Value Driven Data Science: Boost your impact. Earn what you’re worth. Rewrite your career algorithm.

00:00:00 Dr Genevieve Hayes
Hello and welcome to value driven data science brought to you by Genevieve Hayes Consulting. I'm doctor Genevieve Hayes. And today I'm joined by Todd Tressider to discuss some of the unexpected risks of applying data science to retirement modelling. Todd is a former hedge fund manager who retired.
00:00:20 Dr Genevieve Hayes
At age 35 to become a financial consumer advocate and money coach.
00:00:25 Dr Genevieve Hayes
He now runs the popular retirement planning website financialmentor.com and is the author of a range of books on retirement planning and investments, including how much money do I need to retire and the leverage equation. Todd, welcome to the show.
00:00:43 Todd Tresidder
Thanks for having me, Jenny.
00:00:45 Dr Genevieve Hayes
I am a long time finance nurse. My original training before data science was in actuarial studies and finance, and as a result I read finance.
00:00:56 Dr Genevieve Hayes
Books for fun.
00:00:57 Dr Genevieve Hayes
And that was how I came across Todd's work. How much money do I?
00:01:00 Dr Genevieve Hayes
Need to.
00:01:00 Dr Genevieve Hayes
Retire sounds like a retirement planning book.
00:01:03 Dr Genevieve Hayes
But what I found in addition to some really fantastic advice about financial planning and financial modelling is one of the best data science books that I've ever come across because.
00:01:17 Dr Genevieve Hayes
It turns out that in addition to concepts like asset allocation, there's a lot of data science and data modelling that comes behind retirement planning, and if you get that wrong, the consequences can be catastrophic.
00:01:34 Dr Genevieve Hayes
And that's something that I really want to look at in today's episode before we go on a quick disclaimer, listeners should be aware that the content of this podcast is general in nature and is intended for educational purposes only. It does not constitute general or personal financial advice.
00:01:54 Dr Genevieve Hayes
OK, with that out of the way.
00:01:56 Dr Genevieve Hayes
Todd, can you explain to our listeners how you came to write? How much money do I need to retire?
00:02:02 Todd Tresidder
Yeah. So what happened was I started my career from the quantitative side. So I was one of the early pioneers of a quant hedge fund. So that means that we ran everything by the.
00:02:13 Todd Tresidder
Numbers. And so there's two ways that humans make sense out of complexity. One is data numbers, and the other is narrative.
00:02:21 Todd Tresidder
And so I've always come from the data number side. I'm not a math genius, but I'm extremely intuitive with numbers and I have a really good smell test for bull in proper application of numbers at the quantitative.
00:02:34 Todd Tresidder
Hedge fund I was one of the early pioneers of modelling trading systems, which now you've got MIT PHD's doing it.
00:02:39 Todd Tresidder
But I mean, I was doing it back when IBM came out with their first computer with the 8088 processor, which of course is a boat anchor now.
00:02:47 Todd Tresidder
But I was having to input all my own databases by hand, 10 key by touch to build out databases to start modelling the financial markets.
00:02:55 Todd Tresidder
So that was where I cut my teeth and you Start learning really quick. What's valid and what isn't and what's the proper application of data and what is the misuse of data? Because in the financial markets you get punished immediately if you do it wrong.
00:03:08 Todd Tresidder
You can even do it right and get punished because in the end what you're dealing with is probabilistic outcomes.
00:03:15 Todd Tresidder
You're never dealing with deterministic models, you're only dealing with probabilities, and so you have to become really savvy.
00:03:21 Todd Tresidder
Be at what you can know and what is forever unknowable. And so I spent all this time. I was one of the early pioneers of financial modelling in the financial markets, which is, you know, now you see it, these guys are quant traders and they got banks of computers on the trading floors.
00:03:36 Todd Tresidder
And there's all this stuff going on. Back when I was doing it, nobody was doing it.
00:03:39 Todd Tresidder
As it turns out, the.
00:03:41 Todd Tresidder
Stuff that worked back then still works today because it's fundamentally true and it.
00:03:46 Todd Tresidder
Based on basic truths about how the financial markets behave, correct modelling starts with financial truths, and then you bring quantitative discipline to it, not the other way around.
00:03:56 Todd Tresidder
You get a lot of people who come in and do math geek stuff to the data and then they'll completely blow it.
