Scott Sumner  

More on levels vs. growth rates

Patriotism as Political Correc... Jerry Brown's Rail and Tom Saw...

I seem to be one of the few people that are neither a China bull nor a China bear. On one side I'm constantly annoyed by China boosters who claim their rapid growth rate is proof of the superiority of the Chinese model. They often go on to claim that other developing countries should reject the "Washington consensus" and adopt a state-led economic model. And yet the facts show the exact opposite. China is far, far poorer than other ethnic Chinese economies, such as Taiwan, Hong Kong, and Singapore. And those three are all more market-oriented.

At the opposite extreme is people who point out how bad things are in China, without any perspective on the trajectory of change. Some of these people don't seem to know much Chinese history, and think problems like air and water pollution somehow call into question the incredible gains in living standards since the 1970s, when conditions were appallingly bad. If the people in the Congo or Afghanistan suddenly had Chinese GDP/person, do you think they'd be whining about air pollution?

Of course there's a grain of truth in both claims, China really is relatively poor and very polluted, and it really has been growing very fast. But I think both views miss the bigger picture:

Think of Mao's economic policies as a regime capable of producing real GDP per capita equal to 1/20th the global economic frontier (which might have been the US or Switzerland.) Then Deng starts opening up China's economy, and (combined with later reforms) this moves China to their current policy regime, which might be capable of producing a real GDP/person at 50% of the frontier.

There will be a multi-decade transition period where China sees a 10-fold increase in living standards, plus whatever increase occurred during that period in the leading economies. That will be a period of rapid catch-up growth. It does not mean that China has a good economic model in an absolute sense, it obviously does not. But merely by being less bad you'll have a lot of catch-up growth.

Over the next 30 years I expect China to do more economic reforms, and eventually reach a model roughly as efficient as Japan, France, Britain etc., that is, a model capable of delivering about 70% of the US RGDP/person. Obviously that's a guess, and I have no proof, it just seems like the mostly likely outcome, all things considered. But it's also possible they stay at 50%, or go to 100%.

BTW, China is not currently at 50% of US income per person, far from it. My claim is that they'd eventually get there with their current regime, assuming no more reforms.

Many people also seemed confused by my claim that bad economic policies don't affect long term economic growth relative to other countries, except during a transition period. I think most people did understand my claim that if France's welfare state reduced hours worked by 30%, their GDP/person would fall that far, and then growth would resume from that point forward. Rather the concern was that some policies reduce innovation, and hence long term growth.

Let's say African countries had a bad education system, and that reduced innovation. Would that reduce long term steady state growth in Africa? I doubt it. Africa did poorly for a few decades after the end of colonialism, as governance declined in quality. But since that one-time adjustment, Africa has been growing as fast or faster than the US. Even if Africa doesn't produce technological innovations, that doesn't stop 100s of millions of African peasants from getting cell phones. It simply means the level of African income might be only a fraction of the US. But when that level is reached, growth going forward over the millennia should be about the same in both places. A few centuries from now Africa may still have only 5% or 10% of US real income/person, but it might well be richer in absolute terms than the US is today.

There are many developing countries now considered "poor" that are actually richer than Britain in 1890, even though Britain was then one of the richest countries in the world. A developing country like Malaysia is several times richer than Britain was back then, and even Thailand is probably richer.

Levels and growth rates. The key to seeing the big picture is keeping both concepts in mind, and understanding which concept best applies to the specific situation being examined. Different policy regimes produce different long run steady state incomes, as a fraction of the top country. Then when policies change there is a transition period of faster or slower than global growth, as the country adjusts to its new level.

PS. Alex Tabarrok recently linked to an article questioning the accuracy of China's unemployment data:

So how high is Chinese unemployment today? No one knows but it could well be closer to 10% than to 4.1%.
The data doesn't show much variation, which is hard to reconcile with the large variation in real GDP growth. On the other hand I do think the reported rate (4.1%) is reasonable. On my last trip to China (in 2012) I noticed help wanted ads all over the place, in the windows of small businesses. Wages have been rising very rapidly, and factories report difficulty in finding enough workers. The recent slowdown in China may have changed that picture--if so it should show up in the wage data. I'd guess that unemployment is now trending upward.

