Warehouse lending is frequently characterized as a low risk, high yield business, yet there is a shortage of warehouse lenders. The large national lenders have either dropped out of the market entirely, or have restricted their lending to very large customers and very generic product. Many of the remaining second tier lenders focus primarily on early purchase programs for their own product.
Regional and community banks, which tend to be highly sensitive to the needs of their present and prospective customers, are reluctant to rush into a line of business that has been recently dropped by so many of its largest long-term players.
With demand high, concern about lack of yield isn’t likely to be keeping lenders out of the warehouse business. Perception of risk seems to be the more likely cause of the shortage of providers. Risk, however, can be prepared for and managed profitably, but first it needs to be identified.
So, where’s the risk?
To see the risk more clearly, let’s take a minute to look at the business. The warehouse lender’s customer is a mortgage bank that makes loans to consumers, closes loans in its own name, and sells the loans on the secondary market to takeout investors under pre-existing correspondent lending contracts which provide for, among many things, repurchase by the seller of loans that contain defects (including but not limited to fraud) or which fail within a defined period of time.
The customer will generally identify loans it intends to finance no more than 24 clock hours in advance of closing by providing the warehouse lender with a funding request accompanied by the pre-funding documentation required under the warehouse lending agreement. Note that closing has not yet occurred, and that the warehouse lender’s money will move to the closing agent before final documents exist.
After closing, final documents required by the warehouse lending agreement are sent to the warehouse lender. The customer assembles the balance of the investor package, including satisfaction of all open stipulations, and sends it to the designated takeout investor. As soon as the lender’s investor package is ready, the lender notifies the warehouse to ship the balance of the package (principally the original Note) to the takeout investor.
The takeout investor receives the packages from the mortgage lender and the warehouse lender, gives them at least a cursory review, and wires funds representing what it believes to be the correct purchase price to the warehouse. It provides a Purchase Advice, detailing the amount wired to the warehouse, to the mortgage lender by e-mail, fax or on its website.
The warehouse lender applies the wired funds to the mortgage lender’s obligation as provided for in the warehouse lending agreement. Principal outstanding for the particular item will be reduced, and the associated charges will either be paid or billed as stipulated in the warehouse lending agreement.
I’ve used the term “warehouse lending” as a generalization covering pure lending transactions, repurchase transactions and purchase-and-sale transactions. There are differences among the three, but the underlying scenario is the same: the customer chooses, and enters into an agreement with, a buyer, makes product according to the buyer’s requirements, sends the product to the buyer while taking payment in anticipation of a successful sale from a third party, and lets the buyer and the third party settle up once the product is delivered and inspected.
Does this sound like factoring? It should, but many entrants into the warehouse lending field aren’t familiar with asset based lending so they very often limit their review to the customer’s P&L and balance sheet, as they would with any commercial line of credit customer, and think they’re covered. The notion that, in the case of warehouse lending, the primary (and, realistically, the only) source of repayment is liquidation of the collateral seems backwards to a cash flow lender.
The primary repayment source is not merely liquidation of collateral, but consistent and timely liquidation of collateral at or above pricing sufficient to provide a net operating profit from net sale proceeds. Net sale proceeds are what the customer gets after the warehouse lender’s charges are paid.
Take any mortgage banker’s financial statement and see how much you need to deduct from loans held for sale to trigger insolvency. Divide that by the average loan amount for that customer. That’s the number of unsaleable loans it will take to put the customer in the tank, and it is typically not going to be a large number.
It might be possible to mitigate that loss by finding an alternative buyer for each rejected loan, but that will require time. The alternative buyer is also likely to demand a holdback, and 20% of the agreed sale price for a year after purchase is not unusual. The extra time to consummate a “scratch and dent” sale and the holdback can be significant liquidity factors.
My first asset-based customer outside of the garment business was an egg packer. The plant was kept scrupulously clean, but you didn’t want to be downwind of it even on a cold day. As a line worker explained, “the more eggs you put through, the more of them hit the floor.” The mortgage origination business is very similar in that respect, in terms of the percentage (very small) of loans that hit the floor as well as odor of those that do.
