The Coronavirus downturn: social protection under threat

Coronavirus-related lockdown measures have impacted almost 2.7 billion workers, with some countries seeing unprecedented levels of applications for welfare benefits support. 

In response, emergency relief legislation for welfare recipients has been fast-tracked worldwide, from the UK to Brazil. These measures, combined with the growing awareness of Covid-19's differentiated impact along the fault lines of class, race, gender and legal status, rightly seek to address the needs of those most vulnerable. However, these measures often come at a costly price that is easily overlooked, and eventually forgotten. As the UN Special Rapporteur on extreme poverty recently put it, these support packages have left the most vulnerable "short-changed or excluded".

Finding the beneficiaries: a "databased" solution

The current welfare solutions aimed at supporting those most afflicted by coronavirus attempt to capture a novel group of beneficiaries. While the eligibility criteria may vary from one jurisdiction to another the targeted beneficiaries are commonly those in a situation of poverty and/or vulnerability exacerbated by Covid-19. Surprisingly, in some countries, like Colombia, there is no need for purported beneficiaries to sign up or apply - governments believe they can accurately determine who needs extra support from the data they already have.  

On 4 and 5 April 2020, Colombia passed legislation creating a monetary supplement, the Solidary Income (Ingreso Solidario) in response to Covid-19. As was made clear by the accompanying regulations, the Department for Planning - the government ministry responsible for benefits - would select the beneficiaries based on the SISBEN, Colombia's welfare classification system.

Since 1994, all candidates to welfare benefits in Colombia are ranked from 0-100 through SISBEN's algorithm. Their entitlement to benefits will ultimately depend on the score they obtain, with higher scores resulting in inegibility. The SISBEN algorithm and data processing methods are classified, resulting in a welfare black-box. The Solidary Income, like every other benefit in Colombia, relies on SISBEN's scoring. With SISBEN's data as a starting point, the Department for Planning resorts to data-matching with other government databases, including the Departments of Revenue and Health. The selected beneficiaries are subsequently notified via text, and receive the benefit directly to their bank accounts.

Despite the wealth of data held by Department for Planning, it lacked the means to communicate with the Solidary Income beneficiaries: crucially, SISBEN did not house contact number data. To allow for the beneficiaries to be informed via text, a government department decided to lift the prohibition on telecommunications companies to disclose communications data. After a backlash by civil society organisations outlining the risks of enabling the government to access communications data, the prohibition was reinstated. Instead of being told of their eligibility by the government, beneficiaries are told by their banks, and the funds are deposited in their bank accounts. As a result of the individual being completely excluded from the eligibility determination process, it is impossible for anyone to know whether they were in theory entitled to the benefit, but simply failed to receive it due to a government or bank mistake. There are reports of individuals being mistakenly awarded the Solidary Income. The reverse could also be true. 

The Colombian example reflects the limitations of the current use of automation for welfare delivery. In the absence of information relating to the SISBEN algorithm, it is difficult to know whether all intended recipients are being reached. Given that the criteria for entitlement to the Solidary Income aren't public, non-receipt of the benefit cannot be challenged. The Solidary Income can be as punitive as it is exclusionary. The Solidary Income legislation imposes penalties on those having mistakenly received the benefit and failed to report it, despite the fact that mistaken recipients could not have done anything to receive - or indeed avoid receiving - the benefit. 

Exchanging privacy for survival? 

Brazil has launched an Emergency Support (Auxilio Emergencial) benefit aimed at informal workers, small business owners and the unemployed. Some of those eligible need not take action to receive the benefit; others need to register. 

Anyone meeting the conditions set out by the Brazilian government may register to indicate their eligibility. In order to submit the application form, however, they must authorise the government to access their personal data to verify that they qualify. This seemingly innocent request takes a turn for the intrusive when, on the next page, the applicant is asked for his unique taxpayer number, date of birth, and mother's name. 

