Mitigating Homeownership Risk Using SPVs: A Novel Approach to Achieving Diversification

By Marc Biron and Michael J. Seiler

Buying a home is usually the largest purchase one will ever make. Given that people typically buy the most expensive home they can afford and that residential real estate allows homeowners to employ a degree of leverage unparalleled by any other investment opportunity, residential real estate ownership often results in a mal-diversified portfolio. Beyond the extreme allocation weight favoring this single asset class, homeownership also lacks diversification in that it represents ownership in just one home. Imagine the lack of diversification when owning the stock of just one company. Now imagine that single stock represents 70 percent of an entire portfolio of wealth. Because homeownership exemplifies an egregious violation of Modern Portfolio Theory (MPT) (Markowitz 1952), we propose a method to diversify the realistic financial constraint that homeowners expose their portfolio to the idiosyncratic risk of a single, owner-occupied home.

In our model, when a homeowner buys their home, the return on the asset is not determined based on the individual performance of the singular home, but instead on the movement in a portfolio generally representative of a Home Price Index (HPI) for the entire country. This is analogous to buying into an index of stocks across the entire stock market instead of the stock of a single company. The result is a substantial reduction in risk while maintaining a constant expected return.

High net worth individuals can buy into an already diversified portfolios of stock, such as Exchange Traded Funds (ETFs) or they can buy large blocks of individual stocks and create their own diversified portfolios. By extension, wealthy individuals can achieve residential real estate diversification by participating in this diversified Special Purpose Vehicle (SPV) pool or they may not feel the need since their home represents only a small portion of their overall wealth. But, this second option is not the reality for 99.9 percent of the population. Instead, real estate represents the lion's share of their balance sheet in terms of both assets (home equity) and liabilities (mortgage balance). We view this innovation as a way for the majority of borrowers to be made better off by instantly achieving a diversified portfolio, but specifically view this as an opportunity to really move the needle for lower income borrowers and minorities to become first-time homebuyers.

SPECIAL PURPOSE VEHICLE

Secured by a subordinated lien that generally settles at the sale of the home, we form a Special Purpose Vehicle (SPV) to track the US National Home Price Index (HPI). When the borrower eventually sells the home, if their individual home value exceeds the index, the excess is paid to the SPV out of home sale proceeds, and conversely, any underperformance is paid from the SPV to the homeowner. Mortgages in the SPV pool are marked-to market monthly, and since the return to individual homeowners is the weighted average return of the pool, the pool is always kept in balance. Kong et. al (2019) reports that 100 randomly selected loans in a pool reasonably mimics a national HPI. As we later demonstrate, this home diversification strategy eliminates more than 99 percent of mortgage credit risk.

The SPV pool consumer contracts begin with a zero balance, but balances are adjusted up or down monthly based on the relative performance of the homes' marked-to-market valuation and the weighted average HPI of the pool of loans held in the SPV. The SPV is typically settled at the sale of the home, but there is an early termination provision where borrowers can exit if they pay any balances owed, as determined by an appraisal, in full, or receive half of balances owed to them to discourage prepayment of "in the money" consumers.

Through structure, capital or guarantee, the SPV will be investment grade rated or the equivalent and will guarantee all losses to creditors. As such, there is no credit risk to lenders/creditors.

TRANSACTION PROCESS

At mortgage origination, the borrower executes a standard mortgage, fixed or adjustable rate, and with any term to maturity. Also at closing, the borrower executes a home price diversification subordinated second lien mortgage.

Bank lenders can hold zero down payment, as low as 500 Fair Isaac Company (FICO), no Private Mortgage Insurance (PMI) first mortgages with 5 basis points (bps) of average credit losses in portfolio or they can sell the loan to Fannie Mae and Freddie Mac in the ordinary way, but no less than a 3 percent down payment is currently required by the Government-Sponsored Enterprise (GSE) charters. The investment grade SPV guarantees lenders/creditors against any loss. The GSE charters do allow down to 3 percent down payment without PMI, but with recourse to the lender. The SPV would repurchase any defaulted loans from the lender causing lenders/creditors to have practically no credit exposure. Finally, zero down payment mortgages with the SPV guarantee could be funded through private placement or securitization—ultra low risk, high yielding mortgage assets.

