It’s hard to get away from sticky wages
David Henderson quotes from a email by Jeff Hummel:
Most explanations for the current persistence of high unemployment fall into one of three broad categories:
1. Insufficient aggregate demand. This is the position of the Sumnerites who advocate more monetary stimulus and the Keynesians who advocate more fiscal stimulus. While I agree with Scott Sumner that Bernanke’s monetary policy was too tight when the crisis hit during 2007-2009, I think that is irrelevant now. Why hasn’t there been sufficient time for inflexible wages or prices to adjust?
A few points in response:
1. Most other explanations (such as policy uncertainty) also require sticky wages to explain high unemployment.
2. The 2008-09 slump was different from all other deep recessions in one key respect. After the severe 1907, 1921, 1929-33, 1937-38, 1957-58, 1974-73, and 1981-82 slumps, NGDP recovered rapidly. This time the recovery was very slow. Hummel seems to accept my view that the big crash in NGDP (due to bad monetary policy) pushed unemployment from 4.5% in 2007 to 10% in late 2009. Can that explanation explain why unemployment has only fallen to 8.3%?
From a common sense perspective the answer is no. Common sense would suggest that wages and prices should have adjusted by now to the negative nominal shock. But there’s another way of thinking about the problem that makes my argument much more plausible. Let’s start with the fact that we really don’t understand wage stickiness very well. Most people visualize long term contracts that get renegotiated, but it’s also about things like minimum wages (which were raised sharply), government unions, money illusion at the zero rate increase point, protected sectors like education and health care, etc. It’s a very complex labor market. So let’s consider sticky wages a sort of black box that we don’t fully understand.
Now let’s consider the stylized facts:
1. Nominal shocks have real effects, just as the sticky wage theory would predict.
2. Big nominal shocks have big real effects, just as the sticky wage theory would predict.
3. Fast NGDP growth during recoveries is associated with fast recoveries in employment, just as the sticky wage theory would predict.
4. The extremely slow recovery in NGDP after 2009 was associated with an unusually slow recovery in employment, just as the sticky wage theory would predict.
5. Wage growth slowed sharply after 2008, just as the sticky wage theory would predict.
6. Aggregate wage growth never fell below zero, which is hard to explain from long term contracts alone, but can perhaps be explained by combining all the factors discussed above, especially money illusion.
7. Given the combination of very slow recovery in NGDP, and wage growth only slowing from about 4% to 2%, you’d expect a very slow recovery. And that’s what we got.
8. But we did get some recovery, despite the fact that NGDP has been growing at below trend during the recovery. So there has been some wage adjustment.
Given that other theories like “policy uncertainty” also require wage stickiness, I’m not really sure what’s gained from moving away from demand-side models. Bernanke experimented with a historically slow recovery in AD (i.e. NGDP), and he got a historically slow recovery. What else would you expect?
In comment sections of recent posts people pushed back against my claim that the collapse in housing didn’t cause the recession, but most didn’t seem to understand the issues at stake. Lots of people talked about how it wasn’t the fall in home-building alone, but that there were also bubbles in areas like commercial real estate. No, commercial RE did fine until the second half of 2008 when NGDP collapsed. Or they said it was the hit to consumers from mortgage defaults, or the hit to banking. But those are channels that would reduce aggregate demand. And if it’s an AD problem, then ipso facto it’s a tight money problem. It’s the Fed’s job to stabilize AD. If you are going to argue against monetary explanations of the crash, you need a non-AD mechanism, such as misallocation of labor into housing construction. One commenter even mentioned the strong dollar in late 2008. That was supposed to be a counterargument to the Fed causing the crash, as if a strong dollar has nothing to do with tight money!
Tags:
10. August 2012 at 23:23
“But those are channels that would reduce aggregate demand. And if it’s an AD problem, then ipso facto it’s a tight money problem.”
This is the main source of contention between market monetarists and everyone else. Most people do not define low AD as specifically tight money.
