Source: Tradingview and Tellimer Research
Our machine-learning model forecasts a further run-up to US$231.16 over the next five months – representing a 7% increase from current levels. However, this is likely to mark the top of the current bull run. The model indicates a protracted correction, down to US$162.00, by April next year – representing a 43% drop from the forecast highs and 33% from current levels.
The current move has diverted prices over 58% away from their average over the past year, representing the largest divergence on record. Any additional increase in this divergence will become unsustainable over the next 6 to 12 months, and will likely attract some profit-taking and, possibly, even short selling.
Divergence from rolling 1-year average price
Both 2008 and 2011 put in significant cyclic highs around US$170-190 and both ultimately gave way to huge corrections of approximately -70% or more. The current 12-month spike-up has happened with such ferocity that it could be destined for much the same fate. Of course, these types of moves rarely turn on a dime, so timing the correction can be difficult.
The top of the expanding channel that encapsulated much of the price movement from 2015 to 2020 will now act as support to any sizable retracement between 30-35% from current levels, and could even act as a launchpad from which a longer multi-year trend could be built.
Iron ore and steel correlation
The current 1-year pairwise correlation between iron ore and steel is 91%, which is about as tight as any correlation found in finance. It is very close to its record of 92%, set in 2015.
Rolling 1-year correlation between iron ore and steel rebar