It is by now a rite of passage for young technology industries to face a period of definitional uncertainty. Fintech experienced it. So too did biotech. Govtech – an emergent industry of small, innovative technology firms providing products and services for the public sector – is now undergoing this phase in its identity formation. Only two common denominators punctuate the five definitions of govtech included in the UK Technology Innovation in Government Survey: the promise of something ‘new’ or ‘innovative’; and the description of government as client.
Yet a clear oxymoron cuts through these amorphous definitions: narrowly defining government only as the client of the govtech industry risks doing anything but starting afresh. It may instead amplify the historical problems that have accreted around big, unchecked government IT contracts globally, in which procurers signed contracts and companies were left to get on with it. New, nimbler, providers and technologies are not enough alone to remake government technology. For govtech to deliver accountability and efficiency gains, governments must play a more robust and multifaceted role in the emergent ecosystem: as client, yes – but also as skilled procurer, project overseer, and enabler of genuine competition.
Examples of failed or overpriced government IT projects designed and implemented by large technology providers abound. In Canada, problems with a £500 million payroll system left some government staffers unpaid for months while others were overpaid. The lacklustre launch of healthcare.gov is another infamous fiasco. In each such case, it appears that technology firms may have acted carelessly or deceptively. But they were aided by scant technological understanding in the public sector. This information asymmetry enabled poor procurement decisions and failures of oversight. Michael King, UK Local Government Ombudsman, leaves no room for doubt: “The public sector can outsource the service but it cannot outsource its responsibilities.” The language of outsourcing often obfuscates this distinction.
When govtech stakeholders define government as client alone they perpetuate the IT outsourcing narrative – and risk re-entrenching the relationship norms that it facilitated. But the risks and trade-offs involved in a hands-off, client-vendor relationship today outweigh earlier ones. “Bad” IT projects of old were inefficient and spent taxpayer money imprudently. New technology contracts may bring far weightier forms of decision-making and even coercive power, with automated decision-making systems programmed and controlled by new govtech firms being tested in policy areas as sensitive as foster care and migrant relocation.
Tasking government with skilled govtech procurement and supervision does not relinquish all responsibility from the provider. The suggestion – as proposed in an AI Now Institute policy toolkit – that vendors should be contractually obliged to share methodologies and explain algorithmic models to agencies with which they work points in the right direction. But its utility requires sufficient knowledge in government for policymakers to assess the evidence shared with them, or at least access to reliable independent evaluators. As the policy toolkit observes, public procurers “often lack technical expertise to evaluate algorithmic systems, their capabilities, and potential consequences”. Under these circumstances, requiring vendor information-sharing risks becoming a formality, or its civic value dependent on public information requests.
In many countries, policymakers perceive the public sector skills’ gap as intractable and worsening, doubting that a government employee value proposition can be sufficiently competitive to siphon talent from large technology companies or venture capital-backed start-ups. Yet there are promising incentive programmes that could assist in developing sufficient inhouse expertise. The Singapore Smart Nation scholarship offers talented Singaporean students in technology-related disciplines paid university education in return for a period of decently paid public service. The model necessitates interest and skill among secondary school students: building the talent pipeline starts early.
Policymakers often imagine the ‘newness’ of govtech to lie not only in the size of its companies and the innovativeness of its technologies but also in its capacity to boost competition – a mechanism for driving up quality and accountability. But we should not expect real competition to issue from the industry unaided. As I document in Thinking about GovTech, large technology companies are also becoming interested in govtech firms. In 2016, Google acquired Urban Engines, an urban mobility start-up that combines big data and urban analytics. M12, the venture fund of Microsoft – among the biggest public sector technology vendors in many countries – has also invested in govtech. In the US, several large government technology providers including Oracle are buying govtech start-ups to increase their innovative capacity.
The rapprochement between corporates experienced in public sector sales and less connected start-ups may help to bring innovation into the public sector. But as large technology companies acquire successful smaller govtech companies, the relationship may not, in the long run, bolster competition for public sector contracts.
Where competition is an ambition, governments must, then, take on a further role: encouraging govtech companies to go it alone, particularly as they ‘scale-up’. As European Commissioner Margrethe Vestager argues, to grow a business rather than selling to a larger business, “you need money and new types of people so that you don’t feel the pressure to sell”. Govtech entrepreneurs around the world lament a lack of capital availability. Datasets to corroborate the anecdotal evidence await compilation but many investors still evidently perceive govtech as risky and unfamiliar. In the struggle for capital and talent, acquisition may appeal. Will policymakers incentivise or provide alternative financing mechanisms to better bolster long-term govtech competition?
The growth of govtech provides welcome acknowledgement that most governments do not have the capacity to develop and implement the most useful technologies alone. But for govtech to fulfil its foundational promise of breaking with past government technology inadequacies, policymakers must shed the historical one-dimensionality of their relationship with vendors. Government must reimagine itself – not just the type of technology companies with which it works – and accept a robust range of critical functions in the government technology cycle. This evolved relationship necessarily brings with it increased complexity, but we risk otherwise going back to a less accountable future.
Dr. Tanya Filer leads the Digital State Project at the Bennett Institute for Public Policy at the University of Cambridge. She is the author of Thinking about GovTech: A brief guide for policymakers.