00:04:01 Todd Tresidder
So anyway, at age 35, which anybody seeing the picture right now would see that's a long time ago, I'm obviously a lot older than that now.
00:04:09 Todd Tresidder
At age 35, I retired. We sold the hedge fund and I had enough to retire and that was from compounding the wealth I created in the hedge fund.
00:04:17 Todd Tresidder
And I started looking at retirement modelling and I was absolutely shocked. I was stunned for somebody that came from a financial modelling background and from a data background to look at how these guys were modelling retirement planning was just absurdly, laughably bad. It was so missing the.
00:04:37 Todd Tresidder
Nuance and the financial truth that I was just kind of blown away. And so that's when I started teaching, I had I had a financial coaching business at the time.
00:04:47 Todd Tresidder
I just was curious, could I teach people how to do this stuff? So I started a little boutique coaching business.
00:04:53 Todd Tresidder
And one of.
00:04:54 Todd Tresidder
The questions that always came up was how much money you need to retire, and I'd had to figure that out when I sold the hedge fund.
00:04:59 Todd Tresidder
And of course I did it in a completely different way, and so I created this little pamphlet, which was like the grandfather of the book that is now published.
00:05:08 Todd Tresidder
And it went through multiple iterations. It used to sell back when we sold PDF's online for $99.
00:05:13 Todd Tresidder
Or something, you know? First it got passed around a lot. Then I decide. Oh, people have enough value to pass it around.
00:05:18 Todd Tresidder
I'll do it as a PDF and it was sold for a bunch of money as PDFs and eventually Amazon.
00:05:22 Todd Tresidder
Was on, came in and took over the Bookmarket, and so then it dropped to a $9 book, even though it had been through 10 iterations and was infinitely better than the original PDF.
00:05:32 Todd Tresidder
But that's kind of the genesis of it was when I had to do my own retirement planning and figure out how much money was enough and how do you convert assets in a volatile market into an income stream. You won't.
00:05:43 Todd Tresidder
Live. There's just so much stuff that people get wrong. And so again, I just came at it from a very different viewpoint, which was not. I'm a broker trying to sell you assets. I came at it from a guy that.
00:05:54 Todd Tresidder
Had a numbers background in a model building background that had been tested in the financial markets to where I could smell out the flaws.
00:06:03 Todd Tresidder
And so that's where it all came from is that that's kind of the history Genesis story.
00:06:08 Dr Genevieve Hayes
One thing I thought was really interesting when you were talking just then was your comment of how you need to start with the financial truths and then move on to the.
00:06:15 Dr Genevieve Hayes
Quantitative truth.
00:06:17 Dr Genevieve Hayes
Because that's one of the principles of data science. You meant to start with your domain knowledge and then move on to your model.
00:06:24 Todd Tresidder
See, I didn't even know that because I'm not a trained data scientist, right? I had to learn that the hard way. So go ahead. Tell me more.
00:06:30 Dr Genevieve Hayes
I think that's where a lot of the data scientists fall short. The best data scientists are the ones who started with their domain knowledge in whatever their discipline might be.
00:06:39 Dr Genevieve Hayes
I've spoken to a number of people post on this podcast and elsewhere about the water industry, the best water data scientists are the ones who started off as engineers in the water industry and then built.
00:06:50 Dr Genevieve Hayes
Models from.
00:06:51 Dr Genevieve Hayes
That, and I see that in what you're saying here, but a lot of data scientists take their quant skills and then think they're universally applicable.
00:07:00 Todd Tresidder
Yeah, I actually, in my community, I call it the engineers fallacy because I have all these really high end engineers.
00:07:06 Todd Tresidder
Like I'm thinking of one guy out the top of my head. He's a high end Google engineer. He's one of the one of my clients in my wealth.
00:07:12 Todd Tresidder
Building community.
00:07:13 Todd Tresidder
The engineers always make the same mistakes right off the bat. They come in with all their engineering training and they go in.
00:07:19 Todd Tresidder
They optimise models based on the data that's available and I'll say you can't do that. It's not deterministic. You have to understand the limitations of the data and what it can tell you, what it can, and so the correct way to do it is. You start with what are financial truths.
00:07:34 Todd Tresidder
And then you look at the data and you build algorithms and you say within the limitations of the data, can I build an algorithm that will adapt to the known data? But then logically will also adapt to the unknown data that will occur at some point in the.
00:07:49 Todd Tresidder
Sure, right. Because you only have so many regimes, so many economic regimes, only so many types of things that have occurred within the data sets that we have.
