Thanks for the opportunity to use the ancient meme.

One of the peculiar risks of value investing is that the underpriced company that you spent all that time and effort analyzing will be bought out from under you before the gap between its price and intrinsic value is resolved. Ordinarily the acquisition comes at a premium to the market price, but sometimes even this premium price is less than one’s calculated value, and ironically the larger the margin of safety and the more attractive the company’s price, the greater the risk is. It happened to me years ago with Qwest, and the Interpublic Group (IPG) acquisition is another example.

Interpublic Group is an advertising company that I ran across some months ago that reflected this situation. Last December it was announced that IPG would be acquired by Omnicom, a slightly larger advertising firm (and what is with advertising companies having sinister names?). The FTC has given its final approval and the merger is likely to close soon, and unfortunately the price has settled somewhat on the low side of fair value.

Actually time has made a fool of this deal; when it was proposed in Dec 9th, each share of IPG was to be exchanged for .344 shares of Omnicom, but Omnicom was at $102 before the merger was announced and $92 immediately afterwards. At the time this implied a price of $31.65 for Interpublic Group. But as of this writing Omnicom is at $72.65 which puts IPG at almost exactly $25, just a tiny premium over the current share price of $24.85 representing the market’s near certainty that the merger will go through as planned.

However, Omnicom itself is slightly underpriced in my view, offering a free cash flow yield of 12% at current levels and it is actually increasing sales (unlike IPG where sales are flat or declining). So Omnicom is committing a more common risk than an inadequate acquisition price. In general, a company that is underpriced should be repurchasing its stock, a company that is fairly priced should be paying dividends, and a company that is overpriced should be using its stock to acquire cheaper companies. Omnicom is doing it backwards.

Interpublic Group is an advertising agency and PR firm operating in traditional and digital media, as well as related services. As of now it has a market cap of $9.1 billion. It has a cash balance of $1.5 billion but its current assets exceed current liabilities by only $600 million. For the last full year, sales were $10.7 billion and operating income was $1.2 billion. There was a goodwill impairment of $232 billion and excess depreciation of $117 million, resulting in operating cash flow of $1.55 billion. Net interest expense was $80 million (as the company maintains a large cash balance there is substantial interest income), resulting in free cash flow of $1.25 billion after estimated taxes. In 2023 sales were $10.9 billion and free cash flow was $1.17 billion, and in 2022 sales were $10.9 billion and free cash flow was $1.08 billion.

Year to date, sales declined by 3.5% year over year, most significantly in the UK and Asia. I can attribute part of that to the impending merger interrupting client acquisition. At any rate, operating cash flow (net of merger-related restructuring charges) comes to $441 million as compared to $416 million last year. Impressively, expenses declined faster than revenues so margins improved. So, based on a 10% return on investment and flat growth, it would appear that IPG is underpriced or on the low end of fairly priced.

Meanwhile, over at Omnicom, there is a substantial cash balance but no excess cash. For 2024 sales were $15.7 billion and operating income was $2.27 billion. Excess depreciation of $101 million and net interest expense of $147 million produces free cash flow of $1.76 billion. For 2023 sales were $14.7 billion, operating income of $2.10 billion and free cash flow of $1.74 billion, and for 2022 sales were $14.3 billion, operating income was $2.08 billion, and free cash flow was $1.65 billion.

This year to date, sales increased from $7.7 billion from $7.5 billion, operating income was $892 million compared to $989 million, plus excess depreciation of $46 million as compared to $58 million, interest expense of $70 million versus $68 million, leaving free cash flow of $686 million versus $773 million. This includes nearly $100 million in acquisition-related costs. At any rate, based on the 2024 results Omnicom has a free cash flow yield on the order of 12.5%, suggesting that the company is somewhat underpriced.

Furthermore, Omnicom also announced that it is expecting $750 million in synergies, mostly in the form of cost savings, and studies (https://knowledge.wharton.upenn.edu/article/how-to-get-mergers-and-acquisitions-right/) show that cost savings are more reliable than other types of synergies and are realized at an average rate of 60-90%. It does seem surprising that a merger (net of severance costs some of which are yet to be realized) could produce billions of value simply by consolidating some redundant departments, but even if a fraction of the synergies are realized this would enhance the value of an already underpriced company.

So, even though this merger is not the ideal outcome from a value investing perspective, either Interpublic Group or Omnicom would seem to be an enticing candidate for portfolio inclusion.

It is always difficult being a minority shareholder. Well, obviously most people reading this are not going to be the majority shareholder of a public company (and if you are, I am available for consultation at reasonable rates).

But I mean a minority shareholder of a company that has a such majority holder or group that has the power to control the entire company. The defects in such an arrangement are two-fold, one being the fact that the officers and management cannot be replaced for incompetence. For this reason, publicly traded companies with an entrenched majority often trade at a discount that varies according to the market’s perception of management quality.

The other disadvantage in this arrangement is the possibility that the majority holders will enrich themselves at the expense of the outside shareholders. Although Delaware law offers some protection for shareholders against this possibility, this requires costly litigation and in the absence of a clear-cut case there may be some reluctance to pursue this outcome.

Which brings us to GoodRx, a company with two classes of shares, one publicly traded and one not so much. The company is in the business of discount prescription drugs, operating in the shadows of the health insurance industry to provide an alternative nexus between pharmaceutical companies and the public. The company came to my attention during a routine flipping through the Value Line, but the figures in most finance data aggregators doesn’t account for GoodRx’s two classes of shares. Yahoo! Finance, for example, records a market cap of $1.74 billion as of this writing, which corresponds to 361 million outstanding shares. However, the class A shares, which are publicly traded, number only 104 million, and the remaining 257 million are class B shares, which are not publicly traded and have 10 votes instead of one for the class A shares, so the actual float of the company is only $500 million. The class B shares can be converted into class A shares at will, but the holders have only done so in order to engage in share repurchases, for which use the company has expended a considerable amount of the company’s cash.

At least the company has not concealed this intent; GoodRx’s latest 10-K states: “There have been no material changes in the expected use of the net proceeds from our IPO as described in our
Registration Statement. As of December 31, 2024, we estimated we had used approximately $586.4 million of the net proceeds from our IPO: (i) $164.4 million for the acquisition of businesses that complement our business; (ii) $262.0 million for the repurchases of our Class A common stock; and (iii) $160.0 million for the repayment of our outstanding debt obligations. As of December 31, 2024, we had $300.5 million estimated remaining net proceeds from our IPO…” and indeed the balance sheet shows $301 million in cash.

