How forensic accounting analysis can uncover stocks with elevated risk

Rachel Bradley  by Rachel Bradley
  Equity Analyst, Gradient Analytics LLC (a Sabrient Systems company)

At times, there are conflicts of interest between a company’s management and its investors, specifically shareholders, which creates a structural agency problem. So, it is common for companies to try to align the monetary interests of management with shareholders by awarding a large part of executive compensation in equity. As a result, management performance metrics often create an incentive to present the current period financials in the best light possible, even if it is unsustainable in the longer-term. This dynamic creates opportunities for a forensic accounting firm like Gradient Analytics, which specializes in assessing the quality of reported earnings of publicly traded companies to both vet long candidates and to identify short candidates.

An astute investor might ask the perfectly reasonable question, “If publicly traded companies are regulated and audited, then aren’t all reported earnings of passable quality?” However, consider that it is possible for a company to engage in “earnings management” by publishing financial statements that may mislead investors as to the firm’s true financial health without violating accounting standards. Moreover, we would argue that even though two given companies both followed GAAP accounting standards and received clean opinions from their auditor, earnings growth at Company A may be dramatically more sustainable than at Company B. Of course, it can be quite valuable for an investor to know when a firm is showing signs of unsustainability in its reported earnings growth.

In this article, I use three real-life examples to provide a high-level overview of our analysis process, which includes assessing earnings quality, anomalous insider trading activity, audit control risk, and corporate governance. Although Gradient does not employ a team of investigative journalists to knock on doors, interview company employees, or rummage through the trash for evidence, our comprehensive analysis of the published financial statements (and, importantly, the footnotes) can help stack the odds a little bit more in investors’ favor.  Read on....

How can you tell when a firm engages in aggressive accounting practices?

The truthful answer to this question is that to evaluate the quality of earnings reported, an investor needs to either spend the requisite time pouring through all the granular disclosures or they can enlist an expert to do it. Interestingly, many accounting entries are subject to management discretion. Therefore, within GAAP there is a wide range of ways that management can choose to paint the picture of the firm’s financials. So, one firm may use very conservative accounting practices while another may engage in more aggressive tactics, but both could be GAAP compliant and get a clean opinion from the auditor.

Broadly speaking, Gradient uses forensic accounting techniques to assess how accurately the reported financial statements represent the underlying health of the company. Specifically, we look for unsustainable trends that may temporarily result in higher current period earnings by overstating assets or revenues or understating liabilities or expenses. These typically include elevated asset accruals and deteriorating cash flows. They also can include unsustainable working capital trends, sourcing a large proportion of sales from lower quality sources, changes to accounting estimates, a new revenue recognition policy, and other accounting gimmicks that result in current earnings being higher than they otherwise should have been. Often, weak or deteriorating customer demand will accompany these trends, as this can be the catalyst that compels a firm to engage in earnings management.

What does revenue from “low quality sources” look like?

In our first example, we examine recent sales from Codexis, Inc. (CDXS), a life sciences company that uses artificial intelligence to develop proteins. We were primarily concerned that growth in accounts receivable (AR) outpaced sales growth in each of the last six quarters. This is important because when growth in AR disproportionately outpaces sales growth, we view the associated sales as lower quality than normal. In Q1 2020, total receivables increased 58.0% YOY to $23.0 million while sales declined 5.9% YOY to $14.7 million. Consequently, the AR to three-month sales ratio grew to 156.5%. This is not only the highest level in the last ten years, but it is also notably higher than both the level last year (93.2% in Q1 2019) and the historically normalized level. In fact, the current ratio is more than double the trailing ten-year seasonal average of 58.2%.

Our concern when we see such a pattern is that sales appear to be boosted by either loose credit policies or aggressive revenue recognition practices, either of which may lead to deteriorating receivable quality and eventual collection problems. Furthermore, CDXS reported that trailing 12-month (TTM) sales grew 8.7% YOY, but we estimate that absent the unusual increase in receivables, sales would have been much lower such that the sustainable rate of growth is likely closer to just 1.1% YOY.

Not only is the rate of sustainable sales growth likely overstated, but the receivables themselves are also riskier than normal. A recent spike in unbilled receivables drove the growth in total receivables. This is particularly noteworthy because unbilled receivables carry a higher level of risk than other receivables. Companies use unbilled receivables when the timing of revenue recognition differs from the timing of invoicing to customers. While the mere presence of unbilled receivables does not necessarily set off a red flag, the fact that the riskiest part of receivables is the driver of growth in total AR does. It is also important to know that management has discretion over what is included in unbilled receivables. Therefore, sales from unbilled receivables have a much higher risk of negative restatements later if customers dispute any balances once they review the invoice.

