Our investment process is founded on the psychology of error. Our core beliefs are twofold. First, it is much easier to be wrong than most smart people like to believe. Second, smart people are not very good at dealing with their errors. So, what to do about our own mistakes?
We would love to believe we are impervious to bias, and faultless in our interpretation of the behaviour of CEOs, analysts, and investors. But the harsh truth is, we are error-prone too, and occasionally those errors overwhelm us. 2021 is looking like one of those years when we are not quite as smart as we would like to believe. This quarterly review is an attempt to dig a little deeper into our sources of underperformance.
Whenever we do a deep review of our past decisions we are reminded of the brutal effectiveness of the Hit-Rate, Reward-Damage Framework. The ability of a stock to deliver surprisingly good or disappointingly bad news is an amazingly powerful driver of stock performance. A clinical way of deciding what “good news” looks like is to watch analyst behaviour, and in particular their errors. If a stock does something to force analysts to increase their forecasts it tends to go up; if it prompts downward pressure on forecasts, it tends to fall.
This isn’t the only driver of stock prices (delivering consistently high sales growth also helps), but it is very important. We know from experience there are peculiar conditions when this basic driver gets temporarily drowned out (more on this later), but the macro noise about recession, recovery, inflation, and interest rates can only last for so long.
Most of the time CEOs are not very good at fulfilling their promises. They tend to lead analysts to make overly optimistic forecasts, that are generally missed. Since we have been studying analysts’ error, the chances your average company will consistently beat forecasts has been low: around one-in-three. Being good at finding stocks with a better than one-in-three chance of surprising tends to be rewarding. You can view stock picking as a simple game of improving the odds of finding stocks that will surprise.
But this isn’t the whole story. When you are dealing with uncertainty you must also think about your Reward-Damage function: the balance between the rewards of being correct vs the damage done by being wrong.
We view a portfolio as a collection of bets on the future of surprising growth. Some will turn out to be clever calls (they persistently surprise), some won’t. The ideal investment approach has a better-than-average hit rate, with higher-than-average rewards for the stocks which surprise and lower-than-average loss from the “misses”. But investment nirvana is not so easy to achieve.
There can be many ways analysts (and CEOs) can be surprised. Lots of genuinely unpredictable things happen; the last 18 months have been a clear reminder on the inherent unpredictability of the world. However, we believe not all surprises are completely unpredictable. We think we can find some surprises that are predictable from the past behaviour of others.
We prioritise how we approach the problem. First, we focus on the “hit rate” – we look for stocks with a better-than-average chance of surprising. Second, we think about “damage” – we try to avoid stocks where the damage from being wrong looks unusually high. Third, we think about “reward” – we want to get well rewarded for being right. We have tended to be good at the hit rate, not bad at limiting damage, and we are willing to sacrifice some reward for the prospect of a higher hit rate or less potential damage.
Having talked up the Hit-Rate, Reward-Damage Framework I am now going to kick it (temporarily) down. There are two kinds of periods when it isn’t a very useful driver of relative stock price performance. Both of them have been in play over the last 21 months.
First, when almost everything is disappointing. This doesn’t happen very often. It takes something genuinely unpredictable and nasty to undermine the plans of almost every business, regardless of how unusual they are. COVID-19 was one of those rare beasts. In such an environment the damage function matters the most. Damage is related to how deeply a business can miss when it starts to miss. Qualities we have a natural liking for, unrelated to hit rate, become a lot more important – robust, nimble business models, with inherently good core profitability and conservative CEO behaviour. Despite collapsing hit rates our approach can still do OK, but it is a struggle. Our error rate goes up. Despite a lot of stress we were able to navigate last year’s shock reasonably well.
The second difficult environment is when almost everything is surprising. Such periods present us with a different type of problem. If everything is super, then nothing is. Most of the time a lot of companies disappoint: it’s not unusual for two thirds to pressure analysts into forecast downgrades. 2021 has been very different. The rate of disappointment has been very low. By Q3 less than 20% of the biggest 500 stocks in MSCI-ACWI were missing forecasts.
We checked our statistics for 2021. We have done even better (we got down to less than 10%)! But when everything is surprising, having an edge on finding surprises doesn’t count for much. Our troubles were two-fold. First our rewards, second our portfolio construction.
Despite being uncommonly good at finding stocks that have surprised, our rewards have lagged the market. We prefer businesses with the potential for long, persistent surprise. When they surprise, they often do it in smaller increments. Under normal conditions, when surprises are rare, this can be very rewarding. But when any old business can blow the lights out, it’s not so good. Our stocks have tended to surprise, but less than the average.
But most of our underperformance has come from the portfolio construction choices we have made. Two types of recovery stocks have performed very well in 2021: Financials and Energy. Until Q2 we had virtually no financials in the portfolio. We did add a few banks during Q2, and they have tended to do better than your average bank, but our low weighting in financials has been a drag. We have nothing in energy. We do not like the back end loaded risk associated with ESG. Energy stocks have performed very well (for perfectly understandable reasons). Finally, equally weighting our stocks has not helped. Our biggest single detractors to relative performance have come from stocks we own, like Microsoft and Google, where their market-cap weights have contributed significantly more to benchmark performance than our measly equal weights. We have had to contend with portfolio construction issues before, but we haven’t had a year when they have combined in such an unhelpful way at a time when finding surprises also hasn’t counted for much.
The mistakes we have made? There have been a few. We were too slow to react to the mounting regulatory risks in China. Our over-exposure to platform, growth businesses like Tencent, Alibaba, and Meituan was painful in Q3. We have accepted we were wrong, sold and moved on. Late-cycle COVID-19 recovery plays in leisure and travel have not participated in this vigorous recovery cycle enough and have disappointed. We had too many of them.
An unusually large number of stocks are still surprising, but we can sense change. The demand part of the COVID-19 shock has unwound a lot, at a surprisingly fast rate. A stock’s ability to surprise has shifted from the return of surprisingly strong demand to an ability to meet that demand. From Q2-2020 the rate of profit surprises moved in lock-step with sales surprises, indicating a demand led recovery. Since Q2-2021 the two have begun to diverge. Profit surprises, while still prevalent, are noticeably less frequent than sales surprises. This tells us cost pressures and supply chain complexities are starting to matter. Recovery plans are starting to get harder. It is beginning to feel like surprises may start to get a bit more unusual again.
This material is for distribution to Professional Clients only, as defined under the Financial Conduct Authority’s (“FCA”) conduct of business rules, and should not be relied upon by any other persons. Issued by Ardevora Asset Management LLP (authorised and regulated by the FCA). Registered office: 6 New Bridge Street, London EC4V 6AB. Registered in England No. OC351772. Past performance is not a guide to future performance. Care has been taken to ensure the accuracy of this document’s content, but no responsibility is accepted for any errors or omissions herein. The views expressed are our own and do not constitute investment or any other advice and are subject to change. In particular, we are not recommending or expressing an opinion on the merits of buying or selling any securities mentioned herein. Ardevora Asset Management LLP is authorised and regulated in the United Kingdom by the FCA, and operates in the United States as an “exempt reporting adviser” in reliance on the exemption in Section 203(m) of the US Investment Advisers Act of 1940.
You are now leaving the Ardevora website and entering a website not operated by us. Ardevora provides links to external websites that you may find helpful, however the provision of this link does not constitute a recommendation or advice.
Ardevora accepts no responsibility for the accuracy or content of the website that you are about to access, or for any action that you take as a result of accessing this website.
Please ensure that you have read and understand the Terms and Conditions of such third party website and contact the operator of the site should you have any queries.