Kunal and Gaurav are right. In this post, I was talking of the trade-off between rules and principles in policy enforcement.  Actually, when you think of it, rules vs. principles is not a simple dichotomy, but a spectrum of choices.

Imagine that these choices are spread from left to right. At the left end of the spectrum is automation. Rules are enforced automatically, without anyone having the responsibility of enforcing them. The best example  of automation in my stories was actually the turnstile – it automated the task of checking for tickets, leaving very little scope for discretion.

Another example is the jugad “automation” that the Hyderabad police enforced. Blocking off the right turn doesn’t seem like an example of automation, but for our purposes, it is, because it enforced the rules without the police having to intervene.

To the right of automation come rules – clear and transparent rules that leave no scope of discretion to the enforcers. But then, whether to follow the rule or not is still a choice – and ensuring that officials enforce the rule depends on the existence of procedural mechanisms.

As you move further to the right,  you find that the rules have more and more discretion embedded in them. For example, consider the difference between enforcing a red light and ticketing someone for rash driving. The former is easier to enforce fairly than the latter.

At the extreme right of the spectrum is the idea of “principles-based regulation”.  This distinction  between rule-based regulation and principles-based regulation is used most often in the financial sector, so let me use an example from Banking to illustrate.

Banks in most countries have to abide by “KYC” norms – Know your customer. The idea is that they should make  their best effort to know the real identity of their customers. The intent is to prevent money-laundering and financing crime. Regulators could enforce KYC in two ways –  You could state the regulation in general terms – something like “Banks should establish and enforce polices to ensure that they know their customer’s identity at all times. These policies should enforce appropriate weight to the risk that customer represents”.  Or the regulation could specify everything in great detail. It could specify which identity documents the banks may accept, which forms of address proof they should accept, and so on and so forth. The former is principles-based regulation while the latter is rule-based regulation.

Our first preference is to automate the enforcement of those rules, even when they lead to sub-optimal outcomes. Because we can’t rely on the police to enforce speed limits, we overdose on speed-breakers. Because we can’t enforce red lights at an intersection, we build flyovers where none are needed. As the example of the insurance company officials indicates, we’d like to automate hiring and promotion decisions as well, preferably through written exams.

Our financial sector regulation is heavily rules based, much to the annoyance of Percy Mistry and Aadisht Khanna.  And not just our financial sector – our instinct is to demand detailed rules for everything. Indian mutual fund companies have been asked to enforce KYC compliance by collecting an identity proof and an address proof (list specified) and they have all got together and centralized this job . All well and good, except that I am happily KYC compliant even though the address on record is 2 years old and out of date – the AMCs have complied with the rules, but they have not bothered with the principle, which is to know your customer, not just collect documents.

Detailed rules cannot cover every situation and often create loopholes, but we still demand them. Why?  Because we don’t trust our government’s competence or honesty, and with good reason.  You can’t trust politicians with choosing the Chief of Army Staff or the Chief Justice of the Supreme Court, so institute a strict seniority system. You can’t trust bureaucrats to not take kickbacks while buying weapons, so load the purchase requisitions with extremely detailed specs and put in a rule requiring them to choose the lowest bid.

Paradoxically, however, while we don’t trust our government to do anything competently or honestly, it hasn’t stopped us from demanding that the government take on more and more responsibility. This means that our rules are not only extremely detailed, they are also contradictory. I think it was Madhu Menon who once told me that in Bangalore, the excise department required bars to have only one door (to stop you from running out during raids) while the fire department required you to have at least two. When you expect the government to do more and more, it is inevitable that the rules are a tangled mess. Naturally, you end up worsening the problem of corruption that you set out to solve.

As you move to the right of the spectrum, government’s discretionary powers increase. Citizens lose the certainty of clear and transparent rules, but if the principles are applied fairly and consistently, it is a net gain. That is because rather than have rules that guide your every step, you gain the freedom to plan and make decisions, as long as your decisions adhere to the overarching principles.  But for this to happen, three things are necessary:

1. Better ability to detect wrongdoing by citizens after the fact. When traffic management moves from blocking off right turns and policemen directing traffic to traffic signals, you reduce the need for close supervision. But traffic signals work only if the police are capable of catching you if you jump a signal. In Mumbai, they do it by waiting for you past the traffic light and catching you. Because you don’t know whether the police are lying in wait or not, you will obey traffic lights in all circumstances (or at least during daytime) This reduces the amount of supervision that is required.

2. The system should have the ability to distinguish among a sensible decision by a government official that went wrong, an incompetent decision and outright wrongdoing. The first should be protected, the second should show up in your performance appraisal and the third should send you to jail.
3. Our courts and tribunals should have the capacity to interpret principles and hand out sensible decisions consistently and quickly.

 Of course,  our system currently lacks all three. We do not have an ability to detect wrongdoing. Our system does not have an ability to reward government servants for good performance or punish them for incompetence, and our courts do not hand out quick and clear decisions.

 In addition to the above, we need to reduce the number of things government does. We also need to recognize that the government needs to have the right mixture of rules and discretionary power.

 None of these can be done in isolation and if we try, it will make the problem worse. For example, the Jan Lok Pal idea does nothing to reduce the scope of government, and it does nothing about our investigating agencies or courts. Instead, it seeks to bypass the problem by  treating every irregularity – i.e. a violation of a rule as a proxy for corruption. Given how corrupt our system is, this is probably accurate. But the problem is that if we do this, policy paralysis is inevitable.