More on the GDP Mystery

Our last post showed the recovery in GDP since the Great Recession has been abysmal, and we argued that it is a mystery that requires some deep thought and additional research. Let’s go a bit further into the GDP data to try to figure out what is going on.

Breaking out the GDP components is always a bit tricky because they depend on one another. For example, business investment is a function of household consumption. If business investment is weak, it could be because consumption is weak. Nonetheless, there is some useful information when we break out where the GDP weakness comes from.

Ok, now for the charts.

Let’s start with the primary culprit: consumption of services and non-durable goods. They are shockingly weak relative to other recoveries.




Consumption of durable goods is actually doing fairly well, which may reflect aggressive monetary policy that has spurred borrowing especially in the auto market:


It’s still not great, but it looks quite a bit better than consumption of services and non-durable goods.

Many point to weakness in residential investment as the primary culprit. It looks pretty bad, but there are other recessions that had comparably slow growth in residential investment after the recession.


Non-residential investment is a bit tricky, because it is almost a direct function of expected household demand for goods and services. If no one is buying a company’s products, then of course they aren’t going to investment in equipment and structures! So we are not sure how much there is to learn here, but here is the chart:


It looks pretty weak, but again that shouldn’t be a surprise. In any case, it doesn’t look as weak as consumption of services and non-durable goods, which really stands out.

Finally, here is government expenditures and investment:


This one surprised us. Government expenditure and investment have definitely been a drag recently, but they don’t stand out as much as the weakness in consumption in services and non-durables.

Let us repeat: this type of exercise is tricky because all of these components are a function of the other ones — the macro-economy cannot be neatly divided into parts. But there is some valuable information here. The mystery of weak GDP growth over the past 6 years is closely related to consumption, particularly consumption of services and non-durable goods.

Bookmark and Share

20 Responses to More on the GDP Mystery

  1. JMarco on May 1, 2014 at 12:25 pmi

    Ah! how about the money or credit needed to have more consumption? Can you chart wages, wages after taxes, consummer credit, mortgages and then foreclosures during these same periods?

    • Michael on May 2, 2014 at 11:21 ami

      I’ll do you one better. Here is a graph of wages and net borrowing as a % of GDP on the left scale, and the output gap on the right scale. Note the incredible correlation.

      Wages have been falling for 30+ years. (It’s time to reverse our supply-side policies and get back to demand-side policies). Until 2007, household borrowing was able to compensate, but once the housing ATM closed down, we’ve been trying to run on wages alone. But they are too low.

      Wages are almost $1T below their long-term average, while Corporate Profits are almost $1T above their long-term average. Mystery solved. Wages create demand, corporate profits don’t.

  2. Up the down escalator on May 1, 2014 at 1:07 pmi

    You omitted one of the most glaring problems – govt spending after 2007 is the only trend this late in the game with negative slope. And there are no signs in the slope of a turn a around. Yeah, there will be ramp down with cut stimulus but that was years ago. Our population is still growing, in the long run govt spending should track with population growth.

  3. Pinkybum on May 1, 2014 at 2:29 pmi

    Have you two been living under a rock for the last 5 years? There is a massive lack of demand in the economy plain and simple. You also don’t seem to be able to see that the for each series the current recession is less than or only equal to ANY of the previous recessions. For example take 1960 where the government expenditures are about the same after 6 years but Residential Investment was going gang-busters – therefore compensating. NO sector of the economy now is compensating for any other sector of the economy – hence the overall very poor recovery.

  4. Thomas Mullaney on May 1, 2014 at 2:45 pmi

    Would Irving Fisher be so puzzled?

    How do we look compared to recoveries, ours and others, after finaccial crises?

  5. Hugo André on May 1, 2014 at 3:49 pmi

    I find the consumption of services chart somewhat confusing. It shows that the recent recession has been followed by the weakest recovery in consumption of services, the 2001 downturn having the second weakest services recovery with the 1990s recession third weakest.

    What is this all about? I thought we were moving towards an economy that is based more on services and less on manufacturing (and farming) of durable and non-durable goods.

  6. Pinkybum on May 1, 2014 at 5:03 pmi

    I guess I meant 1953 not the 1960 series in the above chart. It’s a bit hard to read series when there are so many!

  7. jonboinAR on May 1, 2014 at 6:35 pmi

    Two terms: Wage stagnation, and unemployment. Ain’t nobody got money to spend on going out to dinner or Six Flags!

  8. Kaleberg on May 1, 2014 at 7:54 pmi

    The causes of the weak recovery are rather obvious to any unreconstructed Keynesian. Are you using the term “mystery” to mean deep or profound or are you actually in the dark?

