## Sunday, December 03, 2006

### Trickle down economics

Two pet hypotheses in India are as follows:

• That economic growth is not reaching "poor people"
• That government is central to improvements in education and health; conversely, that the only way to have educated and healthy people is to have a government spend more and/or spend it better.

Different people share these `pet theories' to different extents. Many liberals are skeptical about the usefulness of more spending by the government on education and health as long as spending consists of a mere intensification of existing programs. This is because there is now ample evidence that existing programs work pretty badly, as is borne out by evidence of absenteeism by health/education workers, the poor learning accomplishments associated with Sarva Shiksha Abhiyaan, etc. Many liberals would support a view that more government spending would help - or is essential for obtaining - better health and education outcomes, as long as fundamental surgery is made to the way in which programs work - e.g. shifting from teachers as civil servants to vouchers.

In recent weeks, fascinating new information about household level health and fertility has come out through NFHS-3. In terms of methodology, NFHS is the best household survey in India - it's the gold standard against which all surveys compare themselves. The first survey was conducted in 1992-93, the second in 1998-99 and the third in 2005-06. Every other survey aspires to NFHS quality resources and methodology. If you are one of those who loves to hate the NSS dataset, you needn't carry this skepticism on to the NFHS: it is the first world class household survey in India.

Kamla Gupta, Sulabha Parasuraman, P. Arokiasamy, S. K. Singh and H. Lhungdim have an article in the EPW of 21 October 2006 titled Preliminary Findings from the Third National Family Health Survey which shows some first findings from NFHS-III for five states (Chhattisgarh, Gujarat, Maharashtra, Orissa and Punjab) [pdf]. The quick summary is: the public health service delivery was terrible, but the health and fertility outcomes got tremendously better. Ila Patnaik has a great article in Indian Express summarising and interpreting their results. In this, she says:

The percentage of infants dying before they attain the age of one has dropped significantly in the last seven years in all five states. It has reduced in Punjab by 26 percent, in Gujarat by 21 percent, in Orissa by 20 percent. When compared to the data from NFHS-1 carried out in 1992-93, Orissa, one of the poorest states in India in terms of per capita income, has witnessed a decline in infant mortality by 40 percent.

For many years, the sense in India was that Kerala and Tamil Nadu had achieved replacement-level fertility (2 children per woman) but fertility in the rest of the country remained stubbornly high. The NFHS findings indicate that over the past 13 years, significant progress in fertility has taken place in all five states. Punjab and Maharashtra have achieved replacement fertility. Women in Orissa, Chattisgarh and Gujarat now average 2.5 children each. These trends in fertility indicate that India will reach replacement level fertility in 2010.

Preliminary evidence shows that the quality of public health services has been worsening. As NFHS data shows, immunisation, which is largely done by the government, has worsened in Gujarat, Punjab and Maharashtra in recent years. This data is in conformity with the data from the report on Reproductive and Child Health Program of the World Bank which found that out of 274 districts in the country, child immunisation declined in 197 districts.

Similarly, indicators of maternal health from the NFHS data show that while antenatal care is now universal in all five states, only 55-75 percent of women are getting the recommended three antenatal visits. Moreover, the report on Reproductive and Child Health Program found that the increase in in-hospital childbirth is caused by a rise in in-hospital births in private hospitals. There has been a decline in in-hospital births in public hospitals. The data on antenatal care and assisted deliveries showed that the pecentage of deliveries assisted by health workers went up from 39.6 percent in 1998-99 to 47.5 in 2002-03, the percentage of women delivering in public health facilities declined from 24 percent to 18.5 percent. The increase took place in deliveries in the private sector, where they rose steeply from 9.4 percent to 21.5 percent.

Moreover, women in richer states were seen to be using public health facilities less and turning to private health. In Andhra Pradesh, the percentage of women delivering in public institutions declined by 9.8 percent, in Kerala by 28.9 percent, in Karnataka by 10.2 percent, in Maharashtra by 9.1 percent and in Tamil Nadu by 15.3 percent. Further, the number of women who received post natal care by public health workers (ANM) through home visits within 2 weeks of delivery also declined from 14.1 percent to 12.7 percent. The Planning Commission's midterm appraisal of the 10th Plan observed that when people first seek treatment, an estimated 70-85 percent visit a private sector provider for their health care needs.

