Skip to main content

The FRED® Blog

The rise of the service economy

One constant throughout economic history is that, as an economy develops, its service sector keeps growing. The graph shows that this is certainly true for the United States. It divides nonfarm payrolls into three categories: government (at all levels), goods-producing industries (mining, manufacturing, construction…), and service-providing industries. Although government is roughly constant, services have far surpassed goods.

Is this bad? Of course not. The standard of living has clearly improved since 1939, when the graph starts. Indeed, goods can now be produced with fewer people—thanks to technological progress and automation…and perhaps also automatization. This transformation allows the economy to direct more of the labor force to enhancing our lives in other ways, such as tourism and entertainment, advanced health care, and anything related to the Internet, all of which are services that were either nonexistent or luxuries in 1939.

How this graph was created: Using the nonfarm payrolls by industry sector release table from the establishment survey, check the series and click “Add to Graph.” From the “Edit Graph” panel, open the “Format” tab and select graph type “Area” and “Stacked.”

Suggested by Christian Zimmermann.

View on FRED, series used in this post: CES0800000001, USGOOD, USGOVT

What’s up (or down) with the yield curve? Analyzing the new most-popular series in FRED

For as long as we can remember, the most popular series in FRED has been the consumer price index (CPI). Well, not anymore. Recently, the series describing the difference between the 10-year and 2-year Treasury constant maturity rates became the most popular. Why this sudden interest? It has to do with the concept of the yield curve: Under normal circumstances, long-term interest rates are higher than short-term interest rates (when annualized), principally because the long term is usually perceived as riskier and so long-term debt demands a higher return. Again, normally, if you plot the interest rates at different maturities, you get an upward-sloping (yield) curve. But if for some reason the short term becomes unusually risky, the curve (or portions of it) may become downward sloping. And why is that important? The graph makes it clear that this kind of yield curve inversion has been associated with impending recessions. (See the gray vertical bars.) As the yield curve gets close to such a situation, there’s going to be a lot of interest in it.

How this graph was created: From the FRED homepage, open the tab “Popular Series,” click on the first one (at the time of this writing, anyway), and expand the sample to the maximum.

Suggested by Christian Zimmermann.

View on FRED, series used in this post: T10Y2Y

If they drive, they will park (Or if they park, they will drive?) Correlation does not always equal causation

This graph shows that the more people drive, the more they park and generate revenue for parking lot and garage operators. While there’s clearly a correlation between these two indicators, it isn’t clear that there’s a straightforward causality between them. In fact, a third indicator may be affecting the other two: the number of cars in use, the size of the road network, economic activity in general, commuting distance… Or maybe it’s a combination of all or some of these. This ambiguity is what makes statistical analysis much more complex than simply looking at correlations in a graph. FRED helps you stay rigorous by allowing you to download data into your favorite statistical software, either with a download from FRED itself (for example, via the “Download Data” link below the graph) or natively from the software of your choice. For starters, you can use this published data list.

How this graph was created: Search for and select “parking lot revenue” and click on “Add to Graph.” From the “Edit Graph” menu, search for “GDP deflator” in the “Customize data” section and add the series, applying formula a/b. Then from the “Add Line” tab, search for and add “vehicle miles.” Finally, from the “Format” tab, place the y-axis of the second line on the right side.

How this data list was created. For starters, you need to (create and) log on to a FRED account. Then, from any account page, click on “Add new” and select “Data list.” Give it a name. Then search for the series, check the series you want, and click on “Add to data list.” Repeat until satisfied. You can make the data list public and will be required to give it a public name.

Suggested by Christian Zimmermann.

View on FRED, series used in this post: GDPDEF, REVEF81293TAXABL, TRFVOLUSM227NFWA

Hyperinflation in Venezuela Using exchange rates to measure out-of-hand inflation

There’s inflation and then there’s hyperinflation, which is when inflation gets out of hand. There’s no official definition, but economists tend to use the “if it looks like hyperinflation, then it is” dictum. When price changes occur rapidly—say, several times within the same day—that’s hyperinflation. When bank notes don’t have denominations large enough to make payments easy, that’s hyperinflation. And measuring hyperinflation isn’t easy, as no statistical office can keep up with the rapid changes in price. One way to track this phenomenon, though, is to look at exchange rates.

The first graph shows the exchange rate between the Venezuelan bolívar fuerte and the U.S. dollar. Quite obviously, something out of the ordinary happened. The bolívar lost value rapidly—so much so that the graph allows us to see only a few recent data points.

