« The generosity collapse | Main | Spain, Ireland, and the US. All breaking Okun's Law. »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Surely you are joking ?
This is an easy one.
Because of much less Social Securtity in the US, workers are weak in a recession.
They have two options: work harder or not work at all and have no or very limited unemployment insurance.
Moreover if they find another job they have to accept a lower wage.
The outlier in the opposite direction is Germany thanks to their Kurzarbeit.

See also:

ws - so that same output is being produced by fewer workers putting in longer hours for less pay, we should see average hours increasing in the US. We don't (this is data on average hours worked in the US from the Bureau of Labor Statistics http://data.bls.gov/PDQ/servlet/SurveyOutputServlet):

Year Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual
2006 34.5 34.5 34.5 34.7 34.5 34.4 34.5 34.5 34.5 34.8
2007 34.5 34.5 34.6 34.7 34.7 34.7 34.6 34.5 34.6 34.5 34.6 34.7
2008 34.5 34.5 34.7 34.6 34.6 34.6 34.5 34.4 34.4 34.4 34.3 34.2
2009 34.2 34.1 34.0 33.9 33.9 33.8 33.8 33.8 33.8 33.7 33.9 33.8
2010 34.0 33.9 34.0 34.1 34.2 34.1 34.2 34.3 34.2 34.3 34.3(P) 34.3(P)
P : preliminary

I said "work harder" , not "work longer". I think that is what higher productivity means.
Also "Kurzarbeit" doesn't always mean less hours. From Wikipedia:
"If an employee agrees to undergo training programs during his or her extra time off, they can often maintain their former income"
So effectively this means "work less hard".
Your second "fake" explanation is actually quite good. But that doesn't help the workers.
So the difference is all about choice.

Off topic: I always try to guess whose post I'm reading and never look until I get to the end. Usually I've made up my mind in the first few sentences, and generally I'm right. Sure, topic often gives it away, but your individual voices are strong clues too. This one had Nick written all over it.

These graphs are an exemplary demonstration of what the "hire and fire", weak safety net US workforce experienced during the crisis, versus the strong, unionized, Kurzarbeit-supported workforce in Germany.

The takeaway is that being unemployed in the US is a sticky condition. Once a worker 'falls out of grace', the chances of getting work again dwindle with every passing month and year. The US stats are in fact optimistic, real unemployment is probably well above 10% - it's just that whole categories of workers have stopped looking for jobs in a trackable way - the benefits time out after ~55 weeks and the prospects are not good.

Contrast that with Germany where it is being employed that is sticky: the country 'buffered' workers in the Kurzarbeit government safety net and had them ready and well-motivated by the time the crisis is over. And Germany is now kicking the pants off the US ...

The US policy is a heartless right-wing approach to those supposedly freeloading, benefit-seeking unemployed workers who are really just caught up in a 'recalculation' or some other structural shift that is their damn fault.

The German policy is to buffer pre-crisis capacity and structure, which assumes that the pre-crisis structure is more or less ready for post-crisis growth.

The graphs tell us a clear story about which policy approach was right economically - let alone socially and morally ...

It's sad in a way, seeing how self-righteous the US conservatives are about all this. They have no sense of responsibility and they have no sense of shame.

Btw., another factor that should be considered are the exchange rates.

The US was pretty much special in that regard, due to the USD being the reserve currency of the world it had an unprecedented inflow of capital during the 2008 crisis.

That strengthened the USD massively - hurting export sectors.

Europe on the other hand had a debt crisis in 2009 which weakened the EUR.

The USD was also kept artificially strong by the chinese buying tens of billions of dollars every quarter, via their massive trade surplus. That predatory mercantilist approach of the chinese kept the dollar artificially strong and the yuan artificially weak - again hurting US exports. (The weak yuan also hurts Europe and the rest of the world.)

Can you extend these graphs back a couple of decades?

My concern with these particular stats is that U.S. employees will jump through whatever hoops they need to in order to NOT lose the relationship with their employer because of the high incidence of employment-based health insurance plans.

If the "rest of the world" (or at least, the rest with some form of public health insurance or public health care) has a more pure labour market, then when declines in employment ought to happen, they do. But for the U.S., I would theorize that there was an overhang of excess employment because of the health care situation. This recession finally forced the separation of those employees, correcting the overhang. The result in recent stats would be the appearance of a sudden persistent drop in employment, that doesn't seem to be subject to the recovery.

The way to find this would be to extend these curves backward. You likely have to move the sync point as well. It might then appear that the U.S. is normal *now*, but for years had an "employment bulge", possibly not matched by GDP.

Can you extend these graphs back a couple of decades?

