I am bored AF of reading AI slop opinions about SaaS and Mag7 companies, so here is something totally different: where I see good value in the commodities industries.
Commodity businesses are an overall mid investment category because they are hugely cyclical and so buying at the wrong time can ruin your future growth trajectory. I suspect this is something everyone about to throw their inheritance into soon-to-IPO LLM developers is going to discover in an incredibly painful manner (lmao). But anyway, commodities do at least have one path to an enduring moat, which is being big and ancient as shit, such that you have economies of scale and assets/permits that can’t be replicated because you got them before environmental regulations were a thing. Hence, the best companies in these industries tend to be market-dominant oligopolies that own their commodity market alongside 2-3 peers (think ADM, Bunge, Cargill, and Louis Dreyfuss for the world’s grain/oilseed supply). Those are the ones I’m gonna talk about.
Even the best of these companies is often a terrible investment because of the cyclicality, and unfortunately the best time to buy them is often when they look like falling knives that are still overvalued and have more to lose (because earnings collapse at cyclical troughs, inflating the P/E ratios; this also makes them look like ‘deep value’ at their peaks if you just look at earnings-based multiple, when actually that’s the worst time to buy). So if the P/E is misleading, what metrics make for a better screen? I have found three particularly useful, and for each, the current value \*relative to its distribution across the 10-year historic range\* is more helpful than the absolute current value itself (you can get this data visualized through various sources; I use Koyfin). IMPORTANT: do not use these to screen all commodity businesses, only to compare the giant proven winners. These values are likely to mean-revert for the survivors across cycles, which gives the predictive power, but often the little guy just goes bankrupt.
Price to book vs 10yr distribution of P/B: the business is cheap relative to its assets at cyclical troughs, because it’s suffering from lower margins and “what have you done for me lately?” return-chasers have moved elsewhere.
EV/Sales vs 10yr distribution: this is less likely to mislead vs PE or EV/EBITDA because it strips out the cyclicality of commodity margins. Lower is better. I like the EV version because it punishes excessive debt, and companies with bad capital allocators in charge often overexpand when they are doing well.
Dividend yield vs 10yr distribution: a confident management team will sustain vs cut their dividend during downturns, which means a yield at or near 10yr highs suggests insiders believe a recovery is coming soon and the business is healthy enough to get there.
Ok, all that out of the way, here are some names that flag green on the metrics I just described and are worth checking out IMO:
\*Ball Corporation and Crown Holdings\*: these two companies make almost every aluminum can that holds your beer, soda, sparkling water, caffeinated gas station beverage, etcetera. Are GLP-1s going to disrupt soda consumption? Maybe. Are they going to disrupt all canned liquid consumption? Probably not. I love CCK and BALL.
\*Hormel, Tyson, and Cal-Maine\*: The protein stack I guess? These are the major US producers of pork, chicken, and eggs and they each exist in longstanding oligopolies. They all stand to benefit from the average American becoming too poor to afford beef and too insecure about their masculinity to eat soy or go vegetarian (this isn’t my investment thesis, it’s just also true). Whether or not the AI capex cycle collapses before diffusion models kill ADBE, people will still go out and buy bacon, eggs, and chicken nuggets.
\*Nutrien and Suzano\*: Fertilizer and paper pulp. Nutrien is a Canadian company with some of the only economically viable supplies of potash outside of Russia, and is a dominant North American and global supplier of potash as well the #2 supplier of nitrogen fertilizer (after CF, which is another of these strong commodity moat businesses, but which has already had a big multiple expansion due to the Iran conflict) and an important supplier of phosphorus. So in other words, they have a dominant position in (read: price-setting for) all of the key chemical inputs needed to grow food. Suzano is a Brazilian company cutting down the rainforest to grow eucalyptus and then grind it down into pulp to make the world’s cheapest paper. Evil, yes, but also a strong business model that is unlikely to be disrupted by AI.
Where I don’t see value right now: Mining/quarries and Oil & Gas. For the former, Rio Tinto, BHP, Vale, Vulcan Materials, Martin Marietta, Nucor, etc are all great businesses that look like they are at cyclical peaks. If they are not at peak, it’s likely because we are indeed going to see a sustained construction boom from a successful continued AI capex cycle - in which case just buy NVDA and TSMC and all that other sexy shit. For the latter, O&G companies looked like a great buy just 6mo ago, but they’ve enjoyed big run-ups due to the war in Iran, and I can’t tell what’s happening will happen with oil next. So, these go in the “too hard” pile for me right now.
Not investment advice, also don’t take investment advice from Reddit.