Skip to main content

Social spending: Managing a $5 trillion challenge

Spending on social priorities, such as pensions, unemployment support, and disability, account for a large portion of government expenditure, yet many services fail to meet customer expectations.

More than a third of government expenditure in Organisation for Economic Co-operation and Development (OECD) countries is assigned to social priorities, including pensions, income and unemployment support, the family, and disability support. The 36 countries collectively spend approximately $10 trillion annually on these areas—more than double the amount allotted to defense and education combined. However, many services still fail to meet customer expectations, and governments need to deliver for less if they are to continue providing the services that people need.

Social spending is already the fastest growing area of government expenditure, expanding, on average, from 18 percent of GDP in 2005 to 20 percent in 2017. Further, it is likely to accelerate in the coming years amid increasing requirements for social services.

Demand will be driven by factors like rising longevity and the disruptive impact of automation in the workplace. The proportion of people aged older than 65, for example, is expected to double by around 2050. Global dependency ratios are set to rise as fewer taxpayers support more beneficiaries.

Incomes, meanwhile, are falling. Two-thirds of households in advanced economies saw their real incomes stall or fall between 2005 and 2014. One reason for this continuing trend is automation in the workplace, which is putting downward pressure on wages. We estimate around 15 percent of workers globally (and up to 50 percent in some advanced economies) will likely switch jobs by 2030. Jobs will also be created, but the cost of a more fluid workplace will be high: workers will need to retrain multiple times in their working lives. As a result, there will more demand for unemployment benefits and transition assistance.

Governments must find ways to overcome the increased pressures of matching their social-spending bills to the outcomes they wish to address. We believe they can do so by first forcing greater clarity into the policy and demand-management outcomes they are seeking—and the trade-offs required to meet them. These outcomes must cascade into the strategy and the organization. Governments can then decide where to innovate across their operations and support functions—for example, through modernizing their use of data and analytics.

A strategic approach to social spending

Governments around the world must think strategically about the challenges presented by a rising demand for social services. Individual countries face particular challenges based on their own tax, demographic, and economic situations. However, we see four dimensions that are common across countries around which policy makers and social-spending-delivery organizations can make a difference (and in some cases, are already doing so). These are policy and demand management, strategy and organization, operations, and support functions (Exhibit 1).

We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com

Clear policy and demand management

Government social spending typically covers four areas: income and unemployment support, family payments, disability-care benefits, and pensions. There is always tension between the outcomes and funding, and this challenge is exacerbated in a tough fiscal environment.

Governments that excel at managing social spending have a clear view of the outcomes they are seeking. They are able to align policy settings to these outcomes effectively. Finally, they accept that there are trade-offs between outcomes and funding and offer transparent guidance on how to approach those trade-offs.

One example of effective prioritization of outcomes that drove innovation in the area of disability care was the introduction in Australia of a National Disability Insurance Scheme (NDIS). The government set an objective around increasing social and economic participation, choice, and control among people with disabilities through the introduction of an insurance scheme for people with disabilities. The scheme upended the status quo, reframing disability as an economic rather than a welfare issue, and led to the abandonment of a one-size-fits-all approach. 1 This objective was prioritized over budgetary considerations: the NDIS involved an almost tripling in funding from around $8 billion to around $22 billion per year. 2 By investing up front, the government was able to win support from challenging stakeholders for transformational change, putting funding in the hands of people with disabilities and creating a market for disability services. In addition to delivering the outcomes sought for people with disabilities, the program is expected deliver significant long-term productivity growth and savings to other government services, more than justifying the up-front investment. 3

Effective strategy and organization

Social-spending agencies refresh their strategies and organizations to deliver on the policy and demand-management outcomes set by governments. This works best where there is a quantitative cascade from policy outcomes to the performance of individual services, a feedback loop into policy setting to show its feasibility or cost, and an organization that is set up to deliver.

A quantitative cascade from policy outcomes to the performance of individual services can be a difficult ask for some policies. Complications include policy areas that are politically charged, such as in child poverty. In those areas, there may be nervousness around producing transparent metrics that political opponents could seize upon to attack progress. Further complications include difficulties in cascading a policy that may cut across services or a legacy organizational structure. Nevertheless, a clear quantitative cascade allows the best targeting of policies and of activities to achieve progress toward outcomes.

