Conversation with Shanta Devarajan, Chief VP, MENA, Economics, World Bank

[FitzFord]

Shanta, whose  condensed cv is a modest presentation of his career in Development Economics at Harvard University and the World Bank, will be with members of, and visitors to,  PFMBoard, will spend the coming Friday morning with us to discuss a crucial set of development issues that he has outlined in a brief note. Both the condensed cv and the discussion note will be posted as well. As usual, there will be exchanges of views, questions and answers that will continue until Thursday 26th. After that the conversation will be closed and remain available online.

Shanta's discussion topic: PFM and public service delivery

"Since the purpose of good public financial management is to improve the effectiveness of government, the discussion will focus at the other end, namely the delivery of public services and ask, how--if at all--PFM can improve their outcomes. First, I will question whether the activities that governments actually spend on are those that they should spend on. Using a simple, welfare-economics framework, I will show that most items in the government budget do not pass the test. In this context, strengthening PFM is like making the trains run faster...to the wrong station. Next, I will show that, even those government interventions that meet the welfare test often fail to deliver what they are supposed to because of (i) poor incentives in the service delivery system; and (ii) capture of public funds by political elites. In these cases, better PFM can help only if they change incentives in the same direction as is necessary to improve service delivery. Specifically, better monitoring of public expenditures can help if the information is shared with people who can influence decisions about service delivery. If it is shared with other government entities, it may increase their power over service beneficiaries and reinforce elite capture. Finally, better PFM for aid projects in an otherwise dysfunctional system rarely helps outcomes and may make matters worse, by increasing government's accountability to donors as opposed to its own citizens."

[SDevarajan]

To get the conversation going, I thought I'd elaborate on the first point I made in my abstract: That governments frequently spend on things that they shouldn't be spending on.  As a principle, governments should spend on things that the private sector, left to its own devices, will not (otherwise, government would simply be crowding out the private sector and wasting the public's money).  In economics jargon, these goods and services are called public goods or goods with externalities.  The classical examples include national defense and lighthouses (although, with GPS technology, the latter have become irrelevant).  But clearly governments spend on more than national defense.  Two of the biggest expenditure items are education and health.  In each of these, there are public-good or externality elements, although there are private goods as well.  For instance, in education, the main benefit from education is that the student earns a higher wage, which is a private good.  There may also be an externality, such as the fact that society as a whole benefits from having a literate and numerate population.  However, governments typically finance and provide all of education, rather than just the externality (which should strictly speaking be addressed with a subsidy to education).  Furthermore, this externality is highest at the primary level, lower at the secondary level and lowest at the tertiary level.  Yet government spending per student is highest at the tertiary level.  This is an example of government spending on private goods (university education) with scarce public resources.  There may be an equity argument for government spending on university education, but this would imply governments should subsidize or give free tuition to poor students, not all students.  In fact, most of the students attending free public universities come from the richest quintiles of the population.  In short, government spending on higher education (sometimes amounting to 1-2 percent of GDP) has very little justification.  What then is the benefit of having PFM systems that track public spending in education closely?  Are we perpetuating a wasteful system?  Let the discussion begin....

Shanta

[Napodano]

Hi, Shanta;

It is great that you start the conversation with a policy issue. One well-established criteria that discriminates a good PFM system from a mediocre one, is the capacity to link budget (and more in general public spending) to policy. Prime Minister Office, Ministry of Finance and the Cabinet collectively need to assess policy priorities, weigh spending options and expected results in and then have a budget that consolidate their policy choices.

To this end, I would like to pose the following question (Question #1):

What is your view on PFM reforms that promote performance-based budgeting in developing countries? After an initial enthusiastic start, only a few success cases have been recorded on a permanente basis. Was it premature for countries which do not have proper economic policies in the first place? Was the reform rejected by Governments with powerful vested interest?  Was too much emphasis on planning indicators a waste of Government attention without a monitoring system to track public service delivery?

[FitzFord]

Question #2

Shanta, will you elaborate on how might the politicians, who may not see the benefits of the system you propose, may be convinced of the superior results that are predicted, and be persuaded that the costs as they may be imagined or real, will be presented in a beneficial
manner in their meaningful time frame?

