Инвестор призывает к диверсификации портфеля на фоне рисков в секторе искусственного интеллекта

This note is about how to play the investment game given what's happening now. Imagine you're playing a game like bridge, poker, backgammon, or chess and have to make your move, and you have a computer that works with you to assess the circumstances and suggest a move. That's what playing the investment game is like for me. Whether or not you have a computer to help or not, I believe that you should: Ask yourself what move to make given how the board is configured (i.e., given the existing characteristics of the market and the influences on it are like). I have been playing the investing game for a long time, and my goal at this stage is to pass along how I'd play it — and beyond that to create a platform that all sorts of people can use to explore playing the investing game anyway they want to play it to learn, see how they would have done, and to do it well. I believe that there are right and wrong ways to deal with the hand that one has been dealt, so when encountering XYZ configuration of circumstances, you should ask yourself, "How should I bet given that configuration of circumstances?" and be able to come up with good answers. I now want to pass along what the existing characteristics of the market look like to me and what I think should be done (and am actually doing) in light of them. How to play the existing set of circumstances What are the most important circumstances today, and how should one place one's bets in light of them? It appears to me, and probably to most everyone else, that we are now in a market environment in which a very limited number of companies in a sector characterized by remarkable new technologies (mostly AI) are dominating the market action. These companies constitute a high percentage of the market cap and have a huge effect on the markets and economy. As with all such times, there is a great deal of excitement, uncertainty, and volatility concentrated in the new technology sector, which are being passed through to the stock markets around the world. So, the swings and uncertainties around this sector matter a lot. There are also other big uncertainties related to other big, important drivers, which are what I call the five big forces: 1) what's going on with debt and money, 2) what's going on with political and social issues that can have big effects on taxes (and other politically-driven influences on markets), 3) what's going on with geopolitical influences on the markets (such as the wars), 4) what's going on with acts of nature, and 5) what's going on with new technologies. I plug those circumstances into my investment system, which thinks about how to place my bets in light of them while I simultaneously do my own thinking about what to bet on. When thinking about how to place ones bets given these circumstances, the most important question to ask and answer is do you want to: a) bet on the new technology even more than is imbedded in the broader stock index (e.g., the S&P 500) by overweighting this new sector or a few companies that you think are the best ones in that sector, b) keep your exposures at about index weights, or c) diversify away from that concentration? While just about everyone wants to buy the best investments and strives hard to do that, there is this new technology that appears to be changing just about everything. History shows that most everyone has failed by putting a high percentage of their chips on the few stocks of the leading companies producing that technology at this stage in the cycle. That is for logical reasons that have always played out in the past. While this new AI technology is unique, there have been many new unique technologies that are analogous to look to as a guide. One should look at them and, if one chooses to ignore them, have good explanations for why this time is different. The risks are certainly high All the past cases of great new technologies have played out in the same ways for the same logical reasons. High risks with great uncertainties are intrinsic to these new technology companies. When we look at how they performed in the past under this configuration we see that even the best revolutionary new technology companies that prospered in the long run (e.g., Microsoft and Apple) got annihilated at similar times along the way, and that, at the time the new technology companies came along (rather than in retrospect), it wasn't easy to say which ones would succeed and which would fail, like IBM. If you look at all those cases, you can see that it is in the nature of remarkable new technology companies to have highly uncertain futures. For example, they either over-invest or under-invest. That's because they will certainly lose if they don't invest well enough to win, and they can't possibly know what's going to happen precisely enough to know whether they're over-investing. Over-investing or under-investing is costly. Also, they can't possibly accurately anticipate all the changes in all the things — including the exogenous ones like the tightening of money, wars, and big changes in taxes — that are going to affect them. So, they all go through big up-and-down waves that first excite and then scare the hell out of and shake out the weak investors, which leads to exaggerated market swings. Further, just as these new technologies and new technology companies were disruptors to those before them, most of them are eventually disrupted by newer technologies and newer technology companies in ways that are impossible to consider, so we should consider the risk that the same thing could happen to these new technologies and technology companies. The implications of quantum computing are some of the known knowns that one might consider. What about the ones that haven't yet been imagined? And what about the risks coming from competitors? For example, China is producing and distributing AI technologies, and Chinese policymakers have a whole different view of economics and AI. We are in a new technology war, and leaders believe that they must win. Their view of AI and its effects on the economy and people's well-being leads them to make it available free or inexpensively because the technology has such a productivity benefit and can raise living standards for the whole. As they see it, profits are less important than the holistic benefit that comes from many people using these new technologies. I would think they are going to compete in international markets as they do with cars, solar panels, batteries, and many other products. This configuration of circumstances looks a lot like that in many cases throughout history that provide lessons. I can't help but think about when the British outcompeted the Dutch in shipbuilding and other important industries at the end of the Dutch Empire and the beginning of the British Empire. Also, there is a geopolitical war going on related to Taiwan that should lead us to at least consider the possibility that China could prevent chips from getting out of Taiwan as a tool for geopolitical warfare. And there are other risks for AI stocks, like the rising risk of wealth taxes and other taxes that would require those who are holding a lot of their wealth in these stocks to have to sell them, or the risk that rising anti-AI sentiments could lead to constraints on companies' advancements. I can give you more things to worry about, and I can give you an equally long list of great opportunities that AI will produce that I want to bet on. I'm not saying how these risks will play out or that one shouldn't bet on AI companies. I am just saying that it is indisputable that there is a lot of concentrated risk in the market and that one should know how to play circumstances like these. From my study of all analogous