Pic of the day

I find this pretty remarkable, though it's somewhat unfair to assign credit or blame to a president for the macroeconomic performance that occurs on his watch. Still...



For the record, Obama is currently at -2.3%, but +1.0% in 2010. I know you would have guessed that G.W. Bush would have the worst record of private sector job creation, but would you have picked Carter for the best?

The seasonal job surge

Casey Mulligan has a provocative piece in the NY Times (here, here). He shows that teenage employment surges in the summer when school lets out as shown in the graph below. This isn't really an astounding finding, but it does seem to imply that a shock to labor supply (the school year ends, making kids available for employment) can affect employment even in the middle of a severe recession. This implies that if government efforts to stimulate demand come with impediments to labor supply they may end up being ineffective or counterproductive. Mulligan makes exactly this claim about ARRA, though what these restrictions are isn't exactly clear.

The argument is interesting, but there's an obvious demand-side explanation for the surge in teen employment during the summer. Summer brings demand for lifeguards, amusement park ride operators, lawn care workers, house painters and other types of jobs typically taken by teenagers. Mulligan's a smart guy, surely he could figure out how much of the surge is due to the opening up of this type of employment in the summer months and how much is due to the release of workers.

[Note: Ryan Avent makes the same argument as me here, with some data.]

A plan to reinvigorate the economy

I approve of Joe Gagnon's approach:



- The Fed launches a new $2 trillion round of quantitative easing with a commitment to further action if the economy remains weak



- The Fed lowers the interest paid on reserves to zero



- The Obama Administration instructs Fannie Mae and Freddie Mac to allow all homeowners current on their mortgage payments to refinance regardless of loan-to-value ratio



- The Obama Administration makes a strong effort for mortgage modifications for distressed mortgage borrowers



- The Treasury renounces its "strong dollar" policy, allowing a broad-based depreciation.



I endorse Paul Krugman's caveat: the Fed should not reassure markets that it will switch course if inflation rises above 3 percent. In fact, the Fed should promise that it will continue to make purchases as long as inflation is below 3 percent for a few years.



I also support the establishment of an infrastructure bank in legislation attached to the highway bill that is under consideration in Congress, and an extension of the payroll tax holiday and unemployment compensation.



Everything but the last three proposals can be done by the Obama Administration and Fed alone, without having to go through Congress. For the last three, Obama needs to be much more specific and forceful than he was in his speech yesterday. The message ought to be, if Congress does not act on jobs, it - not he - will be held responsible for continued high unemployment in 2012. If you can't get the legislation, take the issue.

Initial claims stay flattish

Initial claims for unemployment compensation dropped a bit this week, but continues on a rather flat trajectory.

Initial claims have been drifting down very slowly since the beginning of the year and have been essentially flat the last three months. I've said before that historically large gains in employment have come several months before initial claims fall below 400,000. On the other hand, it seems unlikely that we'll see big gains in employment if initial claims aren't falling. So modest growth in payroll employment (+50-100,000 private) seems a good bet to me for the month of July. I don't know what's going on with Census employment this month so the overall numbers are hard to predict.

What do we do about a slowing economy?

My rosy forecasts for economic growth coming out of this recession are not looking so good right now. Employment growth in March and April suggested a strong recovery. May and June's numbers, however, provide unmistakable evidence that the economy is slowing down. Private sector employment increased by 33,000 in May and 83,000 in June, well below the pace needed to make a dent in unemployment. Hours worked increased by only a tiny bit in June, so that total growth in hours in the second quarter is only 2.4 percent, the same as in the first quarter. Hence my best guess is that GDP grew no better in the second quarter than the first (3-3.5 percent, versus my guess of 4.5-5 percent based on April-May's data). What's going on out there?

Well, it seems as if in the race between the forces of recovery (pent up demand, expansionary monetary and fiscal policy) and the forces of contraction (consumer debt, oversupply of housing, pressure on state and local budgets, general lack of confidence) the forces of contraction have taken the lead. Awhile ago I thought that the prudent course of action was a "small" (on the order of $100 billion or so) stimulus focused on aid to the unemployed and state governments to guard against a slowdown. Now the slowdown seems to be upon us, anyone with a brain would agree that a stimulus package of that size or larger is the way to go. Unfortunately, people with brains seem to be very scarce in the policymaking and opinion-forming communities.

