The Costs of Poor Policy Design

This last week I've been ill and so has much of my family. I gave up on trying to meet work commitments. For a while did little other than clean up my watsapp messages, contacts and messages. One of the really random messages I received was from SBI which said something about logging into "web--banking" (sic) and taking a pledge to celebrate vigilance awareness week. I found the message interesting, because, like many other things, the role of vigilance as an idea, is predicated on the idea that corruption per se is an evil that needs to go.

I was reminded of the origins of the India Against Corruption movement which, for several people, was a watershed moment in India's political history. It brought all kinds of people together - many of whom considered others, within the movement, "corrupt". Today the movement is largely dead. Our current government, in India, was also elected on similar foundations - the return of India's black money. Demonitisation, for those who are celebrating its anniversary this week to co-incide with US elections, was also seen as a frontal assault on the foundations of corruption. Yet, corruption has neither gone away nor reduced according to any efforts to measure it.

Why do we worry about corruption so much? What about it is worse than communalism or or rape for instance, indicators that have also worsened along with corruption, according to many an independent observation in recent times? Perhaps it is the feeling of corruption being profoundly soul-sucking in its effect. Or the fact that it is faced by nearly everyone poor or not. Whatever the reason, one clearly needs to understand what corruption is.

In his recent essay on incentives, twitter's most popular economist - Vivek Kaul looks at the role of incentives and poor policy design vis-a-vis North India's smog problem. Among the other interesting points he makes, perhaps the most important, is the role of unthinking policy decisions.

The truth is that poor policy design impedes economic growth and recovery like nothing else - hell hath no fury like good intentions gone wrong. The road to recovery, is contingent on intelligent social policy design which desperately needs to acknowledge, in terms of design principles, two things: a) the nature of corruption and b) the fallacy of social policy targeting. What do either have to do with tackling a problem like unemployment or underemployment, India's current number one woe? Plenty actually.

One of the biggest criticisms of programmes such as MNREGA (which helps with the unemployment problem) has been corruption and with good reason. It does seem like it is impossible to implement a large-scale social welfare scheme without someone siphoning off the benefits mid-way. Think of the MNREGA as a canvas shoe with a tiny toe hole that wears down the whole shoe a little by little everyday – till it falls apart completely. To understand why corruption exists, one must first think about what corruption is. Corruption is a symptom, like a fever.

 To understand why economies – countries or systems – are corrupt you must be able to appreciate incentives. Human beings respond to incentives . In any resource constrained context (poverty, scarce resources and the like) actors have powerful incentives to act in their own self-interest first, hungry stomachs often lack the ability to be altruistic. At its core, corruption destroys any system, it is unethical and unjust. It also affects the poor the most because the poor cannot buy their way past it easily.

 In order to not be corrupt, one must understand why corruption happens, it is not enough to recognise that it exists. It is also insufficient to say it is wrong because the law says it is wrong. A symptom is a manifestation of a larger problem.

Countries have been and still are extraordinarily corrupt. Officials take bribes from ordinary citizens; they oppress the average citizen for small amounts of money. Why does this happen? In my experience if one is well- fed, largely, one doesn’t grudge food to another. In societies where the average person is underpaid, hungry, over-worked or simply treated badly (as is the case in authoritarian regimes), corruption and violence is often the result.

Sociological studies discuss this phenomenon in the context of patriarchy powerfully – women are often very patriarchal. Why? Because they are extremely oppressed themselves and when they have the chance to exert power over someone else, they do it in the worst ways possible for their own survival or success, in what is still a patriarchal world. Corruption is very similar.

One of the simplest ways, therefore, to stop corruption is to disincentivise it for the individual. Not by punishing or creating fear (although penal compliance is also one way of doing this) but simply by paying fair wages and costs. India’s corruption has gone down, not because it has suddenly become ‘digital’, but because the average person earns more money (per-capita income has steadily increased) and because the average person is more educated.

In other words, the incentives don’t reward small corruption as much as it did earlier, the transaction costs and opportunity cost of small acts of corruption, is now much larger for the average Joe. However, now incomes are shrinking again (in the pandemic and due to really poor policy design) and unless social policy does something about this, corruption is likely to come back in a big way.

At the center of any social policy, such as employment guarantee schemes where many human-incentives are involved, it is always useful to consider why corruption happens in human society. A seminal paper on the subject by Johann Graf Lambsdorff, explains that the idea of ‘reciprocity’ is at the center of it all. In summary it is this, ‘agents (read as corrupt persons) not only are guided by explicit incentives but also serve those to whom they owe gratitude’. 

There are two parts to the corruption puzzle – one side views corruption as rational and often wonders why there isn’t more corruption. Then there is the side that wonders why there isn’t less – this side is concerned with the question of the incentives facing a corrupt individual. Why should he or she reciprocate? After all nobody can force a corrupt individual to do what they said they would when they took a bribe. Law enforcement cannot help, reputations or further opportunities for corruption might not be powerful enough incentives.