I should point out that the registration statement’s “Use of proceeds” section doesn’t mention share repurchases; it mentions acquisition of businesses and general corporate purposes and only a desultory mention that “We may find it necessary or advisable to use the net proceeds for other purposes….” such as, say, providing an exit strategy for a cabal of private equity firms when the share price has declined dramatically since the IPO. (It took me a while to write this article because I was trying to strike the right balance between expressing my opinion and avoiding libel suits).

Actually, what tripped my sensors for GoodRx was its high operating cash flow, which only arises because of the massive amounts of share-based compensation; in fact, the share-based compensation is comparable in magnitude to the quantity of share repurchases in a given year. There are two possible explanations for this, neither of them particularly encouraging. One, the company is finding ways to justify its shareholder repurchases with massive stock grants for some nefarious purpose I’m not quite clear on. Or two, the company is unprofitable enough that it has to pay its employees in stock rather than cash. Normally at this point I would compare GoodRx with its competitors to find an answer, but it seems that they don’t have any publicly traded competitors.

Either way, maybe in 2028 all the class B shares will convert into class A shares so the group may be getting out while the getting’s good, although at the current rate of repurchases the company will have burned through its remaining excess cash by then.

Looking at the figures, operating margins have improved since 2022, as have sales. For 2024 sales were $792 million, operating income was $66 million, net of $99 million in share-based compensation. Depreciation and amortization equaled capitalized software, so by my usual calculation operating cash flow remained at $66 million, and interest expense for the year was a whopping $53 million. The company also finished the year with $450 million in cash on which it earned $23 million in interest, so the interest coverage situation is not desperate, although it seems rather dire for my taste. This leaves $36 million before estimated taxes, or $29 million after taxes. For a float of $500 million this is not unreasonable, although too low for my taste, but let us recall that the company founders, not satisfied with the $150 million in share repurchases they have received this year, are also entitled to about 72% of this money based on their ownership percentages.

In the first quarter of 2025 the company had operating earnings of $23.4 million as compared to $7.4 million the previous year, stock-based compensation expense of $19.2 million as compared to $25 million, net interest expense of $6.7 million as compared to $7.1 million, producing earnings of $12.5 million before estimated taxes, or $10 million after, as compared to $300 thousand and $225 thousand. There were also stock repurchases of $100 million and $153 million last year; I suppose the company is in the habit of doing its repurchases early in the year.

The company’s remaining cash balance, as I said, was $300 million, and normally, as with BorgWarner I would consider it as non-operating and available to the shareholders, but based on GoodRx’s history I think that money is available primarily to the class B controlling shareholders, not the poor outside investors with the publicly traded shares with their claim on only 28% of the company’s earnings.

Obviously I don’t recommend GoodRx; the free cash flow yield is somewhat anemic even setting aside the two tiers of stock, and without setting that aside the yield is certainly not fair compensation for the minority status. It’s more of an object lesson in the dangers of an entrenched majority owner and the foibles of stock screening.

…but nobody to know what it is. This line was originally chosen to describe a company that was floated alongside the South Sea Bubble in England. And although unfortunately this company seems to be fictitious, there are certainly businesses who could stand to be a little more helpful in describing what products or services they are selling.

And so it is with OpenText, which does the following:

OpenText is an Information Management company that provides software and services that empower digital businesses of all sizes to become more intelligent, connected, secure and responsible….With critical tools and services for connecting and classifying data, OpenText accelerates customers’ ability to deploy Artificial Intelligence (AI), automate work, and strengthen productivity…Our products are fundamentally integrated into the operations and existing software systems of our customers’ businesses, so customers can securely manage the complexity of information flow end-to-end. Through automation and AI, we connect, synthesize and deliver information when and where needed to drive new efficiencies, experiences and insights. We make information more valuable by connecting it to digital business processes, enriching it with insights, protecting and securing it throughout its entire lifecycle and leveraging it to create engaging digital experiences….Our solutions also enable organizations and consumers to secure their information so that they can collaborate with confidence, stay ahead of the regulatory technology curve and identify threats across their endpoints and networks…

And this is in their 10-K, so all of the above is true under penalty of perjury.

The trouble is that if we don’t quite know what a company does, we don’t know if they’re any good at it, or if their competitors might be better at it. Warren Buffett famously claims that he never invests in businesses that he doesn’t understand, which includes high technology. And normally software companies have anemic free cash flows anyway, so a value investor would have no interest in them. My theory is that because writing code is considered an operating expense rather than a capital expenditure (with the exception of a product that is somewhere between commercially viable and fully developed (but not post-development updating or maintenance (which tends to be important for business clients))), the company shows a high return on invested capital (unless there has been an acquisition in which case there is a lot of goodwill on the balance sheet), which attracts a lot of algorithm-minded portfolio managers. Furthermore, as a software company can essentially replicate its product indefinitely at no cost, there is a lottery ticket effect for a company that comes up with the next “disruptive” or “killer” app, as the kids say. Even so, for a larger company in a saturated space such as Microsoft or Alphabet, the odds of developing a product big enough to move the bottom line are quite small.

At any rate, OpenText provides exciting cloud services to businesses, including content management, cybersecurity, automation of applications, and IT operations, and derives its income from cloud services and support contracts. The company is a serial acquirer in the field, having most recently purchased Micro Focus in 2023, a company that provides “technology and services that help customers accelerate digital transformations.” This acquisition was for $5.6 billion, mostly funded by debt, while the company’s current market cap is a mere $7.2 billion at this moment. However, last year OpenText resold a large portion of Micro Focus back to a private equity firm for $2.2 billion and used the proceeds to pay down some of the acquisition debt. So far the acquisition seems to be working out well in terms of free cash flow, which is a rare accomplishment for a merger between two evenly-sized companies, per Damodaran’s Investment Fables.

