Using the ALFRED Database: How May It Help You?

When you see the name ALFRED, does the butler in “Batman” immediately spring to mind? Like the fictional character in the storied superhero franchise, the Federal Reserve’s ALFRED provides a service. But unlike Bruce Wayne’s longtime friend and mentor, this ALFRED is a database containing more than 815,000 data series.

ALFRED (for ArchivaL Federal Reserve Economic Data) launched in 2006 as a companion to its older sibling, FRED, the St. Louis Fed’s signature economic database, according to an April 2021 FRED Blog post. While 30-year-old FRED aggregates the latest data releases from sources such as the U.S. bureaus of Labor Statistics, Economic Analysis and Census, 15-year-old ALFRED captures vintage versions of these data as FRED replaces them with the latest numbers. Research Division analysts keep track of scheduled and unscheduled data releases, and ALFRED is typically updated within one business day, according to the ALFRED “Help” page.

Diego Mendez-Carbajo, a senior economic education specialist at the St. Louis Fed, says this information service is one way to take the pulse of the economy. It’s also a good way to study how economic issues evolve, adds Mendez-Carbajo, an economist by training who specializes in active learning with data.

“Decision-makers and policymakers make their decisions based on the data that are available at the time,” he says. “Had they known that the data would be revised, would they have made that decision? Data hindsight is always 20/20.”

St. Louis Fed Seeks to Fill a Data Need with ALFRED

Before ALFRED was created in 2006, the St. Louis Fed’s Research Division was looking for information on original data releases, but it was difficult to find something that offered day-to-day history, according the ALFRED “Help” page. Meanwhile, source agencies traditionally overwrote their older data once new information became available, so it was clear the Fed needed a tool that could show how estimates were refined over time.

After its creation, ALFRED began tracking all the data iterations in FRED, from business activity to employment to gross domestic product and beyond.

ALFRED users can see how data were updated or revised as new information on different historic dates became available. Having access to this data archive helps economists build better economic forecasting models and analyze economic policy decisions, Mendez-Carbajo says. They also can get an idea of how accurate the initial data releases were. If there are large discrepancies between data vintages, a data iteration may contain errors. 

Unemployment Claims and Home Prices: Examples of How ALFRED Works

The Department of Labor’s data series on initial claims for unemployment is an important economic indicator because it shows how many people are applying for jobless benefits and, by extension, how well the labor sector is doing in the larger economy. The first chart below is an example of how data estimates like this can vary over time, with the number of jobless claims being revised from 360,000 on July 15 (blue bar) to 368,000 on July 22 (red bar). The revision occurred after new information became available to the reporting agency. Soon after that, FRED started to display the July 22 data vintage and ALFRED added the July 15 data vintage to its repository.

Part 1: How to create an ALFRED graph of the latest data revisions to initial claims

1. Go to the FRED site at fred.stlouisfed.org.
2. In the Search field in the center under the FRED logo, type “initial claims.”
3. Choose “Initial Claims, Weekly, Seasonally Adjusted.”
4. Once the graph populates onscreen, scroll down to the “Related Resources” section of the page.
5. Click on “ALFRED Vintage Series Initial Claims” (on the right).

Part 2: How to build the ALFRED graph of July 2021 data revisions

1. On the “EDIT BARS” tab, select “EDIT BAR 1.”
2. Under “select a vintage,” choose July 15, 2021, from the drop-down list.
3. On the “EDIT BARS” tab, select “EDIT BAR 2.”
4. Under “select a vintage,” choose July 22, 2021, from the drop-down list.
5. The last step is to select a desired date range. This can be done by choosing the “1Y, 5Y or 10Y” (one-year, five-year and 10-year) tabs at the top of the graph or by entering a custom date range in the boxes to the right of these. For simplicity, the graph above shows a narrow, customized range, from July 10-17, 2021.

Another example of vintage figures can be seen with U.S. home prices, which serve as another key barometer of the economy’s health. The Case-Shiller National Home Price Index has risen steadily in the past year, aside from the winter months, when fewer homes are for sale. More people have remained home during the COVID-19 pandemic and housing demand has increased. The second chart below shows the price increase changing for April from 2.1% to 2.3%, but this is by no means the only revision to home prices in the past 12 months.

The fact is, if FRED contains a data series from a periodic data release, ALFRED can find out exactly what happened along the way. Curious about revisions to real gross domestic product or durable goods orders? They’re in there, just like Alfred in the Batcave. All you need to do is ask.