Paul Krugman and Brad DeLong have written several pieces lately asking how it is that Keynesians seem to have lost the argument on policy to the deficit hawks. The best answer seems to me to be the attractiveness of false prudence in times of crisis. We got into this mess because of over-indulgence. The cure must be to tighten our belts and accept our punishment. Reduce the budget deficit, put a halt to the rise in debt, take our lumps, and things will be all right. Well, that is clearly nonsense: the problem we face right now is that people and businesses prefer to hoard money rather than spend it. The solution must be to stimulate spending, which means that the government needs to borrow more, not less. But something about calling for more "irresponsibility" on the part of government at this time is deeply unattractive to an important segment of our polity (largely, though I suppose not exclusively, that segment that is securely employed such as politicians and journalists). And so here we sit, unable to do the obvious things like extend benefits to the unemployed and states that might get us out of this mess.

When people try to come up with a logical argument for austerity they focus on the effects of austerity on "confidence". David Brooks's column in today's NY Times is a particularly asinine version of this line of reasoning. First he paints an absurd caricature of proponents of stimulus (without naming anyone or citing specifics, of course) as arrogant and overly confident in their theoretical models. Nothing could be further from the truth: we all admit that we're in uncharted territory here (well, territory that has been charted only a couple of times in history) and allow for a wide range of uncertainty in our prescriptions). Then he idolizes the non-egghead, commonsense man on the street who somehow has accumulated more wisdom about the workings of the macroeconomy than the so-called "experts". And he summarizes their wisdom thusly:

You can't read models, but you do talk to entrepreneurs in Racine and Yakima. Higher deficits will make them more insecure and more risk-averse, not less. They're afraid of a fiscal crisis. They're afraid of future tax increases. They don't believe government-stimulated growth is real and lasting. Maybe they are wrong to feel this way, but they do. And they are the ones who invest and hire, not the theorists.

Is there a word for arrogant anti-elitism, the flaunting of ignorance as a badge of honor? Brooks is bathing in this, whatever it is. Would it do any good to point out that there is no evidence that business people are not investing or hiring because of fear of deficits or future tax increases (as opposed to the fear that no one is going to buy their stuff)? Probably not, because reliance on "facts" and "evidence" is too eggheady for the likes of Brooks. But for the record, if people with money were really afraid of a fiscal crisis, they would hesitate to buy government debt. The government would have to pay a premium to borrow - we'd see a big spike in interest rates. The truth? The interest rate on 10-year government bonds is 3.2 percent, which is ridiculously low. The expected inflation rate implied by the difference between the interest rate on regular government bonds and bonds that are indexed to inflation (TIPS) is about 2 percent, the same as it's been for decades.

Even renowned economists are capable of spouting nonsense about the advisability of stimulus. Ed Glaeser counseled in today's Times (the link has disappeared - perhaps he had second thoughts) against the government spending money on infrastructure and other things that we don't need. A waste of resources, he says, echoing the austere gentlemen who ran the British Treasury in the early 1930s as the British economy descended into a death spiral. What is a greater waste of resources - millions of people unemployed, or using some of those unemployed to build highways?

Government spending, either directly or through aid to states and the unemployed, is an essential part of the answer to the problem we face. The Obama Administration needs to make this its number one priority this summer. The Senate has to be convinced to take action now - otherwise the Democrats are doomed, and a Republican dominated legislature is bound to pursue the wrongheaded (but strangely morally satisfying) policies next year that will extend our economic misery for years to come.

Should we panic about jobs?

Analysts and stock markets around the world are clearly panicked about the disappointing May jobs numbers. Paul Krugman is warning once again about a "lost decade" scenario as a result of the G-20's apparent determination to begin a fiscal retrenchment before the recovery has found its legs. Commentators on CNBC Friday took the jobs report as evidence of a jobless recovery.

