Return to education in Rural Vietnam

From: Nguyen Hoang Giang <nghgiang2005@yahoo.com>

Date: Apr 26, 2006 8:30 AM

Subject: [Vsg] Return to education in Rural Vietnam

Dear list,

I am doing a research on the effect of economic structure and economic development on economic return to education in rural Vietnam. Thus, I am finding references on the topic of educational return in Vietnam, especially in Rural Vietnam. Would anyone tell me a reference or research relating to this topic?

Best regards,

From: Matthew Griffiths <matthew_griffiths98@yahoo.com>

Date: Apr 26, 2006 6:26 PM

Subject: Re: [Vsg] Return to education in Rural Vietnam

Dear Nguyen,

Are you looking at adults returning to education or children who have dropped out of school and returned as their families economic cicumstances improve? Or something else? Will help if I can.

From: Adam @ UoM <fforde@unimelb.edu.au>

Date: Apr 26, 2006 6:53 PM

Subject: RE: [Vsg] Return to education in Rural Vietnam

There may be something in the Mekong Delta Poverty Analysis financed by AusAID - I think the final draft is up on their website. So far as I recall, and this is not exactly the same as your question implies, the survey data suggests that incomes rise up to primary level and then stop rising with additional years of study until tertiary.

This can be read off as education being a bought marker of socio-economic status, a result rather than a cause of higher incomes. This is what some sociologists and anthropologists will tell us. But this contradicts assumptions that education adds to human capital and human capital can be understood as an input to production. The econometrics tends to suggest that this is questionable as a universal assumption. And you cannot easily test it as the model you chose will assume much. One result of this is that the statistical results are often contradictory, because the significance tests rely upon the truth of the model, more or less, and are so often (but often unknowably) spurious.

The Sida financed survey of rural Vulnerable Groups in the 1990s shows that welfare (measured in different ways) did not drop much for the less educated, perhaps in part because they had rather good land holdings. This is on our website.

Cheers

Adam

From: Hy Van Luong <vanluong@chass.utoronto.ca>

Date: Apr 27, 2006 12:12 AM

Subject: RE: [Vsg] Return to education in Rural Vietnam

Dear Giang and interested researchers,

In the World Bank volume edited by Paul Glewwe et al. ECONOMIC GROWTH,

POVERTY, AND HOUSEHOLD WELFARE IN VIETNAM (2004), Nga Nguyet Nguyen has a

paper on education with a section on the econometrics of the return on

education, using VLSS 97-98 data.

Hy V. Luong

From: Adam @ UoM <fforde@unimelb.edu.au>

Date: Apr 27, 2006 1:08 AM

Subject: RE: [Vsg] Return to education in Rural Vietnam

I would be interested in Hy v Luong's take as an anthropologist on the

points I was making about problems with data and models. We had some issues

in the Mekong Delta Poverty Analysis work because of the very strong belief

amongst some Vietnamese analysts that returns to education HAD to be

positive, and so the focus should be on educating farmers ...

My basic issue with the econometrics is that of the difference between

correlation and causality and the problems caused by model specification

that treats education as an input to human capital treated in turn as a

factor of production (usually proxied for quality by level of education ...)

given the vast amount of correlation out there in the data.

Adam

From: DHAUGHTON@bentley.edu <DHAUGHTON@bentley.edu>

Date: Apr 27, 2006 4:05 AM

Subject: RE: [Vsg] Return to education in Rural Vietnam

Greetings to all! Paul Glewwe has also published a number of articles on education; on the matter of VLSSs, our two edited books (the first one, Haughton, Haughton, Bales, Chuyen and Nga, Institute of Southeast Asian Studies; the second one at http://www.undp.org.vn/undp/docs/2001/living/lse.pdf has a chapter on education and income) could be pressed into service, and on the issue of causality mentioned by Adam, a very important issue IMHO, work of David Bessler and colleagues on the use of Directed Acyclic Graphs in modelling could be of interest. Best, Dominique

From: Adam @ UoM <fforde@unimelb.edu.au>

Date: Apr 27, 2006 5:20 AM

Subject: RE: [Vsg] Return to education in Rural Vietnam

What is a Directed Acyclic Graph and does it really deal with the problem?

From: DHAUGHTON@bentley.edu <DHAUGHTON@bentley.edu>

Date: Apr 27, 2006 5:27 AM

Subject: RE: [Vsg] Return to education in Rural Vietnam

Dear Adam and colleagues, here is David Bessler's website http://agecon2.tamu.edu/people/faculty/bessler-david/WebPage/Homepage.htm, his paper on world poverty is probably a good source to get an intro to DAGs. Does it deal fully with causality? no, nothing does with observational data IMHO. Does it deal with spurious correlations, direct and indirect predictors? Yes, to some extent. If anyone would like to pursue this further, please let me know I have other papers, albeit not in the area of education. All the best, D.

