Most of us don’t read 500 pages of annual reports every day looking for potential stocks. Rather, we rely on the works of other investors and brokerage reports to narrow down the opportunities. They would screen, analyze, and collate information into a simple and easy to understand narrative. This saves up our time. Instead of spending countless hours doing our own research, we can focus on the critical part of investing: Buying and making money. This is a win-win for everyone. But is it so?
What Influence Us
Most research comes with opinions that can influence our judgment on the attractiveness of an opportunity. If you ask a person to pick a random number between 1 to 100 and subsequently ask him to estimate the price of a stock, the price he comes up is likely to stay close to the number he randomly picks. This is called anchoring bias. A cognitive bias that causes us to rely on a piece of information when making a decision. This bias is at work even if we are aware that 2 events are unrelated. Such as picking a random number and estimating the stock price. When someone recommended a stock and provide all the reasons, you are likely to come to a similar conclusion under the impression that you’ve made the decision without any outside influence. But nothing could be further from the truth.
The market price serves as another potent influence. We are more likely to believe a research as accurate and trustworthy when the share price moves in a direction that agrees with the opinion. This is almost always the case considering most research are written after a company produces a set of outstanding results. Naturally, when a stock is trending up, we want to know the reasons. These research tell us why this is the case. The logic goes like this: The share price is going up because the business is doing well, therefore this information must be true. The price becomes a source of influence on how we judge the quality of the research. We find it hard to disagree when all the evidence is telling us buying is the right decision. This is where we come to the crux of the matter: information asymmetry.
If 10 out of 100 people that analyze a stock think it is a great opportunity, 5 will probably come out and share their opinion. What you don’t see are those 90 people that consider it as a bad investment. No one writes about a stock that they’re not going to buy. There is no incentive to do so. No one is going to read it. You only tell someone about a stock when it has a potential to make money. So what looks like a stock that has a unanimous 100% buy recommendation (5 out of 5), in actual fact only has a 10% consensus (10 out of 100 recommends a buy) when the entire group is taken into account. The opinions of those 90 people matter more but we never see them. They are silent evidence. That’s why there’s little surprise that there is more recommendation to buy than to sell. An opportunity appeals to a wider audience than a non-opportunity. There are always more investors that don’t own a particular stock than those that do. You are more likely to find information that tells you why you should buy than why you shouldn’t buy. This information asymmetry causes us to overestimate the reward and underestimate the actual risk. And if you connect the dots, this is how we become overconfident. The problem goes beyond these.
Valuation — There isn’t a right or wrong way to value a stock. All valuation tools work under the right circumstances. The problem arises when the usability of a tool is stretched to its extreme. Imagine using a hammer to chop a tree. Take the most common valuation tool: P/E multiple. You estimate how much a company should worth by assigning a fair multiple to what it earns now or likely to earn in the future. If a company has an earning of $10 mil, a P/E multiple of 10 will value the stock at $100 mil. This valuation tool has good predictive power when applied to stable companies that enjoy good economics. But it is more commonly used in low P/E companies with poor economics. People confuse low P/E stocks are undervalued with undervalued stocks have low P/E.
Not all low P/E stocks are undervalued, and therefore, deserve a higher P/E multiple just because there is a temporary spurt of earnings growth. Put it another way, multiples are relative; a company with strong economics that sells at 20 P/E can be undervalued just as a company with poor economics that sells at 5 P/E can be overvalued. It all comes down to economics; the long-term return of a business under a normal condition. If I tell you I ran 10 km yesterday, am I a good runner? You can’t tell. There’s nothing to measure against. You need to find out the time I took to complete that. There is a big difference between doing that in 1 hour and 10 hours. Now, replace distance with earnings. Similarly, earnings don’t tell you a thing about the quality of a business. A business that produces $20 mil. in earnings with $100 mil. capital has a different quality than one that produces $1 bil. using $10 bil. of capital. Quality dictates value; value dictates how much a business is worth. Mindlessly assigning an arbitrary P/E multiple while ignoring the long-term return of the business is where most get into trouble.