The main problem is this: when you say “housing & banking crises didn’t cause the drop in GDP”, people think you literally mean that these things could not have affected demand, but what you really seem to mean is that although these things would have major effects on AD the Fed could have easily counter-acted these things, but didn’t, so it was the Fed that caused the drop in AD. Obviously most non market-monetarists would disagree with that as well, but it would at least clear up some confusion.
10. August 2012 at 23:23
Don’t the New Keynesians like to argue that the wage rigidity is (for whatever number of reasons) individually rational? As a coordination failure of some sort? Then there is no prima facie reason to suspect that the whole economy “snaps back” after a single time period of price-setting. Calvo is just the modelling tool, not the actual proposed model of wage bargaining…
11. August 2012 at 02:29
Scott you said:
“In comment sections of recent posts people pushed back against my claim that the collapse in housing didn’t cause the recession, but most didn’t seem to understand the issues at stake. Lots of people talked about how it wasn’t the fall in home-building alone, but that there were also bubbles in areas like commercial real estate. No, commercial RE did fine until the second half of 2008 when NGDP collapsed. Or they said it was the hit to consumers from mortgage defaults, or the hit to banking. But those are channels that would reduce aggregate demand. And if it’s an AD problem, then ipso facto it’s a tight money problem. It’s the Fed’s job to stabilize AD. If you are going to argue against monetary explanations of the crash, you need a non-AD mechanism, such as misallocation of labor into housing construction. One commenter even mentioned the strong dollar in late 2008. That was supposed to be a counterargument to the Fed causing the crash, as if a strong dollar has nothing to do with tight money!”
My only point is-as I mentioned in the other post-is that the stock market was already predicting a recession by July, 2007 after Angelo Mozillo’s form Countrywide Financial’s press conference.
The bear market begun at that point. So this was long before the fall in NGDP
11. August 2012 at 02:34
For illustration, the change in the wage/ngdp gap is able to explain most of the variation in unemployment. Only big supply shocks (like those in 70s) are above the regression line:
http://i46.tinypic.com/f0nbeu.png
11. August 2012 at 03:35
[…] Scott Sumner has this “companion piece”: Given that other theories like “policy uncertainty” also require wage stickiness, I’m not really sure what’s gained from moving away from demand-side models. Bernanke experimented with a historically slow recovery in AD (i.e. NGDP), and he got a historically slow recovery. What else would you expect? Rate this:Share this:EmailTwitterLike this:LikeBe the first to like this. This entry was posted in Uncategorized. Bookmark the permalink. ← Inflation Targeting: Drop Dead! […]
11. August 2012 at 03:42
Yes Mike, the bear market and the recession began long before the fall in NGDP. It was that fall that turned a run-of-the-mill recession into a depression:
http://thefaintofheart.wordpress.com/2012/08/11/it%C2%B4s-all-in-the-framing/
11. August 2012 at 06:26
thanks for the link Marcus, I am going to check it out right now.
11. August 2012 at 06:47
“Lots of people talked about how it wasn’t the fall in home-building alone, but that there were also bubbles in areas like commercial real estate. No, commercial RE did fine until the second half of 2008 when NGDP collapsed.”
It’s not a matter of whether CRE “did fine.” It’s a matter of whether there was widespread over investment in the CRE market that led to a glut of vacant properties in 2008. Have you seen a CRE price chart? See the bottom-left chart on page 5 in this report: https://www.wellsfargo.com/downloads/pdf/com/research/real-estate-and-housing/CREChartbookQ12012_05142012.pdf Are you saying that more AD from the Fed would have prevented CRE prices from tanking? If so, why? What fundamentally changed in the CRE market that would have allowed a price bubble that large to be maintained?
“One commenter even mentioned the strong dollar in late 2008. That was supposed to be a counterargument to the Fed causing the crash, as if a strong dollar has nothing to do with tight money!”
There was a global financial crisis in 2008. Do you know what happens to reserve currencies during global financial crises? They strengthen. This happens mostly irrespective of monetary policy.
I’m all for looser monetary policy to weaken the dollar in 2001-2006, but that’s a different argument. The point about the strong dollar in late-2008 was simply to say that there was one less channel available the offset the effects of the housing crash.
