The FRED® Blog

Seasonal interest rates

When we say “seasonal variation,” we’re referring to fluctuations in the data that follow a pattern according to the time of year. For example, retail trade is always higher just before Christmas. The sale of ski lift tickets is always higher during winter—at least in the Northern Hemisphere. Agricultural output is higher in the growing season. Could this variation also apply to interest rates? It turns out it can, under specific circumstances. In some markets, banks look for liquidity at various times. They typically face regulations that affect what they can carry in their books; depending on the country and other factors, they may have to satisfy these regulations every single day, at the end of the month, or on average over the month. The end-of-month option especially can introduce seasonality in overnight interest rates, as banks scramble to satisfy regulations at very specific times during the year. The graph above shows liquidity in euros, which spikes every last day of the month. The graph below, which covers banks in Denmark, shows a spike every Wednesday. (Special circumstances also apply, such as holidays: Note the spike on Monday, Christmas Eve 2007, for example.) In the lower graph, you can slide the sample window to the right to see that this spike does not always occur: Rules can change over time, as can general market conditions. In fact, at the very end of the sample (Sep. 2012 through March 2013), the spikes actually point down, as banks were trying to get rid of their excess liquidity.

How these graphs were created: Search for “daily overnight” and you’ll find various choices. The two presented above are those with seasonal variations.

Suggested by Christian Zimmermann

Tagged ,

Tracking the duration of unemployment

The latest recession was different from other postwar recessions. One striking feature is how the various durations of unemployment have changed. The fraction of long-term unemployed (>26 weeks) had never been the largest. But now it is the largest by far! Until now, the fraction of short-term unemployed (<5 weeks) has always been the largest. Now it’s second or even third. What’s so peculiar about this recession? Is this a new regime? To truly answer these questions, we most likely have to wait for new data to come in. FRED offers various tools to stay connected. 1. You can create a dashboard that allows you to track statistics. 2. You can place a widget on your web page that reveals the latest data for up to six series. 3. You can subscribe to email alerts for the latest updates of you favorite series. 4. You can put the relevant series in an Excel spreadsheet and refresh the data with a single click (thanks to our Excel add-in). 5. You can come back to this blog post from time to time, and its graph will automatically update with the latest data.

How this graph was created: Find the release table for unemployed persons by duration of unemployment, select the four relevant series at the bottom, and add them to the graph.

Suggested by George Essig

Tagged , , ,

Parallel prices for oil-based fuels

The recent wild fluctuations in oil prices have been reflected in the end-user prices for various forms of fuel. This graph shows average prices at the pump for regular gas, diesel, and heating oil. What is remarkable is that they run nearly parallel to each other, except for a slow drift upward for diesel. These fuels have different patterns of seasonal demand (e.g., high demand for heating oil in winter and gas in summer), so their prices might reflect these variations. Yet, any seasonal price variations appear to be dwarfed by the price variations of the raw material in all three of these fuels: oil. Seasonal changes in demand are smoothed through storage of inventories and through price adjustments. Apparently, though, seasonal adjustments do not affect the prices of these fuels nearly as much as the price of oil does.

How this graph was created: Simply search for “Heating oil price,” then add the two other series. (Btw, the frequency of the series in this graph is monthly.)

Suggested by Christian Zimmermann

Tagged , ,

How did the U.S. economy perform under the pre-Fed gold standard?

The Federal Reserve System, established in 1914, recently marked its 100th anniversary. Every so often, someone expresses a longing for the “good old days” when the United States had no central bank and the dollar was anchored to the gold standard. Under a gold standard, the government promises to exchange its currency for gold at a fixed price; proponents argue that this prevents the government from printing money to finance expenditures, which could be inflationary. So, mainly the supply of gold, and not central bank policy or government spending, determines the value of the nation’s money.

The United States suspended the gold standard during the Civil War, but the Resumption Act of 1875 led to our return to the gold standard in 1879. The gold standard lasted through the creation of the Fed, loosened somewhat during the Great Depression, and eventually was abandoned in 1971.

How did the U.S. economy perform under the gold standard before the Fed was established? I turned to data in the NBER Macrohistory Database to find out.

The first graph plots the U.S. monetary gold stock and an index of the general level of prices from June 1878 to Dec. 1914. (The shaded areas indicate U.S. recessions.) Changes in the size of the gold stock were caused by additions to world gold supply (mainly from mining) and to gold flows associated with payments for international trade and investment. Over much of this period, especially after major gold discoveries increased world supplies in the 1890s, the monetary gold stock rose and the U.S. price level rose in response. However, over shorter periods of one to two years, the relationship between the size of the gold stock and the price level was not so close. For example, the price level rose sharply in the early 1880s and then abruptly declined, despite a smoothly rising gold stock.