To understand these two opposing viewpoints, knowledge of human history matters. John T Noonan’s brilliant book on bribes, offers examples from history, explaining that throughout human history bribes have been a legitimate way to assure outcomes for the individual.

Johann Graf Lambsdorff puts it thus:

 ‘The focus of norms of these early societies refers to reciprocity. Documents ranging from 3000 BC until 1000 AD deplore failure to reciprocate and herald threefold punishment to non-reciprocators. No evidence can be found that public servants should behave differently. No legal codes are known that prohibit gift-taking (Noonan, 1984, p. 9). Citizens would thus not approach public officials empty-handed, nor even gods. They are not outraged about greedy public servants but about those who fail to keep their promises.’

He further says:

‘Noonan describes how the thought of impartial justice is introduced and conflicts with the norm of reciprocity. More and more historical documents show how the importance of gifts starts to be balanced against claims for justice, for those who have little to offer.’

The paper is, of course, more complex than these introductory paragraphs, but what it does establish is that there is an enduring altruistic motive in agents (say employees, for example) to serve their principals (employers, for example). Officials and administration usually want good development work to happen, they just expect reciprocity.

Can reciprocity be built into social policy planning? As a bribe, probably not. As incentives, yes. India’s social policies have long since used incentives successfully. Nearly all of India’s public health worth on the frontlines is carried out by dedicated ASHA workers who work door-to-door for small incentives. There is little reason why better designed incentives at higher levels won’t work as well. The key aspect being "better-designed".

The second aspect of poor policy design has to do with social policy targeting. India, where hundreds upon thousands of people still live in abject poverty, and poverty- reduction programmes are implemented by bureaucrats trained in paperwork and not development, the one thing that is most obvious is this – poverty reduction programmes are meant for the poor. This is the basis of any kind of social programming.

However, how does one identify the poor, the really needy? We’ve tried any number of ways and methods. We’ve had definitions of the poverty line and the politics around that, we’ve had BPL cards based on income and a combination and social and income criteria, household size and we’ve also now tried digitised solutions such as Aadhaar.

 No matter what tool we use, identifying the poor seems to be an arbitrary exercise and there always are ways and means to use these tools to exclude people from social protection and welfare. Surely something must be amiss?

Economists tend to view the issue of targeting as a cost-benefit issue and this lens is of value. Colossal amounts of money are wasted in trying to target and effectively identify India’s poor because the ‘poor’ are looked at as an absolute segment of India’s demography. Unfortunately, poverty is a relative term, although one can be poor in an absolute sense too.

What makes matters worse is that poverty is also multidimensional. In simple terms one can be poor relative to others on one dimension and less so on a different dimension, or can be poor on multiple dimensions at the same time.

It is entirely possible that a person who is “poor” suffers from multiple disadvantages at the same time – for example, women in India often face disadvantages owing to their gender, their knowledge owing to exclusion from schooling and if they belong to a minority religion or are a backward caste (say Dalits for instance) they are disadvantaged three times over. The idea of multidimensionality suggests that focussing on one factor alone, such as income, is not enough to capture the true reality of being poor.

The poverty line idea tries to get around the problem of poverty being relative by fixing a threshold, but it is notoriously difficult to get to a critical value. Then there is the problem of parameters, suppose we decide on a value, what is an appropriate value? Income? What about food or savings? How is one better or worse than the other? Each variable varies and changes also, depending on when and how it is measured? For some variables averages work well, but sometimes averages don’t work well at all.

Governments tend to use indices of visible wealth markers (such as lists of assets - if you have a two-wheeler, a cow, a computer or a sewing machine or two you are not poor) to develop a proxy measure for who is poor, but these are also a problem. As with the BPL, if the populace at large knows what visible wealth markers are being used, they can store wealth in other ways, for example instead of buying gadgets or building a permanent roof they may just buy gold. Some alternatives include social identification of the poor at village level, but the oft-surprising challenge with this is that while people easily identify the really marginalised, they often don’t choose those who are relatively better- off than the absolutely deprived, but those who are still in need of welfare nets.

 The solution is to abandon targeting altogether. Instead of trying to identify and target the poor, social policy programmes should be designed to be self- targeting. It is important to allow social policy design to let people self-select themselves into welfare support programmes.

India’s largest targeted social welfare programme, the Public Distribution System (PDS) for food is also the programme which loses 58% of its supplies to a combination of ghost BPL cards, direct theft and mistargeting, as per a performance evaluation of the scheme by the government. That is a high cost, a cost much higher than the self-selection of some non-poor households into a social welfare programme.

Our current situation is precarious - for our youth, our elderly and nearly everyone in between. We are here because of a combination of short-sighted governmental priorities and terrible policy design.

For too long, India has underinvested in agriculture, manufacturing, education and healthcare basics which are significant priors for a youthful and productive working populace. Yet it is entirely possible to do better even now. All we need is better policy design, i.e. social policy that is not focussed on the theatrics of technology or targeting or keeping corruption away, but instead is intelligent, responsive and humane.

Cat lover. Peaceful bibliophile unless confronted by a fascist.