A new “big data” index of U.S. economic activity

The Fed’s Response to Economic News Explains the ‘Fed Information Effect’

“…In contrast to the Fed information effect, the Fed response to news channel is consistent with all of the evidence in Tables 2-4. The last column of Table 3 reports the business cycle indicator variable from Brave et al (2019), which provides a univariate summary of the state of the business cycle based on a wide variety of economic data releases, with higher numbers indicating a stronger economy. Note that all five observations in Table 3 with negative policy surprises and downward GDP forecast revisions occurred when the economy was weak.…”

“…This has the advantage of being a very simple yet informative measure of economic news for the month, since the nonfarm payrolls release is widely regarded by financial market participants as the most informative of all the macroeconomic data releases (e.g., Andersen and Bollerslev, 1998). 29 Second, we consider a more comprehensive measure of economic news, the "big data" business cycle indicator of Brave et al (2019). This index incorporates extensive information from all of the major statistical releases each month to come up with a single measure of economic activity; it is thus much broader than nonfarm payrolls alone, but has the disadvantage of including data that are released after the FOMC announcement as well as before.…”

“…30 Note that the Brave et al (2019) index for month t − 1 is computed using data released in month t and is reported by the Chicago Fed at the beginning of month t + 1. In regression (6), we use the Brave et al index for month t − 1 as our measure of news t , since it is computed using the economic data that was released in month t. 1990-6/2019, excl.…”

“…Table B.3: Economic News Drives Out the "Fed Information Effect" (excluding Brave et alSame asTable 8except that the variable news t excludes theBrave et al (2019) index. See notes toTable 8for details.…”

Dating COVID-Induced Recession in the U.S

“…We perform a real-time analysis of the health of the U.S. economy induced by the coronavirus. To this end, we take advantage of the Bayesian sequential quickest detection (Bsquid) method of Li, Sheng, and Yang (2020) and a new 'big data' index of economic activity proposed by Brave, Butters, and Kelley (2019). The Bsquid method accurately identifies business cycle peaks and troughs for the five most recent recessions, but within a much faster time frame than official announcements by the National Bureau of Economic Research (NBER).…”