Episode 3 – Artificial Intelligence

[THEME MUSIC]

Hello, everybody. Thank you for tuning in to <i>Market</i> <i>Perspectives,</i> a Mercer Advisors podcast, where we provide a data-driven, common sense perspective on markets, on the economy, and on investing. I’m your host, Don Calcagni, Chief Investment Officer at Mercer Advisors.

Today’s topic is on artificial intelligence and specifically whether or not artificial intelligence can help investors build better portfolios, portfolios that would be capable of outperforming the market and ideally to do so on a risk-adjusted basis.

So why is this important? Why now? Well, the reality is nowadays, you can’t read anything related to markets without getting a sense that AI is going to transform how we invest, that it’s going to transform how we view the world in new and unforeseen ways.

And to that end, over the past several weeks, we’ve actually observed a very significant run-up in AI-related stocks, so much so that according to T. Rowe Price, year to date through May 31, a basket of only 11 AI-related stocks has risen about 15% versus only about 9% for the S&amp;P 500 as a whole.

And in fact, according to the same research by T. Rowe Price, if we were to actually to remove those 11 AI stocks from the S&amp;P, what we would see is that the rest of the index is actually negative for the year by about 6% through the end of May 31.

Now, this has led many investors to consider overweighting AI stocks in their portfolios, perhaps even to invest exclusively in AI-related stocks, or to invest in ETFs that invest in AI stocks, or to invest in ETFs that are managed using artificial technology.

So this is right about when I like to hit the pause button and say, let’s back up. Let’s actually look at some data and see what the data tells us it’s going on here.

So first and foremost, this run-up in AI stocks, it’s important to highlight that this is very recent. This has really happened only over the past several weeks, perhaps the past month or so, where we’ve seen big run-ups in a handful of AI-related stocks.

Now, to be fair, we have seen a run-up in technology stocks this year that began back in January. Some of that is perhaps related to the Fed’s pause that’s coming in June with respect to interest rate hikes.

And we’ve seen the growth trade this year significantly outperform the value trade again, in my view, probably due to the fact that the Fed is getting ready to hit the pause button on interest rate hikes.

But back to AI stocks, the reality is the longer term evidence suggests that this is not part of a longer term secular trend with respect to AI stocks. This is a very short-term phenomenon that has occurred really here just over the past month or so.

Now, what about ETFs? Could we perhaps invest in a, quote, “diversified basket of AI-related stocks through an ETF And perhaps that could help us outperform.” And so what I did is I went and I pulled data over the past five years worth of returns data on a basket of four AI-focused ETFs–

QTUM, QT, ROBO, and AIEX.

All of these are ETFs. You can find lots of information on these ETFs online, on Morningstar, or any other publicly available website. And if we look at those four ETFs over the past five years, what we see is that they’ve delivered a median return of about 7% annually over the past five years.

Now, how does that compare to the S&amp;P? Well, the S&amp;P has actually returned a little over 10% annually over the past five years. And all of this is data through the end of May 31, 2023.

Now, the reality is you should probably benchmark those four ETFs against a technology sector ETF, perhaps a Spider or something like that. And when we do that, what we see is that over the last five years, the technology index as a whole has actually delivered 19.3%

annually. The technology sector obviously has done very well over the past decade or so.

And so again, when we look at these four ETFs that are focused really on AI, AI-related technology, or are perhaps managed using AI, which is what AIEQ does, what we see is that there’s no evidence here that AI-managed ETFs or that AI as a subpopulation of stocks within the tech sector, is outperforming either the S&amp;P or the technology sector as a whole. So again, no evidence here over the past five years of outperformance.

We’ve certainly seen strong outperformance here over the past month or so. Let’s just back up a little bit more and look at some of the data based on where AI stocks are today. When we look at valuations, what we observe is those 11 stocks identified by T. Rowe Price, they trade at a market cap of about 37 times forward earnings. That is a market cap weighted forward earnings of 37 times forward earnings.

Now, how does that compare to the S&amp;P 500? Well, the S&amp;P trades at only 18 times forward earnings. So forward earnings, this is a price to earnings multiple that looks at forward expected earnings. Higher is worse. And we always want the valuation to be lower.

And so to put this in English, AI stocks currently are trading at about twice the valuation of the S&amp;P 500 as a whole. So that’s quite high. That should have alarm bells going off in your head.

