances resilience against asymmetric, idiosyncratic shocks. This is particularly important in a monetary union: its members are especially vulnerable to asymmetric shocks, since there is a limit to the internal flexibility they can rely on to adjust to such shocks.4 The second reason also applies when shocks are not idiosyncratic but symmetric across the whole union, and risk pooling across member states would have no added value per se. National insurance systems create an externality; a country that properly insures itself also helps its neighbours.
The problem at hand is comparable to vaccination. Vaccines are an archetypal example of externalities: through vaccinations, individuals protect not just themselves from infectious diseases but also the people they come into contact with. Hence, with a view to efficiency, it is rational for governments to subsidise vaccinations and/or make them compulsory. Absent this, there will be underprovision of the vaccine. This paper first elaborates upon the compulsory vaccination analogy and then returns to the subsidisation argument.
Prima facie, the vaccination metaphor not only applies to a monetary union but to any highly integrated set of countries, since trade integration suffices to generate externalities of national booms and busts. However, apart from the fact that monetary integration is supposed to promote trade integration, the risk of underprovision of unemployment insurance (our metaphorical vaccine) may be higher in a monetary union. In the short term, unemployment insurance increases the cost of labour. Since exchange rate fluctuations are excluded, nominal wage cost competitiveness is a salient concern in a monetary union. Hence, without some form of coordination, national authorities may be reluctant to provide "costly" unemployment insurance in a monetary union. (In the US in the 1930s, federal initiatives to streamline and support state unemployment insurance were motivated by this classical problem of collective action.)5 In short, a monetary union typically features a high risk of cross-border "contagion" on one hand, whilst on the other hand the collective action problem with regard to "vaccination" is salient. Therefore, it is rational for the members of a monetary union to agree on mandatory vaccination, i.e. a set of minimum requirements with regard to the stabilisation capacity built into their national social and economic systems.
Which minimum requirements should apply? From a preventative perspective, fiscal prudence is a first requirement: member states must not accumulate structural deficits, because that reduces their ability to increase public deficits and incur additional debt during a downturn. However, from a stabilisation point of view, fiscal prudence is only a precondition; welfare states must have an endogenous automatic stabilisation capacity to smooth economic shocks. Automatic stabilisation is associated with the size of the public sector and the level of social spending – the public sector wage bill and social transfers create an inertia effect that reduces short-term volatility. Conceived in this manner, a significant level of pension spending and a large public education sector act as automatic stabilisers.
However, in the domain of public spending, the most effective instrument for stabilisation in case of an unemployment shock is unemployment insurance. It kicks in when people lose their income because of unemployment, which makes it crucially different from pension spending (or from a universal basic income,6 for that matter). Automatic stabilisation is also associated with the average effective tax rate (notably in case of an "income shock"), which is in turn associated with the size of the public sector and social spending. However, the difference in the stabilisation capacity across EU countries is crucially linked with the structure of the benefit system, and notably with the performance of unemployment insurance.7 This provides a first reason why one should focus on unemployment insurance if the internal stabilisation capacity of member states is a matter of common concern. The second reason for focusing on unemployment insurance is that the level and progressivity of taxation is a sensitive national issue, whilst the need for effective unemployment insurance is widely shared across welfare states of different types.
The stabilisation capacity of unemployment benefits depends on their generosity (notably for the short-term unemployed) and their coverage. Hence, a "compulsory vaccination" programme against instability would include minimum requirements with regard to the effective coverage and the generosity of (short-term) unemployment benefits in the participating member states. Do they cover all employees or do large groups remain uninsured, as was traditionally the case in Italy (which explains why the stabilising role of unemployment benefits is so limited in that country)? Are they generous enough to have a stabilising impact without creating inactivity traps?
For national welfare states, unemployment benefits are the metaphoric camel's nose: whether they generate resilience, i.e. whether they function as an effective shock absorber without negative side effects, also depends on general features of the labour markets and the quality of the activation policies. Labour market segmentation that leaves part of the labour force poorly insured against unemployment, or the proliferation of employment relations which are not integrated into systems of social insurance, reduce the stabilisation capacity of welfare states. Poor activation leads to hysteresis rather than resilience. In other words, a "vaccination programme" for resilient national welfare states entails a cluster of principles for labour market policies.8 Obviously, unemployment insurance acts, partly, in a curative way: a medicine is applied when the patient is ill. However, I want to emphasise that the institutions to make this medicine work must be in place well before a crisis arrives and must go beyond the mere establishment of a benefit system per se.