Anything more than an occasional flawed loan will have two effects on the originator – the cash effect of having the loan rejected, and the likelihood of triggering a higher level of QC on the part of the buyer which will add time to the purchase process as well as the likelihood of turning up more loans that can be rejected. Future pricing can be hurt as well, since rejected loans decrease the seller’s pull-through rate, and they cost the buyer review time without allowing the buyer to make a profit.
If a few rejected loans don’t kill the customer right away, they will create a high-maintenance relationship that will, at best, reduce the lender’s profit. Unless the conditions that caused the loans to be rejected are cured, it’s likely that more loans will be rejected, the customer will fail, and the warehouse will become the owner of loans that are most likely worth less than the financed amount.
What causes loans to be rejected?
De facto hedging. In this scenario, takeout investors rely on the underwriting and prepurchase processes to add conditions that seem to justify rejection of loans which had rates locked below prevailing market rates. This is, at best, an immoral practice but it is difficult to prove and is common.
Cash flow issues at the takeout investor. Most takeouts are either aggregators, who buy loans and then sell or securitize them in large blocks, or leveraged portfolio lenders, who finance their purchases by borrowing against the loans they purchase, or a blend of both. In either case, small cash flow disruptions can cause large problems. A rash of defaults requiring advances will cause cash flow challenges even to true portfolio lenders. Cash flow trouble first shows itself as an increasing delay in purchasing loans, and then worsens to a trend of rejections for insubstantial reasons.
Honest mistakes. These happen everywhere, but are found most often when there is insufficient depth or breadth of management, or high turnover among worker bees. The core competencies of sales, secondary marketing, underwriting, closing and shipping need to be adequately represented for a mortgage bank to be consistently successful. A culture which recognizes that honest mistakes happen will generally pass through fewer potential rejects than a “paint-by-numbers” culture. Breadth, depth and stability may result in higher payroll cost that pays for itself by preventing rejects. A company that started up during the boom years may be insufficiently prepared for the volatility we are likely to face for the next several years.
Product falling out of favor. The farther a loan product is from the 80% LTV, 30 year fixed rate full doc conforming standard, the more likely it may find a reduced universe of potential buyers in the secondary market. The old rule of thumb that a loan should not be warehoused if it didn’t have at least 3 potential buyers was widely disregarded during the recent boom years, but it remains a good rule. Not long ago, quite a few mortgage bankers were taking the question “where else can I go for a no-FICO manufactured housing program” seriously, to their peril. The more exotic a loan program is, and the definition of “exotic” seems to be broadening every day, the more vulnerable it is to rejection even if done on a pre-underwritten flow basis.
Closing agent errors, omissions, or malfeasance. The most benign example of this is the practice of some closing agents adding charges to the borrower’s account that are not authorized by the mortgage banker. These charges can bring smaller loans over the High Cost threshold, which makes them unsaleable except as deep-discount scratch-and-dent if they can be sold at all. Getting the consumer to agree to a clean refinancing can be a delicate, time consuming matter with no guarantee of success.
At the other end of the scale is the closing agent that wants wires sent offshore to an entity with a name that is similar enough to a major title insurance company to be taken for an affiliate. That kind of scheme is usually kept going for months while the intended victims get comfortable with it, until suddenly one day millions disappear along with any trail to “Sounds Like” Title.
Long broker chains. It’s often said that the more pairs of eyes that look at a loan, the more likely the loan is to be error-free, but that assumes that the pairs of eyes are focused on something other than pushing it through. Beyond that, it’s human nature for the worker bees, who are always under time pressure, to take the word of a previous worker bee on such time-consuming things as employment verifications, so mistakes get passed on, not corrected.
Insufficient pre-closing quality control. Files are built slowly in the mortgage lending process, and they are contributed to by many people, each of whom is focused on his or her specialty area. Just before the file goes to the closing agent, did anybody make sure that everything was in place, or arranged for? And did the file pass a Reasonableness Test in order to catch, and satisfactorily explain for the record, oddities like 6 months’ reserves for a borrower who barely squeaked through on the income ratios?
