Friday, April 22, 2016

The Problem with Macro in one blogpost

Apparently, Nick Rowe slipped up and wrote something about Macro in plain English. David Andalfatto was understandably quite upset and tried to fix things by adding "a bit of formalism".

Then away he goes:

There is a representative agent (this is not necessary, but makes things easy) with additively-separable log preferences defined over consumption sequences {c(t), t = 0,1,...,∞}, with discount factor β. Let R(t) denote the gross real rate of interest (risk-free) earned on a bond held from date t to date t+1. Assume that all individuals can borrow/lend freely at the risk-free rate.

and then this:

Let me consider an endowment economy where each individual is endowed with a deterministic sequence {y(t), t = 0,1,...,∞}.

OK, everybody got that. Representative agent? check. Perfect capital markets? check, lifetime income fixed and known with certainty? check. Time-separable preferences? check.


People, it would be one thing if models like this fit the data, but they don't.

The consumption CAPM is not an accurate predictor of asset prices, The degree of risk aversion required to make the numbers work in the equity premium puzzle is something on the order of 25 or above, the literature is littered with papers rejecting PIH.

So we are being harangued by a model that is unrealistic in the theory and inaccurate to the extreme in its predictions.

And that's pretty much modern macro in a freakin' nutshell.

Mamba out.

Monday, April 18, 2016

Headline Meme: Florida Memorial Edition

People, eating Floridians is tough work. Just ask the tiger referred to in this classic headline:

Tiger recovering after killing Florida zoo worker

Sunday, April 10, 2016

Angus 3:16: Careful with that Ax, Eugene!

I see that "improve your WB ease of doing business score" policy advice has cropped up again on the interwebs (I'm lookin' at you, Alex T.).

In my own special bombastic way, I've had a lot to say on this topic so I thought I'd collect much of it here and expand on the issue/problem.

1. To the extent there is a correlation between the "doing business" index and per-capita GDP, it is a very loose relationship whereby countries with very similar scores have very different outcomes.

Sadly, this is true of almost all institutional indicators, as Hausmann, Pritchett and Rodrik showed us lo these many years ago.

2. Even though there may be a historical correlation between "doing business" and incomes, it is far from clear that it is either causal or exploitable. The exploitability point is kind of a Lucas Critique point. If we see such a reduced form relationship in historical data, treating it as exploitable (i.e. targeting index improvements as a path to higher incomes) is pretty risky. Incredibly, some countries are making improving their scores a policy goal. Ricardo Hausmann has a nice piece on why this can be problematic. And, Matt Andrews has written a book about how countries are adopting "good looking" but not actually good, governance.

3. As Dreimeier & Pritchett show, it is often the case that the outcomes reported by the index, which is compiled by surveying "experts"  are not really related to the actual experience of business people in the country under study. This is the usual de facto vs. de jure issue that plagues many expert compiled indices.

People, we know that North Korea is poor and South Korea is rich. We know that East Germany was poor when West Germany was rich. But we really do not know in any reliable way, what macro policy advice will ensure development success. Really and truly. "Adopt the observed policies and forms of the rich countries" has been offered as advice by the IFIs for decades with very uneven results. Measures of institutions across countries are converging, but per-capita output across countries is not.

Friday, April 08, 2016

Hey Trump and Sanders: Manufacturing's share of total employment peaked in 1943!!!

I got these data series from Fred, but I am too dumb to get Fred to graph the ratio so I put them into Excel, created the ratio, graphed it and exported to this blogpost. Lost the horizontal axis somewhere along the way but it runs from January 1939 to March 2016. The exact variables are listed below the graph.

People, Manufacturing share of Total non farm employment peaked in November of 1943 (the great war!!) and has by and large been falling ever since. 

It's hard for me, given the historical trend, to put too much blame for the relative shrinking of manufacturing employment on the rise of China, or NAFTA, because it was happening in the 1960s and 1970s well before either of these things happened.

Saturday, April 02, 2016

What if it had been Kremer, Miguel & Trump?

Then last year, things might have gone down a bit like this:

"Some guy is saying a reanalysis of a key study argued that there was no clear evidence for improvement in either school attendance or examination performance when using year-stratified cluster-summary analysis. First of all. Just so you understand, this guy is a total loser. He begged me to be his peer reviewer, I said "NO THANKS." Pathetic!

Second, I never said that about the year-stratified cluster-summary analysis. Never said it! Here's the thing. I use the year-stratified regression models and you know what? It works, it just works. I'm winning, my p-values are so much lower than his, this guy's a joke. And he says the cluster summary! Liar, you can't listen to him, this guy's a degenerate and I don't even need to tell you what he did with the standard errors. I mean, there are things about this guy that I don’t want to talk about. You can look. But I’m not going to talk about them. Very bad guy! Our study, it's terrific. No one else will tell it like it is, but we're telling it like it is. We're going to beat those worms. Terrific!"

hat tip to @Zettel314