Sure, but then we'd be talking about long-run trends in productivity and not what happened in the recession.

eta: Anyway, the employment situation through the 2000's was pretty weak. The US never really recovered from the 2001 recession.

Clearly the jobs market remained sluggish because the minimum wage was raised.

What about outsourcing? Apple's value add is in design. The government announces a stimulus, and Apple picks up the phone and orders another million iPods from China. With a value add of $200 per iPhone sold, productivity goes through the roof, but employment doesn't pick up by all that much.

What RSJ said.

This might be a mystery empirically, with respect to the actual detailed data, but in principle I don't see the puzzle. Consider Ryan Avent, who quotes Rob Shimer as saying: "take any given group within the labour force, and the crisis has essentially generated a doubling of the unemployment rate" (http://www.economist.com/blogs/freeexchange/2011/01/labour_markets.)

One way a uniform geometric increase in unemployment could produce an unexpectedly small decline in GDP is if the original unemployment rates were heterogeneous. Suppose, hypothetically, that high-income occupations have much lower unemployment than low-income ones. Doubling unemployment for both depresses employment much more than GDP, because GDP is a function of employment, not unemployment, and employment has not been scaled uniformly.

You might reasonably ask how such a situation could come about - why don't people move from low-income occupations to high-income ones? (More employment and more pay! What's not to like?) I'll leave that for another day, except to observe that the facts suggest this transition is not so easy.

What Phil said.

I'd like to agree with RSJ about outsourcing playing a large role in the US having a different response than the rest of the G7, but why isn't this outsourcing effect seen more broadly? Are US firms the only firms that outsource, and if so, why? If you're going to rely on the outsourcing story, I think you'll need a ws and White Rabbit "Kurzarbeit" type program to explain why European nations have not outsourced their jobs.

I'm also struck by the record profits generated by US corporations last quarter as the unemployment rate hovered just below 10%. I seem to recall that the record profits were generated by two groups: either multinationals or financial firms. Growth by multinationals would be measured in the GDP, but the attendant growth in jobs would likely occur overseas.

Meanwhile, I believe financial firms contain a disproportionately high number of high-income occupations (or conversely, relatively few employees for the GDP produced, echoing Phil Koop's comment). So the growth in this sector after the sell-off in 2008 would help explain the bounce in GDP but the lag in employment.

I'm not entirely convinced by these explanations but it's a start.

i'd go kalecki here.http://delong.typepad.com/sdj/2009/11/zomfg-wtf-95-third-quarter-productivity-growth-number.html


It's not that U.S. firms are the only ones that outsource, but other nations are constrained by the current account in a way that the U.S. is not.

But I think Phil is "more right". He is making the general point that GDP = GDI, so when GDP doesn't fall as much as employment, you have the simple explanation that low wage jobs are being shed rather than high wage jobs. In that sense, my explanation is just one mechanism by which this can happen -- low wage jobs aren't needed as much to increase GDP.

But then why specifically in the U.S? The current account (and reserve status of the dollar) seems to be the clear outlier.

But you can also argue that the U.S. is an outlier in terms of labor power, and so that is the reason why G7 nations need low wage jobs in order to produce GDP whereas the U.S. does not.

I'm not sure how you would determine which explanation is right, or to what degree each is right. The two seem to go together.

It's definitely the case that job shedding in the States disproportionally impacted lower wage workers, which could have been driven by dollar power (exports got nailed, higher wage industries didn't, Apple being a key example of recent high wage success).

The difference with Europe doesn't surprise me because European social welfare systems are designed to make workers difficult to fire, which is basically a GDP / employment tradeoff: inefficiency hurts GDP, but employment is preserved close to its (already low) pre-recession levels.

My big question then is, were low-wage jobs shed disproportionately in the States as compared to Europe? If so, that's a pretty good argument for this story.

The comparison with Canada is more confusing: strong currency and dynamic labor markets, like the US, plus Canada had high exposure to resources (which tanked), so what the hell? But since Canada did the best in terms of both GDP and employment, answering why it did so well would probably tell you why its productivity was flat too.

Nathan Tankus: thank you for the link.
I think Kalecki (as told by Delong) is right.
But I would add that this works even if it is not intented by the employer.
Suppose you are an employee.
A recession begins. Your employer fires 10% of your colleagues.
You are one of the lucky 90% that can stay. What do you do?
Do you keep having a coffeebreak of 20 minutes instead of the official 7 minutes?
Are you even going to wait untill your boss makes a remark about it?
Does it make a difference if there is a safety net ?
Does it make a difference if politicians argue about reducing the safety net ?
I am not saying that one system is better than the other. I am only trying to explain the differences.
The flip side in Europe is probably higher structural unemployment.
But as a society we have a choice.

not sure of the arithmetic, but how would lower resource prices affect productivity measures?

