Many investors instinctively think in terms of a normalized free cash flow yield when evaluating a company. There are times however, when a company’s cash flows are so complex that sell-side analysts stick to a flawed earning based approach. These situations can lead to exciting opportunities as there are more likely to be market mispricings.

Ocwen (NYSE:OCN ) is one of these companies. Ocwen is one of the largest mortgage servicers in the US, specializing in troubled loans. From 2012-2013 its stock price quadrupled as the company went on an acquisitions binge in the past few years and multiplied in size. Despite the big run up however, I have yet to see a report that properly values OCN based on cash flows. I think the reasons are three-fold: 1) OCN gets lumped with financial companies which tend to trade on earnings, and 2) understanding OCN’s cash flows can be a daunting exercise, and 3) management offers its own measure of cash flows, but it is so different from net income (as the below chart shows) that it adds to the confusion. Despite the aforementioned obstacles, I believe a detailed look at free cash flow is called for as Ocwen is at its core a service company, and the company considers itself maniacally focused on cash flows.

This article will focus on making sense of Ocwen’s free cash flows. I will start with a critique of management’s cash flow measure while developing a more useful way to think about free cash flow yield. I will then try to estimate a sustainable FCF level, and finish with a trading recommendation. As this article is more focused on methodology, those who are uninterested in the nuance of accounting adjustments can skip straight to the trading recommendation at the end.

Critique of management’s measure of cash flows

Management defines cash flow as adjusted cash flow from operations. They start with net cash provided by operating activities on statement of cash flows and subtract advance reductions used to pay down match-funded liabilities. Notice they don’t call it free cash flow — because it’s not.

Table 1: Management’s Adjusted Cash Flow from Operations

While this version of cash flows can be a good measure of OCN’s liquidity, there are three problems that prevent it from being a true free cash flow measure. First, there’s no deduction for capital expenditure. Second, the entire amount of advances should be deducted as it represents a return OF capital, not return on capital. Finally, most of working capital swing & various non-cash item that’s included in Net Cash Provided by Operating Activities (A) should be excluded (as explained later)

Theoretically, the free cash flow number derived after making the above adjustments would simply be net Income + depreciation & amortization — capex, or Owner Earnings as Warren Buffett calls it.

The next few paragraphs will explain these adjustments in more detail.

First, capex. What is the capital expenditure required to maintain for a mortgage servicer? Most people think of capex as investments in property, plant, and equipment. However, for a mortgage servicer a more important capital outlay is the acquisition of mortgage servicing rights (MSR). Because servicing revenues are typically based on Unpaid Principal Balance (UPB), OCN gradually loses its future revenues as a pool of mortgage decreases in principal balance. (this old Seeking Alpha article can explain MSRs much better than I can). The important thing to note is that amortization of servicing rights represents a loss of future earning power that needs to be replaced. In most industries it’s hard to assess maintenance vs growth capex. For mortgage servicers however, a series of MSR transactions in the past years that can give us a sense of replacement costs. As the below table shows (recreated from a Wells Fargo research report), these bulk transactions ran about 2x servicing fee. Applying the same multiple to OCN’s weighted average servicing fee of

35bps would get us to replacement costs of

70bps of UPB.

Table 2: MSR transaction prices

Now we have a rough sense of replacement cost, we can multiply UPB by prepayment rate to estimate a projected UPB runoff amount. Multiply that number by cost of MSR then gets you to a maintenance capex estimate. Table 3 below shows my calculation as of 3Q13. Notice that the resulting estimated replacement cost of

$481mm is much higher than the $236mm of GAAP D&A for the same period.

Table 3: Estimating the cost of replacing runoffs

There’s some room for debate here. One can talk about the use of subservicing contracts to be capital-lite, which Ocwen already does to a large extent with affiliate HLSS. However, that arrangement leads to lower servicing fees and is arguably just a form of financing. Alternatively, Ocwen can ramp up self-origination and reverse mortgages to add UPB. For now let’s just take a mental note for potential levers to improve this maintenance capex need. And yes, the results will be very sensitive to CPR and MSR price assumptions.

Second, the entire amount of advances should be removed from cash flows. As a servicer, OCN has the obligation to advance cash flows when mortgage borrowers don’t pay on time, in order to smooth out the cash flows to MBS investors. Since these advances can get into the billions, OCN borrows from the capital market, then pay down the associated debt (match funded liabilities) when receivables are recovered from the borrowers (here’s an example of advance securitization ). On average, servicers can borrow 70-80% of advanced amount while putting up about 20-30% of their own equity.

Referring back to Table 1 earlier, management subtracts only the debt funded portion of these advances, thus implicitly includes the equity portion in its adjusted cash flow from operations. However, is this really a return ON capital? No, it’s just return OF capital that OCN had previously fronted. Management actually admitted this in their recent investor day:

If you look at our, what generates a cash flows, there are three main categories. Operating earnings would be the biggest single area. The other two large generators of cash for the company, I would describe as return to capital investment. One is amortization of the non-cash expenseThe other component, which is also return to capital is the equity that we need to hold or acquire advances. So, those are the three main components of our cash flow.

For the purpose of calculating a free cash flow yield, we need a return on capital type of number. Therefore we should subtract the entire amount of advances, not just the part funded by debt.

