Federal Reserve Economic Data

The FRED® Blog

Changes in discretionary spending during the pandemic

Data from the Visa spending momentum index

FRED recently added new data for the Visa Spending Momentum Index (SMI), which can offer new insights into US consumer spending behavior during the COVID-19 pandemic.

The FRED graph above shows two SMI data series* available between January 2014 and May 2024:

    • The red bars represent changes in the momentum of non-discretionary spending categories. Those include medical/health, food, and utilities—among several others.
    • The blue bars represent changes in the momentum of discretionary spending, which is everything else (except for spending on restaurants and gas).

The position of the bars indicates the direction of change in spending momentum: Bars above the zero line signal consumer spending was above trend, and bars below the zero line signal consumer spending was below trend. The length of the bars indicates how big those upswings and downswings were.

During March and April of 2020, discretionary spending markedly slowed down. It remained very close to its trend value for the next 12 months and surged above trend for almost a whole year afterward. Between March 2022 and the time of this writing, the Visa SMI has signaled either at-trend or below-trend discretionary consumer spending.

* Learn more about the units of SMI and the definitions of discretionary and non-discretionary spending. Btw, Restaurant and gas spending are included in a total (all-categories) SMI.

How this graph was created: Search FRED for and select “Visa Spending Momentum Index: Discretionary: United States.” From the “Edit Graph” panel, use the “Add Line” tab to search for and select “Visa Spending Momentum Index: Non-discretionary: United States.” Next, use the “Edit Lines” tab to customize the data by typing the formula “a-100.” Do that for both series. Last, use the “Format” tab to change the graph type to “Bar.”

Suggested by Diego Mendez-Carbajo.

FRASER turns 20

Happy birthday, vintage data! Discover econ history with FRASER and ALFRED

Most data users know that economic time series are periodically corrected and revised, and the data in FRED are no exception. But it’s also true that looking at only the most up-to-date data can skew our perception of past economic conditions. If you’re interested in historical data, join our vintage data party!

Today, July 1, FRED‘s younger sibling FRASER celebrates its 20th birthday. FRASER launched in 2004 as part of the St. Louis Fed’s effort to improve access to historical data for policy analysis and research replicability.

ALFRED is the other half of that historical data project: Each time a data series is updated in FRED, the prior version of the series is stored and made available in ALFRED, with all of its prior vintages of data observations and metadata at specific points in time. (Btw, ALFRED celebrates both its 18th and 19th birthdays this week: Its data were added to FRED in July 2005, and the site alfred.stlouisfed.org debuted in July 2006.)

Together, FRASER and ALFRED offer key context for vintage economic and financial data. For instance, let’s say we want to look at the use of credit cards and similar lines of credit during and just after the financial crisis of 2007-2009. We can use consumer credit data from the Board of Governors’ G.19 statistical release to look at the monthly changes in the use of revolving credit. But is that what people thought was happening at the time?

To answer that question, let’s use the ALFRED graph above to compare today’s data with a vintage that’s a little closer to the events in question…

While the general trend is the same, the current, updated data (red bars) show bigger changes and at slightly different times between mid-2009 and 2011 than the vintage data (blue bars) published closer to the crisis. Looking at the original data publications on FRASER, including revision documentation and other historical technical notes, can give us some hints about the cause of these differences.

So this week, celebrate FRASER and ALFRED’s birthdays by checking out historical vintages of your favorite FRED series. You might be surprised at what you find.

Relevant ALFRED series are in the related links at the bottom of any FRED page and it has a help page. Many FRED series also link to their FRASER counterparts, but you can also browse FRASER by title or search it for the G.19 statistical release—and many other data publications.

How this graph was created: Search ALFRED for “REVOLSL.” By default, ALFRED shows a graph with two sets of bars: the most recent vintage (in this case, 2024-06-07) and the prior vintage. From the “Edit Graph” panel, use the Edit “Bar 1” option to select the vintage “2012-06-07” in the vintage selection dropdown menu or type the date in the “As-of date” box. Then change the units to “Percent Change.” Click on the “Copy to all” button to apply the unit change to all the series on the graph. Finally, adjust the start date of the graph to 2005-01-01 and the end date to 2011-12-31.

Suggested by Genevieve Podleski and Hannah Edwards.

How much did the US economy grow last year?

There are two commonly reported ways of computing the annual growth in real GDP.

First is the fourth quarter-over-fourth quarter (Q4/Q4) measure, which compares the amount of production of final goods and services in the December quarter with the amount produced in the December quarter of the previous year.

Second is the year-over-year (YoY) measure, which compares total production for the year with total production for the previous year. Both growth rates are plotted in the FRED graph above. The blue line is the Q4/Q4 growth rate, and the red line is the YoY growth rate.

While the two growth rates look similar over long periods of time, there can be significant differences at any point in time. Which measure should we use? Some examples may help.

Example 1: Higher-frequency data show that the global financial crisis began in 2008, which is shown in the Q4/Q4 annual growth measure. But the YoY measure records the recession as beginning in 2009. For questions about the timing of booms and recessions, the Q4/Q4 measure is probably better.

Example 2: The COVID recession began in the first quarter of 2020: The economy hit bottom in the second quarter, and it recovered substantially by the fourth quarter of the year. The Q4/Q4 measure therefore records a much milder recession than the YoY measure, which uses data for the entire year’s production. For questions about the level of production, the YoY measure is probably better.

How this graph was created: Search FRED for and select “Real GDP Seasonally Adjusted Annual Rate” (series ID: GDPC1). Click “Edit Graph” and, under the “Modify Frequency,” change Quarterly to Annual and change the aggregation method to “End of Period.” Change the units to “Percent Change from Year Ago.” Then click “Add Line” at the top and add the “Real GDP Seasonally Adjusted Annual Rate” (series ID: GDPC1) series again. Under “Edit Lines”, make sure you are editing Line 2. Change the “Modify Frequency” to be Annual and the Aggregation method to be “Sum.”

Suggested by Amy Smaldone and Mark Wright.



Subscribe to the FRED newsletter


Follow us

Back to Top