Macroeconomics — Macroeconomic factors such as interest rate, exchange rate, commodity price etc can have a significant impact on a company’s profitability. So, it is in our interest to predict them. But their complexity makes them highly unpredictable. Many variables are constantly interacting with one another to create an outcome that is nonlinear. Think butterfly effect; a butterfly flapping its wing in New Mexico causes a hurricane in China. So there is a very high chance of getting them wrong when we’re not wired to think in a nonlinear way.
Many will tell you what is their prediction of these macro factors. But what they didn’t tell you is their error rate. It is more important to know the error rate of a prediction than the prediction itself. Error rate tells you the accuracy of a prediction. If a prediction has a 50% error rate, that means the chance of predicting it correctly is the same as flipping a coin. And in the case of macroeconomics, it is normal to have a 30 to 40% prediction error rate going a year out and close to 50% beyond that timeframe.
Now, if someone wants to predict the steel price in order to estimate the earnings of a steel company, then what is the chance of getting both correct? If the probability of being correct on the steel price and the earnings are 60% and 80% respectively (or error rate of 40% and 20%), then the chances of getting both correct are only 48% (0.6×0.8). This is not encouraging. If we push this further, the chances drop precipitously. If a stock requires you to predict 4 macro factors correctly in order to make money and each factor carries a 40% error rate, the success rate is a paltry 13%. But the strange thing is this. The more macroeconomic charts that get thrown into an analysis, the more confident we are about the opportunity. This brings us into persuasiveness.
Persuasiveness — We tend to overestimate an opportunity if the analysis is persuasive regardless of the quality of the evidence. What makes a message persuasive? When it provides more information such as the case with macroeconomic charts. Give someone 5 pieces of information about a company and ask him to make a prediction. Repeat this exercise with 10, 20 and 30 pieces of information and his confidence will grow in every subsequent round without any sharp increase in accuracy. Our poor ability in weighing the quality of information means we tend to overweight evidence that has little relevance.
As a thought experiment, imagine you’ve just read an analysis about a stock. To give you a better understanding of the company, I added a section ‘Background of the Founders’ that chronicle how the founders started the business. The main founder grew up in a poor single-parent family as the eldest of the 7 siblings. Due to hardship, he was forced to work at a young age in a factory collecting scrap plastics to support his family. Things only got worse when his ailing mother passes away. But over those years, through sheer tenacity, the siblings manage to scrape by. Soon, they begin to see light at the end of the tunnel. Bringing with him the experience from the days toiling in the factory, he opened up a mom and pop store together with a few of his siblings. However, it would take many more years of hardships, battling waves of crisis before the business begins to stabilize. At present, the business is on track to hit a major milestone by opening its 100th store next month.
Now, this is a personal story of hardship and perseverance. They are factual information. You are more likely to believe the company will do well after knowing their personal stories compared to before reading them. But does this story tell you how the business is likely to do in the future? Probably not. There can be just as many entrepreneurs with a similar background that goes bankrupt as those that succeed. This is silent evidence again. You won’t find a news that writes about how a founder with a poor background ruins his business.
Persuasiveness is the reason why a 10 pages research looks more convincing than a 1-page research even if the extra information has little relevance to the analysis. Investment track record and maxim falls under this category as well. If I tell you my past 5 predictions have all resulted in 100% gain and throw in a famous quote ‘Be greedy when others are fearful’, my words suddenly sound more convincing to your ears. But does my track record and the quote have anything to do with the accuracy of the prediction and the quality of my decision? No. A quote can be used to justify any opinion. All they do is create good stories. Good stories are memorable. What’s memorable are also information that supports our belief.
Just as it is easy for us to overweight irrelevant information, our mind finds it easy to remember information that confirms with what we believe and forgets those that contradict. Confirmation bias means supporting evidence are more likely to be presented. Making the analysis more convincing than it should be. The most persuasive analysis is one that has little to no conflicting information. In another word, one that doesn’t tell you the risk. Information asymmetry as a result of silent evidence.
Straight line prediction — We tend to extrapolate from a small quantity of information. You are more likely to believe my next prediction given that my past 5 predictions are correct. This is called the law of small numbers. A generalization fallacy where a small sample size creates a biased outcome. This happens in the earlier example. 5 people recommended a stock suddenly makes it a great opportunity.
Imagine trying to predict the trajectory of a missile. As you know, a missile has a similar trajectory as throwing a rock. It goes up before gravity gradually brings it back to the ground. Now, if you zoom in and look at a small section of the trajectory, it is going to look like a straight line instead of a curve. You are going to predict this missile will reach the moon instead of the ground.
