The second-quarter GDP numbers came out. The newspapers and Republicans pounced on low growth and anemic job growth. The Democrats rebut growth is growth and tell us of the steady job gains. How bad is the economy?
Economists know that levels matter, and that long-run growth matters more than anything else. I made a few graphs to emphasize these points.
Start with the level (in logs) of real GDP. (This is an update of a graph I saw on John Taylor's blog.)
If you distrust trend lines, you are wise. But this one reflects a solid historical pattern. Here is real GDP and the 1965-2007 trend through postwar history.
You can see that the economy has quite reliably returned to the trend line after recessions.The 1950s had a steeper trend, but there too the small recessions were followed by catchup growth.
Here is what the recovery is supposed to look like (Again, idea stolen from John Taylor, except I'm using trends rather than "potential GDP'' which I distrust.)
To be fair, I fit the trend through 1980, so I would not use ex-post information. You see that after the severe 1980 recession at the even more severe 1982 recession, the economy recovered to trend, by posting a few years of 6% growth.
The tragedy is poorly expressed in growth rates. By 1987, the economy was back on the prior trend line. We are now 14.5% below the trendline, and each year that goes by like this we lose another half a percent. The average person in the economy is producing 14.5% less, and earning 14.5% less, than if we had followed the path following the 1982 recession.
That's a lot -- and a lot more than the litany of quarterly growth rates suggest.
I used trends, rather than the CBO potential output. If you read how they make it, you're likely to do that too. But here is the same graph contrasting my trend and the CBO's potential
This is tragic. The CBO is giving up on us. The CBO potential, which goes towards a 2.35% long run growth rate, says that what we are seeing now is the new normal. All we can hope for is a modest recovery, and then anemic, sclerotic growth forever after that. The difference between 2.3% and 3.0% adds up fast as the years go by. (And the CBO has been bending the trend line down steadily as the recession goes on. Back in 2005, it's "potential" looked like my "trend." They didn't see a permanent downward shift in level or reduction in growth rates. Look for "potential" to keep declining.)
Well, perhaps the CBO is doing its job as forecasters, saying "here is what will happen if you continue down the present policy path," not "here is where the economy would be if you adopted growth-oriented policies."
What about employment? I find employment more significant than unemployment. Unemployment means job search. It means people answer a survey saying they don't have a job, and are actively searching for a job. It does not count all the people who gave up, or went on disability (effectively ending their careers), early retirement, or are just living in Mom's basement and playing video games. (I don't mean to make light of it. That may be the most tragic, as the chance to accumulate skills is lost.)
Here's a good summary measure, the ratio of employed people to the population
The link between employment and output is productivity. To keep the numbers simple here, I made plots of output per worker. Output per hour, and corrections for demographics and capital use are better, but this is simpler and works about as well. Here is a graph of productivity.
In the short run, capital doesn't change much, so as a rough guide you make more output when you hire more workers (or increase hours) and vice versa. So, GDP = Productivity x workers. To get more workers, we need to make a lot more GDP. The lackluster GDP growth is the other side of the terrible employment coin.
There's more in the graph. In the long run, rising productivity is behind everything good in the economy. It's what gives more income per capita. Rising productivity is the only hope for paying for entitlements and getting out of our deficit trap. It's the main hope for long-run GDP growth, after the empolyment-population ratio reverts to where it should be. Rising productivity comes from new ideas, new companies, new ways of doing business. It isn't all pleasant. Lots of incumbents lose out. Rising productivity is the core of a "growth" agenda as economists understand the word.
You see in the graph that something terrible happened in the 1970s. Productivity, which was behind the large postwar boom, slowed down to a glacial 1% per year. 1982 marked a break in that as well. Productivity started growing 1.69% per year, producing the boom of the late 1980s and 1990s, and incidentally producing large Federal surpluses.
OK, but the far right of the graph doesn't look so good does it. Here it is, blown up, with a 2003-today trend marked in as well.
This is an economists' horror movie. Yes, productivity did rebound. But it seems to be growing slowly as well.
The trends are an economists' horror movie. Real GDP seems not to be recovering at all -- no period of swift growth to go back to a trend. We seem stuck at 2.4% growth forever. The CBO is giving up on us too. Employment will not recover as a fraction of population until the economy recovers. We seem stuck at low employment forever. And now we seem headed to a 1970s productivity slowdown as well.
I don't view this as contentious, outside of Presidential politics. Paul Krugman thinks the economy is pretty awful too.
What to do? If only it were so simple as to have the Fed print up another two trillion dollars, or have the Treasury borrow another $5 trillion and blow it on stimulus boondoggles. We're stuck in sclerotic growth, and to everyone but a few die-hard extremists, that means growth-oriented policies are the only way out.
Disclaimer. Yes, I know there are better ways to measure all this, especially productivity. This is an attempt to paint the basic picture using the simplest numbers. The message is, look at the levels and look at the trends. If you do that with better data, you will have gotten the message.
Data are from the St. Louis Fed's wonderful Fred database, series GDPC96, GDPPOT, EMRATIO.