Sunday, February 3, 2013

Taking Apart 2012 IV Growth

The initial news last week was shocking: real GDP growth in 2012 IV was negative.

Follow-ups in the legacy media downplayed the overall number, and claimed that the components of that number looked good (here’s some of what was said by The New York Times, The Wall Street Journal, and Forbes). Of course, The White House Blog painted a bright picture as well.

I’m usually pretty sanguine about this sort of thing. Not this time … I’ll start with two quick thoughts, and then end with a longer one.

1) Despite the names peak and trough, it’s well known among macroeconomists that peaks tend to be smoother (and thus harder to spot in real time) while troughs are sharper (and easier to spot in real time). Given that we’re in an expansion, a modestly weak quarter is precisely what a peak would look like.

2) By definition, advance estimates are rough drafts. The BEA doesn’t keep a comprehensive data set of past advance estimates, but they do have a download of all announcements and revisions over the last 44 quarters. How often did the advance estimate come up negative without a peak eventually being declared by the NBER Business Cycle Dating Committee? Not once. Now … bear in mind that this isn’t a huge dataset, and it only has one peak in it.

3) Here’s the longer take. The apologists for last quarter have noted that while the overall sum of growth was negative, the important components looked good. I don’t think this is so. Here’s why.

The story usually goes something like this: the big important components were OK, but they were dragged down by a few bad outliers. Fair enough — let’s go look at the components.

GDP growth is typically broken down into 4 major components: consumption, investment, government spending, and net exports. Consumption is the largest, investment is the most volatile.

Then I went and gathered data on the contribution of those 4 components to reported growth for the last 252 quarters (going back to 1950).

Component Contribution
Consumption 1.5
Investment -0.1
Government Spending -1.3
Net Exports -0.3
Real GDP -0.1

Then, I took all 4 elements for those 252 quarters, and ranked them as percentiles. The table shows the medians and interquartile range.

Component 25th Percentile Median 75th Percentile
Consumption 61 75 89
Investment 14 52 81
Government Spending 29 44 59
Net Exports 17 31 48

Clearly, consumption is always the most important component. It is also the least volatile. Investment is the second most important, but clearly the most volatile.

Lastly, I took the 4 components from the first table, and calculated their percentiles for comparison with the second table.










Government Spending



Net Exports



Real GDP



This is far more sobering. Yes, government spending is an outlier — but every component of real GDP growth was below median last quarter!

When had that happened before? The most recent time was 2008 IV – during the worst part of the Great Recession. It also happened in 2007 I and 2006 III — the first two quarters when it looked like something was going wrong with the Bush expansion. Also in 2001 III — the quarter when 9/11 occurred. And 1982 I — the worst quarter of the 1981-2 recession (which was arguably worse than the Great Recession). And 1974 III — right in the middle of the 1973-5 recession (which again, was arguably worse than the Great Recession). And 1970 IV — the trough of the mild 1969-70 recession. And in 1957 II and 1958 I — during the 1957-8 recession (which no expert considers a mild one).

The only false alarms from this signal are in 1968 IV and 1988 III. Both are quarters in which long and strong expansions appeared to falter before recovering their strength.

I can drill down even further. Those 4 components are each divided into two major subcomponents. Of those 8, only Fixed Investment was well above its median percentile contribution. Goods were at the 72nd percentile, while their median in the 71st percentile. And … one of the biggest contributors to making the growth rate higher was Imports. The thing is, Imports get subtracted out of growth … so a bigger contribution to growth means that we didn’t feel flush enough to buy as much foreign made stuff as we usually do. That isn’t good.

At the next level of fineness, there’s 14 components. And what actually was better than the median? Consumption of durables (but not non-durables or services), investment in equipment for firms and residences (but not non-residential structures), and imports of goods and services (both are bad signs).

The bottom line is that it is a very bad sign that the economy was weak across the board in the last quarter.

No comments:

Post a Comment