The wave of artificial intelligence is coming to find a value depression

□ Yu Shipeng, a reporter of this newspaper

The investment style of fund managers is largely the result of the evolution of personal career paths. In this regard, Zhang Kun, a new generation of technology players of Sino Analytica Fund and fund manager of Sino Analytica's preferred return mix, provided a representative "sample". Unlike most fund managers who come from industry researchers, Zhang Kun entered the investment research circle as a strategic analyst. This background makes Zhang Kun a full-market stock picker. However, under the polishing of many years of investment experience, Zhang Kun's investment cognition is also constantly updated, and now he tends to be explosive and emerging industry tracks.

How did his background in strategic analysis make him a market-wide stock picker? In the emerging industry track, how did he meet the two major fields of "artificial intelligence" and "Internet of Everything"? In response to these problems, a reporter from China Securities News recently conducted a face-to-face investment interview with Zhang Kun.

Solve two basic problems

China Securities News: Unlike many fund managers, you entered the investment research circle from a strategist, what role does strategic analysis play in the investment research system?

Zhang Kun: I have a more vivid analogy: the ups and downs of the market are like taking an elevator, and many people rise to a higher place during the bull market. Some people think it's push-ups, some people think it's jumping up by running, but in fact it's because the elevator went up, and the bear market is that the elevator came down. The main job of the strategist is to determine the direction of the "elevator", which is the basis of the modern investment research system.

Specifically, what the strategic analysis is to solve is the two basic problems of "market identification" and "industry price comparison". First of all, analysts will predict the market trend in the future period of time by grasping the macroeconomic trend and industry prosperity, as well as the cyclical changes in different industries and their position in the economy. On this basis, through factors such as liquidity and overall market valuation, the investment cost performance of each sector is judged, and asset allocation suggestions are given, such as over-allocation or under-allocation.

China Securities News: What are the market indicators on which the strategy analysis is based? Can you give a concrete example?

Zhang Kun: The indicators of strategic research seem to be various, but they can be roughly reduced to two categories: "economic" indicators and "monetary" indicators. In real economic operation, the "economy" indicator will lag behind the "currency" indicator, so the strategy research will pay more attention to the "currency" indicator than the "economy" indicator. The most important monetary indicator is the Federal Reserve's monetary policy, and indicators such as inflation levels and commodity prices are derived from this.

For example, we have clearly seen in this round of markets since the epidemic in 2020 that after the Fed's monetary easing, the valuation repair of stocks followed; when valuations rose to a certain stage, the macro economy began to recover, and then commodity prices began to rise, and inflation concerns followed. After that, expectations of tighter liquidity began to emerge.

China Securities News: What is the impact of the background of a strategist on your investment framework?

Zhang Kun: In the early days, I worked as a strategist in both securities companies and fund companies, and macroeconomic analytical thinking did have an impact on my later investment philosophy.

It is not difficult to find that the idea of strategic analysis and the management of the fund portfolio are consistent. The most important point is that the strategic analyst must not only make judgments on the investment cost performance of different tracks, but also have a clear forward-looking vision, which can generally cover the short-term and medium-term market range. Even, the predictions of particularly good strategists can see long-term trends over three years. It's just that the strategy analysis doesn't go deep into the individual stock allocation level.

Because of this, compared with specific industry analysts, strategic analysts naturally have a global vision, that is, from the macro trend and the general trend of the international economy to grasp the investment direction and seize the opportunities of the times. In a fertile land, any healthy seed can bear fruitful fruits. Then, picking a head in the fruitful fruit will naturally be much less difficult.

China Securities News: After changing from strategist to fund manager, how has your understanding of strategic analysis deepened?

Zhang Kun: As far as investment is concerned, a top-down framework is not enough.

Investors who have come from strategists naturally have a high sensitivity to market fluctuations. Correspondingly, when turning from research to investment, it is natural to try swing investment and rely on trend analysis to sell high and suck low. I tried it for a while, but in hindsight the investment didn't work well. On the one hand, that kind of investment is very difficult, you need to pay attention to the disk changes at all times, and the time input cost is very high; on the other hand, in hindsight, every step of the macro trend is clear. But standing at the moment, this kind of strategic analysis can only grasp a vague general direction, and the investment decisions made thereby are not the largest winning rate.

Layout of two major science and technology tracks

China Securities News: The high-growth track has been highly sought after by the market in recent years, and it is also the mainstream track focused on by the public offering, what subdivisions of your layout mainly focus on?

Zhang Kun: In recent years, under the background of economic transformation and upgrading, the state has continuously increased its efforts to support scientific and technological innovation, especially in chips, new energy, artificial intelligence and other tracks, which have become industries that focus on encouraging development. From the perspective of market switching, large funds chasing high-certainty growth tracks, especially small and medium-sized market capitalization or companies with market capitalization below the waist, will be the mainstream investment trend in the future for a period of time.

Among the high-growth tracks, I am most optimistic about the "artificial intelligence" and "Internet of Everything" tracks, and the current portfolio management also focuses on these two major areas. So far this year, these two tracks have not been the most dazzling sectors, and their main rise has not yet arrived. However, the best investment time for such technology companies is precisely when the industry penetration rate is low, such as layout in the 2% to 50% stage, which can capture the excess returns brought by rapid development.

China Securities News: When do you think the big explosion of these two tracks will come?