Turning to the figures, in fiscal year 2024 sales were 5.77 billion, operating earnings were 887 million, plus 808 in depreciation and amortization, minus investing activities of 159, producing operating cash flow of 1536 million. Interest expense was 536 million, leaving $1 billion, or 740 million after estimated taxes, making a reasonable 10% yield of free cash flow. Their fiscal year ends in June for presumably Canadian reasons.

This calculation does not include the $429 million gain on the sale of part of Micro Focus, which OpenText does not consider to be operating even though mergers and acquisitions are part of the company’s core strategy so it may as well be. This stands in positive contrast to Helen of Troy, which considered the sale and leaseback of its premises to be operating even though they’re in the cosmetics and household products business, not the real estate business. At any rate, the gain on sale raises accounting questions; it is surprising that OpenText would be able to turn a 19% profit on assets it held for less than a year, but considering the substantial depreciation and amortization charges that the company takes, it is possible that many of these gains are simply reversals of depreciation charges. Obviously it is in the interest of OpenText to load the divested unit up with as many high priced assets as feasible, to relieve itself of both capital gains and ongoing depreciation charges, but there are limits even to aggressive accounting and I suppose the company should be congratulated for not courting shenanigans.

This figure also does not include the income from the $1.28 billion in cash on OpenText’s balance sheet, which came to $49 million before taxes for the year. Now, on paper, OpenText’s current liabilities exceed its current assets by nearly $400 million, and so not only is there no excess cash available for distribution to the shareholders (which was the happy situation for BorgWarner that many simple-minded editors could not understand), but in theory the company is also going to have to come up with $400 million within the year, so the value of the equity should be reduced by that amount. However, $1.5 billion of OpenText’s current liabilities consist of “deferred revenue.” This category of liability denotes the fact that OpenText’s customers have paid the company up front for services that are to be performed over time, and given the company’s profit margins, even under the worst case scenario of having all of the costs pertaining to this deferred revenue as having yet to be incurred, there should be no additional pressure to come up with more money based on deferred revenues. At any rate, OpenText is in the habit of keeping about $1.2 billion in cash in reserve, so even if it should not be considered available for distribution the income from it should not be ignored.

In 2023 sales were 4.48 billion, operating earnings were 516 million, depreciation was 657 million, capital expenditures 124 million, resulting in 1049. Interest was 364, leaving 685 or $507 after taxes. I should also note that between this fiscal year and 2024 there were charges of nearly $150 million to integrate Micro Focus with OpenText that are in theory nonrecurring but OpenText’s strategy involves frequent acquisitions, though most are not this large.

In 2022 sales were 3.49 billion, operating earnings were 645 million, depreciation was 504 million, capital expenditures came to 93 million, producing 1056. Interest was 151, leaving 905 or 670. There was another acquisition funded by debt in this year.

Year to date 2025, sales are $3.86 billion as compared to $4.41 billion for the first three quarters of 2024, and operating income is $711 million versus $694 million, as expenses are declining faster than sales (which is another promising sign that the Micro Focus acquisition and partial disposition was a good move). Amortization of 480 million and expenditures of 109 produce $1082 in operating cash flow versus $1205 for 2024. However, interest expense was $267 million versus $425 million, leaving $815 versus $780 in pretax free cash flow, or $603 million versus $577 million in after-tax free cash flow, which would be $804 million on a full-year basis.

Interest expense is better covered than before, although the company is still rated BB+ and the outlook is only stable. Unfortunately the gulf between a BB+ bond, which is subinvestment grade, and a BBB- bond, which is investment grade, is roughly as wide as that between Dives and Lazarus, so a credit rating upgrade would be a truly momentous event.

Furthermore the company has made non-trivial amounts of share repurchases this year, so conceivably management sees OpenText as both undervalued and in a state mature enough that its cash needs are predictable enough to allow for repurchases.

So, can we recommend OpenText? Much of the revenue is recurring and the cost of switching IT providers is high enough that companies would not do it lightly. And it is unusual to find a technology company this cheap, although the Software as a Service model is theoretically more difficult to scale than the pure software company model of a company releasing its product into the wild and hoping it finds a habitat and reproduces wildly (if such a model ever existed). Also, the share repurchases suggest the possibility that the company is stabilizing and so even with some opacity as to the company’s operations and competitive niche, OpenText is worthy of consideration for one’s portfolio.

I’m sure someone other than me has observed that a value investment is a value investment that worked out and a value trap is a value investment that didn’t work out. At any rate, it seems to me that it takes more than a high earnings yield to make a value stock. A low P/E can be justified if the company is risky or in terminal decline, which are not inherently disqualifying for a stock, but when combined with a lack of competitive moat a low multiple is well-deserved and is unlikely to signal an opportunity.

Which brings us to Helen of Troy (HELE). This knife has been falling since 2022 when it had a share price of about $245; it consolidated at around $97 between October of 2022 and June of 2023, went up a bit and then back down and then fell off another cliff in June of 2024 to about $55 or $60, drifted sideways for a bit, and then collapsed again because of the tariffs to $24 and change as of this writing, and has not shared in the post-tariff rally. The question becomes at what point, if any, we can declare this share price decline to be overdone.

Actually, that’s the wrong question. First of all, historical stock prices are widely believed to have no predictive power, and obviously the people who thought the company was worth $245 in 2022 are no longer around and willing to act on their belief by bidding the company back up. Second, a company is fairly or incorrectly priced based on its earnings power, not by the path it took to arrive at that price. And third it is entirely possible for the fair price to be zero, particularly in the presence of debt. It may or may be not be a bit harsh to apply this criterion to HELE, as we shall see.

Helen of Troy is a supplier of kitchenware, insulated beveragewear and coolers, and outdoor carrier packs; and also cosmetics, home health care products, and also heaters, fans, air purifiers, etc. to both online and physical retailers. This somewhat chaotic mix of household goods is the result of several acquisitions over the years and whatever competitive advantage Helen of Troy brings from its distribution network is questionable, as sales have been declining in recent years. The company has been focusing on rationalizing its brand portfolio and has incurred significant restructuring charges in recent years, which one is often tempted to treat as nonrecurring but for a serial acquirer undergoing business difficulties I anticipate that Helen of Troy will be in a constant state of restructuring for the foreseeable future.

