PMI Divergence: A Historical Lens on Manufacturing Signals

The November data presents a stark puzzle. The ISM Manufacturing PMI fell to 48.2, its lowest in four months, marking the ninth consecutive month of contraction. This paints a picture of a sector in deepening trouble, with new orders and employment both contracting. Yet, the broader S&P Global PMI indicated expansion at 52.5. This isn't a minor statistical blip; it's a structural split rooted in how the surveys are built. The core investor question is whether this divergence signals a bifurcated recovery-one where large multinationals are struggling while a broader mix of firms is improving-or if it reveals a fundamental flaw in one of our primary economic indicators.

The methodological differences are systematic. The ISM survey is based on its member panel, which is heavily skewed toward larger, often multinational corporations. In contrast, S&P Global's IHS Markit survey is designed to poll a representative mix of company sizes, mirroring the true composition of manufacturing output. This structural bias means the ISM is more likely to capture the struggles of big industrial players, while the broader survey may be more sensitive to the health of smaller, domestic-focused manufacturers. The divergence isn't random noise; it's a feature of the data collection process.

This split has real implications for interpreting the economic signal. For investors, it creates uncertainty about the true state of the manufacturing engine. If the ISM is capturing a specific segment of the sector in distress, while the broader economy is improving, the path for cyclical stocks could be bumpy. Conversely, if the ISM's methodology is consistently distorting the picture, relying on it could lead to premature bearish calls. The bottom line is that two surveys, built on different foundations, are telling two different stories. The market must decide which narrative to believe-or if the truth lies somewhere in between, requiring a more nuanced, company-size-specific analysis.

Methodology as the Key: Unpacking the Survey Differences

The conflicting signals from the ISM and S&P Global PMIs are not a mystery of the data itself, but a direct result of how the data are collected and calculated. The divergence points to a fundamental difference in data quality and reliability. S&P Global's methodology, with its larger, more representative panel and more sophisticated weighting system, consistently correlates more strongly with official government output data, suggesting a higher signal-to-noise ratio.

The core issue is panel composition. The ISM survey relies on a smaller, membership-based panel that skews toward larger, older industrial firms. This creates a narrower slice of manufacturing, potentially missing the broader economic picture. In contrast, S&P Global surveys about 800 companies, including a more balanced mix of smaller and mid-sized firms, providing a broader and more representative view of the sector. This larger sample size also generates more stable data with lower volatility, making it easier to identify true turning points in the business cycle.

A critical methodological difference lies in how the surveys weight their components. The ISM uses a simple, equal-weighted average of its five sub-indices. This can be distorted by counterintuitive components like supplier deliveries, where slower deliveries historically signaled demand strength but now drag the index lower as supply chains stabilize. S&P Global's system adjusts for these distortions by giving more weight to forward-looking components like new orders and production, which are more directly tied to actual output trends.

The bottom line is that the ISM's methodology appears to generate more "noisy" data. Its indices have shown greater month-to-month volatility and a weaker statistical relationship with official output and employment statistics over the past decade. This suggests the ISM may be more sensitive to the specific conditions of its larger, older industrial respondents, while S&P Global's approach provides a clearer, more consistent signal of the underlying manufacturing trend. For investors, this means S&P Global's data offers a more reliable early warning system for shifts in the broader industrial economy.

Historical Context: Lessons from Past PMI Divergence

The September 2025 divergence between the ISM and S&P Global PMIs is not an anomaly. Historical patterns show that methodological differences between surveys can create misleading signals, and that the ISM's contraction signals have often been more reliable indicators of downturns than its growth readings. This episode mirrors past cycles where the ISM's focus on larger multinationals and its specific panel structure led to exaggerated growth signals, while its contraction readings better tracked actual manufacturing trends.

The core of the divergence lies in survey design. The ISM survey, which focuses on larger multinationals, has historically shown considerably higher levels than the IHS Markit indices through 2017 and 2018. This pattern of ISM over-optimism was so pronounced that statistical analysis revealed the IHS Markit data had consistently higher correlation coefficients and adjusted r-squares than the ISM data when compared with official output, factory orders and employment data. In practice, this means S&P Global's survey has been a more accurate tracker of actual manufacturing output over the past decade.

