The baseline view here has been that the global stockbuilding cycle would enter a downswing in 2026 into a low in H1 2027.

The stockbuilding cycle is a repeating fluctuation in the demand for production inputs and goods for final sale caused by inventories periodically over- and undershooting the level desired by manufacturers and distributors.

The approach here is to monitor the cycle using measures of the rate of change of G7 stockbuilding from national accounts data and business surveys.

The national accounts measure places the last cycle low in Q1 2023, implying that the current cycle is well-advanced – see chart 1.

Chart 1

The business survey measure is more timely and recently crossed below zero, consistent with the cycle moving into a downswing.

The original research on the cycle – by Joseph Kitchin, published in 1923 – used data on commodity prices, bank clearings and interest rates. Kitchin attributed the cycle to “psychological causes”, making no explicit link with inventory behaviour.

The cycle is evident in commodity prices because changes in demand for production inputs affect their price as well as volumes – chart 2.

Chart 2

Tech investors often dismiss the cycle as an “old economy” phenomenon but electronic components are a key production input subject to the same inventory-related demand swings as traditional industrial commodities.

There is no single price index capturing the changing nature of electronic components. However, export prices of key supplier economies are a rough proxy. Chart 3 shows that the annual rate of change of an average of Taiwanese and Korean export prices in US dollars tracks the national accounts-based measure of G7 stockbuilding changes.

Chart 3

Chart 4 overlays the annual rate of change of an average of 12-month forward earnings estimates, showing a similarly significant correlation.

Chart 4

Based on the cycle assessment here, the implication is that momentum of component prices and earnings should be at or close to a peak.

Why could this prove wrong? One possibility is that the current stockbuilding cycle will extend. This cycle is judged to be the last of a set of five constituting the current housing cycle (which began in 2009). Final cycles tend to be longer than the 3.5-year average, with one recent example (1986-91) reaching 4.5 years. The next low, therefore, could be delayed until H2 2027, in turn suggesting a later start to the downswing.

Another counter-argument is that strong demand for components is being driven by an ongoing upswing in the business investment cycle – narrowly focused on AI-related spending in the current cycle – as well as peak stockbuilding. This cycle last bottomed in 2020 and has ranged between 7 and 11 years historically, so the upswing could – in theory – continue for several more years. The thinking here is that the current cycle is more likely to be short, reflecting a long prior cycle (11 years) and a drag from the expected stockbuilding downswing.

Cycle analysis indicates that the global economy is in a time window for weakness, suggesting a significant risk that the Gulf War III shock triggers a recession. Real money trends will be key for assessing whether a negative scenario is playing out.

The housing, business investment and stockbuilding cycles average 18, 9 and 3.5 years respectively. The most recent lows are judged to have occurred in 2009, 2020 and 2023, suggesting that the next bottoms will be reached around 2027, 2029 and 2027. All three cycles, therefore, are expected to be in downswings over the next 1-3 years.

Cycle history suggests two possibilities. If the three downswings coincide, a major recession is likely. Historical precedents include the severe global downturns of 1974-75 and 2008-09.

If the cycle lows are spaced out over several years, the template would be the early 1990s – a longer period of rolling economic weakness involving a less damaging recession.

An earlier episode of triple cycle weakness in the late 1950s was also associated with a less pronounced recession but the fall in output on that occasion was limited by strong trend economic growth, reflecting post-war reconstruction.

The impact of shocks on the global economy depends on the cyclical backdrop. Activity bounced back strongly after the 2020 covid shock partly because the stockbuilding and business investment cycles were in time windows to enter recovery phases, while the housing cycle remained in an upswing.

Similarly, economic damage from the 2022 energy shock due to Russia’s invasion of Ukraine was limited by support from the business investment and housing cycles, with only the stockbuilding cycle then in a weak phase.

The timing of the Gulf War III shock echoes the 1973 Arab oil embargo, which hit as the three cycles were peaking and resulted in synchronised and self-reinforcing downswings into 1975 lows.