Helen of Troy’s product lines are largely commoditized, meaning that their products tend to compete on price more than product differentiation. The cosmetics I’m not so worried about for even with recession fears in the market, make-up is about the last thing women will cut from their budget, and generally they are considered consumer staples rather than discretionary. On the other hand, even if people will continue to buy cosmetics, it is not a given that they will continue to buy them from Helen of Troy. Sales in the US account for 3/4 of total sales, and as 63% of their products came from China in fiscal year 2025 (fiscal year ends in February), the tariff concern is not unjustifiable.

Turning now to the figures, just about 47% of Helen of Troy’s sales in 2025 came from Home & Outdoor and 53% from Beauty and Wellness. Sales in the former declined by 1.1% as compared to 2024, and in the latter, 10%, not counting the effect of a new acquisition.

I am often in the practice of counting acquisitions as growth capital rather than maintenance capital for purposes of computing free cash flow, but I think this approach is more justifiable for industrial companies and other capital-intensive businesses where an acquisition can add a new product line or supplier and at any rate it represents its own source of sales and earnings. But in HELE’s case, acquisitions of brand names are part of the company’s innate strategy and generally results in acquiring intangibles, so I think acquisitions might well be charged against free cash flow, particularly as they have resulted in occasional goodwill writeoffs. At any rate, given the size of Helen of Troy’s recent acquisitions relative to its current market cap, the acquisitions should at least be noted.

The figures are further complicated by a gain on the sale and leaseback of certain premises in 2024 which the company credited to cost of goods sold for some reason, and the expense of complying with an EPA ruling in 2023, both of which I have adjusted out. Counting the gains from the sale of property as operating income borders on an accounting shenanigan to me, as the company is not in the business of buying and selling real estate and this gain is not part of the company’s recurrent business activity.

In 2025: Sales $1908 million, gross margin $914 million, operating income (not counting impairments) $188 million, depreciation and amortization $55 million and capital expenditures $30 million, producing operating cash flow of $213 million. Interest expense $52 million, leaving $161 million or $128 million after 20% estimated taxes. There was also a $229 million acquisition, so if that is counted free cash flow was actually negative in fiscal year 2025.

In 2024 sales were $2005 million, gross margin $949 million, operating income (adjusted for real estate gains) was $224 million. Depreciation and amortization was $51 million and capital expenditure was $37 million, leaving $238 million in operating cash flow. Interest expense was $53 million, leaving $183 million, or $146 million after estimated taxes. There were no acquisitions this year.

In 2023 sales were $2073 million, gross margin was $899 million, operating income (adjusted for EPA compliance costs) were $235 million. Depreciation and amortization were $45 million, and capital expenditures were $175 million, leaving $105 million. Interest expense was $41 million, leaving $64 million, or $49 million after estimated taxes. There was also $146 million in acquisitions that year.

In 2022 sales were $2223 million, gross margin was $953 million, operating income was $273 million. Depreciation and amortization was $36 million and capital expenditures were $78 million, producing operating cash flow of $231 million. Interest expense was $13 million, leaving $218 million, or $176 million after estimated taxes. There was also $411 million in acquisitions that year.

The alert reader will note that the acquisitions were not covered by Helen of Troy’s operating cash flows, which were funded by borrowings. And yet despite these acquisitions, sales and earnings have been declining. I should note in 2025 the company repurchased $103.2 million in shares at an average price of $99.34, and in 2024, $55.2 million at an average price of $127.67. Helen of Troy also repurchased $188 million in shares in 2022 at an average price of $220 per share and $203 million in 2021 at an average price of $197. Even allowing that the company can be forgiven for not anticipating the collapse in share prices, three of the repurchases were made with borrowed money. Some companies do recapitalize by swapping debt for equity, but those are generally maturing companies moving from their growth to their mature phase, not ones in such a precarious position as Helen of Troy, which has declining sales, $560 million in market cap, and $900 million in long term debt.

So, returning to our original question of whether the collapsing stock price of Helen of Troy is overdone, $146 million in free cash flow for 2025 gives us free cash flow yield of 26%. If we demanded a 10% return on investment, and we could be assured that earnings would decline at no more than 16% a year in perpetuity, we could in theory find a satisfactory purchase. However, as we have said, Helen of Troy imports 63% of its products from China and even before the tariff announcement the company was seeing weakening demand for its products, which as I have said are commoditized and therefore compete largely on price even though HELE’s strategy seems to involving acquiring brands. Therefore, I find it difficult to predict if sales will decline at 10% or 16% or 26% or more (or less, admittedly), and I would not care to take the risk of guessing wrong.

I could take the backup approach and value the company largely on its balance sheet, but owing the company’s acquisitive nature, most of its long term assets consist of goodwill and intangible assets. Tangible assets total $1.382 billion, while the company’s total liabilities come to $1.449 billion. So, when I asked how much further the share price has to fall, all the way to zero is not the trivial answer it first appeared. Furthermore, even if the company is coming up with $146 million in free cash flows every year, a prudent company would shoring up its balance sheet by paying down debt (and an imprudent company is an unattractive proposition in itself), so it could be years before that cash flow would be genuinely available to the shareholders.

Wherefore I can say that even if catching a falling knife were a desirable strategy in general, I think catching this one strikes me as an unattractive proposition, despite the considerable free cash flow yield, and so I cannot recommend it as a candidate for portfolio inclusion.

I had a surprisingly bad experience at a certain investing website that accepts user submissions and I thought I would speak of it. Last week I spoke highly of BorgWarner and thought a wider audience might benefit from my views, so I submitted my last post to them.

I proposed that the equity was worth $4.8 billion, plus $2.1 billion in the company’s excess cash, plus about $400 million in the present value of the company’s below market interest rate, total $7.3 billion. My editors, including two managing editors, though, argued that if I wanted to add the company’s excess cash to the equity, I would have to remove the company’s long term debt from it as well. Naturally, I found this quite confusing because I thought I already deducted it from the assets to arrive at the equity in the first place, and that under their rules the company was worth less money with the excess cash than without it, but they weren’t having it.

Now, the formula for Enterprise Value is EV = Equity + debt – cash. This equation is so ubiquitous that I can’t really think of a citation for it, so let’s just say The CFA Curriculum (any level), vol. 2. The equation can easily be rewritten as Equity = EV – debt + cash. Now, the equation that the editors are using, equity = EV – debt + cash – debt, I am quite unable to find a citation for.

