Economic design of democracy and bureaucracy

The other day I attended a lecture by Dr Tom Wilkening of the University of Melbourne. Tom specialises in economic design particularly of market interactions with significant complexity, such as auctions with hundreds of thousands of potential bundles of commodities on offer (e.g. mobile spectrum).

Economic design has emerged relatively recently, being a discipline in which economists help to directly design certain complex institutions. In the past, economists would tend to explain what exists and how things happen (pure science), instead of helping in the actual design better institutions. Indeed, as a general rule economists are (rightly) averse to designing institutions because they believe that markets will independently, through iteration and trial and error, determine the optimal designs. Indeed, there are huge risks of unintended consequences of artificially designed institutions. No doubt all this is true, but this argument fails in the case of design of government institutions. Such institutions have ALWAYS been deliberately designed. And because their design differs, their outcomes differ dramatically. We need to learn how to design these institutions better.

I asked Tom whether he knew of any significant work that has been done in the economic design of democratic institutions. He was not aware of such work. I suspect there is none – except my book, BFN.

As those of you who have read my book, BFN, know, my contribution in that book is essentially to ask us to think at the detailed level of incentives that operate within institutions. Given that not much work has been done by any economist in this area, it is quite possible that my work is among the first in the world to ask – and investigate – these issues somewhat systematically. I'm definitely not aware of anyone who has done any work on explaining why, for instance, different democracies or different bureaucracies result in different outcomes – and proposing a new design that will address many fundamental weaknesses.

The importance of economic design in the field of democracy is crucial. Consider, for instance, the Reform Act of 1832. It is my contention that without this act, which changed many incentives within the British political system and democracy, Britain would not have seen people like J.S.Mill as their representatives. Indeed, the outcomes (which are fraying at the edges as the fundamental principles of design are lost) that the British take for granted today are related to the economic design of the 1832 Act.

One of the key features of the Act, apart from rationalising constituencies and providing at least some opportunity for the upper-middle-class to enter politics, was to pay politicians for representing their constituencies. Prior to that it was expected that politicians would "manage" within the powers of allocated to them. That led to considerable amount of corruption and nepotism within the British political and bureaucratic system. Incentives matter. Indeed nothing else matters.

Then consider state funding of elections of the form found in Australia (see BFN for details). What this system does is to allow even extreme ideas to express themselves to the political process, an example being One Nation – a political party grounded on the racist ideology. One may disagree with such ideas, but it is important that they find a way to express themselves – and get booted out in the elections. That way those who could become terrorists in the future are neutralised. An opinion that the voters reject at the hustings is unlikely to arise as an extremist ideology.

But most importantly, state funding of elections reduces the need for the major political parties and politicians to raise funds for elections. No doubt they need some more money beyond what the state reimburses them, but the extra amount is minimal.  As a result, there is a greater potential for good people to enter the system – and no institutional incentive for corruption (unlike in the case of India).

In addition, politicians in Australia paid reasonably well. All these things, together, dramatically reduce incentives for corruption – although they do not eliminate bad policy (or even corruption). But at that stage things like the Ombudsman/Lokpal can come and catch the rogues (e.g. see this).

The need for experimental economics and economic design of political institutions

I believe that the discipline of public choice needs to be greatly expanded to understand better the underlying drivers of democratic institutions and bureaucracies. More importantly, I believe that experiments should be conducted in laboratories of different democratic (and bureaucratic) systems, to help us arrive at a better understanding of how different systems are likely to perform. 

It is well known, for instance, that a bad set of rules/institutions will ensure that even good people become bad, e.g: 

If you find that your class fellow with lower calibre than you has got a job in a government organisation (say, a school) by paying a bribe to the Minister for Education, then you are more likely to also give a bribe to the Minister.  One bad apple spoils the entire bunch. This is also clearly established by the simple game theoretic example of the prisoner's dilemma. Once cheated, people will cheat. 

This way of thinking about institutions is like that of a scientist who examines the causes of malaria. We cannot  determine the cause of malaria by looking at a patient or his environment (e.g. "miasma"). You have to look inside his bloodstream with a microscope, then you have to conduct extensive analysis at the detailed level. Only then can you determine that the malaria is caused by the malarial parasite. 

