Scott Sumner  

Once again, the Fed was wrong

Vargas Llosa on populism and p... Incomes, Spending, and Saving...

About 2 years ago, I predicted that 3% NGDP growth was the new normal. At the time, the Fed predicted significantly higher trend growth (both RGDP and NGDP.)

The Fed has a big NGDP problem. It's becoming increasingly clear that when the labor market recovers, RGDP growth will be very slow, maybe 1.2%. Add in about 1.8% on the GDP deflator, and 3% NGDP growth looks like the new normal, assuming the Fed intends to stick with 2% PCE inflation targeting. Bill Woolsey wins!!
With today's figures, it is now pretty clear that I was right and the Fed was wrong.

(This has been repeatedly true of one issue after another, over the past 8 years. For instance, in December I predicted the Fed would raise interest rates in 2016 fewer than the 4 times they were predicting.)

To estimate the rate of trend growth in real GDP, you need to go back to a period where the labor market was approximately in equilibrium. In the 1st quarter of 2008, the unemployment rate averaged 5.0%. In the first quarter of 2016, the unemployment rate averaged 4.97%, virtually identical to 2008:1. The growth rate in RGDP over that 8 year period was 1.286%/year, just slightly above the 1.2% rate I estimated as the trend rate. And with labor force growth now slowing due to boomer retirements, I expect RGDP growth to slow a bit further, to 1.2% as the new normal.

Why was I able to see something the Fed missed? Because I don't rely on past values of trend growth as being stable. In 2008 and 2009, the Fed erred by assuming that the natural interest rate was higher than it actually was. The performance of the economy was telling us that the Wicksellian equilibrium rate had fallen sharply, but the Fed simply refused to believe it. These low rates were inconsistent with decades of experience. The Fed assumed we'd eventually get back to normal, although now it's becoming increasingly clear that low rates are the new normal.

Conclusion: Listen to the markets.

The same occurred with trend RGDP growth. For well over 100 years, America has averaged roughly 3% per year in trend RGDP growth. Thus the Fed assumed the economy would return to this growth path, after the recession. In fact, we are not doing so---we are in a "Great Stagnation" that (AFAIK) has no precedent in all of American history. Even the 1930s were totally different, as productivity growth was high and RGDP growth bounced back once the unemployment rate fell. But now we have unemployment having fallen all the way to 5%, and growth is still 1.29% over 8 years.

I first saw this coming back in 2011, because I noticed that we had a "job-filled non-recovery" not the jobless recovery postulated in the media. That is, the unemployment rate was falling quickly, but output growth remained well below 3%. Normally, growth is well above trend when unemployment is falling fast. That told me that RGDP trend growth had slowed sharply.

So should the Fed fire all their economists and hire me? Yes. But an even better solution is to fire all their economists and hire someone like Robin Hanson or Justin Wolfers to set up prediction markets for macro variables. Stop relying on government bureaucrats to predict the economy, and instead rely on the wisdom of crowds.

PS. Was the slow growth in Q1 (1.2% annual rate for NGDP) caused by the Fed's decision to tighten policy in late 2015? Probably, at least in part.

Comments and Sharing

COMMENTS (20 to date)
John Hall writes:

My biggest worry about nominal GDP targeting is the risk of changes in trend GDP growth.

ThaomasH writes:

How can we be sure that the lower RGDP growth is "normal" and not the result of firms fearing that in any future AD shock the Fed will again fail to maintain NGDP and reduce investment accordingly? Or asking another way, if the Fed had maintained NGDP growth after 2008, (targeted a rapid return to 5%), would real GDP growth have been as slow as it has been with inflation having been correspondingly higher?

What good would a NGDP predictions market do if the Fed continues to have an inflation rate ceiling?

Mark Barbieri writes:
Stop relying on government bureaucrats to predict the economy, and instead rely on the wisdom of crowds.

This is the best suggestion that I've read in a while. At the very least, allow (and encourage) the markets to develop alongside the bureaucrats so that people can compare their accuracy over time.

Andrew_FL writes:
How can we be sure that the lower RGDP growth is "normal" and not the result of firms fearing that in any future AD shock the Fed will again fail to maintain NGDP and reduce investment accordingly?

The long term trend growth rate is determined by "supply side factors" i.e. money is neutral in the long run.

Brendan writes:


Thanks for continuing to share all your insights! Before reading your stuff here at Econlog I always found Macroeconomics to be marred by false assumptions that were too far abstracted from the principles of Microeconomics.

Do you have any articles that suggest why we are in a period of stagnant growth with regards to NGDP and RGDP? Is it institutional bloat, supply-side, an aging labor force, the growing pains of major shift to a human capital-centric economy because our labor force cannot adapt quickly enough?

Thanks for sharing!

bill writes:

I like ThomasH's comment.

bill writes:

Your prediction of 3% NGDP growth was very good. It's astounding that the people who actually control the levers can be out-predicted! Imagine baseball fans who could better predict the speed of the next pitch than the pitcher himself.

And maybe the prediction on the 4 rate increases is more astounding in a way. Not only did you have to see that the Fed was going to miss its target but you had to predict that they'd be missing it so much that even they'd realize it and react semi-appropriately (the appropriate response would have been to "normalize" IOR by returning it to zero (I put normalize in quotes because almost everyone else that uses the word normalize means to raise rates)).

Scott Sumner writes:

John, Why would that be a problem? That's exactly why I prefer NGDP targeting to IT.

Thaomas, We can't be sure of anything in economics, but it seems unlikely that tight money would still be depressing output 8 years later, when unemployment has returned to normal.

