Description

Thesis Overview

I know what you're thinking - DB is a complete dog and this time is not different. Maybe you're right...but I'm here to pitch DB as a long in light of their recent restructuring announcement, which I think is a massive positive and evidence that after decades of being run by and for investment bankers, management has finally been forced to try and make some money for shareholders. With a retail banker (Christian Sewing) now in the CEO role, Garth Ritchie out as head of the CIB, and two incentivized, operationally-focused activists - Cerberus and Hudson Executive Capital - pulling the strings behind the scenes, DB is appropriately structured and staffed to pull off a turnaround that is not as risky as portrayed in the sell-side and media. 

The consensus view on DB right now is that this story is simply too complicated and fraught with execution risk to be involved, which has led to binary thinking about the outcomes: DB will either execute on all of its targets (8% RoTE, 70% cost ratio etc. by 2022) and the stock will do great, or DB will miss because of some combination of higher than expected restructuring costs, failure to grow top-line, central bank interest rate cuts. etc. and the stock will do terribly. My analysis in this write-up will try and simplify what DB is actually doing, quantify the risk to the remaining business, and accordingly provide a reasonable range of outcomes which I think illustrate that you can buy a sufficiently capitalized DB at a highly favorable risk/reward, without needing management to execute on all of its expense savings targets or for any revenue growth to come through to do very well on the stock (80%+ over 3 years or 20%+ IRRs). There are multiple, straightforward paths to significant upside (management gets all of the €6bn of cost savings, corporate banking revenues actually grow, the regulator lets DB eliminate a large portion of operational risk RWAs which creates excess capital) and with the stock trading at only 25% of tangible book and a liquid, saleable asset in DWS, downside is reasonably well protected.  

Summary of Restructuring Plan

To the surprise of many including myself, DB announced that it is closing down the entire equities trading division and eliminating a sizable portion of its FICC trading business, principally rates trading. The shuttered businesses carry €38bn of productive (excludes operational risk) risk-weighted assets ("RWAs") and nearly €300bn of leverage exposure which is being moved to a non-core unit, elegantly dubbed the Capital Release Unit ("CRU"). These assets consist primarily of DB's entire equities trading book (includes repo financing, liquid cash equities, and structured derivatives with natural roll-off periods) and also a meaningful portion of the bank's FICC RWAs (principally interest rate / macro trading assets employed in institutional flow trading). Against these assets is €3.5bn of expenses and €2.5bn of net revenue. Management believes they can take out €2.5bn of the €3.5bn of expenses in the CRU, as well as another €3.5bn of expenses across the Group, with €1.4bn related to the Private Bank segment. All together, there is about €5bn of tangible equity (13% CET ratio*€38bn of RWAs) to be released and €6bn of expenses to be eliminated here. To pay for this, DB will need to take €2.3bn of incremental cash restructuring charges, and will also of course lose the €2.5bn of net revenue associated with exited businesses.⁽¹⁾ Multiple sell-side analysts are focused on the headline €7.4bn of restructuring charges. This figure includes over €4bn of DTA write-downs, with the remainder arising from non-cash software, real estate, and goodwill impairments. Only €2.3bn of incremental charges will be flowing through CET capital, which is really what matters here.

(1) The net revenue in the CRU may go slightly negative in the coming years as "net revenue" captures interest expense as well as all interest income and fee revenue. Management stated on the strategy call that the pre-tax loss for the CRU will likely be a shade above 1bn in the out-years, so for modeling purposes I've assumed 200m of negative net revenue. 

Nearly 50% of the €6bn cost savings target will be derived from near-term headcount reductions, principally across Equities and the Private Bank, with the remainder mostly coming from (1) shutting down the entire equities trading platform and associated IT / consultants (DB has 10,000 external consultants to help manage these sorts of software platforms and data center needs) and real estate; (2) reduced investment, IT, and real estate expense in the Private Bank; and (3) reduced Bank Levy simply from having a smaller balance sheet. Given the nature of these items and the fact that management is targeting an absolute number rather than a cost ratio (which is often evasive in this industry), I believe it's highly likely that management will at least get close to the €6bn target. If they get halfway there just from headcount reductions, you get €3bn of incremental pre-tax income not baked into the current valuation (more detail later in this write-up) which is worth €15-20bn or more than today's entire market cap. 