The delivery organization should feed back into policy setting. For example, a UK financial-delivery organization tasked with implementing an annual set of policy changes developed a model to show how policy and demand-management outcomes flowed through operational activity. It grouped policies into 30 different features, each with their own characteristics, showing ease of operational and technical implementation and the ongoing running cost. It showed instances in which outcomes could be achieved better through simplified policy. This model allowed better-informed decisions at the point of shaping policy.

Delivery requires an organization to be structured and skilled so that it can cope with changes to policy outcomes—often while implementing head-count reduction. Using agile is a proven way to organize people rapidly to deliver against outcomes amid instability. The playbook to become agile at scale is well established in the private sector—for example, at ING Bank and several other banks. Many public-sector organizations have experimented with agile or moved at scale toward agile. Organizations can choose the pace at which they transform to agile. We typically recommend piloting one service to an excellent standard over six to nine months to create the blueprint for the organization to follow.

Automation can alleviate budget pressures. Our work with one social-spending agency suggests around 40 percent of activities can be automated. The potential upside is a significant productivity boost and improved scalability. This would allow agencies to cope better with spikes in demand for services arising from future economic shocks. Still, several agencies have tried automation pilots and been dissatisfied with the results. Automation requires senior-leadership support to make decisions on difficult head-count reductions. It also requires implementation at sufficient scale to overcome challenges associated with instances in which automating a part of many individual’s daily activities does not sum up to automating the activity of any individual.

One national social-spending agency found that automating 15 tasks would capture 80 percent of the productivity opportunity. The tasks were primarily based on physical, manual, and basic cognitive skills but also included more complex tasks, such as project management, continuous learning, statistical analysis, and negotiation. Identifying this opportunity early allowed the agency to capture productivity gains and, at the same time, reskill employees—as well as reduce full-time equivalents through attrition and prioritize non-automatable skills in new hires.

Social-spending agencies must consider the workforce skills and talent they will need in the coming years. This implies both reskilling (avoiding expensive redundancies) and tailoring hiring to predicted requirements. The United Kingdom’s Department for Work and Pensions is an example in which digital-group leaders are active in social media and tech communities to raise awareness of the work they do and encourage applicants. They acknowledge that while many people will not want traditional full civil-service careers, they may be motivated by the opportunity to spend part of their careers focused on creating societal impacts that the private sector cannot rival.

Our experience suggests that it takes 12 to 18 months to get workforce change, from strategic planning to reskilling-program design and delivery, right.

Streamlined operations

High-performing social-spending agencies are constantly seeking new and better ways to deliver at an operational level, from managing the supply chain to providing high-quality customer service, maintaining compliance and payment integrity, performing asset-and-liability management, and managing citizen registration and eligibility. Here, we dive deep into two important areas: compliance and payment integrity and asset-and-liability management.

Compliance and payment integrity

Better compliance and payment-integrity functions can enable governments to reduce debt, increase payment accuracy, and boost customer and employee experiences. In most countries, 2 to 4 percent of social spending is lost on fraud, waste, and abuse. This represents a multibillion-dollar opportunity.

One social-spending agency defined its vision for 2025 as “the right payment delivered to the right person at the right time, without their active involvement.” It mapped out 45 capabilities required to deliver against the vision and scored itself against each capability. After identifying its strengths and weaknesses, it developed a seven-year road map for moving forward. Other social-spending agencies have captured more immediate opportunities through identification of two or three priority analytics-use cases to reduce fraud, waste, and abuse. Many have leveraged test-and-learn methodologies to drive solutions.

Asset-and-liability management

Governments can capture significant benefits through better management of their assets and liabilities. Pension is a critical area to get right. Governments increasingly face challenges in sustaining their pension schemes because of demographic shifts, strains on public finances, and a challenging investment environment. Two structural approaches—increasing contributions to pension funds and cutting benefits—are politically contentious, although many governments are making bold moves in these areas. A third option, improving investment performance, may be easier to accomplish and should be accompanied by structural reform.

Top performers in asset management achieve 60 percent higher risk-adjusted returns than bottom performers do, at up to a third of the cost. This implies government-run pension funds need to think about how they can achieve more for citizens and identify ways to drive better returns, which will ultimately reduce demand for government support. While clearly not every fund can be a top performer, governments can start by identifying areas of underperformance in pensions.

