WEBVTT - How to Use Narrative Information

0:00:09.080 --> 0:00:14.400
<v Speaker 1>Investors have long been taught that fundamentals drive stock prices,

0:00:15.040 --> 0:00:19.239
<v Speaker 1>revenue growth, profits. As Benjamin Graham taught us, in the

0:00:19.320 --> 0:00:22.000
<v Speaker 1>short run, the market is a voting machine. But in

0:00:22.040 --> 0:00:26.640
<v Speaker 1>the long run it's a weighing machine. But what if

0:00:26.680 --> 0:00:32.400
<v Speaker 1>it isn't? What if narratives are driving market valuations. I'm

0:00:32.440 --> 0:00:35.600
<v Speaker 1>Barry Riddolets and on today's edition of At the Money,

0:00:36.159 --> 0:00:40.040
<v Speaker 1>we're going to discuss how to identify when market narratives

0:00:40.680 --> 0:00:44.760
<v Speaker 1>overtake fundamentals. To help us unpack all of this and

0:00:44.800 --> 0:00:48.240
<v Speaker 1>what it means for your portfolio, let's bring in Ben

0:00:48.360 --> 0:00:53.080
<v Speaker 1>Hunt of Percion. Ben's firm studies narratives and the stories

0:00:53.120 --> 0:00:57.920
<v Speaker 1>that shape markets, investing, and social behavior through the lens

0:00:58.040 --> 0:01:02.160
<v Speaker 1>of information theory. So Ben, let's start with a definition.

0:01:02.560 --> 0:01:06.399
<v Speaker 1>How do you define a narrative in the context of

0:01:06.560 --> 0:01:07.680
<v Speaker 1>markets and investing.

0:01:08.319 --> 0:01:13.040
<v Speaker 2>It's a simple definition. A narrative is an answer to

0:01:13.120 --> 0:01:17.640
<v Speaker 2>the question why why did the market go up today?

0:01:18.160 --> 0:01:20.440
<v Speaker 2>Why should you buy the stock? Why should you sell

0:01:20.480 --> 0:01:23.920
<v Speaker 2>this stock? Why should you vote for this person? Any

0:01:24.080 --> 0:01:28.240
<v Speaker 2>answer to the question why is a narrative? Well?

0:01:28.280 --> 0:01:33.520
<v Speaker 1>Sad you distinguish between something like data, which I arguably

0:01:33.880 --> 0:01:37.560
<v Speaker 1>comes with a storyline attached to it, and just a

0:01:37.720 --> 0:01:43.000
<v Speaker 1>straight up Hey, bitcoin is millennium or digital goals? Like,

0:01:43.040 --> 0:01:46.000
<v Speaker 1>how do you define the difference between this? I guess

0:01:46.080 --> 0:01:49.160
<v Speaker 1>this stock is cheap with a pe of nine? Is

0:01:49.200 --> 0:01:49.760
<v Speaker 1>a narrative?

0:01:50.440 --> 0:01:53.480
<v Speaker 2>Yeah, well why is? Why is a pe of nine cheap? Right?

0:01:53.640 --> 0:01:56.080
<v Speaker 2>Or why is a pe of three cheap or fifteen cheap?

0:01:56.280 --> 0:02:03.240
<v Speaker 2>These are all stories. Any any valuation, any multiple, any

0:02:03.440 --> 0:02:08.359
<v Speaker 2>meaning that you attached to numbers, it's a story. It's

0:02:08.400 --> 0:02:12.160
<v Speaker 2>a It's an answer to the question why why do

0:02:12.240 --> 0:02:14.560
<v Speaker 2>I call it cheap? Why do I think you should

0:02:14.600 --> 0:02:18.320
<v Speaker 2>buy it? Why is this interesting? Those are all why

0:02:18.480 --> 0:02:21.640
<v Speaker 2>questions and the answer Those are all narratives.