Garrett Jones had an interesting explanation:

Workers mostly build organizational capital, not final output. This explains high productivity per 'worker' during recessions.

The recession is a period of low investment--that includes sub-normal investment in organization capital. Consequently some workers become separated. Output holds steady but employment declines. Maybe future output declines too. The organization capital otherwise produced is not measured at its true value by GDP, as only the cost of the worker appears in GDP not the surplus captured by the firm. A more comprehensive measure of output would show a larger decline, OR Output rises relative to other countries and compensates because MC declines: Low productivity workers get laid off. Investment slows, so high productivity workers get re-tasked to producing immediate output.

This latter story comports with my experience in the business world.

You can argue that it is the social safety net in other countries, but that does not quite ring true as the whole truth. Perhaps a contrarian view would be that US business operates in a different way to business in other countries. Having worked for companies around the globe this certainly as some truth to it. In response to a recession I would expect German companies to improve efficiency by by investing in new machinery and developing better products. In response to a recession I would expect French companies to improve efficiency by developing a grand plan and investing in their people. In response to a recession I would expect UK companies to focus on customer service, squeezing out costs and selective investment in business bottle necks. In response to a recession I would expect Japanese companies to focus on developing better products, investing in their people and squeezing out costs.I am not familiar with canadian companys but I suspect they are similar to UK companies. In response to a recession I would expect Italian companies to target product development and papering over the cracks until better times.

In response to a recession I would expect US companies to shed workers and shutter plants.It would be my contention that world business sees workers as a valuable asset while US business sees workers as a resource to be chopped to size. Now look at the responses and ask yourself which of these improves competitivness on a longer term basis. Having said this there are other aspects to US business which I tend to admire.

I see this as a story of openness and stimulus. All of these countries are much more open than the US.

Japan and Germany, being on the extreme end of openness, had a GDP collapse due to a trade collapse.Canada would have too had it not been bouyed by resource prices.

France and the UK both had fairly aggresive stimulus programs (well, the UK did before the austerians took over - look for the UK to fall behind the pack).

And Italy has of course just muddled along as any country saddled with that much debt must do.

And, being a recovering American, I can tell you that I'm pretty sure ws is right: in a fairly closed economy like the US with a miserable safety net what choice do workers have but to sweat blood to keep their jobs?

1. Difference in the way statistics are collected when it comes to unemployment.

2. Workers are hired when firms expect growth. When this growth does not materialize workers have to be laid off. The drop in GDP should be measured against the expectations of GDP-growth. The problem is that this is difficult to operationalize in a way so that all observers agree.

3. Openness, is not a good explanation as it could go either way.
The explanation, is not an explanation perse, merely an observation that the accounting identity holds. The question is what is the relationship that determines that the new equilibrium settles where it does.

4. The difference in social safety (counter-cyclical) net should be visible in the steepness of GDP decline. Why does Canada hold up so well compared to the other 'social safety net'-countries?

What if a shift from construction to manufacturing is a shift from low productivity employment to high productivity employment? I can easily see boeing workers building planes for exports producing a lot of output per worker. I am not so sure I understand why construction workers no longer producing houses like they did in the boom don't involve a lot of sacrific of output.

Also, layoffs continued at the usual rate for the most part (there was a big spike in the spring of 2009.) The drop in employment, however, is almost entirely a matter of the normal rate of separations, and unusually low hires.

If other countries have less churn in the labor force, the normal layoffs and restricted new hires won't cause as much loss in employment. (I don't think.)

Of course, my bottom line is that when NGDP is back on the 1984-2010 trend, come back and we will discuss employment, RGDP and productivity.

Brad Delong comments on his blog:


The answer could be very simple: one could posit that a lot of the formerly employed people added so little value to the enterprise that their departures were not felt as a decline in aggregate output. Under this assumption, remaining workers would not have to "work harder" or "work longer" to maintain aggregate output. Assuming essentially the same aggregate output by the remaining employees, the result would be (1) apparently increased productivity, even though no remaining individual was indeed more productive (2) essentially the same GDP and (3) declining employment.

Another possible explanation is that aggregate output is maintained by outsourcing to foreigners. Their production is added to the company's domestic production, but they are not added to the company's payroll. Result: it appears that fewer employed workers have been responsible for the same output as when a larger body of workers was employed.

Social welfare costs born by companies have no bearing on GDP figures, and the lack of employer-sponored welfare programs does not contribute to unemployment. The existence of such programs does help explain why companies find it relatively more costly to keep marginally productive workers on the payroll.

RSJ, you wrote: "but other nations are constrained by the current account in a way that the U.S. is not."