00:07:58 Todd Tresidder
But there's also other economic regimes that can and will occur in the future that aren't built into the data sets.
00:08:05 Todd Tresidder
There's an art and the science to the whole process. The science is working with the data. You have to prove out your theories.
00:08:12 Todd Tresidder
Based on the data you have, but you can't go the other direction, you can't just start with data and say here let's mine the data and come up with an optimised model and not optimise set of parameter sets for a given algorithm to profit from this data set. You can't do that.
00:08:26 Todd Tresidder
Direction. You have to start from the other side and say what are the financial truths? How do financial?
00:08:30 Todd Tresidder
Markets behave.
00:08:31 Todd Tresidder
And so the great example of this foible, the this mistake and it's a huge mistake, is long term Capital Management.
00:08:40 Todd Tresidder
So in the middle 1990s, long term Capital Management was formed. They had 17 pH D's and Nobel laureates.
00:08:46 Todd Tresidder
And staff, they had more data science and math skill than most university.
00:08:52 Todd Tresidder
Math departments have on staff and they were out of business in four years. They started. They went huge because they were successful right off the bat.
00:09:01 Todd Tresidder
And then as soon as the Russian debt crisis hit, they were suddenly dealing with out of sample data. They didn't have any data for anything to do with the Russian debt crisis. So they hit out of sample data.
00:09:12 Todd Tresidder
Their models did not adapt, they were not designed to be adaptive, and so as soon as the out of sample data hit, it was a valuation driven model of pairs trading.
00:09:23 Todd Tresidder
And so as soon as the pairs diverge, which they're going to do in a default crisis, the pairs that they would normally trade diverge.
00:09:31 Todd Tresidder
So their model said trade, more trade more and so they went bigger and bigger and they just compounded their losses geometrically when they should have immediately cut capital losses, you know, to conserve capital because that's again in the geometric growth, the the asymmetric growth of compounding wealth, right. There's math laws that govern how wealth compounds and it's asymmetric.
00:09:52 Todd Tresidder
And so one of the first rules is you have to cut your losses. And that's not Todd's opinion.
00:09:56 Todd Tresidder
That's just the math truth of how wealth compounds, right? And that's built into all models that are successful long terms. You have to know how to control risk.
00:10:04 Todd Tresidder
Well, they did the opposite and they just blew up like fabulously. It was such.
00:10:08 Todd Tresidder
An huge blow.
00:10:10 Todd Tresidder
Up that it required government intervention because what happened?
00:10:14 Todd Tresidder
They were such a big player because they were such a big profile firm and all these big names and big brands on staff, that money through at them. So when they blew up they got so big, so fast.
00:10:24 Todd Tresidder
That they were actually an influence on the markets. So then the other thing they didn't have in.
00:10:28 Todd Tresidder
The data, which was just beauty.
00:10:29 Todd Tresidder
Right. Is all the traders knew what their positions were because the models were known and so everybody traded against them and multiplied their losses, knowing they had to unwind the positions.
00:10:40 Todd Tresidder
Now, of course, none of that's in the data, but anybody that's in the financial markets knows that's what's going to occur. That's how the markets behave. So these are financial truths.
00:10:50 Todd Tresidder
Right, this is how the markets work. You have to.
00:10:52 Todd Tresidder
Have the market savvy first.
00:10:54 Todd Tresidder
And then what you do is you bring the quantitative discipline behind it to deal with the human foibles, the human errors and judgement that occur because we're emotional human beings.
00:11:04 Todd Tresidder
We make decisions emotionally, not rationally. We'll rationalise them and support them, but our philtres and our brains will logically pull the data and the information that supports our biases.
00:11:16 Todd Tresidder
So you build algorithmic models based on data to create discipline and reinforce, but they have to be built first and foremost on financial truths. And one of those financial truths includes risk management, but it also includes so many other things they.
00:11:29 Todd Tresidder
Yelled at.
00:11:30 Dr Genevieve Hayes
What are the?
00:11:31 Dr Genevieve Hayes
Most important financial truths when it comes to retirement financial planning.
00:11:36 Todd Tresidder
Ohh God, where do I start Genevieve? That's too big of a topic, but they're all over the place. And you build a knowledge of them with experience. But.
00:11:44 Todd Tresidder
Having me pull them off the top of my head, I'm not going to be able.
00:11:47 Todd Tresidder
To just.