0:02:22.320 --> 0:02:27.799
<v Speaker 1>So how do you identify any particular narrative that may

0:02:27.840 --> 0:02:32.040
<v Speaker 1>be driving a particular stock or asset class or the

0:02:32.160 --> 0:02:36.000
<v Speaker 1>overall market? What what tools do you use to help

0:02:36.120 --> 0:02:36.640
<v Speaker 1>discern that?

0:02:38.000 --> 0:02:40.440
<v Speaker 2>Well, let me start by saying what I don't care about.

0:02:40.800 --> 0:02:44.800
<v Speaker 2>So what I don't care about is truth or accuracy.

0:02:45.320 --> 0:02:47.000
<v Speaker 2>I mean, it sounds crazy, but.

0:02:47.560 --> 0:02:49.640
<v Speaker 1>Not at all because you're trying to figure out what

0:02:49.840 --> 0:02:52.680
<v Speaker 1>the crowd believes, whether it's true or not. Well, if

0:02:52.720 --> 0:02:55.480
<v Speaker 1>it affects their behavior, it matters.

0:02:55.280 --> 0:02:58.360
<v Speaker 2>It matters. Right, So I've I've given up on trying

0:02:58.400 --> 0:03:02.400
<v Speaker 2>to figure out what reality is. What I'm trying to

0:03:02.440 --> 0:03:07.639
<v Speaker 2>figure out is how is reality being presented? How is

0:03:07.680 --> 0:03:10.680
<v Speaker 2>it being presented to us? So what I'm looking for

0:03:11.360 --> 0:03:16.839
<v Speaker 2>are elements of presentation. I'm looking for word choice. I'm

0:03:16.840 --> 0:03:20.360
<v Speaker 2>looking for how loud is it being presented to you,

0:03:20.800 --> 0:03:25.079
<v Speaker 2>how frequently it's being presented to you. But most of all, Barry,

0:03:25.200 --> 0:03:28.959
<v Speaker 2>I'm looking for a concept that you talk about in

0:03:29.240 --> 0:03:33.680
<v Speaker 2>network math, and it's called density, And I'm looking for

0:03:33.760 --> 0:03:37.880
<v Speaker 2>how the language is connected to other words. In fact,

0:03:37.920 --> 0:03:41.760
<v Speaker 2>those are the measurements you use in network math. Those

0:03:41.760 --> 0:03:45.600
<v Speaker 2>are where they we talk about betweenness, we talk about

0:03:45.920 --> 0:03:50.960
<v Speaker 2>connectedness and centrality. Those are the measurements. Because I'm looking

0:03:51.040 --> 0:03:53.440
<v Speaker 2>at the presentation, not at the reality.

0:03:53.720 --> 0:03:56.720
<v Speaker 1>So how do you measure density? You hinted at some

0:03:56.840 --> 0:03:59.720
<v Speaker 1>of it, how loud it is, how repetitive it is.

0:04:00.240 --> 0:04:04.560
<v Speaker 1>How do you tell when a specific narrative is beginning

0:04:04.560 --> 0:04:06.560
<v Speaker 1>to exert influence over prices?

0:04:08.200 --> 0:04:11.480
<v Speaker 2>Yeah? So, Bob Sheller wrote a pretty good book a

0:04:12.120 --> 0:04:16.440
<v Speaker 2>couple of years back called Narrative Economics, And the takeaway

0:04:16.440 --> 0:04:19.680
<v Speaker 2>from that book is that you should think about and

0:04:19.839 --> 0:04:25.560
<v Speaker 2>understand narratives in exactly the same way you think about

0:04:25.960 --> 0:04:31.800
<v Speaker 2>and you understand disease. I mean, you use the same measurements.

