Good shorts have three main characteristics:
- We have a high conviction of where the price will end up
- Low volatility in the stock price from today until that point
- Large negative price movement in a short period of time
To achieve these goals, investors typically use one of the following models:
- The valuation approach: Predict revenues, margins and profits and free cash flow. Benjamin Graham taught us that if you are correct about your intrinsic value, the stock price should revert to this within 2-3 years. A good short is the opposite of what is generally considered a good long. The product the company produces should be commoditised, the industry should have low barriers to entry and the returns on reinvested capital low. The valuation should be high compared to either its replacement cost of assets or its long-term free cash flow calculated off normalised margins. Fundamentals to keep track of include the competitive offering as well as industry supply & demand patterns.
- The Steinhardt model: Front-run other fundamental investors by finding situations where you have a variant view to consensus of what is going to happen. The variant view is usually built on an information edge. When your forecast is proven correct, other investors will adjust their view of the valuation of assets and sell the overvalued stock. Steinhardt emphasizes the need to find the exact catalyst that will cause investors to change their minds so that the price corrects in a short period of time.
- The Soros model: Investors have an imperfect understanding of the world. An underlying trend in fundamentals can both affect investors biased view on it, and vice versa. This may cause a reflexive boom by which the price increasingly becomes disconnected from fundamentals. A short should not be initiated until the trend has been discredited, at which point the reflexive bubble will burst. Security prices tend to rise along with the country’s currency, which itself is a function of global money flows.
How the models stack up
Shorting a stock using a valuation approach is risky. First, the investor needs to make long-term forecasts that are better than the average investor. Some factors such as demographics and secular consumption trends may be easy to forecast. Others, such as corporate governance, or the risk that the company will be acquired, are not easily predictable. It is by no means certain that the stock price will end up close to the investor’s calculation of intrinsic value, even if he waits for 2-3 years. Over that time period, the investor will be bombarded with information and price movements that will cause much doubts about the validity of the thesis. If the company is in a commoditised industry, margins are likely to be volatile. Higher margins will cause less informed investors to buy the stock whenever margins rise enough to cause a low current PE ratio. In terms of the valuation of the asset, it is not entirely clear what discount rate to use in the valuation of an asset. Small changes in the discount rate can cause enormous changes in the valuation of a company with open-ended growth prospects. For example, PE ratios in India are almost twice as high as in Vietnam. Can such a disparity really be justified on the basis of expected growth rates or discount rates? If the investor uses the current market discount rate, he is making an implicit assumption that in my view warrants further analysis.
Steinhardt’s model also emphasizes the valuation of an asset. But since the exposure is likely to be short-term with a certain catalyst in mind, it does not matter that much if the valuation is off by say 20-30%. The direction is more important. If the investor is able to gauge whether market expectations for a stock are positive, he can be confident that the stock price will fall once fundamentals start to develop in a negative direction. A question is how an investor can figure out what expectations are built into the price. Should he rely on sell-side consensus figures? Or should he rely on anecdotes that he gleans from other investors when they discuss the stock? Most companies are complex and key variables move in different directions. And over time, complexity builds up. Shorting on variant perception should therefore be short-term and with a specific event in sight.
Soros model is vague. He makes the point that no-one knows what the future holds. The concept of valuation is only useful if the marginal buyer uses it as a criteria for his buy or sell decision. One should not short every stock that has disconnected from fundamentals, only if the false premise on which the trend is built on has been firmly discredited. It does not matter what you think will happen to the cash flows or the valuation. What matters is the fact that other investors are likely to change their minds about whether it is worth owning the asset. If the owners of the stock are primarily value investors, then it makes sense to forecast fundamentals and expected valuation ratios. If most investors are chasing momentum, then one should short the stock once it can be proven that it has lost its momentum. Another strength of the model is the recognition that investors’ own actions may change the fundamentals of the asset. Whether we talk about issuing shares to finance acquisitions that keep up the illusion of EPS growth or reflexivity in a macro sense, it all goes into the equation of whether to buy or short a certain stock.
