COOP: variant fundamental thesis is playing out

This post is a bit delayed, but I wanted to walk through my take on COOP’s Q2 results.

I first recommended COOP at $9.01 when the 10 year UST was 2.39%… I did not expect to be at 1.48% today! With that said, I’ve always argued that COOP was much less rate sensitive than investors appreciated, and I think the Q2 results validated this. That doesn’t mean that it’s been a smooth ride, however, as the stock declined 24% from my initial recommendation before rebounding 44%, giving up those gains, and effectively ending up where it started. I did add shares on the day of my June 3rd website post ($7.88), but it’s certainly been a wild ride. If you look through the stock’s volatility, however, I do think you’ll see solid progress on the original fundamental thesis.

For perspective, COOP reported $118mm of Q2 pre-tax operating income or ~$1.25/share on a cash EPS basis (3% cash tax rate). These results were undeniably aided by over-earning in the originations segment on the surge in refi volumes, but it’s worth taking a step back and thinking about that number… an $8.96 stock just produced ~$5.00 in annualized earnings! The market reacted well to this release (+12% on the day), however there remains a lingering question as to what the true earnings power actually is. The market clearly remains skeptical. We’ll go segment by segment to create our own updated estimate.

Servicing segment

The servicing results were what you’d expect for a quarter where mortgage rates dropped ~30bps – a large negative MSR mark, higher realized CPRs, higher MSR amortization, and declining margins. A declining MSR discount rate assumption muted some of the fair value hit, but the net mark was still $205mm pre-tax. CPRs sequentially increased from 8.2% to 13.0%, but some of that was seasonality. At current rates, CPRs should rise again in Q3 before falling in Q4 and Q1.

COOP reported segment pre-tax income of $70mm and we’ll make a few adjustments to this figure. We can add-back $11mm of Project Titan up-front investments (ending in Q4), add another ~$2mm of remaining Project Titan recurring benefits (to get to $30mm run-rate), and deduct $17mm of excess and non-recurring reverse portfolio “other income” (similar to Q4 and Q1) to get to $66mm of normalized pre-tax operating income. This figure is already burdened for the amortization of the MSR’s fair value write-up, so it correctly reflects the current economic cost of replacing MSR run-off. On an annualized basis – before any additional portfolio growth – this implies $264mm of pre-tax earnings power. This margin profile is a tad below my original 2020 projections on an apples-to-apples basis (4.1bps vs. 4.4bps), but not bad for the current rate environment.

Originations segment

On the flipside, the originations segment produced extremely strong results. Volumes were up 76% sequentially to $10bn on strong refi volumes and full-period Pacific Union contributions. Pre-tax margins on funded volume (a bit different than management’s margin definition) expanded sequentially from 76bps to 105bps. Quarterly pre-tax income was a record $118mm, although Q3 is already trending better with 10% sequentially higher volumes and steady margins. A lot of my original thesis was predicated on unexpected improvement in lending margins, and this has certainly played out faster than I expected.

It’s worth reiterating that my prior through-the-cycle margin assumption for COOP was 87bps, so we’re unsurprisingly a bit above trend, but not far out-of-line. We entered the year with industry margins at 10 year lows, so pricing had a lot of room to improve before getting to excess levels. I saw several examples of industry supply cuts last fall, so the unexpected refi wave caught many off guard. It takes time to re-add staff and capacity, and some have permanently left the industry.

COOP is on track to end the year with ~$35bn of lending volumes, but an estimate of run-rate volumes (with normal refi activity) is probably closer to ~$32bn. At my prior 87bps estimated through-the-cycle margins, this implies normalized segment pre-tax earnings of $278mm.

Xome segment

The integration of the money-losing AMS business continues at a solid pace. Xome’s $10mm of Q2 pre-tax income is a combination of $12mm in non-AMS income and a $2mm loss at AMS. AMS should become breakeven by Q3. Management noted that 14/15 major AMS systems will be integrated at year-end, with one final integration in Q1 2020. Management alluded to run-rate segment earnings of $50mm for 2020. This is well below my prior estimate of $74mm, and I think this reflects a combination of conservatism for run-rate AMS margins and further declines from delinquency-driven exchange volumes. I’ll assume that $50mm is indeed the right target.


