Description

Worthington Industries (WOR) was written up in June of 2020 by Hamilton1757.  The crux of that thesis was that WOR had a large and valuable stake in Nikola (then worth $1.2bln) but was trading as if the NKLA stake was worthless.  I want to provide a hearty thank you to Hamilton1757 for sharing the write-up not only because the situation ended up being quite lucrative but also because it brought what is a great company to my attention.  I’ve reloaded on the stock in recent months as it has sold back off to levels that are, by my estimation, cheaper even than June 2020.

Over these past 2 years, WOR sold it’s NKLA stake for nearly $650m and used that cash (and on-going earnings) to:

• Buy a number of businesses (spending ~$500m over 7 quarters)

• Buy back ~$375m of stock reducing the share count from 55.6m to 48.4m

• Continue paying dividends

• Fund a working capital increase from ~$1bln to ~$1.9bln as steel prices have risen (Net W.C. from $550m to $900m)

Summary:

WOR is a high quality business and has been for decades.  It is currently experiencing the best run of profitability in its history but is trading at all-time low multiples to its balance sheet (and even more dramatically depressed all-time low multiples to earnings but, as its over-earning, that’s less helpful).  The company has a tremendous capital allocation track record and has taken the recent weakness in the stock as an opportunity to substantially ramp their buybacks activities.  If the stock languishes here and even if earnings immediately fall back to a conservative “normalized” level, the company could easily afford to buy back enough stock such that management and I are the only shareholders remaining in 8 or so years.  Under a 5 year holding period, using conservative estimates of normalized earnings and a conservative historic P/E multiple of 12.5x points to an IRR in the mid-teens.  As the forecasts (for multiple and/or normalized earnings power) grow less conservative that IRR could quite realistically rise to 30%.

Overview:

Worthington describes itself as an industrial manufacturing company, focused on value-added steel processing and manufactured metal products.  They produce pressure cylinders for LPG, CNG, oxygen, refrigerant and other gas storage; water well tanks for commercial and residential uses; hand torches and cylinders; propane cylinders; helium-filled balloon kits; speciality tools; and, through their JV’s, complete ceiling grid solutions, laser welded blanks, light gauge steel framing for commercial and residential construction, automotive service stampings, and engineered cabs and operator stations and cab components.  Put more simply, WOR makes: steel cylinders for a large number of applications, tools, ceilings, framings, and automotive parts.

The company is owner-operated being 35% held by the current Chairman, John McConnell (former CEO and son/grandson of the founder) who, from what I can tell, has always treated other shareholders well.  The current CEO, who worked under John as CFO and President for nearly 15 years, owns a further 1.5% of the equity.  Notably, these ownership stakes come despite the company dramatically reducing the share count over time.

Worthington is a good business and has been for many decades.  The company’s history as a public company goes all the way back to 1983 but Bloomberg only has financial data since 1987.  Over the 35 years from 1987, the company has lost money (on a GAAP basis) just one time - in 2009 when the company had to write down steel inventories by $105m and took a $97m goodwill impairment; excluding these factors, the company would have reported positive pre-tax profits.

Prior to 2021, WOR’s Net Income margins have run 3-8% in every year (excluding 2009) averaging ~5% over the entire time.  EPS has compounded at ~6-7% which is accomplished through GDP-like topline growth, stable margins, and buybacks.  Shares have declined from ~93m in 1987 to ~48m today with the rate of buyback accelerating in recent years (10 years ago there were 70m shares).

In addition to the buybacks, WOR has paid dividends every year as a public company and currently yields 2.8% having grown its dividend by 7% annually over the past 5 years (but, which the EPS growth, generally keeping the payout ratio on normalized earnings flat).

In the last 2 years, results have been much stronger than usual.  Net Income margins approached 11% last year and were roughly 7% in the year that just ended May 31st.  While these results probably aren’t sustainable, the earnings they generated (along with the NKLA windfall) have permitted WOR to reinvest in (and repurchase) the business at a more rapid pace than previously.

In the past 2 years, the company has acquired 2 tools businesses (at 9-10x EBITDA with which they think their existing businesses have synergies), a supplier of advanced and lightweight automotive parts (at 5x EBITDA), a producer of motors and transformer laminations for the electrical steel market (which might accelerate WOR’s growth in coming years).

Market Disconnect:

Historically, WOR was much more tied to the steel processing businesses it owns.  Today, steel processing is just 40% of earnings but the company still trades with and like steel mills rather than the diversified industrials it is much closer to.  That said, in the last 2-3 months, even the historically very close correlation with steel equities (which never made much sense but has made even less sense over the last several years) has broken down:

How many steel mills or businesses highly tied to the volatile steel cycle have lost money only once in the last 35 years?  How many have margin and growth profiles like WOR?

The other area where I think the market gets things wrong are the reported earnings.  With a large steel processing business, WOR, obviously, holds large amounts of steel in inventory.  When steel prices go up, WOR’s steel in inventory gains value and WOR generates a gain on the sale of that steel (in addition to the processing profits); when steel prices go down the opposite occurs.  Several times in the last couple years, I’ve seen the market sell WOR hard after some ostensibly bad earnings announcements that can be attributed to the inventory holding gains/losses.

In WOR’s fiscal Q3 report issued on 3/23/2022, WOR reported $57.5m of Net Income or $1.16 of EPS (missing consensus by ~18%).  This was down 46% QoQ and down 16% YoY.  If, however, we add back the inventory holding losses (with a tax effect) earnings were $1.55/share up 2% QoQ (from similar adjusted earnings) and up 70% YoY.  Despite that, and a pretty strong forward outlook (excluding more expected inventory holding losses), WOR traded down 17%.

