Syndicate content

It’s not the How; It’s the Why

Shanta Devarajan's picture
Also available in: Español | Français | العربية

Hardly a week goes by without my hearing the statement, “It’s not the What; it’s the How.”  On the reform of energy subsidies in the Middle East and North Africa, for instance, the discussion is focused not on whether subsidies should be reformed (everyone agrees they should be), but on how the reform should be carried out.  Similar points are made about business regulations, education, agriculture, or health. I confess to having written similar things myself.  And there is no shortage of such proposals on this blog
Reforms are needed because there is a policy or institutional arrangement in place that has become counterproductive.  But before suggesting how to reform it, we should ask why that policy exists at all, why it has persisted for so long, and why it hasn’t been reformed until now.  For these policies didn’t come about by accident.  Nor have they remained because somebody forgot to change them.  And they are unlikely to be reformed just because a policymaker happens to read a book, article or blog post entitled “How to reform…”

Rather, these distortions (shorthand for “counterproductive policies and institutions”) exist and persist because politically powerful people benefit from them.  In Egypt, politically connected firms are among the most energy-intensive ones, and hence benefit from energy subsidies. In Tunisia, restrictions on foreign investment in certain sectors protected former President Ben Ali’s family firms, enabling them to capture 20 percent of economy wide profits. In India, politicians and bureaucrats tolerate teacher absenteeism in public schools because teachers campaign for local politicians and bribe bureaucrats for transfers.  In many countries in Africa, regulations prohibiting entry into the trucking industry remain on the books (and keep transport prices high) because a family member of the head of state owns the monopoly trucking company.

In this setting, telling governments that they should reform these policies, or even showing them how to reform them, is at best unhelpful.  Yet policy papers and reports invariably have a concluding section called “policy recommendations.”  One of my favorites is the report on sector X that always ends with the recommendation, “Government should spend more on X.”  Have they thought about why government up to now has not spent more on X?  Perhaps it is a political equilibrium.  After all, spending more on X means spending less on something else—which sector do they propose government should cut spending on?

Recognizing that distortions are the result of a political equilibrium also sheds light on the refrain, “One size doesn’t fit all.”  This is often leveled at economists, because the same policy reform (trade liberalization for instance) was tried in different countries with different (and sometimes disastrous) results.  But the economic analysis that showed the benefits of trade liberalization applied in every one of those countries.  The politics, however, were very different, so that in some countries, certain groups were able to undermine the reform while in other countries the beneficiaries managed to sustain it.  Dani Rodrik makes a similar point in his book, One Economics, Many Recipes.

If distortions represent a political equilibrium, what can be done about them?  Are all lists of policy recommendations, including those on this blog, useless?  Of course not.  Most (but sadly not all) policy recommendations derive from some analysis that shows the benefits and costs of the policy change.  This analysis, and its link to policy reform, is extremely valuable—not in telling the policymaker what to do, but in informing the general public, the majority of whom are likely to benefit from the reforms.  The analysis can nourish an evidence-based debate among the public, which increases the chances of a political consensus around the reform.  Only if there is a political consensus—not if he reads it in a report--is a politician likely to adopt the recommendation.  Getting this analysis out to the general public, then, is the main purpose of this blog.

But we may be able to do more than just inform the public.  We now have a better understanding of how political elites are able to capture public policies to their own benefit.  With this knowledge, can we design policy reforms to minimize the chances that the new policies will not also be captured?  Jim Robinson, who calls this “politician-proof policy”, is skeptical because we don’t understand the incentives facing politicians.  But in the nine years since Jim published his paper, we have gained a better understanding of political incentives, so it may be worth trying.  Suggestions welcome!


Submitted by Waleed Addas on

Shanta, permit me to add one general comment: when people are used to paying a lower-than 'market price' for a good or service, say for gasoline and have been like that for a longer period of time, any new policy recommendation needs to be presented in such a way that it gives the policy makers some feasible options. A new policy to remove or reduce 'subsidies' may be difficult to implement if it does not also provide serious alternatives. In other words, to reduce the price of gasoline in a country that has no other means of transportation except automobiles would be quite impractical unless these 'policy safe-guards' are carried out first such as alternative means of transportation etc.

