Session 32 (of 42): Market Timing - Mean Reversion and Macro Fundamentals

0:00 Hi, welcome to this my third session on timing the market. Markets reflect the economy, right? At least that's a conventional wisdom. What happens to interest rates, what happens to the economy should ultimately play out in markets. And there are people who take this insight which is a logical one and build on and say if I can forecast where interest rates are going or what the economy is doing, I should be able to forecast markets, right? And the answer they expect usually is sure. Yeah. In this session, we're going to look at the link between fundamentals, macro fundamentals in the market and whether you can use that linkage to actually make money on that on those relationships.

0:41 Ultimately, many of the measures we use to time markets are mind mean reversion measures. What does that mean? We assume things revert back to the way they used to be. In the context of stocks and bonds, here's how they play up. In the stock market, we look at PE ratios, the multiple of earnings stocks trade at. And we look at what stocks are trading at as a multiple of earnings today relative to history. And if this number, the number they're trading at today is much higher, we argue stocks are overpriced. If interest rates today are much lower than they used to be, we argue that they're going to go up.

1:16 Implicit in both is the assumption that things revert back to the way they used to be. We'll talk about when that assumption breaks down, but it's a pretty reasonable one, right? So, let's take US equities and look at PE ratios going back to 1960. So, what I've looked at is price, the level of the index divide by the trailing earnings for the index. What does that mean? On December 31st of each year, I don't know what the earnings will be through December 31st. I note only only through September 30th or the third quarter of the year. I'm taking the trailing earnings and I've graphed out the P ratio. I've also taken the historical P ratio and this is pure statistics. Don't read more than you should into it. And looked at the 75th percentile and the 25th percentile, the third and the first quartile. The 75th percentile is 19.9. That's a PE ratio of about 20. And the 25th percentile is a PE ratio of about 14.5. You're saying so what if you're a believer in mean reversion and you looked at the PE ratio in 2024 which is about the PE ratio is about 23 your argument will be stocks are overpriced why because they're much higher than they should be the PE ratio is much higher than the 75th percentile in contrast you go back to 2011 or much of the 1970s your argument is stocks are underpriced the P ratio is much lower than the 25th percentile and already you can see the seeds of weakness in these approaches.

2:43 In hindsight, of course, it looks like a great strategy. In the 1970s, if you'd bought stocks in 1972, which is when they first dropped below the 25th percentile, you'd have spent an entire decade in the darkness. But that's the basis for the PE ratio. And it's a and since the PE ratio is so widely talked about, it should come as no surprise that there have been modifications of that traditional PE ratio. When you divide the level of the index today by the earnings in the most recent year, one of the problems you run into is because earnings are cyclical. They're volatile. They're up in boom years, down in bust years. There are people who say rather than look at last year's earnings, we should look at earnings over a longer period. That's called a normalized PE. We take the price today and divide by the average earnings for the last 10 years. But when you bring in the earnings over the last 10 years, you bring in another wild card which is if there's inflation, earnings from 10 years ago and earnings today are not directly comparable. So here's what you can do. You can take the earnings from 10 years ago adjusted for inflation. And when you do that, you've gone from normalized PE to what's called the Schiller PE. The Schiller PE is a normalized inflation adjusted earnings.

3:55 And the argument is it's a much better measure of the PE ratio because it smooths things out and gives you a better perspective. As you can see in this graph, I think too much is made of it. For the most part, the three versions of PE ratios move together. Regular PE, normalized P, and Schiller PE. And guess what? All of them at the end of 2024 would have said stocks are overpriced because the PE ratio today is higher than it's been historically.

