Do Chinese Aluminum Import and Export Flows Affect LME Prices?

To start our discussion, let’s look at aluminum flows to and from China (i.e., imports and exports). In 2022, both imports and exports have increased dramatically, which is unusual. In general, low export volumes do correlate with high import volumes and vice versa. However, same-month imports and exports need to be viewed individually, not together or “netted.” This is mainly because troughs in exports usually precede peaks in imports by approximately one to two months.   

The following two charts will demonstrate key support and resistance levels for LME aluminum prices, and how they relate to aluminum shipments to and from China. First, we will look at imports. In both 2009 and 2020, imports increased dramatically within two to three months of hitting a bottom at approximately $1,500. This year has been an anomaly, as imports increased while prices were still at historically high levels.

Please note that this did not occur in 2016. Moreover, the $1,500/mt level seemed to be of little consequence before 2009.

Please also note that peaks in LME prices seemed to have little correlation with bottoms in import demand.

Chinese aluminum exports seem to have a very tight relationship with price. In general, China exports more as prices increase. Likewise, exports drop as prices decrease. More importantly, when aluminum prices peak, aluminum prices are likely near the top. When aluminum prices bottom, exports are also likely near a bottom within one to two months. Both $2,700/mt and $3,300/mt have proven to be inflection points for both prices and exports. It appears that “rest-of-world” demand shuts off at those levels.

For importers and exporters, changes in the spread between LME and SHFE prices are of critical importance. In general, we would expect to see imports increase when SHFE prices are high relative to LME. Similarly, we would expect to see exports increase when LME prices are high relative to SHFE.

The next two charts aim to further clarify why flows of imports and exports seem to change around the $1,500/mt, $2,700/mt, and $3,300/mt price levels. First, let’s look at imports. To do this, we will compare the SHFE minus LME spread to monthly unwrought aluminum imports. Since 2009, peaks in the SHFE-LME spread have generally occurred when LME prices dipped to approximately $1,500/mt.  In general, imports increase when the SHFE-LME spread is at $400/mt or higher, but volumes are inconsistent. For example, in April 2009, the spread averaged approximately $511/mt, and imports hit a then-record volume of 439,900 mt. However, when this spread was higher than $500/mt at several periods between 2011 and 2014, imports ranged from approximately 96,000 to 125,000 mt. More recently, when LME prices hit a pandemic-era bottom in May 2020, the SHFE-LME spread had been on an upswing for nearly 18 months. The spread hit a record-high monthly average of $624/mt in September 2021.

To decipher export trends, we will use the LME minus SHFE spread. Again, we would expect to see exports increase when LME prices are high relative to SHFE. This is by and large true. When exports hit a record-tying 300,000 mt in June 2008, the LME-SHFE spread hit a record-high monthly average of $335/mt in July. Similarly, LME prices also hit a then-record high of $3,317/mt in July 2008.

Between 2008 and the “pandemic era,” this relationship continued to hold. For example, when aluminum prices recovered to $2,700/mt in April 2011, the LME-SHFE spread hit a short-term peak of $143/mt that month, and exports hit a then record of 410,000 mt that same month.

However, during the pandemic era, this correlation has decoupled. In fact, between March 2020 and September 2021, this spread collapsed by nearly $500/mt, yet exports kept rising. We do note that exports really ramped up when the spread started to improve beginning in September 2021.

Conclusion:

Due to lag times in economic data, we would prefer to use real-time prices as leading indicators of peaks or troughs in aluminum flows to and from China. Luckily, we can do that for both imports and exports! However, because export data is released earlier than import data, both prices and the LME-SHFE spread seem to have a better correlation with exports. Except for the 2008 peaks in prices and exports, prices seem to be a leading indicator for potential peaks in exports. This lead time was generally zero to six months.

Extrapolating further, it appears that exports are more price sensitive than imports. More importantly, it appears that prices are more dependent upon “rest of world” demand rather than China.

Import data is still useful, but in our opinion, it’s not as useful as export data. This is because peaks in prices do not correlate well with troughs in imports. Moreover, since 2001, there have been only two very clear instances (2009 and 2020) where the $1,500/mt level triggered a significant surge in imports.