At least you’re acknowledging some of you critics. So we’re finally making progress.
11. August 2012 at 06:49
Does the sticky wage theory make a prediction (back-test, not forecast) about w/NGDP vis a vis unemployment?
Say, for example that if w/NGDP is at the same level as in 2008, unemployment should be at the same level as in 2008 (and differences should be explanable through demographics).
If yes, does it corespond to the stylized facts? If no, why not (why does the theory not make such a prediction) ?
11. August 2012 at 06:53
“There was a global financial crisis in 2008. Do you know what happens to reserve currencies during global financial crises? They strengthen. This happens mostly irrespective of monetary policy.”
A flight to “safety” is not a reason to print more money.
The GOAL of people hoarding cash is NOT to see the cash value be inflated away.
This annoys DeKrugman to no end.
They wish to pretend that since there is an increase in demand, we can increase the supply.
But there is a difference at keeping us from deflation and tking advantage of the money holders.
Afterall, the money holders are the people who money was created for, people who don’t actually accumulate wealth and capital, they just enjoy the side benefits of what and ho money was created for.
11. August 2012 at 07:08
Scott: “Most people visualize long term contracts that get renegotiated, but it’s also about things like minimum wages (which were raised sharply), government unions, money illusion at the zero rate increase point, protected sectors like education and health care, etc.”
I’ve seen such lists of causes of wage stickiness many times now, and they always seem to exclude the cause that, to me, seems most salient: throughout the labor market, there are strong social norms that bias wages toward no-decline. In almost any job, if your boss decreases your pay but not anyone else’s, that’s a good signal that it’s time to clean up your resume (and strike your boss from your references list).
In addition, there is a predominant, and quite reasonable, understanding that as one’s experience grows, one’s value as an employee grows, and so one’s wage at the least ought not to decrease.
These stylized facts seem to me to be logically, if not historically, antecedent to unions, minimum wages and protected sectors. They are the basis of the power of the Money Illusion.
Scott, would you be willing to add “long standing and very strong social norms against wage decreases” to your list of sources of wage stickiness? (You could call it a personal favor for me)
11. August 2012 at 07:40
@ Mike Sax:
“. . . the stock market was already predicting a recession by July, 2007 . . . .” This is an odd remark, since the stock market did not peak until early October, 2007.
“So this was long before the fall in NGDP.” The relevant factor is not current NGDP, but expected NGDP a year or two into the future.
11. August 2012 at 07:43
I would just say that any increase in employment will be an increase in output. If it is increase in output, it is regulated by the Fed, and therefore it is impossible to get employment growth under a strict growth rate-targeting regime because SRAS isn’t level.
You cannot exit from suboptimal production under a growth rate-targeting regime. That makes rate-targeting a great policy for staying on track and a bad policy once the economy slips off the growth path.
Growth rate targeting is difference stationary. This only requires two assumptions:
1) unemployment is suboptimal for output
2) SRAS isn’t level
11. August 2012 at 08:17
Even though it has proven quite difficult to renegotiate residential mortgage terms, from what I understand commercial real estate long term contracts are much easier to renegotiate and so would seem to be less susceptible to price rigidities. Any explanation whether this is so would be greatly appreciated.
11. August 2012 at 09:11
Scott – Does the rising DJIA and five-year breakeven spreads pushing back toward 200 bps make you believe the recovery and jobs may pick up a bit of steam here? Even commercial bank credit is starting to expand again, up 6% y/y, the best so far in this ‘recovery’.
11. August 2012 at 11:41
“The relevant factor is not current NGDP, but expected NGDP a year or two into the future.”
So you know what expected NGDP was in July, 2007?
12. August 2012 at 01:10
Like #1 Brito:
And if it’s an AD problem, then ipso facto it’s a tight money problem.
AD includes Net Worth. The fall in house prices means a HUGE, unprecedented, fall in Net Worth, and thus AD. Not just wages, wealth.
