Social Protection in a Changing World:
LSE-IMF Joint Workshop
Driven by factors such as high inequality, perceived unfairness surrounding the gains from globalization and technological change, and anxiety about the future of work, trust in institutions is waning and social discontent is rising. Adding in population aging and climate change, the present moment surely calls for a re-assessment of existing approaches to social protection to better align policies and instruments with these 21st Century challenges.
As part of its strategic focus, LSE is putting a high priority on research related to the contours of new welfare provision. At the same time, the IMF is re-assessing the nature and extent of its engagement on social spending policies. In Spring 2019, the IMF will present a paper setting out a new strategic framework for engaging on these issues with its member countries to its Executive Board. This is being guided by a growing awareness that social spending is important for inclusive and sustainable economic growth and financial stability, the key focus of the IMF.
Reflecting this convergence of interests, LSE and IMF jointly sponsored a workshop on "Social Protection in a Changing World". The workshop, held at LSE on 2 November 2018, convened leading academics working in the field together with representatives from the IMF and the World Bank. The agenda for the day included discussions on emerging challenges, the role and design of social assistance and social insurance, the balance between universalism and targeting, and financing social protection.
Dame Minouche Shafik, Director of LSE, observed that social safety nets had on the whole performed well in the immediate aftermath of the 2007/08 financial crisis, partially cushioning many people from the effects of the macroeconomic shock, but subsequent constraints on public spending, supervening on already high levels of inequality and combined with continuing low growth and fears about the uneven impact of technological change, had deepened social discontent and led to a rise in political populism in many countries. She proposed that a new social contract was needed, one that reinstated reciprocity and insurance elements of welfare provision, redressed the balance between labour and capital taxation, and instituted effective "predistribution" measures such as wage floors, investments in education and skills, boosting social mobility, and infrastructure developments in disadvantaged regions. Minouche hoped that LSE would continue to be a central player in exploring the contours of this new contract, building on its rich tradition of research on the design and economics of welfare.
Dr Vitor Gaspar, Director of the Fiscal Affairs Department at the IMF, reminded us that growth and stability were at the core of the IMF's mandate, and he observed that these were increasingly recognised as being mutually dependent on effective social protection, and investment in people through health, education and gender equality. High and persistent inequality undermines the sustainability of economic growth, whilst "automatic stabilizers" including social insurance may be more effective than discretionary policies in mitigating the volatility of the business cycle, and also contribute to sustaining political stability. However achieving sustainable inclusive growth is not straightforward. It depends in part on state capacity: the ability to collect taxes, deliver services and allocate public spending in an accountable and efficient way.
The IMF is currently developing a framework to clarify the scope, objectives, and boundaries of its engagement with countries on social spending, including social protection, across its surveillance, program, and capacity development work, with a view to issuing guidance on how to assess the macro-criticality of social spending. Collaborating with partners will be crucial in this work and today's workshop is one step along that road.
Climate change, and climate induced migration, introduce new requirements for policies to be not just economically but also environmentally sustainable, and highlight new needs for protection.
Dr Michal Rutkowski (World Bank), Dr Jeni Klugman (Harvard) and Professors Nick Barr and Ian Gough (both LSE) outlined a range of inter-related social, economic, political and climate challenges for social protection in the 21st century. Many aspects of demographic change are welcome more people living longer, changing gender norms and increased participation of women in the labour market but the design of social protection has not kept pace. Similarly, technological change should increase productivity and free people from drudgery, but if the distribution of skills and work is not addressed, the result will be an increasingly polarised labour market with consequential strains on social protection. Informality and insecurity in the labour market mean social protection cannot be based on the presumption of a stable relationship with an employer.
Climate change, and climate-induced migration, introduce new requirements for policies to be not just economically but also environmentally sustainable, and highlight new needs for protection. Those affected are often already the most vulnerable: climate change is a "threat multiplier".
Social insurance and social assistance Dr Santiago Levy (Inter-American Development Bank), and Professors Armando Barrientos (Manchester), Camille Landais and Robin Burgess (both LSE) debated the functions and forms of social insurance and social assistance. The speakers agreed that although social insurance was traditionally associated with income-smoothing and risk-pooling, and social assistance was considered a primary tool for poverty alleviation, in practice inter-dependencies between insurance and assistance and their goals needed to be understood and taken into account when evaluating policies and designing reforms. This includes understanding the combined incentive effects and behavioural responses, as well as detailed analysis of (potential) beneficiaries including by age, gender and occupational status. Social insurance is sometimes associated with well-organised interest groups, but expansion of social assistance could also be electorally popular.
The rise in informal employment in rich countries, and its continuing prevalence in low and middle income countries, limits the reach of conventional social insurance. Conversely the costs to employers and employees of social insurance are much more visible than its benefits; hence social insurance can itself act to incentivise informality. Santiago argued a new architecture is needed with as large a risk pool as possible, and contributions that are independent of the particular form of employment.
The evidence presented on the positive effects on sustained poverty alleviation of large asset transfers and skill development in rural Bangladesh generated considerable interest. Robin observed that the income transfers in existing social assistance programmes are often very small compared to needs and that the magnitude of the asset in this example was crucial; transfers below a given threshold did not enable recipients to achieve "exit velocity'" from poverty.