Again, there’s nothing wrong with AI stocks conceptually in terms of the promise that they hold for modern life in terms of their applications and things like that. Certainly all of those, I would argue, are probably good things for society as a whole.

But we need to be careful, that does not translate into a winnable investment thesis, not when you would be buying those stocks at twice the valuation as the broader S&amp;P 500 index. Perhaps a little bit more on that here in a few moments.

So I want to pivot, and I want to tackle a related question. And that is, can AI-driven portfolios deliver superior returns to investors? So maybe setting aside whether or not we should invest in AI stocks, can we use AI to build and manage better portfolios for clients?

And so in this case, I want to consider AIEQ. According to AIEQ’s website, it is an ETF that fully utilizes AI to select stocks that it believes have the highest appreciation potential over the next 12 months.

And AIEX has been around for over five years. Its inception date was October 17, 2017. And so I wanted to explore and actually look at the data and just evaluate how has this AI-managed ETF, how has it actually done relative to the broad market?

Well, since inception, the fund has returned only 2.66%

annually versus a little over 10% for the S&amp;P 500 index. So clearly a dramatic underperformance relative to the S&amp;P. Well, let’s dig a little deeper. Maybe on a risk-adjusted basis, the fund outperformed. And what we observe is that’s not the case either.

After we adjust for risk, when we look at the realized Sharpe ratio of the returns for AIEX, we see that it has a Sharpe ratio of 0.24 versus 0.61. And the Sharpe ratio is a measure where higher is better. We always want a higher Sharpe ratio relative to a lower Sharpe ratio.

So what does this mean in English? What it means in English is that this particular AI-managed ETF actually underperformed the S&amp;P 500 index. And it also delivered higher risk to investors than the S&amp;P 500 index did.

So when we actually look at the data, what we see is that the volatility of returns for AIEX was about a third higher than the volatility of the S&amp;P 500 index as a whole. So the moral of the story here is if you invested in this particular ETF, at least over the past five years, you actually earned less than the S&amp;P. And in return, you also had to stomach greater volatility than the S&amp;P over the past five years.

Again, past performance certainly no indicator of future performance. So all of the obligatory disclosures certainly apply here. But again, that’s what the data tells us when we just look at the actual real world returns.

So what are the common sense takeaways for investors here? I’ve thrown a bunch of data out there. What are the common sense takeaways for investors? First, the run-up in AI stocks is very recent. And if you’ve been investing for any meaningful period of time, what you have observed is that the markets have the attention span of a gnat.

And so things are always changing. There’s always new trends. Yesterday, it was crypto. Today, it’s AI.

20 some years ago, it was the internet boom. For a while there, it was real estate, the real estate boom in the mid-aughts. So just be very careful here. This is a recent trend.

There’s no evidence that this is somehow a broader secular trend that is investable for all of us as investors.

The second is that valuations matter. They only don’t matter to those who don’t mind losing money. What you pay for a stock or an ETF is the single most important predictor of what your future return will be on that investment. So price matters. And you want to be really careful right now because the valuations of these AI stocks, like I said, they’re trading at valuations that, on average, are two times higher than the valuation of the S&amp;P 500 index as a whole.

There is a negative correlation between valuations and future returns. The higher the valuation, the lower the future return and vice versa. That is a well-documented finding within both academic literature and in the real-world empirical evidence of realized returns for investors over the past, oh, let’s say 100 years or so.

The final point here is that there is no evidence that AI can actually help investors capture better returns. There is none. And so what’s the takeaway here? What should we be doing? Well, the reality is we should continue to invest in broadly diversified portfolios.

And if you are invested in a broadly diversified portfolio, by definition you already have an allocation to these AI stocks, a globally diversified portfolio that owns many, many thousands of companies.

You already own NVIDIA. You already own Meta. You already own Alphabet, a.k.a. Google. You already own Microsoft.

All right. It’s really just a question of the weights, right? So if you have a diversified portfolio, you already have an allocation to AI-related stocks. There’s no reason to go all in or to overweight those stocks just based on short-term performance.

Well, that’s all for today. I’m Don Calcagni. Thank you for listening to <i>Market</i> <i>Perspectives.</i> Should you have any questions, please feel free to reach out to your advisor here directly at Mercer Advisors. Thank you and have a nice day.

[THEME MUSIC]

MUSIC]