Finally, new development/economic policies need to take into consideration such effects on the whole of society/economy as our future development hangs on this basic premise: to move from the current compartmentalized or specific sectorial view to a more integrative universal perspective.

Submitted by Shanta on

Waleed, thanks for this excellent point. Precisely because of potential political resistance, we should always propose alternatives, rather than a single, "optimal" policy reform. Otherwise the policymaker is working with one hand tied behind his back. I recall some years ago when the Delhi Water Board introduced a management contract for managing water in some neighborhoods. When there was fierce political resistance--people were marching in the streets--the Bank was unable to give the head of the water board other options, claiming the management contract was by far the best. The net result was that the program was cancelled. Incidentally, the source of the political resistance is not just because people are used to paying lower prices. It is often because some people are benefitting from the gap between the subsidized price and market price (by for example smuggling the subsidized good across the border), so they stand to lose a lot of money if the reform goes through.

Submitted by Wolfgang on

Congratulations for such a great post at the occasion of the first anniversary of “Future Development”. I have a comment and a question:

Comment: It is true that outside players are too often naïve about domestic politics. They indeed need to ask the “why” question more often and systematically. However, there are also many cases where domestic players (even the top) were determined to carry out reforms, sometimes due to market pressures (e.g. Indonesia’s fuel subsidy reduction) or other reasons (e.g. Germany’s labor market reforms in 2003/04). In these situations the “How” still matters, sometimes even the “What”.

Question: What is the practical implication of focusing on “why”? As you highlighted in your very first blog, one objective of our work (and blog) should be to help shift the balance of power and provide a platform for an evidence-based policy debate. In the case of Egypt, would this mean that analytical work should focus more on the composition and activities of the energy-intensive industries? Here would be an analogous example from environmental analysis: The World Resource Institute has been mapping forest fires in the world and in Indonesia in real time; they then overplayed this data with the boundaries of concessions (mainly palm oil plantations) and can now point to the culprits of the fires:

Submitted by Shanta on

Thanks, Wolfgang. On your comment, if a leader is determined to undertake reforms, he or she still has to overcome the vested interests that will block them. In the example you cite, it was market pressures that forced the government's hand. But it took a prior decision by government to subject itself to market pressures (e.g. By opening up the capital account). Often, governments are reluctant to do this, citing good empirical work about the problems, because it ties their hands and leaves little room for discretion and hence for rent-seeking. On your point about using analytical work to inform the debate, you're right (of course), but we need to be careful about both the content of the analysis, as well as how it is communicated to the public. Stuti Khemani's work has some evidence, including examples where getting these two things wrong can lead to unanticipated results.

Submitted by Shahzad Sharjeel on

Dear Shanta, now that we have established that 'subsidies' continue to exist because they benefit certain constituencies and that politicians/policy makers are usually among them or reliant on them, the next challenge is to find a win-win for the majority which MUST include the politician. The win for the politician has to be a) her ability to retain her own seat, and b)gain new supporters for the party to make up for the lost support among the 'subsidy losing' constituents. In order to convince the general population that across the board subsidies actually hurt them as the resources spent on subisdizing everyone restricts the government's ability to spend more on human development, the policy makers will first have to engender a sense of ownership regards the national resources among the populace, they must believe that the resources saved will actually be spent on their development and wellbeing, only then they would feel that they have a stake in the system. This gives citizens the incentive to forego a cheaper unit of electricity or a liter of fuel today for better education, health, clean drinking water, better roads, transportation tomorrow. In the absence of this people will continue to grab for the subsidized goodies today. Reforms will continue to be seen as donor-driven. The dots are not difficult to connect but they have to be connected by someone the citizens can trust.

Submitted by Shanta on

Sharjeel, thanks for weighing in. For the reform to benefit the politician, she must be convinced that the voters--the majority of whom will benefit from the reform--will vote for her afterwards. For this to happen, voters will need to vote according to outcomes (attributable to actions by the politicians) rather than along ethnic, tribal or religious lines. This is where information can play a role. If voters are better informed about the performance of politicians, they may shift the way they vote. But such a result is not guaranteed, as the work of Stuti Khemani and Phil Keefer shows. More generally, as you suggest, the fundamental issue is trust. As long as voters don't trust government to deliver better services (and not steal the money), they will not vote for subsidy reform, even if they don't benefit from the subsidies. How can we shift from a low-trust equilibrium to a high-trust one? I think this is a fundamental challenge, including in some MENA countries today.