We do the same thing with stocks. We extrapolate from a few quarters of strong earnings and conclude the company will continue to do well into the future. Since it is in our nature to seek out evidence to justify our opinion, we’ll certainly find information to fit what we see. Ironically, this is what most research do with competitions.
Competitions — Most research only talks about the competitions when they are used as a way justify a higher valuation. Such as using competitors’ P/E to rationalize a stock is undervalued. But how can an analysis be thorough without taking into account what the competitors are doing?
Think it this way. Pick your favorite basketball team. Or a soccer team if you like. Then estimate its chances of becoming the champion this season. You probably going to estimate a 70–90% chance of winning since it is your favorite team. Maybe even higher. I will repeat this exercise with others so all the teams within the league are covered. When I add those chances together, they will almost always exceed 100%. But that can’t be right. There can only be one champion. And the probability of all the teams can only be 100%. A research always overestimates the success rate of a company if it fails to take the competitors into consideration.
The lucidity of information is also at play here. You only hear about how the company is going to succeed, but not what the competitions are going to do in response. Things that don’t get mentioned don’t get remembered. Means you’ll fail to consider the probability of adverse events happening.
Insignificant variables — A research can give you a list of reasons to buy, citing favorable exchange rate, tax benefits, previous one-off surcharge etc. But how significant are these things? How much as a percentage do they contribute to the entire earnings?
All these factors create an illusion that we think we understand the business, industry, and opportunity when in fact we have little knowledge and even less independent judgment than we think we do. Our decision is heavily influenced by the information that is provided to us even after our own research. What’s worse is when most contradictory information rarely shows up. This missing information causes us to become overconfident and overestimate the opportunity at hand.
A Critical Mind
Most analyses are there to persuade you to buy or at least, to convince you to agree with their opinion. There is no right or wrong. But the question is how do you differentiate between a real opportunity and one that looks real but risky? How can you tell a good analysis from bad ones? We tend to judge if an analysis is sound by comparing it to the share price movement. But that can’t be right. If the price goes up next month but down again the month after, what do you make out of that? Is that a good or bad analysis?
The only way to judge the quality of an analysis is to find out the thought process behind the decision. If I correctly predicted 10 stocks in the past, then you are likely to believe my next prediction. But if you ask how did I predict them, and I explain that I search for my lucky star every night — if it’s visible under the dark skies, that’s a buy signal. Now, you’ll probably take my prediction with a grain of salt. Knowing that my chance of correctly predicting the next winning stock is no better than flipping a coin because my process is grounded on unproven theory despite my stellar track record.
Of course, things are not as clear-cut as this in real life. But the key is, finding out how someone makes a decision will tell you more about the quality of the prediction than looking at the outcome. In investing, we can be right for the wrong reasons. Hence, it is easy to be fooled by the outcome as a result of luck. Finding out the how takes you closer to the truth than the what. And you can tell a lot about a person’s thought process and reasoning skill from his research.
Reading someone’s work should be a conversation or an argument between you and the writer. There is no expert and authority in investing. All opinions shouldn’t be taken at face value. Always clarify and ask questions. What assumptions is he making? Read between the lines. What is he trying to say? Be careful of silent evidence. What are the things that should show up but didn’t? Use the list above as a checklist. The more they show up, the more you should be skeptical about the prediction. Quantify things. If someone writes ‘the company has unlimited potential…’, what does ‘unlimited’ means? Can we put a number on that? As far as the universe is concerned, everything is finite. Never hold a fixed view. All predictions are probabilistic. This mean develops more than one hypothesis to explain things. If someone tells you a stock is undervalued because it is going up, while that could possibly be true, it can also mean there’s a euphoria; a self-fulfilling prophecy; or simply a random movement without any cause. In essence, find ways to falsify a hypothesis instead of confirming it. A hypothesis that can be easily falsified has little to no value. You have to have a critical mind when examining someone’s work to avoid getting into risky opportunities.