0:04:32.120 --> 0:04:34.000
<v Speaker 2>You remember when we were talking about COVID, it was

0:04:34.120 --> 0:04:38.680
<v Speaker 2>are not right? How fast does it spread? How quickly

0:04:38.720 --> 0:04:42.600
<v Speaker 2>does it spread? What is the medium in which it

0:04:42.640 --> 0:04:47.200
<v Speaker 2>can spread most easily? It's exactly the same thing here, bear,

0:04:47.279 --> 0:04:50.760
<v Speaker 2>It's exactly the same thing. We're really using exactly the

0:04:50.800 --> 0:04:54.480
<v Speaker 2>same math as you would use to try to understand

0:04:54.640 --> 0:04:59.120
<v Speaker 2>epidemiology and the spread of a virus. So what we're

0:04:59.120 --> 0:05:04.000
<v Speaker 2>looking at, though, instead of the atmosphere or waste water,

0:05:04.360 --> 0:05:07.279
<v Speaker 2>we're looking at the words. We're looking at all the

0:05:07.360 --> 0:05:12.280
<v Speaker 2>words that are out there in terms of text, of

0:05:12.440 --> 0:05:17.880
<v Speaker 2>what's being written, what's being said. This is all unstructured data,

0:05:18.279 --> 0:05:21.560
<v Speaker 2>and we're looking for the presence of certain ideas, clusters

0:05:21.560 --> 0:05:27.479
<v Speaker 2>of words that spread through that medium of media in

0:05:27.560 --> 0:05:30.159
<v Speaker 2>exactly the same way that a virus spreads through the

0:05:30.240 --> 0:05:33.479
<v Speaker 2>air or through the water. If you're a financial advisor

0:05:34.040 --> 0:05:38.760
<v Speaker 2>or you're a market person of any sort and somebody

0:05:38.800 --> 0:05:41.719
<v Speaker 2>comes up to say, yeah, why the why the market

0:05:41.720 --> 0:05:45.440
<v Speaker 2>go up today? And you could say, well, there are

0:05:45.440 --> 0:05:49.200
<v Speaker 2>a hundred different reasons, but basically it was just variants.

0:05:50.080 --> 0:05:53.039
<v Speaker 2>You know, it's just it's just it's just random. It's

0:05:53.080 --> 0:05:56.360
<v Speaker 2>just a random walk. But it went up today. So

0:05:56.400 --> 0:05:59.080
<v Speaker 2>someone will look like people hate that answer because it's

0:05:59.120 --> 0:06:02.440
<v Speaker 2>a crappy story. It's just it's a it's just true.

0:06:02.640 --> 0:06:04.479
<v Speaker 1>But it's not competitive exactly.

0:06:04.520 --> 0:06:09.240
<v Speaker 2>It's not truthy, right, it doesn't it doesn't connect with you.

0:06:09.240 --> 0:06:13.400
<v Speaker 2>I said, well, that's disappointing. I want a story. So

0:06:15.080 --> 0:06:19.560
<v Speaker 2>the stories that are basically there to fill the time. Uh,

0:06:19.760 --> 0:06:25.160
<v Speaker 2>and these are these are typically examples where I call

0:06:25.200 --> 0:06:29.919
<v Speaker 2>them descriptive narratives. They're they're answering the question why, but

0:06:29.960 --> 0:06:34.000
<v Speaker 2>they're describing why something happened.

0:06:35.040 --> 0:06:37.839
<v Speaker 1>Always after the fact, never in advance.

0:06:37.640 --> 0:06:41.279
<v Speaker 2>Always after the fact. That's exactly right, Barry. So what

0:06:41.440 --> 0:06:44.480
<v Speaker 2>you'll find is that I don't know, it's earning season,

0:06:44.680 --> 0:06:49.599
<v Speaker 2>and who comes out first with earnings financials, right, and so, uh,

0:06:50.240 --> 0:06:54.839
<v Speaker 2>you know City and Goldman, Sacks, whoever. They'll report some

0:06:54.920 --> 0:06:59.480
<v Speaker 2>good earnings, the stocks will go up, and afterwards, Kramer

0:06:59.680 --> 0:07:02.400
<v Speaker 2>and everyone else, it's got to tell you why. And

0:07:02.440 --> 0:07:05.599
<v Speaker 2>they'll say, oh, I'm bullish on financials, and they'll give

0:07:05.640 --> 0:07:09.800
<v Speaker 2>you a reason. And that the half life for that