Zhang Kun: The rapid rise in industry penetration requires certain necessary conditions. Taking the artificial intelligence industry as an example, when the following three elements go hand in hand, it may be the time when the artificial intelligence industry breaks out. Similar to the Internet more than a decade ago, this outbreak will greatly change the ecology of all walks of life, and the investment value it brings is also huge:

First, the amount of data should be large enough. For artificial intelligence to train a mature set of algorithms, it must provide enough data. For example, using AI to do image recognition, to achieve high accuracy, the amount of data in machine learning should be more than a few million pictures. From the current point of view, we are already in a digital environment, whether it is a mobile phone or a computer, a large amount of data can be generated every day, and there are relatively fixed scenes (social, shopping, sports, rest, etc.), so the amount of data will not constitute an obstacle to the development of artificial intelligence, but the data in some segments, such as health management, needs to be further enriched. At present, relevant companies have used these data to develop personalized platforms to provide targeted business solutions and improve industry efficiency.

Second, the computing power should continue to increase. In the case of insufficient computing power, once the amount of data increases, the computer's memory is not enough. In recent years, the industry has continued to introduce high-speed GPU computing chips, which is a big step forward. With the gradual improvement of computing power, the data training efficiency of artificial intelligence will become higher and higher.

The third is to improve the level of deep learning algorithms. The theoretical basis of deep learning was proposed around 2004, before which the development of the biological community in the field of algorithms had stagnated for about two or three decades. The principle of the deep algorithm is similar to that of human brain neurons, which greatly improves the recognition of machines, especially the recognition ability of language and images, from the previous upper limit of 80% to 90%. In the future, continuous progress in the field of algorithms will become a key step in the outbreak of the artificial intelligence industry.

China Securities News: As of now, what is the development of the above related conditions in China? Which aspect of the investment trend is clearer?

Zhang Kun: As of now, there have been a number of platform-based companies in China, and they are opening up data to the outside world on their own platforms, and the development trend is relatively clear. The giants who can win in the future may be born from the current companies.

In the industrial chain of the Internet of Everything, the order from upstream to downstream is sensors, communication modules, and data platforms. From the perspective of investment, the front-end sensor or communication module is more elastic and the performance is more explosive; but the most certainty link is the data platform, because the competition barriers are high and have obvious scale effects.

"Up and down verification" improves the chances of investment success

Zhang Kun joined Sino Analytica as a fund manager assistant in May 2014, and after officially starting to manage the portfolio in August 2015, after becoming a fund manager to manage funds, Zhang Kun began to improve the investment research framework towards a bottom-up path.

China Securities News: Since 2015, the A-share market has undergone a new round of bull-bear conversion. How has this market experience affected you?

Zhang Kun: Since entering the market, I have experienced two rounds of big fluctuations in A-shares, one was the unilateral decline in the second half of 2015, and the other was the financial deleveraging bear market in 2018. Looking back now, it is only clear that making money in a bull market does not prove the effectiveness of an investment framework. Therefore, after managing the funds, my research gradually shifted in the direction of micro fundamentals, gradually forming a combination of top-down and bottom-up investment frameworks.

In my opinion, strategic analysis solves the "trend" problem, and fundamental analysis solves the "value" problem. The combination of the two, looking for a value depression in the general trend, is the problem that investment has to solve. Therefore, a confident investment decision is not only supported by macro analysis and micro perspective, but also can often be verified between macro analysis and micro perspective.

China Securities News: How does this idea of macro-micro mutual verification be verified in ordinary research?

Zhang Kun: For example, from the perspective of economic direction and industrial trends, an industry has a lot of room for development, at the same time, in the company's research, we can see that the company's orders continue to increase, and the company's quarterly performance growth rate continues to rise, so as to achieve "consistency from top to bottom", which can basically be judged to be a company worth investing in, such investment is also highly certain, and the probability of winning is relatively large.

On the contrary, if the macro analysis is not verified at the micro level, or the company's fundamentals lack obvious industrial prosperity support, it means that this is not a big market opportunity, or there are blind spots in the decision-making process that need to be further clarified.

At this time, it is necessary to make a decision review, whether the direction of the industry trend analysis is biased, or the company's fundamentals are not thoroughly studied. For example, if from the bottom up, the fundamentals of a company are good, but there is no effective logical support in the top-down dimension, then the company's performance growth is not necessarily sustainable, but only pulsed, phased, and investing in such a company must be unconscious.

Such examples are not uncommon in the market, and the smartphone market that occurred around 2014 is a good example. Since 2010, the trend of large-scale popularity of smart phones has slowly come out, and by 2014 we can see that smart phone sales are growing, and the orders of companies in the industrial chain continue to reach new highs, and finally a number of big bull stocks have emerged in this trend.

China Securities News: After you use the idea of "mutual verification between the top and bottom" to find out the "good company", how to see the "good price"? How to understand the valuation factors of individual stocks?

Zhang Kun: Indeed, valuation has no less impact on investment effectiveness than performance. However, the influencing factors of valuation are complex, in addition to the obvious profit growth rate, business model, entrepreneurial spirit, industry prosperity, industrial chain status and other factors also need to be comprehensively considered.

Specifically, the impact of earnings growth on valuation is short-term, while factors such as industry prosperity and industrial chain status are the main influencing factors of long-term valuation. In recent years, we have seen a leading food and beverage stock with an annual performance growth rate of 20%, but the market once gave a valuation level of 100 times. To a large extent, this is mainly due to the company's channel discourse and industry position. Emerging industries such as artificial intelligence and big data, many companies that lack profits have valuations of up to hundreds of times, which is also derived from their future development trends and industry prosperity.

This article originated from China Securities News