  9. Philip George on May 2, 2014 at 12:08 ami

    This must be the best blog post I’ve read in the past five years.

    I love the way you throw off gems so casually: “For example, business investment is a function of household consumption. If business investment is weak, it could be because consumption is weak.” (Don’t the guys who draw up IS-LM equations ever think of this?)

    Any details of the raw data used to construct these graphs would be very welcome. I could use them in the next drastically revised version of my book, with acknowledgements of course.

    I’m only surprised that the economic theory behind the weak recovery has not yet occurred to you. It of course calls for throwing out much of current macroeconomics.

  10. Pookah Harvey on May 2, 2014 at 6:34 ami

    As you note “all of these components are a function of the other ones” but we only have direct control of one, government spending.

  11. Girod on May 2, 2014 at 8:52 ami

    I think you guys should find the answer in the decreasing level of private domestic leverage (~180% after ~210% in 2007)…

  12. Jason Smith on May 2, 2014 at 10:31 ami

    It’s a bit of a mystery going back more than just this past recession — the later years tend to be lower than earlier years. Secular stagnation? Persistent liquidity trap (a la Japan)?

    • Michael on May 2, 2014 at 11:27 ami

      The later years had the lowest wages as a % of GDP. That’s why growth has been weaker.

      • Jason Smith on May 3, 2014 at 12:34 pmi

        The question is, is that a large variation? The ratio of nominal wages (NW) to NGDP has been roughly constant relative to other dimensionless indicators like the price level. It does look dramatic if you selectively window around the data.

        The ratio of nominal wages (NW) to the number of employees (L) has been essentially given by the price level (P ~ NW/L)** since the 1960s. This is an aspect of sticky wages — total wages change based on the number of employees, not salary adjustments. However, all this does is bring us back to low inflation as the root cause of the declining wage share of NGDP, which in turn brings us to the question of why many developed countries have entered into deflationary/disinflationary “liquidity traps” or “secular stagnation” (Japan, US, EU)?

        Since NGDP = RGDP x Price level, this effect is also apparent in NGDP, per my comment above.

        ** See the end of this post:

        • Michael on May 4, 2014 at 10:30 ami

          I think you might be over-analyzing the situation. The two biggest components of the National Income are Wages and Corporate Profits. The disappearing wages have gone straight to profits.

  13. Mike on May 2, 2014 at 11:12 ami

    By normalizing to the quarter before the recession you are imbedding a definition of “normalcy” into that quarter. But what if that quarter was part of a bubble? Try this: draw a trend line through post-recession RGDP and extrapolate backwards. You will find the current trend lines up beautifully with the trend during the 1990s expansion.

    That is not to say that the trend today is “good or correct” – I’m not smart enough to know the answer to that. However, if you draw a trend in RGDP from 2000-2007 and extrapolate backwards you find that the 1990s exhibited a huge persistent negative output gap. I tend to recall the 1990s as being a pretty good expansion and hence that version of trend doesn’t fit with my idea of “good.”

    • Atif Mian and Amir Sufi on May 2, 2014 at 12:52 pmi

      See here. Great Recession recovery looks pretty bad even over a long term.

    • Pinkybum on May 2, 2014 at 8:00 pmi

      You have the causation exactly backwards. It would seem (see chart from authors) that the housing boom was only allowing household spending to keep up with the “natural” rate of growth in the economy to maintain “full employment”. When the consumer confidence and earnings were shattered by the housing bust suddenly everybody was saving/not spending and then demand was hit all around. The only agency left to pick up the spending slack – the US government – decided to turn to deficit reduction and spending cuts exactly the opposite of what was required.

  14. Dan on May 2, 2014 at 1:38 pmi

    Agreed it is tricky, and I suppose in that respect you tend to see what you want to see, at least you can see what you want to see.

    But the fall in consumption is explained by deleveraging-rising savings rate. though it is mildly interesting to see straight consumption the worst – that is relative to itself in other recessions, relative to durables, in the same time period it still looks better.

    And still the glaring feature remains – if you have weak household/investment – why is the government not strongly countercyclical?

    I don’t see why the government line is a surprise. The initial response coordinated at the G20 and which the US participated in was to be countercyclical, then in 2010 that was abandoned. that looks pretty much like what the charts show.

    There may be and almost certainly are many things going on here, and as you say there are lots of interactions between the components of GDP, but that is the whole point of countercyclical spending – it is supposed to positively interact with the private side.

    I assume your question is in many ways rhetorical, though also inquiring for deeper insight. But I think the very simple Econ 101 insight would have dramatically changed the picture – and why not start with the low hanging fruit…