[SDevarajan]

Answer to question #1

Napodano:  Thanks for your question.  I think the limited success of performance-based budgeting has to do with many of the items you mentioned--poor policy environment, vested interests, excessive focus on planning indicators.  But there is a fourth reason that I've been concerned about.  Whenever we measure performance, we never specify the counterfactual: what would performance have been in the absence of the allocation?  Yet, unless we use the counterfactual, we create a host of problems for PBB.  For one thing, people will choose easy indicators to be measured against (effectively, indicators that would have been achieved anyway--hence the need to specify the counterfactual).  For another, performance would not be measured against what the private sector would have achieved--the point I made in the earlier post.  A third problem--which is also a problem with performance-based aid--is that there is an inherent risk in any development endeavor.  We don't know for sure whether we will achieve the result.  By making budgetary or aid allocations based on performance, we impose all the risk on the recipient, and none on the donor or budgetary authority.  Not a good way to run a railroad.  Shanta

[SDevarajan]

Answer to question #2

Fitz: In answer to your question, the only way to convince politicians (of anything) is to have them believe that a majority of their constituents will vote for the new system (or the results of the system).  This will require two things.  First, that voters are well-informed about the quality of service delivery, say, so they value this (as opposed to the politician's religious or ethnic affiliation) at the voting booth.  Secondly, that politicians know this, so they will start catering to voters along service-delivery lines, rather than by simply expecting them to vote their ethnic group.  Neither of these processes takes very long, especially these days with technology-based information sharing.  The process can be self-reinforcing.  For instance, if people start voting for better schools, and politicians deliver better schools, then the electorate becomes more literate, in which case they are even better informed (they read newspapers, etc.) and they start demanding even better services.  Besley and Burgess have documented this process for India.

Shanta

[FitzFord]

Question #3

Shanta,

There are implications in some national budgets and expenditures information systems (on computers, and accessible in terms of timeliness, clarity of information and application) that are available to the public and designed to facilitate the public's understanding of what is being spent and delivered. Until fairly recent, these systems were non-existent. However, there has been significant improvements in public finance systems that allow the public to be aware of programs and expenditure allocations within budgets that allow citizens to be aware of allocations and programs that are assigned to their localities.

Do you now to what extent the WB promotes the use of such programs and factor their results and the extent to which governments (i.e. politicians)  are systematically called upon to explain their performances (success and otherwise) and program planning and budgets for the future. This is not to put you or the Bank on the spot; it is essentially to get a useful notion of the effectiveness of what looks like attractive tools, and understanding their utilization and impact.

Fitz.

[John Short]

Question #4

Shanta,

Many thanks for spending time on the Board and initiating a discussion on a very important area of PFM.  I must admit I was concerned when I read your first post as I thought it may be too academic (at least for me anyway) and was mightily relieved to see your education example which is a great one.  There is role for the private sector but also cost recovery (and appropriate subsidy in certain cases to provide incentives) where there is public sector provision. 

Please allow me some comments to add to Mauro’s and Fitz’s questions.  I think it is important to distinguish between the policy-making-environment and the technical aspects of delivering policies (and changes in policies after an election). This latter requires development of procedures and processes and the legal basis to implement them as well as the building of capacity (initial and continued renewal).  Ideally they should not change if implemented well.  These do not happen overnight and usually take years (up to 10?).  Linking budgeting with the planning system has to take place in this context and often is not which often accounts for a disconnect between targets and the budget that allocates resources to realise them.  Often the introduction of programme budgeting fails because the preconditions that are needed –budget calendar, a multi-year fiscal framework etc.- are missing. The framework in Jack Diamond’s paper on sequencing (discussed on the Board) is rarely met.  I have been in a country recently where programme budgeting is being pushed by donors at both central and local government levels yet the basics are not in place.  In another, where the processes and procedures backed by a legal framework that worked is being dismantled to give a politician a job!  You mention having an electorate that is savvy – which is important but it is more important that we have a political class that is realistic.  There are many cases including the present in Europe that politicians get elected on impossible promises.  In the UK the issue of higher education student fees is becoming an election topic without the consequences of replacing the reductions in lost revenue to the universities being explained.

[Napodano]

Shanta,

You mentioned education, I want to add health:

Question #5

what do you think of Obama-care in terms of impact on the budget? How much is it 'exportable' to middle income countries?

Mauro

[John Short]

More on Question #5

Or UK National Health Service which is free at the point of use and funded from taxation - which is a dangerous no go political area!  There is also private provision though insurance schemes.

[Lewis Hawke]

Question #6

It is something of a straw man to pose a scenario where a dysfunctional aspect of the PFM system is made more dysfunctional by improving other very specific parts of the system, such as payment processes or internal reporting. A more interesting discussion for PFM practitioners is why the original dysfunctions exist and are not corrected or controlled.

The study of political economy is becoming a more important aspect of PFM, and Shanta's discussion points are great illustrations of how political economy influences can create and amplify system distortions. Good PFM analysis will take a whole system approach and identify the causes and consequences of inefficient or mis-allocation of resources. It will identify options for addressing them and analyse the likely net impact before proposing solutions. Political economy influences can mean that it is impractical or inefficient to treat the cause directly, for example the returns to populist policies that target swinging voters or the lower costs to protect the status quo than for disbursed and disorganised losers from a policy to act in their collective best interest. Good PFM analysis and policies will identify alternative options within the system that will yield net improvements. For example, this could be by strengthening access to information for NGOs representing wider interests,  or establishing institutions such as independent fiscal councils or non-partisan think tanks to highlight the full implications of inefficient or populist initiatives as a means to limit the incentives or galvanise opposition.

The existence of distortions are not arguments for maintaining poorly performing processes to keep the breaks on efficient resource allocation, but arguments for a systems approach and better targeting of the causes of problems, not taking a narrow emphasis on tweaking (or not) flow of funds and internal access to information.

[SDevarajan]

Reply to question #3

Fitz, the World Bank has been emphasizing transparency in public budgeting for years, using modern tools, such as open-data, whenever possible.  For example, we have a practice that we will not grant a budget-support loan to any country that doesn't publish its budget online.  We have also been spearheading citizen engagement in public spending decisions, including citizen feedback on our own projects.  That said, the evidence on the effectiveness of these programs is quite thin.  I think the main point is that, even when citizens know that public resources are being misspent, they have little recourse.  Until we can help with this second part, where citizens can sanction policymakers who are responsible for budgetary allocations, I'm not sure what progress we will see.

[orlandovv]

Question #7

Dear Shanta and PFM practitioners,

There are dozen of initiatives down here at the developing countries to improve service delivery through efficient PFM and Governance Systems. Some I'm personally involved with here in Central America are: the Open Government Initiative, Open Budget, Congressional Open Budget, and along with them PEFA.  My question is what overarching idea should link all these initiatives in a way most people can understand, demand, and support to truly make a lasting difference and have country ownership and donors harmonization?

Kind regards to all,

Orlando Valladares

[STONE]

Shanta, thanks for your time. Anyone of your academic and professional record (and a Kanbur index of 1) is especially welcome here.

It seems that ‘Reg the Taxi Driver’ (http://pfmboard.com/index.php?action=profile;u=1378 ) has a broken GPS (or has suffered from its battery shorting out in his motor-boat in a storm on the windward side of an erstwhile glowing public lighthouse good) and  ‘Sir Humphrey’s Ghost’ hasn’t walked here since his fit of apoplexy when Matt Andrews popped by to warm his slippers at the fireside, so it looks like you have me. (Old-time PFM Boarders know that I am just difficult, some use other, blunter terms). 

Obscure Board-cultural references aside... Question #8

Lewis Hawke, may have made this point better already, but is it not the case that public financial management, good, bad, or indifferent, is policy neutral? 

In a six-phase presentation of the PFM cycle, we journey-people of the (minor) Guild of PFM ‘Bricoleurs’ begin with ‘policy design and review’ and move swiftly on to the ‘techie’ bits of PFM, serenely uninterested (and, of course, disinterested) in policy content. If pressed to opine on whether or not food stamp/child benefit programmes should include non-poor families, do we not and should we not ask only if the policy is of high enough priority rank to justify the triumph of effectiveness of sustained (through subsidy-seeking votes of the non-poor) provision to poor families over short-term concerns of efficiency or economy?

[FitzFord]

Question #9

Shanta,

The Bank is promoting and supporting Citizen Engagement (CE) and have engaged with some academics as well as others in an effort to extend and strengthen CE in order to develop more involvement environment with politicians and their behavior and products, providing timely information and feedback for in planning and implementation AND adjustment of strategy and implementation. To what extent is that effort engaged with financial planning, managing and assessment? Or in the other way around?

Fitz.