Latest / The Best of the Money Show / Investment School: What can decades of financial data actually tell us — and where does it lie?
Transcript
- 0:00Speaker 1: The Money Show.
- 0:01Speaker 2: Investment School. 23 minutes now to eight the time. There
- 0:05Speaker 2: is a phrase that you'll hear in investment. People often
- 0:08Speaker 2: have it as the sort of T's and C's line.
- 0:11Speaker 2: And it goes like this. Past performance is no indicator
- 0:15Speaker 2: of future results. And the point is that someone will say,
- 0:19Speaker 2: we have the best fund. We did the best this.
- 0:21Speaker 2: We did the best that. But they can't predict that
- 0:23Speaker 2: that's what they'll do in the future. Now, is that true?
- 0:26Speaker 2: They're obviously trying to have their cake and to eat
- 0:28Speaker 2: it and then to eat it again. But tonight we're
- 0:31Speaker 2: looking at what does financial data, if you go back
- 0:34Speaker 2: decades in this stuff, what does it tell us? Does
- 0:37Speaker 2: it tell us anything? Can it actually lie? Can it
- 0:40Speaker 2: be a false indicator? Does it give you any hint
- 0:43Speaker 2: of where things will go in the near future or
- 0:45Speaker 2: maybe even in the long-term future?
- 0:48Speaker 1: Dr.
- 0:48Speaker 2: William Wright is a lead data scientist at Differential Capital,
- 0:51Speaker 2: Wayne McCurry Institute. What do I call you, Wayne? Technically,
- 0:54Speaker 2: I'm supposed to call you retired, but I'm going to
- 0:57Speaker 2: say veteran market watcher instead because I think that's more accurate. Gentlemen,
- 1:01Speaker 2: good evening. I really do appreciate the time.
- 1:04Speaker 1: Okay.
- 1:05Speaker 2: William, I'm going to start with you. So whenever you
- 1:07Speaker 2: look at data, why do you need to look at
- 1:10Speaker 2: the data first before you look at any kind of strategy?
- 1:14Speaker 2: Normally in investment, you want to have a strategy that
- 1:16Speaker 2: you follow. That's how most funds work. Why do you
- 1:18Speaker 2: need to look at the data first?
- 1:20Speaker 1: Yeah, absolutely.
- 1:20Speaker 3: Thanks for having me on the show, Stephen. So particularly
- 1:24Speaker 3: from a quantitative lens, when someone comes to you with
- 1:27Speaker 3: this product or the strategy and they say, here's a
- 1:31Speaker 3: back test, you know, this is how the strategy that
- 1:34Speaker 3: we're proposing is done over the past 10 years, 20 years,
- 1:37Speaker 3: 30 years, you know, your next natural question is, well,
- 1:41Speaker 3: what does that mean in terms of performance, right? How
- 1:43Speaker 3: much am I going to get out of this? But naturally,
- 1:46Speaker 3: I actually think the correct question that you should be
- 1:49Speaker 3: asking is, what data did you test this on? What
- 1:54Speaker 3: information did you use to inform the strategy, to train
- 1:58Speaker 3: this model? And there's a whole bunch of sub-questions underneath that.
- 2:01Speaker 3: Was this data clean? Did you make sure that the
- 2:04Speaker 3: correct constituents were used in building all those things out?
- 2:07Speaker 3: Because the answers to those questions will actually tell you
- 2:12Speaker 3: whether the performance that you're getting quoted is even possible
- 2:16Speaker 3: or whether you can actually realize it in the long term.
- 2:20Speaker 2: Wayne, there's another sort of deeper question, I suppose, which
- 2:23Speaker 2: is when you're investing, which is really about the future,
- 2:27Speaker 2: does the past matter? And I mean, the past must
- 2:30Speaker 2: matter because it informs the present would be my first
- 2:32Speaker 2: sort of stab at the answer.
- 2:35Speaker 1: Yes and no.
- 2:36Speaker 4: Look, in investments, I mean, first of all, let's take
- 2:38Speaker 4: a step back and say nobody knows the future. No model,
- 2:42Speaker 4: no AI, nothing. Nobody knows the future. What I've learned
- 2:47Speaker 4: about past data and past instances and past occurrences is
- 2:52Speaker 4: that the past can't predict the future, but.
- 2:56Speaker 1: It does rhyme.
- 2:58Speaker 4: In other words, you can gather information and data and
- 3:02Speaker 4: ideas and.
- 3:05Speaker 1: From past data.
