Paul Krugman’s Free Lunch Theory
There is no cause to worry. The high tide of prosperity will continue
—Andrew W. Mellon, Secretary of the Treasury - September, 1929
The worst is over without a doubt
—James J. Davis, Secretary of Labor - June, 1930
The depression has ended
—Dr. Julius Klein, Assistant Secretary of Commerce - June, 1931
A serious dispute has broken out among economists about fiscal deficits in the United States. Paul Krugman has been making the rounds on the talk shows, arguing that more stimulus spending is required to spur job creation. From his lofty perch at the New York Times, Krugman preaches the Gospel of Spending to anyone who will listen.
[My note: link  goes to Krugman’s The Burden of Debt, which I cite below.]
I see things differently too. I am not going to review Hamilton’s cogent arguments, which you can read yourself. In the wrangling about what is politically possible, bond vigilantes and future interest rates, which determine the cost of servicing the debt, our prospects for economic growth are left unexamined.
But Paul feels I’m using an inappropriate metric [for paying off the debt, and then quotes Krugman]:
Jim gets scary numbers about the debt burden by assuming that we’ll have to pay off the debt in 10 years. But why would we have to do that? Again, the lesson of the 1950s– or, if you like, the lesson of Belgium and Italy, which brought their debt-GDP ratios down from early 90s levels– is that you need to stabilize debt, not pay it off; economic growth will do the rest.
Normally, you’d think that putting off repaying a debt does not make it any smaller…
Perhaps Paul is suggesting that there may be a potential free lunch available from postponing payment in this case, arising from the fact that the economy’s growth rate has historically exceeded the government’s cost of borrowing. If I put off paying another year, with interest the amount I owe grows 2% in real terms, but my income grows 3%, so things get easier the longer I put it off.
Let me go on the record as favoring the consumption of truly free lunches. If everything the government buys really is free, then by all means, let’s have them buy more, and more, and more, and rather than double my taxes, let’s cut taxes all the way to zero…
Krugman’s “free lunch” requires “above trend” future economic growth as Hamilton explains. Such growth can not be taken for granted; this assumption requires scrutiny. Here’s the relevant text from Krugman’s debt burden article—
So: in 2008, with revenues already depressed by the recession and housing bust, the federal government took in $2.5 trillion in revenues. If we assume 2.5% real growth* and 2% inflation, by 2019 that would rise to $4 trillion. So debt service costs due to the next decade’s deficits would be less than 6 percent of revenue under current law.
*Contrary to what some think, we’d actually expect growth over the next decade to be somewhat above trend, as the economy picks up some of the current slack. That’s what the historical record tells us actually happens.
Krugman also made his assumptions explicit on ABC’s This Week with George Stephanopoulos. Here Krugman has an exchange with another George, the bow tie George Will.
KRUGMAN: We’re not going to hit 100 percent [debt/GDP] until a decade from now. And countries have gone above 100 percent. I mean, if you actually ask about the interest cost, particularly inflation-adjusted interest cost, you know, we’re now paying 1.2 percent real interest rate on federal debt. Even if you add 50 percent of GDP in debt, which I don’t think is going to happen, that’s still only a fraction of a percent of GDP in additional debt service costs.
WILL: But even [with] unreasonably cheerful assumptions about economic growth and interest rates, we’re apt to be spending in 10 years $700 billion a year servicing our debt.
KRUGMAN: That’s in a $20 trillion economy. It doesn’t sound as bad as it is.
[My note: Either the transcript is wrong, or Krugman is guilty of a Freudian slip. I think he meant "it isn't as bad as it sounds."]
What $20 trillion dollar economy is Krugman referring to? He is giving us his estimate of GDP in 2019. I will (generously) assume he is speaking of nominal GDP, which stood at $14.266 trillion in the 3rd quarter this year. Real (inflation adjusted) GDP was $12.990 trillion in chained 2005 dollars. Krugman thus assumes that the economy will grow about 28.7% over the coming decade. This requires annual nominal GDP growth of 3.4% starting in 2010 and ending in 2019. What an astonishing assumption!
It is as if Krugman regards the economy as a quasi perpetual motion machine whose gears freeze up once and a while. If we apply a liberal dose of lubricant to grease the wheels, the machine easily starts up again, cranking out GDP & jobs regardless of our present circumstances and our past mistakes.
