Tools › Investing › Economy and Rates › Economic Indicators › Method

Economic Indicators: Method

Open the Economic Indicators tool

Economic Indicators: How to Read It

What it does

The page shows the main statistics of the U.S. economy in one place: interest rates, inflation, the labour market, output and income, money and public debt. For each series it gives the latest value with its date, the change from the period before and from a year earlier, the low and high of the last ten years, and a chart. Overview gathers the headline figures of every group. The page is for reading. It does not forecast, and it does not say whether a number is good or bad.

Every figure is published by a U.S. government agency or by the Federal Reserve. The site keeps a copy and refreshes it on a schedule: the daily and weekly series (Treasury yields, public debt, the dollar index, jobless claims) each business day, the monthly and quarterly ones each week; the page never contacts an agency when you open it. A figure released in the last day or two may therefore not be here yet. The tables show the date of each series' latest observation and the date the copy was retrieved.

The series and where they come from

Interest rates. Treasury par yields, 1 month to 30 years, each business day from 1990 (U.S. Department of the Treasury, Daily Treasury Par Yield Curve Rates): the yearly yield a Treasury security of that maturity would pay if it sold at face value, derived by the Treasury from closing market prices. 10-year real yield, from 2003 (Treasury, Daily Treasury Par Real Yield Curve Rates): the same for Treasury Inflation-Protected Securities (TIPS), whose principal rises with the Consumer Price Index. Effective federal funds rate, monthly from 1954 (Federal Reserve Board, release H.15): the rate banks pay each other for overnight loans of reserves, averaged over the month.

Inflation. Consumer Price Index (CPI-U, all items; not seasonally adjusted from 1913, seasonally adjusted from 1947) and core CPI (all items less food and energy, from 1957), Bureau of Labor Statistics: the cost of a fixed list of what urban households buy, with 1982-84 set to 100. PCE price index and core PCE (excluding food and energy), monthly from 1959, Bureau of Economic Analysis, national accounts Table 2.8.4: the prices of everything households consume, with a mix that changes as spending changes.

Labour. Unemployment rate and labour force participation rate (from 1948) from the monthly household survey; nonfarm payroll employment (from 1939) and average hourly earnings of all private employees (from 2006) from the monthly survey of establishments; all Bureau of Labor Statistics, seasonally adjusted. Initial claims for unemployment insurance, weekly from 1967, Department of Labor.

Output and income. Real GDP, GDP in current dollars and real GDP growth at an annual rate, quarterly from 1947 (Bureau of Economic Analysis, Tables 1.1.6, 1.1.5 and 1.1.1). Real disposable personal income and the personal saving rate, monthly from 1959 (Table 2.6). Industrial production, monthly from 1919 (Federal Reserve Board, release G.17). Retail and food services sales (from 1992) and housing starts (from 1959), Census Bureau.

Money and public finance. M2, monthly from 1959 (Federal Reserve Board, release H.6). Total public debt outstanding, each business day from April 1993 (Treasury, Debt to the Penny). Federal surplus or deficit as a percent of GDP, by fiscal year from 1929 (Office of Management and Budget figures divided by GDP, as calculated by the Federal Reserve Bank of St. Louis). Nominal broad dollar index, daily from 2006 (Federal Reserve Board, release H.10).

The Treasury, Bureau of Labor Statistics, Bureau of Economic Analysis and Fiscal Service series are retrieved from those agencies' own data services. The Federal Reserve Board, Department of Labor (weekly claims), Census Bureau and budget series are retrieved through FRED, the data service of the Federal Reserve Bank of St. Louis; so are the Bureau of Labor Statistics series on a week when its own service does not answer. The notices under the results name the route actually used. Series owned by private companies (stock indexes, survey indexes, mortgage-rate surveys and the like) are deliberately not included.

The method

The page does very little arithmetic, and all of it is one of four operations on stored observations.

Change from the period before. The latest observation less the one before it: the previous business day, week, month, quarter or fiscal year. For a rate (a number already in percent) the change is given in percentage points, written pt: a rate that goes from 4.2 % to 4.1 % has changed by −0.10 pt. For an index or a money amount the change is given in percent, 100 × (new / old − 1). For a count (jobs, claims, housing starts) it is the plain difference.

