Modern politicians, and – if polling is to be believed – modern electorates are preoccupied with law and order. Policing, rural crime, safety on our streets are issues of grave concern to the politicians promising more and more cops, and deliverance from threats to safety and security. At the same time, spending is being reduced, and outcomes are deteriorating in visible ways through reduced sentences, and lower conviction rates. The rule of law appears to be weaker.

In all liberal democratic countries today, rules-based systems for management, scoring and budgeting are proliferating, where algorithms score performance of social functions – like policing – and allocate resources accordingly. Tolerance thresholds are tested to assess the point at which people will reject austerity and cuts, addressing popular reactions with localised and targeted investment. The approach results in a minimum viable level approach to policing.

These rules-based systems are doing something else too. They are scoring the individual, designing incentives programs for job performance, promotion and preferment. The functions and objectives of individual policemen are quantified and reported upon in a thoroughly transparent and often subjectively oppressive way.

Rules-based systems – algorithmically implemented or not – have a performative impact on the service being delivered. Good behaviour is often not recognised within the system, and poor behaviour is sometimes rewarded. While on the one hand, resource distribution in an area might be statistically equitable, officer performance may be optimal, and crime rates may be acceptable, people may perceive there to be a breakdown of law and order. They may detect that the incentives for crime are higher – that there’s less chance of getting caught; that even if you do get caught, you’re not likely to be convicted; that you’ll be out on bail in twenty-four hours and not likely to see a judge for a year or more; that even if you do get convicted, you’ll only serve half your time.

The rule of law is based on two elements, one prescriptive and one invisible. The prescription is the law-book itself: things that are illegal, rules for discipline and punishment, and a court, prosecution and enforcement bureaucracy. The invisible element of the rule of law is the complicit populous, a people who agree with the system, who participate in the system, and who are a part of that system. This needs officials – police officers – who are focused on reinforcing the rule of law, and not just enforcing the law. That is the shift that liberal democracies are making, and the absence of rule of law reinforcement appears to be undermining the rule of law itself.

Of course, this is happening while the actual enforcement of the law is becoming ever more efficient. Fewer cops are needed, fewer stations, fewer prisons. At the same time – the algorithms and the statistics tell us – we are becoming safer, because the crime rates are falling. Every quarter police agencies all over the world release crime statistics showing how things continue to get better; and yet, too often it simply doesn’t seem that way at all.

Jeremy Betham’s idea for the panopticon – a prison design formed in a circle, with an elevated platform in the center from which every cell was visible – has often been referred to in the recent debates on privacy and surveillance. A key element of Bentham’s idea however is that the prison population becomes compliant, as it is always possible that any one of the prisoners is being observed. The rule of law is not merely about the enforcement of law; it is about the willing compliance of the lay population in its enforcement.

In a fascinating paper last year, Bayamlıoğlu and Leenes outlined three areas of concern in the rule of law under the auspices of data-driven decision making: law as a normative enterprise, law as a causative enterprise and law as a moral enterprise. Recognising these three, and acknowledging the poverty of algorithmic decision making in the execution of policing as competent to reinforce the rule of law, how far should we go? A rules-based approach is bound to fail; we need to redesign how we do policing.