Just to be clear, we are not out to identify who is good or bad. Most investors have a genuine passion for investing. Sharing their works should be applauded. But it is our duty to make good decisions. And that comes from thinking well and reduce mistakes. You have to have filters in your mind to avoid doing stupid things. As opposed to making money, it is not losing money that enable compounding to work. Knowing when not to read is as important as what you choose to read.
What Influence Us
Most research comes with opinions that can influence our judgment on the attractiveness of an opportunity. If you ask a person to pick a random number between 1 to 100 and subsequently ask him to estimate the price of a stock, the price he comes up is likely to stay close to the number he randomly picks. This is called anchoring bias. A cognitive bias that causes us to rely on a piece of information when making a decision. This bias is at work even if we are aware that 2 events are unrelated. Such as picking a random number and estimating the stock price. When someone recommended a stock and provide all the reasons, you are likely to come to a similar conclusion under the impression that you’ve made the decision without any outside influence. But nothing could be further from the truth.
The market price serves as another potent influence. We are more likely to believe a research as accurate and trustworthy when the share price moves in a direction that agrees with the opinion. This is almost always the case considering most research are written after a company produces a set of outstanding results. Naturally, when a stock is trending up, we want to know the reasons. These research tell us why this is the case. The logic goes like this: The share price is going up because the business is doing well, therefore this information must be true. The price becomes a source of influence on how we judge the quality of the research. We find it hard to disagree when all the evidence is telling us buying is the right decision. This is where we come to the crux of the matter: information asymmetry.
If 10 out of 100 people that analyze a stock think it is a great opportunity, 5 will probably come out and share their opinion. What you don’t see are those 90 people that consider it as a bad investment. No one writes about a stock that they’re not going to buy. There is no incentive to do so. No one is going to read it. You only tell someone about a stock when it has a potential to make money. So what looks like a stock that has a unanimous 100% buy recommendation (5 out of 5), in actual fact only has a 10% consensus (10 out of 100 recommends a buy) when the entire group is taken into account. The opinions of those 90 people matter more but we never see them. They are silent evidence. That’s why there’s little surprise that there is more recommendation to buy than to sell. An opportunity appeals to a wider audience than a non-opportunity. There are always more investors that don’t own a particular stock than those that do. You are more likely to find information that tells you why you should buy than why you shouldn’t buy. This information asymmetry causes us to overestimate the reward and underestimate the actual risk. And if you connect the dots, this is how we become overconfident. The problem goes beyond these.
Valuation — There isn’t a right or wrong way to value a stock. All valuation tools work under the right circumstances. The problem arises when the usability of a tool is stretched to its extreme. Imagine using a hammer to chop a tree. Take the most common valuation tool: P/E multiple. You estimate how much a company should worth by assigning a fair multiple to what it earns now or likely to earn in the future. If a company has an earning of $10 mil, a P/E multiple of 10 will value the stock at $100 mil. This valuation tool has good predictive power when applied to stable companies that enjoy good economics. But it is more commonly used in low P/E companies with poor economics. People confuse low P/E stocks are undervalued with undervalued stocks have low P/E.
Not all low P/E stocks are undervalued, and therefore, deserve a higher P/E multiple just because there is a temporary spurt of earnings growth. Put it another way, multiples are relative; a company with strong economics that sells at 20 P/E can be undervalued just as a company with poor economics that sells at 5 P/E can be overvalued. It all comes down to economics; the long-term return of a business under a normal condition. If I tell you I ran 10 km yesterday, am I a good runner? You can’t tell. There’s nothing to measure against. You need to find out the time I took to complete that. There is a big difference between doing that in 1 hour and 10 hours. Now, replace distance with earnings. Similarly, earnings don’t tell you a thing about the quality of a business. A business that produces $20 mil. in earnings with $100 mil. capital has a different quality than one that produces $1 bil. using $10 bil. of capital. Quality dictates value; value dictates how much a business is worth. Mindlessly assigning an arbitrary P/E multiple while ignoring the long-term return of the business is where most get into trouble.
Macroeconomics — Macroeconomic factors such as interest rate, exchange rate, commodity price etc can have a significant impact on a company’s profitability. So, it is in our interest to predict them. But their complexity makes them highly unpredictable. Many variables are constantly interacting with one another to create an outcome that is nonlinear. Think butterfly effect; a butterfly flapping its wing in New Mexico causes a hurricane in China. So there is a very high chance of getting them wrong when we’re not wired to think in a nonlinear way.