I don't understand what you mean by this (call me dense). Maybe I'm just not sure how the current account constrains a country like Germany? I think you're onto something by focusing on the US's exceptional current account and reserve currency status, but I can't make the connection to increased productivity. Undoubtedly, offshore outsourcing has played a role, but the US's lower labour standards must be important as well.


Comparing Canada to other G7 nations might be inappropriate given the dependency of the Canadian economy upon resources. Perhaps Australia and New Zealand should be included in the mix. Of course, those latter two economies are much more dependent upon Chinese economic growth, but their addition might be insightful.

Patrick: Brad Delong doesn't answer the question. He only rephrases it:
Why has US productivity become more countercyclical than elsewhere ?

This is not going to get better, until the conventional wisdom changes. For it is conventional wisdom which destroyed the Great American Jobs Machine. These graphs reveal how well and truly dead it is.

What we are witnessing is the agonizing death of the conventional wisdom. The fixes applied to the financial system by the Obama administration had the unfortunate side effect that the conventional wisdom staggers on, undermined by not destroyed by the unemployment crisis.

So it will only be through a war of attrition that the conventional wisdom is dismantled. How long will it take before there is legislative action wrt industrial policy, outsourcing, destructive job market effects from free trade agreements, easing unionization? Years. Yes, years. And all the while the crazies won't be going away, nor the vitriol, nor the guns. Strap yourselves in for a rough ride.

Paul Krugman gives the answer in the second paragraph of his piece on the euro:

Try charting Ireland and Spain. The hypothesis is that all three nations endured a housing bubble, and therefore, due to the inventory overhang, employment in construction is impaired.


In this simple example, the Government announces a stimulus of $600. Apple sells an additional iPhone for $600, but little additional hiring occurs since the iPhone is made overseas for $100, and apple, on the margin, only needs to hire a few more sales staff, for example. So employment goes up by 1 hour, and GDP goes up by $500.

That assumes that, on the margin, imports rise by $100 with no corresponding rise in exports. That is possible only if the rest of the world is willing to accumulate U.S. dollar assets -- the government sells $600 of bonds, of which $100 are accumulated by the exporters.

If that was not the case, then U.S. exports would need to rise by $100 (requiring additional hiring), or Apple's input costs would rise due to the depreciating dollar, (and their value add, or contribution to GDP, would fall for the same reason).

In either case, the deviation between the change in GDP and the change in employment would be less than if the nation was unconstrained in its current account.

ws: I never claimed DeLong provided an answer.

nice point about countries with housing bubbles seeing increased unemployment from the construction industry. On the other hand, didn't the UK also have a housing bubble (albeit somewhat smaller)? Their employment has had a better bounce back.

Thanks for the example. I see what you're getting at with the U.S. being able to run a higher current account deficit and therefore about to outsource. On the other hand, shouldn't surplus countries like Germany always be in the position where they can outsource? The outsourcing might even cause a depreciation of the German currency, but wouldn't this be good for sales/exports and increase GDP? So why didn't Germany exploit outsourcing over 2008-2010?

What you see in the graphs can't plausibly be explained in terms of short term interactions between labor input and GDP output. It's more like the bobbing to the surface of a larger portion of the tip of a long term 'iceberg'. The iceberg in question is the chronic miscalculation of national income and product accounts that Simon Kuznets identified in his scathing 1947 critique of the Commerce Department's first release of the series. Stefano Bartolini and Roefie Hueting have offered more recent updates of the criticism.

What Kuznets decried was the inclusion of vast swaths of what he argued were really intermediate goods and thus should not added into the accounts because that would constitute double counting. Hueting and Bartolini have each modified the double-counting charge to also include what Hueting refers to as "asymmetric entering" and Bartolini as "negative externality growth".

Although there is undoubtedly a large component of asymmetric entering or negative externality growth in the GDPs of all G7 countries, as indeed of all countries, the USA stands out as the country that has worked the hardest to make a "virtue" of social-accounting malfunction through its military-industrial and medical-pharmaceutical complexes and its horrendously bloated finance, insurance and real estate sector.

(Cross posted at Grasping Reality)

A couple of blind stabs:

Perhaps change in unemployment isn't quite the right measure. If countries with more rigid labor markets started with higher unemployment to start with then lower marginal productivity workers could already be out of work. If both overall unemployment rates and production per hour worked converged that might be plausible.

Also, the composition issue raised above strikes me as plausible if unemployment in the US is disporportionately born by lower earning workers compared to other G7 nations the GDP would be less affected than those countries.

For example, if the economy had 20 employees making 100K, 50 employees making 50K and 30 employees making 20K, GDI = 5,100,000. Unemployment is 0%

If 10 20K employees are rendered unemployed then GDI = 4,900,000 a 5.9% decrease, but a 10% increase in unemployment.