- 3:07Speaker 4: But to me, as we're speaking about in this whole thing,
- 3:12Speaker 4: in this whole discussion, past data can't predict future data,
- 3:17Speaker 4: but it can give you an idea of possibilities. And
- 3:21Speaker 4: my experience in investments, first of all, I've never seen
- 3:25Speaker 4: a backtest that didn't work because if.
- 3:28Speaker 1: It didn't work, you wouldn't be seeing it.
- 3:31Speaker 4: Every back test you do gives a fantastic result because
- 3:35Speaker 4: if it didn't give a fantastic result, no one's coming
- 3:37Speaker 4: to pitch that idea to you. Every back test I've
- 3:39Speaker 4: ever seen shows this wonderful future or wonderful past. I've
- 3:44Speaker 4: also learned that it is so critical to get your
- 3:48Speaker 4: starting and your end time, not necessarily so much because
- 3:51Speaker 4: your end time is normally yesterday, but to get your
- 3:54Speaker 4: starting time. You can't just say I've gone back 30
- 3:56Speaker 4: years because 30 years ago might have been the peak
- 3:59Speaker 4: of the investment commodity cycle. It might have been the trough.
- 4:03Speaker 4: of the commodity cycle. You know, 25 years is a
- 4:05Speaker 4: long time. So everyone says, oh, we went back 25,
- 4:09Speaker 4: 30 years. But it actually depends on when you start,
- 4:11Speaker 4: at what stage was the cycle in comparison to where
- 4:14Speaker 4: you are now. So everything in investments, is subject to
- 4:19Speaker 4: interpretation and I suppose intuition by the investment manager making
- 4:26Speaker 4: the decision. But certainly past data, past analysis does help
- 4:30Speaker 4: you in getting to a decision, hopefully the right decision.
- 4:35Speaker 1: But it can never be guiding.
- 4:38Speaker 4: And one of the problems I found with very, very
- 4:41Speaker 4: quantitative studies and very, very quantitative people is that This
- 4:47Speaker 4: is the daughter they work.
- 4:48Speaker 1: This is what they do. And like anyone else.
- 4:51Speaker 4: Who's doing a good job, they thoroughly believe that this
- 4:54Speaker 4: is right and this is what the outcome is. The
- 4:58Speaker 4: outcome it's predicting is going to be the right outcome.
- 5:01Speaker 4: So you have to understand that it literally cannot predict
- 5:05Speaker 4: the future. But as we're discussing tonight, how clean your
- 5:08Speaker 4: daughter is is critically important.
- 5:12Speaker 2: Wayne, is there a good example of this playing out
- 5:15Speaker 2: right now? Yesterday, a market commentator made the point that
- 5:20Speaker 2: she had never seen yields and U.S. bonds as high
- 5:24Speaker 2: as they are now. The Economist had a piece, maybe
- 5:27Speaker 2: it was two days ago, the Economist had a piece
- 5:28Speaker 2: I think yesterday saying it wasn't, the yields were not
- 5:32Speaker 2: that high. They were very high over the last 20 years.
- 5:36Speaker 2: But actually the last 20 years have been quite a
- 5:38Speaker 2: strange time for bonds. And if you looked at the
- 5:40Speaker 2: period before that, then they were quite normal.
- 5:44Speaker 1: Yeah.
- 5:44Speaker 4: Look, I mean, after the global financial crisis, bond yields,
- 5:48Speaker 4: I mean, you went to zero, went very low for
- 5:50Speaker 4: an extremely long time period. So in the last 20 years,
- 5:55Speaker 4: it's probably wrong to say, look, the average for the U.S.
- 5:57Speaker 1: Long bond is 2%. It's currently at 5%. It's a disaster.
- 6:01Speaker 4: So there you should go back and look at more
- 6:03Speaker 4: normalized periods because, you know, after the global financial crisis
- 6:07Speaker 4: was abnormal, but make no mistake, the U.S.
- 6:10Speaker 1: Long bond is high. It's not low.
- 6:13Speaker 4: It is high, and investors should be taking it into
- 6:17Speaker 4: account and maybe be a little bit cautious. But certainly,
- 6:20Speaker 4: it's not as catastrophic as if you compare it to
- 6:24Speaker 4: relatively short or medium-term bond averages.
- 6:27Speaker 1: Okay.
- 6:29Speaker 2: William, how much data do we actually have? Do we
- 6:31Speaker 2: go back 50 years? Do we go back maybe 100?
