Updated: Jun 13
June 13, 2022 | First a big thank you to readers of The Institutional Risk Analyst who liked our post regarding non-controlling interests in the disclosure of public companies (“Memo to Gary Gensler: Beware the “Non-Controlling Interest”). None of the media companies that are in the business of aggregating and re-selling financial data have yet been willing to report on the issue. We remain hopeful. Perhaps the Securities and Exchange Commission and FINRA eventually will look into the obvious disclosure and GAAP presentation issues this anomaly raises.
One wonderful technical supporter at Bloomberg commented that he couldn't change the calculation of equity to include non-controlling interests since it would "affect too many people." Well, you are affecting them right now. Tens of thousands of shareholders of hundreds of public issuers are impacted by this false presentation of equity. We are hoping for a new series including the non-controlling interest.
Not only does the characterization of equity as a “non-controlling interest” strike us as inaccurate in many (but not all) cases, but the question of inaccurate data and derivative metrics raises a lot of nasty questions for the world of “artificial intelligence” or AI and credit. If the inputs into an automated system are wrong, then the output is garbage. Garbage in, garbage out as they say in Silicon Valley. But then, technology companies would never accept the error rates that are considered normal on Wall Street.
Q: Can a new era broker dealer like Robinhood Markets (HOOD) use AI to power retail investment recommendations without checking the accuracy of the data? There are hundreds of issuers that utilize “non-controlling interests” in their public company financial disclosure, for example, yet the data vendors do not inspect the presentation of the data.
In the world of investment banking, you cannot use vendor data for fairness opinions or other materials, for example, because of such anomalies. So, is a typical 2-3% error rate for financial data and metrics OK for retail investors? The issue of data quality and fidelity is crucial not just for consumers, but because it goes to the very heart of risk in the fast approaching credit cycle.
All of the "new" loan approval systems that were developed after 2008 and during the period of quantitative easing (QE) are biased down in terms of the cost of credit. These are systems primarily designed to originate an asset to a minimum level of confidence and sell it to an investor while minimizing the legal liability of the issuer. Loans held in portfolio by a bank, for example, are subject to far more scrutiny and thus incur more cost of origination.
As we’ve documented in The IRA Bank Book Q2 2022, the cost of credit in bank owned 1-4 family mortgages has been negative for five years. As default rates normalize in this pool of large, mostly prime mortgage loans, what do you suppose will happen to the less prime mortgage exposures that were sold into an ABS? FHA exposures, for example, could reach the high teens in terms of delinquency in this cycle -- before home prices actually crack a couple of years from now.
Source: FDIC/WGA LLC
Of interest, last week the Consumer Financial Protection Bureau (CFPB) issued an order to terminate Upstart Network (UPST) from its list of approved “no-action letters.” The CFPB granted special regulatory treatment to Upstart by immunizing the lender from being charged with fair lending law violations with respect to its AI-based underwriting algorithm, while the “no-action letter” remained in force. Upstart requested an amendment to the “no-action letter” that effectively seeks immediate termination.
“Loans originated by Upstart are either held, sold to institutional investors, or retained by bank partners,” notes the CFPB. “In 2021, Upstart’s bank partners originated 1.3 million loans, totaling $11.8 billion.” It will be interesting to see how the various vintages of UPST production perform in a rising interest rate environment. The CFPB, of note, is taking increased interest in technology that supports lenders and the vendors that provide these systems of record.
The CFPB granted UPST a first “no-action letter” in September 2017 and a second letter in November 2020, just when the utopian frenzy of nonbank fintech AI-enabled lending caused by QE maxed out. “No-action letters” typically grant individual companies special regulatory treatment, on certain specified matters, where an agency agrees to not take action for violations of law.
Remember, making good loans is about making good underwriting decisions, not machine learning. As the head of non-QM lending at a top-four bank said of using AI for automated post-close due diligence prior to sale: “No, I know what we are dealing with. Open the door and loan quality will degrade.”
The idea of modifying an automated credit model at the end of a decade of QE may strike some of our readers as significant. We agree. Our skepticism toward the world of automated credit originations is similar to our view of automated appraisals for residential mortgages. They work until they don’t, especially at major economic and market inflection points.