One element of the report that has not received as much attention as it should, however, is the increase in hours worked. The BLS's measure of aggregate hours worked increased by 0.3 (in May, below the increase of 0.4 in March and April but still not bad. For the first five months of 2010 hours worked has increased at an annual rate of 3.4 percent. This is hardly a blistering pace, but it is significantly better than what the US experienced after the 1990-91 and 2001 recessions. Following the 2001 recession, for example, the US did not achieve a sustained pace of hours increases at this level until early 2006. We are certainly not experiencing the kind of rebound we did in 1983-84 when hours worked in some quarters rose at 8 or 9 percent, but 3.4 percent is still consistent with a sustained recovery. Add a guess of 1.5% for productivity growth, and we're looking at GDP growth of 5 percent - in the same range as what the ISM numbers indicate as noted in an earlier post.

One thing that appears to be happening is that the increase in hours worked is affecting the average work week more than employment; as the work week hits normal this summer we should see a pickup in job creation. The big decline in workers working part time as reported in the household survey suggests a shift from temporary work to full-time work. This is obviously good news for the economy, but it does not show up in the headline employment numbers.

That said, however, there are real concerns as to whether the recovery is sustainable at this point given what's happening in Europe. And the economy is in such a deep hole that we really shouldn't be satisfied with GDP growth in the 4-5 percent range, if that's what we're looking at now.

The ADP numbers are missing the increase in BLS employment

ADP reported today that private non-farm employment increased by 55,000 in May. ADP is widely thought of as a good predictor of the BLS figure which comes out tomorrow. But lately the ADP numbers have been way off, as the graph below shows. Last month, for example, ADP showed an increase of 32,000 - two days later the BLS came out with an increase of 231,000. Since January ADP has understated the BLS numbers by about half a million jobs. The difference between ADP and BLS numbers appears to be larger and more persistent now than it was in 2009; I wonder why. At any rate, the fact that ADP is predicting sluggish jobs growth is not convincing evidence that the BLS number will not be large.

Simple extrapolation again

Calculated Risk previews Friday's jobs report. The consensus forecast is for an increase in employment of around 540,000, about 400,000 of which are temporary Census jobs (the Census will start cutting employment in June, so from then on Census hiring will reduce the employment numbers. Since government employment has been roughly stagnant for the past year, the experts are essentially forecasting an increase in private sector employment of 130,000 to 150,000.

The 130-150,000 forecast is an argument that the current recovery will look something like the recovery that began in 2003. The experts could be right. But simple extrapolation of the trend begun last spring would suggest that private sector employment will rise by 60,000 more than it rose last month, or roughly 290,000. That would give us a headline number of close to 700,000, plus or minus whatever happens with non-Census government employment. Is there any good reason to believe that private sector employment grew more slowly in May than in April? I don't see any, so I'll go with +290,000 private sector, +700,000 overall.

The Republicans' jobs proposal

Paul Krugman and Ezra Klein are right, the Republican party is really bereft of ideas. Here's the Republicans' new jobs plan (the one they've rolled out to make everyone forget last week's Medicare plan). A summary:

Empower Small Business Owners and Reduce Regulatory Burdens:

• Require congressional review and approval of any government regulations that have a significant impact on the economy or burden small businesses.
  


• Audit existing and pending regulations to identify and address those that hinder economic growth.

Fix the Tax Code to Help Job Creators:

• Increase American competitiveness to spur investment and create more American jobs by streamlining the tax code and lowering the tax rate for businesses and individuals including small business owners to no more than 25%.
  


• Reform the tax code to allow American businesses to bring back their overseas profits without having to pay a tax penalty so they can invest in our economy and create American jobs.

Increase Competitiveness for American Manufacturers:

• Pass the three pending free trade agreements with Colombia, Panama, and South Korea to create up to 250,000 jobs.
  


• Continue to open new markets to American made products.

Encourage Entrepreneurship and Growth:

• Modernize our patent system to protect our nation’s innovators, discourage frivolous lawsuits, and expedite patent reviews.
  


• Re-Authorize and improve federal programs and approval processes to streamline development of new products. Remove barriers to building a first class workforce so that the United States can compete in the global marketplace and lead the way in technological development and growth.

Maximize Domestic Energy Production to Ensure an Energy Policy for the 21st Century:

• Promote lower energy prices through increased domestic production. Encourage all forms of energy production.

Pay Down America's Unsustainable Debt Burden and Start Living Within our Means:

• Build upon the House Republicans’ Budget by enacting significant spending cuts.

