Importance of statistics in governance

You were in academics for a long time. How is it different being in government?

Let me clarify that it is not that academics are not involved in governance. Most universities are large institutions in their own right. Academics by being virtue of running that institution learn a certain amount of governance. Having said that, principally academics is concerned with research or study areas of focus of his or her interests, whatever they may be. Governance on the other hand is a little more than that. A large part of governance is facilitating the work you may or may not be directly involved with. Let me give you an example. As an economist I have worked in a number of areas of economics — labour, policy and so on. But there may be other areas of economics in which I have not been necessarily or directly interested  in. But in governance you have to be involved with all of them. This is true of academics in general. Look at the problem of governance.
Governance is part of facilitating the work of a large number of technical people. Each technical person is interested in a narrow area of interest. This is true of this organisation. There are people who work like academics who labour in their narrow technical areas like national accounts, sample survey or industrial or agro statistics. There is government dimension of facilitating all their work in an integrated manner. Governance is partly about that. That is also the difference. Government is interested in lot more things than your mere specialisation.

Is it more challenging for you?

Yes very much. It is in the nature of things that you have to interact with a whole range of people. One has to understand what sort and nature of objectives and interests they have and the constraints they face.

You’re the chief statistician and secretary in the department of statistics. How would you explain the importance of statistics in the framework of overall governance?

Statistics plays a number of roles. We have to ask what are the questions that face policy makers. Let us suppose a policy maker says he wishes to do well for his people. So his first question is what should I do? In order to answer that question, he needs to know what’s the nature of the problems confronting his people? The second thing which he wants to know is, lets say,  if there is a drinking water problem. If he has started a programme he will come to us and want to know how he is performing? Fundamentally these are two elements which good governance should be concerned about at any point of time. The organisation of which I am secretary was started to enable an answer to both the issues. PC Mahalnobis was called by first Prime Minister of India Jawaharlal Nehru because India had become independent and there was huge burden of expectations from the people.

To determine directions of meeting such expectations, numbers are a must. The statistical edifice which was created was to help give you answers to those questions. The second part of it is that it must be on an on-going basis, as the sequence of number start coming, it becomes possible to answer whether I am able to meet the objectives I set for myself. So statistics work for good government in the same way your sensory organs work for a person to move or functions for a society.

You took over as the secretary in July 2010. Have you influenced the working in the department?

It is a very difficult question for me to answer. I suspect I see my role being played in two ways. One, I have been concerned about what is the speed with which policy makers can get information relevant to making decisions. One of things which we have been trying to do is to reduce the time lag in production of data. We have succeeded considerably by virtue of fact that in a number of data production enterprises, whether it is National Sample Survey (NSS) or the annual survey of industries, we have significantly reduced the time lag.

What was the time lag earlier?

Last year for the first time NSS results were released within in a year of data collection. Earlier it used to take two to three years. Similarly in the annual survey of industries we have managed to bring down the time lag in the production of results within nine months of collection of field work. But I should say something; these are not entirely my doing.

One of thing I realise at this level it is not possible for any one individual to do anything unless the entire organisation is committed to that task. It is not that these initiatives are new. A lot of these have been around for a while. The other initiative I can take credit for is pushing the organisation into thinking about production of new types of data.

What are these?

Some of these have not yet seen the light of the day. But work is going on. We are going to conduct a study which will help us get better estimates of women’ participation. We are also going to study ways which will give us better estimates of the service sector.

These things take time to bring to fruition. What happens is in organisation like this is you often end up taking credit for your predecessor’s initiatives and your successor will end up taking credit for yours. That is in the nature of things. The organisation has accepted these changes. Many other things like transparency, disclosures etc, I would take partly credit for. These are institutional features. A lot of these have been in place for some time.

What challenges does the department face in data collection?

There are two things about data collection. What is the source of your data? In official statistics there are two methods by which you obtain data. One type of data you get from official institutions. For instance hospitals are set up and they treat patients, they register, they document. You get data on diseases and problems faced by institutions. The quality of data will be as good as the institution itself providing the service. As a poor developing country, our institutions are still in the process of development.

So even a major source can be viewed with a considerable degree of suspicion? They may feel that the particular institution is not of requisite quality. Sometimes we get the flak for the data from this source. Flak is not for us, but it is the generic weakness we have because of weak public institutions. The other way to get around this is to directly  gather data going to people and conduct surveys and census. But we have to keep in mind they are enormously labour intensive. In order to implement them, you will need somebody who will go and ask questions and to conduct a census, you will need that many more people and for surveys it can be less. There are issues of design and coverage and so on. Not all questions can be gathered by surveys.