HT: Angus

Comments and Sharing

COMMENTS (20 to date)
Brooks Alvarado writes:

I have a couple questions which are not about the main point (which is the distinction between levels and growth rates). It's about the numbers, so to speak.

If I google "rgdp per capita" and follow the first link, I get the World Bank.

Two points:

1) It says the US RGDP per capita in current US$ is 54,629. Comparing to the UK, France, and Japan:
UK : 45,603 / 54,629 = 83%
France: 42,736 / 54,629 = 78%
Japan : 35,194 / 54,629 = 66%

First, is this data reliable? Second, is 70% just a ballpark figure? I'm fine with using 70% as a ballpark figure, I'm just asking to make sure I understand where this is coming from.

2) What's the top country?

Different policy regimes produce different long run steady state incomes, as a fraction of the top country.

Talking about "fraction of the top country" and comparing every other country to the US could make the reader think the US is the top country. According to the World Bank table above, it's not. Norway is tops, followed by Macau, Qatar, Australia, Denmark, Sweden, Singapore, and then the United States.

There's some definite oil wealth pushing up RGDP there. But there's also countries without significant mineral resources, as far as I know. Also, the two countries immediately following the USA are Ireland and Iceland. Interesting all around.

Thomas B writes:

I'd argue that you're missing a key point: if every country has bad policies, there may be no growth at all - and in fact there was practically no global growth at all in the 3,000 years prior to about 1600AD. There were reasons other than policies, but the fact is that growth is not a given.

So, growth rates equalize, but the growth rate itself is not a given. How do we reconcile these?

I propose that the overall, global growth rate can be very much affected by policies: the more places with good policies, the faster; the more with bad, the slower.

Wealth levels increase with good policies, and decrease with bad ones. But wealth also flows. If policies are good in one place, it makes places with bad policies better off. Very likely, the reverse is also true: places with bad policies are bad for places with good policies.

E. Harding writes:

"Different policy regimes produce different long run steady state incomes, as a fraction of the top country. Then when policies change there is a transition period of faster or slower than global growth, as the country adjusts to its new level."
-Ah, but what if a country banned all imports of foreign technology, and banned all domestic innovation? Wouldn't that affect long-run growth? And why does China have faster growth than India, anyway? I don't think it's just a matter of levels.

Also, as someone else pointed out on the previous post on this topic, if Britain had banned all coal-burning in the 17th century, wouldn't that have radically changed world economic history and growth?

Tom Brown writes:

Scott, Jason uses his model to estimate a true Chinese unemployment rate between 5 and 10% (looks like about 6.5%).

AlexR writes:

Scott, you seem to think there is some background growth rate, something like a Solow residual, that is exogenous to all conceivable factors and necessarily equal across all countries and timeframes. All observed variation in growth rates from this background rate, across time and place, you attribute to various "one-time level changes." I won't ask what empirical support you have for this view; clearly there can be none. This is a choice on how to parse reality, which may or may not be a useful way of organizing facts.

What practical benefit do you see in it? Does it help predict anything? When you label growth as "catch up," does this help with figuring out when the "catching up" will end?

If an economy has output of y in an inititial period and thereafter experiences geometric mean growth of g per year over the next t years, this can certainly be described as a "one-time level change" of [(1+g)^t -1] y. But does that characterization make it any easier to predict output at t+1? Does the proposition "true growth rates never change; any observed change is really due to a transitory level change" hold any prospect of testable predictions?

Scott Sumner writes:

Thomas, I'm not at all convinced that the lack of growth in ancient times was due to bad policies. I think it had more to do with a lack of science and advanced technologies.

E. Harding. I don't believe there is any feasible policy regime that would hold down growth forever, it would simply affect the level (which is itself hugely important, to be sure.) Technology would get smuggled in.