4:22 But key here to remember is when you say higher than it's been historically is you're drawing drawing on the assumption that things revert back to the way they used to be. That's with stocks. Let's look at interest rates. This is a graph of the T-bond rate, the 10-year T- bond rate going back to 1927. Again, if you take this graph, you can draw the 90th, the 75th, the the 25th, and the 10th percentile. You're going to get a range on rates and there will be periods where interest rates are below that range. And here again, mean reversion kicks in. If interest rates are higher than they've been historically, let's say the 90th percentile is 5%. And you're at 7%, the argument is rates are going to are more likely to come down. If rates are lower than they were historically, let's say they're 2%, the historical historically the 25th percentile is closer to 4%. The argument might be rates are too low, they've got to go up. And here again, the weaknesses kind of kick through, right? In the last decade, for instance, you'd have been out of the entire bond market with this view because rates dropped below historic norms and they stayed lower. Now, of course, the reason they stayed lower is there were structural shifts in the economy that kept rates low. And of course, for conspiracy-minded people, the Fed kept rates low. But whatever the reason, just because rates are lower or higher than expected doesn't mean that they will move back to the norm at least in the near term. So with both stocks and bonds, there is this assumption of mean reversion. So that's one way to think about market timing is to draw on mean reversion. Assume things will revert.

5:59 and a surprisingly large number of market dimemers. That's the basis for their argument that stocks are expensive or cheap is comparing P ratios today to P ratios historically or interest rates today to interest rates historically. A slightly deeper way of thinking about market timing is say look ultimately what you see in the market has be a reflection of what's happening in the economy. What's happening to interest rates, inflation, GDP growth and over time there have been rules of thumb that have been developed that claim to tell you what's going to happen to stocks based on what happens to those fundamentals. I've heard people argue that the best time to buy stocks is when interest rates are dropping when in or when the long-term interest rates are really low or when the yield curve is upward sloping. Basically, what you're capturing there is are long-term rates much higher than short-term rates or when the economy is strong. Now, intuitively, all those things seem to make sense, right? When interest rates are low, of course, prices should be higher. When GDP growth is strong, earnings growth should also be stronger.

7:04 But to see how noisy these for these indicators are, here's what I did. Let's take rates. Let's suppose you go back in time and look at years in which rates dropped by more than 1% between 0 and 1% and also look at years where rates went up, rates going down, rates were going up and then look at the returns across time. You can already see if you look at the last column that this is a very noisy indicator.

7:31 It's true that when rates drop between zero and 1%, your returns are 18.4%. Right? But before you get too excited, when rates drop by more than 1%, which should be better, right? The annual return is only 9.14%. When rates go up by 0 to 1%, your returns are 9.84, but they go up by more than 1%. They're up only 6.25%. You might look at this and say this is, you know, this means when interest rates go up, it's bad for stocks. And interest rates go down, it's good for stocks.

8:04 That might be the conclusion you reach, but you can see it's a very noisy relationship and there are lots of things that can cause you not to earn the returns you see on paper. When interest rates go down, is it good for stocks? Perhaps. But again, this is also I mean, I'm looking at returns in the following year. in the same year you might get a much stronger relationship but because you have to observe what happens and then invest. So I didn't clarify that better before I described this table. This is what you're doing after you've observed what happened to rates in the previous year. Knowing what's happened to rates in the last year doesn't help you that much in the market in terms of forecasting what stocks will do.

8:48 There's a study by bunch of researchers who looked at a strategy of in of switching from stock to cash based on the level of the table rate. And while they concluded that having this strategy would have added about 2% in excess returns before transactions, cost and taxes, there's a danger here of making money on paper on these strategies because it's a follow-up know look at the same strategy discovered. Much of that excess return came from one time period 1950 to 75.

9:18 You took that out all of the predictability disappeared. So when you look at these studies that claim to time the entire market based on rates, take a deeper look. You see what about the level of rates? Aren't lower rates better for returns than higher rates? Well, this graph what I have is the level of rates on one axis and what stocks do in the other. Right? Now if you have a strong positive rel, in other words, when you expect when rates are low, you expect returns to be higher.

