If you want to know why some countries are rich and others are poor, a good starting place is this graph, which I took from the first edition of David Weil’s quite marvelous textbook on Economic Growth:
Because I took this from Weil’s first edition, these data are several years old, and some countries might have moved up or down the ladder since then. But the overall picture is clear: More capital per worker means more output per worker, and more output per worker means more income per worker. This relationship — in fact, the nearly linear relationship that you see on the graph — is just what standard economic theory predicts. It’s nice to see that prediction so powerfully confirmed.
Capital here refers to physical capital — the machinery, factory space and office equipment that allows workers to be more productive. A garment worker with a sewing machine produces more blouses per hour than a garment worker with a needle and thread; therefore the garment worker with a sewing machine earns higher wages. (A good rule of thumb is that workers are paid about 2/3 the value of what they produce.) If you want rich garment workers, you need a lot of high-quality sewing machines. If you want rich farm workers, you need a lot of high-quality tractors.
And if you want more high-quality tractors, you need more factories producing them — which means fewer factories producing consumer goods like, say, cars. That in turn requires consumers to rein in their immediate appetites, spending less and saving more. In other words, capital accumulation is driven in part by frugality. Forgo that car purchase and the factories will produce another tractor instead. That’s why tax policies matter. A tax that punishes saving (e.g. the estate tax) is a tax that encourages spending. Because of that tax, the rich buy more cars, the factories produce fewer tractors, and farm wages fall. If you doubt that farm wages depend heavily on the production of tractors, have another look at the graph.
Do not, however, jump to the conclusion that if, say, Nigerians had access to Japanese levels of capital, then Nigerian wages would rise to Japanese levels. Part of the reason Nigerians have so little capital is that capital is used less efficiently in Nigeria, so people choose to accumulate less of it. To move up this ladder, you need to do more than just accumulate capital — you’ve got to be the sort of country where capital is worth accumulating. What that entails will be a topic for a future post.
Steve,
Beautiful graph. First thing that springs to my mind is the circularity problem. Why does the higher wage not lead to a higher value of capital, rather than that a higher value of capital leads to a higher wage?
I am guessing it has something to do with that capital is traded internationally whilst labor domestically and your point that capital is valued more where it is used more efficiently?
There are three questions I have – first of all, can you prove that the estate tax leads to higher spending IN PRACTICE? Sure the assertion makes sense in theory, but (in particular when death is involved) human beings aren’t nearly as rational as economists conveniently assume for the sake of their models.
Secondly, is there any way of quantifying how large this effect is? Is it maybe so small as to be insignificant?
And thirdly – does the estate tax have any benefits? What are the implications for the economy if capital accumulates in an extremely top-heavy fashion (which is to some degree countered by the estate tax)? Why do you think the negatives outweigh the positives?
There is evidence that high inequality leads to lower growth, after all.
http://www.economist.com/economics/by-invitation/guest-contributions/reducing_inequality_can_be_growth-enhancing
Bravo! The post of the year.
capital accumulation is driven in part by frugality.
In other words, the only way to increase productivity is to increase your savings and reduce your consumption.
Do you think that a policy of printing money and lending it helps reach this objective or works against it?
I do not see a linear relationship, your units on the graph are in logs. This is also a cross sectional observation, so this does not tell us anything about capital accumulation or how a country’s growth rate of GDP depends on capital.
1. But if everyone’s saving, who’s buying the tractors?
2. But if people don’t buy cars, who will build car factories?
Martin-1: As I said in the last paragraph, you must not infer causality from this graph. What’s happening is that a) productivity affects both capital and income and b) capital affects income — so what you’re seeing here is partly due to the fact that capital and income have a joint cause (productivity) and partly due to the fact that capital affects income. I’ll try to sort all of that out in a post later this month.
Maurizio Colucci:
In other words, the only way to increase productivity is to increase your savings and reduce your consumption.
Do you think that a policy of printing money and lending it helps reach this objective or works against it?