This result flies in the face of the two pet hypotheses cited above. If you believe that the lot of poor people is not improving, then this evidence is inconsistent with this position, because it suggests that poor people were a lot better off in 2005-06 when compared with 1998-99. If you believe that government spending and/or program design is important, then this evidence is inconsistent with this position: the NFHS (and other sources of evidence) show that the government system did badly, but that people got healthier and had fewer kids anyway.

What is going on? I think the main insight is that the health of the people reflects lots of things. It reflects nutrition, sanitation, knowledge, private purchases of health services and the outcomes delivered by the public health system. It is by no means controlled exclusively by the public health system; when people talk about improvements to the public health system as the only channel to having a healthier population, this is flat wrong. When people get richer, they buy better food, better sanitation and cleanliness, more knowledge (e.g. education within the family), and services of private doctors / hospitals. India has been experiencing powerful economic growth, which is trickling down to poor people. So even though the public health system is doing badly, health outcomes have improved, amongst poor people.

What I'm saying is not at all surprising when you think in terms of common sense. When people get richer, they have fewer children, buy more soap, buy more vegetables, are more likely to go to a private doctor when faced with a health problem, and are more likely to buy education services thus inducing more knowledge within the household. This makes them more healthy. But this common sense flies against the orthodoxy in India, which equates "the health of the public" with the spending on and the design of "public health programs", and assumes that poor people are not sharing in economic growth. NFHS-3 suggests that if you completely froze the spending on the Ministry of Health in nominal terms, and just had high GDP growth, you would most likely continue to get strong improvements in health outcomes.

Such an understanding - where health outcomes are not equated to the public health system - is consistent with the history of health in Europe, where a great deal of improvements in health took place owing to rising incomes feeding into nutrition, cleanliness and private purchases of health services. Such an understanding is also consistent with analysis of NFHS-1 and NFHS-2, where the basic story which emerges is that the presence of a Primary Health Centre does nothing for health.

1. NFHS3 survey, is it just about poor persons or all households? If it is about all households, perhaps your conclusions about its implicaitons for the poor households need to be checked?

2. NFHS is about all households. It's a random sample of what India's households look like.

3. Wouldn't it be interesting to do a sub-analysis focusing on households with a defined limited purchasing power, and compare their across the decades (adjusting for inflation) ?

4. *compare their health

5. When you think of issues like infant mortality or fertility, the overall average makes a lot of sense. Rich people anyway never have high infant mortality. So the overall infant mortality for the population ends up reflecting (a) the poverty rate and (b) infant mortality amongst the poor. Similar issues influence fertility.

NFHS-3 is very recent. This EPW article is the first time any data has come out. As my post suggests, even the broad aggregated data is rich with counter-intuitive results, and improves our knowledge considerably.

6. Sir,
I'm sure there is a lot of improvement in the health and fertility but to the bigger question of whether the economic growth is reaching poor people in the right proportion.I'm not sure..I guess if there is a way to compare urban and rural progress in this aspect,it would be helpful.Don't think you think teh government plays an important role in relaying this growth to the poor?

7. No, I think infant mortality is a great way to measure the genuine quality of life of poor people.

Think about it: the upper class does not have children dying at a young age, barring levels found in OECD countries. Whatever deaths you are seeing beyond 15 per 1000 are deaths amongst poor people.

As a thumb-rule, any infant mortality rate above 65 per 1000 is really unacceptable - the most basic access to food, cleanliness and health services gives numbers better than 65. What we have seen in India is clearly a case where the observed infant mortality is a consequence of poverty & deprivation. Ergo, any improvements in infant mortality that are taking place are caused by improvements in food, cleanliness and health services accessed by poor people.

The non-obvious step here is to notice that GDP growth can (possibly) do 4 things. It can yield better nutrition for poor people; it can yield better cleanliness for poor people; it can empower poor people to buy better health services; it can involve better production of health services in the public sector.

In India, we know the last one did not happen. But we know that health & fertility got better. Therefore the first three elements were at work: poor people got richer and were buying more vegetables, soap and private medical services. Better GDP to better health can take place with no help whatsoever on the part of the public health system. Conversely, if you freeze the nominal spending of the Ministry of Health, these improvements will continue to obtain.

8. Excellent post Ajay. It's rising per capita wealth that drives improvements in a wide range of human conditions. And because the free exchange of goods and services provides the best environment for economic growth, it's logical that this would also be the case for health related matters.