China is growing faster than India because it's in a catch-up phase, due to previous policy changes. India's also had some policy changes, but far less dramatic than China.

To, I don't understand his model, so I can't comment.

Scott Sumner writes:

Alex, You said:

"I won't ask what empirical support you have for this view; clearly there can be none."

I'm not going to even respond to that, you seem to have your mind made up.

As far as how it's useful, it inoculates me from saying erroneous things about China, which (as I explained in my intro) I see so many other people doing, based on either levels or growth rates.

emerich writes:

Almost completely agree with your comments about China, which I've been observing for several decades, now from Hong Kong, though I think there's good reason to be skeptical of official statistics. Policy in China is far from ideal, and still much worse than the U.S. let alone Hong Kong or Singapore or Australia, but it grew so fast for 35 years because Deng's policies were a huge improvement over Maoism. And people forget that Tokyo made news for its awful pollution, with stories about oxygen dispensers for pedestrians on sidewalks, in the 70's, and L..A. was the poster child for polluted cities only a decade earlier.

Can't say I understand why there's some universal growth rate to which all economies converge once you strip out "catch up", but I look forward to further discussion about that. Why wouldn't an economy with policies more friendly to innovation and entrepreneurialism have a higher long-term growth rate than one less so?

BC writes:

Here is one thing that puzzles me about economic "productivity". Suppose Africa produces only apples and the rest of the world produces only phones. Now, suppose due to innovation phone production increases but apple production stays constant. The outward shift of the phone supply curve probably causes the relative price of phones measured in apples to fall and the relative price of apples to phones to increase. Similarly, diminishing marginal utility of phones leads to increased relative marginal utility of apples. In short, Africans can buy more phones with the same number of apples, so their income, well-being, etc. all increase. I'm fine with that, but it seems strange to say that their "productivity" has increased. Yes, they may produce more utility or output measured in constant year dollars (due to changes in *relative* prices). But, their physical productivity, the number of apples, has not changed at all!

Am I thinking about this correctly? Can a phone maker's productivity affect an apple grower's economic productivity even if physical production of apples is totally unrelated to phones?

E. Harding writes:

Just for curiosity, here's a sampling of a bunch of African countries and two Asian countries:!ctype=l&strail=false&bcs=d&nselm=h&met_y=ny_gdp_pcap_pp_kd&scale_y=log&ind_y=false&rdim=region&idim=country:NER:TGO:GMB:BEN:ZMB:UGA:TZA:ZAF:SWZ:SEN:STP:RWA:NGA:NAM:MRT:MLI:MWI:MDG:LBR:GIN:GNB:KEN:LSO:ETH:ERI:BDI:CIV:COM:USA:CHN&ifdim=region&hl=en_US&dl=en&ind=false
How consistent is this with Scott's model of things?
BC, excellent question.

BC writes:

Here is another puzzle. Consider two workers A and B that produce phones. Suppose A's phone productivity increases, while B's remains constant. The increase in supply of phones decreases the price of phones relative to all other goods. Thus, B's economic productivity measured in constant-year dollars falls, even though he produces just as many phones as before! (Again, the *relative* price of phones has fallen, so inflation adjustments don't change that.)

When one person's productivity increases, it increases the (economic) productivity of people producing other goods and decreases the productivity of people producing the same good!

Brandon Berg writes:

Those numbers are not PPP-adjusted. Take a look at the "GNI per capita, PPP" table at the same site. US is still not the top, but every country above it is higher due to special circumstances (Qatar, Norway, and the UAE get large portions of their GDP from energy exports, and Singapore and Hong Kong are city-states). Singapore probably owes some of its lead to policy, but it's unlikely that a large country with similar policies would have a per-capita GDP quite that high.

Of course, the US is still behind where a hypothetical economy with ideal policy would be.

Yes, this is basically what I was asking for in my comment to your last post. Thanks!

