Description

Apr 2 2021:  Steel Partners Holdings (SPLP) is collection of decent quality assets with some hair trading at ~3x earnings – that’s Russian resource company territory!  There’s less than two turns of financial debt offset by securities – this isn’t a levered equity. 

It’s a partnership and so investors get K-1s.  It’s illiquid and trades less than $1m a day, although there is a little additional liquidity in the preferreds which also offer a terrific risk reward.   And it is controlled by Warren Lichtenstein and Jack Howard who own a combined ~60%.   Lichtenstein has a, errrrrr, mixed reputation with investors.  The biggest risk is a take under - although appraisal rights are written into the partnership agreement in the case of a squeeze out. 

I think we’ve reached an inflection point with more of the growing operating earnings accreting to equity.  Steel spent the last ten years taking out minorities in its controlled positions - a buyout is the next logical use of cash.  And we have an option on a large growth asset in WebBank.  

Lichtenstein’s fund, Steel Partners, got caught in illiquid holdings during the financial crisis and wasn’t able to meet redemptions.  Steel Partners Holdings was created to provide liquidity to LPs and listed in 2012.   Lichtenstein and Howard spent the last decade cleaning up the holding structure and buying in minorities/associates of their core listed positions, mostly at attractive prices following the financial crisis and during the industrial recession of 2015/2016.  The units are flat to where they listed in 2012, but minorities have been bought in and consolidated revenue and EBITDA have more than doubled.  

The partnership agreement retains the incentive structure of the hedge fund.  Lichtenstein and Howard get a 1.5% management fee levied on partners’ equity (book value).  And then they get a 15% incentive fee, paid in units, on the increased market value of the equity subject to a high water mark (this isn’t factored into my earnings multiples).  Paying them to get the stock up would be just fine with me!  There are some other related party transactions and reimbursable expenses, but none of this looks egregious.  The management fee has been running at $8m per year, expensed through the financial statements, and isn’t too much more than what the CEO and CFO would earn in a conventional C-Corp.  The unit price has been stuck in the mud – 2017 was the only time in the last six years when an incentive fee was paid, to the tune of $9m.  So far, Lichtenstein has been more interested in growing the pie and minimizing taxes than he is in clipping the fees.

The earnings are roughly 60% Handy and Harman, a collection of decent industrial assets where Steel bought out the 30% listed minority in 2017 at an EV of $900m or 9x EBIT.   This seems like a good benchmark for thinking about intrinsic value.   The other 40% is WebBank, a quickly growing Utah chartered industrial bank (FDIC insured but not regulated by the Fed) that serves as a partner bank/rent a bank for a collection of market platform lenders (fka as peer to peer), customer financing solutions and private label credit cards.  Most of the loans are sold on to third parties with WebBank retaining $750m of loans (4x equity) yielding 17% as of 19Q4 – this is a normal credit card rate and historical loan losses have been restrained.  As of 20Q2, WebBank executed a deft pivot away from market place lending and now predominantly holds SBA guaranteed PPP loans.

The company just announced that their 2020 10-K will be delayed because of “non material” accounting errors at SLI, a manufacturer of power converters and motion control.  SLI is 10% of sales and the errors relate to internal controls, revenue recognition and mis-marking inventories and receivables.  Mmmmm, bad but not the worst.  The stock is unchanged, reinforcing that this is non material, or at least that no one is paying attention. 

·         $350m of bank loans which is sub 2x EBITDA.

·         $150m of 6% Series A cumulative preferreds are due 2026.  Currently trading at $20 offering a 7.5% current yield and a ~10%+ YTM.   These are callable at $25.  The preferreds were offered for payment in the acquisition of minority interests at Handy and Harman and Steel Excel and so may have uneconomic sellers.  In 20Q2, Steel paid the distribution in kind. 

·         SPLP owns 5% of Aerojet Rocketdyne which I mark at the current share price of $47 vs. the $56 at which Lockheed is offering to buy the company (big anti trust risk).   They own 12% of Aviat (AVNW) which makes wireless networking products.  And, on a fully diluted basis, they own 49% of Steel Connect (STCN, fka ModusLink).  All together, at current prices, these shareholdings are worst $262m or $11.30 per share.  Steel made a lowball bid for Steel Connect that didn’t go through and is selling out of Aviat.