0:07:09.960 --> 0:07:14.120
<v Speaker 2>sort of narrative. It's a week or two. I mean,

0:07:14.200 --> 0:07:19.600
<v Speaker 2>it's it. And what you want to do with that, honestly, Barry,

0:07:19.720 --> 0:07:22.920
<v Speaker 2>is you want to fade it. You want to look

0:07:22.960 --> 0:07:25.840
<v Speaker 2>for that to appear in your Bloomberg email in the

0:07:25.880 --> 0:07:32.200
<v Speaker 2>morning saying, oh, market experts bullish on financials because blah

0:07:32.280 --> 0:07:36.160
<v Speaker 2>blah blah. And when that drum beating gets pretty loud

0:07:36.280 --> 0:07:40.080
<v Speaker 2>when it's appearing in your morning email, I want to

0:07:40.120 --> 0:07:42.800
<v Speaker 2>fade it. You don't want to press it. You want

0:07:42.800 --> 0:07:46.640
<v Speaker 2>to fade it because this is just the ordinary business

0:07:46.720 --> 0:07:49.080
<v Speaker 2>of Wall Street. You've got to have an answer to why.

0:07:49.680 --> 0:07:54.520
<v Speaker 2>It's almost usually just variants or some combination of idiosyncratic stuff,

0:07:55.000 --> 0:07:57.080
<v Speaker 2>but you've got to come up with some blanket why.

0:07:57.520 --> 0:07:59.880
<v Speaker 2>And if that gets a lot of play, I want

0:07:59.880 --> 0:08:02.880
<v Speaker 2>to go the other way. Now there's a there's another

0:08:02.920 --> 0:08:07.040
<v Speaker 2>type of narrative though, where I want to press it.

0:08:07.760 --> 0:08:13.080
<v Speaker 2>And so that's what we call prescriptive narratives. It's not saying, oh,

0:08:13.200 --> 0:08:17.800
<v Speaker 2>the Fed is dubvish. It's saying, you know, Mohammad l

0:08:17.840 --> 0:08:21.680
<v Speaker 2>Arion will come out or I don't know, Trump will

0:08:21.720 --> 0:08:26.640
<v Speaker 2>come out and say the Fed should be dubvish. You

0:08:26.680 --> 0:08:30.400
<v Speaker 2>see the difference. It's not it's not describing what's already happened.

0:08:31.000 --> 0:08:35.520
<v Speaker 2>It's trying to lay the framework for what should happen.

0:08:36.000 --> 0:08:38.720
<v Speaker 1>So so let's delve into that. What sort of tools

0:08:38.720 --> 0:08:44.960
<v Speaker 1>are you using to identify these narratives, be they descriptive

0:08:45.120 --> 0:08:49.680
<v Speaker 1>or or prescriptive. How how much lead time do you

0:08:49.800 --> 0:08:55.640
<v Speaker 1>get to analyze you know, this giant corpus of noise

0:08:55.880 --> 0:09:00.360
<v Speaker 1>and commentary and data that is produced every single day.

0:09:00.440 --> 0:09:06.120
<v Speaker 2>Well, this is what good traders, honestly and and investors

0:09:06.640 --> 0:09:11.000
<v Speaker 2>have always done. They've they've they've always made these sort

0:09:11.000 --> 0:09:16.800
<v Speaker 2>of assessments of the news and the stories that are happening.

0:09:17.400 --> 0:09:19.680
<v Speaker 2>So this what we're trying to do is we're trying

0:09:19.720 --> 0:09:22.880
<v Speaker 2>to externalize what good traders and i'll call it short

0:09:22.960 --> 0:09:27.800
<v Speaker 2>term investors have always done. They've always internalized this. So

0:09:29.679 --> 0:09:32.320
<v Speaker 2>anybody can do this. You start looking for the words

0:09:32.360 --> 0:09:38.960
<v Speaker 2>that are trying to prescribe versus describe. That'll you you

0:09:38.960 --> 0:09:40.880
<v Speaker 2>can you can pick this up in whatever you do.

