More on the Output Gap

As I said, Bullard’s thinking on the output gap mirrors that of a lot folks.

Tyler Cowen points to this from Sober Outlook.

A key part of the post says

This output gap, thought often taken for granted, may in fact be much smaller. Some recent work by Barclays Capital has already shown that the potential GDP (red line) is basically extrapolating the bubble of the pre-crisis era and is therefore unreliable.

The CBO number crunchers wish to believe in their ability to compute the potential GDP of the US economy.  The reality is that nobody knows how much output the US economy is truly capable of. But there are signs that the gap may now be below the CBO’s assertion.

In one sense the bolded paragraph is true. No one knows for sure exactly what you could get out of the US economy, especially if you want to get the highest value added production. This is part because its not completely clear what “highest valued added production” might even mean in thoroughly consistent way.

However, it we take a step back and ask more or less how much could you squeeze out the US economy, in terms of easily recognizable production we could probably back of the envelope this without too, too much trouble.

For example, if we started building lots of cars at roughly the prevailing mix as in 2007, it would be difficult to come up with a concept of “highest value added production” under which that would be the right thing to do.

If nothing else overwhelmingly the odds are that the mix of cars should be tilted slightly more towards high fuel efficiency models if one wanted to maximize value.

Nonetheless, if we produced 600K F-150s we can be pretty confident that we could exchange the marginal F-150 for not too many fewer unskilled labor hours than we could in 2007.

That is, we take the very last F-150 to roll off the lot and we put out the call to America and say, of everyone out there who is willing to work the most hours for me in exchange for this truck. Its unlikely the number will be much lower than it was in 2007.

Which in economic terms is to say if we use unskilled labor – which is raw human time or the basic unit of life – as the numeraire good, the marginal valuation of an F-150 at 600K units is not likely to be that much less than it was before the crisis.

All that is to say that we could back of the envelope this sucker and we wouldn’t be too far off.

However, I think the even deeper point comes from a self-link in the post

Barclays Capital: …the CBO estimates that output was at potential during the housing bubble years and that any deviation from the trend established during those years represents an output gap. In contrast, our view is that the housing bubble pushed the economy above its potential; thus, we believe the output gap is much smaller.

In fact if one zooms in on this chart, it starts to make much more sense than the original BEA/CBO analysis that assumed the US was producing at capacity prior to the financial crisis. The Barclays chart shows that economic output between 2004 and 2007 had been growing at a rate above what it is normally capable of. One can only interpret this production level above capacity as simply funded by credit, particularly real estate driven credit.

The key question – and maybe someone at Barclay’s wants to have this conversation – is what reason is there to believe that the housing bubble punched the economy above its potential?

My main reason for asking is simply that I just don’t know of any actual empirical evidence to suggest this is the case or what a consistent narrative about this would be.

However, to nudge folks forward into the conversation I will note that standard economics – sans Barro -  says the exact opposite. A housing bubble should drive output below its long run trend.  Ditto for a credit bubble.

Why?

Well, because as people got wealthier they would attempt to cash in on some of that wealth. IE, cash-out-refinances, etc.

They would then use those funds to buy things.

One of the things people like to buy is not working or investing in the future. Both working and investing in the future are costly and folks only do it because there is a return – which is determined by the actual marginal products of capital and labor.

As you get wealthier the marginal benefit of that return is worth less to you and so you do less of it. You work less. You invest less.

Working less directly subtracts output from the economy. Thus making it smaller.

Investing less leaves us with a smaller capital stock meaning we are able to produce less in the future, thus subtracting from future output.

So, both of these factors depress output.

Now, what happens when the bubble pops. In a smoothly functioning economy those trends reverse. First, you start saving more because suddenly you are not as wealthy as you thought.

Second, you start working more to because (A) You can’t live off cash-outs anymore and (B) you have to repay the cash-out you already performed, when you thought you were richer.

In this case what we would see is production roar back towards the trend line it was on before.

In a fundamental way we would say that the output (the potential output) of the economy is determined by tastes and technology, neither of which has changed. And, so the long run path of the economy has not changed.

The only alternative is to suggest – as Stephen Williamson might be – that the housing bubble or its predicates were themselves a technology that in fact boosted the productive potential of the economy.

Once they are gone we are left with an inferior technological stock.

To be more specific, the idea would be that collateral is a productive asset solely on the basis of it being collateral because it facilitates mutually beneficial trades that could otherwise have not occurred.

I am sympathetic to this view. Indeed, I used to quasi-hold it and I still think its qualitatively true. I just don’t think that its quantitatively important. The stream of housing services that could be obtained in a high collateral world is higher, but its not that much higher and we are still receiving the stream.