The second graph shows the volatility of the price level during this period by plotting the year-over-year percent change in the price level (i.e., the inflation rate) alongside the gold stock. We can see the inflation rate fluctuated widely in this period, from about -10% in some years to about +10% in others. Over most of the period, the average inflation rate was low—close to zero—but the annual fluctuations in the rate were large, much larger than they have been in the United States in recent years. This price index for the 19th and early 20th centuries was constructed using different methods and with less information and precision than current price indexes use, which must be kept in mind when making comparisons between the pre-Fed era and modern times. However, beyond the volatility in the rate of inflation, the pre-Fed era was marked by several recessions, as well as serious banking panics and other financial crises. Thus, the historical evidence indicates that neither a gold standard nor the absence of a central bank guarantees economic or financial stability.

How these graphs were created: The “Academic Data” section of FRED contains data series constructed or contributed by academics and other non-official sources. The NBER Macrohistory Database includes historical series collected by researchers at the NBER and others. For these graphs, find the series “Monetary Gold Stock for United States” in the NBER Macrohistory Database under the subcategory “Money and Banking” or by using the tag “gold.” Select the range 1878-1914. Add the “Index of the General Price Level for United States” series to the graph (from the same database) and assign this series to the right y-axis in the “edit data series” window. For the second graph, simply change the units of the price level series to “Percent change from year ago.”

Suggested by David Wheelock.

Tagged ,

Churning in the labor market

The U.S. labor market changes quite a bit, with hirings, firings, gains, and losses. The graph above represents this dynamic situation: In red (in negative territory) are all the job separations and in blue are all the new hires. The end result is the net creation of jobs. The series used here are not seasonally adjusted, so one can readily see strong patterns—both throughout the individual years and during recessions and booms. It is also remarkable how small the net effect is with respect to the two series. The labor market always moves, and the net gains happen at the margin. But it could be different, of course: Separations could occur mostly or even only during recessions and hires could occur only during booms. But the reason that doesn’t happen is that not every region or every sector of the economy follows the same pattern as the overall economy. Also, even during recessions, people frequently change jobs and businesses need new workers, just as businesses can close even during booms. There is a lot of churning out there.

How this graph was created: Look for “Hires: Total Non Farm” (level in thousands, not seasonally adjusted) and graph that series. Then add the series “Total Separations: Total Nonfarm” (also level in thousands, not seasonally adjusted). Transform the latter series by applying the formula -a. Then choose graph type “Area” with “Normal” stacking.

Suggested by Christian Zimmermann

Tagged ,

Inflation across space

The national consumer price index averages the prices of many goods across the entire nation. But just because this index gets all the headlines doesn’t mean that all prices evolve in the same way. Certainly, there can be short-term differences in price. But it’s generally underappreciated that there can be longer-term differences as well. The graph above illustrates this by comparing a few MSAs in the United States: The differences in their price evolutions show that inflation at the local level is not entirely driven by a common currency. The same applies to countries under a monetary union, such as those from the European Monetary Union in the graph below. Put more simply, using the same currency does not mean the inflation rate will be the same everywhere. Why? First, the basket of goods used to compute the local price index may differ. Second, the relative prices of the local goods may vary, foremost housing and transportation. And third, the local business cycles are not synchronized, meaning that inflationary pressures may vary from place to place.

How these graphs were created: Play with the tags until you find your preferred set of series. For the first graph, it is useful to set the geography type to “msa” and choose “annual” for frequency. Then you can easily narrow down the set of series. For the second graph, “nation” and “eurostat” are good starting points. Once you have a list of series you’re happy with, select them and click on the “Add to graph” button.

Suggested by Christian Zimmermann

Tagged , , , , , , ,

How labor market flows changed

After the most recent recession, the volume of workers switching their employment status from unemployment to non-participation (that is, not in the labor force) and vice versa increased dramatically, reaching levels not seen in the previous two recessions.

The first graph shows the flows from non-participation into unemployment (NU) and from employment into unemployment (EU), both normalized by population. In the past, these two flows closely tracked each other and the contribution of both non-participation and employment to unemployment was roughly equal. However, in the Great Recession, the contribution to unemployment from both series increased significantly, with the contribution of non-participation becoming substantially larger, peaking far above the EU flow and taking a much longer time to return to lower levels.

The second graph shows the flows of workers that have switched from employment and non-participation into unemployment as a fraction of total population. We can see that the flows from employment to non-participation (EN) have decreased in the past recession, a behavior in line with previous episodes. The flows from unemployment into non-participation (UN) show a much stronger response, increasing to historically high levels after 2008.

