I fail to understand some things. Please educate me here and I mean it seriously.

What will the majority of people do in services? Health Care/unchecked housing construction as in the US? Tourism as in a tourist country like say Maldives/Singapore? Remember that the life expectancy is also creeping up, so you need to occupy people for a longer amount of time.

I think political-economists dismiss the importance of demographics primarily because of a general wariness of what they construe as social engineering (since according to econ theory, a policymaker cannot CHANGE human behavior, only EXPLOIT and MANAGE it). Having worked at a place which reccommended policies to deal with demographics, I would say that occaisionally we were guilty of that overreach.

Lo mismo digo yo. Which is just Spanish for the feeling is mutual.

“I am not saying that policy changes can’t alter the demographic dynamics of a country — reducing infant mortality, delivering better healthcare, and improving education can definitely do that. However, my (admittedly unsubstantiated) claim is that, on the margins, this change is lower than can be expected from TFPG-enhancing reforms.”

OK, Ok. Let’s just say that the beast walks on two legs: one of these is demographic and the other is institutional (in the broadest sense of this term). Of course at some level getting maximum benefit from something like TFP is very interesting, you only have to look at the US/EU differential in the uptake of ICT, and then the subsequent leveraging that the Scandinavian economies have made of it, to see that.

Basically I am a macroeconomist, and if I bend the stick in the direction of demography it is only because so many people tend to imagine that it has no importance. Essentially I think this idea comes from some flawed correlation studies which were done back in the 1980s. I think what has put the demographic question firmly back on the agenda again is the emphasis of David Bloom and others on the fact that it is *age structure* and not *size* which is the important thing. What you then need to do is ‘marry’ this understanding to a revamped version of demographic transition theory (ie one which takes into account the two-step longevity transition, and the arrival of lowest-low fertility).

If you like ‘big picture’ stuff, this is my first stab at putting my ideas together on all this:

http://www.edwardhugh.net/homeostasis.html

Essentially it is a ‘work in progress’, since I am not sure that I have all the details right yet. This is my objective over the next twelve months, to knock it all into shape.

Anyway, please stick around, I would appreciate an ongoing debate, and not only is there a lot to do, there is also a lot still to understand. (Incidentally I hope to have a post up on obesity in India, and why it could well be a big problem, over the weekend or early next week).

I agree with all your points. In some ways, I was making the point about the determinant of economic growth in order to outline the need and scope of reforms.

In my work, I have done quite a bit of research on productivity growth using company-level data (sales being a proxy for output, PP&E a proxy for capital input, and employee hours a proxy for labor input) accross several big countries. The purpose of doing this painstaking exercise was firstly to compare the results with the available macro data, and secondly to delve into the determinants of TFPG by looking at variables like level of competition, regulatory structures, innovation (although it is very hard to find good quantitative data on this), corruption, etc. Competition was by far the most significant determinant — not only at the national level but also at the industry level (yes, I actually went through the companies one by one and categorized them!). My measure of competition was simply the concentration of sales in a particular industry (more concentration implying less competition). Furthermore, I found that the fastest-growing economies had productivity which was driven by TFPG, not capital deepening. So my point here is simply that I have a renewed degree of faith in the residual as a ‘catch-all’ term, as well as the major driver of the economy.

Returning to the points you made in your reply, I fully agree that demographic dynamics are probably equally (if not more) important driver of growth. However, I do not think that there is much that can be done at the policy level to effect change here substantially. On the other hand, policy changes CAN have a direct and large impact on TFPG — hence I would say that is the best use of the reformers’ political capital.

I am not saying that policy changes can’t alter the demographic dynamics of a country — reducing infant mortality, delivering better healthcare, and improving education can definitely do that. However, my (admittedly unsubstantiated) claim is that, on the margins, this change is lower than can be expected from TFPG-enhancing reforms.

You bring up another important point which hadn’t occured to me — which is the INTERPLAY between demographics and productivity growth. I have to think some more about this - in the mean time, thanks for the stimulating discussion.

“You mention ‘using a lot less people’. What do you envisage the population to be employed in? IT enabled services? Manufacturing?”

Basically what I am arguing is the the transition from an underdeveloped to a developed economy can be seen in the career structure of the population. From having relatively few in services, a few more in industry, and the vast majority in agriculture you need to move to an economy were the great majority work in services, some work in industry, and a small percentage remain in a very productive agriculture. Obviously there is no iron law here, and the balance will vary from one model of development to another. In France, for example, more people still work in agriculture than they do in the UK.

A very interesting and substantial comment. Thanks.

Broadly we agree. Just a few points.

“For the past three years these have been running at 6% and 2% respectively — hence the 8% rate we have had.”

Yes, but the labour force component should now start to accelerate, since it is not only an age issue, but is about getting non-productive population into productive activities, and this is also gender related. As you suggest this is the part that isn’t automatic, since you have to have your labour market expanding fast enough to create the jobs to allow this to happen. Normally in the DD there is a certain circularity though, since the increase in saving which comes from having less dependent children makes funds cheaper and more available for investment, although again you still need efficient capital markets to make sure these funds get through to the projects that need them.

“EU and Japan both have a shrinking labor force, and the US and China are barely growing”

This is not entirely correct. Japan has a declining potential labour force, so does Italy and Germany (and incidentally Russia, and places like Lithuania, Estonia, you need to look at Eastern Europe). France, the UK and Spain have (thanks laregly to immigration still expanding workforces, in some cases on a quite expansionary basis. The US too still, thanks to higher fertility and immigration has a growing potential labour force. They need to create about 300,000 jobs a month just to absorb the impact of this. China is about to stabilise. Then China will not be able to increase output simply increasing the number of people working, it will have to go for improving the *quality* of the work done (as you say productivity) or in other terms move up the value chain.

“It is notoriously hard to figure out the driving forces of productivity growth - but crudely, it can be broken down into capital deepening (substituting capital for labor) and an abstract term which economists call ‘total factor productivity growth’”

Yes, well quite. Moses Abramovitz used to call the residual “the measure of our ignorance”. How you interpret this residual depends on
your theoretical model. On one account (which Griliches used to go for) there shouldn’t be a residual, it is all measurement error. On another, Schumpeterian, account it is a kind of technological ‘rent’ you get while the others catch up.

Let me put a couple more possible factors which influnce the total productivity (or income per capita) of a society. One is life expectancy. Here it is important to consider average life expectancy at the age of say 20, since the huge presence of infant and child mortality in the pre-transition societies skews the data otherwise.

Now the Demographic Dividend is also associated with a continuous increase in life expectancy, from say an expectancy of 25 more years (on average) at 20, to one of 45 more years, and then to one of 65 more years (to age 85). Now the interesting thing is that the first of these leaps is hugely beneficial from a productivity point of view, while the second leap (which is happening now in Europe) may be less beneficial since the child dependency situation is increasingly swapped for an age dependency one.

Also we, as individuals, have a lifetime productivity profile. This varies from occupation to occupation (cricketers peak earlier, and judges later) but if we consider averages, then it is normally thought that in skilled work people take off after the mid twenties, and peak in the mid to late 40’s. After this individual performance is normally considered to decline.

So obviously any society with very ‘thick’ generations in the 25 to 45 age group is a very productive one. This would tend to be a peak growth point.

I don’t know whether this is suitably clear. This is one of the topics which I myself am working on.