The phrase “free lunch” can be defined like this—
The economic theory, and also the lay opinion, that whatever goods and services are provided, they must be paid for by someone - i.e. you don’t get something for nothing. The phrase is also known by the acronym of ‘there ain’t no such thing as a free lunch’ - tanstaafl.
The phrase has it’s origins in a 19th century bar practice used to lure morning drinkers—
Free lunch was a commonplace term in the USA and, to a lesser extent in Britain, from the mid 19th century onward. It wasn’t used to describe handouts of food to the poor and hungry though, it denoted the free food that American saloon keepers used to attract drinkers. For example, this advertisement for a Milwaukee saloon, in The Commercial Advertiser, June 1850:
At The Crescent… Can be found the choicest of Segars, Wines and Liquors… N. B. - A free lunch every day at 11 o’clock will be served up.
Free lunches, often cold food but sometimes quite elaborate affairs, were provided for anyone who bought drink.
The lunch is free, but the drinks are not. Thus, you don’t get something for nothing. But in Krugman’s world you do get something for nothing, sort of. We can pile on the debt to restart the economy because future growth will pay for it. (Well, OK, not quite all of it, but why split hairs?) This akin to perpetual motion.
Free lunches are quite popular although, like Unicorns, they do not exist. For example, when CERA tells you that
- [Oil] supply evolution through 2030 is not a question of resource availability.
- IHS CERA projects growth of productive capacity through 2030, with no peak evident
they are actually telling you that you can eat & drink for free today and every day until 2030. As Hamilton notes in his response to Krugman, let me go on the record as favoring the consumption of truly free lunches. Who doesn’t want a truly free lunch? I’m sure CERA report sales are brisk.
Why might I think that the kind of growth Krugman envisions is impossible? Rather than resort to subtle arguments, I will deploy overwhelming brute force.
This is an economy & a society*
- whose automobile industry is a shambles
- whose manufacturing base has been in decline for decades
- where “full” employment (~5%) likely will not return until sometime in the 2016-2020 period (see Figure 4 and the discussion below)
- where non-wealth-creating consumption (PCE) makes up about 71% of GDP
- whose total (public & private) debt-to-GDP ratio is north of 350% and is now rising faster than defaults, asset sales or payments can pare it down due to increased public debt (see Figures 1, 2, and 3 and the discussion below)
- where real income actually declined over the last decade
- where “over the past 10 years, the private sector has generated roughly 1.1 million additional jobs, or about 100K per year. The public sector created about 2.4 million jobs.”
- where home prices, which were the principal source of household wealth in a bubble economy, have likely not yet hit bottom
- where household wealth has fallen precipitously as a result
- whose overbuilt residential investment sector (new housing) will stagnate for many years to come
- where there are 23.1 square feet of shopping center space for every American
- where rising health care, food and energy costs are eroding household disposable income
- where over 45 million Americans lack any kind of health insurance
- where the cost of an college education has increased far in excess of the inflation rate for decades
- where wealth & income inequality is at its highest level since 1928
- where 49 million Americans went hungry at some point in 2008
- where 1 in 8 Americans and 1 in 4 children receive food stamps
- where total bank credit of all commercial banks is contracting, not expanding
- where several regional banks fail each week
- where the government insurance system for depositors (FDIC) is insolvent
- where “enormous budget deficits in nearly every State in the Union are ‘wreaking havoc’ on government employees, the services they provide, and the residents who need them most”
- where the Central Bank (aka “the Fed”) is accountable to no one, and is thus free to print money for any purpose it deems important, including monetizing the debt or bailing out too-big-to-fail (TBTF) banks
- where Congress has been “captured” by those they must regulate across all industries, and so does literally nothing or practically nothing in the name of reform
- where the Executive branch of government (i.e. the Treasury) mostly serves the interests of highly over-leveraged TBTF banks
- whose politically powerful TBTF banks drain huge amounts of capital away from wealth-creating activities
- whose negative current accounts (trade) balance exploded during the years leading up to the financial crisis and still shows a substantial deficit (exports versus imports)
- whose declining oil production peaked in 1970
- whose concomitant dependency on oil imports has been rising for decades (see trade balance)
- whose government debt must be funded by Asian & European central banks and oil exporters
- which is involved in one costly but pointless occupation (Iraq) and one pointless but costly foreign war, and is on the verge of expanding the war (Afghanistan)
* Not all items are documented (linked), but ample documentation could be provided
That’s enough—you get the idea. The list is not comprehensive. I didn’t stop because I ran out of material. Some indicators more time-sensitive than others, so improvements will vary over time if the situation is reversible. For example, home prices should bottom out sometime in the next few years. On the other hand, nothing will fix our declining oil production. Some items are sensitive to political circumstances, e.g. the non-accountability of the Fed.