There are important mitigating differences from the 1973 shock. The oil price rise has been much smaller, while the oil intensity of GDP has fallen significantly. The 1973 shock occurred against a backdrop of double-digit G7 money growth, ensuring an inflationary outcome – current expansion is still low. Surging inflation forced major monetary policy tightening, the combined result being a severe real money squeeze – see chart 1.

Chart 1

Real money trends appeared modestly supportive before the current shock: global / G7 growth had firmed into early 2026, suggesting that economic expansion was on course to hold up through Q3.

The mechanical impact of higher energy and other costs on consumer price inflation will ensure a sharp slowdown in real money momentum into mid-year – chart 2. The extent of the decline will be key for assessing the likely degree of economic weakness. As noted, modest money growth argues against significant “second-round” inflation effects but central banks are hinting at precautionary tightening, which would magnify monetary weakness.

Chart 2

A further risk is of “endogenous” monetary tightening if the Gulf War III shock interacts with recent problems in private lending, leading to a generalised reduction in credit availability. Such a shift could be signalled in ECB and Fed loan officer surveys due in late April and early May respectively.

Country real money numbers through February suggest that US economic prospects were improving absolutely and relative to other majors before the shock – chart 3.

Chart 3

Japanese monetary weakness continues to argue that BoJ policy tightening – via large-scale QT as well as rate hikes – has been misguided. As expected, core CPI inflation – ex. food and energy – has fallen and is below 2% even stripping out the impact of government subsidies.

Recoveries in Eurozone and UK real money momentum have stalled at unimpressive levels, suggesting dull economic prospects before the shock. Within the Eurozone, readings are similar across the large economies, with France no longer a negative outlier.

Chinese real money momentum has slowed but may hold up better than elsewhere going forward, reflecting stable interest rates and government intervention to limit price rises. A strong balance of payments position, partly stemming from a still significantly undervalued currency, is generating monetary inflows.

Cyclical equity market sectors had started to underperform before the shock. The cyclical / defensive relative is correlated with the stockbuilding cycle, which is not expected to bottom before late 2026 at the earliest – chart 4.

Chart 4

 

The long-standing forecast here has been that the three key global economic cycles – stockbuilding, business investment and housing – would enter downswings by 2026-27.

In the event that the downswings were synchronised, the result would be a major recession, on the scale of 1974-75 or 2008-09.

If the downswings occurred successively, the result would be a long period of rolling weakness including a less severe recession – the early 1990s scenario.

Chart 1 is an exhibit used in 2018, suggesting – based on average cycle timings – a recession bottoming in 2019-20, a slowdown into 2023 and a major recession around 2027. The former two played out.

Chart 1

The housing cycle is already in a downswing, the stockbuilding cycle is at or close to a peak, while the business investment cycle is well-advanced. The Gulf War III energy price shock could be the trigger for synchronised weakness.

There have been six oil price “shocks” since the 1960s, defined here as episodes in which the spot price rose 50% or more above its two-year moving average for at least three months. All were followed by a significant contraction in G7 industrial output and five were associated with a US recession, as determined by the NBER – chart 2.

Chart 2

Spot Brent of $100 per barrel represents a 37% premium to the two-year moving average. Oil would have to be sustained at about $115 for three months for the current spike to qualify as a “shock”.

The oil intensity of GDP has fallen dramatically since the 1970s, suggesting less vulnerability to shocks. This could explain why the most recent shock – in 2022 – did not involve a US recession.

An alternative view is that a recession was avoided because the cyclical backdrop was less unfavourable and there was pent-up demand due to pandemic disruption, reflected in a large monetary overhang. The stockbuilding cycle turned down in 2022 but the housing and business investment cycles were in late upswing and recovery phases respectively.

The prior shocks caused short-term inflation spikes and were associated with upward pressure on interest rates, contributing to a fall in nominal narrow money momentum. G7 real money contracted ahead of maximum economic weakness. Real money data will be key for assessing the extent of current economic damage.