·         The pension is $526m gross and $183m net and discounted at 4% as of 2019.  Steel has been making large pension contributions and expects to continue making large pension contributions on the order of $30m per year.  Higher rates would ameliorate this.

The company financials are messy but Steel provides adjusted EBITDA figures that reconcile well to the cash flow statement (i.e. are real).  I’ve simplified the financials by grouping the Industrial, Energy and Corporate segments together and charging them with the full consolidated gross capex (there’s hardly any capex at WebBank) and then charging them for equity comp.  The WebBank adj. EBITDA figures are equivalent to EBT.  

At $14/unit, SPLP trades at 4x earnings on a trailing basis and 3x earnings on a normalized basis.   Some of my normal adjustments are guesses.  The additional earnings come from

1.       1)  The bankruptcy and liquidation of API Group which cost the Industrial Segment $16.5m in 2019.

2.       2)  Normalizing capex to $30m – there’s been some growth capex flowing through Handy and Hardman and API, but this should be coming down.

3.       3)  Normalizing WebBank earnings for the strategic pivot.  We don’t have any info on what WebBank is making on their current PPP and non guaranteed loan portfolio.  But WebBank earned ~$20m in 20Q3 pre tax pre provision income and has been writing back loan losses taken earlier last year.

4.       4)  Interest expense is coming down as the company bought out some preferreds, paid down some of the credit line and benchmark rates are down.

5.       5)  I assume a normal tax rate of 20%, but the tax situation is complicated by the partnership structure, the frequent write downs and losses (I assume to minimize taxes) and the NOLs.   Cash taxes have been lower reflecting reported net income which has been lower.

Adjusted EBITDA reconciles well to cash flow.

Over the past four years, cash flow has been eaten up equity growth at WebBank, pension payments, interest payments and excess capex.  And growth in BV has been additionally impacted by write downs and repurchases.  

But I think this changes going forward with the earnings/FCF accreting to equity.  And so, for the last 9M, during COVID, we have seen SPLP repurchase $40m of preferreds, take down the revolver by $50m and then buy back $18m of units at $9/sh in December (8% of the share count).  As SPLP closes out its minorities (STCN is the last one) and starts gushing cash flow, it seems logical that they either take a shot at the remaining SPLP units outstanding or get the share price up and start benefitting from the incentive structure. 

·              SPLP is still responsible for ~$30m a year in pension payments for the next couple of years, although maybe this changes with interest rates.

·              As of 20Q1, SPLP moved $50m of debt at API which is guaranteed by the parent into Other Current Liabilities.

As an additional check on value, Steel bought in the minority at Handy and Harman (the bulk of the Industrial segment) at an EV of $900m in 2017, they bought in the minority at Steel Excel (the Energy segment) at an EV of $85m in 2017 and then they bought Dunmore for $70m in 2018.  Add $186m in WebBank BV and you get a total EV of $1.2B or double the current market EV.  Deduct ~$300m in net liabilities and you get $900m in equity or roughly triple the current value.  This seems like a floor to me.  I’m not a lawyer but I like the valuation based on Steel’s acquisition history because it seems germane to any argument around appraisal value.   

It seems to me that SPLP is conservatively worth 4x the current unit price.  At $14/unit, there is plenty of room for Steel to make a lowball offer and minority holders to win, even as we haggle over price.

Industrial Segment:  The bulk of the Industrial segment is Handy and Harman.  In 2017, Steel bought in the 30% minority at Handy and Harman at $37.10 or $454m in total equity and $887m in EV ($234m in net debt and $199m in net pension adjusted for taxes at 25%).  HNH guided to 2017 EBITDA of $113m - $136m and capex of $20m - $30m.   So EBITDA – capex of $101m at the midpoint and 9x EBITDA – capex.  Pretty cheap!   

The segment financials look like this – my numbers are about $10m a year below the company reported adj. EBITDA – the difference is pension, stock comp and other.  For what its worth, Steel cut expenses aggressively in 2020 and I would expect some of these costs to come back.