Other indicators, for example those illustrating growing poverty, may not directly affect GDP growth. These points are meant to show a society going rapidly downhill.
A New Household Spending Spree—Why Not?
Recall that personal consumption expenditures (PCE) make up 71% of GDP (item #4 on the list). For the economy to grow 28.7% in the next decade consumption must keep up. Look at Figures 1-3.
Figure 1 — Private debt levels in the United States as a percentage of GDP from Credit Suisse. (You can safely ignore the Credit Suisse contention that debt levels in the United States are nothing to worry about.)
Figure 2 — Household debt levels in the United States as a percentage of GDP from Credit Suisse as above. The possible scenario indicates household de-leveraging accompanying a rise in the savings rate.
Figure 3 — Debt-driven contribution to demand (left-scale, red line) and unemployment (right-scale, inverted, blue line) from Australian economist Steve Keen’s 4 Years of Calling the GFC, published December 1, 2009 at his website Debt Watch.
Taken together, the 3 graphs clearly show that—
- private debt levels, including household debt, rose precipitously from the early 1980s on to support the rapid economic expansion. (For example, real GDP grew 22.4% in real terms during the Bubble Era 1998-2007; growth in nominal terms was 37.5%.)
- “debt has little impact on demand when the debt to GDP ratio is low–such as … in the US from the start of WWII till the early 60s. But whenever the debt to GDP ratio becomes substantial, changes in debt come to dominate economic performance [as reflected in unemployment in Figure 3]” (quote from Steve Keen)
You can see in Figure 3 that the debt-driven component of demand has fallen off a cliff and is now negative for the first time since the years immediately after World War II. In Figure 2, you can see the reasonable de-leveraging scenario whereby households repair their balance sheets over the next several years—assuming, of course, that the household is still intact, the adults in the household are still working, and payments on the mortgage are not delinquent (renters need not worry about the latter).
From a debt perspective, the prognosis for expansion of personal consumption expenditures in future years is grim. The huge loss of wealth Middle Class Americans incurred when house prices collapsed will dampen consumer spending for years to come. Finally, the Baby Boomers, the most spendthrift generation in history, will start retiring in 2011. Over half of the Boomers are fiscally unprepared for retirement.
Yet Paul Krugman foresees GDP growth averaging 3.4%/year which gives us a $20 trillion economy (nominal GDP) in ten years. It would appear that Krugman applies “free lunch” theory not only to public debt, but also to private debt accumulated during the Great Moderation by businesses and households.
In so far as personal consumption expenditures are 71% of GDP in real terms, I must ask Krugman the now-famous question Barney Frank asked a heckler at a townhall meeting: Paul, what planet are you living on?
When Will Full Employment Return?
It’s mighty hard to jack up your personal consumption when you’re unemployed or underemployed. As everyone knows by now, the “official” unemployment rate stands at 10.2% (the U3). The U6 measurement stands at 17.5%. This number includes involuntary part-time workers and the marginally attached. From the Bureau of Labor Statistics—
Marginally attached workers are persons who currently are neither working nor looking for work but indicate that they want and are available for a job and have looked for work sometime in the recent past. Discouraged workers, a subset of the marginally attached, have given a job-market related reason for not looking currently for a job. Persons employed part time for economic reasons are those who want and are available for full-time work but have had to settle for a part-time schedule.
[My note: "have a job-market related reason for not looking for work" — meaning they have completely given up looking for work because there's no hope of finding it.]
Since Krugman foresees an enormous economic expansion in the next decade, it behooves us to ask when “full” employment may return. For the sake of argument, let’s set the “full” U3 employment number at 5%. Please bear in mind this important statistic cited by Ben Bernanke in 2003—
Because new workers are always entering the labor force, the U.S, economy needs to create something on the order of 150,000 net new jobs each month just to keep the unemployment rate stable.
When can we expect to see this (or a higher) level of job creation?
That’s a good question, Ben. In item #3 on The List, I speculated that full employment might return sometime in the period 2016 to 2020. How did I come up with those dates?
The earlier date, which is a mere 7 years from now, derives from the Wall Street Journal’s Scarred Job Market Expected to Weigh on Economy. This article was written on October 8, 2009, before the 3rd quarter preliminary GDP number came out (+3.5%) and was subsequently revised down (+2.5%). It was also written before the “official” unemployment rate hit 10.2% that same month.