From: Adam @ UoM <fforde@unimelb.edu.au>

Date: Apr 27, 2006 4:31 PM

Subject: RE: [Vsg] Return to education in Rural Vietnam

Dear D

Had a look at the DAGs. Sounds like what you are told in Stats 101 - graph the data in various ways and see what should fit, then run a regression. This risks the 'Measurement without Theory' taunt a la Koopmans and Mitchell. But so what?

For the culturally sensitive out there who read this list, and usually dismiss economics as silly, then what this is all about is the empirics associated with arguments about the epistemological assumptions that support belief in a transcendental logos as a basis for action - where does the notion of causality come from that 'we' construe as underpinning intentional change, aka classic 'Western'. policy logic? Dominique is saying that this cannot come solely from the data, which IMHO is a classic and very robust position.

And what if (as I think is basic to Vietnamese, Dao Lao and so on) there is no cultural predilection to belief in a transcendental logos, in fact the reverse? Pragmatism thus read as simply a belief that abstract judgements are somewhat nonsensical. Which suggests that policy in Vietnam and policy in Canberra are, ontologically, different. IMHO.

Or, to quote a good joke, 'try the soup'.

Regards

Adam

From: Jim Cobbe <jcobbe@mailer.fsu.edu>

Date: Apr 28, 2006 4:57 AM

Subject: Fwd: RE: [Vsg] Return to education in Rural Vietnam

I think this discussion somewhat misses the point about attempts to measure rates of return to education. There are several problems in trying to produce meaningful estimates, all currently insurmountable in practice, in my opinion. First, one cannot in reality measure some of the most important inputs to education, namely talent and effort by the pupils/students, and quality of instruction. People have tried using various proxies, but they are none of them satisfactory, IMHO. Second, to obtain an estimate one needs to know the return, i.e. the average increment to earnings attributable to an extra amount of education, over the average working lifetime. That strictly speaking requires a projection of average earnings by education for 40 years odd, something we cannot produce with any accuracy, apart from the issue of potential selection bias [persons who are more productive for other reasons may tend to acquire more education]. Third, for policy guidance, what one would really want to know is the marginal rate of return, not the average -- and estimating that would be even more intractable.

There are ways to get crude estimates, and people [and agencies like the World Bank] use them all the time. Just don't bet on their accuracy or reliability; as a project officer for US A.I.D. once said to me, "we all know that by changing your assumptions a little you guys can produce any rate of return I want." Regards, Jim

From: Adam @ UoM <fforde@unimelb.edu.au>

Date: Apr 28, 2006 5:38 AM

Subject: RE: RE: [Vsg] Return to education in Rural Vietnam

Well, I think the point remains.

If people - like you - define the return as " i.e. the average increment to earnings attributable to an extra amount of education" then they are assuming a causality where a change in earnings is attributed to a change in education. Education changes human capital and changes inputs to production. If these inputs are paid more then earnings rise and so extra education is rewarded.

But if you think that it may be household socio-economic status (rather than an individual's personal earnings) that is crucial to behaviour, then that status may be marked by increased outlays on education, such as a form of conspicuous consumption. In this case the change in education is attributed to a change in earnings. And the education, on examination, does not seem to change productivity. It is there to reproduce social differentiation and positioning, as Veblen argued.

The former is the common assumption. I think much Vietnamese data seems to support the latter. Common (but far from universal) Vietnamese cultural assumptions are the former - people are poor because they are uneducated rather than people are uneducated because they are poor. My sense, having interviewed an awful lot of farmers in my time, is that it is the latter. Poor people spend their money on other things. And this is what the surveys tend to show. Interesting to compare it with health care.

Another line, which you also allude to, is that much of the education on offer to farmers often does not pay, from their point of view.

Rather complicated!

Cheers

Adam

From: Jim Cobbe <jcobbe@mailer.fsu.edu>

Date: Apr 28, 2006 6:08 AM

Subject: RE: RE: [Vsg] Return to education in Rural Vietnam

I think we are very close to agreement, if not in total agreement [see the reference to selection bias, which is perhaps interpretable as an obfuscatory economist way of saying one does not know the direction of causation]. Of course there is a consumption element in all education [hey, I work at a 'party school' in the US!], and absolutely those with higher income buy more education for their children, on average. We don't know for sure how much of the reason they do that is because they think they need to for social reasons, and how much because they think it will help their children [and therefore maybe them in old age or disability] economically. We can only observe; and the observation is close to universal that both (a) the children of the better off on average are more educated, and (b) on average, those with more education have both higher incomes and better social status. Given (b), to try to produce a private rate of return estimate for education is a reasonable objective, even if [IMHO] unreliable; not knowing the causation means that attempts at using such estimating methods to produce genuine estimates of social rates of return are spurious -- although it is the social rate that one needs for policy purposes.

In theory perhaps just rather complicated, but I'd say very complicated in practice, with the result that most people just assume more education is a good thing. Cheers, Jim

From: DHAUGHTON@bentley.edu <DHAUGHTON@bentley.edu>

Date: Apr 28, 2006 6:22 AM

Subject: RE: RE: [Vsg] Return to education in Rural Vietnam

Hello everyone, agreed, we seem to be more or less on the same wave length.

