Theranos Trial: United States vs. Elizabeth Holmes, Day 5 (Finance Edition)

September 14, 2021 marked Elizabeth Holmes' fifth day in court and her second day of trial. The government continued questioning Theranos' former corporate controller, Danise Yam aka So Han Spivey, focusing on Theranos' lack of revenue and high net losses. The defense countered claims of desperation by arguing Theranos had ample cash (from investors) and "always made its payroll." (Earlier, Yam testified Theranos could not pay vendor invoices in 2009 until Ramesh Balwani provided a personal loan.) To explain Theranos' 2009 incident, defense attorney Lance Wade cited 2008-2009's worldwide financial crisis, i.e., events not unique to Theranos. Though Yam's testimony addressed arcane matters such as accounting statements, revenue projections, and options pricing, the jury never appeared uninterested. Having prior experience in a Fortune 500 company's finance department, I understood Yam's testimony but didn't consider it favorable to either party. Holmes' intent to deceive remains an unsettled question.

Before discussing Yam's testimony, an accounting primer is useful. If you're not familiar with "revenue recognition," "409A," "fair market value," "GAAP," and "accumulated deficits," don't worry. I won't explain those terms; instead, I'll explain their purpose. If I've done my job properly, you'll see both sides used Theranos' finances to engage in meaningless posturing. 

"There are lies, damn lies, and then there are statistics." -- attributed to Benjamin Disraeli 

Every company looking for investors needs financial statements capable of passing an outside firm's audit. Small companies can use P&L (profit and loss) spreadsheets and a "cash basis" method (money in, money out); however, once debt or equity appear, the "accrual" method is more common. What's the "accrual" method? It's a more lenient accounting standard allowing businesses to show income even if they haven't been paid yet. But why allow anyone to show income in advance of being paid? 

For Starbucks, accounting is mostly straightforward: you charge a certain amount for a cup of coffee, and customers pay you when they purchase your product. If you're Oracle or SAP, it's not as simple. Your customer might buy a product, but the product might be discounted 99% from list because a multi-year service contract is purchased simultaneously. (Is the listed price a sham? Can the customer cancel the service contract with 30 days' notice?) Perhaps your customer buys a ten-million dollar piece of hardware. Over what period of time can the customer pay the full amount owed before the revenue becomes speculative? When can Oracle book revenue if a customer buys three items tied together but each having different contractual terms? (e.g., hardware including a multi-year software contract plus access to a direct client support line) It depends, but let's focus on why accounting rules exist, including ones that allow different interpretations using the same numbers.

Every product needs to be sold, and every major company relies on a marketing and sales department. Though movies celebrate lawyers and documentaries praise compliance managers, within companies, these people are despised because they can hold up a sale and therefore a salesperson's commission. Unlike other professionals, salespersons usually work under contracts where most pay is commission-based. It is understood by all that if a sale or a certain number of sales are not completed monthly or quarterly, the salesperson will leave the company. With such high turnover, loyalty and ethics are non-existent in most sales departments. Insiders understand this dynamic and rely on internal compliance teams to rein in salespersons who might otherwise backdate contracts, forge signatures, offer unwise discounts, or lie to customers about contractual terms. (By the time the lie is discovered, the salesperson has already collected his/her commission and left the company, often with a glowing recommendation.) 

Making matters worse, finance departments--where compliance managers are often based--also need oversight, lest they offer unsustainable incentives to assist sales, especially at quarter-end, when everyone is wheeling and dealing. A three-year 0% loan might help push an older model off the lot, but if an entire dealership's books contain 0% loans, then its survival depends not on ingenuity or customer service, but on the direction of interest rates. Negligence in accounting can cause as many problems as accounting shenanigans, and both have caused bankruptcies, Enron being the most famous. 

Complex accounting rules encourage and require sales to work with legal and finance teams to finalize deals, but if a company’s accountants are corrupt, strict accounting rules will not prevent corruption. Enron used "SPVs" (special purpose vehicles) to hide unsustainable accounting practices. Unless a forensic expert knew exactly where to look, s/he could not discover Enron's fraud until after the fact, and perhaps even then, only in exchange for a CFO's reduced criminal sentence.

"The analysis the accountant came up with showed that [Rebecca] Mark's business was earning a mere 2% return on equity--a pathetic amount...[but] Mark's analysis showed...that the international business had been a success, producing over $1 billion of cash and earnings and making a 12% compound annual return over its history... 'Figures lie, liars figure,' says one of the accountants who worked on the analysis... Yet the accountant went on to note an even more astonishing fact: viewed through their respective prisms, they were probably both right." -- The Smartest Guys in the Room (2003), hardcover, pp. 261 

[The first analysis evaluated the assets based on cost, current cash flow, and current market value, whereas Mark's analysis included accounting structures designed to book earnings immediately via monetization and/or securitizations.] 

But neither Theranos nor Elizabeth Holmes committed accounting fraud. The government isn't alleging Holmes "cooked the books." It's alleging Holmes lied to investors, patients, and/or the public about her sponsors and a specific piece of hardware after pharma partners abandoned Theranos in 2009. Government lawyer Robert Leach proved Theranos lacked revenue in 2012 and 2013 and had accumulated deficits; however, the government failed to prove intentional fraud because Holmes never demanded any particular revenue be recognized a certain way. 

