The tombstone of a famous short seller in a Boston cemetery reads: “I was right, but my timing was off.”
Short selling is rightfully considered the most challenging of all investment strategies. While many factors contribute to the difficulty of short selling, the fundamental challenge is the need for almost impeccable timing.
This is because the collapse of fraudulent stocks, such as Wirecard, frequently occurs only after a period of considerable pain for short sellers in a preceding period of excessive hype. Detailed forensic research cannot solve this timing dilemma.
The solution lies not in the depth of analysis, but in the breadth because timing risk necessitates diversification. To improve performance, short sellers need more extensive lists of suspect companies coupled with systems to pinpoint those exhibiting warning signs of an emerging price weakness. Until the advent of AI, this was impractical, even for large research teams.
Why short selling is difficult
Short selling is difficult for a range of structural, psychological, and practical reasons.
On the structural side, the math is unfriendly due to the asymmetric risk profile associated with short selling. When you buy a stock, the worst-case scenario is that it goes to $0. You lose 100%. When you short, there's no upper limit on how high a stock can go. The losses are theoretically unlimited.
A single bad short (especially in a short squeeze like GameStop or myriad other meme stocks) can wipe out an account many times over. This forces very tight risk management, but squeezes can happen extremely fast. Since some of the most obvious shorts in the market are meme stocks, most recently in uranium, rare earths and AI, this is a common problem.
Another structural problem is the market’s strong long-term upward bias. Stock indices trend upwards over time due to economic/earnings growth and inflation. Since the path of least resistance is mostly up, shorting is like swimming against the current. The majority of shorts tend to get squeezed or slowly bleed out even if fundamentally "right" in the long run.
The risk of a short squeeze is typically greatest in the most fundamentally obvious, and hence the most crowded short trades. When too many people short the same stock, any positive catalyst (good earnings, news, retail buying or even market manipulation) triggers rapid covering. The price explodes upward causing more covering in a self-reinforcing feedback loop that's brutal for shorts.
As a consequence short sellers are highly motivated to find shorts that are undiscovered by the market. This is not only time consuming but, almost by definition, means that short sellers often target stocks that are popular and therefore rising. Whereas momentum favors the long investor, it handicaps the short seller.
Shorting is expensive
Making matters worse, shorting stocks is costly. One must borrow shares to short them, and this costs money (borrow fees plus interest). Popular shorts (especially over-hyped or meme stocks) become "hard to borrow," with fees spiking to 50% or even higher on an annualized basis.
Brokers can also recall shares suddenly, forcing one to cover at the worst possible time. Dividends and corporate actions frequently add to the cost of shorting. If the stock pays a dividend while you're short, you pay it to the lender. If the stock price does not drop sufficiently enough to offset the dividend payment, as invariably occurs, short sellers will be out of pocket.
These structural and practical factors explain why empirical studies show that short-side strategies have a lower win rate than long-side ones. Profits come in sharp bursts (crashes), but drawdowns are frequent and painful.
Frequent losses mean that the psychological pressure associated with short selling is intense. This pressure is intensified by the perception that short selling is seen as anti-social. The market celebrates rallies but vilifies shorts. Short sellers are frequently seen as attacking companies and come under fire themselves with the threat of lawsuits.
Psychological pressure heightens the risk of cognitive dissonance and poor decision making. For this reason, short sellers typically require greater conviction in their trades than long-only investors. They undertake exhaustive research, with reports often running to hundreds of pages, to ensure they have strong conviction in their shorts.
They publish their research for free in an attempt to encourage weakness in the stocks they have shorted but of course this raises the ire of both the company and investors who are long the stock. Short-selling is a lonely road.
In short, deep, fundamental research can prove a thesis, but it fundamentally fails at the most critical variable: timing. For short sellers to survive the irrational market, a different, more scalable approach is required.
The case for forensic AI
No matter how exhaustive the research, no matter how many people read the research, a short seller does not control the market. As such, fundamental research effort helps little with timing. Since the timing window on short trades typically remains narrow, this is a low probability approach.
In fact, the more exhaustive the research, the fewer trades a short seller will identify. This means that beyond a certain point there will be diminishing returns on deep-dive forensic analysis as increasing research intensity reduces the ability to spread risk across trades.
Somehow investors need a more extensive list of suspect companies coupled with the necessary tools to pinpoint those exhibiting warning signs of an imminent price breakdown. The research effort can then focus on those opportunities.
Until the advent of AI, this kind of trading approach was a tall order. There were simply not enough hours in the day to sift through thousands of companies to identify the often overlooked signs of financial vulnerability that lie deep within a company's financial statements. Now, with the advent of forensic AI such as that provided by Transparently.AI, this information is instantly available.
AI-powered stock screening
Transparently.AI's proprietary accounting risk scores are designed to surface these hidden vulnerabilities, providing short sellers with a critical screening edge.
For example, a company exhibiting heavy revenue smoothing - a red flag deep within the accruals - might instantly flag a 90% risk score, compelling the investor to investigate that single trade before hundreds of others.
Transparently.AI's model delves into the intricate details of financial reporting, analyzing hundreds of factors to assign a risk score (from 0% for high-quality accounts to 100% for extreme risk) and a corresponding rating (A+ to F). These scores are not reactive; they are predictive, designed to identify accounting manipulation risk that often precedes significant underperformance and adverse events.
The true power of this approach lies in its ability to act as an early warning system. By scrutinizing accounting practices, our model can detect subtle red flags that might otherwise go unnoticed until it's too late.
Using Luca, Transparently’s GenAI investigator, short sellers can further refine the screening process, allowing the short seller to filter the initial high-risk list down to those exhibiting the specific type of fraud or vulnerability that best matches the trader's expertise: be it fraud, valuation arbitrage, disruption due to technological change or shifting consumer habits, balance sheet fragility or flawed management /governance.
Short sellers can then screen high-risk companies for emerging price weakness. Once the forensic AI and the price screening is complete, the short-seller can then apply their particular skills to go deeper and uncover a greater number of shorts that appear ripe for price declines.
Conclusion
For short sellers, accounting quality is not just another metric; it's a foundational signal that can unlock significant opportunities. This is true regardless of shorting style.
Transparently.AI's accounting risk scores offer a powerful tool to identify companies with vulnerable accounting practices well in advance of public disclosures or market reactions. By focusing on red flags like smoothing, accruals, and valuation gaps, our model provides the foresight needed to position a short book effectively, capitalizing on the inevitable market correction when the time is right.
Accounting irregularities always come to light but timing is critical and the risk on timing means that proper diversification is key. This proactive approach allows investors to move beyond reactive analysis, transforming hidden accounting risks into tangible investment advantages.