9:47 You should expect to see that show up as much better returns when rates are low and you don't. This looks like a short gun blast. In fact, the line I fit through was at a very slightly positive slope. In fact, the years where T-bond rates were higher, the returns tended to be higher in the following year. Here again, I'm not looking at returns by looking at the same year. So the level of debond rates rate you observe it and you look at what stock returns are in the next year not much to learn. So you tell me rates are low now what should I do next year and stocks I really can't tell you much because the relationship between rates being low or high and what stocks do in the subsequent period is diffuse the reason of course intuitively is the following. If rates are high, stock prices already reflect those rates. For you to make returns, stocks have to be mispriced given the level of rates.

10:46 There are of course people argue that T-bond rate itself is not the indicator you should be looking. You should be really looking at what stocks are priced to earn which they capture with an earnings to price ratio. The inverse of the P ratio and compared to the T- bond rate. It's a model called the Fed model. It's been around for decades. There are people who swear by it still. And what they're looking for is an earnings yield that exceeds the T- bond rate. So here's what I did. I went back in time and I broke return data down into that different based on that difference between the earnings yield and the T- bond rate. So greater than 2% 1 to 2%.

11:24 So as you go further down, these are the years where the earnings yield is less than the T- bond rate, which at least in the Fed model is a bad sign for stocks. I'm going to let you make a judgment based on the returns you see across time, but I don't see a pattern. It's true when earnings yield is more than 2% higher than the T- bond rate, my returns are 13.85%. You're saying this is great, but your returns are almost almost as good when the earnings yield is 1% below the T- bond rate, which suggests again an incredible amount of noise in this relationship. I have never been a great believer in the Fed model. Maybe I'm missing some nuance here, but comparing the earnings to price ratio, the debond rate might give you some nice talking points at the next cocktail party you go to. But as a market timing device, it hasn't worked that well. And you can cap see that in terms of the variance in returns over time. You think what about the economy? So if you look at the link between what happens to the economy measured as the change in real GDP and the stock return the year again it's an incredibly noisy relationship. The notion that when the economy is doing well stocks do well might be widely held but there's very little basis for it.

12:43 Stocks do well when the economy is doing badly. Stocks can do badly when the economy is doing well. And the reason again is it's all relative to expectations. Here what I've looked at is observing what the economy is doing and then looking at the stock return the next year. Again the key is the reason you don't want to look in the same year is there's not much you will learn by you know you can do if you learn that when the economy is strong stocks do well in that same year. You need to be able to have an observable number to act on. So what I've done is looked at the the data based upon GDP growth during the year during a year from really bad years less than 0% to really good years and then looked at the return in the following year. Again focus on the average return there is no relationship between how well or badly the economy was doing last year and what stocks will do in the next year. By now you see the pattern right with every macroeconomic indicator the linkage is weak or often in the opposite direction as intuition is suggest. So if your intuition say when rates are low and the economy is strong I should invest in stocks hold back because the data doesn't back that up. Second, if you think that you can go out and forecast macro variables, I'll concede the ground to you because if you can forecast the level of interest rates next year and what GDP, maybe there's a basis for you to make money, but forecasting macro variables is really difficult to do.

14:14 Finally, watch out for structural changes. We talked about interest rates staying low in the last decade between 2010 and 2019. We also talked about why many people attribute that the Fed or central banks. I don't think that's the reason why interest rates were low. You know why rates were low in the last decade? Because inflation dropped and real GDP growth was anemic. Interest rates were low for that reason. That was a structural change. And those people who missed that structural change ended up being out of both the stock and the bond market for much of the last decade because they held their breath waiting for rates to go back up. So to summarize, here's what the session's been about. Is there a linkage between macro variables like interest rates and the economy and markets? Obviously, but much of that linkage is contemporaneous, which means in years in which interest rates go down, stock prices tend to go up. But that's not good enough, right?

15:12 Cuz to be able to trade on something, you got to be able to observe it. So in this in this session, hopefully you've seen how difficult it is to establish a linkage between what's happened to these fundamentals in the past and what will happen to stocks in the future. I hope you found the session useful and I thank you very much for listening.