Only #2 (tax cuts) is what I would call a major policy proposal. It might have some impact on employment and growth in the long run, but it's not going to help the economy here and now. None of the other proposals could be expected to have a substantial impact on employment (the one number cited, 250,000 jobs created from trade pacts, is the equivalent of one good month's job creation; a drop in the bucket). The last element (spending cuts) would surely reduce employment considerably.

A serious proposal to stimulate growth now would include: support for another round of quantitative easing by the Fed, creation of an infrastructure bank, federal government support for state government budgets, and most importantly, somehow resolving the foreclosure crisis that continues to depress the housing market.

Aargh again!

Robert Reich pooh-poohs the March employment numbers:

The US economy added 162,000 jobs in March. Great news until you look more closely. In March, the federal government began hiring census takers big time. These are six-month temp jobs, and they tell us nothing about underlying trends in the labor market... A census-taking job is better than no job, but it’s no substitute for the real thing. Bottom line: This is no jobs recovery.

But Reich has also been a major proponent of a "new WPA" to fight unemployment. Hey, the government has just announced a plan to hire a million workers on a temporary basis for the next six months. It's called "Census" rather than "WPA," but you'd think Reich could gin up just a bit more enthusiasm for it nevertheless.

And yet...

ADP reports private employment down 23,000 in March. But: that doesn't count Census workers, of which the government probably hired around 100,000 this month; ADP has been 70,000 too low on average since October; ADP numbers weren't affected by the blizzard in February while BLS's were artificially suppressed. So it's not hard to see BLS reporting +200,000 even in light of ADP's numbers. But I won't be satisfied with anything less than 250,000.

Monetary policy, bond markets, and the job-rich recovery

Hyman Minsky argued that monetary policy affects the economy in the following way. The Federal Reserve reduces short-term interest rates. There's now no profit to be made sitting on money, so financial managers shift funds out of short-term, risk-free assets into long-term risky assets like corporate bonds. The increased demand for bonds pushes their prices (which were low at the trough of the recession) up and their yields (which were high) down. Credit spreads (the difference between risky rates and risk-free rates) therefore plunge. Businesses can borrow more cheaply, so begin to undertake investment projects that they had delayed due to the recession. Consumers also face lower borrowing costs, so they loosen up, and the recovery is underway.

The story implies that a plunge in credit spreads, say as measured by the difference between the yield on Baa corporate bonds and 10-year Treasuries, should normally precede a recovery in employment. Look at the historical record:

1974-76: Baa-Treasury spread peaks at 3.31 in January 1975, falls to 2.49 by July. Strong growth in employment begins that month.



1981-84: Baa-Treasury spread peaks at 3.82 in October 1982, falls to 1.79 by August 1983. Strong growth in employment begins in April 1983 with spread at 2.89.



1990-94: Baa-Treasury spread peaks falls, rises, falls again in early stages of the "jobless recovery." The last peak, at 2.25, comes in October 1992. From there it falls to 1.29 by December 1994. Strong growth in employment begins December 1992.



2000-2004: Another jobless recovery, with Baa-Treasury spread bouncing around until it peaks at 3.79 in October 2002. From there it plunges to 2.03 by May 2004. Strong growth in employment begins in March 2004.



So what have we seen this time around? A much more severe jump in credit spreads (up to 6.0 in December 2008) and a much more severe recession. But there's been a continuous drop in the spread beginning in March 2009, and the spread is now at 2.65.



Today the WSJ reports on the rally in corporate bond markets, and it sounds like it could have been written by Minsky (or Barbera) himself.

Investors flooded risky companies with money in March even as the government prepares to shut down a key engine driving one of the greatest corporate-bond rallies in history. A total $31.5 billion in new high-yield debt, otherwise known as junk bonds, hit the market through Tuesday, exceeding the previous monthly record in November 2006. Partly propelling the activity: The Federal Reserve's massive mortgage-buying program, which comes to an end Wednesday. By buying $1.25 trillion of mortgage securities, the Fed absorbed a flood of assets that otherwise would have needed buyers. That kept money in the hands of investors, who went searching for something else to buy. The Fed's underpinning encouraged investors to seek riskier, higher-yielding securities. A natural choice: corporate bonds...