I think it was the “- cash” part that was throwing them off, because I am a firm believer in the concept of excess cash, of which BorgWarner has $2.1 billion based on the usual formula. And for the sake of avoiding double counting, I scrupulously removed any of the interest on the excess cash, so that $4.8 billion is what the equity of BorgWarner would be worth if the company had no cash at all.

We may check the math as follows: Enterprise value is also the free cash flow to firm discounted at the weighted average cost of capital. The weighted average cost of capital is the long term debt weight of 5.7% * 3.7 billion, plus 12.5% * 4.8 billion/8.5 billion, or just about 9.5%, and $811 million per year capitalized at 9.5% is $8.5 billion again (unsurprising since I’m just reversing the weighted average cost of capital formula). The value of the equity is based on a discount of the free cash flow to the equity at 12.5% and matches the one computed from the enterprise value equation (as it should since I supplied the numbers myself), and now the cash must be dealt with somehow. In fact, the enterprise value equation does not distinguish between excess and non-excess cash so the raw equation is even more generous to my point of view.

And here, again, is where my editors insist for whatever reason that I subtract the debt again, which, in addition to being based on the equation conjured from their own fevered brains, also messes up the enterprise value based on weighted average cost of capital as it essentially assumes that the company will pay off its long term debt and yet continue to pay interest on that debt for some reason.

In fairness, a company cannot operate with no cash at all, and even though BorgWarner is cash flow positive and has an untapped $2 billion line of credit it may find it useful to keep some cash on hand. A rule of thumb is 2% of annual sales, which in BorgWarner’s case works out to $280 million, but the company could give up at least the $1.8 billion in cash left over. And obviously a company can have excess debt, current or long-term, and any cash used to pay that down would be unavailable to the shareholders, but BorgWarner is not such a company and therefore there is no reason to assume that its long-term debt will be paid off at any point in the foreseeable future.

I was trying to explain my reasoning, although perhaps not as well as I might because I was still stunned into inarticulacy that they were using the special EV = equity + 2*debt – cash formula that their editorial policy has apparently forced on them, when they declared their decision was final. Suffice to say, I will not be renewing my trial membership.

My faithful readers will recall that I thought Garrett Motion, maker of turbochargers and related auto parts, was an attractive investment, and indeed it was until the tariff monster was awakened. And while I was looking at it, I compared it to BorgWarner, another company in the auto parts sector, which had comparable R & D expenditure rates and a strategy of technological development. More to the point, it presents an even more attractive valuation based on its free cash flow yield.

BorgWarner (BWA) is a globally positioned producer of engine and drivetrain components, and also invests significant R & D expenditures in order to remain a technology leader in its space. The company has maintained a massive free cash flow yield over the last few years and of its $5.8 billion market cap as of this writing, $2.1 billion consists of cash, nearly all of which is “excess,” or not necessary for the company to carry on its business, and the earnings yield on its operating equity is highly enticing.

BorgWarner has recently refocused its strategy away from a largely electric-vehicle focused approach in favor of a more balanced use of its entire portfolio of offerings. As a result, its strong cash flows in the last few years have accumulated on its balance sheet and, in my view, this cash will be deployed most efficaciously into substantial share repurchases, to the benefit of the share price.

Company Overview & Strategic Position

BorgWarner operates in several fields in the automotive parts sector: turbos & thermal technologies are about 40% of sales, drivetrain devices a slightly lower proportion, powerdrive, including all-electric car technologies, about 15% of sales, and battery & charging systems, the last 5%. Total net R & D in 2024, according to the latest annual report, came to about 700 million or 5-6% of sales which is comparable to Garrett Motion, which I have written about before and still like) one of its competitors in the turbocharger space; however, more than half of that R & D was allocated to powerdrive and batteries despite their lower presence in the sales mix.

BorgWarner’s strategy starting in 2021 was to go all-in on electric vehicles, and the company even spun off its fuel systems division in 2023. However, in 2024 the company determined that adoption of electric vehicles was “volatile” compared to their expectations, and indeed the operating income from both the powertrain and battery segments were negative in 2023 and 2024. As a result, the company refocused its strategic efforts towards growth along its entire portfolio of offerings, including turbochargers, transmissions, etc. alongside developing its electric offerings. For this reason, I anticipate that there is scope for reduction in both R & D and capital expenditures (including acquisitions), resulting in further incremental improvement in free cash flows.

Valuation

Methodology

The auto parts industry is cyclical, which makes calculating a company’s prospective earnings power a complicated process, so I will explain my method for doing it:

Starting with the figures in the latest annual report, in the last year BorgWarner reported earnings of $338 million, but this was net of a $646 million goodwill writeoff which resulted from the company’s disappointed expectations in various acquisitions pertaining to its aggressive electric vehicle strategy.

Stepping back from that, BorgWarner’s operating income without the writeoff was $1192 million, and the company has $3.7 billion in debt outstanding with an average interest rate of 2.8%, producing interest expense of $105 million. However, this interest charge reflects that BorgWarner’s debts were issued at interest rates that are lower than current rates (for example, they have borrowed 1 billion euros at a rate of 1% until 2031). To focus on the company’s prospective as opposed to historical earnings power, I should adjust their pro forma interest expense to reflect the current rate of 5.7% for BBB+ rated bonds, which is BorgWarner’s credit rating. I should note that the company’s latest borrowing in August of 2024 was at an average rate of 5.2%. The pro forma interest expense comes to just about $211 million per year, leaving just under $1 billion in estimated pretax earnings. As the company has a global footprint, estimating its tax rate is difficult but the company’s provision is about 23% on average, leaving just about $755 million in after tax free cash flow, which is an impressive free cash flow yield of 20.7% of its effective market cap. I will point out that the above free cash flow figure does not include any income from BorgWarner’s enormous cash balance, as I consider that income to be non-operating.