00:11:47 Dr Genevieve Hayes
Throw them down like that one line that really stood.
00:11:50 Dr Genevieve Hayes
Out to me in.
00:11:50 Dr Genevieve Hayes
Your book was statistics don't work when you're a sample size of one. I really love that.
00:11:56 Dr Genevieve Hayes
Mine because I think a lot of data scientists forget that the vast majority of the techniques they're working with are based on.
00:12:04 Dr Genevieve Hayes
Changes. So once you get down to a sample size below a certain point, everything's obviously going to fall apart.
00:12:12 Todd Tresidder
Oh yeah. I mean, that's one of the huge mistakes in retirement planning, right? Is they'll talk about average results and average life expectancy and average this and average that.
00:12:22 Todd Tresidder
And yet, for every human being, they're a sample size of 1 and they have one life, so they don't have a mathematically or statistically derived outcome.
00:12:32 Todd Tresidder
And so you have to play your probabilities and your risk management completely different unless you.
00:12:40 Todd Tresidder
Slough off the risk via things like insurance. Right. Insurance is where they're increasing. They're taking and pooling a large data sample into a statistically valid sample size, and then they profit from a percentage of it, right?
00:12:53 Todd Tresidder
So that's essentially what life insurance is the actuaries well, you know this is a former actuary. They'll figure out what the pricing is on life insurance. And I'm just going off of.
00:13:03 Todd Tresidder
Rendered life insurance. Right? I'm not getting into whole life which it creates all kinds of complication. They'll figure out life expectancies and probabilities price out the life insurance, and then on a.
00:13:13 Todd Tresidder
Large pool of.
00:13:14 Todd Tresidder
People, statistically, there's a profit margin involved.
00:13:17 Todd Tresidder
Plus, the life insurance company then has the float on all of the premiums that are paid on the policies.
00:13:23 Todd Tresidder
Which then they can compound in the background so they have a profit on the policy and they profit on the float.
00:13:27 Todd Tresidder
And that's one of the reasons they're one of the most profitable companies out there, which is why you have.
00:13:32 Todd Tresidder
People like Warren Buffett and his mentor before him built their investment management companies wrapped inside an insurance company, so they wrapped the investment management inside because they're building wealth off the float of the premiums. It's statistical. But in in terms of you and me, I can't plan my retirement the same way.
00:13:53 Todd Tresidder
And here's another thing that's going on that's absolutely fascinating retirement planning. As we speak, there's an absolute biotech revolution going on that's just accelerating. And the insights that are already coming in, in terms of long.
00:14:05 Todd Tresidder
Levity and ageing science is just now beginning, and it's growing geometrically and the expected lifespan is for even me at 62 as we record, this is probably much, much much longer than what has historically been true.
00:14:27 Todd Tresidder
For any expectation that we would have in retirement planning, there's already multiple models of expanding your expected lifespan by 1015% individually and on mouse models, they've combined them for a 50% life extension.
00:14:40 Todd Tresidder
And this is already done, and they're just at the starting point. So the implications for retirement planning are enormous, and that comes back to the sample size of 1.
00:14:51 Todd Tresidder
You could go through and probabilistically figure this out, but one of the risks everybody takes is life expectancy, and so that varies in ways that don't even fit math models, right, because you have to understand what's going on in the world.
00:15:04 Todd Tresidder
Around you too.
00:15:06 Dr Genevieve Hayes
And I mean with a life insurer, if one person lives to 105, they don't care because, well, if everyone else has died at age 75 or whatever, the life expectancy is, they've got their money.
00:15:18 Dr Genevieve Hayes
But if you personally lived to 105 and you haven't made proper plans, you're stuffed.
00:15:24 Todd Tresidder
Ohh yeah, and that's a conservative.
00:15:26 Todd Tresidder
Number I think so. Anyway, this is it's really fascinating stuff when you dig into it, when you start to understanding like the limitations of what's known, what's forever unknowable, what are the current developments.
00:15:40 Todd Tresidder
What are deterministic models? What can the data tell you? What does the financial truths tell you and you start weighing all these out.
00:15:46 Todd Tresidder
You have to really think through it. There's a lot of nuance to it, and yet it's taught so simplistically.
00:15:53 Dr Genevieve Hayes
For those listeners who haven't come across a traditional financial model for retirement, can you give the listeners an overview of how financial modelling is traditionally done?
