Kansas Fed Composite Index

The Kansas Fed Composite Index is the headline result of the Tenth District Manufacturing Survey from the Kansas City Fed.

It covers manufacturing plants in the Tenth District states: Kansas, Colorado, Nebraska, Oklahoma, Wyoming, western Missouri and northern New Mexico.

The monthly survey started in July 2001, with quarterly history back to 1994. It draws about 90 to 130 responses each month through an online questionnaire. Plants are chosen by geography, industry mix and size.

It is a diffusion index that tracks how widespread growth or decline is across plants. It covers 3 horizons: month-ago, year-ago and expected in six months. So it is not a national gauge, not a services measure and not a gauge of output volume.

Responses are collected online from the 3rd Wednesday to the following Monday each month, a window of five or six days. Results appear once a month on Thursday, usually the 4th Thursday, at 11:00 AM Eastern, with November and December shifted 1 week earlier. Each release includes a press release, an industry summary, a history chart, result tables for 3 horizons, selected comments and history files. Month-ago and expected values are seasonally adjusted, with factors reset each January, so history is slightly revised.

The composite averages 5 equally weighted parts that capture different factory pressures.

• Production: change in plant output.
• Volume of new orders: inflow of new demand.
• Number of employees: change in factory hiring.
• Supplier delivery time: change in delivery times; slower times read positive.
• Inventories: Materials: stocks of raw materials.

In April 2020 the month-ago composite fell to -30, the lowest on record. Production was -66, new orders -60, employees -32 and materials inventories -15, while supplier delivery time rose to +24 because disruptions lengthened deliveries. Since slower deliveries enter with a plus sign, they partly offset the drop.

Activity then improved to -19 in May 2020 and to 2 in June 2020. This shows the index tracks breadth of change, not output volume.

1. Diffusion index: for each question subtract the share reporting a decrease from the share reporting an increase; ignore no-change answers. For example, among 100 plants, 30 report increase, 45 no change and 25 decrease, so the index is 5.
2. Seasonal adjustment: adjust month-ago and expected values only, not year-ago values, with Census X-13 ARIMA-SEATS. For example, a diffusion index of 5 with a seasonal factor of -2 becomes 3.
3. Composite average: average the 5 adjusted parts and round to a whole number; the headline is the month-ago result. For example, July 2024 parts of -12, -21, -12, -3 and -17 average to -13. An illustrative set of 3, 8, -1, 4 and -4 totals 10, divided by 5 equals 2.