The revival of bond fortunes has roots in the Fed's decision, around Thanksgiving 2008, that may have done more than anything else to encourage more investors to take a flyer on bonds. On Nov. 25, the Fed announced it planned to buy debt and mortgage-backed securities issued by housing-related governmentsponsored entities such as Fannie Mae and Freddie Mac. The program pushed mortgage-security prices higher, giving fixed-income managers an incentive to sell to the Fed. In return, they had a flow of cash that had to be put to work. With Treasury debt yields at record lows, the best alternative remaining was corporate debt. "That was the big turning point," says Ashish Shah, head of global credit strategy at Barclays Capital. "That's what drove money into credit."

The Fed expanded this program on March 18, of last year, to buy $1.25 trillion in mortgage securities, along with $200 billion in debt of Fannie and Freddie and up to $300 billion in long-term Treasury debt. The expansion fueled the second leg of the rally, which hasn't stopped.

The behavior of credit spreads in the last year looks very much like it did in 1975-76, 1982-83, and 2003-04. The change in employment has been following a path similar to that in 75-76 and 82-83. It's taken us longer to hit zero this time around because the declines were so enormous during the recession. With monetary policy having delivered a 340 basis point reduction in credit spreads over the last year (compared to 82 in 75-76, 203 in 1982-83 and 176 in 2003-04), is it unrealistic to think we're on the verge of a strong recovery in employment? I don't think so.

March employment and the Census

Calculated Risk runs the numbers and says if employment grows 200,000 in March, that's a pretty weak report. I'm inclined to agree, since March's number has to make up for the blizzard-suppressed February number, but at the same time perhaps beggars should not be choosers. I'm a little puzzled at CR's attitude toward Census workers. Beginning in March the Census Bureau will start hiring workers by the gajillions. CR calls this a "distortion" to the data. I say a job's a job - if Congress had passed a big jobs bill that created the same number of jobs, would CR call that a distortion?

February jobs report

I'm afraid I'm going to have to cry foul on the Squawkbox talking heads. The BLS reported payroll employment down 36,000 in February and the unemployment rate unchanged at 9.7%. The analysts, who for weeks have been saying that the February blizzards could reduce the payroll number by tens of thousands, maybe over a hundred thousand, saw the number and concluded that the blizzard effect had been much negligible after all. Apparently they had fixed in their minds that the "true" number was likely to be zero; when the number came in less negative than they had predicted, they revised downward their estimate of the blizzard effect rather than revising upward their estimate of the true number. I thought the blizzard effect was going to be in the area of -100,000 - though who really knows - and so I thought the payroll employment number was going to be in the area of 0 to -100,000. There's no information in the jobs report to change my estimate of the blizzard effect, so I'm going to say the "true" number was around +64,000. In other words, a pretty good report and an indicator that the jobs recovery is progressing on schedule.

Evidence for a substantial blizzard effect is that the biggest job-losing sectors was construction at -64,000. Also, the stock market responded positively to the report - traders may be interpreting the report as suggesting the beginning of sustained employment growth.

Bob Barbera's view of the report was more positive than the other commentators - no surprise there. He noted that we've now had two consecutive months with employment growth in the household survey of 541,000 and 308,000 respectively. That's phenomenal news.

Mark Zandi moved the goalposts. He's conceded that March and April are going to be good months for employment, but the true test is May and June. I want to rewind the tape to a few months ago and see if he's being consistent with his past statements.

Meanwhile, the report revised December's and January's payroll numbers: in December the BLS now estimates that the economy lost 109,000 jobs rather than the 150,000 originally estimated, and in January the job loss is 26,000 rather than 20,000.

But Fed forecasts are as wrong as everyone else's!

The employment report is great news, right? Not according to many prominent commentators. Here's Paul Krugman:

Good news is bad news

Today’s unemployment report was good news. But in a real sense good news is bad news, because this month’s not-too-bad number deflates the sense of urgency.

The fact remains that realistic projections show unemployment staying disastrously high for many years. The chart above is from the minutes of the Fed’s Open Market Committee. Unemployment above 8 percent in the fourth quarter of 2011; above 7 percent in the fourth quarter of 2012.

That's fallacy #2 in my taxonomy (see post below). The Fed's forecasts are liable to be as wrong as everyone else's. Let's examine, for example, the Fed's forecasts from October 2007, two months before the start of the recession.



