Federal Reserve Economic Data

Recessions, like most things, have beginnings and ends. In the U.S., the beginnings and ends of recessions are determined by the National Bureau of Economic Research (NBER) Business Cycle Dating Committee. They announce the months and quarters* when overall economic activity has reached a peak and starts to contract and when overall economic activity has reached a trough and starts to expand again.

The NBER determined the beginning and end of the COVID-19-induced recession as February 2020 and April 2020, respectively. The Federal Reserve Bank of St. Louis has created daily data series to help consumers of economic data better understand the story here and the possible choices for dating recessions.

The FRED graph above shows three daily series that date the start and the end of the COVID-19-induced recession. In the graph, the data points have a value of 1 when the economy is in recession and a value of 0 when the economy is in expansion. The three series report the same total number of days (60) during which overall economic activity was contracting, but the choices determining the beginning and end dates are different:

• The blue line marks the start of a recession on the first day of the month determined to be the peak month. It marks the end of a recession on the last day of the month before the trough month. This timing is reflected in the shaded areas shown in FRED graphs to represent recessions. For the most recent recession, these dates are February 1, 2020, to March 31, 2020.
• The yellow line marks the start of a recession on the 15th day of the peak month, and it marks the end of a recession on the 15th day of the trough month. For the most recent recession, February 15, 2020, to April 15, 2020.
• The red line marks the start of a recession on the last day of the peak month, and it marks the end of a recession on the last day of the trough month. For the most recent recession, February 29, 2020, to April 30, 2020.

Each series can help consumers of FRED data tell the story behind different sets of daily or weekly numbers. For example, this FRED graph shows that, in 2020, weekly initial claims for unemployment benefits grew ten times between the second and the third week of March. Interpreting the timing of that increase in relation to the start and end dates of the COVID-19-induced recession naturally depends on when those recession dates are set. So, FRED presents three choices to tell that story.

*For a discussion of how the 2022 recession lasted two months but spanned two quarters, see the FRED Blog post “Discrepancies in dating recessions.”

How this graph was created: Search for and select “NBER based Recession Indicators for the United States from the Peak through the Period preceding the Trough.” From the “Edit Graph” panel, use the “Add Line” tab to search for and select “NBER based Recession Indicators for the United States from the Peak through the Trough.” Repeat the last step to add “NBER based Recession Indicators for the United States from the Period following the Peak through the Trough” to the graph. To change the style and color of the lines in the graph use the “Format” panel.

Suggested by Diego Mendez-Carbajo.

The FRED graph above tracks the price of eggs. Specifically, a dozen large grade A chicken eggs, in U.S. dollars, for each month since January 2021.

This graph is a histogram, otherwise known as a bar graph, so the length of each bar represents the full price. Merely glancing at the heights of the bars shows us that the price fluctuates over time, with a low of $1.47 and a high of $2.52 as of April. This runny volatility is why food prices, along with energy prices, are often excluded from monetary policy analysis.

Cracking open this first concept is fairly simple, but the FRED Blog team hatched the idea of asking another question: What would the graph look like if we purchased that same carton of eggs with bitcoins instead of U.S. dollars? The graph below shows this. Because a bitcoin is worth so much more than a carton of eggs, we multiplied the price by 100 million to express it in so-called satoshis, which is the smallest subunit of bitcoin.

The price fluctuates quite a bit, between 2829 and 6086, which is much more than it did for the U.S. dollar price. Plus, you’d need to add a bitcoin transaction fee, which has been about $2 lately, but which can spike above $50 on occasion. Hopefully, if you were making this purchase with bitcoin, you’d put many many more eggs in your basket.

How these graphs were created: First graph: Search FRED for “price of eggs”; the series we use should appear at the top of the results. Restrict the sample period to start in January 2021. From the “Edit Graph” panel, use the “Format” tab to choose “Bar” as the graph type. Second graph: Take the first and add a series to the first line by searching for and selecting “bitcoin price” and applying formula a/b*100000000.

Suggested by Christian Zimmermann.

No doubt about it, mortgage rates are up. The FRED graph above shows the rates for the most popular fixed-rate mortgages: the 15-year and 30-year. Every data point is the average rate offered at that point in time for new mortgages. Although the graph shows the recent data, at this point very few people are actually paying these increased rates.

One could say that current mortgage holders are enjoying a good deal: They’re paying a lower rate, and inflation is higher. Inflation matters because mortgage debt is nominal. So, if inflation increases all prices (and in particular wages), paying a nominal debt such as a mortgage becomes much easier. One might then consider that even new mortgages are also a good deal when there’s inflation.

The previous low-rate mortgages were not set during a time of higher inflation, and those who set the rates must not have anticipated the higher inflation to come. But the new mortgage rates now include the anticipation of higher inflation, and thus this inflation advantage is factored into the mortgage rate.

To look at the data behind this argument, we use the FRED graph below. Here, we deflate each mortgage rate by the corresponding “breakeven” rate, which takes into account the anticipated average inflation from the point of measure over the relevant number of years. (Unfortunately, there’s no 15-year breakeven rate, so we average the 10- and 20-year rates using FRED’s fancy tools.)

The result is actually not that different. The real mortgage rates are still significantly up. The reason is that inflation expectations over such long horizons (15 to 30 years) have not moved that much, likely reflecting a general expectation that inflation won’t last. Our last FRED graph documents those expectations.

How these graphs were created: First graph: Search FRED and select the 15-year mortgage rate. Once you have the graph, use the “Edit Graph” panel’s “Add Line” tab to search for and add the 30-year mortgage rate. Second graph: Start with the first, use the “Edit Line” tab for the 30-year mortgage series: Search for and add the “Breakeven 30-year” series and apply formula a-b. In a similar way, add two series to the 15-year mortgage line: “Breakeven 10-year” and “Breakeven 20-year,” and apply formula a-(b+c)/2. Third graph: Use the interest rate spreads release to select the relevant breakeven rates and click “Add to graph.”

Suggested by Christian Zimmermann.