There are statistical techniques(PC algorithm, for partial correlation, see

Tetrad homepage at Carnegie Mellon) that attempt to identify the direction

of a link (from poor to educ or the other way around), but they are no

panacea, and in some cases the methodology will return a verdict of

"unclear direction", I suspect it would do that in the case of poor and

educ). Best, Dominique

From: Markus Taussig <markustaussig@mac.com>

Date: Apr 28, 2006 7:11 AM

Subject: Re: [Vsg] Return to education in Rural Vietnam

Earnings might also not capture the full economic effect of education. In a country like Vietnam, education investment likely has a substantial element of signaling aimed at addressing risk aversion, i.e. securing not just more income, but more stable, predictable income.

From: Adam @ UoM <fforde@unimelb.edu.au>

Date: Apr 28, 2006 4:07 PM

Subject: RE: RE: [Vsg] Return to education in Rural Vietnam

Yup. Rat phuc tap.

There is a colleague of mine at the University of Melbourne who has been

paid by the Reserve Bank to play with a very large Australian incomes etc

panel (40,000, if I recall correctly). Very lucky man. He told me that the

recorded correlation of incomes between brothers was LESS than that between

strangers. He concluded that the image we have of social stability and so on

(the 'cake' metaphor) was simply imaginary: mobility as proxied by incomes

variation within small kin groups was in fact far higher. Whether we

ourselves should conclude that, of course, is entirely another matter.

But we still come back to the cultural issues. We all have these beliefs.

What we are adding to the VSG debate, from which the professionally

culturally sensitive seem so far strikingly absent, is what this data shows.

We all know that data is relative to observational theories and logics,

which are contingent. Whether this makes it transcendentally true, such as

the statement that all statements are relative, is not something we engage

with (or we should not). But, the data understood as such does not clearly

show that there is a return to education similar to, for example, putting

money into the bank. It is as consistent with the data that people treat

bank deposits as items of consumption ("look, I've got a check book, aren't

I successful" - see the film Tommy BTW) rather than the other way around.

Yet the strong belief, easily taught and internalised, amongst many

Vietnamese social scientists, who may just influence policy, is that

correlations between the return to education and earnings shows that more

money spent on education WILL increase incomes and so GDP.

And the recent Congress, as the recent medium term trend, is towards

'developmentalism', 'correct and implementable policy', 'we do this and we

know that that will happen', and so on. Otherwise why have perhaps the most

effective man around as the new Premier. Risky? Perhaps, does this sit

easily with other cultural traits? And have we been here before ('Rational

priority to the development of heavy industry')? Wait and see.

When will the culturally sensitive come out and play? Can they cope with

economists who want to play, perhaps parallel play, perhaps even in the same

sand pit?

Adam

From: jon mcintyre <jon.mcintyre@gmail.com>

Date: Apr 29, 2006 6:58 PM

Subject: Re: [Vsg] Return to education in Rural Vietnam

This is a really interesting discussion, especially given the debates about the return to education in other countries such as the US, and how these debates may make use of rather thin statistics on all sides of the discussion.

What interests me, especially in the Vietnamese case, are possible confounding factors in the causal chain. Off the top of my head something that could confound the education-income relationship may be number of, or quality of relationship with, relatives abroad (or in urban centers) that may simultaneously relate to both income and education in a way that disrupts a clear causal relationship between the two. I suppose if someone could accurately quantify such relationships they could be accounted for with multivariate models, but I'm an anthropologist and I tend to doubt such quantifications.

Just something to think about from a qualitative researcher that dabbles in quantitative methods.

From: Adam @ UoM <fforde@unimelb.edu.au>

Date: Apr 29, 2006 8:09 PM

Subject: RE: [Vsg] Return to education in Rural Vietnam

This is very interesting too.

On the things that can happen when there is this sort of underlying complexity, and there is the risk issue that Marcus mentioned, is that people who assume that there is a simple relationship and run regressions obtain spurious results, especially if they are working for a Professor, or a client, with strong priors so they need to produce a statistical result where the significance tests are passed, so that they mine the data and select functional forms to suit.

But this often results in a literature full of journal articles that have been refereed and passed, but which disagree with each other. That can be taken as an indication that the underlying model - such as the common human capital one - is awry. But each article can (as often is) used to support a particular policy position, or donor stance.

Adam

From: Peter Kiang <peter.kiang@umb.edu>

Date: Apr 30, 2006 4:40 AM

Subject: Re: [Vsg] Return to education in Rural Vietnam

some of you interested in this thread might want to look at Peter Taylor's

recent book:

Unruly Complexity: Ecology, Interpretation, Engagement. Chicago: University

of Chicago Press, 2005.

also see:

http://www.bemidjistate.edu/dsiems/courses/ConsBio/Complexity.html

pk

Return to top of page