Chicago Fed National Activity Index

The Chicago Fed National Activity Index (CFNAI) is a monthly composite gauge of national economic activity. The Federal Reserve Bank of Chicago builds it from 85 inflation adjusted national series in 4 groups. It is not a survey and it has no panel, so there is no sample size in the survey sense.

The index tracks real activity around its long run trend. A value of 0 marks trend growth, with higher values pointing above trend and lower values pointing below trend in standard deviation units. CFNAI is a coincident indicator of expansion and contraction, not a Chicago gauge and not a forecast. It is also not a direct gauge of inflation, though firm or soft activity can relate to inflation pressure.

Each release covers one full reference month of underlying data. About one third of the 85 series are unpublished at release time, so those values start as model estimates. The release arrives at 8:30 in the morning Eastern Time toward the end of the following calendar month. Release timing can shift or combine when source data are delayed.

Each release reports 3 headline numbers, with the monthly index plus the 3 month average CFNAI-MA3 and the 3 month breadth gauge. It also shows the contribution of each of the 4 groups and counts of positive and improved series. It includes revision notes and charts plus the schedule for the following release. Values are revised with each release as actuals replace estimates and as sources revise inputs.

Weights are refit with each release, with a small effect in practice. The monthly index is volatile, so analysis rests on the 3 month average and breadth, never on one month alone.

CFNAI moves with broad real activity across 4 groups, each with a fixed weight share. Shares sum to full weight and each group reacts to the cycle in its own way.

• Production and income, 31.9%: It captures factory and mining output, capacity use and real income. It turns negative fastest in recessions and positive first in recoveries.

• Employment, unemployment and hours, 32.8%: It captures payrolls, unemployment rates, hours and claims. It often lags, with its weakest contribution near the end of recessions, and higher unemployment pushes the index down.

• Personal consumption and housing, 13.5%: It captures real spending, real retail sales, housing starts, permits and mobile home shipments. The publisher finds no clear recession pattern for this group.

• Sales, orders and inventories, 21.8%: It captures real factory and trade sales, new orders including ISM orders and inventory ratios. Higher inventory ratios push the index down, and the group turns with production.

Each series is filtered into stable form, with jobs and output expressed as growth rates. Other series use differences, logs or levels, as listed in the method. Each filtered series is centered to mean 0 and scaled to standard deviation 1. This puts all series in common units before they are combined.

The index equals that month data times a weight vector from the first principal component. That component captures the shared movement across the 85 series. The result is reset to mean 0 and standard deviation 1, so readings are steps above or below trend.

Companions smooth the signal, with CFNAI-MA3 as the 3 month average. Diffusion nets positive weight share minus negative or neutral share over 3 months. Missing series use model estimates and weights are fit on trimmed data but applied to full data. Inputs arrive already adjusted for seasonality by source agencies, with no added adjustment by the publisher.

Do not confuse CFNAI with Chicago PMI, which is a separate regional survey. Treat cycle thresholds with care since they were set with hindsight on revised data.