It is important to note that the preceding argument – drawing upon the metaphor of compulsory vaccination – is not premised on the idea that the eurozone would itself be equipped with automatic stabilisers. However, it is no coincidence that vaccination is being subsidised both in countries where it is mandatory and in countries where it is not. Economic theory teaches that goods and services with positive externalities should be subsidised to reach an optimal level of consumption. In the eurozone, it would be rational to associate reinsurance of national schemes (granting a European subsidy to national systems when the need is high) with minimum requirements for the stabilisation and activation qualities of these national schemes. In other words, reinsurance (which creates a temporary subsidy to keep the national "vaccine" against economic volatility affordable) and a compulsory vaccination programme would go hand in hand. Such an insurance device would create a fiscal union of a special kind, which is politically easier to obtain than a full-fledged budgetary union. The EU will not develop a federal budget like in the US or Canada, but a relatively small insurance premium could have the same stabilising impact.
From theoretical principles to practical policies
Admittedly, the vaccination metaphor needs empirical validation. We have to assess the extent to which the fragility of the eurozone is due to problems in the financial sphere, to which banking union and related proposals are the first answer. We have to consider the extent to which shocks are asymmetric or synchronised. We have to examine how significant the shock-absorbing role of unemployment insurance at the national level really is, and next, how significant its positive, cross-border externalities are. Space constraints prevent me from thoroughly digging into the empirical literature on these questions here. However, it seems fair to start from the following presumptions.
First, shocks in the eurozone are relatively well synchronised, except for their amplitude;9 this means that we are confronted with a peculiar mixture of asymmetry (in terms of amplitude) and synchronisation, pointing to the need for both cross-country risk pooling and intertemporal smoothing. The vaccination metaphor is compatible with this observation: resilient national welfare states are a matter of common concern, and their resilience may need support, partly (but not only) because of asymmetries.
Second, the externalities associated with adequate unemployment insurance and a resilient national welfare state are not such that we should set up a complex cooperative scheme at the eurozone level to deal with normal, relatively mild swings in the business cycle. These externalities only become relevant as a matter of common concern and justify collective action in the case of large shocks.10 In other words, a cooperative fiscal arrangement should address the risk of large shocks. The crux is that its interventions should not be based on political discretion, but on pre-established, self-executing mechanisms. The automatic nature of stabilisers is key.
The rationale for self-executing mechanisms is simple: prevention is better than cure. Although the eurozone has developed a degree of solidarity since the crisis, it only came about after difficult intergovernmental negotiations. Solidarity was not ex ante rooted in the European fabric; it occurred ex post. This has two drawbacks. Organising solidarity ex post in an intergovernmental setting implies repeated ad hoc negotiations about burden sharing and conditionality, which easily leads to polarisation among governments and their electorates. Ex post solidarity is also more expensive than ex ante solidarity if the latter has a preventative impact. This certainly applies to economic instability: economic swings are driven by expectations. The expectation that a shock absorber will do its job when an economy is threatened by a significant shock contributes to the prevention of severe shocks. With a view to resilience, risk reduction and risk sharing reinforce each other.
This preventative dimension also explains why "private insurance mechanisms" through international financial markets need complementary "public insurance mechanisms" through fiscal transfers. International markets will be less prone to panics and sudden stops if public insurance mechanisms are expected to cushion serious shocks. This is exactly the message to be found in much of the IMF work on the eurozone, which views the banking union, capital markets union and fiscal stabilisers as complementary devices. The "private" insurance provided for national economies by well-functioning international credit and capital markets needs the support of a "public" inter-state insurance mechanism based on a fiscal capacity.11 In the United States, so goes the IMF analysis, it is the complementarity of private insurance against idiosyncratic shocks hitting the individual American states and public insurance provided by the federal tax and benefit system that does the job. Although a banking union may in itself be most urgent, the capital markets union and automatic fiscal stabilisers are complementary solutions that together add up to more than the sum of their parts. Hence, even if the design of automatic stabilisers entails a complex discussion, it should not be postponed. One day, the concurrent existence of a banking union, a capital markets union and automatic stabilisers may constitute a formidable institutional asset for EMU – and thus for the European project at large.