Handy and Harman is a collection of diverse industrial assets.   From the 2016 10K, the pieces look like this.

·         Joining (Lucas Milhaupt) makes brazing alloys and sells into general industrial markets.

·         Tubing (Handy Tube) makes seamless steel tubing coils, with a specialty in small diameter (less than 3mm) and length (longer than 5,000 ft.).  Sells into petrochem, shipbuilding, aero/defense, transportation and semi equipment.

·         Building Materials (OMG) makes fasteners, adhesives and other for commercial roofing and resi roofs, decks and landscaping.

·         Performance Materials (JPS) makes sheet and mechanically formed glass, quartz, carbon and aramid materials for composite applications – aerospace, PCBs, commercial construction, auto and body armor.  HNH owned 39% of public but delisted JPS and bought out the remainder at a $70m equity value in 2015.  The 2016 performance for JPS is trashed by impairment charges of $32m and whatever other restructuring charges.

·         Electrical Products is the combination of SLI bought in 2016 for $165m or 9x EBIT and EME bought from Hamilton Sundstrand (UTX) for $63m.

The other piece of Industrial was API Group, a UK laminate and packaging company that Steel bought out through its listed sub CoSine in 2015 for $85m in 2015.  In 2018, Steel added Dunmore, a film and laminate manufacturer for $70m.  

In 2019 Steel reported big losses at API and in 20Q1 put API into liquidation.   I’m not sure what happened.  Reading through some trade press and internet snooping yields “unforeseen costs relating to its relocation to a new facility and a significant drop in demand for their products, due to unfavourable market dynamics and a shift toward more environmentally sustainable products.  The drop in demand was also attributed to tobacco customers shifting to lower-cost, alternative packaging, and competitive pressures due to overcapacity in the industry.”  There was a major customer loss in 2018 and API mismanaged moving its main US facility from Rahway NY to Lawrence KS.   It sounds like some of the work is getting moved to Dunmore which is still standing. 

I looked at the UK Companies House reporting for API to get a sense of what was going on.  The Companies House holds financial filings for all UK companies – very useful but holding structures and consolidation issues can make it difficult to know what you are looking at.  The Companies House reports for API show that 1) performance was flat to public financials from 2014, when Steel bought them out, to 2017; 2) this is a structurally low margin biz and 3) things fell apart in 2018.  “Restructuring and reconfiguration activity prevalent in 2017 and 2018 is now mainly complete.  2019 will be a difficult year for the group and will reflect the full year impact of the reduction in activity with a major tobacco customer which resulted in a reduction in the Laminates Europe workforce in the first half of 2019.  Underlying trading conditions remain challenging as businesses along the case manufacturing, erection and packing supply chain seek to internalise product wherever possible to both protect and enhance margins.”   API never filed for 2019 but the Steel filings reveal a loss at API of $16.5m.

Low profitability at API explains the margin decline in the Industrial Segment.   If you back out API, Handy and Harmen has been very steady at HSD margins.  The Industrial Segment financials, adjusted for API, look something like this. 

Energy Segment – is the old Steel Excel.   SPLP bought the 36% it didn’t own at an EV of $85m in 2017 ($177m in equity minus $64m in cash and $28m in listed stakes).   Steel Excel was a vehicle for rolling up well completion services.  Also includes Steel Sports which runs youth sports leagues and I think is small.   I don’t have much to add – this did $15m in adj. EBITDA in 2019 and so possibly upside as oil prices come back.  

WebBank – is a Utah chartered industrial bank.  What is an industrial bank?  An industrial bank is not considered a bank for Bank Holding Company Act purposes – i.e. it is regulated by the state authority and the FDIC but not the Federal Reserve.  It holds FDIC insured term deposits but no demand deposits.   In practice, WebBank is an asset light, regulatory light internet bank that that takes term deposits at competitive rates and then lends through a collection of marketplace lenders (LendingClub and Prosper), buy now pay later (Klarna and Paypal), consumer financing and private credit cards (Dell and Yamaha) and then other online retailers (Fingerhut).  Most importantly, this is an originate to distribute model and WebBank holds very few of these loans on its own balance sheet.  The loans are generally put back to the merchant partner or into syndication within a couple of days.  For 2019, WebBank held an average loan balance of $637m vs. $23.9B of sold loans for 2019.  And the average yield on its loan book was 16.7% - this is normal credit card territory and far removed from the most usurious rates on marketplace loans (often triple digits) or the ~60% at World Acceptance. 