The 48 surveyed economists expect the economy to bounce back from four quarters of contraction with 3.1% growth in gross domestic product at a seasonally adjusted annual rate in the just-ended third quarter.
Expansion is seen continuing through the first half of 2010, though at a slower rate. But the massive downturn means the labor market will take years to heal. On average, the economists don’t expect unemployment to fall below 6% until 2013…
“Never before has business shed so many workers so fast, so many people failed to find work who are looking for work, and so many dropped out of the labor force as in the current circumstance,” said Allen Sinai at Decision Economics….
On average the economists — not all of whom answered every question – expect the unemployment rate to peak at 10.2% in February . But even once the employment situation stops getting worse, economists expect recovery to come slowly. “It could take until 2014-15 before we see a 5% handle on unemployment again,” said Diane Swonk at Mesirow Financial.
In line with their overly rosy view of when joblessness would hit 10.2%, the economists surveyed take a more optimistic view of when unemployment will fall below 6% than the situation warrants. The outlook is slightly worse than Diane Swonk predicts if everything that can go right does go right (Figure 4).
Figure 4 — From the WSJ article. Unemployment was about 5% when the recession started. How long would it take to return to the employment levels we had at the start of the downturn? It will take until the end of 2016, but only if the economy starts adding jobs immediately at a furious pace well above the 150,000 new jobs we need every month just to stay even.
The December, 2016 date shown in Figure 4 may be too optimistic. In Welcome to the New Normal, John Mauldin analyzes past jobs data to determine how many jobs the economy must add each month to get to 5% unemployment by 2014.
Let’s assume that we would like to get back to a 5% unemployment rate. That would not be stellar, but it would certainly be better than where we are today. Five percent unemployment in late 2014 will mean 8.1 million unemployed. To get to 5% unemployment we will have to create 14 million jobs in the five years from 2010-2014. (163 million in labor pool minus 8 million unemployed is 155 million jobs. We now have 139 million jobs, so the difference is roughly 15 million.) Plus the equivalent of 3 million jobs that Rosenberg estimates, just to get back to an average work week. And maybe the extra 1.5 million a year I mentioned above.
But let’s ignore those latter jobs and round it off to 15 million. Let’s hope that by the beginning of next year we stop losing jobs. That means that to get back to 5% unemployment within five years we need to see, on average, the creation of 250,000 jobs per month. As an AVERAGE!!!!!
[My note: look at Mauldin's complex analysis to get the full context of the quote.]
And has the economy ever added an average of 250,000 jobs per month in any of the last 10 years? NO!!!!!
If you take the best year, which was 2006, you get an average monthly growth of 232,000. If you average the ten years from 1999, you get average monthly job growth of 50,000. If you take the average job growth from 1989 until now, you get an average of 91,000 a month. If you take the best ten years I could find, which would be 1991-2000, the average is still only 150,000. That is a long way from 250,000.
To make a long story short, Mike Shedlock (Mish) of Global Economic Analysis takes Mauldin’s review as a starting point and concludes that it may be 2020 before the U.S. reaches 5% employment, assuming everything may things go right. That’s where I got the higher date in item #3 on The List. Given the staggering jobs growth required, and the economic challenges ahead, the 2020 date does not seem out of line.
From a jobs perspective, Krugman’s assumption of real (inflation-adjusted) 2.5% annual GDP growth over the next 10 years seems ludicrous. The Princeton economist is very concerned about unemployment, as he should be if his rosy visions of future GDP growth are to come to pass. In The Jobs Imperative, Krugman writes—
If you’re looking for a job right now, your prospects are terrible. There are six times as many Americans seeking work as there are job openings, and the average duration of unemployment — the time the average job-seeker has spent looking for work — is more than six months, the highest level since the 1930s…
Yes, the recession is probably over in a technical sense, but that doesn’t mean that full employment is just around the corner. Historically, financial crises have typically been followed not just by severe recessions but by anemic recoveries; it’s usually years before unemployment declines to anything like normal levels. And all indications are that the aftermath of the latest financial crisis is following the usual script. The Federal Reserve, for example, expects unemployment, currently 10.2 percent, to stay above 8 percent — a number that would have been considered disastrous not long ago — until sometime in 2012…
So it’s time for an emergency jobs program… Meanwhile, the federal government could provide jobs by … providing jobs. It’s time for at least a small-scale version of the New Deal’s Works Progress Administration, one that would offer relatively low-paying (but much better than nothing) public-service employment. There would be accusations that the government was creating make-work jobs, but the W.P.A. left many solid achievements in its wake. And the key point is that direct public employment can create a lot of jobs at relatively low cost. In a proposal to be released today, the Economic Policy Institute, a progressive think tank, argues that spending $40 billion a year for three years on public-service employment would create a million jobs, which sounds about right.