This historical reliability suggests a pattern: when the ISM hits a 5-year low, as it did in September, it may be sending a more credible signal of economic weakness than its headline number alone suggests. The divergence in September, where the ISM PMI sank to 47.8 while S&P Global's hit a five-month high of 51.1, is a textbook example of this dynamic. The ISM's contraction signal, while extreme, may be more aligned with the reality of a manufacturing recession that both surveys agree was extended into the third quarter.

The bottom line is that investors should treat the headline PMI numbers with caution and look deeper. The historical evidence points to S&P Global's methodology-its larger, more representative panel and focus on US factories-providing a more stable and accurate picture of the sector's health. In a period of conflicting signals, the survey with the stronger track record of correlating with official output data is the one to watch.

The Bifurcated Recovery: What the Divergence Reveals

The conflicting signals from the ISM and S&P Global PMI surveys are not a statistical anomaly. They are a clear map of a manufacturing sector in transition, revealing a sharp bifurcation between winners and losers. The S&P PMI's expansion, driven by its broader sample of smaller and mid-sized firms, captures strength in domestic-focused sectors like consumer goods and tech-linked producers. This reflects the resilience of areas tied to US capex and the digital infrastructure build-out, where investment pipelines remain robust. In contrast, the ISM's continued contraction points to persistent weakness in the traditional industrial core-capital goods, chemicals, and heavy manufacturing-still grappling with global demand softness and tariff uncertainty.

This divergence creates selective opportunities. For investors, it's a signal to lean into firms benefiting from the new US investment cycle. Companies leveraged to digital infrastructure, semiconductors, and the retooling of domestic supply chains are the emerging winners. Their fortunes are tied to the S&P PMI's strength in new orders and production. The risk, however, lies in the old global goods cycle. The ISM's weakness is a reminder to be cautious on export-heavy manufacturers and structurally challenged end-markets like autos, where global demand remains soft. The data suggests these firms are still stuck in the doldrums.

The bottom line is that a single manufacturing headline is meaningless. The real story is one of corporate fortunes diverging. The market is already pricing in this split, rewarding firms in the new cycle while punishing those reliant on the old. For a selective recovery to be durable, the S&P PMI's strength must translate into broader employment and capacity utilization. Until then, the divergence is the investment narrative: a sectoral winner-take-all dynamic where exposure is defined by which cycle a company serves.

Risks & Guardrails: When the Divergence Breaks

The PMI split is not a minor data quirk; it is a signal of a fractured manufacturing recovery. The divergence reveals a sector in two minds: domestic-focused firms, particularly those serving the data center and industrial automation boom, are seeing strong new orders growth, the fastest in 20 months. This group is insulated from global trade frictions and is driving the S&P Global expansion. Conversely, traditional heavy manufacturers are grappling with tariff impacts, weak exports, and destocking, pulling down the ISM index. For investors, the critical risk is that this split is not a sustainable equilibrium. The ultimate guardrail is the health of the broader economy, which both surveys must eventually reflect.

The first failure point is a sustained reversal in the S&P Global's new orders growth. That headline is the engine of the expansion narrative. If this surge in domestic demand falters, the divergence collapses into a single, negative signal. A drop in S&P's new orders would directly undermine the story of a resilient, homegrown manufacturing recovery and signal broader weakness. It would force a reassessment of the entire sector's momentum, regardless of the ISM's contraction.

The second, more systemic risk is that the ISM's contraction deepens into a severe recession signal. The ISM's latest reading of 48.2 in November shows the downturn accelerating. If this trend continues, it could trigger a broader market reassessment of economic momentum. The S&P Global's expansion would then appear as an outlier, a fragile bubble in a deteriorating macroeconomic environment. This is the scenario where the survey with the larger, more representative panel-historically shown to have a stronger relationship with official output data-proves to be the more reliable leading indicator.

The ultimate guardrail, however, is historical precedent. The evidence shows the ISM has a track record of sending exaggerated growth signals in past cycles. This is a structural flaw, not a one-off error. It means the ISM's contraction readings, while painful, may be more accurate than its expansion signals. The risk is that both surveys fail to capture a looming downturn because they are measuring different parts of a complex, tariff-reshaped economy. The S&P Global might miss the export and heavy-industry weakness, while the ISM might miss the domestic demand surge. When the divergence breaks, it may not be a clear signal of recovery or collapse, but a confirmation that the old manufacturing economy is gone, and the new one is still being defined.