Money, Prices & Policy

Definitions

Year-over-year growth

100 × (value this month ÷ value exactly twelve months earlier − 1). If either month is missing, the result is missing.

Real fed funds rate

The effective federal funds rate minus core PCE inflation for the same month.

Pearson and Spearman correlation

Pearson's r measures linear co-movement. Spearman's ρ applies the same formula to ranks, with tied values sharing their average rank, so it captures any steadily rising or falling relationship and is less sensitive to outliers. Both use every month where the two series are available.

Lead and lag

At a lead of k months, X in month t−k is paired with Y in month t, so a positive k means X moves first. Leads run from 0 to 36 months. A month belongs to an era according to the date of its Y observation.

Rolling correlation

Computed over a 36, 60 or 120 month window ending in each month, at the chosen lead. A window needs at least 80% of its month pairs present.

Regression

Ordinary least squares of Y in month t on X in month t−k, with an intercept. The page reports the slope, intercept, R², the number of months, the standard error of the slope and its t-statistic.

Summary statistics

Standard deviation uses n−1. Skewness and excess kurtosis are the bias-corrected sample versions used by common statistical software. Percentile is the share of all months in a series' history at or below its latest reading.

Eras

Glossary

ADF test

Augmented Dickey-Fuller test for a unit root. A low p-value means the series is stationary.

Bonferroni correction

Multiplies each p-value by the number of tests run (here 12 lags) to rule out lucky results.

Core inflation

Inflation excluding volatile food and energy prices. Headline inflation includes them.

CPI

Consumer Price Index for All Urban Consumers, published monthly by the Bureau of Labor Statistics (BLS).

Fed funds rate (FF)

The overnight rate banks charge each other, which the Federal Reserve steers as its policy rate; averaged over each month.

FRED

Federal Reserve Economic Data, the St. Louis Fed's free database, read by software through its API (application programming interface).

Granger causality

Whether one series' past improves forecasts of another. It measures predictive power, not cause and effect.

HAC standard errors

Heteroskedasticity- and autocorrelation-consistent (Newey-West) standard errors, which allow for overlapping months.

Index base period

Price indexes set a base period to 100: 1982-84 for CPI, 2017 for PCE. Growth rates do not depend on it.

Log scale

Equal vertical steps are equal percentage changes, so steady growth plots as a straight line.

M1

Narrow money: currency held by the public, checking deposits and, since May 2020, savings deposits.

M2

Broad money: M1 plus small time deposits and retail money market funds.

NBER recession

A recession as dated by the National Bureau of Economic Research. Shaded on the charts.

p-value

The chance of a result this strong if the test's starting assumption were true. Below 5%, or 1%, counts as significant.

PCE price index

Personal Consumption Expenditures price index, from the Bureau of Economic Analysis; the basis of the Fed's 2% target.

pp

Percentage points, the plain difference between two rates: 3% to 4% is a 1 pp rise.

QE

Quantitative easing: Federal Reserve bond buying to push down long-term rates, in 2008-14 and 2020-22.

r and ρ

Correlation, from −1 (perfectly opposite) through 0 (unrelated) to +1 (perfectly in step). Defined above.

R²

The share of the variation in Y explained by the fitted line, from 0 to 1.

SA and NSA

Seasonally adjusted, or not. Adjustment removes regular calendar swings such as holiday spending.

Series code

FRED's identifier for each series, such as M2SL. Sources lists them all.

Skewness and excess kurtosis

How lopsided a distribution is, and how heavy its tails are. A normal distribution scores 0 on both.

Standard error and t-statistic

The uncertainty around an estimate, and the estimate divided by it. A t beyond about ±2 is conventionally significant.

Stationary and unit root

A stationary series returns to a stable average; a unit root makes it drift, and drifting series can correlate by coincidence.

Zero lower bound

The floor near 0% for the policy rate, where the Fed held fed funds in 2008-15 and 2020-22.

Other abbreviations

B and T, billions and trillions of dollars; mo, months; n, months used; k, lead in months; Std dev, standard deviation; CSV, comma-separated values.

Caveats

• M1 was redefined in May 2020. The Federal Reserve moved savings deposits into M1, which roughly tripled its measured level in a single month. That is a change in definition, not new money. M1 growth from May 2020 onward is still drawn, dashed and labeled, but it is excluded from every statistic on this page.
• October 2025 CPI does not exist. The Bureau of Labor Statistics did not publish it because of the federal government shutdown. Headline and core CPI growth for that month are left blank, not filled in.
• End dates differ. The fed funds rate and the recession indicator run one month past the money and price series.
• Overlapping windows. Year-over-year rates share eleven of their twelve months with their neighbors, so consecutive readings are far from independent. Standard errors and t-statistics here assume independence and overstate precision; read them as descriptive.
• Revisions. FRED revises history. Figures reflect the data shown in the header.
• Correlation is not causation. Co-movement, even with a consistent lead, does not show that one series drives another. Common causes, policy reactions and changes in definitions can all produce it.

Sources