- 6:34Speaker 3: Well, just like everything, it depends.
- 6:37Speaker 2: And it depends on where you are, I'm sure. Exactly.
- 6:39Speaker 3: It very much depends on the exchange and the market.
- 6:42Speaker 3: the market with the longest history of trustworthy data is
- 6:47Speaker 3: the New York Stock Exchange. And it doesn't go back
- 6:49Speaker 3: as far as you might think, considering the history of,
- 6:52Speaker 3: you know, finances and institution. Reliable, trustworthy, continuous data only
- 6:57Speaker 3: goes back to 1925. Then you have the NASDAQ, which
- 7:02Speaker 3: is 1972. And I mean, even if you look at
- 7:06Speaker 3: different asset classes, like, you know, bonds, as we were
- 7:08Speaker 3: just talking about, There's history going back many, many decades,
- 7:12Speaker 3: but detailed, consistent transaction level data only goes back to
- 7:16Speaker 3: the 2000s. Currencies, you have a whole different ballgame entirely
- 7:21Speaker 3: because you have a switch in regimes, right? You have
- 7:24Speaker 3: the gold standard in 1971 switching to floating currencies. So
- 7:28Speaker 3: you're basically having data sets that effectively describe two different
- 7:32Speaker 3: financial worlds.
- 7:34Speaker 2: I'm one of those people who's not scared of a
- 7:36Speaker 2: big history book. The bigger, the better. I'm absolutely petrified
- 7:39Speaker 2: of spreadsheets, but that's a different story. I would presume,
- 7:45Speaker 2: based on that, that the more you know, the better,
- 7:47Speaker 2: which would mean the longer the history, the better. Is
- 7:50Speaker 2: that true?
- 7:51Speaker 3: Not necessarily. The longer the history, the more statistically significant.
- 7:56Speaker 3: But that doesn't necessarily mean that your results are going
- 7:59Speaker 3: to be better, right? Because markets are what we would
- 8:03Speaker 3: call non-stationary. They're they're dynamic and not only do they
- 8:07Speaker 3: change the rules that, you know, basically define the game
- 8:11Speaker 3: can change. And the currencies are a perfect example of that, right?
- 8:16Speaker 3: Because you have a large data set defining, you know,
- 8:20Speaker 3: a different version of what foreign exchange is going up
- 8:23Speaker 3: to 1971, then paradigm shift. And all of a sudden
- 8:28Speaker 3: you have a new data set that defines something entirely different.
- 8:31Speaker 3: So if someone comes to you and says, well, I've
- 8:33Speaker 3: got something that works over 50 years or whatever. If
- 8:38Speaker 3: they haven't given you a justification, how do you handle
- 8:42Speaker 3: that exact inflection point? Then that's something you should be
- 8:46Speaker 3: incredibly weary of.
- 8:47Speaker 2: And then back to that bond yield example, longer history
- 8:50Speaker 2: there is better.
- 8:52Speaker 3: Yes, and that's also due to the nature of the
- 8:54Speaker 3: specific market itself. Yes.
- 8:58Speaker 2: Warren, Wayne, are there Different types of survivorship bias. I mean,
- 9:04Speaker 2: some things seem to, you know, you can, look, you
- 9:08Speaker 2: can make, you know, you get lies, damned lies and statistics.
- 9:11Speaker 2: You get the same with history, by the way. You
- 9:13Speaker 2: can use it to prove anything. Is the same true
- 9:15Speaker 2: for an investment case? You can use the history to
- 9:17Speaker 2: prove anything.
- 9:19Speaker 1: Yeah, look, I mean, you could always find some statistics
- 9:21Speaker 1: to prove your point.
- 9:23Speaker 4: And that's why, you know, you still need, I suppose
- 9:26Speaker 4: you could say, the human touch in investing. But, you know,
- 9:29Speaker 4: survivorship bias, as you mentioned this, Stephen, this is a
- 9:33Speaker 4: massive thing. You know, we once again, as we're talking,
- 9:38Speaker 4: went back and studied past auto when I was still working.
- 9:41Speaker 4: And we said, right, we know the top 10 asset
- 9:44Speaker 4: managers in South Africa. We know their 10-year performance record.
- 9:48Speaker 4: We know the average for the top 10 and each year.