I do know that unemployment in the US is concentrated among those with lower educational acheivement, I don't know to what extent this is true of other countries. Though, if it is different, why is that?

Lastly, I assume this is real GDP, how does NGDP change vs inflation look?

David Fitzsimmons: The UK is there, and they had a housing bubble, and it looks more like France than the US.

Just to be clear, the first graph is *nominal* GDP? Or real GDP? The distinction seems important here. I suppose I would like to see both.

It's real GDP.

I've always wondered about people talking about unemployment as bad and job creation as necessary but haven't really found qualified people to ask about it. Happened on this post and I think someone here can help me understand from the level of the article and the replies.

If we can produce all the same stuff without it, shouldn't we just work less? With better infrastructure and increasingly advanced manufacturing isn't it inevitable that there won't be enough work to go around and the best we can do is try and delay this transition while we think of something?

I almost switched to a 3 day week myself to do my part. Stuff is so cheap that you really don't need a lot of money to live very comfortably. People are just insane and think they need more than a warm roof, A to B transport and good produce from the grocery store. Another way I find life very affordable is buying really high quality goods that last me decades and often buying those used.

Are there any groups working on making this transition for developed economies? The extra free time would actually increase consumption, no? And less hours billed means companies are more efficient at producing stuff to fill that time? Or am I missing one or three huge things and the world is not at all as I see it?

Obviously, the workers who are now unemployed were not very productive. Otherwise, the US would not be able to maintain it's GDP. Everyone knows someone who shows up at work every day and does nothing. Looks like US companies have dropped a lot of dead weight.

"If we can produce all the same stuff without it, shouldn't we just work less?"

Yes, yes, yes, yes, yes, yes, yes! But truth be told, the idea is heresy to economists -- heresy! -- because the thought that "there won't be enough work to go around" is a Fallacy (with a capital "f" because it's "one of the best known fallacies in economics").

There's a very old idea in economics (310 years old at least) that free trade or new machines or other improvements in productivity will always lead to more jobs in the long run because the cheaper prices will increase effective demand.

There's another idea, not quite so old, called the Jevons Paradox or rebound effect that improvements in energy efficiency will lead to more, rather than less, total consumption of fuels.

Those two ideas are actually the same idea. But one version of it implies that growth is unlimited while the other version implies definite limits to growth. So what we have here is a paradox of paradoxes. The solution to the riddle (and it's really more of a riddle than a paradox) is... we just work less.

Okay, so if working less and having goods and services to meet demand is viable, how do we get from here to there?

Right now if I tell someone "Hey, you can make the same amount but work 3/5 the time and spend the rest of it enriching yourself and your family!" just as likely as not they'd respond "Ya... or I could make 2/5th more for doing exactly what I'm used to".

How do we overcome that and is anyone planning for it at all or is it just going to be a big mess? Is it even possible to plan for?

Does working less and producing the same create economic growth through additional leisure time in which to consume?

Does it create more opportunity for interesting diversity in the market as there is a greater pool of possible entrepreneurs once free time is a common thing?

For those reasons, would countries that do this first get an edge somehow in the long run or is that same "yah, you do that, I'll keep working and get a bigger slice of the pie" logic applicable on the national scale too?

Is there a good book or three that tackles this rather than just trying to reconcile it with existing thinking and models?

Try charting Ireland and Spain.

Interesting idea. Here's employment and GDP. Haven't looked too closely at them, but Ireland is just dreadful in both graphs, and Spain isn't much better.


Well... I just happen to have the horse right here! (in the form of the manuscript of a book, no less). The book is called Jobs, Liberty and the Bottom Line and it outlines the institutional and social-accounting arrangements that could facilitate a transition to the "leisure society" (leisure being understood as self-directed activity, rather than as entertainment and idleness).


That's my long answer to questions 1 and 5. The short answer to your questions 2, 3 and 4 is: just like any innovation, it depends on how well it's designed and implemented.

Harmless: if productivity doubles, and people decide to consume the same amount of goods and work only half the time, they are not unemployed. "Unemployment" means "not working *and looking for work*"

The demand for labour halves, and the supply of labour halves too. No problem. No unemployment.

No "heresies", no "fallacies", no nothing. A doubling of productivity means people can either consume twice as much, or work half as much, or any combination in between. Historically, most people have chosen to work a little bit less and consume a lot more.

Stephen. Spain is even weirder than the US! Employment fell much more than GDP. A big increase in productivity!


I wonder if it has to do with the way those GDP numbers are calculated. I don't know much about how it's done, so this may be all wrong.