[ Update 8/20/2018: The Venezuelan government announced economic policies designed to stem hyperinflation. The immediate impact is to raise the exchange rate to 6 million bolívars to the U.S. dollar (up from about 200,000). Time will tell whether these policies are successful. ]

One remedy for visualizing the wide range of values is to use a log scale, as the second graph does. A graph with a log scale will show with a straight line any data that increase at a constant rate. If the data increase at an increasing rate, the line moves steeper (i.e., it becomes convex). A few things are remarkable. First, there have been long periods of constant exchange rates, owing to the government’s policy of setting those rates. (See the several straight lines.) But recently, the rise in the exchange rate has been accelerating. (See the several steps up and eventually the vertical line.) This behavior in the data is characteristic of hyperinflation, which is obviously not sustainable.

How these graphs were created: For the first graph, search for and select “Venezuela exchange rate” (the monthly series) and click on “Add to Graph.” For the second graph, adjust the first graph in the “Edit Graph” section: Use the “Format” tab to select “Log scale” on the left.

Suggested by Christian Zimmermann.

View on FRED, series used in this post: EXVZUS

Dips in durables and nondurables Tracking consumption dynamics since the Great Recession

Household consumption contributes about 70% of total U.S. production. So, movements in consumption over time and the make-up of products purchased are essential for understanding GDP.

In general, final consumption is procyclical. That is, in good times people spend more and in bad times they spend less. During the recent 2007-09 Great Recession, real personal consumption expenditures dropped by 2.4% from pre-recession levels. But did people cut back the same amount on spending for all goods? The first graph splits aggregate consumption into durable and nondurable goods: Durable goods include, for example, cars and furniture and nondurable goods include, for example, food and clothing. We can see that consumption of durable goods is more volatile. During the recession, durable goods expenditures dropped by 12%, while nodurable goods expenditures dropped by 3%. (This doesn’t add up to the total 2.4% drop because there’s also the consumption of services, which isn’t shown here.)

The second graph separates durable goods expenditures into finer categories.

Among these durable goods categories, expenditures on motor vehicles and parts dropped the most—by as much as 24%. Expenditures on furnishings and durable household equipment follow, with a drop of 16%, although this category hits bottom at the end of the recession. Both categories recovered to their pre-recession levels only after 2013. In contrast, expenditures on recreational goods and vehicles and on other durable goods decreased only mildly.

Why did vehicle sales drop so much? One reason could be that a car purchase requires a larger payment than other durable goods. Another explanation is that the decline in house prices deteriorated households’ credit scores, making it more costly to finance a car purchase with a loan. The sinking housing market also would have driven down demand for furnishings and durable household equipment. And, simply, credit may have generally been more difficult to obtain.

Expenditures for nondurable goods fell by 4% on average, less than they did for durable goods. The third graph shows four categories. Interestingly, expenditures on gasoline and other energy goods slowly and consistently fell, hitting bottom in late 2012 and gradually recovering by 2016, which is likely due to the falling car sales shown in the previous graph.

Overall, expenditures related to vehicles and houses account for most of the drop in consumption in the 2007-09 recession.

How these graphs were created: All these series can easily be found in FRED using the Release View. From the FRED home page, select Releases from the main search box and click the link for gross domestic product. You’ll then see the sections of the GDP release. Select “Section 2–Personal Income and Outlays” and then “Table 2.3.3. Real Personal Consumption Expenditures by Major Type of Product, Quantity Indexes” (quarterly frequency). The resulting page will be a table with all the series associated with the real personal consumption expenditures release and all graphs can be made from this page.

First graph: Select “Durable goods” and “Nondurable goods” and click “Add to Graph.” Using the scroll bar at the bottom or the date range at the top, adjust the time horizon of the graph to start in 2006:Q1. From the “Edit Graph” menu, adjust the units to be “Index (scale value to 100 for chosen date)” and set the date to be 2007-01-01. Click “Copy to all” to copy these changes to all the series.

Second graph: Select the four subcomponents under the durable goods category and click “Add to Graph.” Follow the same steps as in the first graph to adjust the time and index base value.

Third graph: Select the four subcomponents under the nondurable goods category and click “Add to Graph.” Follow the same steps as in the first graph to adjust the time and index base value.

Suggested by Sungki Hong.


Subscribe to the FRED newsletter

Follow us

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