- 9:51Speaker 4: But I said, you know, what we must do is
- 9:53Speaker 4: actually go back and look at these averages with all
- 9:57Speaker 4: the participants, not just the ones who sacrificed. And the average, obviously,
- 10:03Speaker 4: the top 10 change all the time over any particular
- 10:05Speaker 4: time period. And the average is actually quite a bit
- 10:08Speaker 4: less because all the true underperformers are either taken over,
- 10:12Speaker 4: go out of business or change their names or do something.
- 10:15Speaker 4: So their data disappears from the current running average. So
- 10:19Speaker 4: this survivorship bias disappears. is actually a really, really significant thing.
- 10:24Speaker 4: It's the same as saying, look, the top 50 companies today,
- 10:27Speaker 4: let's look at their performance.
- 10:29Speaker 1: Those are the top 50 that have survived.
- 10:32Speaker 4: If you just look at their performance going back, it
- 10:35Speaker 4: doesn't take into account the 30 that didn't survive. So,
- 10:38Speaker 4: I mean, this is actually a very, very material thing,
- 10:41Speaker 4: and it's not something people think about at all, in fact, sometimes.
- 10:45Speaker 4: Is there.
- 10:46Speaker 2: Does this have a big impact then, William? I mean,
- 10:50Speaker 2: does it lead to, like, can you, what I'm trying
- 10:54Speaker 2: to get to is, can you actually, using maths, account
- 10:57Speaker 2: for the bias, the level of bias?
- 10:59Speaker 3: Yeah, and in certain cases it can be quite significant.
- 11:02Speaker 3: I mean, there's been studies that have shown, you know,
- 11:05Speaker 3: depending on the market you're in, it can cost between 2%
- 11:09Speaker 3: and 8%, you know, return year on year. And What's
- 11:14Speaker 3: also interesting is that, you know, we were just talking
- 11:16Speaker 3: about regimes. The effect varies depending on the regime that
- 11:20Speaker 3: you're in. So when it was, you know, way back
- 11:25Speaker 3: in the 1920s and the 1940s, if you, you know,
- 11:29Speaker 3: study the effect of this survivorship biased phenomenon on stocks,
- 11:34Speaker 3: it barely has any impact, at least on the New
- 11:37Speaker 3: York Stock Exchange. And then progressively as the turnover increases
- 11:42Speaker 3: with time, it starts to have a greater impact. And
- 11:45Speaker 3: that just accelerates with electronic trading in the 2000s and
- 11:50Speaker 3: higher frequency trading as well.
- 11:52Speaker 2: Doesn't it always. We're talking investment school about data. Does
- 11:57Speaker 2: past performance, is it any indicator of future results is
- 12:00Speaker 2: really the question. We're looking at decades of financial data.
- 12:04Speaker 3: Dr.
- 12:04Speaker 2: William Wright is the lead data scientist at Differential Capital,
- 12:07Speaker 2: the kind of person you'd like to have in the
- 12:08Speaker 2: room when you have this conversation. Wayne McCurry has the
- 12:11Speaker 2: kind of experience you'd like to have in the conversation.
- 12:14Speaker 2: He's a veteran watcher of our market. Wayne McCurry is
- 12:19Speaker 2: a veteran market watcher.
- 12:20Speaker 1: Dr.
- 12:20Speaker 2: William Wright's lead data scientist at Differential Capital. We're talking
- 12:24Speaker 2: about what decades of financial data can actually tell us. Wayne,
- 12:28Speaker 2: we do need to make sure the data we're looking
- 12:29Speaker 2: at is actually clean. In other words, we need to
- 12:32Speaker 2: make sure we haven't lost any of it. It's handling
- 12:35Speaker 2: all of the data correctly. Companies can change names and things.
- 12:39Speaker 2: You need to actually spend a lot of time on that.
- 12:42Speaker 2: You need to make sure that your data is of
- 12:43Speaker 2: a very high quality before you even start to look
- 12:45Speaker 2: at it.
- 12:45Speaker 1: No, no, hugely so.
- 12:49Speaker 4: I mean, we had some very, very clever, well-qualified and
- 12:55Speaker 4: well-experienced people working in our quants department. I would reckon
- 13:00Speaker 4: they spent 40, 50% of their time looking at the
- 13:03Speaker 4: quality of the data before even putting it into a model.