Such causal thinking is hard. Thinking about symptoms is easy. There exists a strong tendency among our "educated" classes to focus on the symptoms and not on the causes. Causal analysis generally requires more rigorous, more intellectually demanding thought processes, and more careful observation. In the case of institutions, it requires understanding INCENTIVES. 

In India we have a continuous stream of bad policy because people try to address symptoms, not causes. But symptoms of corruption (for instance) can't be removed without addressing causes. The challenge before educated Indians is this: can they think systematically? Can they think deeply? Can they explore causes and not jump to conclusions by looking at symptoms?

Just because you find that corruption exists, and currency notes of a high denomination might facilitate such corruption does NOT mean that high denomination currency notes CAUSE corruption. And so on. 

Addendum

Alvin E. Roth. "THE ECONOMIST AS ENGINEER: GAME THEORY, EXPERIMENTATION, AND COMPUTATION AS TOOLS FOR DESIGN ECONOMICS", Econometrica, Vol. 70, No. 4 (July, 2002), 1341–1378 [PDF]

Examples of bad thinking

Just today some of my interactions on the social media illustrate the challenge we face:

Srivatsa Krishna, a young IAS officer (about whose work I have heard many positive things) found objections with my suggestion that the IAS must be replaced with a modern non-tenured (contractual) bureaucracy.   

SrivatsaKrishna

@sabhlok thx beg to differ. UK, canada, singapore all have tenured civil services. Issue is not with the IAS but with people who staff it

I responded:

sabhlok 

@SrivatsaKrishna Small differences in incentives in the institutional design lead to massive differences in institutional outcomes. 

To this he wrote: 

SrivatsaKrishna 

@sabhlok no amount of change in incentives can be a substitute for either personal value of honesty or ability to show spine at d right time

To which I responded:

sabhlok

@SrivatsaKrishna Srivastava, these are not mutually exclusive. I think of these as of one piece. "Heroism" is not needed. Sense is. 

The point here is that we DON'T want India's bureaucracy to depend on the existence of "heroes" who will "sacrifice" for India. I DETEST the idea of sacrifice. It is false, unsustainable, hypocritical. Let's make the system such that it will perform wonderfully without heroes and martyrs. And that simply can't be done with a tenured civil service (or at least not with a host of other reforms which will address the many serious flaws – in terms of incentives – of tenured civil services).

Similarly, in response to my suggestion that state funding of elections is the first step to removal of corruption in India, I found a Lok Satta member providing the following comment (note that JP does understand this issue very well, so this member doesn't represent LS, and hasn't read JP's work in this area.)

LoksattaMember Loksatta follower 

@sabhlok there are too many causes for corruption.. first we need to have a framework to stop people eating money.

To which I responded:

sabhlok Sanjeev Sabhlok 

@ LoksattaMember When you have malaria, you don't put band aid. You kill the malarial parasite. Address causes, not symptoms. 

Indeed, Vishal of FTI has said this even better:

If you have a cricket team which cannot play cricket then where do you focus your energies on? getting the right team or putting checks on what cricketers are doing? You need a good cricket team and then have balanced checks. Having only checks will never win world cup.

But Vishal represents a minority in India: the few who can think of CAUSALITY, not symptoms.

I invite experts in economics who understand economic design to examine BFN and let me know of other comparable economic analyses of democratic institutions. In BFN I have tried to be like the scientist who explores the CAUSES of malaria and identifies the malarial parasite.

BFN outlines the PRECISE causal mechanism by which corruption (and bad policy, more generally) is generated in India. I trust those with academic interests will explore the hints in that book and study this very important matter further. Time permitting I will outline a systematic framework for theoretical and empirical investigations into this area that could inform the analysis and economic design of democracies and bureaucracies. Such findings would be invaluable to the world in the future.

I'm struggling with time and health, so I'd appreciate if someone with solid experimental economics and game theoretic skills can work with me – voluntarily of course – and explore the suggestions (and hypotheses) I have on this subject, in systematic detail.