9. One would hope after looking at this study, UPA would come to its sense. Instead of throwing good money after bad (like thousands of crores on the corrupt employment schemes) and spending time and resource on finding new ways for government intervention into people's lives would do more to reform the economy and to create unskilled and skilled jobs.

Somehow I doubt that.

10. Ajay,

It’s really very pleasing to see the improvement in health indicators in the third round of the NFHS. It’s even more gratifying to be able to attribute the improvement to the high levels of economic growth in India. However, having said that I would also like to point out the lesson that’s stored in this correlation for the Ministry of Health and Family Welfare. That is, the urgent need for the MOHFW to strengthen its public health programs. The trickle down effect of India’s growth is certainly making it more possible for families to get better health care but, having said that, I would also like to direct our attention towards the millions of people whose health still depends on the effective functioning of the government’s public health system. I work for an organization that has just finished conducting a detailed analysis of maternal and childcare issues in the ten of the least developed districts of India. The data just started rolling in and it only substantiates my point about the pressing need for a better public health system. This is because, we found that while there are now a number of people that have the facilities to get better health services, there are many others that are either not as well informed, or lack the resources to get them. This results in many people turning up at the local government health centers and unless those centers become our first and most effective line of defense, we will not be able to achieve significant improvements in the health sector. Additionally, we will need time to achieve the high levels of economic growth that can usher in a sea change in the health of our population, given our current issues with infrastructure and such. Like you mentioned in your article, people will first have to become rich, and then hopefully, they can get better medical attention. This can take a substantial amount of time to happen. However, the government currently has the resources to bring about a large-scale change in health indicators. Thus, it’s important that the Government of India understands the urgency of the issue and makes appropriate amendments to its current public health practices. Your example of the European countries and the improvement in the health of their populations spurred by the high levels of growth in those nations is certainly useful in furthering the cause of greater economic growth in India, but we must realize that their governments have also put in place a strong and more effective public health system.

In sum, we should and must appreciate the impact that economic growth is having on the health of our population, but that does not mean that the government can shy away from its responsibility of building a robust public health system.

11. First about the IMR - I am wondering if the data provide any information on female infanticides/foeticides. Punjab, especially has come under heavy criticism recently for its skewed m/f ratio. 2-2.5 children sounds great (probably) but at what cost? I would like to know how many female foetuses are killed to achieve this great number. This relates to your emphatic assertion about the education of the masses. While I completely agree (for I would be a fool to deny) that our governments fare pathetically in their functions, I am yet to see any convincing result that tells me that private enterprises are sufficiently better. It is not enough to base our judgments on what is observed. Yes, women might choose a private over a public institution and then we can also show how 'satisfied' they are with the services. But when probed deeper, things might look entirely different. The prevailing perception of 'private-necessarily means-better ' (thanks to the rapid advent and advocacy of the institutions, similar in the lines of the Cato and the Heritage Foundation in this 'poor, developing' country of ours :-) - may be one of the many 'irrational' factors of people turning to such services. Sporadic reports are now aplenty providing information of the clandestine operations of these private entities, many of which are also involved in female infanticides/foeticides.

"When people get richer, they buy better food, better sanitation and cleanliness, more knowledge (e.g. education within the family), and services of private doctors / hospitals." -

I agree that when the purchasing power increases, people's options also increase. BUT, people do not necessarily buy and eat better food. That is a false assumption. My earning capacity makes it possible for me to buy a pizza from 'Pizza Hut' in Delhi (though I would detest going to a PH outlet in the US - and here are the socio-cultural and economic aspects), but does my purchase of a pizza make my consumption pattern better than a person who can afford to buy 'only' lentil, onion, tomato and some milk? While the consumerist patterns cannot and SHOULD NOT be dictated, the assumptions behind the advocacy for private enterprises seem quite off the mark.

Also as a researcher, I expect a much more nuanced stance and attitude towards the governmental sector from a person who also has a doctoral degree. Things are not so black and white. The language makes a difference and I sometimes wonder whether the only purpose here is to promote the profit-oriented enterprise and nothing else or the purpose is to pool together people from different streams of thought to do something about issues. If it is the former - then this lambasting, derogatory, black-and-white tone and picture is fine. But if it is the latter then may be, pointing out the limits of both the governmental and the private sectors seem crucially important.

Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticising me is perfectly okay; uncivilised language is not. I delete any comment which is spam, has personal attacks against anyone, or uses foul language. I delete any comment which does not contribute to the intellectual discussion about the blog article in question.

Please note: LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.