[My note: If we are in a Depression, and I think we are, then I support some kind of WPA project to get people working. It sure beats Citigroup debt guarantees. I do not lack compassion except where criminally negligent bankers are concerned.]
Krugman is clearly aware that sustained, high unemployment will hamper his immaculate recovery. Let’s review. The free lunch theory goes like this:
- Massive stimulus (debt-financed government spending) is required to grow GDP and create new jobs
- Such spending will enable 2.5% real GDP growth in the next decade
- Servicing the resultant multi-trillion dollar debt will not be a problem because the economic expansion will boost tax receipts, which will obviate the debt-servicing problem and “stabilize” the debt (see Hamilton’s article cited at the top). Throw in 2% annual inflation, and everything should work out just fine.
Are We In A Depression?
The phrase “economic Depression” is not well-defined. We only have one good example in the 20th century, and it’s clear that what is happening now is not the same as what happened in 1930-1933 (e.g. the TBTF banks did not fail because they were bailed out, we are not in a deflationary spiral due to massive government intervention, etc.). However, a Depression might be minimally defined as a prolonged, painful economic event preceded by unsustainable debt levels and characterized by sustained high unemployment.
We are in some new kind of Depression or we are not in one as defined above. You can’t have it both ways. See Dave Rosenberg’s How Is This Not a Depression? for one view of the situation. I agree with much of his assessment.
Expansion of the Fed’s balance sheet and the $787 billion stimulus have not prevented terrible economic conditions. And we can’t ignore the trillions of dollars that have been spent to stabilize Big Finance. Monies spent on the banks and via the fiscal stimulus are said to have averted a second Great Depression.
Our failure is clearly indicated by the Chicago Fed’s National Activity Index (CFNAI-MA3, Figure 5).
Figure 5 — This is a three month moving average (MA-3) “The index is a weighted average of 85 indicators of national economic activity. The indicators are drawn from four broad categories of data: 1) production and income; 2) employment, unemployment, and hours; 3) personal consumption and housing; and 4) sales, orders, and inventories.A zero value for the index indicates that the national economy is expanding at its historical trend rate of growth; negative values indicate below-average growth; and positive values indicate above-average growth. When the CFNAI-MA3 value moves below –0.70 following a period of economic expansion, there is an increasing likelihood that a recession has begun.”
Here is the relevant passage from the Chicago Fed on determining the end of recession–
When the economy is coming out of a recession, the CFNAI-MA3 moves significantly into positive territory a few months after the official NBER date of the trough. Specifically, after the onset of a recession, when the index first crosses +0.20, the recession has ended according to the NBER business cycle measures… The critical question is: how early does the CFNAI-MA3 reveal this turning point? For four of the last five recessions, this happened within five months of the business cycle trough.”
Figure 5 shows that the CFNAI-MA3 indicator is now going South again when it should be headed North. We’re nowhere close being out of this downturn.
Krugman wants a revival of FDR’s Works Progress Administration (WPA). That would seem to indicate that he thinks we’re in a Depression. But if we’re in a Depression, it would seem that no amount of government spending, absent a mobilization for World War III, is going to enable the kind of GDP growth Krugman envisions. Reviving the WPA is an act of desperation. It is the last resort, a last ditch effort. This is what you do when all else has failed.
Not so, says Krugman. We just haven’t spent enough.
I think Krugman’s position is incoherent. A Depression, however it is precisely defined, is a prolonged, painful economic event. That is what distinguishes it from a normal business-cycle recession. If we’re in a Depression, and we need a WPA, it also seems clear that we’re not going to get 2.5% real annual GDP growth over the next decade. If we’re not in a Depression, then why do we need a WPA?
Apparently, Krugman’s answer is that there is indeed such a thing as a free lunch. We can get both free drinks—unlimited deficit spending—and free food—vigorous GDP growth thereafter which “stabilizes” the debt and does away with the problem of servicing (paying the interest on) the debt. I do not believe his economic growth story. If only reality worked the way Krugman thinks it does.
Contact the author at firstname.lastname@example.org