Understanding this indicator
Markets often start to change before the most familiar economic statistics make the shift obvious. Purchasing managers’ surveys offer another vantage point: they ask manufacturers about conditions such as orders, output and employment. Bringing several surveys together can make a broad change in manufacturing easier to recognise than following one country alone.
This indicator uses the manufacturing PMI series in the sample as published survey readings. It does not turn them all into price-style annual returns. After adjusting for differences in trends and variation, principal component analysis identifies the strongest movement they share. The resulting line condenses the surveys into a common manufacturing cycle.
The indicator helps track whether manufacturing conditions are broadly improving or losing momentum. New orders, inventory decisions and production plans link manufacturers to commodity demand, transport and company earnings. These connections make the series useful alongside market-based cycles: it offers evidence from business surveys about the economic activity behind asset prices.
Compare its direction and turning points with equities and commodities, allowing for differences in timing. A market upswing accompanied by improving survey conditions tells a different story from one with little business-sector support. Because the composite is rescaled, its zero line is not the 50-point threshold used for an individual PMI; the main focus is the direction and relative position of the shared cycle.
The green line isolates the 3–4-year Kitchin cycle from the historical data; it is a fitted cycle, not a forecast of the next turn. The lower chart shows the 7–10-year Juglar cycle and the 15–25-year Kuznets cycle. Comparing these time scales can reveal why a shorter upswing sometimes sits within a much longer period of weakness. The two main-chart axes use different scales, so compare their timing and direction rather than the height of one line against the other. Historical curves may change as the data are updated.