CREDIT RISK ANALYSIS METHODOLOGY

Ruthless default describes the theoretical case where borrowers default on their mortgage the moment their loan-to-value (LTV) ratio exceeds 100. But extant experiments and empirical studies show that due to a myriad of reasons, people do not begin to default until they are underwater by around 25 percent. To update this analysis and ultimately analyze the reduction in mortgage credit risk due to home price diversification, more than 1 million mortgages in the public Freddie Mac database over the period 1999-2020 are examined to determine mark-to-market loan to value (MMLTV) at foreclosure and at final foreclosed home disposition.

From the Freddie Mac database, foreclosure and loss rates were measured by MMLTV at foreclosure and disposition. Then, with foreclosure and loss rates mapped to MMLTV, the foreclosure and loss rates by LTV and FICO score band at origination are modeled using the monthly Zillow national Home Price Index (HPI) data over the past 25 years. All of the rolling 12-month, 48-month, and so forth to 84-month defaults are examined. The worst loan holding period was conservatively used to arrive at projected foreclosure and loss rates using a traditional 20% down payment (80 percent traditional Qualified Mortgage LTV), at 3 percent down payment (97 percent maximum GSE LTV) and zero down payment. Loan loss severity is conservatively grossed up by 20 percent to account for a liquidation value discount.

Because the counterfactual is impossible to observe (that is, there is no way to determine how a defaulted borrower would have subsequently behaved with this diversification product when they benefit from having their home value adjusted higher and the MMLTV reduced), a true analytical methodology for this type of historical analysis is challenging. That said, every borrower had to be reanalyzed each month during the entire historical period using the mortgage origination date, local ZIP code level HPI and the US HPI to adjust every home to an MMLTV, where the home values in underperforming ZIP codes were adjusted up (and MMLTV, adjusted down) and conversely, the home values in outperforming ZIP codes were adjusted down (and MMLTV, adjusted up). After each month, an adjusted default (and loss) rate is determined by using the MMLTV and the actual historical default rate associated with each MMLTV bucket to derive the new portfolio default rate and loss rates based on the probabilistic rates for each month.

Determining a historical default rate by MMLTV requires a numerator (number of defaulters) and a denominator (total number of borrowers) for each MMLTV bucket. While loan balance is available for all homes each month, the challenge is that there are no observable historical market prices for homes, which means deriving monthly MMLTVs for homes throughout history is challenging. The numerator for the default rate by MMLTV bucket is relatively easy to identify. For non-defaulters, the historic home value is derived by using the purchase price and date and the appropriate ZIP code HPI and US HPI for the time period from origination to each historical observation month, which assumes that every home price in the ZIP code changes at the same rate. The denominator to determine default rate by MMLTV is derived by adding the defaulters to the non-defaulters for each MMLTV bucket for each month throughout the 20+ year history.

RESULTS

The average loss rate associated with the standard 20 percent down payment mortgage when the home's price is diversified is practically zero. In fact, the national housing market would need to decline by 20 percent (the down payment), plus approximately 25 bps before defaults start to become meaningful at 125 percent MMLTV plus loan amortization. Never in recent history has the market declined so dramatically.

Home price diversification reduces credit losses by more than 99 percent. Furthermore, losses for zero down payment home price diversified mortgages is conservatively estimated at 5 bps annually. To illustrate, it is a rare event over the 20+ year period examined that the national HPI declines by enough to drive MMLTV enough to where defaults become meaningful. Moreover, over the stress case, which covers the Great Financial Crisis (GFC) from 2007-2012, credit losses associated with a zero down payment would only have been 25 bps. These low loss rates compare to through-the-cycle traditional (20 percent down payment or PMI) mortgage credit loss rates of 29 bps, based on average loss rates reported in Fannie Mae annual reports from 2009 to 2020.