- 13:06Speaker 4: trying to get some deduction from that model. So it
- 13:11Speaker 4: is vitally important that you know what you're looking at,
- 13:14Speaker 4: that it's clean data. And I mean, quite a few
- 13:16Speaker 4: people just do, you know, just get this raw data
- 13:20Speaker 4: and they don't take out the exceptions because quite often
- 13:23Speaker 4: when you looked at your feed from whichever service provider
- 13:26Speaker 4: you were doing, you could look at a graph and
- 13:29Speaker 4: automatically you could see that's an error. You say, I mean,
- 13:32Speaker 4: that just looks crazy. And you know that it's an
- 13:35Speaker 4: error and it eventually rectifies itself. But what I found
- 13:40Speaker 4: the biggest problem with using, let's call it quantitative people,
- 13:45Speaker 4: is they haven't got market experience. So they can't say,
- 13:49Speaker 4: this looks wrong or this doesn't feel right because they
- 13:54Speaker 4: haven't lived in the market as close as, say, market
- 13:59Speaker 4: participants have. So you've got to also assist them and
- 14:04Speaker 4: look at the data. But getting clean data is critically important.
- 14:08Speaker 2: William, I don't know. Do you identify as a data person?
- 14:12Speaker 2: You're a data scientist. So, I mean, I'm guessing you
- 14:14Speaker 2: can defend your corner if you like.
- 14:16Speaker 3: Well, I mean, last time I checked, I think there's
- 14:21Speaker 3: an attitude in sort of quantitative circles and sort of
- 14:25Speaker 3: the opposite on the subject matter side of things is
- 14:28Speaker 3: that somehow you can divorce the one from the other. Somehow,
- 14:31Speaker 3: if you just have enough data, then magically you can
- 14:34Speaker 3: solve any problem. Similarly, that if you have enough experience
- 14:39Speaker 3: in a domain, then, you know, the feelings that you
- 14:42Speaker 3: have will, you know, be able to lead you to
- 14:45Speaker 3: the correct result where, you know, I think in actual fact,
- 14:48Speaker 3: it's the combination of those two things that makes, you know,
- 14:52Speaker 3: the output most powerful. And I think you can see
- 14:56Speaker 3: it in so many other areas outside of finance, even
- 14:59Speaker 3: it's in, you know, science and engineering. I think those,
- 15:04Speaker 3: prove those cases. So I think why not apply it
- 15:07Speaker 3: to finance as well?
- 15:08Speaker 2: That's why you need a manager to manage everybody together. William,
- 15:12Speaker 2: what happens when data goes wrong? I mean, you can
- 15:14Speaker 2: have data that is wrong and then it's used to
- 15:18Speaker 2: create a model or something. The results, I presume, can
- 15:20Speaker 2: be catastrophic.
- 15:21Speaker 3: Yeah, absolutely. And it's not the sort of thing that's
- 15:25Speaker 3: specifically focused in this domain. I'll give two quick examples,
- 15:29Speaker 3: if I may. One is a hedge fund called Axa Rosenberg,
- 15:34Speaker 3: which in 2007, there was a coding error that resulted
- 15:38Speaker 3: in incorrect data being fed into a risk model. At
- 15:42Speaker 3: the end of the day, no one reported it. The
- 15:45Speaker 3: problem continued for about three years. And when it wasn't covered,
- 15:49Speaker 3: it wasn't reported to executive management and to investors. That
- 15:54Speaker 3: cost the firm $ 200 million, and the founder got a
- 15:57Speaker 3: lifetime ban on trading. in, I suppose, in a way
- 16:03Speaker 3: more consequential. Similar problem happened in 2010 when two Harvard
- 16:09Speaker 3: academics wrote a paper that was used as a justification
- 16:15Speaker 3: by governments across the world to implement austerity measures, saying,
- 16:19Speaker 3: You know, if your GDP debt to GDP ratio reached
- 16:22Speaker 3: a certain level, growth tanks. And it turns out that
- 16:27Speaker 3: when someone tried to replicate the research years later.
- 16:30Speaker 1: They couldn't do it.
- 16:32Speaker 3: And it was a single mistake in an Excel spreadsheet.
- 16:38Speaker 1: Wow. Yeah.
- 16:39Speaker 2: The left must have had fun with that.
- 16:40Speaker 1: Yeah.
- 16:43Speaker 2: Wayne, when you are looking at back tests and when
- 16:45Speaker 2: you're looking at information and you're trying to go through it,
- 16:48Speaker 2: are there particular questions that you should be asking yourself,
- 16:50Speaker 2: particular things that you must think? Are there particular things
- 16:53Speaker 2: that go through your mind when you're looking at this
- 16:55Speaker 2: kind of data?
- 16:57Speaker 4: Well, look, anything that you look at in investments, you
- 17:00Speaker 4: must be critical. You must be skeptical of the outcome.