Universalism and targeting
McKnight (both LSE) eschewed the binary distinction between universalism and targeting and instead encouraged us to think about different degrees of universalism and different forms of targeting. Moreover the extent to which social protection is pro-poor' depends not only on the transfers themselves but also on how the revenue is raised to pay for them: the progressivity of taxes, contributions and benefits need to be analysed together. Abigail's recent evidence based on analysis within a panel of high-income countries over a period of up to 40 years supports Korpi and Palme's original "paradox of redistribution": when net cash transfers are more concentrated on lower income households (ie, more targeted), reductions in inequality and poverty are lower.
A range of explanations for the paradox have been put forward, amongst them the extent to which political support for transfers may be undermined when an increasingly small proportion of the population are seen to be beneficiaries. Political sustainability was one of three criteria proposed by David for the assessment of social protection, the other two being efficiency in reducing poverty (including poverty gaps), and labour market and demographic behavioural effects.
Evaluating Universal Basic Income against these criteria produces an unfavourable result: with low poverty-reduction efficiency, potentially damaging behavioural effects and low political sustainability in the face of "free rider" problems. By contrast, "tagging", that is, concentrating transfers on groups with clearly identifiable characteristics associated with increased risk of poverty children, elderly people, those who are unemployed or disability scores well on David's criteria, and is superior to conventional income-targeting in terms of political support and behavioural effects.
Many participants commented on the renewed willingness and commitment that became apparent through the day to work across traditional boundaries to tackle the challenges of providing social protection in the 21st century.
Financing social protection
The final session of the day picked up a number of themes from earlier discussions the implications of increased informal employment and limited state capacity, the overlapping functions of social assistance and insurance, the joint significance of financing and transfers in achieving redistribution, and the central importance of political sustainability alongside economic sustainability. Each of the speakers tackled different problems and proposed solutions. Dr David Coady (IMF) put the case for consumption taxes as an important source of revenue for developing countries, and argued that they could be combined with transfers based on proxy means testing (PMT) to achieve "progressive universalism". Although limitations of PMT were acknowledged, these were mitigated if it was used to differentiate levels of transfers rather than to determine eligibility overall.
Dr Barry Herman (New School University) proposed that countries should be able to ring-fence spending on a social protection floor, certified by the ILO and IMF and backed by a compact with creditors, for example through a GDP-linked bond. This would insulate the social protection floor against the claims of creditors if and when the country experienced a debt crisis. It could be justified on economic grounds because of the strong countercyclical benefits of social protection spending, but also on social grounds (embedding the need to meet Sustainable Development Goals) and political grounds (enhancing political stability and confidence in government).
Dr Andrew Fischer (Erasmus University) argued that external donor-funding of cash transfer programmes could be problematic. Whilst donors believe they are contributing aid to support social protection floors, and expect a degree of accountability for the programmes they are funding, recipient governments resist the heavy-handed conditionality associated with some previous interventions by international financial organisations and seek to preserve room for manoeuvre to meet domestic political economy priorities. Depending on the consequent adjustments and spending substitutions, the overall effects on poverty can be positive or negative.
Reflections on the day
An encouraging aspect of the workshop was the atmosphere of open and constructive dialogue. Many participants commented on the renewed willingness and commitment that became apparent through the day to work across traditional boundaries to tackle the challenges of providing social protection in the 21st century. This encompasses disciplinary boundaries within academia for example, between economics, social policy, development studies and political science and the boundaries between academia, NGOs and international organisations such as the IMF and the World Bank. It also encompasses the divisions and tensions that sometimes exist within organisations between departments responsible for policy development and the teams charged with implementation in particular countries and contexts. As we develop initiatives to follow up on the workshop we have the importance of these crossdisciplinary, cross-sectoral and cross-functional links very much in mind.
Another clear message from the day was that lack of clarity over concepts and terminology hampers effective dialogue. One person's universalism is another person's targeting (for example, with reference to taxable child benefits); one person's social insurance is another person's social assistance (for example, where contributions are nominal or payments income-related); and the delineation of "social protection" itself is unclear. Evolving a common vocabulary and set of definitions will be an important to ensure future exchanges can be as productive as possible, based on a critique of the contents of proposals rather than the labels attached to them.
Finally, and related to both the previous observations, there was striking consensus on the desirability on the one hand of articulating a clear vision of what social protection should achieve, and on the other hand, of avoiding dogma when it comes to identifying the mechanisms to deliver those goals. Ideas matter, and can set the framework for, and give the impetus to, the identification of new solutions. But it does not make sense to start from a fixed commitment to a particular type of solution such as targeting or consumption taxes or universalism given the diverse functions of social protection (poverty alleviation, income smoothing and insurance, human capital development, counter-cyclical macroeconomic effects, and promoting social and political stability to name but a few); given the necessity to articulate solutions with what already exists in any given country, for example recognising path dependencies with respect to existing social insurance schemes; and given that the context of state capacity, social provision (especially of health and education), and labour market policies with which social protection interacts varies so widely from place to place.