Does a time gap between your instrument and endogenous regressor matter?*

Was browsing the 2015 NEUDC program for potential seminar speakers and came across this fascinating paper now titled “Instrumental Variables in the Long Run” by Casey & Klemp.

The piece argues that the large literature using IV to estimate the “long run causal effect of institutions on growth” may be dramatically overstating the effect!

This is due to the fact that in,

” IV regressions where the endogenous variable is measured later in time [KG here: they mean later than the instrument is measured] estimate the long-run effect divided by the persistence of the endogenous variable. We define ‘persistence’ as the causal effect of historical levels of the endogenous variable on its current level. Our analysis, therefore, shows that accounting for the persistence in the endogenous variable is crucial for estimating long-run causal effects when the endogenous variable is observed after the effect of the instrument.”

 

An example of the situation they are describing could be using settler mortality in 1700 as the IV for institutions in 1990 (sound familiar people??)

The upshot for the institutions and growth literature is that institutions like constraints on the executive have low persistence so the IV coefficient on them is large for the wrong reason.

The authors employ their technique that accounts for the persistence in the endogenous regressor and find,

“In our preferred specification, a change in constraints on executive power in 1700 from the lowest to the highest possible score on the standard index leads to a 1.2 standard deviation change in 1990 income per capita. While sizable, this effect is less than half as large as the coefficient generated by the conventional IV regressions.”

 

Wow. this is a very cool paper. Highly recommended.

*Better Title:  “Who is going to make it? We’ll find out in the long run. We can handle some resistance if our love is a strong one.”

PS: note that the literature on institutions and growth is actually a literature on institutions and income levels. I have never understood the universal misclassification here.

2 thoughts on “Does a time gap between your instrument and endogenous regressor matter?*

  1. Pingback: Tuesday assorted links - Marginal REVOLUTION

  2. Pingback: Links round-up | The CSAE Blog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s