Mark, Yes, that's exactly the first step that I have suggested.

Brendan, I'm not certain, but it seems to be a global phenomenon. I suppose it's partly bad policy, and partly due to the shift to a service economy. There are also measurement questions.

Bill, Thanks, but I owe it all to the markets. I find it funny that the markets can predict what the Fed will do better than the Fed itself.

Tyler writes:

"The long term trend growth rate is determined by "supply side factors" i.e. money is neutral in the long run."

Andrew, there are two sides to the coin. If firms think there is going to be inadequate demand (for whatever reason, poor monetary policy, excessive saving or 'secular stagnation') then they are going to adjust capacity accordingly. You can't have solid RGDP growth without normal demand.

BC writes:

George Will likes to say that American politics works well with 3% growth but, at 2% growth, our institutions are not good at allocating scarcity. See for example here: []. I don't know what the consequences of 1% growth will be, Trump vs. Sanders elections for the next few decades? I hope not.

Matthew Moore writes:

Scott - you should look for ways to better publicise these predictions. In particular, I suggest starting a specific and public central collection, starting now, of your forecasts compared to the Fed.

You may well have been consistently right, but a cynic would think you are just cherry-picking your best predictions, and it would take a lot to prove that idea wrong. By ex ante labeling those statements you are happy to be part of your prediction set, you avoid this problem.

Scott Sumner writes:

Matthew, Thanks, but I don't view myself as a good forecaster. Beating the Fed at forecasting is like taking candy from a baby. All I had to do is look at the fed funds futures last December, and I could see that the Fed was probably not going to raise rates four times this year.

Chris L writes:

Wasn't the current economic condition predicted decades ago? Where I work baby boomers are retiring in droves. This is a major brain drain that has to be putting a downward pressure on overall productivity. Also, all the advice is as one nears retirement, start moving investments to safer havens, so that explains lower interest rates as people are moving into more govt bonds along with sluggish pressures on stock markets.

I know I read predictions about this decades ago.... so why is this all a mystery now?

LK Beland writes:

Great stagnation indeed.

From 1970 to 2010, real GDP per hour worked increased by 1.7% per year. The 4-year standard deviation is 0.6 percentage point per year.

From 2010 to 2014, it has averaged 0.2% per year. This 4-year average is 2.5 standard deviations from its long-term mean. That's low.

Chris L might be right that older workers might be at cause (brain drain when retiring, or maybe simply not being as productive as they grow old). The move away from risky investments might also hamper productivity.

Or is it because people coming back to the workforce have been unemployed for too long and became "rusty"?

ThaomasH writes:

@ Andrew FL,

The long term trend growth rate is determined by "supply side factors" i.e. money is neutral in the long run

If it's not neutral in the short run (and how can anyone look at 2008-16 and say that if has been neutral), it can't be truly neutral in the long run, either. Does not the possibility (likelihood, given current Fed policy) of future Fed errors in not maintaining NGDP trend depress investment in human and physical capital. Those are "supply and demand factors."

LK Beland writes:

Interestingly, the other two "major" non-oil OECD economies at the production frontier (looking at GDP per capita PPP), Switzerland and Ireland, did not suffer from the same slowdown in the growth of GDP per hour.

The "Great Stagnation" did not take root there yet.

Jesse C writes:

I'm not holding my breath until the "experts" to agree to face off against markets.

This isn't like Garry Kasparov playing against the internet's voted-for next move. Markets are more intelligent, the game more complex, and the experts have more to lose.

Andrew_FL writes:

Tyler, ThaomasH, you're both stuck on the same fundamental fallacy. The only reason the magnitude of the stream of money payments matters to the amount of physical production is because the entrepreneurial process of price discovery is just that, a process, rather than working like Walras' imaginary auctioneer. What you're both really saying, is that you just don't think market pricing works, that prices can get permanently stuck, not merely long enough to temporarily depress output, but to do so forever. That's absurd.

Andrew_FL writes:

"But it is alarming to see that after we have gone through the process of developing a systematic account of those forces which in the long run determine prices and production, we are now called upon to scrap it, in order to replace it by the shortsighted philosophy of the business man raised to the dignity of a science."

Hayek, in The Pure Theory of Capital, as if anticipating this very conversation. (Emphasis added)

ChrisA writes:

On the low rate of monetary expansion depressing real growth - I agree that it is strange that after 8 years we would still be seeing this effect. But I can't help feel that it is related. My take is that perhaps low inflation reduces risk taking - I wonder how much of capital investment is not because of a desire to "become rich" but to protect capital? In previous high inflation times, there was large financial repression, with interest rates significantly below inflation, so you had to invest. With today's interest rates you are only a percent or so below inflation so the penalty for keeping the money in the bank and waiting for a good opportunity seems less. This is true of individuals, but also firms. So capital investment is less and productivity lower.

There is also the opposite theory, when there is high inflation people see real assets increasing in nominal terms at high rates and then they want some of that nominal increase, so they go and invest.

Related issue to the one above - investments returns are calculated in nominal economics, if you see that your product prices are going to rise in nominal terms, even if zero in real, you can still get a positive IRR in nominal dollars.

Final issue is the opposite of the sticky wage - people maybe are more willing to stick at their current job in a low inflation environment - in an inflationary environment if you see your wage declining in real terms you would be more vigorous in looking for a new job, even if nominally it was being held. Or you form a union and agitate for higher wages. Owners see this which means more productivity increases are needed to offset increasing wages.

I do wonder then if we did see a significant inflationary expectation increase (to say 4%) whether we would soon see increasing productivity and then higher GDP/capita increases.

Comments for this entry have been closed
Return to top