Change from a year earlier. The same comparison against the same month, quarter or year one year before; for a daily or weekly series, against the last observation on or before the same date a year earlier. For a price index this is the inflation rate over the year:

Infl = (Indexnow / Indexago − 1) × 100

When the agency did not publish the month a year earlier, there is nothing to compare with and the table shows a dash; the headline figure then uses the latest month that has a year-earlier value, and says so.

Yield spread. One yield less another on the same day, in percentage points: Sp2y = y10 − y2 and Sp3m = y10 − y3m. An inverted curve means the short yield is above the long yield, so the spread is negative.

Breakeven inflation. The 10-year nominal yield less the 10-year real yield on the same day: BE = Ynom − Yreal, or equally Yreal = Ynom − BE. It is the average yearly rise of the CPI over ten years at which an ordinary Treasury note and an inflation-protected one would pay the same.

Low and high. The smallest and largest observation in the ten years ending at the series' latest observation.

Charts. A chart covers the period chosen. A long series (in practice a daily one) is drawn with at most 600 points: the first and last observation are kept and, from each of about 300 equal runs of the observations between, the one with the lowest and the one with the highest value, so every peak and trough survives; the caption says how many were drawn. The tables always use every observation. Grey bands mark recessions as dated by the National Bureau of Economic Research (the months after a business-cycle peak up to the trough). Show As (Advanced) redraws monthly, quarterly and yearly indexes, amounts and counts as the change from a year earlier or from the period before; rates and daily and weekly series keep their level.

What it assumes

That the stored copy is what the agency published on the retrieval date; the agency's own release is the authority. That a "year earlier" for a daily series may be the business day before the same calendar date. That quarterly GDP growth is quoted as BEA publishes it, compounded to an annual rate, so a quarter of 0.8 % growth appears as about 3.4 %. That seasonally adjusted series are used where the agency's headline figure is seasonally adjusted, and the unadjusted CPI for change over a year, as the Bureau of Labor Statistics does.

When it misleads

Revisions. Payrolls, GDP, retail sales, industrial production and the seasonally adjusted price indexes are revised, sometimes by a lot, months or years after first publication. The page shows today's version of history, not what was known at the time.

Different dates in one table. Yields are a day old, prices and jobs a month, GDP a quarter, the deficit a fiscal year. Read the Period column before comparing rows.

Month-to-month noise. One month's change in payrolls or claims is within the sampling error of the survey more often than headlines suggest. A change from a year earlier is steadier but slower to show a turn.

Nominal amounts. Retail sales, M2, debt, hourly earnings and GDP in current dollars rise with prices as well as with volume. Only the series marked real are adjusted for prices.

Breakeven inflation is a price, not a measurement. It contains premiums for inflation risk and for the lower liquidity of TIPS, which move for reasons unrelated to expected inflation.

The yield curve and recession bands. The bands record what happened; a spread on a given day is a fact about that day's prices. The page draws no conclusion from one about the other, and neither should a chart alone.

A missing month. The Bureau of Labor Statistics did not publish the October 2025 household-survey and CPI figures, so comparisons that need that month are blank.

Worked example

Overview with the figures stored on 13 November 2026.

Yield spreads on 12 November 2026: the 10-year par yield was 5.10 %, the 2-year 4.86 % and the 3-month 4.32 %. Sp2y = 5.10 − 4.86 = 0.24 pt and Sp3m = 5.10 − 4.32 = 0.78 pt. Both are positive: the long yield was above the short yields.

Breakeven inflation on the same day: the 10-year real yield was 2.86 %, so BE = 5.10 − 2.86 = 2.24 % a year.

Inflation over a year: the CPI was 336.940 in September 2026 and 324.800 in September 2025, so Infl = (336.940 / 324.800 − 1) × 100 = 3.738, about 3.7 %. October 2026 (337.511) cannot be compared because October 2025 was not published.

Payrolls: 159,293 thousand in October 2026 against 159,128 thousand in September, a change of +165 thousand.

These numbers change each time the stored copy is refreshed.

Notices

This product uses the FRED® API but is not endorsed or certified by the Federal Reserve Bank of St. Louis. (FRED® API Terms of Use)

This product uses the Bureau of Economic Analysis (BEA) Data API but is not endorsed or certified by BEA.

Recession periods: National Bureau of Economic Research, US Business Cycle Expansions and Contractions.

A calculator for education and analysis, not investment advice.