Methodology

About the Data

The Federal Reserve Bank of Philadelphia’s MBOS is a monthly qualitative survey that asks high-level executives about the direction of change in activity at their manufacturing firms located in the Third District (Delaware, southern New Jersey, and eastern and central Pennsylvania) for the current period as well as their expectations for changes in the future. The MBOS was launched in May 1968.

Survey Design

Each month, we email participants and collect their responses electronically. Typically, we collect responses over an eight-day period that closes the Monday prior to publication. Respondent participation is voluntary.

For the core monthly questions of the survey, we ask manufacturers to indicate the direction of change (increase, no change, or decrease) from the prior month for general business activity as well as for these specific indicators: new orders, shipments, unfilled orders, delivery times, inventories, prices paid, prices received, employment, and average employee workweek. We also ask firms to indicate their expected direction of change over the upcoming six-month period for all these indicators and capital expenditures. See a sample survey questionnaire for the core questions.

Respondents also have the opportunity to provide general comments and responses to special questions that we repeat less frequently than the core monthly questions, including our quarterly Price and Inflation Expectations Survey. See more information about our special questions in the MBOS FAQs.

Summary of Responses — Diffusion Indexes

We present the MBOS results as diffusion indexes for each current and future indicator (general activity, new orders, etc.), and the results are seasonally adjusted. Diffusion indexes aggregate the individual qualitative survey responses from the firms into quantitative measures, which facilitate tracking conditions over time. Current general business activity, the survey’s headline index, and its future counterpart are derived from distinct questions in the survey, not weighted averages or composites of other indicators.

For each indicator, we calculate percentages of responses for increase, no change, and decrease; seasonally adjust each of the three components; then construct the seasonally adjusted diffusion index as the percentage of firms reporting an increase minus the percentage of firms reporting a decrease. For example, if 40 percent of the firms indicated an increase in general business activity from the prior month, and 15 percent of the firms indicated a decrease, then the diffusion index for current general business activity would be 25.0. By construction, each index can take values between -100 and 100, inclusive, and has a midpoint of zero — a value that indicates that the percentage of firms reporting increases was the same as the percentage reporting decreases.

Revisions

We revise the MBOS historical data once a year to incorporate updated seasonal adjustment factors. In January of each year, we calculate new seasonal adjustment factors for each indicator’s components using data through December of the previous calendar year and apply those factors to the history of nonseasonally adjusted data (see Seasonal Adjustment for more detail). We publish the revised history files one week prior to the regular January release.

These new seasonal factors typically generate minor revisions to the seasonally adjusted data for prior years. These annual revisions also mean that values found in the historical data files currently available to download will typically differ slightly from values cited in archived PDF reports from prior years.

Seasonal Adjustment

Due to atypical volatility at the onset of the pandemic and the long, steep recovery period that followed, we have been excluding data from 2020 and 2021 from our analysis of seasonal factors. To revise seasonal adjustment factors for prior years and forecast factors for the current year (2026), we:

• calculated new seasonal adjustment factors using the U.S. Census Bureau’s X-12 procedure based on data from January 1976 through December 2019, appended with data from January 2022 through December 2025.
• calculated factors for each month of 2020 and 2021 by interpolating between each month of 2019 and 2022.
• applied revised seasonal adjustment factors to the nonseasonally adjusted data from January 1986 through December 2025.
• forecasted seasonal factors for each month of 2026 using the X-12 ARIMA method.