As an aside, can Malaysia still be considered a "developing economy?" With a per-capita GDP (PPP) of $24,000, it's solidly middle-income, and rapidly gaining on Greece and Portugal. Or do you just mean "developing" in the sense of still undergoing catch-up growth?

Brandon Berg writes:

I think this can be proven by contradiction. Suppose a difference in growth rates can in fact be sustained indefinitely. This implies that the per-capita GDP gap grows indefinitely. With reasonable assumptions, so does the wage gap.

But the wage gap can't grow indefinitely. Eventually it grows to the point where the poorer country's wages are so low relative to the richer country's that it becomes an attractive destination for foreign investment, which leads to catch-up growth.

Furthermore, technology is global. Even if a government suppresses any domestic innovation, the economy still benefits from innovation elsewhere. For example, North Korea still has computers.

Of course, this assumes certain limits to the government's awfulness. If it completely isolates the economy, keeping out all new technology and foreign investors, that could do it. But this strikes me as a purely academic point.

David Michael Myers writes:

[Comment removed pending confirmation of email address. Email the to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Robert Schadler writes:

One comment:
It seems close to Hayek's "fatal conceit" or hubris to think international comparisons of household income and PPP are much more than notional. With the devaluation of the yen, it is clear that exchange rates are significantly distorted by government policy, with each government affecting its exchange rates differently. How then is there an apple-to-apple comparison possible, especially at a refined level?
PPP tries to get at the "reality" of economic well being but then implies that with "purchasing power parity" people in the compared countries want to purchase a similar basket of goods. This is often quite false, due to differences in cultural tastes.

I think "unemployment statistics" should be used with particular caution, especially when comparing countries with very different levels in the monetized employment market. "Unemployment" on a family farm (and even a family business) may be virtually impossible to make comparable to other unemployment stats. And with China (and other countries) one has to wonder about the government skewing the results. If we expect govt failure and distortions generally, and especially with some govts surely making fine statistical comparisons based on their stats is a strange exercise. Goat entrails may yield greater insights.
Put more starkly, the private sector provides some stats, in some countries, relevant to assessing unemployment. But is there enough of those to make global and a great many dyad comparisons? I think not.

Scott Sumner writes:

Sorry for the slow replies everyone.

Brooks, I'm using PPP adjusted numbers.

Emerich, The reason a country's growth advantage is not permanent is that the further behind another country falls, the easier it is to catch-up by borrowing innovation from other countries. If not, countries would diverge without end, but the rapid growth in cell phone use in Africa suggests that is not plausible. US tech is helping Africa, with a delay.

BC, Cell phones have definitely boosted African farm productivity. They help African farmers learn quickly which markets offer the best prices. Other technologies from the West like fertilizer and pesticides also help. Recall that western technology triggered the Green Revolution in developing countries.

Brandon, Yes, I guess it's a borderline case, although middle income countries are usually still considered developing. I suppose the Thailand example is better.

Robert, I agree with those observations.

Scott Sumner writes:

E. Harding, Your link doesn't work. Can you state the question in words?

Brandon Berg writes:

Here's E. Harding's link. I was able to get it by double-clicking the link, which selects the entire paragraph, allowing me to copy and paste it.

BC writes:

Scott, my point was not to debate whether Africans' physical productivity has actually increased. My point was about the concept of (economic) productivity itself because even *if* African farm physical productivity was unaffected by phones, relative price changes (caused by increase in phone supply) *in and of themselves* can boost economic productivity.

I think this is an artifact of what Arnold Kling calls "thinking of economies as GDP factories". Intuitively, we think of "*product*ivity" as measuring what one (physically) *produce*s. However, because economic productivity is measured in constant-year dollars (not in apples, phones, etc.), it actually measures *what one can trade for* in exchange for what one produces. Hence, it is impacted by relative prices, not just physical quantity. By the way, this could be a feature not a bug in the definition of economic productivity because one really is better off when relative prices change in one's favor. It's just confusing intuitively to call this "productivity".

Rob writes:

Would be good to get your thoughts Scott.


Comments for this entry have been closed
Return to top