Our challenge is we run amongst not just the developing countries but in a world which consists of well accepted survey systems. The demand for information on a range of issues-living standards, sanitary conditions, what we are eating etc, are done by us. There are few surveys globally which meet this kind of scale and challenge. But still, given the expectations, we are invariably falling behind. One difficulty with statistics in India, as with other developing countries, is you are collecting data in systems where governance structures are weak. And, your quality of data is only as good as the structure of the governance through which data is collected.

What governance structures are you talking about?

Our ability to track which industries are producing what goods, to know how much they are producing, to have a system to be able to go back and check with the industry how much production took place. If you notice, these are not statistics people (that collect data). (They) only compile data, which are collected from various sources. A lot of this data is reported under administrative arrangements. Those administrative set ups and their ability to track the sectors also need to be looked at.

Can you tell us how a survey works?

Next year we are going to conduct a survey on total sanitation and drinking water both in the urban and rural sectors across the country. In order to carry out this assessment at the request of few ministries which wanted additional information, we started work for this around July-August 2011. We constituted a committee headed by an eminent academic who knows this field well with representatives from civil society and other departments. This group had the mandate to devise survey instruments for the purpose and design of the survey and identification of slums. This group prepared a report which culminated in an exercise last month where we called senior officers from field offices and  representatives of state governments (as it is done with collaboration with states) to come for a meeting in Delhi. Local issues are raised. Surveys instruments and designs are finalised. These are taken by NSSO offices and trainings are conducted. That will start from May. Survey work will start from July this year and will take six months. Data processing may take six months to a year, depending upon of how good it is. Also data validation takes some time to plug inconsistencies in the information collected. To the question how many people are involved depends on the survey. Typically it is conducted in 40 regional offices and then there are sub offices which are close to 200.

There are challenges for the chief statistician of India.

There are a number of areas in which data collection currently is inadequate. India has been growing rapidly during the last few years. Some growth has taken place in the service sector. We do not have as yet an efficient system of tracking nature of growth and their service to the Indian economy. One of things we are plan to launch is a regular survey of service enterprises. This is extremely challenging because unlike the manufacturing sector for which surveys have been conducted for the last 60 years, the service sector is very heterogeneous. Designing survey instruments which can be used to canvas information about service enterprises is challenging. Some data is available from the ministry of corporate affairs. We want other characteristics like how many people they employ, what types of input they use, management practises, technology. Other challenges: a large population in the country is engaged in self employment. It becomes useful to understand how they spend time. So we are conceptualising time-use surveys.

Where have been these pilot surveys done?

It was done during the late 1990’s. Unfortunately, it could not progress further as there was need to standardise classification systems to be used. This type of information is extremely useful. There has been demands from the academia and government asking us whether we can do this again. What we have done is to start a group working on the experience of pilot projects and international experience and come out on what it takes to do a survey in India. Then we will raise funds for it.

As there are lot of expectations and demands for different surveys, how do you handle human resource constraints?

It is a major challenge. We have been thinking on how best to innovate in different ways. One innovation has already been done in the ministry. We are committed to releasing a new product consumer price index. The difference between this CPI and the ones done in the past is that we have included specific groups - agricultural labour and industrial workers. This one will be an aggregate price index and all India in character and represents all types of consumers, self-employed, casual labour etc, and capture different types of market. This was a huge challenge. In order to gather this data, samples were collected from village, blocks and towns, something never done in the past. So what we did was to enter into a partnership with the department of post. A post man goes to every lane and every village. This was a fairly major exercise involving this ministry and the department of post was undertaken to identify the postal structures which could be used and train them for data collection. This has been released for the last year. So it has been successful in one way. But we continue to face challenges. Postmen are not trained to work the way our investigators do. We also have to work out innovative partnerships of this type with other institutions depending on the subjects. Prevailing prices are the easiest for the post man to gather. Most consumer goods can be explained to a post man. There may be other subjects we have to tie up with other institutions.

Which are these subjects and what other partnerships do you foresee in coming months?

For the time use survey, we are exploring use of network of women study centres created by the UGC. There are 30 to 40 such study centres across the country. Some of the survey will be done by the department while others  could be piggybacked on centres and networks elsewhere. When we collect data from villages, we use Anganwadi workers. So we have to engage in strategic tie-ups with other institutions who can compliment our work. All these have to be specially tailored. It is imperative to explore other partnerships.