How should automatic fiscal stabilisers be organised? Several proposals have been published, some of them linking stabilisers to national unemployment schemes.12 These proposals typically imply that member states contribute to a common fund that disburses money to member states affected by negative shocks, e.g. a significant increase in unemployment. A research consortium led by the Centre for European Policy Studies (CEPS) examined different variants of a European Unemployment Benefit Scheme.13 The complexity of setting up a genuine European Unemployment Benefit Scheme (GEUBS), even if it only complements existing national schemes, should not be underestimated. Moreover, any European scheme should preclude permanent transfers to member states and avoid a structural redistribution of resources between countries. It should instantiate a pure insurance logic, covering risks that affect all countries participating in the scheme to the same extent. Simultaneously, since there is a degree of synchronisation of the business cycles across the eurozone, such a scheme must be able to organise intertemporal smoothing. That is, it must be able to incur deficits and issue debt, in one way or other. My conclusion from this research is that it is easier to meet these conditions and to implement such a scheme if it takes the form of reinsurance of national insurance schemes rather than as a GEUBS.
The difference between a GEUBS and reinsurance is twofold. First, in a GEUBS, individuals who are short-term unemployed receive an individual benefit from a European fund, whilst reinsurance operates with lump sum budgetary transfers between a European fund and member states. Second, in the reinsurance model, member states receive transfers on the basis of a trigger (for instance, based on the deviation of current short-term unemployment in the member state from its past trajectory in that same member state). In a GEUBS, there is no specific trigger for the scheme to start disbursing money (any short-term unemployed individual in a participating member state receives a benefit, independent of the level or growth rate of short-term unemployment in that member state). It is not happenstance that there is no trigger in a GEUBS, but this needs some explanation.
Imagine a European fund paying basic unemployment benefits to unemployed citizens, which can be topped up by national governments; these European transfers constitute individual rights that cannot be made conditional on a trigger, such as the fact that short-term unemployment exceeds a certain level or that its growth rate is higher than a certain percentage.14 In contrast, reinsurance does not interfere with the individual benefits that citizens receive, but simply cashes out budgetary transfers to member states in support of national systems. A trigger is a natural feature in such an approach. It allows more flexibility in the design, for instance with a view to address large shocks rather than any cyclical movement. A trigger also creates additional possibilities to fight institutional moral hazard, as explained in the next section. Reinsurance also seems the less complicated option.
Solidarity and moral hazard
Solidarity is always intrusive. If the aim of European solidarity is to contribute to stabilisation, a logical corollary is that the stabilisation capacity of the national socio-economic systems must be sufficient: maintaining (and, in some countries, reinforcing) the stabilisation capacity of national systems becomes the self-evident political quid pro quo for organising European support. Moreover, the possibility that member states become lax with regard to the activation of the unemployed – and (re)employment policies in general – because they will benefit from EU funding generates an obvious risk of institutional moral hazard.
However, we should not become obsessed with moral hazard, which is unavoidable in any insurance context. If we want to eliminate the faintest possibility of moral hazard, we will never be able to organise insurance and reap the benefits of collective action. On the other hand, we should not dismissive about moral hazard: we should address it and find solutions to minimise it.
The risk of moral hazard can be reduced through financial mechanisms. For instance, in the case of reinsurance, the trigger can be based on relative deviations of the national level of short-term unemployment from its historical trajectory; this would ensure that a high structural level of short-term unemployment would not be supported by the EU. The trigger could guarantee that the fund only intervened in cases of significant deviations from a country's historic profile. "Experience rating" could be introduced to minimise the risk that a member state becomes a permanent beneficiary of such a scheme. The downside is that the more stringent these regulations, the weaker the insurance mechanism, but they can be essential for political support. Moral hazard can also be reduced by establishing minimum requirements on the quality of the member states' activation and employment policies. If these minimum requirements are effective, more room is created for a powerful insurance mechanism. For solidarity to be effective, it needs to be somewhat intrusive.
