I figured since I made fun of John Cochrane's defense of Jeb Bush's claim of 4% growth, I should maintain some balance and make fun of Bernie Sanders' economic proposals (see e.g. here or here). From the second link, we can get derive an estimate of the projected civilian labor force (CLF) and use the "quantity theory of labor" model to estimate the NGDP growth we'd expect. While the Sanders version only says NGDP growth would be 8.7% ~~5.3%~~, the labor force increases would actually imply an even higher NGDP growth of 10.5% (averaged over 2016-2026):

The dotted lines are the log-linear extrapolations for the IT model (blue) and the data (yellow), the projection based on the CLF figures derived from the Sanders campaign (green, BS) as well as the NGDP number of ~~5.3%~~ from the campaign (red, BS₀). Note that this model also implies inflation would be 7.9% (and therefore average real growth would be 2.7%).

...

I read the 5.3% as a nominal growth figure (you can see why I'd make such a mistake!); nominal growth in Gerald Friedman's projection to 2026 is more like 8.7% (to 43 trillion from 18 trillion) -- and therefore 3.4% inflation.~~Disregard the red line in the above graph~~. Edited the above post to reflect this.

Here is some good commentary at EconoSpeak on the projections themselves.

Corrected the red line (finally).

...

**Update 19 Feb 2016**I read the 5.3% as a nominal growth figure (you can see why I'd make such a mistake!); nominal growth in Gerald Friedman's projection to 2026 is more like 8.7% (to 43 trillion from 18 trillion) -- and therefore 3.4% inflation.

**Update 20 Feb 2016**Here is some good commentary at EconoSpeak on the projections themselves.

**Update 22 Feb 2016**Corrected the red line (finally).

Jason,

ReplyDeleteIf the CLF grows extremely quickly without corresponding real growth, is there just some massive increase in unemployment, or is the capital stock falling really fast?

I would think that the ITM Solow model would exhibit some

interestingpredictions, given "2.7%" real growth and extremely high CLF growth.True!

DeleteThe real result is that the "QTL" underestimates real growth and overestimates inflation.

I do want to try the more accurate model and see what comes out.

At first I thought you were calling two of the projections "Bull Shit" ... and then I realized what it stood for. Unfortunate initials for Bernie...

ReplyDeleteIt was kind of a double entendre.

DeleteWell it may be BS but if we could get real growth to 2.7% with inflation of 8% we could easily inflate away our debt. A bit OT, but I am going to send you a very interesting graph showing the very strong correlation between M0 growth and inflation across countries. If this is as strong as it looks, and the ITM model (to my understanding) also predicts that M0 printing should lead to inflation, it seems that countries could easily get whatever inflation rate they want just by tailoring M0 growth rates. Would this not be a good thing for Japan?

ReplyDeleteM0 growth only leads to inflation if k > 1. In Japan, M0 expansion leads to deflation.

DeleteAlso, see this:

http://informationtransfereconomics.blogspot.com/2015/06/the-quantity-theory-of-money-as.html

Wait- then how does M0 growth compare to "helicopter drop", where my understanding was that excess money is printed and spent either on transfers to the population or on infrastructure. As I recall, you mentioned previously in the ITM framework that spending money to increase demand would likely result in an increase in inflation.

DeleteTodd,

DeleteYou need to keep track of magnitudes. The US has k > 1, but it is close to 1, so printing money doesn't cause much inflation because log P ~ (k - 1) log M0. Giving it to people would be money-financed fiscal stimulus, which would raise NGDP by some amount. The NGDP increase would also increase inflation a bit, but that goes as

logNGDP ... so it would be tiny unless you did a "helicopter drop" that was large compared to NGDP.O/T: Feynman's wrong here, isn't he? 400 millionths of a centimeter to 700 millionths of a centimeter is 4-7 um which is mid-wavelength infrared (i.e. not visible), isn't it?

ReplyDelete