No products in the cart.
Private Lending (9) – Two Proprietary Algorithms
This time, we’re going to talk about two proprietary algorithms we’ve developed internally to better underwrite deals.
In addition to the LTV discussed in the previous section, we also calculate two other things: a power of sale/foreclosure cushion and the homeowner’s repayment ability. It’s highly likely that nobody uses our methods to calculate these two things, but we absolutely must use our own methods and adjust our underwriting conclusions based on the results. Calculating these two factors involves gathering a lot of data discussed in the previous sections, ultimately leading to conclusions about whether to lend, how much to lend, and how to lend.
Let’s start with the simpler one: calculating the homeowner’s repayment ability. Generally, for private mortgages, if the true LTV is low enough, proof of income from the homeowner isn’t required. However, if the LTV is hovering around 75% or the homeowner’s repayment ability is particularly strong, the homeowner will typically provide proof of income voluntarily. Banks usually calculate TDS/GDS (you can google this to find out what they are), but we list out each monthly debt from the credit report one by one. In addition to calculating the minimum payment, we also calculate the total debt amount and add the monthly payments for the first and second loans. Then, based on the homeowner’s family situation, number of people, lifestyle, etc., we estimate their minimum living expenses to determine their monthly consumption level. We then calculate the best, moderate, and worst-case scenarios based on their income situation. Generally, in most cases, it turns out that the homeowner’s family can’t pay off their debts even if they don’t eat or drink. This makes the conclusion obvious: they will eventually default (even though the reality is that there’s a high probability that a new lender will take over after the loan maturies and buy us out, but we don’t gamble on that). We give the other party two options: either provide more properties for cross-collateralization or push the LTV down further to the point where it can survive and power of sale/foreclosure. If they agree, great; if not, we walk.
So, do you see it now? The key to reviewing credit reports is not to look at the borrower’s score, but to examine the specific details of each debt and liability! A high credit score doesn’t guarantee that the borrower can repay the money!
This brings us to the second algorithm, which we internally refer to as the cushion calculation.
This is thanks to our other line of business, which is real estate, particularly in multifamily. From acquisition to renovation, from financing to daily operations, from leasing to evicting tenants, from legal battles to selling and estimating property values and exits, we manage everything internally year-round (this is what we’ve been forced to do after several very expensive lessons). So, for each of these stages, we have firsthand, real data. Most of the properties we operate are single-family, duplexes, and multifamily units, therefore, we never rely on estimates from real estate agents, mortgage brokers, lawyers, property managers, or appraisers (sometimes their optimistic estimates are less than half of the actual expenses! Trusting others’ conclusions will only lead to losses for yourself). So, based on the pitfalls we’ve encountered over the years, we’ve listed out every expense incurred after defaulting, according to the most realistic circumstances. This includes everything from sending collection letters, obtaining rights to sell, legal fees, court costs, eviction fees, property management fees, vacancy costs, various overdue extension fees, extra interest, fines, costs incurred during the power of sale/foreclosure period, selling costs, discounts etc and so on. We calculate these standard expenses for each power of sale/foreclosure.
Furthermore, we go back to the appraisal report and extract any details regarding additional expenses needed for property, such as water stains, mold spots, minor foundation issues, and any cosmetic repairs needed for the property’s overall condition. We determine the budget based on our actual construction experience.
Once all of the above is calculated, we can figure out, in the worst-case scenario, the true “cushion” when selling the property after a prolonged legal battle. This figure is completely independent of the LTV because these expenses can quickly push the LTV to a point where what may seem like a good deal is actually not feasible.
In other words, besides considering the LTV, it’s crucial to calculate the actual dollar figure. You’ll suddenly realize that for lower-value assets, you need to push the LTV even lower, for moderately valued assets, you need to examine the specific cushion amount, and for overvalued assets, you need to aggressively lower the LTV, perhaps even more so than for lower-valued assets (you can guess the logic behind this).
Finally, after reviewing the application, the property, the borrower, the title, and calculating the true LTV and safety cushion, we can come to a conclusion: whether to lend, how much to lend, at what price, under what terms and conditions, and what the timeline would be in case of default, and how to maximize the recovery of principal and interest if borrowers go default. After going through this process, reading all the documents, an inexperienced underwriter might take around 4-6 hours to reach a conclusion, whereas seasoned professionals like us can finish it all in about an hour. So, many people send us their applications for review, considering it as buying insurance for themselves 😂.
With the basics covered, we can finally move on to share some real (crazy) stories! Reality is always more dramatic than the movies. Till next time!