In WOR’s fiscal Q4 report issued on 6/22/2022, WOR reported $78.5m of Net Income or $1.58 of EPS (beating consensus by a whopping 89%). WOR also raised their dividend 11%. These kinds of earnings misses by the street (89%, -18%, 130%, 29%, 39% over the last 5 quarters) speak to the street's confusion about the business.  As just one example, in the run up to the earnings release, Bloomberg put out a piece implying that steel industry data was diverging strongly from estimates for WOR which called for 26% YoY revenue growth - Bloomberg was implying that growth estimate was too high - in actuality, WOR reported 55% revenue growth.  In addition to its 89% EPS beat, though, WOR took a $92.8m loss on steel inventory holding during the quarter.  Tax effected, this was a $71m ($1.44/share) headwind to earnings.  On an underlying basis the company did ~$3.05/share of EPS in Q4 alone.  The stock trades at $43.

Valuation:

WOR’s accounting is complicated by the fact that much of its value is in highly profitable JV’s which aren’t consolidated.  This obfuscates the accounting a bit.  People like EBITDA so I’ll note that Bloomberg’s headline trailing EBITDA shows $485.6m while its trailing Net Income is a very similar $425.4m.

A more accurate picture of WOR’s “EBITDA” can be calculated by subtracting (from headline EBITDA) minority interest (with a tax effect) and adding back earnings from affiliates - both of which come below the EBITDA line.  WOR’s historic EBITDA results (along with these adjustments) are shown below:

At present, the company is trading at roughly 5x trailing "Adjusted EBITDA".  This value understates EBITDA as it does not add back interest or D&A from the JV entities (though it also does not subtract I+D&A on the much smaller minority interest).

On an earnings basis, things are pretty similar.  The business trades at <6x trailing Net Income and ~15x its lowest Net Income over the last 10 years.  It trades at ~10x its average earnings over the last 10 years but, as mentioned, WOR should be a more valuable business at present given its acquisitions over the last 24 months.

I would also keep in mind that WOR is running at or slightly below its average level of leverage at <2x Net Debt to EBITDA despite the fact that WOR should see a ~$500m (25% of market cap, 60% of debt, 15% of EV) working capital unwind in the coming years.  In fact, consensus expects >$500m of FCF (relative to $230m of Net Income) in the 12 months to 5/31/2023.

Now, obviously, WOR is over-earning at present with Net Income margins of 7% (vs. a long-term 5% average) on revenue running well above trend in the LTM - how much it’s over-earning is the important question.  Given WOR’s stable long-term growth rates and margin profile, we can get a pretty good estimate.  If we assume WOR has grown the top-line at 3% since 2019 (in line with its historic rate but conservative given all the acquisitions in the past few years), we’d get a revenue profile of ~$4.1-4.2bln (not far from consensus for 2023).  At WOR’s average historic Net Income margin of 5%, this would be ~$210m of Net Income (~10% behind 2023 consensus but I don't put a lot of stock in the sell-side on this name), putting the business at ~10x earnings and implying that earnings are returning to their trailing 10-year average level ($209m).  WOR’s average P/E over the past 10 years (including the Covid and post-Covid period when its multiple has been much lower) is 12.5x earnings.  From 2010-2019 the average rises to 15.4x and from 1990-2022 the average is 16.5x.

Going back to an average P/E multiple (12.5x-16.5x) on an assumed level of normalized earnings ($210m) that is, if anything, conservative would see 25-50%+ upside in WOR (giving no effect to the working capital release).

Alternatively, WOR is trading at or below its lowest price to book levels in the last 10 years:

Trading up to the average here would result in 80-90% upside and probably more accurately reflects the changes to the working capital position and growth in the business that are not being captured by what is a largely backward looking P/E analysis.  Given WOR’s capital allocation history, they aren’t going to let that capital sit idle (and, in fact, have recently entered into a $175m receivables financing arrangement which will allow them to access this working capital if it doesn’t draw down naturally).

While we wait for a re-rating, WOR should continue to grow revenues in line with its historic average (3-4%), pay holders what is now a roughly 2.9% dividend that grows at a mid to high single digit rate (call it 7%) and buy back stock.  To maintain a 7% dividend growth rate with 3-4% Net Income growth, the company would need to buy back 3-4% of its stock annually.  At the current valuation that amounts to just $70m/year (vs. an average of $150m the past 4 years) which, along with the dividend is an annual return to shareholders of ~$120m on $200m+ of earnings - certainly doable while leaving plenty of cash for acquisitions.

All in, EPS should be growing at 6-8%+ annually (3-4% topline + 3-4%+ buyback) while we get paid 2.9% (also growing 6-8% annually) in dividends.  The total return should be 9-11%+ if the valuation stays the same and capital is reinvested at rates comparable to what the company is currently generating.

The math on that becomes pretty compelling.

If we assume a 12.5x P/E (last 10 year average) exit multiple and a 5 year hold, the IRR is 16%:

And it rises to 22% at a 16.5x P/E (lifetime average multiple) exit multiple:

If instead we use a P/B analysis (which would remove the potential conservatism of the P/E analysis and implicitly provide for a release of the excess working capital) the returns would be better at something around a 30% IRR or a ~4x over 5 years assuming the company goes back to its 10-year 2.6x P/B average.  This analysis may be generous as it assumes buybacks occur at a constant $2 billion market cap though going to a constant $3 billion market cap only lowers the IRR by roughly 1%.

Catalyst

Nothing specific so, time, I guess.