Thread by @ScottIrwinUI

1. Weekend Reading: Continuing theme of the value of USDA reports. Fast forwarding to recent research rather than my older stuff. Up next is, "Are USDA reports still news to changing crop markets?" 2019 Food Policy paper. Several great co-authors. sciencedirect.com/science/articl…

2. Background: old question with a new twist. People have been questioning the value of USDA crop reports forever. Been examining that question pretty much my entire academic career. In the last decade or so a new twist on that line of attack emerged.

3. Last decade has seen the rise of "big data" companies that use satellite remote sensing data, lots and lots and lots of other weather data, and super sophisticated statistical modeling in order to forecast acreage and crop production. Some see this as making the USDA obsolete

4. Well, if the big data revolution has made USDA crop reports obsolete, then it should be obvious in the reaction of grain futures markets to the release of USDA reports. Obsolete = no impact. That is testable.

5. We wanted to do a comprehensive test of all major USDA acreage, stock, and production reports for grain markets. Since some of the reports are released on the say day we could not easily use price volatility tests. Need market surprises for each report.

6. Fortunately, I had quite a bit of the market expectation data laying around an industry friend helped with the rest. A big thanks to that person! Charts show Average absolute size of market surprises for all major reports from 1984-2016 for corn, soybeans, and wheat.

7. First major finding was that absolute size of market surprises for the different USDA reports did not decline over time, which you would expect to see if the big data hypothesis was correct.

8. We then ran price reaction tests using the market surprises. Really nice models because we could disentangle price impact for all the major reports in one model. That had never been done before.

9. What did we find? One example here using rolling windows for the price reaction tests. Either find no change in market impact of USDA reports or that impact increased per unit of surprise in recent years. Take that bid data hypothesis!

10. Lots of evidence in the paper that USDA acreage, stock, and production reports have INCREASED in importance to the market in the last decade. This actually makes sense economically given the increased market volatility since 2006.

11. I have been incredibly fortunate to work with truly great ag economists during my career. This paper is a wonderful example. Olga Isengildina-Massa of VPI and Berna Karali of Georgia did the heavy lifting on the paper and I was lucky to be along for the ride. Thanks!

@threadreaderapp unroll