cases and for logical reasons, I am confident that the risks are high and that the best way to play such a configuration of circumstances is to: *Diversify well As you probably know, my mantra is diversify, and my "holy grail of investing" is to *strive to have 15 good uncorrelated investments that are risk balanced. Or said another way: *A well-diversified portfolio of good bets will outperform (have a higher return to risk ratio that can be engineered to have a better return given the same amount of risk) a concentrated bet, and the more the level of risk is concentrated in one area of the market, the more one should diversify, especially if the markets are driven by a revolutionary new technology which inherently produces great uncertainties." That isn't an opinion; it is a mathematical certainty. For example, if I compare one bet that has a return to risk ratio of 0.3 (let's say 6% with an 18% standard deviation, which is what it's generally assumed to be for equities) and to having 5, 10, or 15 uncorrelated bets, I would have the same 6% return with risks as measured by standard deviations of 8%, 6% and 5% respectively. So, with 15 good uncorrelated investments, I'd improve my return-to-risk ratio by a factor of 4.3x (from a 0.3 ratio to a 1.29 ratio). If you choose, you could lever that up and get much higher returns with the same amount of risk. That's a fact. I confidently believe — because of my back-testing, because of my delivering the actual returns I delivered over my over 50 years of playing the investment game, and because of probable logic — that having excellent diversification of bets and gearing them to one's desired volatility will produce much better returns over time than having the concentrated bets that most investors are inclined to have. More specifically, by diversifying well, one can produce a much better risk-to-reward ratio than by having any concentrated bet, and by gearing that to one's desired level of risk, one can get a higher return with the desired level of risk than by following any other process. Because I am passing this approach along, it is my now "not-so-secret" way of achieving investment success. Still, it is extremely rare that I encounter investors who think about their investment strategy in this way—i.e., it is extremely rare that I encounter investors who think about portfolio construction (i.e., about how a well-structured diversified portfolio of bets would compare in performance against a concentrated position in stocks in companies in a great new transformative industry). Most are just thinking about whether those stocks and that industry are going to do well and how to bet on them. There is a huge difference between the performance results of those who do think about portfolio construction and those who don't. So, I will be sure to convey my more complete thoughts on how to do this well at another time. For all these reasons, it seems to me that reflecting on how to play one's hand with this configuration of circumstances should lead one to ask oneself how big of a concentrated bet one should have and then diversify. The returns look to be low While it is indisputable that the risks are high, I am now going to give you an opinion, which could be wrong, that the prospective returns are low. That assessment of prospective future returns is based on my analytical work related to valuations and my bubble indicator's readings: the real returns in equities over the next 5 to 10 years look to be about -5 to -10%, though there is considerable uncertainty around those numbers. It appears to me that these stocks are long-duration assets that are very risky because it is very difficult to see reliably far into the future, and they appear expensive and are in weak hands. A question from my research team on this topic In a recent meeting, one of the members of my research team asked me: "Why do you think the markets are wrong in being configured the way they are today? How do you know that the lack of diversification that exists in the market today doesn't exist for good reasons — i.e., because some investors think the expected returns on AI stocks will be very high, because such concentrations in the index happen naturally when one industry constitutes such a high percentage of the market capitalization, or because when there is a lot of enthusiasm for an industry, that leads a lot of investors to buy these stocks without doing smart and reliable calculations of what the future earnings will be and how these earnings should be priced in the stock price?" My answer Prices go up for all sorts of reasons, and they are not all good ones. Some investors think about the price and push prices higher because they think the prices are attractive relative to the fundamentals, some investors are long in these stocks because they recognize this as a great new technology and they see the prices of these stocks rising as confirmation that these are good stocks, and other investors have an index exposure which has put them in the position of having large weightings in these stocks. As I see it, you can wrestle with these sorts of questions in order to try to decide what you want to do, or you can recognize that you don't have to wrestle with this question because you simply don't know enough to confidently make a bet. *You can simply say, "I don't know enough to make a bet" and then not make a bet. What gets people into trouble is thinking that they have to form an opinion and that their opinion is worth something, when it's more likely that they can't form a reliable enough opinion to bet on. (Footnote: To be clear, I'm not recommending avoiding making bets — and besides, you can't avoid making bets because you have to put your money either into some investment or into cash, which most people think is the least risky investment but is surely the worst investment over time. I am recommending knowing how to diversify your bets well, even when you have no tactical views of which markets are good or bad. The way to do this is by having a well-balanced strategic asset allocation mix that you hold when you have no tactical views that you are confident to bet on. But that's a subject for another time.) So, I believe that *Knowing what you don't know in order to decide when not to make a bet is as important as knowing what you do know to make a bet. Said more simply, I believe in the following principle: *Since it is generally too difficult to know enough to justify concentrated bets, it is best to have a diversified portfolio of only one's most confident, uncorrelated bets and to engineer the portfolio to the desired risk level. That is my "holy grail of investing." At this moment, given the existing configuration of circumstances, I don't think anyone knows well enough what's going to happen in this current technology-driven market to place a big, concentrated bet. In my opinion, avoiding concentration and being diversified is the best way to deal with not knowing. I realize this is contrary to the theory you may read in a textbook, which essentially says the markets are efficient, so you should "just trust the market." In summary, the fact that we currently have an unusually concentrated market centered on a revolutionary new technology should lead us not to confuse our excitement about the new technology with the attractiveness of the new technology stocks and to throw caution to the wind and have a concentration of high-risk, high-correlation bets, especially when we can get equally attractive returns with much less risk through smart diversification. PS: While I won't share my positions or tactical views with you because I don't want to be your investment advisor, I will soon share some key perspectives behind them with you, including my bubble gauge readings and the logic behind them.