While zero down payment, no PMI, 500 minimum FICO score mortgages at likely lower interest rates would be compelling to all market segments, home price diversification directly addresses home affordability and equality. Furthermore, home price risk is significantly reduced, protecting wealth with an inherent Net Present Value (NPV) exceeding 12 percent of the home's value, as demonstrated below.

Adverse Selection

If all homes in the country carried this type of diversification protection, then by definition, the SPV pool would exactly match the national HPI. Even though the SPV pool will consist of a sub-sample of homes across the country, if the two are perfectly positively correlated, the SPV pool and national HPI would effectively be identical. Adverse selection would suggest this will not be the case. For instance, a homeowner might be more likely to seek this type of diversification protection if he believes his home's price appreciation might not keep pace with that of the national average. If homeowners are correct and if adverse selection became a problem, the SPV pool might eventually become depleted. To control for adverse selection, instead of the home's price being tied specifically to the national HPI, it is instead compared to the price movement in the SPV directly. The larger the number of participating borrowers, the closer the sample represents the population and the higher the correlation coefficient between the national HPI and the SPV portfolio.

Moral Hazard

Because the return to homeownership is based on the performance of the pool and not on the idiosyncratic price movements of the individual home, concerns over moral hazard arise. For example, the homeowner is less incentivized to properly maintain the residence since money invested in the home is paid solely by the individual homeowner, but increases in price are eventually spread across the entire SPV pool.

The mechanism to mitigate moral hazard risk and maintain incentives for home maintenance and improvements is to credit individual homeowner accounts for value added improvements made over time. In this sense, moral hazard concerns are controlled for in the same way as Shared Appreciation Mortgage (SAM) companies, who face similar challenges. To keep the SPV pool in balance, when an individual homeowner adds measured value as determined by a final appraisal, her account is credited for the value added for improvements, and all other accounts are proportionally debited to keep the pool in balance. Conversely, as determined by a final appraisal, any deferred maintenance is debited from the individual homeowner account and the remainder of accounts in the pool are proportionately credited.

Another potential moral hazard we control for is the concern that since the return to the consumer is based on the return on the SPV pool, and not on the homeowner's individual residence, a seller is not naturally incentivized to maximize their selling price as they normally would be. Again, consistent with the general practice of SAMs, the seller's value is based on an appraisal performed at the time of sale, not on the actual sale price of the home. If no principal-agents problems exist, then these two numbers should effectively be the same. By basing the SPV pool credit/debit to the seller on an appraised value, the seller remains fully incentivized to maximize the sale price of their home because they get to keep every additional dollar yielded from the sale.

Regret Aversion

Probabilistically speaking, it is expected that half the homes will outperform the SPV pool causing those homeowners to regret having to pay the SPV at the time of home sale, while the other half of the homes will underperform the SPV pool resulting in the receipt of additional funds by the homeowner at the time of sale. Kahneman and Tversky's (1979) Prospect Theory shows that loses hurt twice as much as gains feel good. In this sense, the weighted average utility (or satisfaction) of all homeowners in the pool will be higher than if homeowners purchased in the traditional way bearing all the downside risk themselves.

And this is no different from a diversification strategy in the stock market. An investor who holds a well-diversified portfolio might regret not originally putting all his money in Amazon, but that is only true because the investor can now safely look back and see how Amazon has performed. Hindsight Bias does not change the fact that the wise strategy at the time of investing was to choose a diversified portfolio of stocks since one does not know ex-ante which stocks will do well and which ones will do poorly.

