Big Data: Central Banking in the New Decade
As the United States nears eleven consecutive years of expansion, a national record, the Federal Reserve is determined to prolong the country’s growth. Consecutive FOMC statements have reiterated the willingness of the Fed to act as appropriate given new developments in the economy. Right now, however, it seems the appropriate thing to do is nothing. Hemmed in by the uncertainties of a global trade war that shows no meaningful sign of abating, the Fed has adopted a wait and see approach to monetary policy.
After years at the zero-lower bound, the path of the fed funds rates was quite clear in the mid-2010’s: up. As unemployment fell to rates last seen in the 1960’s and inflation failed to reach the Fed’s symmetrical target of 2%, the path grew more uncertain (St. Louis FRED). Having relied heavily on forward guidance as a policy tool for the last decade, the Fed suddenly had none to give. Financial markets that live and die by the Fed were just as confused, offering diverse–to put it politely–guesses as to where interest rates were headed.
Although it was apparent economic growth was slowing as a result of trade uncertainties and inflation remained below target, strong job reports and consumer confidence confounded the decision making of the FOMC. It certainly did not help that a record-long government shutdown in early 2019 prevented timely compilation and release of economic data the Fed relies on to shape its monetary policy decisions. Fortunately for the Fed however, things have changed since the last government shutdown. The exponential increase in data collection by companies and use of big data meant the Fed could procure similar data from private sources. Using consumer spending figures from First Data, a card payment processing company, the Fed was able to see consumer confidence was strong despite the shutdown and larger economic uncertainty. Government data released afterwards reached broadly similar conclusions according to a study by Fed economists (Aladangady et al., 2019).
Although the Fed already uses private data, such as the PMI from ISM and Case-Shiller from S&P, the experiences of the last government shutdown show the many unexplored ways the central bank can refine its understanding of the US economy. One of the biggest advantages is timeliness of the information. Using big data, the Federal Reserve can spot economic trends far faster than official government data. Many federal reports only update once a month while official GDP figures arrive quarterly. Payroll processors on the other hand update payroll data weekly, providing the Fed with higher frequency data and timelier updates that improves forecast accuracy and enables more refined adjustments to past data (Cajner et al., 2018).
Employment, wage, and payroll data are some of the most important economic indicators utilized by the Fed and help to paint an accurate picture of the economy at any given point. In a speech last October, Chairman Powell suggested the “new measure, had it been available in 2008, would have been much closer to the revised data, alerting us that the job situation might be considerably worse than the official data suggested.” Given such knowledge and assuming other indicators could have been similarly tracked, then-Chair Ben Bernanke might have been less confident in claiming the subprime mortgage crisis was self-contained and unlikely to spillover to the broader economy (Bernanke, 2007).
Another advantage of utilizing data from firms like ADP and First Data is the significantly increased level of granularity. Although the Federal Reserve sets monetary policy appropriate for the US economy at large, using nationwide trends as reported by government data can obscure localized developments and increase susceptibility to one-off events. For example, Aladangady et al. used payment processing data to analyze the effects of hurricanes in 2017 on spending. Using similar data to analyze localized spending data before and after a major event, economists would be much better able to assess whether GDP variations were the aftermath of an isolated event or the beginnings of an economic trend. Given the increased focus on pulling workers from the fringes of the economy and concerns about the appropriateness of policy by demographic, granularity also allows the Fed to see where the weakest regions and industries are long before official data.
Beyond the benefits in terms of speed and granularity, big data also decreases the amount of effort involved in collecting the data. Many official reports depend on survey data or physical visits by government staff. Surveys have suffered from declining response rates, both from households and businesses. Combined with demographic trends such as the decline in use of landlines, government agencies will only encounter increasing difficulties in procuring representative data using surveys. Physical visits aren’t difficult given the multitude of transportation options but are time-consuming, manpower-heavy, and limit the frequency with which data, such as the CPI, can be updated. Big data will decrease the effort involved in traditional data collection, provide real-time updates on spending trends from payment processors, and significantly decrease the amount of personnel required to maintain relevant information.
There are concerns associated with the Fed utilizing private data when making monetary policy decisions. Although the instances mentioned in this article show broad similarity between the private and public sources, it does not necessarily follow that will be the case for all the economic indicators used by the Fed. While the additional information is undoubtedly helpful in refining the Fed’s picture of the economy, it may exacerbate volatility if short-term developments turn out to have no long-term economic impact. In the most likely scenario, big data from private sources can be used to complement the multitude of economic indicators already tracked by the government.
Another concern is the cost of accessing private sources of information. The Federal Reserve paid for access to the First Data economic figures, which has since elapsed. While the ADP worked with the economists at the Fed for their study, there is no reason to believe the company would allow continued access without some sort of financial arrangement. If the central bank comes to rely on private sources for their economic data, they may eventually find the financial costs of access too high. Given the uniqueness of the data such as payment processing information, the Fed is likely to have trouble replicating it on their own.
Even given these concerns, big data’s influence on monetary policy decision making is likely to increase as time goes on. Chairman Powell has already publicly discussed exploring uses of big data to help provide timely updates on economic data. While analysts, pundits, and researchers have written much about the rise of big data and AI for commercial use, American government usage has been much less talked about. There may be no better place to start than the public institution that likes to say “monetary policy is data dependent” (Powell, 2019).
Works Cited
Aladangady, Aditya, Shifrah Aron-Dine, Wendy Dunn, Laura Feiveson, Paul Lengermann, and Claudia Sahm (2019). “From Transactions Data to Economic Statistics: Constructing Real-time, High-frequency, Geographic Measures of Consumer Spending,” Finance and Economics Discussion Series 2019-057. Washington: Board of Governors of the Federal Reserve System, https://doi.org/10.17016/FEDS.2019.057.
Bernanke, Ben S. “The Subprime Mortgage Market.” Board of Governors of the Federal Reserve System, 17 May 2017, www.federalreserve.gov/newsevents/speech/bernanke20070517a.htm.
Cajner, Tomaz, Leland Crane, Ryan Decker, Adrian Hamins-Puertolas, Christopher Kurz, and Tyler Radler (2018). “Using Payroll Processor Microdata to Measure Aggregate Labor Market Activity,” Finance and Economics Discussion Series 2018-005. Washington: Board of Governors of the Federal Reserve System, https://doi.org/10.17016/FEDS.2018.005.
“Consumer Price Index Frequently Asked Questions.” U.S. Bureau of Labor Statistics, U.S. Bureau of Labor Statistics, 25 Apr. 2019, www.bls.gov/cpi/questions-and-answers.htm.
“Federal Reserve Issues FOMC Statement.” Board of Governors of the Federal Reserve System, www.federalreserve.gov/newsevents/pressreleases/monetary20200129a.htm.
Powell, Jerome. “Speech by Chair Powell on Data-Dependent Monetary Policy in an Evolving Economy.” Board of Governors of the Federal Reserve System, 8 Oct. 2019, www.federalreserve.gov/newsevents/speech/powell20191008a.htm.
Rugaber, Christopher. “How's the Economy? Fed Increasingly Turns to Private Data.” AP NEWS, Associated Press, 5 Feb. 2020, apnews.com/843c9b4096c78fbe93165ed4f792af57.
“Unemployment Rate.” FRED, 7 Feb. 2020, fred.stlouisfed.org/series/UNRATE/.