Computers and convergence

by John Q on May 16, 2004

When I first studied economics ( a long, long time ago) the textbook explanation of why income differed between countries was based on capital. In the simplest version for example, that of Harrod and Domar), rich countries had a bigger stock of capital than poor countries, and the problem was one of accumulating sufficient capital to catch up. In more sophisticated versions, rich countries had more modern capital stocks, and therefore benefited from embodied technological progress.

Even when I was a student, this kind of thinking was already being superseded by notions such as human capital theory[1]. Still, I’ve never seen a really convincing refutation. It strikes me that computers and the Internet provide one, at least as far as differences among developed countries are concerned.

Taking the United States as the leading country, by how many years does its stock of computers and Internet connections lead that other members of the OECD? Take a relatively poor and poorly-connected country like Spain for comparison. In 2001, only 9 per cent of the Spanish population was connected to the Internet, a figure that elicits the description “technological backwater”. Although estimates vary widely, it seems reasonable to suggest that the US reached this figure around 1995. It also seems reasonable to suggest a similar (roughly two-generation) US lead in relation to computers. That is, in the sector where its lead is arguably greatest, the US is six years ahead of Spain.

Now, if we suppose that labour productivity grows by around 2 per cent per year (this is generous), the combination of larger capital stocks and embodied technological change might account for a difference in income per person of around 12 per cent. This is comparable to the difference associated with differential timing of business cycles (two years of boom vs two years of recession), that is, it can be disregarded for most practical purposes.

I’m leading up to an argument that, in an important sense, the process of convergence among developed countries is complete, or close to complete. Differences in income per person, on average and for different groups within society, reflect differences in individual and social choices rather than the kind of differences typically considered in growth theory. I’ll try to develop this further in subsequent posts.

fn1. Despite its origins in Chicago, and the unattractive nature of its central metaphor, human capital theory leads directly to social-democratic policy conclusions.



Q 05.16.04 at 10:36 am

Interesting post. _the process of convergence among developed countries is complete, or close to complete_

On factors of production: For example, New Zealand will always be a poorer country because it has a small population miles from anywhere with no oil (low economies of scale, high transport costs). Australia will always have a problem because of the lack of water and rivers. Italy will always be stepping over ancient Roman settlements. UK will always have a problem because of the lack of space. Given equal mineral resourse, water, equal economies of scales, then converge is made more likely. USA has minerals, economies of scale, water, space, great weather – it is a potent mix.

Standard of living? For example, you might be happier working for half the money living in Nice, than twice the money, living in Stockholm. Equilibrium could quite possibly be a situation where low paid people in Spain are happy to save nothing because they are going nowhere, but high paid people in Frankfurt must save 20% of their income so they can visit Spain every so often.

Maybe the stable equilibrium is the unequal monopoly? The current usage of the US dollar is as a store of wealth and medium of exchange, which is being challenged by the Euro currently and maybe by other currencies in the future. This means that IR policy in the USA affects the rest of the world, as it affects so many international investors.

Can you demonstrate how the average New Zealander will earn the same purchasing power as the average American?


rube legendre 05.16.04 at 3:38 pm

Can you demonstrate how the average New Zealander will earn the same purchasing power as the average American?

Japan per capita GDP vs. US since 1960?


Atrios 05.16.04 at 4:56 pm

Yes, the “geography” (broadly defined) of the US, including the immense variety of mineral deposits, is an often overlooked source of our prosperity.


Giles 05.16.04 at 8:30 pm

“I’m leading up to an argument that, in an important sense, the process of convergence among developed countries is complete, or close to complete. ”

Which leads to the question of whether a process of divergence may now begin.

The interesting case in point would be countries, like south america, that have generally had high investment in education and a couple have had income per capita roughly equal to the US’s (Argtentina and Venuzuala)at some time in the last century and yet ended up diverging.

The second problem is that some countries like Korea have higher internet connection rates than the US and incomes that differ by a significant amount.

The third is the sotchasitc argement – maybe growth is a random process and convergence a rationalisation of a convergence to the mean process. In which case, the question is, is spain going to converge to a nice mean like the US or a nasty mean like morocco?


Giles 05.16.04 at 8:35 pm

alos provides an interesting counter refutation

most countires closest run in with the US were 20 or 30 years ago.


Jason 05.16.04 at 10:24 pm

Ummm, are you looking at the same graph as me?

I followed the link, and it looks to me like most countries highlighted bar (when they reached the highest % of GDP per capita) was quite recent, in the last 5 years of the survey.

One thing I’ve noticed from my stay in America is that people here work MUCH longer hours than people elsewhere (and this magnifies even more so if you take into account the longer commutes here). When I see Norway at 85.7% of US GDP, and I know that the hours spent on work there (for the cross section of society I know, engineers/professors, admittedly a biased sample) are about 60% of those in the US, I’d say it’s actually the US that needs to catch up to Norway.


andrew 05.16.04 at 11:12 pm

Jason’s point above has been made elsewhere (the last Economist issue I believe). France, for instance, has 5% higher output/work hour, productivity, that the US when you remember that they work about 5-8 hours less a week.

Which is John’s point – 32 hour weeks and 2 month vacations are in trade for more income.


Jason 05.16.04 at 11:30 pm

I wasn’t attempting to refute John’s post, my comment was more directed at Giles’ link and followup.


