I have been reading Hayek’s The Road to Serfdom, which has been a useful reminder of what it means to have rule of law, and why it matters.
(I’ve previously offered a systems perspective on the survivability of rule of law under assault, as well as a view that law is essentially a mediating format that lowers the cost of interpersonal interactions.)
Hayek’s core observation on this subject is that the rule of law is impersonal: at least in principle, the resolution of a legal matter proceeds through the same steps regardless of the identity of the parties to the case.
Elements
From a computer systems perspective, I might break down this definition into three distinct elements:
1. The resolution of any given legal matter is algorithmic in nature, with clearly identified inputs, outputs, and processing steps.
2. The overall process allows for the possibility of exceptions and/or failures, which are handled by some kind of associated appeals process.
3. The algorithms don’t have steps that depend on the specific identities of the parties.
In practice, there are exceptions to these characteristics; nevertheless, this is not a bad summary of the “design center” for a legal system. If a system works mostly like this, we give it the “seal of approval” for conformance to rule of law. If it mostly doesn’t work like this, we say it’s lacking in that dimension.
The first item is violated whenever:
there is no identifiable process (arbitrary decisions), or
there is an identifiable process that is not actually followed in some or all cases.
The second item is violated whenever:
there is no appeals process for failures in the “main” process, or
the appeals process is not actually followed in some or all appeals.
The third item is violated whenever the process has an implicit or explicit branching structure that depends on one of the parties. At the most extreme, this could be a so-called bill of attainder that actually refers to a specific person by name, and either grants favors or exacts penalties.
More reasonably, but still a possible cause for concern, are laws that have different steps selected depending on different categories of persons: men vs. women, wealthy vs. poor, black vs. white, tall vs. short, etc. However, some categories are legitimate for the purposes of law: for example, minors are distinguished from adults in many situations.
Note that this issue of selecting different resolutions is different from having different effects. For example, consider a law that includes a $500 penalty (or reward). That amount of money will likely have a very different effect on a poor person than on a rich person, which may be an issue for the authors of the law to consider. However, that difference in effect could well be defensible. In contrast, now consider a different law that first sorts out the parties into a “rich path” and a “poor path,” with different penalties or rewards on each. That might be a structure that is introduced for a noble purpose (ensuring similar impact, regardless of wealth) or for dubious purposes (ensuring that rich people aren’t inconvenienced). Regardless of the motivation, such a structure would be inconsistent with rule of law.
Predictability
Why does rule of law matter? What do we gain from it, and why is it worth defending? It’s all about predictability, but of a specific kind. We need to distinguish here between the inherent unpredictability of any specific case vs. the overall “process” predictability of many cases. In law, we can know what the steps are in that process without being able to know for sure what the outcome will be for a given set of inputs.
This situation is analogous to the way that we can read a program and understand it completely, without necessarily knowing exactly what outcome it will yield in any particular execution. Sometimes that uncertainty is due to randomness in the algorithm itself, and sometimes that uncertainty is due to our incomplete knowledge of one or more inputs. (I have written previously about the possibility of using program execution as a model for the illusion of free will.)
Even though we don’t know the legal outcome in advance, we can have confidence (a) that we know what steps will be executed, and (b) that in the absence of formal processes to change it (e.g. via legislation), it will be the same steps executed that were executed for similar situations the previous day, as well as being the same steps that will be executed for similar situations the next day.
This predictability is good for business and investment, but it's also reassuring at a personal level. If we are not confident in rule of law, everyday life necessarily becomes an exercise in low-level paranoia, trying to avoid triggering an engagement with the arbitrary legal system. Actual encounters with the law-enforcement apparatus become especially fraught, and more so if agents are unidentifiable and unaccountable.
Rule of law is separable from democracy. Both are appealing to the governed people, and both are currently under assault in the US. However, there is no requirement that a rule-of-law system be democratic, nor that a democracy operates following the rule of law. Both democracy and rule of law are impediments to tyranny, and the avoidance of tyranny was important to the authors of the Constitution. I plan to consider a systems view of democracy in a future essay.