As governments look for ways to get individuals to save more for retirement, the ability to sustain useful returns will be increasingly important. In Australia, 9.5 percent employer contributions to superannuation schemes are compulsory. Fund assets total more than 2.8 trillion Australian dollars ($1.97 trillion). Australia is consistently ranked in the leading countries for best practices in retirement policy. Yet in January 2019, the country’s Productivity Commission identified several asset-management challenges, including widespread underperformance; a long tail of small, underperforming, high-fee funds; excessive insurance costs; and poor governance, including trustee misconduct. These all lead to balance erosion—lower savings, compounded, meaning more Australians will end up on the state-funded pensions in retirement.

The key message from the Australian experience is that governments around the world must do more to analyze and diagnose challenges in their pension funds. Inaction could have serious implications for private-pension balances, and calls on public pensions, in the future.

Another approach being adopted by many countries is to shift pension structures away from defined-benefit models, which promise a specific income, to defined-contribution schemes, in which the pension depends on the amount paid in. In addition, it has been common to encourage investment in private-pension schemes (Exhibit 2).

We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com

Support functions

Effective delivery requires efficient support functions. Capabilities in IT, data and analytics, delivery support, behavioral economics, and back-office functions are equally as important as frontline knowledge and policy. Two areas of activity making waves right now in evolving the management of social-spending organizations are data and analytics and use of nudges.

Data and analytics applied to delivery and policy

Among the most important recent developments is the ability to leverage data and analytics to create insights that can pinpoint improvements to service delivery and shape future policy decisions.

Applied to service delivery, data and analytics can support a more personalized service. Segmenting citizens early in the process means that those who have a clear eligibility decision can undertake a shorter citizen journey. This allows the bulk of operational staff to focus their efforts where the eligibility decision is either marginal or complicated. Positive impacts include improved citizen satisfaction, lower fraud and error rates, and fewer “extreme bad outcomes,” which are when the agency has clearly made the wrong decisions in complex cases (Exhibit 3).

We strive to provide individuals with disabilities equal access to our website. If you would like information about this content we will be happy to work with you. Please email us at: McKinsey_Website_Accessibility@mckinsey.com

Further, the next wave of social-spending-policy design will be underpinned by data and analytics that can better predict policy effectiveness and cost and then measure their real-world impact. This, in turn, will provide the raw material for future improvements. Analytics can be particularly effective when citizen data are shared across government departments. Citizens have multiple interactions with a government, so joining these can create additional insights. In one European example of sharing data across departments, diverse agencies were able to offer joined services when an individual was released from prison, combining housing and welfare to encourage reintegration into society and reduce reoffense.

The techniques to share data and apply analytics have matured substantially in the past two to three years, and there are detailed blueprints agencies can work from. We typically recommend launching a data-and-analytics transformation by proving its value in one service in parallel with building the operating model to scale and sustaining that value. Value can be proven in as little as four weeks by establishing a “data studio” in which operations, policy, and data specialists are brought together to uncover actionable insights for a particular outcome. Proving the value in this way provides the detailed blueprint for scaling to the wider organization and creates pull at the senior-leadership level for data and analytics services.

Nudges: Proven tools with many use cases

Some agencies are employing behavioral-economics techniques to influence citizen behavior—in particular, to reduce fraud and reduce demand for services. Practices have matured in recent years, leading to a library of use cases and proven methods of setting up small and inexpensive “nudge units” to provide inputs to delivery.

Nudges have been proven in fiscal settings, and social-spending agencies are picking up on that trend to discourage benefit fraud. Changing communications to emphasize social norms can encourage timely payments. In the United Kingdom, a government pilot to improve late tax payments showed that changing letters to explain that most people in their local area had already paid their taxes boosted repayment rates by around 15 percentage points, worth £160 million nationally. Private-sector examples have shown that simple options, such as reversing the order of forms, can have an impact. For example, signing at the beginning, rather than the end, of a form can lead to a 42-percentage-point reduction in cheating.


Citizens increasingly demand that governments do more with less. In addition, the sheer size of social-spending budgets demands that governments prioritize effective and efficient policy and service delivery.

To prepare, a reasonable course of action is first to test that the agency has sufficient focus and clarity on its target outcomes and then test how its priorities fare across the 18 levers we have identified. Agencies can compare their aspirations with global best practices and identify the changes and capabilities that will have highest incremental impact for their outcomes.

About the author(s)

Melanie Brown is a consultant in McKinsey’s Sydney office, Damien Bruce is a senior partner in the Melbourne office, and Mike McCarthy is an associate partner in the London office.

Related Articles