There also is evidence of unfavorable aging within receivables. In Q1 2020, past due receivables represented 3.4% of total AR, which is up from just 1.3% in Q4 2019. So, not only is recent sales growth likely overstating the sustainable rate of growth, but there is also a higher risk of restatements and write offs.

Anomalous insider selling and other red flags:

Gradient also looks for insider transactions that may be suggestive of negative sentiment within the firm. When we see anomalous volumes of insider stock sales, we investigate to see whether it is broad-based (as a single insider bailing out is more likely to be driven by personal liquidity needs) and then whether it has other characteristics that we associate with increased risk of share-price underperformance.

Chief among those are option exercises that occur early relative to our option-timing model’s expectation – which takes into account the tendency of insiders to exercise options and sell the acquired shares earlier following periods of rising share price and of higher volatility, as well as a couple dozen other factors that are relevant to the issue – and the characteristics of Rule 10b5-1 trading plans. In fact, academic research has shown that the use of 10b5-1 plans on average tends to precede share-price underperformance, but our additional work suggests this is driven by the subset of plans that deviate from the ideal of regularly timed sales of uniform volume (e.g., 15,000 shares on the 15th day of every quarter).

To provide a recent example, in April 2019, we issued a report examining insider activity at Amarin Corp. (AMRN). We found a big spike in sales volume from Q3 2018 to Q1 2019, ranking first among all quarters in the trailing five years in both sales volume and market value by insiders at the company. In addition, no individual seller accounted for more than a third of the volume (indicating that selling was broad-based among several insiders); the sales were preceded by early option exercises on average; all of the sales were executed pursuant to trading plans; and the sales reduced the beneficial ownership of the sellers by an average of over 50%. The stock subsequently fell precipitously.

In addition to assessing the quality of reported earnings and insider behavior, Gradient also considers a variety of other signals that speak to the tone and culture of reporting within the firm. Specifically, we review any audit control red flags. Among others, these include a pattern of restating earnings much lower than initially reported, filing SEC reports late, high turnover of CFOs, unusual incidents with the auditor (e.g., short auditor tenure, significant vote against ratification, unusual change in auditor fees, etc.). We also stay abreast of correspondence with the SEC, litigation activity, and any significant deviations from expectations.

Additionally, we consider a variety of corporate governance issues. The most important among these, in our view, is executive compensation practices, but we also review the structure of the board, shareholder rights, and best practices among the industry. If firms tie management compensation to near-term performance metrics (like growth in sales or non-GAAP adjusted EBITDA), then we know which metrics are most likely to be “managed.”

As mentioned above, many performance metrics are tied to accounting entries that are subject to the discretion of management and its use of estimates. The context of how companies choose to evaluate executive performance can be immensely helpful when disaggregating normal (sustainable) performance from any cosmetic (unsustainable) enhancements.

Earnings quality analysis can augment and improve most investing strategies:

Once we dissect a company and view it from the variety of angles mentioned above, we have a much better grasp of the economic health of the underlying business. For example, let’s say that a firm displays red flags in each of the areas we discussed:  customer demand is dwindling from competitive pressure; the firm has deteriorating cash flows and appears to be engaged in some form of earnings management; several insiders have divested a material portion of their beneficial ownership in the firm; and to top it all off, they have filed late reports with the SEC and gone through two CFOs over the past year. Such information can be crucial to investors.

Our long/short hedge fund clients may use such an analysis either for a new short idea or to avoid a hazardous long position. On the other hand, an investor might use a more favorable analysis to help identify a company with solid earnings quality or to add conviction to a long thesis.

Material earnings management often slips past auditors and investors take the loss:

Reporting can be complicated, which is why we apply a detailed and rigorous method to evaluating companies. Even though the stock market tends to be highly efficient, it can still be irrational at times and reflect asset values that appear to be mispriced. For investors who want to protect their assets and mitigate risk in their portfolios, it is not enough to solely rely on assurances from management and the auditor opinion. One report estimated that external auditors catch as little as 4% of fraud cases. Earnings quality analysis can help bridge the gap and uncover assets with an elevated risk of underperformance. Next, we examine a recent case involving a grossly mispriced security.

