Vitreous Glass: A Low-Growth, High Dividend Yield Stock with Incredible Returns on Equity and Incredibly Frightening Supplier and Customer Concentration Risks
Vitreous Glass is a stock with some similarities to businesses I’ve liked in the past – NACCO, cement producers, lime producers, Ball (BLL), etc. It has a single plant located close enough to a couple customers (fiberglass producers) and with an exclusive source of supply (glass beverage bottles from the Canadian province of Alberta that need to be recycled) and – most importantly – the commodity (glass) can’t be shipped very far because the value to weight ratio is so low that the price of transportation quickly exceeds the price being charged for the glass itself (absent those shipping costs). The stock is also overlooked. It’s a microcap with fairly low float, beta, etc. It’s also a simple business. And capital allocation is as simple as it gets. The company pays out basically all the free cash flow it generates as dividends. And operating cash flow converts to free cash flow at a very high rate, because the company spends very little on cap-ex.
That’s most of the good news. The one other bit of good news is the stock’s price. As I write this, the stock trades at about 3.60 Canadian Dollars. I did a quick calculation of what seemed to be the normal trend in dividends these past few years. The company pays out a quarterly dividend that pretty much varies with quarterly cash flow from operations. So, it’s not a perfectly even dividend from quarter-to-quarter. But, it seems fairly stable when averaged over 4, 8, 12, etc. quarters and compared to cash flow from operations. If we do that – I’d say the current pace of dividends seems to be right around 0.36 cents per share each year. In other words, the dividend yield is 10%. There are other ways to estimate this. For example, we can use the price-to-sales ratio (EV/Sales would be similar, the company has some cash and no debt) and compare it to the free cash flow margin we’d expect. Then assume that all FCF will be paid out in dividends. Again, we get numbers suggesting future annual dividends are likely to be a lot closer to 10% of today’s stock price than say 5%.
There is some bad news though. One bit of bad news is the difficulty I’ve had in this initial interest post verifying certain important facts about the business. In preparation for this write-up, I read 3-4 different write-ups of this stock at various blogs, Value Investors Club, etc. I read the company’s filings on SEDAR (the Canadian version of America’s EDGAR). While I believe the information in the blog posts to be true -they’re getting those facts from somewhere, I can’t independently verify certain things about the supply agreement, the specific customers buying from Vitreous Glass, etc. Having said that, nothing I found in the accounting and in the filing overall really seemed to contradict what I read in the blog posts. And I did notice some stuff in the accounting that matches up pretty strongly with the way the business is described in some of those blog posts. Let’s start with some of the things that match up well.
Vitreous Glass has an exclusive supply agreement with what is essentially a part of the Canadian government in Alberta (it’s the province’s mandatory glass recycling program). Vitreous Glass has to take whatever it is supplied. So, it’s not that Viterous Glass is choosing to buy how much waste glass its customers need. Instead, Vitreous Glass has to buy ALL of the waste glass collected in Alberta. And then Vitreous turns that waste glass into material that can be used by fiberglass producers within a couple hundred miles of the company’s waste glass processing plant. There’s some stuff mentioned in these write-ups that jives well with unusual accounting items I noticed. I’ll mention those that may help to explain the company’s business model.
So, one logical question would be why don’t customers just “in-house” the conversion from post consumer waste glass to a fiberglass input. The answer may be that Vitreous Glass is really operating a staging area as much as a manufacturing plant. The company’s accounts strongly suggest this fact. It sometimes holds a lot of inventory. And the inventory is almost all raw materials. This makes perfect sense if two things are happening here: one on the supply side and one of the demand side. Let’s say hypothetically that waste glass is produced at a fairly stable level each year. This would make sense. I don’t think the people of Alberta are using vastly different amounts of glass beverage bottles each year. Supply could then be very, very stable. Like, Vitreous could be getting 2% more or 2% less waste glass sold to them each year. But, it wouldn’t vary by like 20% in any year. However, the demand from fiberglass manufacturers would vary a lot. This is because demand for things like housing starts – this fiberglass is being used in insulation – can swing 20% or more a year quite frequently. Well, if Vitreous glass is required to accept all the waste glass offered by the region – then, it will sometimes have its factory increase supply by 2% in a year where demand drops by 20% (or a lot more). This would increase inventory levels by more than 20% obviously. The reverse could also happen as well. If the market for waste glass was competitive – others were bidding against Vitreous Glass to buy the used beverage bottles – then, financial results would be quite variable. However, if Vitreous Glass is the only one taking waste glass in the region (which is what all write ups of the stock claim) then this changes the volatility of results a lot. In fact, I would expect that both the input cost of actual waste glass arriving at the plant and the output price – excluding certain transportation costs which probably would vary more – could be remarkably stable. In fact, they might actually be surprisingly stable on a non-unit but rather plant wide basis. Let me explain why.