Why isn’t fraud on the list?
We have all heard the assertion that “the only risk in warehouse lending is fraud, and fraud is rare in the mortgage industry.”
An underwriting manager I used to work with liked to bet people lunch at Mimi’s, the most expensive place within walking distance, that any of his underwriters, picked at random, could find fraud in any loan file, also picked at random, in four hours or less. He would have grown fat on free lunches if most people weren’t fully aware that mortgage fraud is nearly universal and it goes on because no real-world underwriter commonly gets more than a few minutes per file to find it.
“The big, ugly fraud is what we catch,” said the underwriting manager. “We just hope that the little ugly fraud doesn’t hurt us too badly.” Fraud is always there. It’s how it’s dealt with that prevents, or causes, fraud-related losses.
Other risks to lenders
There are additional exposures that warehouse lenders take and, while most of them are related to the volume of lending and average amounts outstanding, two of them are not.
Both of these risks unrelated to lending volume are related to litigation, and neither is well known outside the warehouse lending community.
The first, and potentially most damaging, can be summarized by the phrase “you are who you fund.” First Alliance, a takeout investor and retail/wholesale lender, failed in its defense against charges of predatory lending, was slapped with a large judgment, and filed bankruptcy. The plaintiff class saw that Lehman Brothers had a relationship with First Alliance that included warehouse lending, sensed a deep pocket, sued Lehman Brothers on the theory that it “enabled” First Alliance’s predatory lending practices, and won a judgment in the amount of 10% of the judgment against First Alliance.
Lehman’s appeal was pending in the Federal system when Lehman filed its own bankruptcy, and although it is not likely that many of the connections between Lehman units and First Alliance were duplicated in other warehouse lending relationships, the case set Federal precedent in areas that can be expected to come up again. It is not entirely a coincidence that the warehouse lending subsidiaries of many large players in the financial markets were disbanded after the likely outcome of Lehman’s appeal took color from the increasing publicity about subprime and predatory lending.
This risk, especially to a new entrant into the warehouse lending field, like all of the risks set out in this short article, can be mitigated, although the First Alliance/Lehman risk is more complicated than it seems at first glance. It is a topic that needs to be dealt with at length. Suffice it to say, for now, that adequate protections can be put in place, and should be effective as long as they are strictly observed.
Another litigation risk comes from lending to mortgage bankers who are required by their regulators to have a warehouse line in place at all times. The lender who cancels such a banker’s only warehouse line has, in effect, put the banker out of business by causing its license to be revoked. If “New York minute” hadn’t been in the language already, it would have got there to describe how long it would take a New York banker to sue its warehouse lender for cancelling its last remaining line.
Adequate preparation can mitigate this risk, but no litigation risk can be done away with entirely. As with any other risk, the risk of being forced to defend against lawsuits, whether well-grounded or frivolous, needs to be factored into pricing.
Market Risk and Liquidity Issues
This is familiar ground to all lenders, and the situation in warehouse lending is the same as in any other kind of lending. Margins can be friend or foe depending on how much planning you put into them, but you can generally assume that your demand will be strongest when margins are weakest.
What’s different in the warehouse lending world is the likelihood of large swings in line utilization, which often happen for no predictable reason. While you can assume, usually correctly, that lines will ebb at mid-month and flood at month end, a wide variety of financial and non-financial factors can cause a customer’s utilization to peak quickly at any time. After all, in most markets $1 million in borrowing corresponds to only four loans, and even a small mortgage banker can generate a pop of a dozen or so “extra” loans on short notice. Warehouse lines consequently require more reserve liquidity than most commercial lines, as well as more attention to keeping unused commitments to a minimum.