Early in 2020, an anonymous researcher disseminated a detailed report alleging that Luckin Coffee (LK) materially overstated 2019 sales. On 02/03/20, LK issued its initial response which described these allegations as misleading and false. Specifically, LK said that the methodology of the report was flawed, the evidence unsubstantiated, and the authors had “malicious interpretations of events.” To be sure, anonymous researchers and activist short sellers have been known to make false or exaggerated claims to profit from market reaction. However, shortly after on 03/19/20, LK formed a special committee to oversee an internal investigation into concerns brought up during its 2019 audit. Specifically, the firm was exploring the possibility that employees had fabricated transactions to overstate sales.

On 04/02/20, LK told investors that the internal investigation indeed uncovered that the firm had materially overstated sales in 2019. LK employees misrepresented sales by buying a massive amount of vouchers that customers could exchange for product. The individuals involved directed the funds supporting these transactions to LK through a number of third parties that included fake buyers and obscure companies. The amount of fabricated transactions aggregated to ¥2.1 billion in sales from Q2 2019 to Q4 2019. This was highly material as the firm only reported ¥1.5 billion in sales in Q3 and ¥909.1 million in Q2.

As of this writing, LK has not yet filed its 2019 Annual Report so there is no public data for periods following Q3 2019. On 07/01/20, LK disclosed that the rate of overstatement increased during the year. Chronologically, LK inflated revenues by ¥0.3 billion in Q2, ¥0.7 billion in Q3, and ¥1.2 billion in Q4 2019. LK had also inflated expenses by ¥1.3 billion (or ¥0.2 in Q2 2019, ¥0.5 billion in Q3, and ¥0.7 billion in Q4 2019).

Following the special committee’s investigation, LK dismissed several employees based on evidence showing their participation in fabricated transactions. These included the CEO, COO, Chairman of the Board, and 12 employees. Another 15 employees were subject to other disciplinary actions, and LK is ending its relationship with the third parties involved. From 01/02/20 to 04/06/20 the stock lost 88.5% of its value.

The Nasdaq exchange suspended trading on 04/07/20 and delisted the firm on 06/29/20. LK shares now trade over the counter. Interestingly, there are reports that the former Chairman of the Board may still effectively remain in control of the company. Moreover, LK also removed the director that was the head of the internal investigation as well as two other independent directors, and there are questions surrounding the independence and loyalties of the two directors that replaced them.

Our key takeaways from the LK case study are twofold. First, investors should be skeptical of growth numbers that appear too good to be true. Second, we need to be careful with unprofitable companies that are raising money. While it is extremely common that firms need access to capital to scale up and become profitable, we should apply a higher degree of due diligence and skepticism to those firms.

The importance of an unbiased analysis:

Gradient Analytics provides investors with information about the sustainability of reported earnings growth, anomalous insider trading, audit control risk, and corporate governance. Our sweet spot is finding companies that are trying to squeeze every penny out of current period earnings without necessarily engaging in criminally fraudulent activity. Clients use this information to develop either a short or long thesis or to reduce exposure to a potentially hazardous long position already in a portfolio.

While some activist short sellers publish “research reports” for free to the public, this practice is often unreliable and biased, with the primary intent of supporting the author’s existing short position in the stock. The report on LK was an exception in that it was of quality and accuracy. However, it involved an extensive “boots on the ground” investigation that included data from 92 full-time and 1,418 part-time staff to run surveillance and record store traffic for 981 store-days on 620 stores. This group was effectively able to “rummage through the trash” to find “smoking gun” documents, so to speak.

On the other hand, rather than undertake such massive surveillance efforts, Gradient Analytics relies upon an evenhanded and impartial approach to evaluating financial statements and disclosures in the notes. Gradient does not allow its analysts and employees to trade in the stocks on which we write during the coverage period or for one full year after ending coverage. This keeps our research independent so we can give investors an unbiased view of a given company’s financials.


Disclosure: At the time of this writing, the author held no positions in the securities mentioned.

Disclaimer: This newsletter is published solely for informational purposes and is not to be construed as advice or a recommendation to specific individuals. Individuals should take into account their personal financial circumstances in acting on any opinions, commentary, rankings, or stock selections provided by Sabrient Systems or its wholly owned subsidiary, Gradient Analytics. Sabrient Systems makes no representations that the techniques used in its rankings or analysis will result in or guarantee profits in trading. Trading involves risk, including possible loss of principal and other losses, and past performance is no indication of future results. 


Earnings Quality