The Compound and Friends - The thing about historical data (with Tadas and Ben Carlson)

Episode Date: April 8, 2019

"As market participants we all want more data, but once you understand where some of the data comes from you have to take it with a big grain of salt." Stock market data today is abundant and mostly free. However data purporting to go back decade, if not centuries, has a lot of issues with. Ben Carlson, institutional asset management chief at Ritholtz Wealth Management does as much research as anybody using historical data. So he has a unique perspective on what weight we should put on historical data hen making investment decisions today. You can read more about Ben's take on using long term data at his blog, A Wealth Of Common Sense: https://awealthofcommonsense.com/2019/03/real-estate-vs-the-stock-market/ Enable our Alexa skill here - "Alexa, play the Compound show!" https://www.amazon.com/Ritholtz-Wealth-Management-LLC-Compound/dp/B07P777QBZ Talk to us about your portfolio or financial plan here:  http://ritholtzwealth.com/ Obviously nothing on this channel should be considered as personalized financial advice just for you or a solicitation to buy or sell any securities. Please see this 3,000 word terms & conditions disclaimer if you seriously need this spelled out for you.  https://thereformedbroker.com/terms-and-conditions/ Hosted on Acast. See acast.com/privacy for more information. Learn more about your ad choices. Visit megaphone.fm/adchoices

Hey, Ben, it's Tadous. How's it going? Hey, Tadous. Hey. Hey, we both wrote about a piece talking about the equity risk premium versus real estate over a couple of centuries. And it kind of got me thinking about a topic which I know I've written about and you've written about as well.

And that's really kind of thinking about data, equity market data over not just decades, but centuries. And so I guess one of my first questions to you is, you know, how should we really think about that data? I mean, we've really only had computers and computerized data for, you know, really not that long. But, you know, people purport to have data for stocks going back all the way back to the 1800s. Yeah. And unfortunately, I think a lot of people take it at face value, too, that those numbers must be exact. Because if you think about it, the Shiller data that he uses for the CAPE ratio goes back to 1871. But I read about it, and I think in one of his books or maybe the New York Times or something, he actually came up with the CAPE ratio in 1988 when he wrote an article about it.

So obviously they had to piece a lot of that data together. And even Schiller said going into the real estate crisis, one of the reasons that it was so hard for people to understand what was happening with real estate prices is because no one had ever bothered to put that data together. And he was one of the first ones to do it. So not only is the data old, but the collection of the data is relatively new, and obviously they're using data collected at the time. But I wish I had a time frame to give in terms of here's a drop dead and when you can use it or can't use, but I think you just have to take it with a huge grain of salt.

No, I think that's absolutely right. I mean, I think, you know, it's amazing because I think you hit it on the head. People take that data as kind of as gospel. And I know, you know, for a long time, the Ibbotson data in terms of, you know, stocks, bonds and bills was, again, one of those things that were just kind of taken as, you know, taken as gospel. And so, you know, I think it's probably too big of an ask for the average investor to kind of, you know, dig into the details and be skeptical about it. But I think you're right in terms of, you know, kind of the general rule being, you know, anytime you see somebody purporting to have, you know, centuries of data, I think you have to, you know, take that, like you know, anytime you see somebody purporting to have, you know, centuries of data,

I think you have to, you know, take that, like you said, take it with a big grain of salt. So I remember a few years ago reading a story at the Chicago Business School website about how they put together the first CRISP data, that CRSP database. And they did it in the 1960s. And they basically didn't have any computers at the time. And they were trying to piece back together to the 1920s, all these stock market prices. And they didn't't have any computers at the time, and they were trying to piece back together to the 1920s, all these stock market prices. And they didn't really have a lot to go on. And they basically said they were starting from scratch, and they were trying to work backwards. And I think you can use that sort of data to provide something of a range of outcomes maybe and think in a probabilistic way. But I think if you're trying to use that data to show this is exactly what's going to happen because something happened back then,

it's such a stretch because not only is the data suspect in a lot of ways, but I think it's hard because back at that time, no one had the data either. So having the data in some ways can kind of change the way marker participants think about it. No, absolutely. And just for people's reference, the CRISP database is essentially the gold standard for academic research. So, you know, that just to give you some just to give people some context, you know, that's, you know, just about every academic paper you read today that, you know, uses uses equity data is in all likelihood using the crisp database. And of course, the other element of this is even the data that is good that we have now that all the quants are using, unfortunately, everyone has that same data in most cases, which is kind of good and bad, I guess.

We've democratized the use of data. And I'm always shocked at the amount of free data that's available today. I wrote a post about this a number of years ago, some of my favorite websites for free data, because people ask all the time. But that's kind of a double edged sword because the fact that it is free now means that everyone has it in most cases. And so it kind of comes down to how do you use it and what is how do you create your analysis on that data, not just what is the data itself? How do you create your analysis on that data, not just what is the data itself? Yeah, no, that's why hedge funds are in a, you know, arms race to try and find, you know, alternative data that isn't yet captured by, you know, a lot of these free databases. So that's why they're, you know, looking at satellite data and, you know, shipping data and all sorts of stuff, which, you know, is, you know, way beyond my pay grade. And it's crazy how that stuff becomes table stakes. At a certain point, you get a first mover advantage and then everyone uses that. So again,

it comes down to how are you going to actually use the data in your analysis to sort of look at the past through a lens of probabilities and then apply it to the present and then think about the future, which obviously is one of the things that makes this whole game so interesting, I guess. Yeah, I think the challenge for, you know, the bigger challenge, I think, is more that I find to be, and I would probably characterize it as being more conceptual, is even if you assume that the data that we have going back is, you know, pristine, let's just say it's pristine, Let's just say it's as good as we get. You know, to me is to what relevance is stock market data from the 1890s or early 20th century, which is essentially, you know, dominated by railroad stocks. How is that relevant to the

stock market today? I mean, to me, it's like, you know, it's almost it's like apples and oranges. Well, I think that's the other big lesson that I always get from looking at market history is it gives you a sense of how the structure of the market changes over time. And I think that's important for people to understand. And I think in one of your blogger pieces where you ask other bloggers questions, I think you asked something along the lines of, what if we had 10,000 years of data? What would it show us or how would it change the way you feel? And it's kind of amazing because even if we, I think even if we did have that much data to go on, it still in some ways wouldn't help because people would just be looking for different ways to use it

and sort of data mining from those 10,000 years. So I think it always is just, it's really hard. And again, it helps you think through probabilities, not certainties in a lot of ways. Yeah, no, I think that's right. I mean, I think the, the, you know, the, for quants, I think the kind of the existential question is, you know, what's changed, you know, have we had, are we in a completely different environment where the previous data is, if not irrelevant is, you know, should be greatly depreciated. So I think that's something, not just quants, but I think all investors have to deal with, whether you're a quant or whether

you're a fundamental investor, is the valuation I'm using on this stock today, is it relevant? So I think that's always the challenge for anybody making any sort of investment decision. Yeah. So I mean, I give those researchers credit who painstakingly go back through and put those data series together because I think it is, in some ways, it's almost more entertaining and interesting than anything. But I just don't know how useful it is. So that's kind of my big takeaway is just be careful when you, especially going back that far with your data and trying to make really legitimate extreme, you know, predictions off of it. Yeah, no, I think you're right. I think you have to, I think you have to look backwards,

but the biggest decision for anybody investing is how you take that, how you take that data, analyze it and, you know and apply it to the future. And I think, like you said, we're dealing with probabilities, not certainty. Right, there are no front tests. Yeah, absolutely. All right, well, Ben, thanks very much for joining me, and we'll talk soon. Thanks, Addis.