The reason why lenders continue to spend big dollars on data consumption and validation is the need to limit the process errors to a certain level so as to avoid spikes in credit costs. But it is axiomatic that the production that is sold to investors is always inferior to the production retained. Thus the big risk to originate-to-sell lenders is repurchase claims.
As we noted last week, it was the retention on balance sheet of UPST production that spooked investors (“Update: Upstart Holdings & Cross River Bank”). When investors see a nonbank issuer like UPST and its partner banks retaining exposures that were meant for sale to investors (aka, the “victims”), then you can understand the natural feeling of consternation. In the world of retail, this is what is known as "expired stock" -- especially when interest rates are rising.
Falling market liquidity in the private markets for asset backed securities (ABS) adds a certain urgency to news that UPST has decided to make significant modifications to its credit model. Given the way UPST handled its involuntary foray into warehousing unsecured consumer loans, we look for more revelations in the future. But the big picture facing the ABS markets is of greater concern.
New issue volumes for ABS actually rose sharply in May 2022, but the absolute value of $35 billion in new issuance remains well-below 2019 levels. Since January of 2020, ABS issuance has been throttled, a decidedly negative indicator for liquidity risk and the U.S. economy. Likewise the bid for non-QM mortgage loans is also moderating after a torrid 2021, raising more operational issues for residential lenders. Notice that total mortgage issuance of all asset types is now below $200 billion per month, as shown in the SIFMA data below, half of levels a year ago.
The May data for Treasury issuance, added to the record tax receipts, pushed the government’s cash balances over $1 trillion in May, but is now trending lower. Meanwhile, the crowd looking for risk-free returns at the Fed’s reverse repurchase window (RRPs) is growing toward $2.5 trillion. Notably the crowd does not include the dealers and banks that were the intended participants. The negative stigma of borrowing from the Fed remains too powerful as disincentive.
The RRP facility is now a policy challenge for the FOMC. We suspect that the usage of the RRP facility is going to become a problem because the crowd of money market funds looking for risk-free returns is only going to grow as rates rise. The only way to reduce the level of utilization in RRPs, it seems, is to push down the awarded rate and literally force participants back into Treasury bills. The chart below shows 30-day T-Bills vs the award rate for overnight RRPs.
This desire for higher risk-free returns may only be a transient problem, however, since we fully expect the 10-year Treasury note to start falling back towards two percent yield in short-order. Even as the FOMC forces up the front of the Treasury yield curve, the back end is more likely to rally. Could we see a rally in bonds and some down marks on servicing assets in future even as mortgage rates rise?
The 4x growth in the duration of the Fed’s MBS portfolio since last June describes the magnitude of the credit risk which resides inside all fixed income assets. The FOMC moved loss rates on 1-4 family loans down by an order or magnitude over the past four years. Given the huge amount of credit exposure embedded in private financial assets over a decade of QE, and the closely related demand for safe assets, the true risk free rate is probably less than zero. Ponder that as Chairman Powell tries to wind-down QE.
The Institutional Risk Analyst is published by Whalen Global Advisors LLC and is provided for general informational purposes. By making use of The Institutional Risk Analyst web site and content, the recipient thereof acknowledges and agrees to our copyright and the matters set forth below in this disclaimer. Whalen Global Advisors LLC makes no representation or warranty (express or implied) regarding the adequacy, accuracy or completeness of any information in The Institutional Risk Analyst. Information contained herein is obtained from public and private sources deemed reliable. Any analysis or statements contained in The Institutional Risk Analyst are preliminary and are not intended to be complete, and such information is qualified in its entirety. Any opinions or estimates contained in The Institutional Risk Analyst represent the judgment of Whalen Global Advisors LLC at this time, and is subject to change without notice. The Institutional Risk Analyst is not an offer to sell, or a solicitation of an offer to buy, any securities or instruments named or described herein. The Institutional Risk Analyst is not intended to provide, and must not be relied on for, accounting, legal, regulatory, tax, business, financial or related advice or investment recommendations. Whalen Global Advisors LLC is not acting as fiduciary or advisor with respect to the information contained herein. You must consult with your own advisors as to the legal, regulatory, tax, business, financial, investment and other aspects of the subjects addressed in The Institutional Risk Analyst. Interested parties are advised to contact Whalen Global Advisors LLC for more information.