Third, most of working capital and non-cash adjustments should be excluded. In general, whether to include these items are a bit of a judgment call on a case by case basis. For OCN, I ended up excluding most of these items as 1) the working capital numbers have a heavy non-economic component and too lumpy to be considered recurring, and 2) most of the add-backs to arrive at GAAP cash flow from operations actually have economic impact. As an example of OCN’ working capital being non-economic, in 9mo13 OCN received $265.6mm of cash from decrease in receivables and other assets, net, it turns out that the bulk of this amount is simply accounting mirage consisting of increase in other asset, which has to do with loans already sold to Ginnie Mae but OCN was forced to consolidate (see note 13 of 10Q). As for non-cash adjustments, a quick examination of the cash flow statement reveals various non-cash items that are economic in nature so I opted not to adjust for them. Examples of this type include gain on sale of loans, as well as origination/purchase/sale of loans.

So what does FCF look like?

Table 4 below shows the various adjustments I made to management’s adjusted cash flow from operations to get to a free cash flow number that’s useable for valuation purposes. As an extra check, table 5 ties the FCF number with how an investor might calculate free cash flow starting from the net income line. Either way, FCF is way below management’s cash flow measure. It is also much lower than net income, mostly because accounting amortization understates true replacement cost.

Table 4: Reconciling between GAAP CFO, Management CF, and FCF

Table 5: FCF = Net Income +D&A — Capex

FCF for LTM 3Q13 was

$182mm, or $1.3/share. I expect 2014E free cash flow to be about $470mm, or

$3 per share, which, at the current price of

47/share means a respectable FCF yield of

7%.

Growing toward an sustainable level

Now that we have a FCF measure, the next questions are naturally will OCN grow its FCF per share in the next few years? and What’s a normalized FCF number look like? In the next couple years at least, I expect to free cash flow and earnings to explode upward. Banks are still looking to sell MSRs for capital reasons and OCN has cited a possible $400bn UPB pipeline for the next 1-2 years. Industrywide, OCN and its competitors expect about $1 trillion of UPB available for servicing transfers from the big banks.

The real question is what happens in 3-5 years when banks are done selling MSR? OCN is currently losing 15%-16% of its UPB per year, if it can’t continue to find a pipeline of loans, it risks becoming a melting ice cube. What are sustainable levels of revenue, earning, and free cash flows after the wave of bank MSR sales are done?

We can think of what’s a sustainable level for OCN by working backwards: 1) estimate an amount of new UPB that OCN can comfortably add each year. Think of this as the amount that OCN can afford to let runoff so that ending UPB = beginning UPB. 2) assume a runoff rate to back into a normalized UPB level. 3) use that UPB to project cash flows.

First, what is a reasonable amount of new UPB that OCN can add a year when the dust settles? OCN has 2 major ways of sourcing MSR — by originating loans, or via MSR transfers from third parties. On the MSR transfer side. My view is that banks will get in the habit of transferring delinquent loans to a handful of special servicers (Ocwen, Nationstar, and Walter Investments), as these players have much better expertise servicing troubled loans, giving them cost advantages. So the key is new delinquencies. Federal Reserve data (chart below) shows that pre-crisis about 6% of total loans each year go from current to 30 days or more delinquent. Consider that there’s close to $10 trillion of mortgages outstanding, I believe OCN can comfortably get $100bn per year by just grabbing its fair share of new delinquent loans and ramping up its own origination (most likely non-QM, but that’s for another article). Put another way, in a steady state, the company can lose $100bn of UPB a year and still maintain its revenue base constant.

Take this mental exercise a step further. $100mm of runoffs a year at a rate of 15% implies that at $667bn of UPB, OCN can indefinitely sustain its revenue base. If we assume a lower runoff rate then UPB could be much higher. For illustrative purposes however, below is a stab at what FCF might looks like with $667bn of UPB. Notice that FCF is LOWER than 2014E despite UPB being higher! This is mostly because 1) MSR acquisition is likely higher in the future as competition bids up prices, 2) financing cost are likely to be higher with higher UPB and higher interest rates.

Table 6: a Normalized scenario

The point isn’t to pin down specific numbers, as any discussions about normalized level inevitably involve layers and layers of assumptions. However, I think the analysis here is enough to show that under a reasonably conservative set of assumptions, OCN can grow its free cash flow to a sustainable level that is much higher than 2013 levels. Keep in mind there are lots of potential upsides here. Lower CPR is a big one, but management can also shift business mix between agency and non-agency, improve delinquency with preemptive methods to lower cost, strive for more operational efficiencies, ramp up origination/reverse mortgages, share buybacksand so on.

The key risks, on the other hand are continually high runoff rates (in my mind unlikely), higher financing cost (likely), higher cost of buying MSR (likely), and regulatory/legal issues (uncertain).

Trading Recommendation: OCN is an attractive company but I consider the stock fairly valued at $47/share. I would be quite comfortable however, selling cash collateralized puts:

Jan 15 put with strike of $45. The bid price is currently $5.3. If the put triggers you would effectively be buying OCN at a price of $39.7/share. If it doesn’t get exercised you get a yield of 11.8% for waiting.

Conclusion

OCN is a fascinating company at the intersection of industry and regulatory changes. The accounting and corporate structure can get complex but investors need to get a rough sense of valuation metrics as to not be driving blind. I hope this article provides Ocwen (and other mortgage servicing) investors a mental framework that bypasses accounting quirks associated with earnings, thereby allowing them to focus on the industry drivers.

Disclosure: I have no positions in any stocks mentioned, but may initiate a long position in OCN, over the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.

Since you’ve shown interest in OCN, you may also be interested in