Of course, for a cyclical company like an auto parts manufacturer, one year’s results are not a reliable measure of earnings power; it could be that 2024 was a particularly good year. One should consider Borg Warner’s earnings power over a complete business cycle, and applying a 5.7% interest rate, pro forma free cash flow figures for those years (taking data from previous 10-K filings) were, starting in 2023, were 2023: 539; 2022: 543; 2021: 743, and 2020: 397 (and 891 in 2019 but that properly belongs to the previous business cycle). The average figure over the length of a business cycle was $595 million per year, or a yield of 16.3%. BorgWarner’s long term debt has been stable since 2020, but interest rates were substantially lower before this year so actual free cash flows were higher. But again, as we are looking at prospective earnings power we should probably apply the present higher interest rates. But even in the pandemic year of 2020 the company managed a free cash flow yield of over 10% based on the current effective market cap.

I spoke earlier of the goodwill writeoff that BorgWarner took in 2024. As I stated before, the company’s aggressive pursuit of expansion in electric vehicle products included a number of acquisitions, and the above calculations do not count them against the company’s cash flows. However, in my opinion, although the acquisitions were regrettable with the benefit of hindsight, I anticipate that BorgWarner’s management will going to be more circumspect about purchasing growth in future. Therefore it would be somewhat unfair to ding the company’s future earnings prospects based on its past mistakes, especially as the pace of acquisitions slowed considerably in 2022 and 2023 and ceased completely in 2024 even as large amounts of cash have built up on the balance sheet.

Speaking of the cash balance, I would describe nearly all of the $2.1 billion in cash on BorgWarner’s balance sheet as “excess” cash. Excess cash is cash that a company holds that is not needed for the company’s operations and could be distributed to shareholders without affecting the company’s cash needs. The mode of calculation is as follows: excess cash is total cash minus current liabilities plus noncash current assets (or zero, whichever is greater). In this case, total cash is $2.1 billion, current liabilities total $3.6 billion, and noncash current assets total $4.4 billion, meaning that essentially all of BorgWarner’s cash is available to distribute to shareholders. This is hardly surprising for a reasonably mature cash flow-positive business like BorgWarner, particularly as the company has an unused $2 billion credit facility to address liquidity needs. And to the best of my knowledge none of BorgWarner’s creditors have imposed any legal restrictions on dividends or share repurchases (yes, I read the bond indentures). And as I stated above, none of the income from the company’s cash holdings was added to the free cash flow to firm/equity in order to avoid double counting.

Price Target

So, putting it all together, we have $600 million in average annual earnings over the last business cycle. Applying a conservative multiple of 8 times gives us a market cap of $4.8 billion. Add to that the approximately $1.8 billion in excess cash and $375 million representing the present value of the company’s below-market interest rates on its long term debt, produces a target market cap of about $7.2 billion, or a share price of $33 on the low end, which compares favorably to the share price as of this writing of $26.45.

Potential Risks

Obviously the most visible risk is the uncertain tariff situation. However, as I stated before BorgWarner has a global footprint, with only 25 of its 84 properties located in North America. Moreover, the United States represents only 16% of BorgWarner’s net sales, and indeed North America represents about 16% of global auto sales in the first place. And although 20% of BorgWarner’s property, plant and equipment is located in China, where the trade war is presently happening, again not all of those exports are directed to the United States so hopefully the present tariffs regime may not affect more than a single digit percentage of BorgWarner’s business. Of course, some of BorgWarner’s non-US customers could later seek to export their cars to the United States and get caught in the tariff net, but the effects of that are unpredictable.

But for what it’s worth, BorgWarner’s stock price has tracked the broader indexes pretty closely since the tariffs were initially announced so at least the market doesn’t seem to believe the company is more exposed that any other American company.

Beyond tariffs there is the possibility that the company could go on another ill-considered acquisition spree, even though recent experience may have scared the management team away from that course. Another possibility is that adoption of all-electric vehicles may in fact occur faster than BorgWarner has been observing, which would diminish the value of the company’s existing portfolio of products. But in my view the transition to an all-electric transportation fleet will take decades if it occurs at all, and meanwhile plug-in hybrids, which use many of BorgWarner’s existing suite of offerings, will be with us for a considerable length of time.

Conclusion

So, I would argue that BorgWarner’s prospective earnings power as measured by free cash flow yield is attractively high and the company is undervalued at the current price. Furthermore, the company is no longer disdaining its non-electric-vehicle portfolio, and there is room to save some research and development and capital expenditures on the electric vehicle product lines. Also, the company has recently accelerated share repurchases ($402 million in 2024 alone) and has the resources to apply billions more, which is always a good use of cash for an undervalued company. Therefore, I can strongly recommend BorgWarner as a candidate for portfolio inclusion.

Disclosure: As of this writing, long BWA and GTX

Insnurance companies

According to some guy from Columbia…

In my final article on this series of whether companies in certain sectors can be bought based on price to book ratio, I examine insurance companies. Recall that my original thesis is that, according to some well known financier whose name I forget, banks have assets that are impossible to value, and insurance companies have both assets and liabilities that are impossible to value. Even so, it is conceivable that any errors in valuation could be in the owner’s favor as easily as to their detriment. More to the point, the assets of these companies tend to consist largely of financial investments rather than, say, industrial equipment, so their book value tends to be a more reliable guide to their actual value and often times may be quoted regularly in the market. And so the naive strategy of evaluating these companies solely on their book value might be effective.

As we have seen, the correlation coefficient between a bank’s price/book ratio and its performance in 2024 was practically nonexistent, but when using tangible book value there was slight evidence of low price to book ratios producing outperformance when outliers are eliminated but at least in 2024 higher than average price to book ratios were not punished and even showed a tendency to be rewarded, which came as a surprise.

Turning now to insurance companies, the first complication is which book value to use. Insurance companies, particularly life insurance companies, make actuarial assumptions about life expectancy, interest rates, expected return on investments, and so forth. These assumptions are subject to change, of course, and as they change the value of the company’s liabilities change with them, sometimes dramatically, even though the company’s assets and near term cash flows are unchanged and a future shift in actuarial assumptions could erase the effect of the current ones instantly. The effects of these changes in actuarial assumptions are subsumed in “Accumulated Other Comprehensive Income” on the balance sheet, as these changes are not considered operating, since an insurance company is in the business of selling insurance and investing assets, not actuarial services.