Submitted by Joel S. Hellman on

I just love it when economists recognize that it all comes down to politics in the end! But as you rightly ask, so what? What tools do aid agencies like the World Bank have to tailor their support to political realities within the scope of their clearly apoliical mandates? You hit on one of the key answers: the provision of technical advice to reform minded politicians, bureaucrats, and the informed public. I always find that this obvious answer is increasingly "pooh-poohed." In the rush for us to do political economy analyses and to promote politically feasible reforms, we sometimes forget that we will never have enough information to adequately diagnose the full complexity of political contexts as they relate to individual reform decisions. Political scientists (as well as political pundits) have an awful record trying to predict policy choices (just see the recent work of Tetlock). So perhaps we should be more comfortable with the idea that our role is to bring information about alternatives to reform-minded politicians, bureaucrats and the public and in the hopes of helping them duke it out politically on a more informed basis. We know that they will lose in many instances, but that is the nature of the reform business.

One other thing strikes me about your post: you talk a lot about policies, and not so much about institutions. Most of the examples you cite are of bad policies and or Jim Robinson's notion of politician proof policies. Policies don't make themselves. They are forged within institutions. And one thing we really don't understand is the process of institutional change. We know that institutions are critical for development. And we sort-of know what good institutions look like (though there are many diverse forms in different contexts). But what we know very little about is how institutions change. Institutional change is quite different than policy change (though the two are obviously related). Acemoglu and Robinson see institutional change as a historically induced process affected by "critical junctures." But what do we really know about critical junctures (other than we sort-of know them after we have seen them)? And I think that the critical junctures story hides a more interesting process of gradual institutional change, the dynamics of which we know very little about. Giuven that the World Bank is perhaps one of the largest investors in institutional reform in developing countries, we should invest far more in understanding institutional change. That seems to be a key take-away from your post. Sounds like we need more "institutional economists" (or does that sound too self-serving to you)?

Submitted by Shanta on

Joel, agree on both points. First, we should not spend a lot of time understanding winners and losers. Simply collect the information about the consequences of policies, and put it out there in a manner that voters can access. Second, it is institutions that determine policy change. But not only do we know very little about how institutions change, but I think our ability to affect that process from the outside is limited. While I agree that we should understand the process of change better (and I'm delighted that we have a chief institutional economist now), I think we should continue to collect information about the consequences of policies and put it out there, while studying institutional change.

Submitted by Nachiket Mor on

This is a great post Shanta. While I am certainly no expert, I feel that in complex democracies such as India where even astute politicians cannot accurately guess what the electorate truly want and are often surprised when they lose their deposit, and the process of building consensus is so slow and opaque that often it seems that there is no movement at all, "big and bold" reform efforts may not be the way to go, but instead a carefully considered series of "butterfly effect" moves which are obliquely aimed. This requires a great deal of patience and time frames stretching over multiple decades.

Submitted by Shanta on

Nachiket, thanks for your comment. We should be putting the information about both kinds of reforms--bold and incremental--out there. Some of it will have impact, and others will not. But it's hard to predict ex ante which is which. Shanta

Submitted by AminaSemlali on

Dear Shanta,

Excellent post, as always. It is indeed a must that the public learn about the benefits of subsidy reforms.

In MENA citizens have resisted subsidy reform and sadly the poor often seem to share the belief that the subsidies at hand are their best option - while the economic/political elites continue reaping the main benefits.

In addition to concretely showing the populations that there are in fact more effective alternatives to subsidies the question is how -in environments often based on mistrust- societal mindsets can be changed and new norms adopted?

How to make a society (and within that society - especially the ones most unwilling to give up their own perks) not only intellectally but also emotionally understand that an improvement in the wellfare of also the less fortunate segments (through e.g. replacing subsidies with more juste social safety nets) will eventually improve also your own life and the lives of your children? - and if nothing else that it is good because it is good in itself? How to replace a "grab as much as you that you can" mentality with one more based on trust?