Taking the best from each approach
It is simple, really: asset prices are determined by supply and demand. “Valuation” is just a concept we use to predict what the company might be worth in a private market transaction. A target price may also help us to manage the fear we experience when stock prices fall and buy more of it when the price is lower. But it would be naïve to think that a rational calculation of discounted cash flows to the firm is the basis for all stock price movements. Fundamentals are therefore only part of the puzzle, and for a good short we need to take into account both fundamentals and investors’ reaction to them.
The factors that may cause investors to buy a stock are:
- Momentum or “a good chart”
- An exciting story that captures the imagination of investors
- Support from the sell-side research
- Open-ended revenue growth thanks to a competitive product or a long-term consumption trend that takes a while to spread
- Expected one-off improvements in business fundamentals: a new product, new regulations, a planned restructuring of the businesses, etc.
- Margins improvements due to industry supply & demand, FX rates or interest rates
- Improvement in the balance sheet due strong cash flows
- Money supply growth or reflexive processes
If we want to minimise buying pressure and maximise selling pressure, then in line with Soros’ approach we should look for instances where: stock price momentum has stopped, an exciting story has been firmly refuted, sell-side has stopped covering the stock, it does not have the potential for open-ended revenue growth, margins are unlikely to go higher, the balance sheet is deteriorating and the money supply is contracting.
The best types of shorts
The ideal situation is when you can control the release of information that proves to the world that a company is a fraud. All the reasons for buying have then been discredited and investors will sell as soon as possible in order to salvage their investment.
Investments that less informed investors have bought on the basis of a low valuation multiple may provide a fertile field for shorts. If the reason for the low valuation is technological obsolescence or consumption trends working against the company, any exciting story will probably already have been discredited. A low valuation multiple is hence the only good reason for owning the stock. Once earnings are proven illusory, investors may be slow to react and the result will be a slow grind lower with little volatility. If earnings disappoint due to a build-up of accounting tricks that can no longer be maintained, investors are unlikely to react positively.
Shorting stocks with commoditised products where industry supply & demand determine the margins the company is allowed to earn can be dangerous. Industry dynamics are difficult to forecast, and the volatility may be more than most investors are able to stomach. Companies in cyclical industries are hence more difficult to short and it requires an enormous amount of confidence that fundamentals are really deteriorating. Risk-reward can be improved by focusing on stocks with a very high ratio of price-to-replacement cost such that even in the most optimistic scenario, the PE ratio is unlikely to look attractive. If industry supply-demand is likely to diverge in the near term, then all the better. A benefit of shorting stocks in cyclical industries is that commodity prices, or supply-and-demand can shift very rapidly and thus free up that money to be invested in other situations.
Shorting stocks due to a high valuation and despite open-ended growth does not appear to be sensible. It is difficult to estimate the discounted cash flows of an asset when most of the value comes from cash flows 10 years hence. And if it is difficult for you, how difficult is it for the average retail investor? The story needs to have been discredited – not only by you, but by the average investor. It is simply unrealistic to think that retail investors or even the average mutual fund manager will be as rational as a dedicated short seller. If it is difficult to tell exactly when growth will stop, the investor is likely to face a lot of volatility, even if he is ultimately proven correct. For open-ended growth stories, it is helpful if sell-side analysts has already given up on the stock and fundamentals have started to deteriorate, as evidenced by falling revenue and rising debt levels.
The risk-reward of shorting a stock becomes better when excess money supply growth turns negative, and investors are forced to sell stocks in order to deleverage. Large outflows from mutual funds or ETFs have a similar effect, although such events may be hard to predict in advance.
Note to self
- Do not rely on valuation alone when shorting stocks
- Only short stocks when a key reason for buying the stock has been firmly discredited or will be discredited in the near term
- Frauds are better than fads, which are better than cyclical shorts, which are better than valuation shorts with open-ended growth
- Falling revenues and rising debt is a great combination
- Don’t short stocks in the face of very high growth in excess money supply