Monetary policy drives NGDP growth, and NGDP growth (per worker) is by far the most important determinant of nominal wage growth. (The other determinant is labor share of GDP.)
During the past 4 years, NGDP growth has been running at 4.05%/year, well below the historical norms. So why is wage growth running at only about 2.5%/year? The answer is simple; payroll employment has been growing at 1.78% over that same period. The predicted growth in average hourly earnings is 2.27%, whereas actual wage growth has been running at 2.47%. The recent puzzle is why is wage growth so high, not low.
[If you look back over the past 8 years, you have NGDP growth averaging 3.92%, employment growth at 1.66%, predicted wage growth at 2.26% and actual wage growth at 2.23%.]
So why are so many people puzzled over the lack of wage growth? They make the fundamental error of confusing micro and macroeconomics, unfortunately a common mistake among Keynesians. Factors like international trade, tight labor markets, etc., may affect relative wages in particular fields. That’s microeconomics. But at the macro level, nominal wages are mostly driven by monetary policy. (Of course real wages are a different story, heavily influenced by productivity.)
There is no mystery as to why wages have recently been growing at 2.5%, not the 3.5% to 4% typical before the Great Recession. NGDP growth has recently been running around 4%, which is below the 5.4% trend before the Great Recession.
Wage growth is almost exactly what you’d expect given recent trends in monetary policy. There is no mystery to explain. The lesson here is to never confuse macroeconomics and microeconomics; they are radically different fields. Micro models cannot provide macro answers, for the simple reason that the factors that determine real prices are radically different from the factors that determine nominal prices.
PS. This Paul Krugman post is a typical example of the way most people think about wages. Krugman focuses on several microeconomic type explanations for slow wage growth, while missing the elephant in the room (NGDP growth).
PPS. David Henderson has a couple recent posts (here and here) where he is rightly skeptical of some explanations for the wage puzzle, but perhaps not skeptical enough. Caroline Baum has a recent article that also fails to mention NGDP growth, but does correctly note the link between slow real wage growth and sluggish productivity.
READER COMMENTS
Kenneth Duda
May 11 2018 at 5:58pm
My goodness you make this sound so simple. Of course, average wage growth *has* to be NGDP growth minus workforce growth, unless labor share of income is changing, which it isn’t.
Why isn’t this obvious to more people??
Effem
May 11 2018 at 6:01pm
I think people mostly lament the falling wage share of wealth. Why is it that wealth can grow faster than NGDP?
David R Henderson
May 11 2018 at 7:10pm
@Scott Sumner,
Excellent.
What you would really need is the number for growth of labor hours, right? And you figure that employment growth is almost perfectly correlated with growth of labor hours, right?
Kenneth Duda
May 11 2018 at 11:17pm
Effem: If you take a simplistic definition of “wealth” as the market value of your assets minus your liabilities, then the reason that wealth can grow much faster than NGDP is that asset prices can rise and fall based on expectations of future earnings from those assets, and based on prevailing interest rates. For example, if interest rates are 2% and I expect a share of stock to earn $1 a year indefinitely, then (as a simplistic approximation) the fair value of the stock will be $50. If interest rates rise to 5%, and expectations of future earnings fall to $0.50, then my share will fall to $10. I just “lost” 80% of my wealth. If rates move back down and expectations of future earnings move back up, that “wealth” will reappear.
-Ken
Scott Sumner
May 11 2018 at 11:51pm
Ken, I agree. In fairness, the ratio of wages to NGDP/worker can change a bit from year to year, but over a number of years it’s a pretty simple relationship.
David, Yes, labor hours would be the best number, and yes, they are highly correlated. Of course the labor share of national income can move around a bit from year to year. But over 4 or 8 years it’s a pretty good predictor.
Plato’s Revenge
May 12 2018 at 4:23am
I guess many people (me included) prefer a ‘forward looking’ explanation — NGDP is a statistic calculated ex post. A productive way to frame this is to wonder how on earth the Fed managed to hold down NGDP growth in the face of healthy employment growth and inflation formerly being anchored at ~2%. But, of course, we already know this…
What we also want to understand is why split between employment and (hourly) wage growth makes both moderate, while microeconomically, we might expect employment to stall to ‘make room’ for wage ‘inflation’. I.o.w., does work really look in as short supply MACROeconomically as it seems to look MICROeconomically?
Bruno Duarte
May 12 2018 at 5:44am
I’m glad we’re finally on the same page. So, via the wage channel, low interest rates are deflationary if we don’t account for consumer debt, correct?
Bruno Duarte
May 12 2018 at 8:08am
Here’s a blog post I wrote about it:
The puzzle remains on why low unemployment is not resulting in higher wages.
Whereas some conclude low interest rates imply stagnant NGDP, that in itself restraining wage growth, others argue (without much success) demand shifting abroad lowers wage competitiveness by limiting overhead growth.
Paul Krugman argued wage hysterisis as a behavioural reaction to overleverage prevents businesses from competitively bidding labour. I’d add labour became less demanding of high wages absent Labour Union action, as it happens.
Compiling both arguments, we could almost conclude low interest rates are deflationary.
As a by-product of my thesis on EFSI and Cohesion, available in print at the IES, I discovered a high correlation (>0.89) between Producer Price Index (PPI) and Productive Capital – but not Capital Stock.
The observation results from Ben Bernanke’s comments while Fed Chairman on how capital depreciation may hold back utilised capacity – as anecdotally demonstrated here with the IT industry.
Low interest rates are not an incentive to higher prices to either supply or demand. For supply, lower overhead costs (that result from high umeployment rates) and lower investment costs lower the incentive to increase prices. If prices are stagnant and interest rates are lower for longer, consumers may act as with deflation and postpone spending.
The negative correlation between Fed Funds Rate and Consumer spending (<-.37) plots a divergent chart between consumer spending weighted by Fed Funds Rate and the Rate itself (indexed to 1990). The higher correlation (in the US, >.69) between Manufacturing Productive Capital and PPI, and the latter’s high correlation with CPI (>.98) leads to believe the right policy mix to avert deflation is somewhat related with increasing deductions for depreciated capital. [Results hold better for Europe as DG ECFIN statistics account for economy-wide productive capital]
Oddly enough, that’s what the Trump Administration is doing.
—
This led me to believe firms started by increasing headcount before increasing capital spending – and firms did so: plot the chart of private consumption, new hires and gross private domestic investment for the US, indexed to end of the 2009 Recession, and you’ll see new hires tracks consumer spending for three quarters (before soaring as firms anticipate a trend) while investment dips until Q3 2009.
https://eunomicsblog.wordpress.com/2018/05/12/wages-inflation-and-jobs/
Anisha Meka
May 13 2018 at 9:33pm
Since, wages depend on both the supply and demand for labor, macroeconomic policy can raise wage growth by trying to create the optimal level of demand for workers.Furthermore, there should be policy implemented that raises the productivity of the workers and allows them to apply for higher paying jobs. There should be an upscale of wages over time and growth should not remain flat.
Andrew_FL
May 14 2018 at 10:49am
Outline the specific process whereby “the labor share of GDP” causes anything
Comments are closed.