However, insurance companies are in the habit of reporting their book value under GAAP and their book value not counting this accumulated other comprehensive income. By an extraordinary coincidence, of the fifty or so insurance companies that formed my sample, every single one of them reported a higher book value with this adjustment. Apparently, not a single insurance company has ever had its actuarial assumptions move in its favor in the history of the American insurance business. (Or, more likely, if they have the company sneakily counted it as operating income to make themselves look better). At any rate, in my opinion accumulated other comprehensive income, or more accurately, accumulated other comprehensive losses, should be included in book value, as it represents an economic loss, even if unrealized as yet, and a constraint on the insurance company’s business activities. This is also the view of a noted professor at Columbia School of Business.

Also, some insurance companies have goodwill, and eliminating that to arrive at tangible book value is another possible adjustment .

That said, what are our results? The Value Line, from which I draw my samples, divides insurers into life insurers and property and casualty insurers, the latter being much more numerous. There are also reinsurers, but not many of them so it is difficult to draw conclusions from such a small set.

Among life insurers I was pleased to discover that in 2024 the naive strategy actually showed some signs of success, with a correlation coefficient of about 0.2 for both raw and tangible book value. Unsurprisingly, book value modified for accumulated other comprehensive income showed a weaker correlation. As I’ve said, given the numerous possible determinants of investment performance, one can hardly expect price to book ratio to be the sole factor in returns, but a correlation coefficient this high at least suggests that something is going on.

Among property and casualty insurers, I was disappointed to find a correlation coefficient of about 0.015, essentially consistent with no correlation. However, by dividing the outcomes into quadrants as before, I find that a higher than average price to book ratio was almost twice as likely to produce a lower than average return as a higher one, while a lower than average price to book ratio was more likely to produce a low return. Also, apparently 2024 was a difficult year for the average insurer, as returns were skewed to the downside.

So, to conclude, at least for 2024 the naive strategy of buying banks or insurance companies based on price/book ratio did not show reliable correlations apart from among life insurers. However, when purely considering if the price/book ratio was above average there were some minor indications of the validity of the strategy; with property and casualty insurers high price/book ratios were more likely to underperform, while with Midwestern banks, high price/book ratio banks showed above average returns but curiously that advantage disappeared when looking at tangible book values, to be replaced by a slight advantage for low price/book ratios, a sign that goodwill writeoffs are not overdue at least. And for non-Midwestern banks low price/book ratios had a slight advantage once outliers are removed, but curiously the high price/book outliers were more likely to outperform, which is counterintuitive and requiring further investigation. And of course this study should be extended to prior years.

As a followup to my previous article where we discovered that at least in 2024 the naive strategy of buying banks based solely on price to book ratio did not perform as advertised. However, the Value Line from which I drew my sample also treats Midwestern banks as a separate category of bank, for reasons knowable only to them. True, Midwestern banks are more likely to be small or medium sized and regional, but in my previous article I focused on similar non-Midwestern banks, as large investment or money center banks likely have a distinct competitive niche that makes them less susceptible to being examined purely on their balance sheets. Also, I found some repeat entries from last week in this set of banks, meaning that depending on when one looks a bank can simultaneously be Midwestern and not Midwestern.

For the non-Midwestern banks, recall, neither regression analysis nor the simple sorting into quadrants of above and below average indicated that this strategy showed any chances of outperforming. Nor were results noticeably different when using tangible book value, which at least implies that most banks are not sitting on overdue writeoffs of goodwill, which I suppose is promising.

But now focusing on Midwestern banks, using raw book value the naive strategy fared no better in 2024, with an r-squared value of 1.72%, which is entirely consistent with the relationship between price/book value and returns being nonexistent. And removing outliers resulted in no improvement, as three of the banks with a price/book ratio above 1.8 performed better than the average for all banks under consideration and two such banks underperformed.

Looking at the quadrant approach, we see that in fact twice as many banks with an above average price to book ratio earned an above average return. So much for price/book ratio mattering! Low price to book was divided evenly.

However, when we switch to tangible book values, we see that the low price to book ratio strategy has some stirrings of validity. The r squared value was 2.75%, which is still quite small but considering how many factors actually go into bank returns, ferreting out the effect of a single causative factor is not easy and 2.75% is certainly further from zero than we’ve seen earlier.

And looking at the quadrant we see dramatic results indeed. The apparent “advantage” in high price to book banks more or less disappears, and is replaced by an apparent advantage in low price to book banks. So this suggests, at least subtly, that for Midwestern banks some goodwill writeoffs are indeed overdue and that the effort banks go through to report tangible as well as raw book value is not wasted, at least in the Midwest.

Obviously, more years and more potential for confounding variables should be examined, but it is encouraging to see that our naive price to book strategy is not entirely off base. And we haven’t even started on insurance companies yet.

A wise financier whose name escapes me once wrote that the trouble with banks is that they have assets that are impossible to value, and the trouble with insurance companies is that they have assets and liabilities that are impossible to value. Given the current situation with insurance companies and natural disasters, this is a difficult statement to gainsay, and likewise valuing mortgages and derivatives is difficult even during non-financial crisis times.

However, the possibility did occur to me that even if valuation of such companies is error prone, the error might in fact be nonbiased, which is to say that one has the same odds of guessing too high as too low. And so, as most valuation techniques of banks and insurance companies center on the balance sheet, could the naive strategy of assembling a portfolio based solely on buying companies with a low price/book ratio and avoiding the companies with a high price to book ratio be a valid strategy?

In order to investigate this, I computed the price/book ratio of all the banks in the Value Line investment survey (which hopefully is a representative sample) as of roughly February 2, 2024 and examined their one year performance. I decided to focus on the smaller community or regional banks, ignoring the massive money center or investment banks. And then, because banks are complicated by goodwill and other intangible assets, I did the same with tangible book value, which most but not all banks are kind enough to report in their earnings announcements. I then ran a linear regression analysis, as data mining analysts are known to do.

Before I reveal the results, I should try to discuss why banks and insurance companies are more transparent to balance sheet analysis than other companies. The reason is the assets of, say, an industrial company, are custom-designed to produce a particular product, and the value of those assets on the balance sheet may have nothing to do with their actual value in the economic sense, which is determined mainly by the demand for the company’s product and its competitive position. However, banks’ assets are generally marketable and fungible, as a bank’s “product” is simply the use of money, and the size of its portfolio of assets is generally a reasonable measure of its ability to do so. Insurance companies operate in an analogous way; they profit by taking in premiums and investing them, thus pocketing any investment income produced between when the premium is collected and when claims are paid.