They can't be just comparing values of output between countries based on exchange rates, because that would be subject to change as currency values fluctuate for all sorts of reasons, including manipulation and speculation. Presumably, they must be trying to get some kind of "real" output, using some kind of GDP deflator or something.

So what happens if they use a deflator that includes house prices? House prices go down in the U.S. more than elsewhere, so it looks as if the "price level" has fallen or not risen as fast as it would have otherwise, Then, when they calculate the value of what is produced and "correct" it with that deflator, the value of all the other things produced looks good.

It looks to me as if Spain and Ireland might show a bit of a similarity to the U.S. GDP looks slightly less relatively disasterous than employment to me.

Nick: So really the disparity in the graph above is more of a cultural problem than an economic one?

"Historically, most people have chosen to work a little bit less and consume a lot more."

That's true -- except "history" appears to have stopped in the U.S. sometime around 1980. And it had slowed down a heck of a lot in the 1950s, '60s and '70s.

In 1958, BLS economist Joseph Zeisel called the long-term decline in the industrial workweek, "one of the most persistent and significant trends in the American economy in the last century." Yet the average manufacturing workweek in March 2010 was six minutes longer than the average had been for August 1945.

In the broader economy, if the trend of annual hours worked that had prevailed from 1909 to 1957 had persisted, average annual hours in 2009 would have been 14% lower than they were. They could have been even lower than that, taking into account the sectoral shift into the service sector (where hours are traditionally shorter) and the much increased labor force participation of women with young and school-aged children.

Robert Reich argues that it was the stagnation of wages that drove the increase in two-income families. I tend to view it as a question of the stagnation of hours driving the stagnation of wages.

Nick: I guess it does look like one could put together a story revolving around overbuilding, a collapse in construction employment and difficulties shifting employment from construction to export-oriented sectors. In both countries, the obstacle is trying to engineer the necessary depreciation. Spain is stuck with the euro, and the US is stuck with the yuan peg.

But I don't know how to get increased output/employment ratios out of that story.


Okun's Law has always been the puzzle. Why should productivity decline in a recession? The standard answer (for Keynesians and Monetarists) has always been labour hoarding. Aggregate Demand falls, firms don't need all the workers, because they can't sell the output, but firms don't fire them, because it will be difficult to hire them back again when demand recovers. So it keeps the workers on the books, but not really working.

Why should construction be different?

A1: because construction workers have general human capital rather than firm-specific human capital. So it's easy to re-hire when demand returns.

A2: because the construction firms know that demand won't return for decades.

harmless: forget the graphs. What you are talking about is the improvement in productivity that happens over decades, and how people choose to respond to it. The graphs are an economic puzzle about the short run.

Spain is really really weird. Much weirder than the US. Spanish productivity increased by about 5%-6%!!

It has to be a combination of two things: the absence of labour hoarding (that explains why productivity didn't fall); plus, construction has lower productivity than the rest of the economy(????) (that explains why productivity rises when you destroy the construction sector.

Time for a new post "Spain is even weirder than the US"?

"But I don't know how to get increased output/employment ratios out of that story."

The mystery dissolves if you distinguish between wealth, understood as goods and services produced for final consumption by individuals, and claims on wealth. Statistically, it's not easy to make those distinctions because, for example, some financial services are indeed services while others are hocus pocus (see Inside Job for examples). Back in the 1930s, when productivity estimates were first being made, the economists relied on physical outputs and it took long lead times before the data became available. Then, when the Commerce Department started issuing quarterly national income accounts, the switch was made from physical outputs to money transactions deflated by a price index, so much more timely analysis and tracking could occur.

People seems to forget that any gain in expediency is accompanied by a loss in depth and accuracy. The joke here is that however much people may want the numbers themselves to tell a story, there is also a meta-story about the numbers and how they came to have the characteristics they have. When the story the numbers tell is contradictory, it's a good time to go back and look at that other story. Here's a few suggestions:

"National Income: A New Version," Simon Kuznets, The Review of Economics and Statistics, Volume XXX, Number 3, August 1948. See especially pages 155-160.

"Objectives of National Income Measurement: A reply to Professor Kuznets," Milton Gilbert, George Jaszi, Edward F. Denison & Charles F. Schwartz, The Review of Economics and Statistics, Volume XXX, Number 3, August 1948. See especially pages 182-189.

"The Conceptual Basis of the Accounts: A Re-examination," George Jaszi, in A Critique of the United States Income and Product Accounts, NBER, pages 13-148. See especially pages 69-76.

"Extended Accounts for National Income and Product," Robert Eisner, Journal of Economic Literature, Volume 26, Number 4, December 1988, pages 1611-1684.