Because this hasn’t happened in the US before, Net Worth is seldom (never?) included as a driving variable in econ models used for making policy. This in a case where a usually non-driving variable changes by such a magnitude that it becomes a driver — and all models which ignore this variable become less valid, if not invalid.
Also like Brito, your reliance on the house price peak of 2006 not immediately being followed by a crash makes those of us more comprehensively looking at the economy feel you always deny the AD impact of the house price crash.
Wage stickiness is much less important than Net Worth impact on Ag Demand. But house prices are also sticky, tho not as much over the long term. The 2 year delay from house price peak before crash was significantly caused by house price stickiness … getting unstuck.
Because house buyers included owners wanting to live in the houses, and financial speculators “banking” on the prices increasing to make them financial profit. From 2006 until Lehman, it wasn’t clear to the house speculators that they were in deep deep DEEP do-do, and had to bail eating what losses they had. But once the speculators realized they had missed the peak (or even bought, foolishly, at the peak), they tried to bail as fast as they could.
Simultaneously, all the boomers with house equity saw their own nest eggs drop in value, so their buying behavior changed.
Despite not having big changes in their income.
So, I still agree with your policy prescriptions, but disagree, strongly, on your methodology of downplaying Net Worth effects.
12. August 2012 at 06:57
[…] Sumner sorta responds to a criticism of nominal GDP targeting that I’ve been raising. Recall that I’ve been arguing […]
12. August 2012 at 14:02
A tangential point I don’t see talked about much outside the MarketMonetaristsphere is debt stickiness. It can take years and years for debt burdens to adjust from the old 5% NGDP growth world, to the new 3.5% NGDP growth world, given the effect that the initial nominal shock had on wage growth.
12. August 2012 at 16:50
[…] Some commenters ask why I have such a strange definition of monetary policy, attributing any change in NGDP to monetary policy. Heh, I’m just using Captain Ben Bernanke’s definition. […]
13. August 2012 at 09:14
Tom:
AD includes Net Worth. The fall in house prices means a HUGE, unprecedented, fall in Net Worth, and thus AD. Not just wages, wealth.
I don’t think that’s right, Tom. AD is a flow concept. It deals with how many dollars are spent over a given period of time.
Net worth is a stock concept. It deals with a snapshot of the value of a person’s property at a given time, which is related to spending, in the sense that the value is estimated via past or expected future expenditures. But that doesn’t mean AD includes net worth.
If I own a home today the value of which is estimated by the top potential buyer to be worth $300k, or if I own a factory today the value of which is estimated by the top potential buyer to be worth $200 million, then these value estimates are NOT a part of actual AD today. AD for today would consist of the total number of dollars spent today.
14. August 2012 at 05:40
[…] rate of 8.4 percent and has not returned to the previous trend, something unheard of for any other recession in the past century. The fall in NGDP has harmed the economy in many ways, most importantly by making it harder for […]
16. August 2012 at 08:36
Since I’m new to this “blogging” business, I figure I’m allowed to promote myself in other blogs’ comment threads, right? Here is a story of sticky wages, told in pictures.
16. August 2012 at 08:59
Neal,
I liked your “story of sticky wages, told in pictures.”
I have a question about the quote from Scott in your post…”NGDP shocks are like a game of musical chairs. Remove 9% of NGDP relative to trend, and you’ll have 9,000,000 unemployed workers sitting on the floor. It’s that simple.”
Why 9,000,000 workers ? Why not (just making up a number here ) 90,000 ceo’s taking a wage cut ?
Some peoples wages are stickier than others. lol
Also what if we did something similar to what they do in Germany ?
Instead of laying off workers we cut pay by 9%. But then the government makes up the cut with partial “unemployment” insurance of 9%.
All other things being held equal wouldn’t that make up the NGDP trend shortfall ?
16. August 2012 at 11:37
Why 9,000,000 workers? Because that’s just how our economy is structured. If we had some insurance similar to what the Germans have, we would see (private) wages fall as total wages fall; that wouldn’t stop the nominal income drop. Better to just not let nominal incomes fall.