Europe is a union of welfare states with no intention to become a federal welfare state. However, in this endeavour, we are considering a well-known problem of federal welfare states, where unemployment benefits and employment policy are managed at different levels. There is an institutional risk of moral hazard when a central government is responsible for unemployment benefits while the states, provinces, regions or municipalities are responsible for activation. In this respect, it is interesting to look into countries such as the US, Canada, Germany, Belgium, Austria or Switzerland. In all these countries, institutional moral hazard, whether implicit or explicit, is an issue of politics and policy. There are a wide range of solutions, including minimum requirements on the quality of policies, more or less complex financing models and direct control through coordination mechanisms, among others.15
Ever since the European Employment Strategy was launched in 1997, "coordination" has been part and parcel of the Union. The 2014 Youth Guarantee, which is closely connected with the Employment Strategy, can be seen as a European quality assurance system regarding activation. These mechanisms are too "soft" to underpin a European reinsurance of national unemployment schemes, but the possibility of a European reinsurance could also be the trigger to make them more ambitious and to give them more bite. If its logic is based on insurance, this would create a legitimate quid pro quo. Binding commitments can leave leeway for differentiation in the concrete policies; the essential point is that commitments are complied with, not that they are elaborated in detail. There is no need for a homogeneous European social model, but there is a need for an agreement on some key functions that it has to serve.
Hence, there is an intrinsic link between the debate on the European Pillar of Social Rights and the debate on eurozone stabilisers. It is imperative to sort out the priority fields in which convergence is necessary. A priority field for convergence is unemployment insurance: this concerns principles with regard to the generosity and coverage of unemployment benefits, as well as principles of activation and broader principles of labour market regulation that allow universal access to unemployment insurance for all workers. This priority follows from the preceding analysis, but it is also reinforced by the observation that trends since 2009 point to reductions and a dispersion in the quality of unemployment insurance systems across the EU with regard to benefit generosity, benefit coverage and active labour market policies.16
The vaccination metaphor signals two policy conclusions. The first conclusion aligns with the argument that vaccination should be compulsory: resilience in the face of economic shocks must be a feature of all eurozone member states. Resilience requires sufficiently generous and universal unemployment insurance and adequate activation policies for the unemployed. In itself, this first conclusion is significant; it underscores the importance of a number of principles set out in the European Commission's proposal for a Pillar of Social Rights, independently of the creation of eurozone stabilisers. The second conclusion aligns with the argument that vaccination must be subsidised to reach its optimal level: in the case of large shocks, fiscal support for member states in dire straits is necessary to keep the vaccination programme affordable. If such a mechanism is set up ex ante, risk reduction and risk sharing reinforce each other. Reinsurance, rather than a GEUBS, promises the best approach to attain that dual goal.
This paper elaborates on a section of a forthcoming ECFIN Discussion Paper titled Structural convergence versus systems completion: limits to the diversity in the European Economic and Monetary Union; the research was funded by the ECFIN Fellowship Initiative 2016-2017.
1 It can be argued that EMU already contains insurance functions, notably the TARGET2 system; see W. Schelkle: The political economy of monetary solidarity: understanding the euro experiment, Oxford 2017, Oxford University Press.
2 For a discussion and qualification of this argument, see P. De Grauwe, Y. Ji: Flexibility and Stability. A difficult trade-off in the Eurozone, CEPS Working Document No. 422, April 2016.
3 C. Alcidi, P. D'Imperio, G. Thirion: Risk-sharing and Consumption-smoothing Patterns in the US and the Euro Area: A comprehensive comparison, CEPS Working Document No. 2017/04, May 2017; and G. Thirion: European Fiscal Union: Economic rationale and design challenges, CEPS Working Document No. 2017/01, January 2017.
4 This argument is based on the optimum currency area theory referred to earlier and can be challenged on the following ground: monetary union facilitates the cross-border integration of capital markets. An integrated capital market allows risk pooling across countries – a key observation in theories of international risk sharing; see C. Alcidi et al., op. cit. for further references. Hence, capital market integration reduces the need for fiscal risk sharing. The counterargument, in turn, is twofold: first, capital market integration, whilst important as a risk-sharing mechanism in the US, is limited in the eurozone and bound to emerge only gradually; second, risk sharing via capital market integration and fiscal risk sharing are complementary mechanisms, as I will discuss below.
5 S. Simonetta: What the EU Can Learn from the American Experience with Unemployment Insurance, in: Intereconomics – Review of European Economic Policy, Vol. 52, No. 3, 2017, pp. 142-148.