Inflation discourages people from holding cash, but cash is not the only store of savings, so I doubt this is a big factor.
Except for your final paragraph, this post has a very Mises-esque flavor to it. Mises felt that savings (and therefore capital accumulation) were the only way to generate wealth in an economy.
I recall reading in the blogosphere a few weeks ago someone suggesting that poorer countries must first accumulate a lower level of capital before they can reap the gains of higher levels of capital. So, Nigeria would have to first “become China” in order to eventually “become Japan,” and so on. This makes intuitive sense to me, but I’ve never read a comprehensive analysis on the topic.
Steve - This is more or less the subject of my research. What I found, looking at developing economies since 1980, is that those which tried to impel growth by accumulating large capital stocks mainly failed. Big investment, particularly when done using government funds after a commodity price boom, does not lead to take-off.
It’s certainly true that output growth and capital growth go hand-in-hand, but I’m strongly disposed to believe that both are led by a third cause.
My findings, of course, are not necessarily generalizable to developed economies.
Cool graph. Now to quibble:
1. Intrigued that Landsburg chose a graph focusing on physical capital but not human capital. Is human capital harder to measure? Less explanatory?
2. The graph is labeled "Capital per worker" and "Income per worker" "“ which I take to mean average capital per worker and average income per worker. I wonder how this graph would change if we looked at median levels of capital owned by worker, and median levels of income received by workers.
After all, a nation of slaves and slave owners might fit well on this graph "“ with high levels of "Capital per worker" corresponding with high levels of "Income per worker." I'd be curious to hear what the workers at Foxconn think of Landsburg's "good rule of thumb "¦ that workers are paid about 2/3 the value of what they produce."
3. I'm curious how "worker" is defined for purposes of this graph.
4. And, as ever, I'm curious about how "income" is defined. In particular, is the value of government services counted as income?
Hypothesis: Europe has higher level of taxation than the US, but uses these resources to invest in human capital (education, health care) at a government level. Weil excludes these benefits from measures of "income," while including comparable income that US workers earn and spend for these purposes.
One thing that always confuses me is that one person’s consumption is by definition another person’s income. If people are not buying cars what incentive do I have to buy equipment to produce cars? Or take a more extreme example suppose everybody saves 100% of their income, then consumption will be 0, and no incentive to produce.
Salim Furth:
It's certainly true that output growth and capital growth go hand-in-hand, but I'm strongly disposed to believe that both are led by a third cause.It's certainly true that output growth and capital growth go hand-in-hand, but I'm strongly disposed to believe that both are led by a third cause.
I hope I did not imply otherwise! Total factor productivity, of course, drives both output growth and capital growth. This is exactly what I promised in my final paragraph to address in a future post.
I wish these numbers were normalized to the cost of land in the respective countries.
My hypothesis: both these numbers are proportional to cost of land.
(1) Wages: a worker is an economic entity whose main expenses are housing (proportional to local land costs), food (proportional to cost of agricultural land, again) and entertainment (again,…). In the presence of unemployment, wages for workers without unique skills can’t significantly exceed these expenses.
(2) Capital: “Capital here refers to physical capital "” … factory space …”, which I suspect is the main component. (Again, it’s proportional to cost of land + cost of construction (wage), which is proportional to cost of land according to (1). Sewing machines and tractors are not that expensive compared to buildings and fields.
In response to nobody.really, the income per worker is about $70,000. Given that this is old-ish, that does not represent median income, which is about $45,000. I presume this is mean income.
Steve you wrote: “Inflation discourages people from holding cash, but cash is not the only store of savings, so I doubt this is a big factor.”
Printing money sends a false signal to the market that there is more capital available than there actually is. Economic agents both increase their consumption and start on long term projects when society hasn’t accumulated enough capital to complete these projects (eg building a house that needs 100 bricks when there are only 80 bricks available) leading to their inevitable collapse.
Read Full Article »