Many will tell you what is their prediction of these macro factors. But what they didn’t tell you is their error rate. It is more important to know the error rate of a prediction than the prediction itself. Error rate tells you the accuracy of a prediction. If a prediction has a 50% error rate, that means the chance of predicting it correctly is the same as flipping a coin. And in the case of macroeconomics, it is normal to have a 30 to 40% prediction error rate going a year out and close to 50% beyond that timeframe.
Now, if someone wants to predict the steel price in order to estimate the earnings of a steel company, then what is the chance of getting both correct? If the probability of being correct on the steel price and the earnings are 60% and 80% respectively (or error rate of 40% and 20%), then the chances of getting both correct are only 48% (0.6×0.8). This is not encouraging. If we push this further, the chances drop precipitously. If a stock requires you to predict 4 macro factors correctly in order to make money and each factor carries a 40% error rate, the success rate is a paltry 13%. But the strange thing is this. The more macroeconomic charts that get thrown into an analysis, the more confident we are about the opportunity. This brings us into persuasiveness.
Persuasiveness — We tend to overestimate an opportunity if the analysis is persuasive regardless of the quality of the evidence. What makes a message persuasive? When it provides more information such as the case with macroeconomic charts. Give someone 5 pieces of information about a company and ask him to make a prediction. Repeat this exercise with 10, 20 and 30 pieces of information and his confidence will grow in every subsequent round without any sharp increase in accuracy. Our poor ability in weighing the quality of information means we tend to overweight evidence that has little relevance.
As a thought experiment, imagine you’ve just read an analysis about a stock. To give you a better understanding of the company, I added a section ‘Background of the Founders’ that chronicle how the founders started the business. The main founder grew up in a poor single-parent family as the eldest of the 7 siblings. Due to hardship, he was forced to work at a young age in a factory collecting scrap plastics to support his family. Things only got worse when his ailing mother passes away. But over those years, through sheer tenacity, the siblings manage to scrape by. Soon, they begin to see light at the end of the tunnel. Bringing with him the experience from the days toiling in the factory, he opened up a mom and pop store together with a few of his siblings. However, it would take many more years of hardships, battling waves of crisis before the business begins to stabilize. At present, the business is on track to hit a major milestone by opening its 100th store next month.
Now, this is a personal story of hardship and perseverance. They are factual information. You are more likely to believe the company will do well after knowing their personal stories compared to before reading them. But does this story tell you how the business is likely to do in the future? Probably not. There can be just as many entrepreneurs with a similar background that goes bankrupt as those that succeed. This is silent evidence again. You won’t find a news that writes about how a founder with a poor background ruins his business.
Persuasiveness is the reason why a 10 pages research looks more convincing than a 1-page research even if the extra information has little relevance to the analysis. Investment track record and maxim falls under this category as well. If I tell you my past 5 predictions have all resulted in 100% gain and throw in a famous quote ‘Be greedy when others are fearful’, my words suddenly sound more convincing to your ears. But does my track record and the quote have anything to do with the accuracy of the prediction and the quality of my decision? No. A quote can be used to justify any opinion. All they do is create good stories. Good stories are memorable. What’s memorable are also information that supports our belief.
Just as it is easy for us to overweight irrelevant information, our mind finds it easy to remember information that confirms with what we believe and forgets those that contradict. Confirmation bias means supporting evidence are more likely to be presented. Making the analysis more convincing than it should be. The most persuasive analysis is one that has little to no conflicting information. In another word, one that doesn’t tell you the risk. Information asymmetry as a result of silent evidence.
Straight line prediction — We tend to extrapolate from a small quantity of information. You are more likely to believe my next prediction given that my past 5 predictions are correct. This is called the law of small numbers. A generalization fallacy where a small sample size creates a biased outcome. This happens in the earlier example. 5 people recommended a stock suddenly makes it a great opportunity.
Imagine trying to predict the trajectory of a missile. As you know, a missile has a similar trajectory as throwing a rock. It goes up before gravity gradually brings it back to the ground. Now, if you zoom in and look at a small section of the trajectory, it is going to look like a straight line instead of a curve. You are going to predict this missile will reach the moon instead of the ground.
