Mind the (Spending) Gap

March 24, 2014

We all know that households cut back on spending dramatically during the Great Recession. Are they spending now? Has spending caught up to the trend the United States was on before?

The red line in the chart below plots retail spending in real terms in the United States from 1992 to 2013. We want to get a sense of the trend in spending so we plot spending on a logarithmic scale, and we subtract off the 1992 level to start the line at zero. A logarithmic scale is informative because a straight line in the chart would imply that spending was growing at a constant rate in real terms.


Prior to 2007, spending was growing at a constant rate of about 3% real growth per year. The black dots show the pre-2007 trend and where we would be if we had continued on that trend through 2013. The Great Recession is plainly evident in the chart: see the sharp decline in retail spending in 2008 and 2009 that took us well below trend.

So are we catching back up to our previous trend? Absolutely not. In fact, in 2013, it looks like the gap may be getting even larger. The gray arrow shows that we are not even close to the trend we were on before the Great Recession hit. This is unusual. In most recessions, strong growth takes us back to the trend we were on before the recession hit. Something about the Great Recession is different.

Why aren’t we getting back to trend? One answer often given is that the housing boom artificially boosted spending from 2002 to 2006, and so the trend we were on was an unrealistic benchmark that could not be sustained.

But the data contradict that story. There is no evidence that spending was above trend from 2002 to 2006. In the chart above, the red line doesn’t go above the black dots during the housing boom. Further, there is no evidence that the economy was overheating in terms of capacity from 2002 to 2006. House prices were booming, but other measures of inflation were steady.

Instead, the chart above may be evidence corroborating the worrisome “secular stagnation” view. The housing boom fueled household spending, but that spending only kept us on the same path we were on before – and it was done by enticing debtors to borrow and spend out of ephemeral housing wealth. When the housing boom disappeared, the permanent adjustment downward in the chart above suggests that we were already on a secular decline in household spending that the housing boom masked temporarily.

Bookmark and Share

16 Responses to Mind the (Spending) Gap

  1. FR on March 24, 2014 at 8:19 ami

    Maybe I am missing something, but why the log transformation. Data in real terms would indicate that retail spending in real terms has recovered to 2007 levels.


    • Atif Mian and Amir Sufi on March 24, 2014 at 9:24 ami

      That’s not the right way of thinking about it. We want to return to trend, not level. We typically return to trend in recoveries after recessions. Not happening this time.

      • FR on March 24, 2014 at 9:49 ami

        Thanks for the reply.

        I realize that we need to revert to trend. I was thinking that maybe it is too early to establish a trend after the shock.

  2. Jimbo on March 24, 2014 at 9:53 ami

    How would we see improvements in technology for such a data set? Could advances in technology be limiting how quickly we replace consumer goods? How would we see that in the data, if ever?

  3. Jimbo on March 24, 2014 at 9:55 ami

    Another observation is that. The divergence from the trendline looks like it began around 2004.

  4. BananaGuard on March 24, 2014 at 11:20 ami

    OK, so I’m gonna guess, based on the savings rate, that we are not saving more, so that the widening gap in real spending is not the result of an increased preference for assets vs consumption. Ah, then I look at a chart of real disposable income and sure enough, there’s a kink starting in 2008 and we have never gotten back to trend there either. Then I look at employment and sure enough, we fell away from trend and never recovered. (That’s a more common pattern in employment, but the drop this last time was bigger, and the new trend, just as in the case of spending and income, is shallower than before the recession.)

    • rjs on March 25, 2014 at 3:46 ami

      exactly…you cant expect the cart of retail sales to pull the horses of employment and disposable personal income…

  5. Avante Guard on March 24, 2014 at 12:12 pmi

    Are you plotting nominal retail sales or real retail sales? Disinflation will cause nominal retail sales to fall below trend, though I think that there is also a fall off in real sales. Looking at a similar graph using both real PCE and real disposable income, there is a bump up in both income and spending between 1998 and 2008 followed by a decline.


  6. Patrick MacAuley on March 24, 2014 at 8:45 pmi

    Another theory is that the long-term trend has changed since about 1997-2007. Certainly globalization has reduced the comparative value of American labor and this would be expected to eventually drive down US retail sales.