00:16:05 Todd Tresidder
Traditionally what you do, and I'm going to say this sarcastically, right, so you'll hear the sarcasm in my voice, but traditionally what you do is you work a lifetime to, you know, to age 65 or whatever the retirement age is.
00:16:16 Todd Tresidder
Right. So you work a lifetime to 65, you have your career, you scrimp and save to save a certain percentage of the money you earn, you put it in tax deferred accounts and you hand it to your genius financial planner, who has this magical asset allocation formula that's going to give you a pot of gold at the end of the rainbow. And then at this Inflexion point, as called your retirement age.
00:16:36 Todd Tresidder
You suddenly go from earning in a productive life where you scrimp and save, and now suddenly you're doing nothing, and now you're in this amortisation equation off of your savings.
00:16:47 Todd Tresidder
And the magical asset allocation is going to provide this safe withdrawal rate that has been actuarially determined from past history.
00:16:54 Todd Tresidder
But unfortunately, that's usually U.S. history, which is the prom queen of the economic world for the data period over which it was analysed, and that same data does not apply to any other country, and the safe withdrawal rate.
00:17:07 Todd Tresidder
The US fails on every other country and on and on and on. I mean, I could just go in circles on this.
00:17:14 Todd Tresidder
And yet that's what's done for probably 99% of financial plans out there. Now here's what they'll do right is they'll come in and they'll say.
00:17:22 Todd Tresidder
Ohh, but we're so smart. We're data scientists and we've got the cutting edge, so we're going to bring in a Monte Carlo simulation. So what's a Monte Carlo simulation? I mean, you know, you're smiling.
00:17:33 Todd Tresidder
That I'll just.
00:17:34 Todd Tresidder
I'll just kind of ramble it for a second, and again, I'm being intentionally simplistic and sarcastic to make a point. There's more to it than what I'm saying, right? But I'm trying to drive home a point here.
00:17:45 Todd Tresidder
And so money colour simulation is the Randomising pass data or the randomising data entirely based on certain statistical assumptions, and they're trying to create a Monte Carlo distribution of your risk of ruin or success, and ruin is defined as living longer than your money, and success is determined as.
00:18:05 Todd Tresidder
Having more money than life and so they'll do it as a Monte Carlo simulation. But again, now you've got to come back to assumptions and financial.
00:18:12 Todd Tresidder
Use is data random? No, it's not OK volatility clusters. OK. And that's a you asked me for financial truth.
00:18:21 Todd Tresidder
So there's another financial truth, OK, volatility clusters, both your largest winning days and your largest losing days are all below trend line. They're all in bear markets and so.
00:18:33 Todd Tresidder
You know, volatility rises during bear markets, correlations of assets all rise during bear markets. Correlations rise towards one. These are basic financial truths.
00:18:41 Todd Tresidder
And so if you randomise data in a Monte Carlo simulation, guess what you're doing? You're violating financial truth. That's not actually the way markets work.
00:18:50 Dr Genevieve Hayes
So you're ignoring the autocorrelation.
00:18:53 Todd Tresidder
Yeah, there's serial autocorrelation in the data at one time frame, and then there's also mean reversion at a longer time frame.
00:18:59 Todd Tresidder
They both coexist within the same data set, but they exist in different time frames, so that's why, like valuation, investing is valid. So Warren Buffett's model is valid. Basically, valuation investing is.
00:19:12 Todd Tresidder
Fancy term for mean reversion in the data.
00:19:15 Todd Tresidder
Only there's fundamentals behind it that help you determine when you're finding a value, etcetera, etcetera. But you can even take in terms of a mean reversion data.
00:19:24 Todd Tresidder
There's there's valid models that look at assets that have declined 90% or more and what are the probabilities of the rise over the next five years and things like that. They're very well done, right? So you can do it as a simple mean reversion.
00:19:36 Todd Tresidder
Model or you can bring in valuation criteria and suddenly you've got that. But that exists in one time frame. Serial autocorrelation exists at a much shorter time frame and his trade will using trend volume.
00:19:48 Todd Tresidder
And so there's these different models that exist in different things based on different principles in the data, which are really financial truths.
00:19:56 Todd Tresidder
So that was a bit of rambling, did I?
00:19:58 Todd Tresidder
Hit your question.
00:19:59 Dr Genevieve Hayes
Oh, yeah, yeah, yeah. What you're saying? I find it hilarious because when I was doing actuarial work, Monte Carlo simulations come up all the time.