The sell-side came out swinging against the restructuring plan because they have no incentive to stick their necks out and make a bold call on a name that has done nothing but go down for years. Much of the sell-side community and many large institutional holders are also panicked about revenue attrition across the remaining DB franchise which is derived from a general lack of understanding regarding the distinction between true corporate business (i.e. Fortune 500s) and institutional business (i.e. asset managers and hedge funds). 

Institutional vs. Buy-Side Business

One top 10 shareholder was quoted in the FT as saying DB is making a mistake because hedge funds like to trade FICC and Equities with the same bank so now DB will unintentionally lose tons of FICC trading revenue too. This sort of thinking gets at the heart of the misunderstanding. DB does not make any money on institutional / buy-side business, which is highly price sensitive and leverage-intensive (a significant problem for DB given their 4% leverage ratio), and precisely why they are closing down almost the entire institutional function. More specifically, corporate customers do not care if DB trades cloud stocks or lends money to hedge funds, nor do they care if DB trades interest rates for Brevan Howard. 

The essence of DB's new strategy is to focus on global corporate clients who come to DB for transaction banking (e.g. foreign exchange, cash management, trade finance, working capital facilities etc.), corporate lending, and the occasional M&A assignment or DCM/ECM issuance (executed in the Investment Bank).⁽²⁾ Most of this business is fee (not narrow bid/ask spread) business, but even more importantly it is relationship business. This provides for better margins - corporate banks across the industry tend to run at c.40% pre-provision margins vs. 20% for Investment Banks - with attractive levels of capital deployment and overall returns on equity. Importantly, Corporate Banking functions are a GDP+ type growth business and are well-insulated from alternative service providers. This stands in stark contrast to "flow" S&T activities which are increasingly under pressure from unregulated / lightly-regulated entities like Citadel who are not subject to onerous VaR restrictions or capital requirements. 

(2) It's worth noting that some of these activities (i.e. foreign exchange) which might ordinarily be thought of as corporate businesses are sometimes executed using capital resources in the Investment Bank, with the associated revenue split how management sees fit; this is important to flag because a bank run by investment bankers will typically fight - successfully - to have a favorable revenue share in their business rather than the corporate business where the relationship originated. This can have the effect of overstating Investment Banking revenue and making the at-risk revenue in the Investment Bank appear higher than it is in reality.

In the modern era, institutional business is a commodity offering that is almost exclusively a function of price rather than long-standing relationships. A large portion of daily volumes go through the brand name asset managers and hedge funds, and these guys do not really care if you took them out to dinner last week. Every bank can execute essentially any trade in liquid markets and spreads are basically non-existent. It helps a lot if you have a higher-margin custody or fund administration offering to cross-sell (JPM) or an absolutely dominate position and requisite scale in a particular product (MS or UBS in equities) but even these banks earn, at best, high-single RoTEs in their Markets / trading businesses based on reasonable estimates. Banks will also do prime brokerage (cheap short-term lending) to win trading business but the former is extremely leverage-intensive and both carry low overall profit margins. The banks with the highest returns in their Investment Banks - JPM and BAML - also have an enormous base of corporate customers, as well as wealth management clients, who drive high volumes of activity across all sorts of trading products (FX, cash equities etc.) often at fee-based rates as opposed to at the bid/ask spread. 

DB has a nice base of corporate customers who drive business to the Investment Bank, but the IB has long been disproportionately sized relative to its corporate order flow. On the pure institutional side, DB does not have a particular expertise or scale advantage in many products, perhaps with the exception of high-yield debt given its position in leveraged finance. To compensate for their shortcomings, DB did a lot of leverage-intensive prime brokerage financing to win institutional trading business but had limited scale and no high-margin business to cross-sell in an environment of razor thin spreads - hence why they've just reported that the standalone Investment Bank (which includes higher-return M&A and underwriting) is a c.2% RoTE business. DB used to disclose Global Markets as a standalone segment in 2016 - those financials revealed a business that struggled to be breakeven, and the industry has weakened considerably in the last 3 years, making the situation untenable for players like DB. 

What's at-Risk?

Probably the main concern among investors is whether an apparent large-scale restructuring will result in DB losing Corporate and/or Investment Banking business because (i) it is no longer a true "full-service" bank and/or (ii) bad press will result in customers getting nervous and finding new providers. There is a related concern that as a result of these concerns, DB will not be able to hit its 2022 targets - which rely on some revenue growth particularly within the Corporate Bank - and therefore the stock is uninvestable. As I alluded to in the introduction, I actually do not think DB has to grow revenue in the Core bank for the stock to be close to a double, and I have actually assumed some modest revenue decline in the Core bank in my base case. I will expound on this further in the valuation section but this context is worth noting here as well. 