"The Politics of Social Accounting: Public Goals and the Evolution of the National Accounts in Germany, the United Kingdom and the United States," Mark Perlman & Morgan Marietta, Review of Political Economy, Volume 17, Number 2, 211-230, April 2005.

"Accountants and the Price System: The Problem of Social Costs," Donald R. Stabile, Journal of Economic Issues, Volume XXVII, Number 1, March 1993.

"Productivity as a Social Problem: The Uses and Misuses of Social Indicators," Fred Block and Gene A. Burns, American Sociological Review, Volume 51, Number 6, December, 1986, pages 767-780.

"Environmental and social degradation as the engine of economic growth," Stefano Bartolini and Luigi Bonatti, Ecological Economics, Volume 43, Issue 1, November 2002, Pages 1-16.

"Why environmental sustainability can most probably not be attained with growing production," Roefie Hueting, Journal of Cleaner Production, Volume 18, Issue 6, April 2010, Pages 525-530.

Chris S said: "Can you extend these graphs back a couple of decades?"

Maybe not the exact chart you are looking for but try this link with Yves Smith:


The chart is at about the 5 minute 30 second mark and titled productivity and average real earnings.

I'd listen from about the 4 minute and 15 second mark to about the 7 minute mark.

"The US is an even bigger puzzle if you think that business cycles are caused by productivity shocks."

more like "... caused by positive productivity shocks and other things that cause price deflation and not handled correctly in terms of the medium of exchange.

Yep. I think David Fitzsimmon's suggestion was a great one!

Stephen - I know you said it was "OECD data" but can you be a bit more specific? The OECD stats page never seems to want to give me monthly stats, and there is more than one variant of GDP.

I can do a bunch of the grunt work of getting stuff, but references to specific data sets, and maybe a starting URL would be helpful. Of course, if the data is the really detailed good stuff that came via a University subscription of some kind, that might mean I can't repeat these myself.

I got them from here:


No subscription necessary. The GDP data are of course only available at the quarterly frequency, and some countries - Germany, for example - don't seem to have monthly employment data.

eta: Yes, there are different variants. I used the ones that corresponded with the series StatsCan publishes.

The US, Ireland, and Spain, faced a collapse in demand due to the destruction of credit. They were operating on a model of debt-financed consumption, fuelled primarily by their housing bubbles. Their housing markets crashed and with that went the HELOCs and the investment properties, etc. For more on this, try Steve Keen. He has working dynamical models of modern economies where final demand is equal to GDP plus the change in debt.

The bottom line is that Canada is still expanding debt. The government backstops virtually all new mortgages. Hence all the warnings from Mark Carney about how 10% of the population will go under if interest rates rise 1.5%. The problem will continue to get worse every year, but eventually we will hit the same wall where a lot of consumer debt can no longer be financed, let alone repaid. Then we'll understand a bit better what happened in the US. Australia is hitting this wall right now.

One thing that will be different is that when banks foreclose in Canada, the government is automatically liable for any difference between value of the mortgage and the market value of the house. So you'd better pray the housing market does not revert to anything like the historical mean with respect to incomes or rents. If it does this country is cooked.

To me a big increase of productivity in Spain is not at all weird.
Where else do you have a 3 hour lunch break (from 2 pm to 5 pm)?
see e.g. http://gospain.about.com/od/spanishlife/f/siesta.htm
Time enough for a barbecue (and cheap wine).
What do you think the productivity is in the afternoon (i.e. after 5pm)?
Seems there is room enough for improvement.
Do you wonder why the punctual Germans are not eager to bail out Spain ?

Oooh. I think the bad accounting is the answer. The US has an especially bloated "useless" sector -- financial, military, etc. -- and an especially high tolerance for accounting fraud. In other words, the GDP numbers for the US are simply lies.

The siesta is an accomodation to the weather. What do you think productivity would be from 2-5pm given the weather? I would expect most people to spend some of that time sleeping, which kicks the productivity post siesta up.

European labour market regulation does what regulation so often does: it protects incumbents. (Young people were already significantly locked out of employment, this has got worse.) Since the risk of hiring new folk is much higher, European employers take more action to retain "known quantity" workers. US employers are much more confident about hiring new workers, so are much less concerned with retaining existing workers. So, ceteris paribus, European employment drops less in downturns. The cost is that their employment growth is (much) lower over the long term.

As for the rest, the Fed did it. The dramatic, unexpected drop in inflation made a lot of employment contracts "badly priced" in real terms and wages are sticky, but US employment is not. While the shift upwards in expectations about the future value of money in a situation of considerable uncertainty and credit problems choked off a lot of investment, further discouraging employment. In a "de-leveraging" economy with flexible employment, labour is what got shed.