Furthermore, it is not uncommon for new companies, especially ones investing heavily in R&D, to show net losses. (Amazon is the most famous example of a successful company investing heavily in R&D while suffering consecutive years of net losses.) Though Wade didn't mention Amazon, the defense might in its closing statement. Wade spent considerable time emphasizing Theranos' R&D investments in order to show operating expenses were the reason for most net losses (rather than Holmes' desire to enrich herself). Yet, an Amazon comparison would be inappropriate. Of course Theranos spent most of its money on R&D. It's a goddamn scientific research company trying to re-invent medical diagnostics. What else should they be spending investor funds on? Artwork for the company's walls? Private jets? (The government tried raising Holmes' private jet expenses with Yam but failed to elicit specific numbers.) Regarding the use of accounting statements, both sides earned failing grades, but under the reasonable doubt standard, two "F"s favor the defense. (The burden of proving intent is on the government, not Holmes.)

The government then pointed to variances in Theranos' analytical metrics, alleging its once 9.5 billion USD valuation was based on stock sales, not actual or expected revenue. On this point, the government performed better, but only because Wade highlighted a part of a document out of context, giving the government an opening. 

Before we discuss Wade's miscue, we should mention the government's attempt to implicate Holmes in Theranos' pricing of stock options. Attorney Leach argued Holmes inflated the value of Theranos and its privately-owned shares to gain investor funds on false pretenses. (Assuming the founder of a company desires majority control and outside investors, a stock option valued at ten cents a future share is less advantageous than one valued ten dollars a future share.) 

To assign intent to Holmes, the government displayed fundraising documents with different share prices: 0.15, 0.185, 0.5, 3.00, 15.00, and 17.00 USD/share. The most popular rounds were at fifty cents and seventeen dollars, with latter sales generating about 730 million USD. Simply put, Theranos received cash upfront in exchange for less insider ownership of the company. If everything worked out, outsiders would receive 5 to 40 times their initial investments. No one complained about stock option pricing when they bought in, and it's hard to see how anyone could have challenged internal projections. Unlike publicly-traded companies, a private company is just that: private. Its stock price is whatever it says it is. 

Exempt from the scrutiny of public markets, a private company looking for investors must project future revenue, often with little visibility. (Now perhaps you understand why banks require collateral.) In 2014, Holmes, along with Yam, Balwani, KPMG, and another accounting firm, decided Theranos would earn 100 million USD in 2015; 200 million in 2016; 300 million in 2017; and 500 million in 2018. What were these numbers based upon? Hope and investor funds. On December 22, 2014, when Holmes received an email with the numbers above, she responded, "100 mil for 15," i.e., 100 million USD for 2015? She said nothing about 2016, 2017, or 2018. Why did Holmes specifically mention 2015? Presumably because she knew 100 million USD of investor funds had been deposited October 31, 2014.

Holmes is not an accountant. She's never worked a low-level finance job. She didn't know the difference between cash received from investors and bona fide revenue. A week before New Year's Eve, she was presumably concerned only with the following year, when she knew she'd have to make good on her public statements about Theranos. From my perspective, the accounting documents proved only inadequate board oversight. Of course a conflict of interest exists when a founder or majority shareholder of a private company approves revenue projections and therefore her company's worth. That's why a board of directors exists, preferably containing at least one person with an accounting or finance background. 

As stated above, Holmes would have had a successful day were it not for an oversight by her lawyer. In attempting to show Holmes' projections were reasonable, and, furthermore, that private company valuations vary significantly, Wade showed the jury a document in which independent accountants pegged Theranos' projected value between 1.9 and 9.5 billion USD. The lower number was based on an "income approach," and the higher number probably based on the price some investors paid for Theranos' stock. Alas, Wade omitted a crucial fact: in the same paragraph as the 9.5 billion USD number was a comment declining to give that number any weight. The higher valuation was "not assigned any weight" in part because its method valued all shares, including preferred shares, equally. On re-direct, Leach pounced, red-lining the relevant part of the document used by Wade. 

Still, it would take the day's second witness, Erika Chung, to nudge the government closer to its required legal threshold of "beyond a reasonable doubt." I will wait until Chung concludes her testimony before writing further.

© Matthew Mehdi Rafat (2021) 

Bonus: my favorite accounting example is from Alex Berenson's The Number (2004), which explains why ethical accounting is so hard: 

"You and your neighbor buy identical new cars on the same day. You take good care of your car, regularly changing its oil and putting it in garages instead of parking it on the street... Your neighbor is much less conscientious. Three years later you have spent $1,000 more maintaining your car than your neighbor. But you figure that your car is worth $2,000 more than your neighbor's, because it is in better shape. The way you see it, you've saved $1,000. Your neighbor disagrees. He thinks the cars are worth the same. He thinks you've wasted $1,000 on unnecessary maintenance. Who's right? You may not be able to tell until the cars are sold. But if you were a public company, you would have to estimate your car's value every three months--and so would your neighbor... Now, imagine that instead of owning one car, you own thousands of garbage trucks. If you underestimate how quickly those assets are losing value, either accidentally or deliberately, you will wind up overestimating your earnings [because of incorrect depreciation]. And it will be essentially impossible, until you actually sell the used trucks, for anyone to know what you've done."