Below is a breakdown of DB's 2018 Corporate and Investment Banking revenue: 

• €2bn of Equities trading revenue (all moved to CRU)
• €5.4bn of FICC revenue (c.€500m now in CRU, leaving €4.9bn "at-risk")
• €400m of Equities Underwriting
• €1.1bn of Debt Underwriting
• €500m of M&A Advisory
• €3.8bn of Transaction Banking

We know that €2bn of equities trading revenue and €500m of FICC revenue will be eliminated as those assets are wound-down in the CRU. This leaves c.€10.5bn of revenue in the legacy CIB across FICC, M&A / Underwriting, and Transaction Banking.

As I have attempted to explain, DB's Transaction Banking customers, who are generally the CFOs and treasurers of large corporates - do not know or care if DB trades equities or rates, so long as they can facilitate foreign exchange, global payments, cash management, etc. and provide them with attractive trade finance products. These products are deeply embedded, often highly customized, and based on long-standing relationships. This allows us to rule out the €3.9bn of Transaction Banking revenue as being at-risk.

Given the extent of DB's corporate relationships, which typically revolve around Transaction Banking, DB is often the go-to house for regular corporate debt issuance (note: this is how someone like HSBC gets its DCM business as well). Some of the €1.1bn of Debt Underwriting revenue also comes from DB's leveraged finance practice, which is typically driven more by private equity clients more so than corporates. Notably, DB is not paring back its S&T capabilities in high-yield debt - DB has a top-5 position across most leveraged loan league tables and probably does decently well from an RoTE perspective in the product. Financial sponsors will not care about this restructuring: they choose underwriters / syndicates based on (1) credit terms, (2) price, (3) knowledge of the underlying credit / banker relationship. DB still has an equity underwriting capability but even if they didn't, sponsors wouldn't turn away DB from a leveraged loan syndication because DB can't take their portfolio company public someday. In the sponsors world, all of these capital markets activities are competitively bid by multiple investment banks and typically come down to price. Relationships exist to some degree, but these are really siloed, independently-bid products. 

DB also has €900m of revenue across M&A and Equities Underwriting. Both of these products, particularly the former, are driven in large part by understanding of the underlying business and personal relationships built over a long period of time. Perhaps even more importantly, however, success in these businesses is driven by corporate lending, which DB is actually committing to grow. M&A and Equities Underwriting are pay-to-play businesses where cheap revolving credit facilities (or attractive trade finance terms in the Transaction Bank) can go a long way to win mandates. DB still has plenty of lending firepower, and is also retaining an equity sales force and research team to execute ECM activities. The only reason an issuer might not give DB a lead-left mandate following this restructuring is because they don't like that DB won't have a full S&T function to help trade the equity on the secondary market. This is certainly possible at the margin, but most IPOs today are highly liquid as soon as they are listed and the processes operate according to a highly standardized template. In any event, Underwriting has slim profit margins, typically operating at 80%+ cost / income ratios. You could write-off the entire DB's entire ECM franchise and it would have maybe a 20bps impact on the Group's RoTE.

The most obvious bucket of at-risk revenue is the €5bn of remaining FICC revenue. This figure is a bit misleading however - at the end of 2017, DB started combining "financing" (i.e. commercial real estate lending and investment grade lending to corporates to support direct IB relationships) in FICC. DB's financing business generated €2.2bn of revenue when it was last disclosed in FY2017. Accordingly, the true FICC trading business in 2018 actually did about €2.8bn of revenue, or €2.3bn given transfers to the CRU. Additionally, within this €2.3bn, there is some amount that is being driven by corporate business (e.g. foreign exchange forwards / interest rate swaps for corporates) and not institutional activity (e.g. junk debt trading). Again, this is important because corporates are not the ones who will care if DB trades equities or not. We don't know how much of this €2.3bn is actually coming from revenue splits from the Corporate Bank and therefore not pure institutional business, but it's likely a meaningful amount. 