Inflation numbers were a source of constant headaches for policy makers. Will there be some relief this year?

That will be hard to say. Now you are not talking to me as a statistician  but as an economist. Inflation numbers indicate a few things. They indicate supply and demand. What has happened is our public policy is creating demand in certain types of areas. Let me hypothetically take the example that we are pursuing an inclusive policy and let us also assume we are successful in that policy. So purchasing power does reach the hands of people, especially at levels earlier not reached. They will demand goods and services. But the supply of goods and services is not automatically forthcoming. So you will see a rise in their prices. Secondly, as the economy grows people will demand higher value goods. If you look at items of household consumption, people are consuming milk, dairy products, pulses, meat and eggs more than was the pattern 30 years ago. This trend is going to  continue. Our challenge is to meet the increased demand. If we cannot meet this increased demand you will see a sharp rise in prices. How do you manage is a challenge for policy makers and the establishment. There are two approaches — long run is the investment in infrastructure and short run is through a better understanding of areas where demands are going to be raised and putting in place effective means of ensuring supply.

Dr C Rangarajan has made a number of recommendations to streamline statistics collection. How much of it has been implemented?

The recommendations have been made for the statistical system as a whole. The Indian statistical system is a federal system. It is structured in the following way — there is a central statistical system and a state statistical system. In the central statistical system we are the nodal ministry and there are establishments in a large number of ministries. What I can say is that what pertains to this ministry has been implemented. A large chunk of what pertains to other central ministries has been implemented but not all. For the states, a number of other ideas are being worked out.

What about streamlining IIP data?

We are in the process of revising the base year for IIP. This time, we may constitute a special working group for revision of the IIP, as was done in the WPI (wholesale price index). We have already done some internal work from our standing committee on methodological improvement.
There is also a report of Unido and the United Nations Statistical Commission on revised standards for industrial statistics, specifically for IIP. In the last five years, we have put in a number of mechanisms to ensure the ASI (Annual Survey of Industries) results come out more quickly. So, we will examine if it is more feasible to put into place other administrative arrangements to improve the (IIP) index.

What about volatility in IIP figures?

The problem of data quality is more substantial because it is a function of a variety of things-of the system through which it is collected, the agencies from where it is collected, the resources available, and the mechanics that are followed while reporting. This needs to be constantly reviewed and assessed, which is a separate exercise. When an index is prepared, it has certain limitations. If data collection is through official channels, you can examine the mechanics by which it is be collected and seek to improve their internal collection mechanisms. If data collection is through sampling exercises, an effort can be made to reduce sampling errors, both of which are currently under way.

You had ordered an audit into IIP figures?
That was for the old series. The audit reported a number of structural aspects of the IIP production which has been discussed in the NSC. Some of the interim conclusions noted as the base revision exercise of the new series. The new series has come out with base of 2004-05. But we would like the base year to be revised as it is already six-seven years since 2004-05.

So when do you plan to revise the base year from 2004-05?

We have already started the base revision exercise for IIP. As part of the exercise, all information and knowledge from that audit process and from the other criticism that continues to come up as people used against the IIP series, will be taken on board to see how to improve the IIP data. An expert committee set up will answer a lot of questions. There are a number of possible answers they can choose from. They can choose a year and give a different method. They are indices which do not have a fixed base, in which base is automatically revised either every year or periodically following a set of criteria. Those possibilities exist. First is the fixed base index and second is chain base index. These are options which are available for the committee to examine in light of institutional and administrative feasibility, whatever is doable. They can ask for a more dynamic base updating process.

Isn’t the department of statistics the ‘eyes and ears’ of the government as it looks at every section of life?

Very much. This why it is difficult to separate out work of statistics from the work of the government. Statistics happen because government works.

How far does the quality of poor data or statistics affect the working of government or decision making process?

There are two things when we say poor quality of statistics. One is, does the data have systemic flaws in it? Second, the data may not have systemic flaws but processing errors. Addressing the issue of systematic flaws requires focusing on what sort of problems are leading to such flaws — is there improper understanding about information being collected, is the organisational design in which the information being collected defective? These are questions statisticians need to answer.

Should more experts be brought into governance?

Experts bring in a certain approach which is different from government specialists. Experts have different strengths in terms of technical areas. Governance requires both. Governance is about experts and generalists and they can talk to each other and plan policies.