Why do some laws fail?

Legislation is produced in vast quantities by law-makers in the UK and elsewhere. Much of it is, broadly, effective. But a significant proportion of that legislation fails to achieve its policy goals, or has unintended consequences. Or both. In Making Laws That Work I seek to study, and learn from, these failures.

I describe four broad categories of failed laws. First, the ‘damp squib’. Some legislation simply fails to produce the desired result. The legislation is passed, but the world doesn’t change (much). There are many examples of such failures.

A common cause of damp squib failures is that the people who are intended to benefit from the legislation, or to comply with it, simply are not aware of it. Or do not understand it. Or are unable to do what it requires. Laws can only be effective if the people whose behaviour they are intended to influence know what the law enables (or requires) them to do. And they are able to do it in practice. Many consumer protection laws are damp squibs: they confer rights on consumers that can only be enforced through court processes that are too complex, expensive and slow to be viable. Many family laws are also damp squibs, for similar reasons – the people intended to benefit from the law do not have the information and resources needed to obtain those benefits in practice.

The book explores some ways in which we can avoid damp squibs: improved information about entitlements (or obligations) that is provided to people as and when they need it, in a form they can understand; simpler processes for exercising rights; and ensuring institutions have the capacity to implement the law.

One good example of a law that avoids damp squib failure through smart legislative design is New Zealand’s child support law. If a parent (often, the mother) who is caring for a child has to go to court to obtain child support from the other parent, as is the case in many countries, that process is often too difficult and too expensive to be viable. And even if a mother does go to court, no money flows until the court decides the case (which can take many months). Under the NZ child support regime, an initial ‘formula assessment’, done swiftly and at negligible cost, determines a default amount that must be paid each week. If either parent objects, they can seek a review by an official, who has a limited ability to adjust the default amount. If a party still objects, they can go to court, where a judge has a broad discretion to adjust the amount payable. Importantly, throughout this process the weekly sum already determined must be paid: support does not wait till the legal process is complete.

Second, the ‘overshoot’ – the law that delivers what it was intended to achieve, but fails to stop there. A recent example is New Zealand’s 2021 responsible lender regime. New rules intended to protect vulnerable borrowers led to significant delays in approving home loans, and discouraged lending to many middle-income borrowers. Within months the Government was announcing changes to curb ‘unintended consequences’.

Third, ‘nasty surprises’. This self-explanatory category of failures is also pretty common. One famous example is the window tax introduced in England in 1696. It initially raised substantial revenue. But people do not like paying taxes. Over time, buildings were constructed with fewer windows. And existing windows were blocked up. The results of this can still be seen in many English cities. Tax revenue fell. And – the nasty surprise – reduced light and air in buildings had significant adverse effects on health. By 1851 the law was gone – but only after causing harm to many people, especially the working poor who flocked to cities during the industrial revolution.

Fourth, the backfire: laws that produce results that are the exact opposite of what was intended. In Bombay, under the Raj, the Governor became concerned about the number of poisonous cobras. A bounty was introduced on cobra skins. Enterprising locals began breeding cobras on the outskirts of the city: this was much easier than catching and killing the wild ones. Large sums were spent on the bounty. But there were no fewer cobras. Realising the policy was a failure, the Governor cancelled the bounty. The cobras in the farms were now worthless – so they were released. Result: more cobras then before. Nor is this a one-off example: there are many other examples of failed laws of this kind.

There are two main ways in which we can reduce the risk that laws will fail. First, we can adopt a more structured and systematic approach to gathering information relevant to proposed legislation: information about what is happening now, and about how behaviour is likely to change if the legislation is passed. In the book I explain how this might be done, and provide checklists to guide this work. Second, we need to bear in mind the limits of our knowledge and the uncertainty of our predictions about the future. We should aim to design ‘adaptive legislation’ – legislation that enables us to track what is happening over time, and adjust the law as we obtain new information, or as circumstances change in ways we did not anticipate. The book explores techniques for designing legislation to adapt in this way.

In all of this, I emphasise the importance of the ‘user interface’ of the law: the institutions that apply, interpret and enforce the law. It is impossible to design laws well without paying close attention to these institutions, and how they work in practice. Laws are like the code running in the background on your computer or your phone; the institutions that implement the law are the hardware and user interface through which we interact with that code. What matters in law, as in technology, is the outcomes that users experience. Bad code pretty much guarantees bad outcomes. But even the best-written code will fail to deliver good outcomes if the code and the hardware don’t work well together, or if the interface is complex and difficult to use. Legislation needs to be designed in tandem with its institutional interface – implementation cannot be treated as an afterthought, to be dealt with down the track after the law is made.

There is a closely related point: the importance of default outcomes, and the ‘stickiness’ of those defaults – how hard it is to move away from the default. Well-designed laws and institutions make it as easy as possible to get to the outcome that the law is intended to achieve: either by setting that outcome as the default, or by making it easy for people to get from the default to that outcome. Technology is important here – it is easier than ever before to gather and apply information to tailor defaults to particular circumstances. Think variable speed limits, updated for road conditions. Or tax systems that adjust earner deductions to ensure the correct amount of tax is paid in each year, or that generate automatic tax refunds without the need to file a return.

We need to explore how we can reduce complexity as experienced by the user of the law. Here too, interfaces and defaults are critical, and technology plays a growing role.

If those of us who are involved in the business of law-making hold ourselves to the same standards that we expect of other professionals whose work affects public wellbeing, then we don’t have the luxury of continuing to design laws based on inadequate information and analysis, and disproportionate confidence. We need to do better. The book makes some modest practical suggestions for how we might do that.