So, if there is a competitive market for the supply of something – which is the situation all of us investors are most used to seeing – then, microeconomic analysis on a per unit of volume basis makes a ton of sense. It makes sense to imagine that the average retailer of soda makes a gross profit of 6 cents a can or whatever. But, does it make quite as much sense to assume that Ball – a company that provides ALL the supply of cans to certain beverage plants – always makes a gross profit of 3 cents a can or whatever? Well, it does and it doesn’t. Obviously, what matters to both Ball and the bottling plant sited near Ball’s beverage container factory is really the return on investment of those entire plants as projects. If volume assumptions were very different, then supply agreements on a per unit basis would have to be very different too. The more a plant supplies, the less it’d have to charge per unit. The less it supplies, the more it’d have to charge per unit to justify the investment. And because of the co-dependence, this calculation matters to both the company and its customer. Ball doesn’t want to build a plant to supply a customer unless it can get an adequate return on its investment. And the customer isn’t really making the decision of producing just one more can of Coke or not. The important decision is the entire planning of a bottling plant including who to use as the supplier of beverage cans to the plant. The critical microeconomic factors here are the expected returns on investment of each of the plants – the beverage can plant and the bottling plant – on a factory wide (rather than per can of soda) basis. But – and this might not be the best example, because the market for soda is super non-cyclical – it’s still true that after you output these cans, price and quantity is responding to demand for soda along with competitive factors and such on a per can basis. If people want to consume 5% less soda this year and 5% more healthier alternatives or whatever, this does have a meaningful impact on the economics of the bottling plant and that can pass through to the beverage can plant as well.
My guess would be that the situation with Vitreous Glass is – if anything – less variable in terms of the profitability as opposed to the volume of business being done. Whereas I think it’d be easier for the profitability of some things – like cans of soda – to vary for the entire plant, I think the situation at Vitreous Glass is more likely to be far bigger variations in physical quantities of stuff passing through the plant than we’ll see in terms of plantwide profit and losses. There are some indications of this in the footnotes to the financial statements I’ve read. It certainly seems like what’s happening is that Vitreous Glass is taking all of the variability in the chain of production of waste glass to fiberglass. It’s basically freeing its supplier (the waste glass collection program) and its customers (the fiberglass producers) from ever having to think about inventory levels. Vitreous Glass simply takes everything the collection program wants to get rid of (which, obviously, is all the glass bottles used in the region each year) and then only supplies its customers on a just-in-time basis. The company shows extremely minimal levels of finished inventory. If we were to visit the Vitreous Glass plant, I think we’d find that on each of our visits the amount of stockpiled waste glass would vary. Like, I think it might be possible to literally eyeball some extreme variations in physical inventory levels. And this variation should be based on the demand for fiberglass. When housing demand increases, demand for fiberglass insulation would spike. When demand for fiberglass insulation spiked, demand for fiberglass inputs (waste glass) would spike. However, the supply of waste glass can’t spike. That’s because waste glass is a byproduct of the amount people are drinking. And people in Alberta aren’t going to drink more or less simply because a lot more new homes are being built. So, production levels at the Vitreous Glass plant would spike along with housing activity. The company would be depleting inventory. This is because Vitreous can’t buy more or less waste glass depending on the demand for waste glass. It always has to buy whatever is on sale. And whatever is on sale is probably pretty inflexible because the amount of drinking done each year is probably pretty inflexible.
This would have some strange results for Vitreous Glass. One, cash flows would depend a lot on inventory levels – raw material piles – being increased or decreased. However, the economics on the supplier and customer sides really could be stable on like a plant wide basis. This is – sort of – what the historical numbers show. I say “sort of” because the results do show some variability in earnings. However, I can’t see how the variability in earnings is actually as high as what I assume the variability in demand for fiberglass insulation would look like. So, profit seems to vary less than the demand from customers here. This is believable in co-dependent economic relationships. The two parties share the volatility more than you are used to seeing in supplier-customer relationships that are more purely transactional and competitive.
One supplier and two customers account for like 90% of Vitreous Glass’s sales. By the way, because there are big fixed costs at any plant, this means that 1 supplier and 2 customers account for more than 100% of profits here in the sense that Vitreous’s profits are completely dependent on one supply agreement and two customer relationships. The loss of that supplier or of those customers would send Vitreous into the red. That’s obviously the risk here. There’s supplier concentration and customer concentration.
Though, to be fair, it’s bi-lateral. The recycling program is only using Vitreous Glass as its way of getting rid of this waste. And the customers are – possibly (I can’t confirm this based on just the report I’ve read so far) – relying on Vitreous Glass for even more (like 100%) of their source of waste glass than Vitreous is depending on them (more like 35% to 50%) of its total sales.
This kind of bilateral dependency can cause some weird “moat” stuff. So, Vitreous already has the customer relationships, the supply agreement, and the plant. Although it might seem logical for a customer to “in-house” waste glass conversion – there are reasons they wouldn’t do this.
