Technology, it’s often said, enables governments to do “good things”: advance their efforts to achieve the Sustainable Development Goals (SDGs), for example. And, it’s as often said, to do “bad things” – such as political surveillance. What’s really being talked of here’s the relationship between technology and policy.
Technology certainly enables governments to do things more efficiently and more effectively than would otherwise have been the case, for good or bad. Their policies, their plans, their goals in general – and who influences those – will determine what they do with it, and how.
This week, I’ll take a look at that through the prism of social welfare or protection: benefits, entitlements, efforts to redress social inequality and reduce poverty. That’s not so widely talked of as public policy arenas like education or employment, but what happens there is representative of many.
Note that I am only talking here about technology and the policy frameworks that guide governments. Technology also changes the underlying contexts - the economic and social structures and behaviours - to which those frameworks apply. That should affect the policy choices that are made but it doesn't undermine policy's central importance.
My context here is social welfare. In rights terms, that includes Article 9 of the International Covenant on Economic, Social and Cultural Rights, which says governments should ‘recognise the right of everyone to social security, including social insurance,’ and the first part of Article 11, which calls on them to ‘recognise the right of everyone to an adequate standard of living.’
What this means, obviously, differs greatly, country to country, because of different policies and national incomes.
Some European countries established extensive ‘welfare states’ after the Second World War, which (among other things, including healthcare and education) aimed to redistribute wealth and income through entitlements (pensions; sickness, disability and unemployment benefits; support for the extra costs of children, etc.). Others, like the United States, focused more on ‘help to those in need’ or, perhaps, those they thought ‘deserving’.
But national income’s also, obviously, critical to this. Rich countries and those whose economies are growing can do more to support vulnerable individuals and social groups than poor countries and those whose economies are stagnant. The SDGs locate this, therefore, as part of ‘ending poverty’: ‘All people must enjoy a basic standard of living’ their Agenda says, calling for ‘nationally appropriate social protection systems and measures for all, including floors,’ with by 2030 ‘substantial coverage of the poor and … vulnerable’ (Goal 1.3).
My focus in what follows is mostly on extensive welfare systems, because there’s more scope there to see what’s happening – but the principles, I think, apply more generally.
Social welfare (or protection) systems require three things: a policy commitment, broad public consent, and effective means of implementation. The last of these include effective means of identifying those needing or entitled to benefits (accurately, maximising coverage of those entitled, minimising fraud) and cost-effective means of delivering them (whether cash or kind).
The policy focus here’s a spectrum. Some governments are intent on maximising welfare, others on minimising costs. In recent years, governments in most countries with extensive welfare systems have been more interested in cutting taxes than increasing benefits. There’s been a swing in popular opinion, too, away from improving quality of life to providing basic safety nets.
So where does digital fit in?
How do digital technologies fit into this? In broad terms, they offer governments the chance to do things differently: improve efficiency and cost-effectiveness, target resources and expenditure. More precisely?
Establishing identity, for example, is critical to enabling welfare systems. Available resources should be provided to those they’re meant for and all of those they’re meant for. Otherwise they’re wasted or they’re less effective. This is why SDG 16 calls for governments to ‘provide legal identity for all, including birth registration.’ Digitalising identity facilitates entitlements.
Big data likewise help governments identify who’s in need – not merely individuals, but broader demographics (social groups, localities) and intersectionalities (the interactions between multiple factors contributing to deprivation: age plus disability, for instance). Resources should, as a result, end up being better targeted.
And digitalisation can make it easier for people to access the benefits that they’re entitled to. They can apply for them online or on the phone, for example, rather than in writing or at public offices. Claims can be processed through faster automated systems. Entitlements can be cross-checked through different databases. Response times can be reduced, helping those in sudden or in urgent need.
That’s all good, then?
Well, yes and no. In principle, each of those examples can improve social protection systems and the welfare of their beneficiaries. But, as with digitising other public policy arenas, each has problems.
Identity systems are complicated, and there’s much debate at present about how best to implement them. The scope for violations of privacy and for surveillance is obvious, especially when multiple datasets are interlinked. Government databases can be hacked by outsiders or abused by government officials, threatening minorities or diverting funds to those who aren’t entitled. All ID systems find it harder to handle groups like refugees and migrants. And in all systems some people fall through holes those systems haven’t fixed. Think how difficult you’d find life if you couldn’t prove you’re you.
Big data do indeed help governments to target the right individuals and social groups, but only insofar as the data they rely on are accurate or sufficient. Available datasets frequently under-represent particular groups within society – migrants and ethnic minorities especially. Some groups are deliberately marginalised or avoid drawing attention to their presence for fear of repression. Data analysis also depends on the questions asked of it: if women’s rights or those of a particular ethnic minority are unimportant to a government, for example, big data analysis won’t identify their needs. Everything here too depends on policy (see below).
Digitalisation can also make it harder rather than easier to access benefits. Many intended beneficiaries have poor digital skills and few digital resources (computers, even phones). If access is made digital by default, often promoted as efficiency, they’re not enabled but locked out. Algorithmic decision-making is rigid and inflexible: it’s hard to get redress when the computer says “no”, especially when officials are taught to trust computers more than human judgement. Systems can be set to fail applications for trivial reasons in order to save costs: many who’re turned down will not appeal when they’re rejected.
It all depends on policy
The point I’m making here – and this extends to other areas of the public realm – is that it all depends on policy: on what governments want to achieve.
If they want to maximise welfare, digital technologies can help them do so by identifying areas of need (including hidden need), targeting resources more effectively, ensuring that resources reach their targets (and as high a proportion of them as possible) quickly, efficiently and in ways that work best for those beneficiaries.
If they want to minimise costs, digital technologies can help them do so by restricting the range and number of beneficiaries, making benefits more difficult to access, reducing the scope for interventions that correct injustices arising thanks to rigid algorithms, and slowing down the speed of payments. If they want to use digital welfare for surveillance, or as part of ‘social credit’ systems, they can do so too.
As I mentioned earlier, the trend lately in government has been to minimise costs rather than to maximise welfare, to move from structures of entitlement to basic safety nets (or less). It’s often said that efficiency savings can be used by governments to improve effectiveness. In today’s climate, they’re more likely to be used for cutting taxes.
Four pointers for policymakers
First, we should recognise that there are no ‘technology solutions’ here. The problems addressed by social welfare policies aren’t due to digital divides. Digital technologies can help efforts to address them, but that’s not guaranteed and they can also hinder. The key determinant of what will happen is not digital but political: what do governments want to achieve?
Second, this is complex. Too much literature about it is promotional (from those selling ‘technology solutions’) or focused solely on risks (such as a recent report by the UN’s former special rapporteur on poverty). We need to look at both together, because they’re interlinked. Considering the nice without the nasty, or vice versa, polarises policy rather than enabling its development.
Third, the most important people here are those who need social protection. They’re almost always missing from discussion. Tech consultancies naturally focus on cost-effectiveness and selling kit. Rights advocates are not surprisingly concerned about privacy and surveillance. The poor themselves are mostly focused on where they get the money that they need for food today and medicine tomorrow. Their voices should be central.
And fourth, nothing’s immune here from coronavirus. Recovery from COVID will put enormous pressure on public finances. Those who are marginalised could bear the brunt of that in cuts to the social protection they rely on. The policy challenge is to ensure their interests are protected or better. The technical challenge is to help governments that accept that challenge to maximise welfare even with reduced resources.
Image: "Social security", by fabricatorofuselessarticles on Flickr Commons..