The fact that both UN and NU flows are larger than usual, but of a roughly similar magnitude, implies that the labor market has become more dynamic on that margin. The levels of labor market variables (employment, unemployment, and non-participation) are very sensitive to the dynamics of labor market flows. Therefore, understanding the causes for the recent evolution of the UN and NU flows is central to understanding the dynamics of labor market variables in the past recession.

How these graph were created: Go to the labor force status flows category. For the first graph, find the flow called “Labor Force Flows Employed to Unemployed” and graph the seasonally adjusted values. Select the “Graph” tab and scroll down to the “Add Data Series” option. In the keywords box, search for “Labor flows,” scroll down to “Labor Force Flows Not in Labor Force to Unemployed,” and add the series as a new data series. Select the “Add Data Series” option again and search in the keyword box for “Civilian population.” Select the first series and add the data under the “Modify existing series option” for data series 1 and 2. Now, select “Edit Data Series.” Under the “Create your own data transformation” option, type the formula a/b and click “Apply.” Do this for each data series. Finally, restrict the time period to 1990 through 2014. To make the second graph, repeat this process but this time select the flows “Employed to Not in the Labor Force” and “Unemployed to Not in the Labor Force.”

Suggested by Maxiliano Dvorkin and Hannah Shell

Tagged , , , ,

“Log in” to regional data

FRED has compiled regional U.S. data for many economic indicators. The vast amount of regional data can make searching through the categories a bit overwhelming. But there are simpler ways to find what you want: You can search the releases if you have a good idea what you’re looking for. You can also use tags to quickly narrow down your search results—for example, by selecting specific geographies and geography types. Here, we look at per capita income for a sample of metropolitan statistical areas (MSAs) across the U.S.

In the graph above, we took the natural logarithm for each series, for the following reason: If the economic aggregate you’re looking at is growing at a constant rate and you have a sufficiently long sample, the data series will look convex and may give the impression that growth has been explosive. But if you take the natural logarithm, then a constant growth rate will look like a straight line. The graph above uses the natural logarithm: All four metropolitan areas have a kink around 1980 with a subsequent slowdown. The graph below doesn’t use the natural logarithm: That kink is not visible. Also, below it looks like San Francisco is taking off and separating from the others. Above, the distance between the lines can be interpreted as percentage difference, and it is clear that in relative terms San Francisco stays within range.

How this graph was created: Go to the list of MSAs for which FRED has per capita income, select the series you want, and click the “Add to graph” button. That’s the bottom graph. For the top graph, go to the graph tab and, for each series, expand “Create your own transformation” and select “Natural Log.”

Suggested by Christian Zimmermann

Tagged , , ,

Components of M2

What is money? Well, there are many statistical definitions and FRED’s new release tables can help us sort them out. The release table for monetary aggregates shows us the various components of M2, the broadest monetary aggregate currently measured. The major components are represented in the graph above, as shares of total M2. The strictest measure of money is currency, in red at the bottom. Add to that checkable deposit accounts in banks and elsewhere, and you have M1. Add savings accounts, small time-deposit accounts, and money funds to M1, and you get M2.

The graph shows how the composition of M2 has changed over time. Deposit accounts in banks used to be much more important. This has changed as the financial industry and its customers have become more sophisticated. Also, regulatory changes as well as amendments to accounting rules have had a direct impact on measurements or have nudged market participants to hold liquidity or savings differently. But currency has been largely unaffected by such changes.

How this graph was created: Go to the Money Stock Measures release, choose a table from the top, click the series you want graphed, and click on “add to graph.” Then, open the tab for the currency component and move it to the bottom of the pile by clicking on the “down” button. Finally, under graph settings, set graph type to “bar” with stacking set to “percent.” You will notice that the early data series have only currency; thus, start the sample in January 1959.

Suggested by Christian Zimmermann

Tagged , , ,

Quits by industry

FRED recently introduced “release views,” which make it much easier to split an economic aggregate into various components or categories. Here, we use the Job Openings and Labor Turnover release to examine quits and hires by industry. In the graph above, it is striking how the ranking of industry quit rates remains the same no matter how well the economy is doing. Also, the quit rates of some sectors respond more strongly as the economy improves. Naturally, one is more likely to quit a job when it’s easier to find another. This is confirmed by looking at the industry hiring rates in the graph below, where the ranking and trend of the lines are the same as above. See the spike for government hiring around 2010? That corresponds to temporary workers hired for the decennial census.

How these graphs were created: For each graph, go to the Job Openings and Labor Turnover release, find the right release table from the top list, check the industry series you want, and click on the “add to graph” button.

Suggested by Christian Zimmermann.

Tagged , , , , , , , , ,

Recently Viewed Series


Subscribe to our newsletter for updates on published research, data news, and latest econ information.
Name:   Email:  
Twitter logo Google Plus logo Facebook logo YouTube logo LinkedIn logo