- 17:03Speaker 4: You must look at the methodology used in it. You've
- 17:07Speaker 4: got to ask critical questions about it. But just to
- 17:10Speaker 4: come back to what William said, it is so vitally
- 17:12Speaker 4: important that, let's call it, quants and experience work together
- 17:17Speaker 4: because no one knows about it. No one knows the
- 17:20Speaker 4: answer to investments. I mean, there is no holy grail.
- 17:23Speaker 1: And one portion....
- 17:26Speaker 4: Cannot exclude the other portion. I mean, it is a
- 17:29Speaker 4: team effort. But with anything in investments and maybe with
- 17:33Speaker 4: anything in life, you've got to be critical with whatever
- 17:37Speaker 4: is presented in front of you and really get to
- 17:41Speaker 4: I suppose, the foundation of the study that's done. What
- 17:47Speaker 4: data did they use? What type of conclusions did they
- 17:50Speaker 4: come to? What time periods? Because as I mentioned earlier on,
- 17:54Speaker 4: to me, my biggest single experience and knowledge in markets
- 17:59Speaker 4: is that by and large, 90% of markets, if not
- 18:03Speaker 4: a bit more, is indeed cyclical. And it depends on
- 18:07Speaker 4: where you are in the cycle. And when you're analyzing
- 18:10Speaker 4: back data, where you started analyzing the back data at
- 18:13Speaker 4: the peak or the trough of any particular cycle, peak
- 18:16Speaker 4: or trough of interest rates, peak or trough of dollar
- 18:19Speaker 4: strength or dollar weakness. All of these things are so
- 18:22Speaker 4: critical in coming to some sort of conclusion when you're
- 18:26Speaker 4: looking at past data.
- 18:29Speaker 2: William, it's nice to look at a big lot of
- 18:32Speaker 2: history and think, well, The history tells us, and therefore,
- 18:36Speaker 2: this is what's going to happen. In practice, what do
- 18:40Speaker 2: you actually need to do? Because, I mean, we think
- 18:44Speaker 2: of history as a long time, but 100 years in
- 18:46Speaker 2: markets is nothing.
- 18:48Speaker 1: Absolutely.
- 18:49Speaker 3: And I think not everyone needs to be equipped with
- 18:55Speaker 3: the statistical metrics and rigor that they would need to
- 18:59Speaker 3: determine the quality of the data that a particular you know,
- 19:04Speaker 3: model was built off of. But I think, you know,
- 19:06Speaker 3: just having those sort of common sense questions that we've
- 19:10Speaker 3: discussed today, you know, that Wayne also mentioned, thinking about,
- 19:15Speaker 3: you know, have the numbers ever been corrected at any
- 19:19Speaker 3: given point in the past? You know, is your index
- 19:25Speaker 3: constituent set clean? You know, if you keep those things
- 19:29Speaker 3: in mind when, you know, approaching any sort of investment offer,
- 19:33Speaker 3: I think That can provide you with more value than, well,
- 19:37Speaker 3: how does this perform?
- 19:40Speaker 2: Wayne, we've got 30 seconds left. You talked about the
- 19:44Speaker 2: value of experience. Does that mean we shouldn't ever trust
- 19:47Speaker 2: our money to a younger person?
- 19:49Speaker 1: No. not at all. Not at all.
- 19:51Speaker 4: Because you have to get younger people together because they
- 19:55Speaker 4: come up with new ideas. But you must have a
- 19:58Speaker 4: combined team combining talents from many, many different fields, many fields.
- 20:03Speaker 4: Many different age groups. Young people come up with good ideas.
- 20:07Speaker 4: Old people are maybe too scarred and are too conservative
- 20:10Speaker 4: and don't want to jump at opportunities because they've been
- 20:12Speaker 4: burned so many times in the past. So you need
- 20:15Speaker 4: a combination of things. You must understand also one critical
- 20:19Speaker 4: thing is that, you know, all boats go up in
- 20:24Speaker 4: the rising tide. So you've got to say, is this
- 20:26Speaker 4: person actually a good manager or are they just surfing
- 20:29Speaker 4: the wave? Because when the tide goes out, it needs
- 20:32Speaker 4: to use swimming naked.
- 20:34Speaker 2: Wayne McCurry, thank you so much. Veteran market watcher, Dr.
- 20:38Speaker 2: William Wright, the lead data scientist at Differential Capital on
- 20:41Speaker 2: Investment School on the Manisha.