We do the same thing with stocks. We extrapolate from a few quarters of strong earnings and conclude the company will continue to do well into the future. Since it is in our nature to seek out evidence to justify our opinion, we’ll certainly find information to fit what we see. Ironically, this is what most research do with competitions.
Competitions — Most research only talks about the competitions when they are used as a way justify a higher valuation. Such as using competitors’ P/E to rationalize a stock is undervalued. But how can an analysis be thorough without taking into account what the competitors are doing?
Think it this way. Pick your favorite basketball team. Or a soccer team if you like. Then estimate its chances of becoming the champion this season. You probably going to estimate a 70–90% chance of winning since it is your favorite team. Maybe even higher. I will repeat this exercise with others so all the teams within the league are covered. When I add those chances together, they will almost always exceed 100%. But that can’t be right. There can only be one champion. And the probability of all the teams can only be 100%. A research always overestimates the success rate of a company if it fails to take the competitors into consideration.
The lucidity of information is also at play here. You only hear about how the company is going to succeed, but not what the competitions are going to do in response. Things that don’t get mentioned don’t get remembered. Means you’ll fail to consider the probability of adverse events happening.
Insignificant variables — A research can give you a list of reasons to buy, citing favorable exchange rate, tax benefits, previous one-off surcharge etc. But how significant are these things? How much as a percentage do they contribute to the entire earnings?
All these factors create an illusion that we think we understand the business, industry, and opportunity when in fact we have little knowledge and even less independent judgment than we think we do. Our decision is heavily influenced by the information that is provided to us even after our own research. What’s worse is when most contradictory information rarely shows up. This missing information causes us to become overconfident and overestimate the opportunity at hand.
A Critical Mind
Most analyses are there to persuade you to buy or at least, to convince you to agree with their opinion. There is no right or wrong. But the question is how do you differentiate between a real opportunity and one that looks real but risky? How can you tell a good analysis from bad ones? We tend to judge if an analysis is sound by comparing it to the share price movement. But that can’t be right. If the price goes up next month but down again the month after, what do you make out of that? Is that a good or bad analysis?
The only way to judge the quality of an analysis is to find out the thought process behind the decision. If I correctly predicted 10 stocks in the past, then you are likely to believe my next prediction. But if you ask how did I predict them, and I explain that I search for my lucky star every night — if it’s visible under the dark skies, that’s a buy signal. Now, you’ll probably take my prediction with a grain of salt. Knowing that my chance of correctly predicting the next winning stock is no better than flipping a coin because my process is grounded on unproven theory despite my stellar track record.
Of course, things are not as clear-cut as this in real life. But the key is, finding out how someone makes a decision will tell you more about the quality of the prediction than looking at the outcome. In investing, we can be right for the wrong reasons. Hence, it is easy to be fooled by the outcome as a result of luck. Finding out the how takes you closer to the truth than the what. And you can tell a lot about a person’s thought process and reasoning skill from his research.
Reading someone’s work should be a conversation or an argument between you and the writer. There is no expert and authority in investing. All opinions shouldn’t be taken at face value. Always clarify and ask questions. What assumptions is he making? Read between the lines. What is he trying to say? Be careful of silent evidence. What are the things that should show up but didn’t? Use the list above as a checklist. The more they show up, the more you should be skeptical about the prediction. Quantify things. If someone writes ‘the company has unlimited potential…’, what does ‘unlimited’ means? Can we put a number on that? As far as the universe is concerned, everything is finite. Never hold a fixed view. All predictions are probabilistic. This mean develops more than one hypothesis to explain things. If someone tells you a stock is undervalued because it is going up, while that could possibly be true, it can also mean there’s a euphoria; a self-fulfilling prophecy; or simply a random movement without any cause. In essence, find ways to falsify a hypothesis instead of confirming it. A hypothesis that can be easily falsified has little to no value. You have to have a critical mind when examining someone’s work to avoid getting into risky opportunities.
Just to be clear, we are not out to identify who is good or bad. Most investors have a genuine passion for investing. Sharing their works should be applauded. But it is our duty to make good decisions. And that comes from thinking well and reduce mistakes. You have to have filters in your mind to avoid doing stupid things. As opposed to making money, it is not losing money that enable compounding to work. Knowing when not to read is as important as what you choose to read.