Tough questions for us all. In any case, I believe that edu-tainment could be of some help in shedding light on specific development issues and in changing attitudes and behaviors. (I am including a piece I wrote a while back "Fighting Poverty in the Arab World - with Soap Operas?" I touch briefly upon the possible role they could play in regards to subsidy reform.

Amina Semlali

Submitted by Shanta on

Thanks, Amina. Your earlier post has inspired some of us to work with Turkish soap opera writers to include information about regional integration in their scripts.

Submitted by Stuti on

Why the “why” has practical implications
Shanta, Thanks for this great post which has attracted insightful comments from key quarters on this perplexing question of what do we do when politics is the problem. Nachiket’s comment suggests that understanding the “why” can be more complex than putting it down to vested interests, in contexts of competitive politics such as in India. For example, I don’t think we understand why electricity subsidies are politically popular, even though “we” (as in technical experts who have examined the benefit incidence of subsidies) think it must have something to do with elite capture because the greater share of the benefit goes to large landowners. However, the elite suffer from lack of electricity (for which they then pay by using private generators), when bankrupt utilities cannot produce and distribute enough of it, and it’s the poor voters who are ready to protest at the first sign of repeal of subsidies. We under-invest in understanding why because it doesn’t stand our test of “how will we use this knowledge to give policy advice to a Minister”. So we continue giving policy advice that can’t be followed because of political constraints. Perhaps until the reform-minded leaders turn to us to ask how to overcome the political constraints, we won’t make a serious effort to understand the “why” and give technical policy advice on what to do about it. But then, I suppose, few think economists are in a position to give such advice. I hope we can change that opinion with the burgeoning work on understanding citizens’ political behavior and how it shapes the selection and incentives of leaders at the helm of multiple jurisdictions within government.

Submitted by Shanta on

Stuti, excellent points as usual. One question: why should we wait for someone to ask us for this information or research? If we think it's important, we should be doing it, and disseminating it, anyway. Our job is to do the analysis that will help end poverty. If that involves understanding why electricity subsidies persist, we should prioritize that analysis (over analysis that simply concludes that subsidies should be reformed).

Submitted by Philipp on

This is a great post, and I was just about to write a line about "great to see this, and even better that much of political science has already spent years and years thinking about cui bono type questions". But Joel beat me to the punch already.

I think even for technical advisors a lot can be gained by thinking about winners and losers of potential reform alternatives. Not necessarily because that way you'll find an "incentive-compatible", "politician proof" silver bullet policy that "works with the grain", but because it helps avoiding advice that is obviously silly and unrealistic.

More fundamentally, I wonder why of all things development policy is so wedded to the idea that there is a thing called a "policy" that exists independently of reality and is then soiled by "politics", i.e. those politicians and their schemes. Actual policies are not the Platonic ideal-type of how a government ought to do something, it's how things actually happen after all the players had their say - most of whom are probably street-level bureaucrats who don't even know what the "policy" is - and it all takes place in an institutional context that doesn't just implement the policy but defines what it is in the first place.

Submitted by Shanta on

Philipp, thanks for weighing in. No argument on your first point. Indeed, that is why I'm insisting that we ask the "why?" question first: to avoid unhelpful policy advice, such as "cut subsidies" or (dare I say it?) "increase public spending on education". As to the distinction between policy and politics, I think there is one (even though the French word--politiques--is the same for both). Strictly speaking, policies (or government interventions more generally) are justified on two grounds: (i) if there is a market failure (public goods or externalities); or (ii) to promote greater equity in society. Most policies exist because, at some point, they could be justified by one of these two rationales. The problem is that these policies have been captured by the non-poor (who have greater political power) so that they don't even achieve the market-failure-correction or equity-enhancement that they were originally intended to do. In order to reform policies to achieve their rightful purpose, we need to understand the political forces that are preventing them from doing so. That is why I would like to keep a distinction between policy and politics.

Submitted by Anand on

Shanta, Joel, Philipp, Stuti, et. al.

Thanks for stimulating a discussion on policies, politics and institutions and why we continue to do so poorly in integrating these into our approach to development.