0:09:41.160 --> 0:09:44.640
<v Speaker 2>What we're able to do though today is because there's

0:09:44.679 --> 0:09:48.200
<v Speaker 2>now big data, as because there's big compute, is we

0:09:48.200 --> 0:09:51.840
<v Speaker 2>can read everything. Humans we're all limited in what we

0:09:51.840 --> 0:09:53.800
<v Speaker 2>can read, what we can pay attention to. In our

0:09:53.880 --> 0:09:58.080
<v Speaker 2>little corner of the world. What's possible today is to

0:09:58.160 --> 0:10:03.520
<v Speaker 2>read everything into animal lies exactly, this sort of word choice.

0:10:04.120 --> 0:10:07.439
<v Speaker 1>So what sort of output do you get from your

0:10:07.920 --> 0:10:14.280
<v Speaker 1>big compute that's scraping everything, Yeah, sifting through all the language, like,

0:10:14.440 --> 0:10:19.080
<v Speaker 1>does it identify specific words? Does it identify themes? I'm

0:10:19.080 --> 0:10:23.880
<v Speaker 1>assuming artificial intelligence and big data analytics is a key

0:10:23.960 --> 0:10:26.640
<v Speaker 1>part of this. What does the output look like to you?

0:10:27.200 --> 0:10:30.280
<v Speaker 2>Well, let me start with where it starts, because it's

0:10:30.280 --> 0:10:34.480
<v Speaker 2>not that the output comes from what you put into

0:10:34.520 --> 0:10:36.720
<v Speaker 2>it and what you want to put into it, and

0:10:36.760 --> 0:10:41.120
<v Speaker 2>this has to be human generated. Are well, what are

0:10:41.160 --> 0:10:43.960
<v Speaker 2>the ideas? What are the narratives that I care about?

0:10:44.960 --> 0:10:47.720
<v Speaker 2>Here's why that's important, Barry, because let's say you're a

0:10:47.760 --> 0:10:50.600
<v Speaker 2>value investor. You've got a vision, you know, you've got

0:10:50.600 --> 0:10:54.280
<v Speaker 2>a view on a stock, and like all value investors,

0:10:54.280 --> 0:10:59.520
<v Speaker 2>you think the market has not recognized the story about

0:10:59.559 --> 0:11:03.360
<v Speaker 2>this stuff that I think is really important. So what

0:11:04.120 --> 0:11:09.800
<v Speaker 2>you're looking for is not how loud is that story

0:11:09.920 --> 0:11:13.000
<v Speaker 2>right now? That story is dormant. You're looking for that

0:11:13.120 --> 0:11:17.160
<v Speaker 2>story to start being told. So you can't start this

0:11:17.320 --> 0:11:21.200
<v Speaker 2>from asking AI, Hey, AI, what are the prominent stories

0:11:21.280 --> 0:11:24.920
<v Speaker 2>right now, because it can only tell you what's being

0:11:26.080 --> 0:11:31.200
<v Speaker 2>spread right then and there. What I find the way

0:11:31.240 --> 0:11:34.160
<v Speaker 2>to really make money with narratives is to say, what

0:11:34.320 --> 0:11:39.360
<v Speaker 2>narrative is dormant today? I want to see when it

0:11:39.400 --> 0:11:43.320
<v Speaker 2>starts to get discovered. Is that discovery phase when the

0:11:43.400 --> 0:11:48.880
<v Speaker 2>market quote unquote wakes up to a story that you've

0:11:48.920 --> 0:11:52.600
<v Speaker 2>been looking for the market to wake up to. That's

0:11:52.640 --> 0:11:58.720
<v Speaker 2>how I find you can most reliably make money from

0:11:59.120 --> 0:12:00.000
<v Speaker 2>narrative investments.

0:12:00.800 --> 0:12:03.280
<v Speaker 1>So I get the sense we're still in the early

0:12:03.400 --> 0:12:08.280
<v Speaker 1>days of narrative investing as a key strategy, at least

0:12:08.360 --> 0:12:13.120
<v Speaker 1>in terms of using AI and big data. What do

0:12:13.160 --> 0:12:17.400
<v Speaker 1>you see developing in this space? What's the narrative about

0:12:17.520 --> 0:12:18.440
<v Speaker 1>narrative investing?