One, the cost savings wouldn’t be high. Let’s pretend waste glass as an input is 80% of the cost of fiberglass insulation (I have no reason to believe it’s that high, but such a high number helps demonstrate my point). Now, let’s assume that – as some blog posts say – the cost of transport is 3 times the cost of the waste glass. That means 60% of the cost of the final product is transport costs and only 20% is the pre-transport cost of the waste glass. On top of that, Vitreous Glass would have some economies of scale by having two customers instead of one. It’s unlikely competing fiberglass manufacturers would agree to a co-venture of producing waste glass themselves. So, what’s left here in cost savings? 20% of the cost of the insulation? Not really. Because you’d lose the economies of scale. I have no idea how big those would be. Fixed costs at the plant seem quite significant to me. On top of that, Vitreous Glass has been in this business for like 20 plus years at the same location. There are clear indications of deflation here. The most likely explanation for deflation on the cost side is experience gains at the plant. These could be significant. Could they be 2% a year. 1%? I don’t know. But, it’s very likely that a newly built plant by an operator who hasn’t done this before would take several years to lower costs to the point the incumbent is at. And then we have the capital costs. These fiberglass manufacturers are offloading the speculative inventory needs on to Vitreous Glass. Right now, it seems to be a just-in-time process for the plants. It wouldn’t be if you were dealing directly with the waste glass collector. You’d also need to invest in the plant. It’s very likely that the capital costs of building a new plant are really high versus the tiny amount of needed future cap-ex at an old plant.
I’m not saying Vitreous can’t lose one or both customers to in-housing overnight. It could. But, this plant clearly has tremendous returns on equity. So, any observer would say the logical thing is to cut out the middleman. However, I’m seeing a lot of reasons not to do that. And the reasons I’ve talked about here are the hard reasons. I think there are also “soft” reasons not to do this. One, it’s hard to cooperate with a competitor. Two, it’s usually undesirable to complicate the running of your plant by taking a streamlined just in time process and making it one where you’re holding inventory. Three, cutting out the middleman would require some sort of negotiations (with the supplier of glass bottles for recycle). Four, the supplier is not a for profit entity and you (the customer) are. Five, it’s just easier to bid for your supplier and buy their entire old plant, existing supply agreement, long tenured manager and plant manager, employees, etc. You’d get whatever experience there is from doing that. But, six, if one customer bid to buy Vitreous – you’d have the risk of the other customer wanting to bid. The incentives would actually be doubled, because Vitreous is supplying competing local plants. So, a bidder would want any cost savings for themselves plus they’d want to deny any cost savings to a competitor if they believed such cost savings of cutting out the middleman really did exist.
I have no idea if Vitreous Glass’s customers have ever considered bidding for the company, have ever considered a co-venture with other possible users of waste glass, etc. But, it just doesn’t look like a situation where it definitely makes sense to cut out the middleman. I also don’t know if they could. Vitreous does not – at least in the report I read – go into the kind of detail on the length and terms of its contracts with customers, suppliers, etc. the same way that a company like NACCO does.
Vitreous Glass has some of the same risks that NACCO has. The customer concentration is more extreme. And NACCO owns the coal deposits. Vitreous Glass relies on a supply agreement. On the other hand, Vitreous Glass is in a business – providing a fiberglass insulation input – that could last a lot longer than coal. Vitreous Glass also has clearer capital allocation than NACCO does. Incentives here are good – as they are at NACCO, but probably even better here – with the CEO and his wife owning close to 40% of the company and the CEO being compensated primarily via a bonus tied directly to the level of cash flow from operations. Vitreous Glass also will have higher returns on equity, because it doesn’t need much reinvestment (while NACCO expands one consolidated mine, buys expensive equipment for its limerock business, etc.).
For these reasons, Vitreous Glass looks a lot like a bond that yields 10%. But, is it as safe as a bond? No. Definitely not. And the upside doesn’t much compensate for the lower safety. The company might – primarily through cost savings at the plant – increase dividends by like 2% or more a year indefinitely. But, that’d just make it a 10% real yield. That does compare favorably to what like a 2% real yield (assuming 2% inflation) on very long-term investment grade corporate bonds. Would I rather own Vitreous Glass than a 30-year corporate bond? Yes. Would I rather own it than the S&P 500? Maybe. But, unless I learn more – NOT at like a 100% S&P 500 allocation versus 100% Vitreous Glass. There are real risks here. However, one good thing for investors is that Vitreous Glass has pure “business risk”. If it fails, it’ll fail for reasons that won’t cause any problems for the rest of your portfolio.
The risks at Vitreous Glass have zero correlation with whatever risks are already in your portfolio.
So, right now, Vitreous Glass does seem like a really good diversifier. It might add some returns to your portfolio. And it won’t add any general market risk at all. So, it’s a totally idiosyncratic add to any portfolio. If you could find a basket of 10 stocks just like Vitreous Glass – you’d have yourself a solid portfolio that wouldn’t move with the market. Unfortunately, I can’t even think of like 4 stocks that are similar to Vitreous Glass in price, customer concentration risks, etc.
Given the risks here, I’m not yet completely sure a dividend yield of like 10% or so is high enough. But, I am sure I’d like to do more research on Vitreous Glass.
Initial Interest: 80%