GDP growth of 1.8-2.5% in 2008 and 2.3-2.7% in 2009; unemployment averaging 4.8-4.9 percent in both years. Um, how did that work out? And should I bring up the Fed's embarrassing forecasts in October 2008 - after the collapse of Lehman Brothers, when the bottom was falling completely out of the economy? Here's what the Fed's consensus forecasts were for 2009:

GDP: -0.2-1.1%
Unemployment: 7.1-7.6%

Oog!

Happy days are here again (you can tell by the smile!)

Just a few comments:

1. We're in much better shape than we thought we were. The BLS revised the job loss numbers down 79,000 in October and 80,000 in September. The labor market does not seem to have stalled in the third quarter, it kept improving at the pace it began in the spring.

2. Average improvement in the job loss numbers since February = +73,000 per month. If this trend continues, job growth will be +208,000 by February 2010. That would be phenomenal.

3. Several commentators have cautioned that this month's data will be revised, so we can't take it at face value. Yes, revisions are possible. But repeat Bob's mantra: at business cycle turning points, "revisions are in the direction of the inflection." That is, we're more likely to see an upward revision than a downward one.

BINGO!

BLS report on November employment situation:

Unemployment rate = 10.0%

Change in payroll employment = -11,000

I believe I called this one.*

* I will now claim to have been confident in the forecast I made a week and a half ago. Those hesitations you detected in more recent posts were figments of your imagination.

Curious headline

Every month around this time the WSJ reports breathlessly on the Bureau of Labor Statistics release on state unemployment rates with headlines like "Unemployment in States Remains High." The story starts:

Unemployment rates were little changed in most states in October, as a recovery in the labor market remained sluggish across the country.

The Labor Department reported that 19 states and Washington DC experienced jobless-rate decreases, while the rate rose in 14 regions and was unchanged in 17.

Well, we kind of knew this already didn't we? The national numbers did come out on Nov. 5 and showed a steady unemployment rate. What, you thought that unemployment would be plummeting in most states while the national number was unchanged?

If current trends continue...

... job growth will turn positive in January. An employment change number today in the -200,000 range is consistent with that trend.

But job growth needs to be around +140,000 for the unemployment rate to fall (usually); that won't happen until around April.

Employment growth (percent)



Bob says that "at business cycle turning points, revisions are in the direction of the inflection." So when the jobs numbers are revised next year we may find that job losses were smaller last summer and this fall than we think they are now, and the period of job growth would come earlier.

** Update **

Today's employment situation report says 190,000 jobs lost in October: right on target. Furthermore, job losses were revised substantially downward (fewer losses than earlier estimated) in August and September. Unemployment up to 10.2% - very bad, but it doesn't change my story.

Trip down memory lane

I've got no new insights on the employment numbers. But I'm curious, how have the benchmark revisions changed my understanding of where we are and how fast the labor market is going to return to growth territory. First, a trip down memory lane.

On November 6, in the wake of an employment report for October that showed a loss of 190,000 jobs, I presented a graph of the change in payroll employment. The graph drew a line up from the worst month of job losses (January 2009) through the job loss figures since that date. I argued that if current trends continued, employment growth would cross zero in January and be positive in February.

On December 3 I undertook a similar analysis and found that employment growth should be positive in February, maybe a bit negative in January.

On December 4, following the release of the November employment numbers, I guessed that following current trends job growth would be somewhere around 130,000 in January and 200,000 in February. Optimism grows.

On January 9 I projected trends from quarterly data and guessed that we'd have +50,000 in January and +100,000 for February.

On January 21 I used a VAR to forecast employment from initial claims data and predicted that January's employment would be +135,000.

The BLS has now revised its figures for April 2008 - March 2009, which changes the seasonal adjustment factors that it uses for the data after March 2009. The effect of the revisions are shown in the two graphs below. The decline in jobs during the recession was much worse than we thought previously, but the pace of the turnaround has not changed much. Using data from January-December 2009, I would have predicted an increase of 97,000 jobs in January before the revisions and 58,000 after the revisions. The actual number, -20,000, is lower but not much lower. It might have been slightly positive but for really bad weather in January: 75,000 jobs were lost in construction and 14,000 were lost in leisure/hospitality. Almost all other sectors of the economy showed employment increases.

So I think we're still on track. Most of my previous crude forecasts had shown January to be a mediocre month and February to be the first month of substantial gains.