0:12:21.760 --> 0:12:24.040
<v Speaker 2>Barry? This is, like I say, this is an old idea.

0:12:25.000 --> 0:12:27.280
<v Speaker 2>So if you if you go back and kind of

0:12:27.280 --> 0:12:29.480
<v Speaker 2>look at your Wall Street history, or if you talk

0:12:29.520 --> 0:12:33.520
<v Speaker 2>to kind of traders today, what they're always looking at

0:12:33.640 --> 0:12:37.400
<v Speaker 2>is the news that comes along, and they're trying to say, hey,

0:12:37.400 --> 0:12:40.960
<v Speaker 2>do I fade that you know? Is that story worn out?

0:12:41.920 --> 0:12:45.920
<v Speaker 2>Is it is that? Is it topping over? Or does

0:12:45.960 --> 0:12:49.960
<v Speaker 2>this story have legs? Is this the start of its spread?

0:12:50.720 --> 0:12:55.520
<v Speaker 2>Like a virus. So it's an old idea. What's possible

0:12:55.520 --> 0:12:59.120
<v Speaker 2>today is to quantify it. What's possible today is to

0:12:59.320 --> 0:13:03.560
<v Speaker 2>external life what good traders have internalized in the past.

0:13:04.280 --> 0:13:09.520
<v Speaker 2>So it's not that it's you know, inventing cold fusion

0:13:09.679 --> 0:13:13.520
<v Speaker 2>or really doing anything that's new in the world. What

0:13:13.559 --> 0:13:16.160
<v Speaker 2>it is able to do, though, is to make that

0:13:16.200 --> 0:13:20.800
<v Speaker 2>sort of analysis, to look at not what reality is,

0:13:20.880 --> 0:13:24.840
<v Speaker 2>but how reality is being presented, and make it available

0:13:25.080 --> 0:13:28.959
<v Speaker 2>over a much wider scope, not over just that little

0:13:29.000 --> 0:13:31.880
<v Speaker 2>area that the good trader really knows a lot about,

0:13:32.600 --> 0:13:36.520
<v Speaker 2>and make it available over a lot more publications and

0:13:36.559 --> 0:13:41.160
<v Speaker 2>things that are being written because humans can't read everything. Really,

0:13:41.200 --> 0:13:44.079
<v Speaker 2>I think it's a way of I hate to call

0:13:44.120 --> 0:13:48.400
<v Speaker 2>it democratizing investing. I really hate that idea, but it

0:13:48.440 --> 0:13:51.600
<v Speaker 2>really is a notion of creating a new instrument so

0:13:51.760 --> 0:13:57.200
<v Speaker 2>every investor can understand, well, what are the stories that

0:13:57.280 --> 0:13:59.040
<v Speaker 2>are being told about the world today.

0:14:00.120 --> 0:14:04.960
<v Speaker 1>Wrap up technology has enabled us to read much more

0:14:05.559 --> 0:14:09.400
<v Speaker 1>than we were capable of reading as individuals. In fact,

0:14:09.480 --> 0:14:12.920
<v Speaker 1>the machines can read everything, and we can use those

0:14:12.920 --> 0:14:19.640
<v Speaker 1>machines to identify dormant narratives that might lead to increases

0:14:19.880 --> 0:14:24.880
<v Speaker 1>in either a particular stock or asset class or market

0:14:25.920 --> 0:14:27.080
<v Speaker 1>fair statement.

0:14:27.560 --> 0:14:31.920
<v Speaker 2>Very fair. What you're looking for is what's new, what's changed,

0:14:32.480 --> 0:14:36.480
<v Speaker 2>because when the narrative changes, that changes behavior.

0:14:37.040 --> 0:14:52.640
<v Speaker 1>Fascinating stuff. I'm Barry Riddtolts. This is Bloomberg's at the Money,