This labour market segmentation in favour of incumbents is also why Europe had lower falls in employment, but a lot more riots.

America has virtually no labor movement left and very little in the way of labor protection compared to other G7 nations. I believe that a large percentage of the workforce is simply working more hours for the same pay as before. I personally know of many people who work 50+ hours when before the recession they used to work 40 hours per week. Most corporations have had layoffs, and after that happens the remaining workers will generally do whatever it takes too keep their job, especially when the labor market is so awful. This doesn't show up in BLS data because these hours are off the books. A huge percentage of the population just accepts employer wage theft, because the alternative is no wage at all. Productivity per worker is now exceptionally high, and capital is laughing at labor all the way to the bank.

There are a lot of plausible suggestions in these comments. The US is different from the other G7 countries in lots of ways. How to tell which one is right?

The only way I can think of is to increase the sample size. Stephen's new graphs, following David Fitzsimmons' suggestion, show that Spain is just like the US, only even more so. And Ireland is a bit abnormal too. What do these three countries have in common? It's not siestas, and it's not labour laws and social safety nets. I think it's a big construction slump.

See my new post.

rp: Why is the country "cooked" if banks only get current rates + 1.5% on their mortgage investments? They don't have to forclose on everyone, it wouldn't be in their best interests to cause that much instability. Just because they can raise rates doesn't mean they have to?

Would it be catastrophic if they simply treated it as if they bought rather too many long term bonds at a coupon rate below what it ultimately was but based new lending on the current prime? They might be less globally competitive but would that not be offset by the comparative stability in our economy as a haven from risk (opposite the historical view) in what would presumably still be quite tumultuous times?

"The bottom line is that Canada is still expanding debt. The government backstops virtually all new mortgages. Hence all the warnings from Mark Carney about how 10% of the population will go under if interest rates rise 1.5%. The problem will continue to get worse every year, but eventually we will hit the same wall where a lot of consumer debt can no longer be financed, let alone repaid. Then we'll understand a bit better what happened in the US. Australia is hitting this wall right now."

Sounds like somebody's listening. Today's Financial Post:
"Sources say rules now being discussed would add 100% of condominium fees to the list of expenses that is measured against income to decide whether a buyer can afford a mortgage. Currently, only 50% of the fee is considered. The move has the potential to squeeze thousands of consumers out of the market."

Or maybe the market for high rise shoe boxes will come down to a more rational level?

I like Nirad's comment.

Could it be that many Americans are working a lot of unpaid & undeclared overtime?

Also, US inflation statistics are kooky, one reason (of many) being because they take into account increasing raw processor speeds of computers without taking into account the counter-productive effects of software and operating system bloat.

There's no mystery here. Corporations simply adjust their head counts on the basis of revenues and non-payroll expenses in order to meet financial ratio targets. Revenues drop, employees get deleted from the payroll until the sales per employee ratio meets the target value.

And there was a huge spike in energy prices here in the USA in 2008, right around the time the employment started to drop precipitously. Fuel costs rise - to balance out the energy cost increase, workers are deleted from the payroll. Fuel costs rise, automobile sales plummet - auto workers are laid off.

There's no mystery here - you're simply not looking at all the time series or taking into account the feedback loops by which corporations are managed. Was there a huge spike in energy prices in the other G7 nations in 2008, or was the USA the only one hit?

Energy prices rose everywhere; we all buy oil from the same world market.

As predicted, this post has generated some noise in the US blogosphere, but many seem to be missing Nick's point. For example, what is so special about the US that it has zero marginal product workers while that's not the case elsewhere in the G7?

The graphs paint a very simple picture to me; we Americans have trimmed payroll fat that was contributing little or nothing (in aggregate) to our GDP. True, fewer workers are doing more just to sustain output but the disparity in GDP vs. employment shows that it was far from equilibrium going into he recession. If it were, a precipitous drop in employment would have had a lock-step drop in output...but it didn't so our companies (not necessarily the country) are better off (in aggregate). As for those who lost their jobs - what a great chance to re-train, start a business, invent something....innovate!!!!

That answer begs the question of why that didn't happen in other countries. Moreover, US employment growth in the decade preceding the crisis was fairly anemic; hard to see how much fat could have accumulated.


The post is about US productivity growth. Yet only the two components of a an aggregate GDP/labour productivity ratio are furnished in chart format. I guess blog-discussion participants are sufficiently numbers-experienced to readily estimate-imagine the 3rd latent chart. As for non-economists?

Would be interesting to see similar numbers for the USA broken down by sector (business, manufacturing, and public).

Take cover! Krugman link incoming

Here's a colleague that argues that the 1981-82 recession is not any different that 2007-09 recession in terms of labor productivity.


The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad