Saturday, April 12, 2014

Leadership of State agencies in India requires engineers and not drivers

I recently wrote a column in the Economic Times titled Engineers, not drivers, about the problems of malfunctioning government agencies in India and the problem of constructing State capacity. 

Thursday, April 03, 2014

Opportunities at Macro/Finance Group, NIPFP

The Macro/Finance Group at NIPFP has openings in analytical economics and public policy work.


We require individuals with a Ph.D. in economics or finance, with an interest in original research in our fields. These would be contractual appointments for a period of two years. One or more publications in international journals would be helpful, as would be the ability to carry research from inception all the way to publication.

Policy research associates

We require people who can participate in large complex projects in the field of public policy, ranging from envisioning the future to implementing it.

Policy work is highly inter-disciplinary, spanning the fields of public economics, finance, economics, law, and public administration. Deep finance practitioner knowledge is important. Equally important is looking beyond the present and envisioning the future, with first principles reasoning that is informed by international experience. We welcome interest in these positions by people with strong capabilities in some of these areas, and curiosity about the others. The ideal candidates would have read the Percy Mistry, Raghuram Rajan and FSLRC reports, and have familiarity with the things being talked about in this blog.

Quantitative research associates

We require individuals with a Masters degree in economics, finance, public policy or statistics. The work involves participating in academic research projects in the fields of macroeconomics and finance, and practical macroeconomic policy analysis. Desirable features include: domain knowledge; knowledge of computer programming, ideally in R; experience with CMIE databases and datastream. Some of the people who have done this for a couple of years have gone on to do Ph.D. in economics or finance at top schools.

Research programmers

We require a senior person who would play a dual role. On one hand, he would be responsible for an existing system which includes linux desktops, linux servers, and a project management system based on redmine and svn. This is expected to take up roughly 20% of effort. The prime focus will be participating in development work of complex analysis of economic and financial data. This development work is primarily in R. It includes building internal tools and also packages released as open source, such as our R packages fxregime, eventstudies.

Such work could be particularly appropriate for a person who is at present a computer engineer but desires knowledge of economics and finance.

Generic features

The Macro/Finance Group at NIPFP is a conducive research environment including a modern office. Compensation is generally superior to that seen in government academic institutions. The policy and the quantitative teams are strongly interconnected with significant spillovers of knowledge. We are a research environment: non-hierarchical, low politics, high IQ, high intensity and high involvement.

Please look up our websites:

If this interests you, please contact Anurodh Sharma (anurodh54 at with your resume by 20 April 2014, where you clearly identify where your interests and capabilities lie.

Thursday, March 20, 2014

Financial reforms : What should the next government do?

Think India Foundation is setting up a series of TV shows on various aspects of public policy. Percy Mistry, Rashesh Shah, Dhiraj Nayyar and I were on their Financial Policy show, which was shot in English and Hinglish. I was impressed at the quality of the show; I had not thought this possible on television. And, my new ambition is to speak in Hindi like Dhiraj does.

It will air on the coming weekend (22 and 23 March, Saturday and Sunday), as follows:

Sat – 12pm & 8pm
Sun – 10pm

Sat – 10am & 8pm
Sun – 11am & 6pm

Sat – 7:30pm
Sun -1pm

Sat – 8:30pm
Sun – 8pm

ETV UP, ETV MP and ETV Bihar
Sat – 6pm
Sun – 5pm

ETV Rajasthan and ETV Urdu
Sat – 5pm
Sun – 5pm

Monday, March 17, 2014

CBI preliminary enquiry on MCX-SX becoming a stock exchange

This is a good time to ruminate on the notion that CBI or Lok Pal should have independence. Underdevelopment is where the police are more dangerous than the criminals. Coercive power coupled with independence is a recipe for trouble. What is needed is far more thinking on how to do sound public administration, on the right mix of independence and accountability. Anger about corruption is not a useful source of good thinking in public policy.

This is also a good time to ruminate on the lack of rule of law in the determination of fit & proper.

Human capital will fuel high Indian GDP growth for the coming 40 years

In a recent column in the Economic Times -- My generation -- I argue that the outlook on human capital in India is bountiful for the coming 40 years.

Friday, March 07, 2014

12th research meeting of the NIPFP-DEA Research Program (correction)

The 12th Research Meeting of the NIPFP-DEA Research Program will take place at the India Habitat Centre in Delhi on 13 and 14 March.

The program design is up on the website.

The previous post which stated this had a mistake on the URL. Please ignore it.

Wednesday, February 12, 2014

The Bombay police: A failure story

A few days ago, I woke up at 4 AM under an onslaught of the shouting of large beefy men instructing a crew, using megaphones, in a movie shoot in Film City. I thought to myself "this must violate some law".

For a while, I tried to be a free rider, thinking "Someone else will complain". But it was 4 AM and clearly nobody had complained. So I thought I should call the police and complain.

  1. I used my (Airtel) cell phone and dialled "100". The cell phone said this was an imaginary number. I rotated the phone by 90 degrees but this also did not work.
  2. I tried to dial "022-100" but this gave the same error.
  3. Some websites said that 112 is an omnibus emergency number. I tried 112 and 022-112. Neither worked.
  4. I broke into cold sweat thinking that in Bombay, I actually have no means to call the police in an emergency using my cell phone.
  5. I hunted for other ways to reach the police. There is no rapid access mechanism using new technology: You cannot send in a complaint by email or IM. You can chat with an Amazon customer support person by IM or on email, but you can't do this with the Bombay police.
  6. I started hunting for a phone number for the police on the web. There were large numbers of websites. It was not clear which to use.
  7. If you google for "Mumbai police" and click on the first link -- -- it takes you to a fashion store.
  8. I hunted more on the web and got hold of a few numbers and started trying. The first two shooed me away.
  9. The third one was willing to listen to me, but not in English. The only languages that he would speak were Hindi and Marathi. I happen to know some Hindi and some Marathi, but a large number of migrants to Bombay speak neither. It is not good to have government interface that does not grok the lingua franca of India.
  10. He heard me and said "okay". He did not say "This is illegal and we will shut it down". 
  11. He did not give me a ticket number. He made no attempt to take my phone number or email address. Nobody contacted me in the end to tell me what was the disposition of my complaint. I had no idea what happened. The noise blared on.
And I understood why free riding did not work. Nobody had complained because there is no mechanism through which anyone can complain.

There is no public good as fundamental as the criminal justice system, and we in India are simply not trying hard enough.

Tuesday, February 11, 2014

Difficulties with Special Guidance and General Guidance in the IFC

by Anirudh Burman, Pratik Datta, Suyash Rai, Arjun Rajagopal, Shubho Roy.

Administrative and regulatory agencies in India currently use multiple legal instruments to regulate. This creates multiple problems for those regulated:

  1. Identifying all applicable subordinate legislation;
  2. Reconciling sometimes overlapping and inconsistent subordinate legislation; and
  3. Distinguishing between binding and non-binding statements/regulations.

For example, a person investing through the FDI route has to reconcile the applicable FEMA regulations, FDI Press Notes and the Consolidated FDI Policy. All of these have been communicated via legal instruments, but there is uncertainty over which instrument is operative in any given situation. This increases compliance costs and legal uncertainty.

The draft Indian Financial Code (IFC) proposed by the Financial Sector Legislative Reforms Commission introduces a completely new way of thinking about regulatory governance in India. Perhaps the most critical component of this framework of regulatory governance is the proposal that regulators will regulate only through ``regulations''. The use of a single instrument to regulate (instead of the present plethora of circulars, orders, letters, instructions etc.) will ensure legal clarity for regulated entities and therefore facilitate a rule-of-law framework in financial regulation.

However, two provisions in the IFC have the potential to upset this carefully constructed framework. Section 56 (General Guidance) allows regulators to issue ``general guidances'' with respect to the operation of the IFC and regulations made under it, the operation and objectives of the regulator, and other matters on which the regulator wants to provide information or advice. Section 57 (Special Guidance) allows any person to ask for a ``special guidance'' with regard to transactions or activities governed by the IFC. The provision on special guidances is similar to the existing system of advance rulings under the Income Tax Act. It should be noted that a need for a system of advance rulings under the Income Tax Act exists because of the lack of clarity of our taxation regime. If the U.S. experience with such guidances is any indication, both these provisions have huge potential to dilute regulatory certainty.

U.S. experience with guidances

The US Administrative Procedures Act contains uniform administrative procedures to be followed by all federal administrative agencies. Section 553 of the APA requires a public notice and comment procedure to be followed whenever an agency makes a ``rule''. However, this requirement does not apply when an agency issues ``interpretative rules, general statements of policy, or rules of agency organization, procedure, or practice''. This exception has been used by agencies to regulate through interpretative rules and policy statements. (For more reading on the subject, see: Anthony (1992); Recommendations of the Administrative Conference Regarding Administrative Practice and Procedure (1992); and Rakoff (2000)).

As interpretative rules and statements of policy do not require a public notice and comment procedure to be followed, an agency finds it less costly to regulate through these instruments than to frame rules. As an example, the number of US Food and Drug Administration regulations adopted in accordance with the APA reduced 50 percent in mid-90s compared to early 1980's (Rakoff (2000)). Anthony (1992) and Rakoff (2000) point out that interpretative rules and statements of policy become de facto binding for the following reasons:

  1. The staff of the agency starts relying on the interpretative rule or statement of policy in its work. This happens even though some agencies explicitly state that interpretative rules are not binding statements.
  2. The language of the interpretative rule or statement of policy indicates that the agency will follow it.
  3. The language of the interpretative rule or statement of policy indicates that it is of a binding nature, or will be regularly applied against regulated entities.

U.S. Courts have tried to counter this development by stating that: a general rule, an agency can declare its understanding of what a statute requires without providing notice and comment, but an agency cannot go beyond the text of a statute and exercise its delegated powers without first providing adequate notice and comment. (Fertilizer Institute v. EPA F.2d 1303),

Courts however, do not provide a good solution to this problem for two reasons:

  1. Judgements are case-specific, and it is difficult to enunciate a broad principle that substantially whittles away agency discretion, and
  2. In many instances, it is difficult to determine which exact parts of an interpretative statement or statement of policy are in effect, a disguised regulation, and those that are not.

One alternate method of ensuring some restraint on regulation through interpretative rules (and that also functions as a feedback loop) is that of ``petitions''. Section 551 of the APA defines a ``rule'' to include interpretative statements and statements of policy. Section 553 allows any person to petition for the issuance, amendment or repeal of a rule. This provision has been used to petition for changes to the interpretative rules. The US Supreme Court has held that agencies can deny a request made in a petition only on reasons grounded in statutory law.

What should we change in the IFC?

The U.S. experience clearly shows the pitfalls of giving agencies powers to ``explain'' their objectives and operations. Granting agencies the power to issue communications without following a defined legal process opens the door to chaos in regulatory governance. It incentivizes over-regulation (as regulation without notice and comment is cheap) and leads to a breakdown of the rule of law. The power to request and grant special guidances only exacerbates this problem. It is therefore important that a regulator's power to issue guidances be restricted.

To achieve this, certain changes need to be made to the IFC.

Remove Section 57 on special guidances

The U.S. experience with respect to ``general'' interpretative rules highlights how giving agencies an option to avoid regulation making in fact leads to regulatory chaos. This problem is only going to be exacerbated with a system of special guidances.

Every time a regulator issues a special guidance it interprets the law and applies it to the facts present in the application for guidance. This interpretation will, in many if not most cases, be taken as evidence of how a regulator applies the laws to specific facts. In a good rule of law system, any such interpretation should then be applicable generally to all regulated entities. The lack of general applicability creates legal inconsistency and reduces clarity. This is bad for any system of regulatory governance.

Restrict the use and content of general guidances

The purpose of guidances is to aid regulators in explaining regulations to regulated entities and consumers. They should not become substitutes for regulation-making. In an ideal world, there would be no need for a guide to regulations as they would be self-explanatory. Guidances should therefore function as ``regulatory guides''.

At present, Section 56 of the IFC provides that general guidances can be used for the following purposes:

  1. The operation of this Act and any regulations made under it;
  2. Any matters relating to functions of the Financial Agency;
  3. Meeting the objectives of the Financial Agency; or
  4. Any other matter about which the Financial Agency finds it appropriate to provide

All of the above provisions leave scope for a regulator to ``legislate'' regulations through general guidances, in the same way as agencies in the U.S. have done. Unlike the practice mandated under the U.S. APA, financial regulators under the IFC will have to go through a notice-comment procedure even for general guidances. However, they do not have to prepare a cost-benefit analysis of the same. Therefore, ``regulation'' through guidances is still a less costly proposition than framing regulations. Moreover, guidances in India are likely to become binding in India in the same ways as interpretative rules and statements of policy in the U.S:

  1. The staff of the regulator starts relying on the guidance and treats it as binding;
  2. The language may indicate to regulated entities that the regulator will implement the guidance; and
  3. The language indicates that it is of a binding nature, or will be regularly applied against regulated entities.

Guidances should therefore be used only as regulatory guides i.e. they should only aggregate and present existing regulations and publicly available orders of regulators for the benefit of regulated entities. For example, regulators may prepare a regulatory guide aggregating and summarizing all applicable regulations and approval/decision orders for say, a person who wishes to get an approval for starting a mutual fund business. For doing this, the following legal framework is proposed:

  1. Guidances/regulatory guides should be published by regulators for the sole purpose of enhancing the readability of regulations to regulated entities or specific classes of regulated entities.
  2. Guidances/regulatory guides may contain clearly distinguishable additional text only for the following purposes:
    1. For introducing the subject of the guidance and the relevant regulations/orders; and
    2. For connecting portions or extracts of orders to enhance the document's readability.
  3. The guidance/regulatory guide should contain a clear disclaimer stating the following:
    1. Any additional text contained in the guidance/regulatory guide does not constitute legal advice and is not binding on regulated entities.
    2. The regulator will not rely on the additional text in the guidance/regulatory guide in dealing with regulated entities.

This will:

  1. Prevent regulators from accidentally or deliberately adding ``binding'' content in guidances, and
  2. Consequently create greater incentives for regulators to write high quality regulations that are clear and comprehensive.

It will also force regulators to think of how to develop world-class websites where different regulations, orders, and even relevant FSAT judgements can be presented to a user seamlessly. A good example of a financial regulator's website performing this function is the page on ``Regulations guidance and licensing'' of the Monetary Authority of Singapore (link).

Create a system of review for `disguised regulation'

There should be a review mechanism to prevent against regulators accidentally or deliberately adding binding content to guidances. In the U.S., courts have declared invalid interpretative rules that are ``legislative'' in nature and bind regulated entities. A similar power to review guidances should be given to the FSAT. The FSAT should be able to direct the regulator or directly redact portions of guidances if:

  1. The communication is couched in mandatory language;
  2. The communication indicates that the regulator will treat the communication as binding in its interaction with regulated persons;
  3. The language of the communication strongly evidences a binding intent; or,
  4. Regulated persons are led to believe that adverse consequences will follow in case the communication is not complied with.

Create a framework for petitioning regulators for changes in regulations and guidances

In the U.S. petitions have served as an instrument to force agencies to review their rules, interpretative rules and policy statements. We envision a similar role for petitions under the IFC. Additionally, a framework for petitioning regulators can create a legitimate and transparent feedback mechanism through which regulated entities and consumers can talk to regulators. While in the U.S., courts review the denial of petitions, we envision a system of petitions that is softer in nature and focuses on ensuring a functioning communication mechanism between regulators, consumers and regulated entities.

Any person should be allowed to petition for the issuance, modification or repeal of a regulation or guidance. Regulators should allow for petitions to be submitted through their website and on paper. All petitions should be made available on the regulator's website, and the regulator should respond to every petition on its website in a time-bound manner. The regulator may agree to the request made in the petition, or deny it. However, the denial should be on limited, legally defined grounds:

  1. The petition requires the regulator to do something it has no legal power to do;
  2. The petition does not require the issuance, amendment or repeal of a regulation;
  3. There is no way, in spite of reasonable efforts being made, to verify factual claims made in the petition that form the basis of the request to issue, amend or repeal a regulation;
  4. The regulation to be amended or repealed, or the subject matter of a new regulation is under review before the FSAT or any other court; or
  5. The regulator agrees with the substance of the petition, but does not have the resources to implement the petition. In case the regulator denies the petition on this ground, the regulator must state when it proposes to act on the petition.