By any normal bank metric, the returns are extraordinary.   This is a combination of being truly asset light – no physical presence and the web presence seems janky – and occupying this unique, profitable partner bank niche.  But it is my guess that this niche was getting more competitive.  In 2019, the growth in fee income came down and WebBank holding more loans for its own balance sheet at slightly higher interest rates. 

In 20Q2, WebBank made a deft pivot to PPP lending, termed out two to five years, while running down its market place lending.   The leverage and loan balance blew out backed by a PPP liquidity facility (other borrowings).  $2,147m of $2,590m total loans are PPP loans.  This leaves $443m of “at risk” loans, down from $750m in 19Q4.   Crucially, the PPP loans are SBA guaranteed and WebBank has no risk.

We don’t know what WebBank is earning on its PPP loans.  But for 20Q3, which should be a relatively clean quarter post PPP transition, WebBank earned $19m in pre tax pre provision income – so this seems like a good starting point.  As of 20Q3, WebBank already had $43m of allowances or ~10% of the “at risk” loan book.  And as of 20Q3, WebBank was writing back loan losses.  The credit loss history seems consistent with the ~15% rate on their loan book and so I am willing to overcome my knee jerk skepticism and believe that the loan losses are in the vicinity of reality.  

I see two issues at WebBank.  The first is that I’m not sure if there is any lasting franchise value in the marketplace lending channel that will allow WebBank to continue growing at a high ROE.  WebBank doesn’t do anything on the customer side.  But I suppose I don’t get why any bank should earn excess returns (I probably leave too much cash in my checking account).  Yes, WebBank seems to have a minimal cost structure and existing relationships.  And in a 2018 interview, Gilles Gade, CEO of competitor Cross River Bank of NJ, made clear that there is a lot of compliance risk and overhead that discourages new entrants. 

For what it is worth, KKR led a $100m round in Cross River at a $1B valuation in 2018.  This valued Cross River at 7x BV and 25x earnings (Cross River has a similar ROE near 30%) based on regulatory filings.  Cross River is a much better property than WebBank – Cross River seems to actually be building a customer franchise.   On the other end, Lending Club recently bought out partner Radius Bank for $185m or 1.7x book and 29x earnings.  So that seems like a good range for the valuation – 1.7x to 7x book – Ha!

The other issue is regulatory – I would expect the Biden administration and Rohit Chopra at the CFPB to take a deep look at fintechs and consumer finance.  This could involve tempering down on rates or consumer offerings or forcing them to become more bank like (thus obviating the need for WebBank).  More specifically, Utah has no usury laws, but WebBank, through its marketing partners, was making loans to customers in other states.  In 2016, Colorado sued and in 2020, WebBank and Cross River paid a $1m fine to Colorado.  Not surprisingly the Trump administration OCC had clarified that banks are the “true lender” of these financial products and thus can skirt state rate caps.  And so the situation is currently unsettled.  Adhering to the appropriate regulatory environment around state rate caps seems technically manageable, but seems like it would reduce excess profitability across the industry. 

Repurchases – In December, Steel repurchased 2m units at $9/unit.  From 2015 – 2019, Steel bought back 3.7m units mostly in the mid to high teens. 

Lichtenstein also made opportunistic open market purchases at some bottom barrel prices over the course of 2020.  Roughly 100K units each at $5 - $6 (May 2020) and $9 (Dec 2020).  Jack Howard and director Eric Karros also made small open market purchases and, more recently, SVP Gordon Walker has been buying the preferreds a little above $20.

This report is for informational purposes only and does not constitute investment advice.  The author may buy or sell securities mentioned at any time.   Please do your own work.

Catalyst

cheap cheap cheap, ramping cash flow accruing to equity, WebBank growth gets recognized by the market, management buy out