Another factor to plan for is anticipated dwell time – the number of days required between the time you fund a loan and the time you settle it off the line. Most contracts call for a maximum of 30 days, but as a practical matter such a limit is unenforceable because dwell time is between your customer and the takeout investor. If you are assuming that, because the contractual maximum dwell time is 30 days, that you can fund warehouse lines entirely on DDA funds, your cost of funds projection will be understated. 90-day money needs to be factored in, and aggressive loss reserves taking effect after 90 days need to be factored into your cost projection as well.
As we have seen with the rapid implosions of many of the subprime lenders since 2006, a takeout investor can be issuing press releases about record growth and roaring prospects one week, and locking its employees out in preparation for filing bankruptcy the following week. In such an environment, traditional financial statement analysis is inadequate. The best defense a warehouse lender has is to create the ability to constantly know what is going on between each takeout investor and all of the warehouse customers who deal with it in terms of product type, dwell time and reject rate.
Another class of counterparty is the Errors & Omissions and fidelity bond carriers. Warehouse lenders usually require E&O and fidelity coverage from all of their customers and, if you aggregate this information by carrier, some interesting numbers are bound to result. My personal opinion of requiring E&O and fidelity coverage has been not to expect to collect on it, but to use it as a screening device – if the customer can’t get approved for the minimum required coverages, then the carriers’ underwriters may have picked up on something my underwriting might have missed. But it’s still helpful, if only to have a ready answer for the FDIC, to know which carriers you are relying on for coverage, and what the aggregate exposures are.
The Volume Trap
Real estate lending is different in every state, and every mortgage bank has a slightly different way of doing business. The idea that one size fits all in warehouse lending has not held up in practice, but most lenders had to learn the hard way that “economy of scale” is difficult to apply successfully to the warehouse lending business. A small operation with a dozen or so customers and a small, stable, cross-trained staff can compete successfully with “Charmat bulk process” warehouse lenders by offering a high level of customer service without compromising standards because it relies on knowledge rather than generalized systems and checklists.
By 2006, most of the warehouse lenders were using monolithic soup-to-nuts software from one of a very few vendors and competing aggressively with each other for market share. If you are funding $20 million a day and your software, or any of your infrastructure, is not giving you 100% of what you need, you are forced into the position of living with what you have.
You can’t put all activity on hold while you rely on a third party provider to fix the problem, and then start up again days or weeks later. Risk increases if it is impossible to quickly change monolithic software. When the subprime market began to melt down, all warehouse lenders needed good information on their exposure to subprime loans in general and Pay Option ARMs in particular, but most systems were not set up to report on loan types that were not common when the software was designed. Considerable midnight oil was burned by people working to get their data warehouse software to yield up data on loan types it wasn’t designed to report on. Meanwhile, the smaller, high-touch lenders were making corrections because they knew their customers and the loans they financed.
Another characteristic of the boom years was many large lenders’ practice of hiring inexperienced personnel, training them only in their narrow roles, and trying to manage by exception. As long as volume kept rising, the percentages stayed in line but in the end, $10 million of impaired collateral is $10 million of potential loss.
Can an entrant to the warehouse lending start small, grow slowly, and make consistent profits? It’s not exciting, maybe, but it is likely. Concentration on short term market share leads, as one credit officer put it, to “cornering the market on bad customers.”
Opportunities to make money, and opportunities to lose money, abound in the warehouse lending field. Many aspects of warehouse lending, such as reliance on liquidation of collateral as the primary source of repayment, are counterintuitive to experienced cash flow lenders. Success turns on addressing elements of risk that may be unfamiliar but which can be identified and provided for in advance.
This article has tried to identify the major elements of warehouse lending risk in a general way. To address the risk appropriately, it is best to have an analysis done of existing safeguards, procedures and automation in order to continue to rely on proven resources, adding no more new policies, procedures or systems than necessary.
Warehouse lending can be very rewarding to a lender who approaches it with one eye on identifying and dealing with risk and the other eye on cross sale opportunities, such as cash management and deposit products. Community banks may find that reverse economies of scale – avoiding complicated new software in favor of staffing with a few well-rounded employees rather than an “army of clerks” – can be strongly to their advantage.