Is It Ever Too LATE?

This is just a friendly reminder that even the very best IV analyses can only get us local average treatment effects (LATEs). OLS estimates, on the other hand, are average treatment effects (ATEs). This makes it very difficult to compare the two. See this post by Marc Bellemare for a nice description of the issue. My thoughts: Sure, if you have reason to believe that responses to the treatment will be the same for everyone..or if you believe that everyone in your sample is a "complier," sure, they are comparable. Keep in mind though that either of these instances are likely to be rare. I personally do like to see OLS estimates even in IV papers. But I would urge you to just be careful when making comparisons. The fact that an IV estimate is different from an OLS estimate does NOT necessarily imply that the OLS estimate suffers from endogeneity bias. 

AEJ: Applied vs. AEJ: Policy

What's the difference? Where should you submit? Matt Notowidigdo, AEJ: Policy co-editor, very generously gives us the scoop on twitter. So helpful! I had no idea that AEJ: Policy competes more with AEJ: Macro than AEJ: Applied these days. 

Thank you, Matt, for all of this! 

And now a plea to the editors of other journals: I would be so, so grateful if you could do the same for your journals. I even promise to blog about it. ;) 

Careful with IVs...

You know when you just have a feeling about something. Maybe it's a gut instinct that comes from nowhere, but more often, the feeling comes from your experiences. Anyway, for years I have had that type of feeling about instrumental variables (IV) analyses. Yes, I understand the math behind how IV techniques can work wonders in terms of correcting bias. But after playing with IVs in various contexts, using various data sources, for various papers, etc., I have gotten a sense that IV results are just not as robust as I would like. 

Well, it turns out that my feeling is probably right, and Alwyn Young, an economist at LSE, has shown this formally in a recent paper about inference in IV models. Yes, you should read the paper, but since you're probably busy with writing or grading final exams these days, have a quick look at Marc Bellemare's post about the article. These are some of the most important points, as described in Marc's post: 

1. Conventional tests tend to overreject the null hypothesis that the 2SLS coefficient is equal to zero.

2. 2SLS estimates are falsely declared significant one third to one half of the time, depending on the method used for bootstrapping.

3. The 99-percent confidence intervals (CIs) of those 2SLS estimates include the OLS point estimate over 90 of the time. They include the full OLS 99-percent CI over 75 percent of the time.

4. 2SLS estimates are extremely sensitive to outliers.

So what to do? Does this mean we should abandon IV analyses and just report OLS results? Absolutely not. A good exogenous IV can correct some pretty major problems in the OLS. I guess what I would say is just to be careful. And to always report both the IV and OLS results!

In the meantime, a fun activity: come up with new slogans for OLS! Check out some recent candidates from the world of twitter!  

More Insight on Where to Send Short Papers

Remember I blogged about where to send short papers? That new AER journal specifically for short papers has arrived! Read all about the new journal, American Economic Review: Insights here. 

Short papers or long papers, I am happy that there will be another journal of AER quality! 

Cool Data Alert: Distances Between Zip Codes

Here it is. Thank you to the NBER for putting it together and especially to Jean Roth for creating the Stata files. 

Twitter, Now a Source of Data Not Just Procrastination

If you have Stata 15, you can now very easily import Twitter data directly into Stata using the command twitter2stata. More details here. Tweet that!

Need a Dissertation Topic in Applied Micro?

Browse through this website, Our World in Data. Not only can you check out some really fascinating relationships in the world, but you will get some sense for the data sets that exist out there and what types of information they contain. Anything spark your interest?