Skip to main content

The FRED® Blog

A clearer picture of housing equity before the crisis

This graph shows housing equity in the United States. The way it’s shown here, housing equity appears to have undergone an extremely unhealthy evolution: rapidly accelerating run-up, sudden and brutal crash, and another rapid run-up. There’s no doubt the housing crash has been significant; after all, housing equity was cut by half. But the alarming run-up shown in this graph is to some degree an optical illusion. Indeed, an increase in the 1950s isn’t equivalent to a same-sized increase in the 2000s because the level of the series was dramatically different. For a clearer picture, we’ll use the natural logarithm of the series.

Now, the run-up around 2000 looks like a normal part of a trend that’s continued for more than half a century. The illusion shown in the top graph can occur whenever a series grows over time. Think of the principal on a savings account that accumulates interest. Soon enough, the effect of compounding interest kicks in and the principal appears to explode, even though it’s still growing at the same interest rate.

How these graphs were created: For the first graph, search for the series name. For the second, expand the “Create your own data transformation” option in the graph tab and choose the “Natural Log” transformation.

Suggested by Christian Zimmermann

View on FRED, series used in this post: OEHRENWBSHNO

This recession was different

Most recessions share common characteristics, but not the most recent one. To illustrate this, we use a little known and used feature of FRED: setting a common index value and examining a period before and after that point. In the graph, you see four versions of the same series, civilian unemployment. Each series is centered on a different recession peak date, with a value of 100 for these start dates. The graph also shows data for 60 months before and 80 months after those dates.

The period before the start dates reveals nothing remarkable, but the most recent recession deviates from the other recessions after the start date: The unemployment rate shoots up much higher, and despite a steeper downslope the unemployment rate has yet to reach a value that would be expected from a normal recovery. (By the time 80 months had elapsed from the other recessions’ start dates, the unemployment rates had essentially returned to where they started.)

How this graph was created: Find the “Civilian Unemployment Rate” and modify the units to “Index (Scale value to 100 for chosen period).” For this graph, we use “U.S. Recession Peak” (vs. the “Trough” or another “Observation Date”). The default will be the peak of the most recent recession. Then choose the “Display integer periods instead of dates” option. Choose an interior period range of -60 to 80. Add this unemployment rate series three more times, performing the same manipulations but selecting different recession peaks.

Suggested by Christian Zimmermann

View on FRED, series used in this post: UNRATE

The seasonality of e-commerce

We know that retail sales move with the seasons, but what about e-commerce retail sales? FRED has the data, so we can take a look. The graph shows this series is also very predictable. The general trend is a straight line, with a pattern of increasingly large spikes in the fourth quarter. But e-commerce seasonal patterns are particularly striking: Retail sales in general always rise in the fourth quarter, but e-commerce sales do so even more intensely.

How this graph was created: Search for “E-Commerce” and select the “Sales Share” series without seasonal adjustment.

Suggested by Christian Zimmermann

View on FRED, series used in this post: ECOMPCTNSA


Subscribe to the FRED newsletter


Follow us

Twitter logo Google Plus logo Facebook logo YouTube logo LinkedIn logo
Back to Top