Corporate expense jumped by $10mm sequentially, but this was due to one-off items that should reverse in Q3. After adjusting for those items, corporate drag was $70mm or $280mm on an annualized basis. Management alluded to an additional $50mm of overhead opportunities, but I’ll ignore that for now as we don’t have great detail yet. Management also alluded to opportunistic unsecured debt pay-down starting later this year. They expect to retire the $592mm of 6.500% 2021 notes ahead of the July maturity date (these are currently callable), which implies over $300mm of annualized debt pay-down going forward (I previously expected $250mm/year). Interest savings should therefore reduce corporate drag to $270mm next year with additional upside from any opex reductions or faster debt repayments.

Thinking about rate sensitivity

The decline in rates logically produced two somewhat offsetting impacts: (i) a negative MSR mark and (ii) higher lending profits. Although these are roughly offsetting from an economic perspective, understanding this relationship has historically been difficult for investors because we’re trying to compare an immediate balance sheet impact with a somewhat transitory earnings tailwind.

We can still try to put things in perspective. The MSR mark was $205mm. Q2 lending income was $118mm, or ~$50mm higher than my normalized quarterly estimate for the business. If the lending platform is indeed a 1:1 natural hedge, we would therefore want to see another ~4 quarters of excess lending profits to fully “recover” the negative MSR mark ($205mm รท $50mm). This doesn’t seem out of line, as we’ve seen prior historical periods of “over-earning” last longer than 4 quarters.

Management also alluded to 30% of its existing portfolio that could save $225/month by refinancing. Not everyone will take advantage of this opportunity, but COOP’s Home Intelligence app should drive industry-leading conversion rates. This implies $100bn of MSR refinancing opportunities. At COOP’s existing refinance recapture rate of 50%, there’s an embedded $50bn lending opportunity that should continue flowing through origination volumes and drive excess profitability over a sustained period. Taken together, it’s not difficult to see the negative MSR mark fully recovered over a reasonable time frame.

Capital allocation

Management largely reiterated its current view on capital allocation. They will always look at bulk MSR acquisitions, however the current focus is on reducing unsecured leverage and continuing to drive operational efficiencies. Corporate leverage was 5.2x at Q2 and the target remains <5.0x. I would guess that management wants to get below 4.0x before entertaining buybacks or dividends.

It’s also worth reiterating that the Pacific Union deal was clearly a home run for management. COOP paid $128mm for a $25bn MSR portfolio and lending platform that is doing over $10bn/year in volumes at solid margins. If the acquired MSRs are worth at least 50bps, then management opportunistically purchased this lending platform for free.

Updated view on current earnings power

If we add all of our segment estimates together, we should get $322mm of pre-tax operating earnings or $3.43 of run-rate cash EPS. As a reminder, this figure is properly adjusted for the economic cost of replacing its MSR portfolio.

My original investment view was that COOP’s 2018 earnings were highly obfuscated and not representative of the ongoing earnings power of the business. I laid out a path to get to $3.69 of earnings power in 2020, and Q2 results are suggestive that we’re right on track. There have been a few puts and takes, but most importantly the underlying earnings power has shown strong resiliency during this period of declining rates.

So although investors shouldn’t be annualizing and capitalizing the Q2 cash EPS figure of ~$5.00, it should be much easier to bridge to $3.00+ in run-rate earnings generation. This sounded overly optimistic 6 months ago, but most of the evidence is already here today. I continue to think that this situation’s complexity remains a major hurdle to the stock properly re-rating. Once the balance sheet is properly de-levered, however, an initiation of a buyback should somewhat force the market’s hand. KKR’s presence remains interesting.

Tangible book value declined to $15.95/share, but this metric continues to ignore the off-balance sheet subservicing portfolio and fact that this business should earn excess returns on a run-rate basis. At $8.96/share, you’re getting a 40% run-rate cash earnings yield – I continue to find this very attractive.