For the purposes of illustration, assume there is €2bn of true institutional FICC revenue that is now at-risk because institutional clients will balk at using DB to execute or structure their trades given the absence of a cash equities / prime brokerage platform. How bad would that be? Realistically, DB needs some sort of FICC presence because (1) it's difficult to execute FX trades for a corporate if you're not already in the market / holding inventory; and (2) DB has a sizable debt underwriting practice due to its significant presence in leveraged finance, and it's a bit more difficult to win Lev Fin syndications if you can't make a market in the securities. But I am nearly certain that DB makes essentially zero profit in its pure institutional FICC business - the pro forma Investment Bank has a 90% cost ratio on adjusted basis (excluding favorable fair value adjustments in 2018), and this is with a large Financing business (tends to run at c.40% cost ratios) and M&A / Underwriting practice (80% cost ratios). By difference, I think this says the institutional FICC business running through the Investment Bank is probably being run at cost and is really there to support DCM underwriting and other real Corporate business. There may be marginal pressure on the Investment Bank top-line due to clients doing less FICC trading with DB (beyond the rates trading that is being wound-down), but (i) like essentially all institutional, "screen-based" business, activity will still be driven by price and execution capability on the specific security; and (ii) DB actually has some expertise in the FICC businesses that it is retaining (e.g. HY debt trading).

To sensitize the impact of a large-scale decline, I have assumed in my model (see below) a 30% impairment of the €2bn of institutional FICC revenue over the next few years, which represents €600m of revenue loss, as well as a 50% impairment of the €400m of Equities Underwriting business, representing another €200m of revenue loss. Altogether, I assume €800m of revenue loss in the PF Investment Bank, which represents a -3% revenue CAGR thru 2022 off 2018 adjusted revenue. 

Valuation

Put aside the restructuring targets for a moment. Today, DB has an excellent Corporate Banking franchise in which it already makes close to a 10% RoTE or €1.2bn of attributable profit. At a 10x multiple (HSBC trades at 12x), the Corporate Bank accounts €12bn of value or nearly all of DB's current market value and there is additional opportunity for cost take-out within the segment given a 70% cost ratio vs. peers in the 60-65% range. You also get:

1. An 80% stake in a decent, growing Asset Management business (DWS) which carries a value to DB shareholders of €5bn in the market
2. A profitable (but underperforming) German Private / Retail bank with about €8bn of tangible equity in it (worth c.€4bn of market value at 50% of tangible book), also with significant cost take-out potential
3. What will basically be an M&A advisory / Underwriting business (the new Investment Bank) with a reasonably-sized FICC trading capability accounting for €15bn of tangible equity (c.€2bn of market value at a 5x P/E). Importantly, the size and scope of the Investment Bank will no longer imperil the Group.

To "create" the Core bank, DB has loaded the CRU with relatively liquid assets and €1bn of associated pre-tax losses (this also represents anticipated stranded costs, worth roughly €-8bn of market value when capitalized at 8x). 

All together, these assets are conservatively worth somewhere around €15bn before any realization of the €6bn Group cost reductions. At a current market value of €13.7bn, I think at worst you can buy into DB at fair value with a free option on any semblance of a successful restructuring; if management can get half ot the targeted savings, it would be worth €18bn assuming the after-tax savings are capitalized at an 8x rate, less the c.€2bn of incremental cash restructuring costs. This alone would be good for a double in the equity. 

Given that the vast majority of the expense reductions are simply headcount / consultant reductions and the elimination of the Equities platform, and given that DB is already assuming €1bn of stranded costs in its public plan, I believe that an investment thesis that underwrites management getting most of what it's targeting on the expense side is reasonable. 

As an offset to this, I have also modeled a revenue attrition scenario as outlined above in which I assume the PF Investment Bank suffers from annual revenue declines of 3%. I assume no revenue growth in any segment except for Asset Management / DWS, which I grow at 2% p.a. in-line with standalone consensus estimates. 

Below I have pasted a summary of my valuation, as well as snapshot of the Group P+L / balance sheet. A few additional things of note:

• I assume the PF Investment Bank continues to run at a c.2% RoTE and have assigned it zero value.
• Core revenue declines €600m or 3% total vs. management projections for overall Core revenue growth. This is driven by the revenue attrition I have modeled in the Investment Bank.
• I have assumed management only gets €5bn of the €6bn target despite my assumption that Core revenue growth is negative. This is based on what I believe are conservative cost ratio targets for each segment.
• Management is targeting a 70% cost ratio and 8% RoTE target for 2022 - I am well below both of those targets as shown below. 
• Overall, this scenario results in 85% upside or 23% IRRs through 2022, assuming excess capital is valued at 1.0x.

Catalyst

Execution - no question management needs to show some progress on the cost reductions and pace of the CRU wind-down before they get any credit for the restructuring. Given the current valuation / perception I do not think much progress needs to happen for the stock to pop.