6 A pan-European basic income is the most promising answer to the need for pan-European transfers, according to P. Van Parijs, Y. Vanderborght: Basic Income. A Radical Proposal for a Free Society and a Sane Economy, Cambridge MA 2017, Harvard University Press, pp. 230-241. They concede that its stabilising properties might be less than an EU-funded unemployment insurance scheme, but consider it a superior solution for other reasons. I find this argument unconvincing, but that discussion is beyond the scope of this paper.
7 For the association between stabilisation and the average effective tax rate, see Figure 3 and Figure 4 in M. Dolls, C. Fuest, A. Peichl: Automatic stabilization and discretionary fiscal policy in the financial crisis, in: IZA Journal of Labor Policy, Vol. 1, No. 4, 2012. They find that the association with the effective tax rate is stronger in the case of an income shock than in the case of an unemployment shock. For the importance of the structure of the benefit system, in which unemployment benefits are key, see M. Dolls, C. Fuest, A. Peichl: Automatic Stabilizers and Economic Crisis: US vs. Europe, in: Journal of Public Economics, Vol. 96, No. 3-4, 2012, pp. 279-294.
8 This fits into a broader argument on the role of social investment in enhancing the resilience of welfare states; see A. Hemerijck: Making Social Investment Happen for the Eurozone, in: Intereconomics – Review of European Economic Policy, Vol. 51, No. 6, 2016, pp. 341-347.
9 P. De Grauwe, Y. Ji, op. cit.
10 For arguments in support of this, see G. Thirion, op. cit., p. 6, and the references cited therein.
11 See C. Allard, J. Bluedorn, F. Bornhorst, D. Furceri: Lessons from the crisis: minimal elements for a fiscal union in the euro area, in: C. Cottarelli, M. Guerguil (eds.): Designing a European Fiscal Union. Lessons from the experience of fiscal federations, Abingdon 2015, Routledge; and D. Furceri, A. Zdziencicka: The Euro Area Crisis: Need for a Supranational Fiscal Risk Sharing Mechanism?, IMF Working Paper No. 13/198, 2013.
12 For a discussion of the rationale of a European insurance scheme and a comprehensive bibliography, see M. Beblavý, G. Marconi, I. Maselli: A European Unemployment Benefit Scheme: The rationale and the challenges ahead, CEPS Special Report No. 119, 2015; and H. Oksanen: Smoothing Asymmetric Shocks vs. Redistribution in the Euro Area: A Simple Proposal for Dealing with Mistrust in the Euro Area, CESifo Working Paper No. 5817, March 2016.
13 For the results of this research, see M. Beblavý, K. Lenaerts: Feasibility and Added Value of a European Unemployment Benefit Scheme, European Commission, 2017, available at http://ec.europa.eu/social/main.jsp?catId=738&langId=en&pubId=7959.
14 Two nuances are important in this respect, but they do not change the conclusion. First, it is in principle possible to conceive of a scheme whereby citizens are entitled to individual benefits that are funded by the EU when a trigger exceeds a certain threshold but funded by the member states when the threshold is not exceeded. This would be quite different from the typical GEUBS proposals, since it presupposes that the EU imposes a specific and detailed uniform basic benefit system on member states (which they can only top up), without necessarily funding it all the time. However, it would also reduce the symbolic "European" nature of those benefits. The second nuance is that in existing multi-tiered systems, individual transfers and triggers are combined. In the United States, for example, the federal level supplements the unemployment insurance systems of the states by paying individual unemployment benefits to American citizens on the basis of a trigger (and occasionally on the basis of discretionary political decisions). However, these individual benefits constitute a federal top-up of basic state systems rather than a federal basic scheme topped up by the states. In such a scheme, the combination of an individual-level intervention and a trigger becomes a feasible option. In contrast, the GEUBS varieties most often presented are "basic" schemes, topped up by the member states. A GEUBS based on a "top-up" by the EU would create intractable problems.
15 For a comparison of eight countries in which different levels of government are accountable for unemployment benefits and the activation of the unemployed, see F. Vandenbroucke, C. Luigjes: Institutional moral hazard in the multi-tiered regulation of unemployment and social assistance benefits and activation. A summary of eight country case studies, CEPS Special Report No. 137, Brussels 2016, Centre for European Policy Studies.
16 European Commission: Employment and Social Developments in Europe. Annual Review 2016, Luxembourg 2016, Publications Office of the European Union, p. 63, pp. 68-69 and Table 1.1.