Net Present Value Analysis

Another way to frame the decision to diversify the price of your home is to perform a NPV analysis. To calculate the NPV, we need the discount rate at which all future net cash flows will be discounted. The annual risk premium without home price diversification is equal to the long-term historical real estate return of 3.71 percent minus the risk-free rate on a 15-year Treasury of 1.78 percent, or 1.93 percent. Since the average volatility in the Zillow national HPI is 47 percent less than the volatility in the Zillow national home price volatility over 25 years after controlling for the diversification benefit, the risk premium is reduced by 47 percent.

If we assume a typically priced home of $300,000 and an average holding period of 15 years, then at the historical average annual growth in home prices of 3.71 percent, the home's price after 15 years will be $518,120. If we then discount this future home price back to the present value using the slightly lower risk-adjusted discount rate, the present value would be $342,255, $42,255 higher than the resulting unadjusted present value. After deducting the net present value of all fees associated with price diversifying the home, the resulting NPV to the homeowner is $37,353. Alternatively stated, the borrower would be better off by $37,353 in today's dollars if they price diversify their home.

Note that since there is no publicly available data at the individual home level, this analysis was done at the ZIP code level. Because of the added diversification benefit, the reduction in discount rate would be even greater at the individual home level, which would have yielded a higher NPV. The reduction in risk would reasonably be expected to reduce mortgage interest rates, which in turn would ultimately result in an even higher NPV for borrowers. But, even without the financing benefit, homeowners without a mortgage, which represent 33 percent of the national market, would greatly benefit from a home price diversification approach.

While compelling, an NPV analysis may seem esoteric to some. However, one way to think about the benefits of reduced home ownership risk is the opportunity cost of funds saved via residential real estate. If a homeowner wanted to keep the overall risk of his portfolio constant, then the newly realized reduced risk of homeownership could allow for increased risk-taking in the remaining mixed asset portfolio (for example, investing in higher beta stocks and/or using more leverage).

CLOSING THOUGHTS

High-LTV home diversified mortgages carry at most 5 bps of through-the-cycle credit losses versus about 30 bps of traditional Qualified Mortgages. This dramatic reduction in mortgage credit losses underpins our proposed value proposition to consumers and creditors. Additionally, due to the reduced risk in home equity, the homeowner gains an additional NPV exceeding 12 percent of home value.

The implications to the housing finance industry are profound. Zero down payment (3 percent if funded by GSEs) price diversified mortgages without PMI are now feasible at perhaps lower interest rates to the consumer.

As an immediate opportunity, lenders can streamline refinance the existing $1.5 trillion in Private Mortgage Insurer and Federal Housing Administration (FHA) mortgages—saving them ~$100/month, substantially reducing foreclosure risk, all while providing additional home price protection.

NOTE

1. Seiler (2014, 2015a,b, 2016) demonstrates that morality plays a substantial role in the resistance to default as does fear of lender recourse, among other factors.

REFERENCES

Kahneman, Daniel, and Amos Tversky, 1979, "Prospect Theory: An Analysis of Decision Under Risk," Econometrica, 47:2, 263-292.
Kong, Kwun Ho James, Rakesh Reddy Gunukula, Xinjie Dai, and Da Lu, 2019, "Home Value Index (HVI) Risk Reduction Leveraging Random Pools," Cornell University. Working Paper.
Markowitz, Harry M., 1952, "Portfolio Selection," Journal of Finance, 7:1, 77-91.
Seiler, Michael J., 2016, "The Perceived Moral Reprehensibility of Strategic Mortgage Default," Journal of Housing Economics, 32, June, 18-28.
Seiler, Michael J., 2015a, "The Role of Informational Uncertainty in the Decision to Strategically Default," Journal of Housing Economics, 27, March, 49-59.
Seiler, Michael J., 2015b, "Do as I Say, Not as I do: The Role of Advice versus Actions in the Decision to Strategically Default," Journal of Real Estate Research, 37:2, 191-215.
Seiler, Michael J., 2014, "The Effect of Perceived Lender Characteristics and Market Conditions on Strategic Mortgage Defaults," Journal of Real Estate Finance and Economics, 48:2, 256-270.