Brian Wilder 05.17.04 at 4:15 am

The comments on how many hours Americans work are interesting; another factor in American prosperity is China: China, by maintaining its currency at about one-third its purchasing power parity rate, has flooded the U.S. with cheap manufactured goods. Clothing, toys, all sorts of things, have declined significantly in price over the last ten years.

Now Chinese economic growth is pushing on commodity prices, as China imports vast quantities of oil, wheat, etc. For the moment, the huge stockpiles of American dollars are sufficient to pay for this trade, but at some point the Chinese may want to “raise” prices for the Americans and Europeans and “reduce” prices for their commodity imports, by depressing the value of the dollar. Letting the yuan rise and the dollar fall would raise domestic Chinese living standards substantially. Although it might undermine the export-led investment boom, at some point, a domestic investment boom will take over.

This may happen slowly or suddenly. If suddenly, the U.S. will take a big dent, as the world realizes that this is a Chinese century. China, together with its East Asian periphery stretching from Siberia to Burma, has scale and resources far beyond that available to the U.S. China won’t be within spitting distance of the U.S. or Europe, in terms of GDP per capita, when it moves up from 2nd place to become the largest economy in the world, but the impact on U.S. and European incomes of pressure on commodity prices, global warming, etc. could be substantial.


Giles 05.17.04 at 5:22 am

ok manybe most was too strong but 6 out of thirteen, the UK, Holland, Sweeden, belgium Canada and France all reached their peak more than 20 years ago.

I think any strict convergence theory would want a more than 50% success rate.


Giles 05.17.04 at 5:43 am

And the obvious point is that if you have a lower participation rate, then you invariably have a higher productivity since only the most productive workers are employed. If they us only employed Bill Gates its productivity would be about 5 billion per worker.

However, the problem with Frances higher productivity is its persistently higher unemployment rate – so any comparison needs to introduce some consideration of social/work equality. I high productivity worth it if the cost is a large alienated section of the population lying unemployed?


John Quiggin 05.17.04 at 6:18 am

Giles, it’s certainly necessary to take account of the average quality effect arising from different rates of participation, and I plan to take that up when I do my next post on this topic. But I don’t think it’s sufficient to overturn the general conclusion that output per hour has converged.


Scott Martens 05.17.04 at 11:34 am

I’m curious about the effects of foreign exchange on these figures as well – the ones from Wendell Cox at Demographia. Since a high-valued dollar lowers the cost of imports for Americans, it seems to me that exchange rates might have a larger effect on PPP ratios than it would initally appear. Has the now much lower dollar exchange rate raised import prices for Americans? The dollar has fallen sharly over the last couple of years, enough so that there might have been quite a lot of convergence with Europe and Canada recently.


Giles 05.17.04 at 4:22 pm

Scott – the US is still a relatively closed economy so depreciations in the dollar do not have as large or quick an effect on as they would on more open economies such as the UK and the Netherlands.

So unless the dollar stays depreciated for a long time do not expect there to be much effect.


Giles 05.17.04 at 4:38 pm


I looked at this with data from

Using all the productivity measures, European convergence stopped a while ago (although more recently than in the demo data I posted above). In particular the convergence of the poorest countries generally stopped first and, in some cases, preceeded the rise in high unemployment (although I haven’t checked this properly).

Most importantly you would then expect that the countries with the highest unemployment would have had the strongest convergence. But this is clearly refuted by the data on countries like Spain, Portugal and Greece – all of which had high unemployment and stalled productivity convergence. Ireland by contrast had low unemployment and high convergence.

This sugguests that the quality effect is not the most significant factor. More important I think is the age cohort effect – old workers are more productive than young workers. Thus countries like France, with a relatively old workforce tend to be more productive than say Spain and the US where young workers make up a larger proportion of the labor supply.


John Quiggin 05.17.04 at 9:11 pm

giles, I’m not sure if we’re disagreeing. You say convergence has ended, I say it’s complete.

I’m looking at the GDP/hour series from the dataset you mention, and it seems consistent with the claim I’ve made. That is, taking the US as the base, the leading European countries have values around 1, some higher, some lower.

As you observe, Spain has done poorly on the same measure in recent years, but this may be due to things like the age-structure of the workforce.


Giles 05.17.04 at 9:51 pm

But why has it ended? As I understand it you are arguing that this is because convergence in whatever the accumulable factor has been completed. But under most normal models this would imply that the countries that were poor initially would end convergence latest. In fact from the raw data it looks to me like the poor countries finished converging before the richer European countries and stopped at a lower level. And adding in age/cohort and labor quality effects may well strengthen rather than weaken this conclusion.

I am therefore suggesting that if you argue that it is complete it must be a) conditional and/or b) include to some “exogenous” event (as opposed to absolute convergence in some “endogenous” factor).


John Quiggin 05.17.04 at 11:20 pm

I suspect we’re starting to run into sample size problems. The data set contains only about 10 poor countries (below 80 per cent of US GDP/hour worked in 1979), and of those, Finland and Ireland have clearly caught up to the main group since then. Most of the others showed fairly strong catchup in the 1980s, then slowed down. But we’re really only talking about one business cycle in maybe eight countries.

Looking at the experience of the other European countries, and history in general, I’m willing to bet on further catchup in the future for these countries.

But (despite the fact that my post referred to Spain) my main point is about the completion of convergence among the leading dozen or so countries in Western Europe and North America.

Comments on this entry are closed.