What is FRED? Short for Federal Reserve Economic Data, FRED is an online database consisting of hundred of thousands of economic data time series from scores of national, international, public, and private sources. FRED, created and maintained by the Research Department at the Federal Reserve Bank of St. Louis, goes far beyond simply providing data: It combines data with a powerful mix of tools that help the user understand, interact with, display, and disseminate the data. In essence, FRED helps users tell their data stories. The purpose of this article is to guide the potential (or current) FRED user through the various aspects and tools of the database.
A Brief History…
FRED began in the early 1990s as an offshoot of the long-running legacy at the Federal Reserve Bank of St. Louis of providing monetary data to help better understand the Fed’s policy decisions. Before the popularization of the World Wide Web, the data were provided in list form on a dial-in, electronic bulletin board system. The data were organized into categories containing roughly 300 data series and expanded from there. Perhaps surprisingly, FRED did not begin as part of a grand scheme or strategic objective. Rather, it grew over time in a very organic way. St. Louis Fed staff who were involved either directly with the FRED project or working on the periphery developed tools for the database in an independent, ad hoc manner. One example was in 1995 when a member of the Research Department staff decided to write the code that would officially take FRED “online.” Further work would bring the following FRED developments:
ALFRED (ArchivaL Federal Reserve Economic Data) goes live, offering users the ability to access vintage data for many of the available FRED series.
FRED Graph is introduced alongside the ability to download data in an Excel format.
Unit transformations allow users to change data from levels to percent change, among others.
GeoFRED provides mapping tools for FRED data.
Published Data Lists go live, allowing users to post created groupings of data.
The FRED API is released, allowing software developers access to the database functionality.
Frequency aggregation is developed, allowing users the ability to aggregate data from a higher frequency to a lower frequency (e.g., from monthly data to annual data).
The Android app becomes available in Google Play.
These tools, in combination with the raw data, have culminated in a widely popular database that is accessed by over two million persons per year from almost every country in the world.
The Research Department has a firm commitment to the growth of the database. Since its inception, FRED has contained many of the more popular figures reported by the Board of Governors, Bureau of Economic Analysis, Bureau of Labor Statistics, and Census—among others. Through time, FRED has expanded its collection to include many more international, national, and regional data series. More recently, it has become clear that data relevant to other topics and geographies must also be included if FRED is to best serve its users so the data content will continue to grow and evolve. Naturally, care will be taken to add data in a thorough and prudent manner.
Certain data, as it travels through time, is subject to revision. Anyone who follows GDP long enough will be familiar with the BEA revising (up or down) their quarterly released figure. The FRED database always contains and displays the most recent revision—or vintage—of the data available. But FRED’s real-time relative, the aforementioned ALFRED (ArchivaL Federal Reserve Economic Database), captures all of these individual revisions to a data series. This means that collectively, FRED and ALFRED data can literally be used as a data time machine, allowing users access to the precise data that their predecessors used. Researchers often attempt to replicate results of previous academic papers or use data to “train” economic models; in these instances, the relevance of these FRED tools becomes clear.
The data are accessible from a variety of different hardware and software, with the primary point of access being the FRED website. From the homepage, users can choose to search for the data by typing in their search term or alternatively can browse the data through other organized points of access. Data are browsable by (1) source (the institution or company that produces the data), (2) release (the document associated with the data’s publication), (3) category (a list of data topics as organized by the FRED staff), (4) latest update (the most recently updated data), and (5) tags.
While the categories structure has traditionally been the most popular, the recent expansion of the database has made this method of organizing data increasingly difficult to use, as the number of series in any one category can grow quite large. Tags are seen as the next, more-efficient evolution of this structure. By assigning metadata concepts to the series, as opposed to assigning the series to the concepts, FRED allows for a more flexible cataloging structure that will both accommodate further expansion of the database and help users more quickly and intuitively search these data.
Access doesn’t stop at the website, however. Starting with the FRED application programming interface (API) and building from there, the FRED team and private contributors have developed a suite of applications and programmatic wrappers that make accessing the data much simpler for those with specific computing preferences. For example, staff used this tool to develop the FRED add-in for Microsoft Excel. The add-in gives Microsoft Excel users access to all of the data and tools available in FRED. Users can search for, download, and update data; create charts; and perform frequency aggregation and units transformations, all without leaving their spreadsheet. For users on the go, the FRED apps for the iOS and Android platforms allow access to almost all of the FRED tools in the palm of your hand. For those who use devices outside of these platforms, the FRED mobile website provides an excellent alternative.
The FRED team also works on creating tools for software design. To date, these efforts have allowed third-party developers to create accessibility from R, STATA, MatLAB, RATS, and EViews. FRED is also used in Pearson Education’s MyEconLab, a facility designed to help students understand the intricacies of economic data and theory. Above and beyond this, there are several programmatic wrappers and toolkits available for the FRED API including Java, .NET, PHP, Python, and Ruby. Staff are always exploring the development of further products.
Having the data that users need and putting it in their hands is certainly helpful; but, to truly help persons tell their data story, you need to provide the necessary tools to work with the data. FRED provides several simple but powerful tools that aid the user in viewing and presenting the data in a manner that is both accurate and understandable. By far, the most popular way to view a time series in FRED is in a line chart. FRED allows the user to not only display their chosen data in this format, but also completely customize the aesthetic of this chart. Users can adjust the fonts, colors, and line weights, among other attributes of the chart. Line charts are helpful for viewing trends over time, but not always the best at comparing recent observations. For this purpose, FRED can also chart data in pie, bar, and scatter plot form. Beyond this, it can also be helpful to view data in its geographic context. GeoFRED allows user to view data at the state, MSA, and county level.
It’s often the case that we need to use more than one series to tell a particular data story. One could imagine a yield spread, for example. In FRED, this calculation is accomplished easily by first pulling the two series into the FRED chart and subtracting one from the other. Going further, we could imagine charting both the 10-year Treasury rate and an AAA corporate bond rate and subtracting the Treasury rate from the bond rate to create our spread. While this is a simple example, users can interact many series on a single chart. When thinking about creating series such as the Taylor rule, this can be quite a handy tool.
Data are often reported in units that might not be the most conducive to analysis. In FRED, units are easily modified. If users view the FRED series for “real gross domestic product in billions of chained 2009 dollars,” they can quickly change this series to “percent change from a year ago” with one simple selection. Change, percent change, and compounded annual rate of change are among the many selections available.
As with units, frequency can also be an issue. It is often the case that, for purposes of estimation and graphing, the frequencies of all data series must be equal. FRED’s built-in frequency aggregation tool makes this a simple matter. Users can quickly and easily aggregate a data series from a higher frequency to a lower one.
At its worst, data can be intimidating to deal with and difficult to understand. To help alleviate these potential obstacles, many series in FRED contain notes that help the user understand or interpret the data in question. These notes are sometimes an off-site link to more information. Other times they can be full explanations about how a series is constructed or should be interpreted. FRED staff will continue to develop these notes as time goes on.
While we are constantly striving to make FRED as user-friendly as possible, we know that everyone needs help from time to time. Whether a user requires assistance using one of our tools or is simply trying to understand an aspect of a particular data series, we are always available to assist.