Crucially, the platform has no privacy policy - and Brazil's Data Protection Law is not yet in force. Those seeking to obtain Emergency Support cannot know for what other purposes their data will be used, who it will be shared with, and where and for how long it will be stored. On 28 April, the entry into force of key sections of the Brazil's Data Protection Law - including those outlining data rights - was postponed. It is notable that the interim legislation postponing the Data Protection Law also addressed practical, non-privacy related provisions regarding the Emergency Support. 

Individuals will struggle to exercise their data rights even after the Data Protection Law comes into force. After a series of vetoes to specific articles in the Data Protection Law - a power held by the Brazilian President - Bolsonaro significantly eroded the legislation's potential. As a result of these vetoes, Emergency Support recipients submitting data subject access requests to learn of the fate of their data may have their information disclosed to other public and private entities, potentially resulting in unforeseen reprisals or black-listing by other agencies. Unsuccessful applicants are likely to face higher hurdles: if the selection process for Emergency Support is automated, under the Data Protection Law, they will have no right to challenge the decision.

The long-term impact

Even where Covid-19 inspired benefits systems can be presumed effective, history tells us that they are likely to be short-lived. 

Nowhere was this made clearer than during aftermath of the 2008 Great Recession. A study looking at changes in welfare policies across the EU after the 2008 crisis revealed that, despite initial measures to protect citizens and workers from the effects of the crisis, structural reforms aimed at reducing public spending undermined social protection. The US responded to the crisis by relaxing eligibility requirements and increasing available funding, until in 2011 the Budget Control Act applied public spending cuts.

In a world with shrinking public budgets, pressure to optimise the use of taxpayer's money is high. The use of technology - or automation - in the allocation of public funds is one solution which governments turn to in order to ensure that resources are efficiently used. Automation has the potential for good where efficiency is defined as the allocation of resources to those who need them. However, as automation in welfare benefits has often shown, the modern meaning of efficiency does not equate with securing the wellbeing of the many, but cracking down on the perceived criminality of the few.

The post-Covid-19 cycle of welfare restrictions has been already foreshadowed in the US, with the president of the American Enterprise Institute stating that any fundamental change to the US benefit system in the wake of Covid-19 would be a mistake, potentially creating “an entitlement system that in good times leads to people passing up the opportunity to work”.  That statement is illustrative of the regrettably persistent narrative that some people are poor as a result of their personal failings, and not the intersection of structural, broader forms of marginalisation. 

Welfare benefits cuts are not, however, an inescapable consequence of the pandemic: the Covid-19 crisis has the potential for being a catalyst for positive change. Spain recently announced plans to roll out universal basic income on a permanent basis, with economists debunking the long-standing myth that universal basic income is simply unaffordable. Similarly, temporary fixes may pave the way for the overhaul of intrusive and outdated welfare systems. The Irish government has temporarily suspended the requirement for welfare applicants to obtain and present the long-decried Public Services Card (PSC) in order to access an unemployment benefit, revealing the PSC to be an unduly burdensome and ultimately unnecessary requirement. 

It is possible for short-term welfare measures to comprehensively address the needs of those most vulnerable while also protecting their rights. To do so, welfare systems must ensure that the following conditions are met: 

• Eligibility criteria are fair, transparent and lawful, and should not be solely managed through automated systems
• Individuals are informed, in the clearest possible terms, about how their data will be used before they are asked to submit any personal information, if it will be subject to any automated activities, how long it will be stored for, as well as provided information about any third-parties with whom their data will be shared and for what purpose 
• Application forms and platforms only require the applicant to enter information which is strictly necessary to assess their entitlement to the relevant benefit
• The databases where personal data is stored are secure and protected from unauthorised access, destruction, use or disclosure
• Any data-matching and/or data-sharing mechanisms used are prescribed by law, and use of personal data for these purposes is strictly forbidden where it was not consented to by the individual
• Where systems and/or databases are integrated as part of the entitlement determination process, applicants are told what specific databases are involved, and which entity is responsible for the integration 
• Where automated scoring systems are used, applicants must be told of the variables that will be used to determine their entitlement, and they should have the opportunity to ensure the data sets/points used to come to a decision are accurate
• Any decision to deny benefits is subject to review and/or appeal, and the appropriate process is clearly explained to the applicant including relevant deadlines