Whether the problem that we should be addressing is the what, the how, or the why (and the when) is one way to frame the broad question of how the Bank can be more effective in its development impact. It is true that for much of our history we have functioned as though the logic of Pareto and Kaldor-Hicks economic efficiency was self-evident to all and advice on policy would overwhelm vested interests. That has clearly not worked. But we are, I would like to believe, an intelligent organization and should learn from the experience of failure to reform as well as failed reform initiatives. Anybody working on a country for any length of time should be able to comprehend that quite often (not always) the obstacles are political and that there are strong vested interests in the status quo. We should learn from such experience and consider the nature of political and institutional incentives and see how and if the equilibrium could be shifted towards a better development outcome.

So I would say the more interesting question for discussion to add to your list, is – the Why Not – why we do not change our own behavior in the face of failure to have greater development impact. That we do not attempt this more seriously is, frankly, a reflection of our comfort with the status quo – the incentives within the Bank rewards optimism bias because (a) it is easier to put together a project or program that ignores or underestimates the political obstacles and (b) we are not truly held to account for the wasted effort of a failed program or project.

If we were prepared to do better in this area, it would require a deeper understanding of the nature of institutions and institutional change, as Joel points out. Even with such knowledge, it is likely that only under some circumstances would we be in a position to influence positive change - economic development is too complex a process to be easily directed. So the question is whether we are able to recognize minor or major critical junctures and play our role intelligently in catalyzing change. At present, I think not. Our approach to institutional reform is often as blind to political and institutional realities as was our approach to policy advice. But it would be fascinating to change this paradigm and unleash some creative multi-disciplinary thinking

Submitted by Shanta on

Anand, thanks for raising the important question of development assistance. I think the source of the problem is that, to deliver financial assistance, we always require there to be a well-defined (and unique) program of how the money will be spent. That is how the receiving government is held accountable. But with institutional change, there is no unique, or even linear, relationship between inputs and outputs. It would be better if we could try a variety of interventions, each of which has some chance of bringing about change, but none of which is guaranteed to succeed. And our knowledge assistance could be even more experimental. We should be putting knowledge and evidence that is salient out there, and allowing a more evidence-based debate to take place, without necessarily knowing where the consensus will end up. I realize this is a far cry from how we currently do things, but let me suggest two, incremental changes that could point us in this direction. First, for knowledge products, instead of our standard requirement that there be a concept review (where the linear relationship between the product and outcomes is specified), we could have requirement that, in order to stop a knowledge activity, the manager needs to write a two-page note arguing why it should be stopped. This would allow many more risk-taking knowledge activities to flourish. Incidentally, this is the approach that innovative organizations like Google use. Secondly, for financial products such as a project, instead of specifying a single design, we would give a range of designs, to allow for different configurations of the politics of the project. This is what I wish we had done with the Delhi water board project mentioned above. Cheers, Shanta

Submitted by Prakash Kashwan on

Shanta - Thank you for a great post, which has prompted a very stimulating discussion. Here are my two cents related to the discussion above of the politics of it all. Examining the history of reforms that have led to pro-poor (that is, pro-majority) outcomes shows that we have to challenge the underlying model of the individual that is often employed in policy research, including some of the research that does talk about politics. Although the model of a 'boundedly rational' individual is well understood, the ‘bounds’ are often traced to 'culture', information processing abilities etc., and very rarely to politics, including the politics of defensive strategizing on the part of the citizens who are expected to assert their rights. Second, in line with the comment above by Philipp, the scope of (anticipated) policy drivers also have to be widened to include social and political mobilization. The ballot box model has proven to be a highly unreliable marker of accountability politics, to borrow from Jonathan Fox's work.

Submitted by Shanta on

Prakash, thanks for those comments. I don't think we need to deviate from the rational model in understanding the behavior of individuals as voters. The reason poor people vote for politicians who fail to provide basic public services such as health and education is that they don't believe the politician will deliver these services. Instead, they vote for politicians who promise targeted benefits, such as a road project where they can get jobs. As a result, the politicians doesn't provide quality public services, as is elected and re-elected. We have an equilibrium whee everyone is behaving rationally, but the equilibrium is a low-level one. The introduction of information about the quality of public services, and its link to politician behavior, may serve to disrupt this equilibrium and, over time, move the economy to a higher-level one.

Add new comment