The business climate indicator: an updated tool for assessing the state of the economy

Chart 1: Aggregate business climate indicator based on the monthly business survey

The real economy is constantly evolving, but monitoring it day-to-day remains difficult. It takes several weeks to compile so-called “hard” data (GDP, production, employment) before it can be published. Consequently, the Banque de France’s monthly business survey (MBS) is an invaluable tool for closely monitoring the state of the economy, assessing the economic situation and anticipating short-term developments.

From qualitative data to a composite indicator

During the last week of each month and the first week of the following month, the Banque de France conducts a survey of 8,500 business leaders across France through its network of branches and economic centres. These surveys pull together the sentiments and expectations of business leaders obtained during an interview in which they are asked about changes in their production, cash flow, order books, inventory, input costs and selling prices over the past month, as well as their expectations for the coming month and their constraints (supply, employment, etc.).

Based on their answers, balances of opinion are calculated by weighting the difference between the proportion of firms reporting an improvement/increase and the proportion reporting a deterioration/decrease for each given variable. These indicators, widely used in opinion surveys (OECD, 2003), summarise the prevailing trends as perceived by businesses and enable their monitoring over time. By capturing business leaders’ experiences and expectations, they provide reliable signposts indicating future developments and contribute to the Banque de France’s growth forecasting models.

These balances of opinion are consolidated into a composite indicator: the business climate indicator (BCI). The Banque de France has recently updated its BCI and it now summarises the state of business activity in a single figure. It indicates periods of expansion or slowdown by referring to a long-term average benchmark reference of 100.

An aggregate indicator for monitoring the economic situation

Over the recent period, the aggregate indicator has reflected generally moderate economic conditions, marked by a slowdown since 2023. The indicator has remained below its long-term average since then, despite a recovery in 2025 (Chart 1).

The indicator exhibits a certain volatility over the short term, linked to the sensitivity of firms’ responses to their environment. The trend is therefore smoothed to better identify the phases of the economic cycle. 

Although it is not intended to directly forecast overall economic growth – measured by GDP – the aggregate indicator demonstrates a significant correlation with quarterly GDP. It therefore presents a quick and consistent overview of the economic situation. As it is published monthly around ten days after month-end, it allows for far more reactive monitoring than the initial GDP growth estimate published quarterly by the Institut national de la statistique et des études économiques (INSEE, the National Institute of Statistics and Economic Studies) around thirty days after quarter-end.

Contrasting sectoral dynamics

In addition to the overall indicator, sectoral BCIs facilitate more detailed analyses by highlighting the heterogeneity of economic conditions. Industry, market services and construction obviously do not exhibit the same dynamics. Some sectors appear to be more sensitive to demand cycles, while others are more affected by specific supply or cost constraints. This sector-specific approach allows the reader to go beyond the aggregate indicator and to gain a better understanding of the economy’s current underlying structure.

Chart 2 shows trends in the major sectors (industry, market services and construction) since 2021, obtained by smoothing the sectoral indicators. Following a post-Covid rebound, the economic environment has deteriorated, particularly after the outbreak of the war in Ukraine. The downturn in the construction sector was slower to materialise: its BCI trended down below the 100 mark at the end of 2023, staying under 95 ever since and showing no real signs of recovery. Market services broadly followed the same dynamic as industry between 2022 and mid-2024. From that date onwards, we see an improvement in industry driven by a rebound in production in certain segments such as the defence and aeronautic sectors. These developments closely reflect the data from successive monthly business surveys.

Chart 2: Trends in the business climate indicator based on the monthly business survey, by major sector of activity

An updated method, focused on economic data, for a better interpretation

A more detailed analysis may be obtained by examining the contributions of the main balances of opinion that make up each indicator. Until recently, for the three major sectors of industry, market services and construction, the Banque de France’s sectoral BCIs were based on principal component analysis (PCA), which is an effective statistical method for summarising large amounts of data but which can have limitations when used for analysing short-term economic developments.

Combining PCA with smoothing tends to attenuate and delay cyclical economic turnarounds. Furthermore, the implicit weightings derived from PCA are not particularly clear and sometimes introduce balances that are less directly linked to immediate business activity, thereby making it difficult to interpret changes in the indicator. For example, incorporating several employment-related balances, such as current workforce and projected workforce for the following month, can result in an indicator that closely tracks employment trends and disregards the broader dynamic of the sector in question.

Chart 3: Breakdown of the industry BCI (deviation from its long-term average; in percentage points)

The new approach adopted for industry and market services relies less on the combination of balances chosen and more on the economic data they contain (for a detailed presentation of the new approach, see the MBS methodological note). In practical terms, we use a partial least squares regression to identify the balances of opinion in the monthly business survey that are most closely linked to changes in production as measured by INSEE. Each balance is then weighted according to its ability to explain these changes: the more it correlates with changes in production (in industry) or activity (in services), the greater its weight in the final calculation. Finally, the resulting BCI summarises the most relevant information for monitoring economic activity, with a key advantage: it allows us to anticipate changes in production approximately one month before the data are officially published.

Directly aggregating balances (or sectoral BCIs when compiling the overall BCI) makes them easier to interpret, and as it is based on explicit weightings, the factors that influence business leaders’ assessments (such as cash flow problems or a decline in order books or an upturn in production) can be very precisely explained. Consequently, the BCI provides a more immediate reflection of changes in business leaders’ responses, without the lag effects associated with smoothing, and ensures that the indicator is more transparent and can be more readily interpreted. The new back-cast data series are available on Webstat, the Banque de France’s statistics portal. 

This change of methodology means that the determinants of business sentiment can be analysed in greater detail. For example, Chart 3 sets out the contribution of key factors affecting business sentiment in industry, month by month, since January 2025. It reveals the perception of persistently deteriorating cash flow, which contributes negatively to the indicator. Conversely, the labour market’s contribution has generally been positive throughout the period. Variations in the indicator are mainly due to fluctuations in production and demand, which, while negative in the first half of 2025, have proved to be more favourable since then.