Borrowings Suggest Small Company Owners Face Higher Risk

Analysis uses business credit card loans to gauge market perception

The value of the publicly traded companies that make up the S&P 500 reached over $30 trillion in 2020. The same year, the Federal Reserve estimated the total value of private entrepreneurial equity in the U.S. at more than $12 trillion.

While returns on individual stocks and indexes such as the S&P 500 are easy to calculate, the lack of publicly available market data on privately held firms makes it far harder to grasp the risks of, and returns on, these privately held firms.

A working paper from UCLA Anderson’s Francis A. Longstaff and University of Delaware’s Matthias Fleckenstein proposes a methodology for estimating private equity returns, making use of credit card securitization data. They then plug this data into a structural credit model to solve for the implied risk and return of equity investments in these private enterprises.

Financial Data on Private Firms Can Be Elusive

Previous attempts to estimate such figures have landed all over the map. Studies have found average returns on equity in private firms can fall anywhere between 5% and 60%. The authors explain this wide variation is likely due to researchers having to rely on “incomplete, noisy and potentially biased data” on the various metrics that might illuminate private equity’s earnings. For example, the privately held firms the public often hears about — hot startups in a venture capital fund’s investment portfolio or a family firm targeted for a buyout by a private equity titan — make up a tiny fraction of private firms in the U.S. And they are likely not representative of this broader class of privately held entrepreneurial firms, which are often smaller and have fewer employees (think your local hardware store rather than, say, WeWork).

But if you’re a potential investor, good luck finding financial data on the local hardware store.

The authors address this challenge by using data from card issuer Advanta, the secondary market prices of its securitized credit card loans, that covers a wide range of private firms — everyone from television producers to contractors to independent online retailers. This data is useful for two reasons: One, credit cards, along with bank loans, are the most common source of financing for small entrepreneurial firms, and two, these credit card loans are packaged into securities that behave much like bonds.

Since junk bonds can be used as a window into the equity value of leveraged companies, the pricing of securities backed by these firms’ credit card borrowing could also yield insights into the risks and rewards of ownership stakes in this pool of small private companies. Indeed, the credit spread on pools of Advanta business card loans behaves similarly to corporate bond spreads, yielding a certain annualized return as cardholders pay their balances, along with interest and fees, while also being subject to risk of default if cardholders are delinquent in their payments. In particular, it resembles lower rated bonds in the BBB and high-yield segments, suggesting that the market sees entrepreneurial credit risk as driven by the same factors affecting the broader business sector.

Examining the Data Over a 10-Year Period

Advanta, a financial company, offered credit cards to business owners and their employees. The data covers the period from August 2000 to December 2010; at an approximate midpoint, the end of 2006, the firm had more than 1 million accounts, with an average balance of $6,000 and credit limits ranging between $5,000 and $25,000. Advanta’s credit card receivables were securitized into a portfolio that the authors, for simplicity’s sake, treat essentially as a single portfolio (in reality, there were three tranches with different risk and return levels).

As a first step, the authors calculate credit spreads for the Advanta business credit card securitizations. This allows them to compare how the market priced the credit risk of Advanta’s customer base of privately held firms, versus credit risk for other asset classes. Their analysis shows that, especially in a downturn, the market pegs entrepreneurial credit as significantly higher risk than household risk.

Before the financial crisis, credit spreads for Advanta and other consumer credit card securitizations (i.e., personal credit cards issued by American Express or Bank of America) were similar. However, this began to change in 2008 when Advanta’s credit spread reached a maximum of 1,750 basis points (17.5 percentage points), nearly double the nearly 900-basis-point maximum that consumer credit card securitizations reached during the crisis. Over the entire 10-year sample period, this difference in credit spread averaged out to about 284 basis points for Advanta and 124 basis points for other credit card issuers.

How Does the Market Perceive Entrepreneurial Credit Risk?

This still leaves the question of whether the different behavior of these credit spreads reflects the reality of greater credit losses in the Advanta bonds, or if they result from investor perception — how the market perceives entrepreneurial credit risk compared with other types of household or consumer credit risk. To better understand this, the authors control for key structural variables of the different securitization pools’ credit spreads, such as the monthly payoff rate and the portfolio yield. They find that even after controlling for these structural risks, a significant credit premium remains for the Advanta pool, of about 166 basis points.

“The market views entrepreneurial credit risk as fundamentally different in nature from that of conventional household or consumer credit risk and prices it accordingly,” the study suggests.  This market perspective does not appear to be unfounded: Further analysis by the authors suggests that entrepreneurial credit risk responds to different macroeconomic variables than other types of credit risk. For example, changes in credit spreads for consumer credit securitizations are linked to economic shocks that directly affect households, such as inflation or employment conditions.

By contrast, changes in the Advanta credit spread are linked to changes in housing values, measures of financial volatility, such as the VIX, and broader macroeconomic indicators, such as industrial production or consumer confidence. Their results dovetail with previous research showing that the propensity for entrepreneurship is linked to individuals’ access to capital, which for smaller entrepreneurs often means borrowing against their personal assets, such as their home or financial portfolio. These findings therefore support the authors’ hypothesis: “Entrepreneurial risk is linked to collateral values and the risk inherent in individuals’ wealth.”

A Risk and Return Model for Other Markets

Finally, the authors take the Merton model, a time-tested formula that links a firm’s equity and debt values to estimate credit risk, and solve for the expected return of private equity based on these estimated credit risk figures. They arrive at a 14% expected return, which translates to a premium of between 10.4% and 11.7% above the risk-free rate during the 10-year sample period. They also find volatility of around 95%, which is comparable to the volatility of stocks in the smallest size quintile over the same period. The returns from these small firms are also rather volatile, as illustrated by the value of its beta, a metric that measures the volatility of a single stock relative to the stock market as a whole. The market’s beta is definitionally 1, while the authors calculate an average beta of 1.21 for the Advanta companies, implying higher volatility. This figure is consistent with previous estimates in the literature.

Although their data set covers a limited set of small enterprises, it nevertheless offers a data-driven glimpse into an important but poorly understood sector of the U.S. economy, one that “represents a substantial fraction of aggregate household wealth,” the authors write. Furthermore, they note, the same approach to estimating risks and returns can easily be applied to other markets where securitization data is available.