So, how did my regression turn out? I have to confess, not at all convincingly. Using pure book ratio, the r squared turned out to be less than 1%, and eliminating outliers it actually went down. Tangible price to book ratio, outliers or not, fared no better. The r squared, recall, measures how much of the variation in the dependent variable (investment returns) is determined by the independent variable (price/book ratio). A value of 1% implies that hardly any relationship between the two exists.

However, the r squared is not the end-all of analysis, and much like the Spanish Inquisition, when you have data you want to torture it until it says something. So, I divided my 161 banks into four quadrants based on whether the price to book ratio was below the mean or above and whether the one year return was below the mean or above. Using the raw price to book ratio, 38 banks had a low price/book ratio and below average returns, 40 banks had a low price/book ratio and above average returns, 38 banks had a high price/book ratio and below average returns, and 45 banks had a high price/book ratio and above average returns, which more or less confirms what the regression told us, that at least in 2024 price/book ratio had essentially nothing to do with performance, and indeed the number of banks that broke the rule outnumbered the ones that followed it.

However, I did mention outliers before; there are times when the reported price/book ratio is so unrealistic as to be “obviously” wrong. I decided that any bank with a price/book ratio below 0.5 or above 1.8 was so far beyond the pale that it was unlikely that investors were seriously using it as a criterion. This eliminated three of the low price/book, low return banks, three of the high price/book, low return banks, fully 7 of the high price/book, high return banks, and none at all of the low price/book, high return banks. Curiously this has the salutary effect of making the proposed strategy look slightly better, but at the cost of eliminating a disproportionate number of high return banks, which obviously is exactly the opposite of what we want.

Applying the same method to the tangible book value, we find as follows:

Notice that there are more below average tangible price/book ratios. This is not surprising as a bank’s tangible assets can always be less than its raw book value but can never be greater. However, we still see that whether the price to book ratio is low or high, the odds of outperforming the average is the same as underperforming.

But, applying our search for outliers again we see that low price/tangible book value stocks are made to look slightly better.

Perhaps 2024 was an unusual year for banks, but it may be that investors who assign a high price to book ratio are actually seeing something in those banks that, if the year bears out their expectations, will justify a price advance and presumably an even higher multiple. Or it may be that a naive and cheap to execute strategy like low price to book is naturally too simple and exploitable to reliably produce outperformance. I may have to repeat this experiment with insurance companies or other years of historical data.

BlueLinx (BXC) is a wholesaler of construction materials, mostly wood products such as lumber, siding, and plywood. What interests me, though, is the company’s enormous cash reserves. Of its $800 million market cap, nearly $500 million consists of cash, the fruit of a very profitable 2021 and 2022, which the company seems to have no idea what to do with. Obviously my suggestion would be either a massive special dividend or repurchasing of shares (the company has repurchased $120 million so far, but it could handle plenty more), but the point is that for a cash-flow-positive business with a credit facility available to cover liquidity needs, most of that cash may be considered “excess,” and can essentially be deducted from the company’s market cap when analyzing it.

In terms of financial theory, a company is basically worth the money it has plus the present value of the money it will have. Obviously a company needs to keep some cash on hand for liquidity needs or other conveniences, but the issue is one of return on assets. Ordinarily every asset inside a company is contributing to that company’s return on assets, and the attractiveness of the company depends on whether its assets exceed, equal, or fall below the investor’s required return, and since that return is invariably higher than the short term interest rate for cash, a company should try to operate with as little cash tied up as possible. However, if there is “excess” cash not needed for the company’s operations, that cash is not required to meet any return on assets hurdle, because it can (and usually should) be distributed to the shareholders. I should point out that at the beginning of 2021, the company operated with the princely sum of $82,000 in cash on its balance sheet.

So, if we reduce the company’s market cap to $310 million, are the company’s earnings adequate? I did point out that 2021 and 2022 were remarkably profitable years. In fact, the company’s net income was $300 million in each of those years, and if those results were repeated the company would have a P/E ratio of…one. However, the market recognized those results as largely exceptional, with the share price actually declining in 2022, and as the company is essentially tied to the home construction market, which is apparently in a bit of a slump right now, this is understandable.

At any rate, in 2023 sales were $3136 million, gross profit was $527 million, operating income was $138 million, net interest expense (the company only started reporting interest income and expense separately in 2024–again, not used to having lots of cash lying around) was $24 million, leaving $114 million, or $92 million after taxes.

In 2022, sales were $4450 million, gross profit was $833 million, operating income was $439 million, interest was $42 million, leaving $394 million, or $311 million after a provision for taxes.

In 2021, sales were $4277 million, gross profit was $778 million, operating income was $438 million, interest was $46 million, leaving $392 million, or $311 million again.

It is unfortunate that BlueLinx’s gross profit margins tend to decline alongside sales, but that is the nature of operating as a warehouser and distributor. However, the company has been able to reduce inventory levels by nearly 30% between the end of 2022 and 2023, freeing up an additional $140 million in cash which, although not a suitable measure of long term free cash flow generation, as I have written before does suggest that management is not unaware or unresponsive to the slow housing construction market, even though selling, general & administrative expenses haven’t declined significantly since 2022.

At any rate, for 2024 to date sales were $1494 million versus $1613 million for the first half of 2023, gross profit was $250 million versus $269 million, operating income was $52 million vs $72 million, interest expense was $24 million versus $23 million, leaving $28 million versus $39 million, or $22 million estimated net income after taxes versus $31 million for the same period last year. Furthermore, depreciation exceeded capital expenditures by $7.6 million when historically they have tracked each other more closely, which is a source of additional free cash flow, not to mention a small further reduction in inventory.

So, if in a slump in the housing construction market the company is still able to produce, say, $50 million in free cash flow, that looks like a fairly anemic return against a market cap of $800 million but really quite impressive for a market cap of $310 million. And if the company is agile enough to take advantage of a resurgence in the market, as they were in 2021 and 2022, there may be other significant earnings windfalls in the future. However, I am not sure whether the housing industry in the United States has really reached its nadir, or whether the company enjoys a substantial competitive advantage against any competitors, so I would conclude it to be an attractive candidate for investment but not an unqualified buy.