Beyond the Streetlight
by Byrne, Corrado, Kennickell
REMARKs by Christopher D. Carroll
Formerly of the Division of Research and Statistics: 1990-95
Johns Hopkins University (1995-2023)
Federal Reserve RS100 Conference
November 8, 2023
https://econ-ark.org/materials/beyond-the-streetlight
Thank you to all Federal Reserve Staff
- Past, Present, and Future
- For doing a vitally important job as well as humanly possible
Measurement
When I was on the job market, measurement was uncool (in academia)
- "Theory Before Measurement ..."
- Wilcox (1992): 'researchers spend little effort [to understand the data]'
But at the Fed
- measurement was ALWAYS cool!
If I had to choose between data and theory
- It's data all the way!
Measurement = Information Production
Primary:
- Survey of Consumer Finances
Secondary:
- Industrial Production
- Capacity Utilization
I would add a Tertiary:
- books $\in \{$Green, Teal, Beige$\}$
- staff memos ...
What Good Is Information? (Part 1)
Greenspan Fed Called the 'new economy'
- 1997 ERP: 1994-96 pty growth 1.3 pp > NIPA
- (Steve Braun, R&S alum, made these calcs)
Proposition Drew a LOT of Criticism
- From mainstream economists (esp. of type Wilcox (1992) criticizes)
- From political left and right
What Good Did It Do?
- Allowed room for monetary ease in East Asian Crisis (1997)
- 'opportunistic' disinflation
Was Greenspan Fed Just Lucky?
- Or has the Fed gotten better over time?
- That's one point of better measurement
- Maybe everybody has gotten better over time?
- Measurement improvements make everybody better off
2.5 bp increase in FF in Will Du's 'HANK and SAM' model.
What Good Is Information? (Part 2)
Utility cost of business cycles directly proportional to c
- $\Delta V \approx u^{\prime}(c) \times \Delta c $
- Better measurement $\Rightarrow$ less uncertainty $\Rightarrow$ higher welfare
- LLM/NLP for relevant corpuses and search terms
- Better micro data
- linked to macro through RDC