Can technology help the State spend less? As France plunges towards an abysmal public deficit of more than 6%, the question of modernizing public services is on everyone’s lips. Over the past year, 54% of net job creations are attributable to the public sphere. Has the perceived quality of service increased as a result? No, the feeling is actually quite the opposite. It is therefore urgent to question the productivity of this sphere. Direct public production encompasses areas as varied as they are important. Nearly 40% covers the provision of sovereign services – defense, security, justice, general administration. A quarter, education in public establishments. And 16%, care in public hospitals.
But how is the value of this non-market production calculated? By definition, it cannot be determined by sales revenue. It is, in fact, evaluated by the sum of production costs, that is to say the salaries paid to public personnel as well as the costs of other inputs and the depreciation of capital. Without sales, productivity must be measured against non-financial elements, namely processing time and case volumes processed. So how can we increase this productivity? More effective management can certainly help. But the bulk of productivity gains actually come from better software.
This is where the problem lies. The software equipment of the French administration does not meet our requirements for the public service. Certainly, services related to citizens have been improved. The remarkable Justif’Adresse allows, for example, to obtain your national identity card or passport without producing proof of address – it carries out an automatic comparison with the bases of energy suppliers. But these advances distract attention from another problem, that of tools for civil servants and contract workers.
Often brutal digital transitions
In November 2021, more than 3,000 magistrates signed a forum in The World on the deployment of Cassiopée, for Application chain supporting the information system oriented towards criminal procedure and children. Poorly designed, this tool – which is added to six other different applications! – generates enormous excess work for justice personnel, instead of saving them time. The user experience is also degraded.
We need to improve the way digital transitions are carried out in the public sphere. They are often too brutal. Cassiopeia gradually replaced software called Winstru, but the files closed on the old program were never transferred to the new one. The servers that ran Winstru were already down, so information was irretrievably lost. Some large projects entrusted to digital service companies have also suffered multiple delays and turned into financial pits. This is particularly the case with the Louvois software which managed military payroll.
Even more serious: certain essential processes in the public sphere depend on aging software or databases, developed by very small companies or associations whose future is not assured. Department of the Prime Minister, the interministerial Digital Directorate (Dinum), created in 2019, has the mission of developing the State’s digital strategy and piloting its implementation. Having become more political and less technical, it is doing remarkable work but it must be strengthened.
The development of state scale-ups, agile structures, capable of producing code and purchasing solutions that are more efficient than existing tools, is an absolute emergency. Generative artificial intelligence applied to the public service is also a virgin field that Guillaume Kasbarian, the new Minister of the Civil Service, Simplification and Transformation of Public Action, should explore. The pace of diffusion of AI in education is still far too slow, while the personalization of teaching enabled by these tools is a great opportunity to be seized.
.