    There is no law of economics that says 3% real US growth has to last forever. And even if GDP could grow at 3%, the dynamics of globalization would tend to drive down the share of GDP received by the median American worker.

    • Richard on March 24, 2014 at 10:16 pmi

      Yet the trend has been extremely consistent in the US over the past 100+ years (through all sorts of economic regime changes): http://conversableeconomist.blogspot.com/2012/02/remarkable-consistency-of-long-run-us.html

      Given no change in the social superstructure (it’s not like the US has turned from a capitalist democratic to a Communist dictatorship; no matter what certain right-wing fanatics may think), this failure to return to trend is a conundrum (and worrying).

      • Patrick MacAuley on March 25, 2014 at 2:31 pmi

        Richard, I agree with you that this “failure to return to trend” is worrying. The impressive long-run trend through the 20th Century was largely the result of technology advances which continue unabated in the 21st Century. (I read http://www.physorg.com daily to view the amazing torrent of scientific progress.)

        Although the US economy has changed over the past several decades, the most dramatic changes have been elsewhere in the world, especially in Asia. As a result of the new globalization the average US worker is way over-priced, and our $475 billion trade deficit directly reduces aggregate demand which could produce good blue-collar jobs.

        Not only is the American worker over-priced but relative to the emergent Asian economies, our environmental requirements and the rules of international trade are working against us in the global game. The world has changed and our growth trend will probably change too, as global production costs equilibrate.

  7. Ralph Musgrave on March 25, 2014 at 4:15 ami

    I’ve read Summers’s IMF speech about secular stagnation, and regard it as fatuous nonsense. He basically claims that recovery is near impossible because of the drop in various forms of spending, some mentioned by Mian and Sufi above.

    So what’s to stop increasing household spending by cutting taxes? Absolutely nothing! And for those of a left of centre persuasion, what’s to stop government upping its spending? Absolutely nothing!

    Of course there are the economic illiterates who think that increasing government net spending increases the debt. The reality (as Keynes pointed out) is that increased government spending can be funded by new money. And as long as that “print and spend” policy doesn’t go too far, inflation will not ensue.

    Indeed, new money has been shoveled into the economy in unprecedented amounts via QE over the last two years, and hyperinflation is nowhere to be seen.

  8. jcb on March 25, 2014 at 10:18 ami

    By this logic, we ought to have re-attained the pre-1929 trend line of retail spending after the first Great Depression, including the WW2 boom. Otherwise we would have been below economic capacity since the last depression.

    Run a trend line for an identical period (1914-1929). Were we at trend in 1945? in 2007?

    It seems to me that a “trend line” is an artificial construct that depends on largely arbitrary beginning and end points. If the end point is a bubble, then the years after will automatically be “below trend.”

  9. Abraham on March 25, 2014 at 9:11 pmi

    Again, I just do not get it! First, the slope of the two lines in household spending is the same after 2009. Therefore household spending is growing at the pre-crises trend rate of growth. In other words, the recent gap is constant in percentage terms. I do not think that is so bad. Simply look at the sharp break (reversal) in the graph! Second, I have a hard time with the claim that the Fed is not doing enough. Well. interest rates are at historically low levels, the federal funds rate remains at the zero-lower-bound, the balance sheet of the Fed has increased by 400% and there is a threat of deflation. Have you heard of the liquidity trap? Third, is not the fiscal cliff a drag on growth? What about US total factor productivity? What about the hysteresis effect of the greatest recession since the great depression? What about the fragility of the financial system? What about a depressed demand for loans as a result of incomplete deleveraging? I wonder if the authors would lend me money at todays interest rates to buy a home when I am unemployed and I am still overly indebted?

    As the authors suggest, what the Fed has done so far has proven ineffective! Yes, that is the painful thruth, whether we like it or not. Monetary policy has reached its limit. There is no monetary or fiscal space left. Just imagine the crude counterfactual of the Fed had done nothing or really caring about inflation: positive federal funds rates in real terms, no quantiative easing at all. Imagine this together with a much more pressing fiscal situation and no deleveraging of households had taken place. The “new deal” is not working anymore. Keynesians are dead! The classical and monetary economists too! The only way out of this crisis is not to make things worse. Market forces must be left to do the job (otherwise we will really be all dead in the long-run!). Things must settle down by themselves. More government interventionism would do more harm than good.