00:20:07 Dr Genevieve Hayes
And that was what I loved about your book. It's basically taking everything that I've learned and making me realise, yeah, this applies at a insurer level, but it does not apply at a personal level.
00:20:20 Todd Tresidder
Correct. Yeah. And so I don't want to be excessively belittling. I was being playful to make a point. They aren't totally useless.
00:20:28 Todd Tresidder
They're part of a whole picture of knowledge. What happens, what goes wrong is when people present things as truths, when they're just a piece of understanding of an idea. So Monte Carlo simulations.
00:20:41 Todd Tresidder
Have value, particularly if they're done well and people understand the limitations of them. They have value to portray risks and probabilities.
00:20:49 Todd Tresidder
Is if you understand what they're doing where I have a real problem with them though, I'll give you. I'll give you client examples.
00:20:55 Todd Tresidder
So let's say we're coming into the 2021 top or we're coming into the 1999 top.
00:21:01 Todd Tresidder
We're coming into a period characterised by high market valuation, so you can measure it as Cape ratio, which is cyclic adjusted price earnings ratio.
00:21:10 Todd Tresidder
You can measure it by any number of models, right? Because ultimately the value of any asset is discount present value of the cash flows, and there's a variety of models that will discount those present values of cash flows into different forms, and they all come up with highly correlated results, right? So I'm not gonna hang up on the model.
00:21:26 Todd Tresidder
But you come in with these financial models and they'll show extreme high valuation. Now we know a financial truth is all major bear markets and all extended periods of non performance in the markets are preceded by a period of high valuations and or record low interest rates. Combination is particularly lethal.
00:21:47 Todd Tresidder
And so.
00:21:48 Todd Tresidder
If you come into a period of high valuations, if if you look at simulations of safe withdrawal rates without Monte Carlo, you'll see that all the lowest periods follow periods of high valuation.
00:21:59 Todd Tresidder
Why, of course, because volatility clusters on the downside and the precursor is the high market valuation. These are all financial principles, financial truths.
00:22:08 Todd Tresidder
And there's reasons why they're true that we can get into, but they are truths, OK? They're how the financial markets work. It's how pricing works in the financial markets.
00:22:17 Todd Tresidder
It's not random, as commonly taught in academia, it's just not perfect, and that's why you can make it look random with improper use of statistics.
00:22:29 Todd Tresidder
So anyway, I would have clients come to me at these periods of high valuation and they'll say, well, my financial planner says I'm safe. There's only a 5% or 7% risk of failure.
00:22:40 Todd Tresidder
Using these advanced Monte Carlo simulations and I'll turn around and say, well, did your financial planner tell you that you happen to be sitting at the exact precipice of that 5% data set?
00:22:50 Todd Tresidder
In other words, you're almost 100% risk of failing because of the time period and the market characteristics you're in.
00:22:57 Todd Tresidder
And they'll be like, no, I don't understand what you're talking about, you know? And so then I would explain it, and they go back to the financial plan and the financial planner wouldn't know anything about it.
00:23:05 Todd Tresidder
You can't look at a 5% seven percent, 10% risk of ruin on a Monte Carlo simulation and determine that you're safe because those periods of ruin exist. You're a sample size of 1.
00:23:18 Todd Tresidder
And certain situations precede those periods of ruin, and so you could actually have a 20% risk of ruin, but you could have a Cape ratio that's in single digits.
00:23:30 Todd Tresidder
Your risk of failure, even if your numbers show, is at 20% of ruin are probably close to.
00:23:35 Todd Tresidder
Zero because you've got risk management in terms of valuation that says that none of those probabilistic outcomes are likely to occur.
00:23:43 Todd Tresidder
You're more likely to occur at a very high safe withdrawal rate, much higher than you're actually forecasting you're actually going to net net build wealth, even though your risk of ruin is higher than the guy that had 10% or 7% risk of ruin.
00:23:56 Todd Tresidder
But he was at a high valuation market. You know, again you have to start with financial truth. You have to understand what the data can tell you, what it can, and you have to understand how to work with all this.
00:24:05 Todd Tresidder
Stuff intelligently, it's.
00:24:07 Todd Tresidder
Just like you said, it's not. It's not commonly talk.
00:24:10 Dr Genevieve Hayes
What I'm getting? I mean, if you're saying to someone the Monte Carlo simulation says the 95% of the time you'll be fined 5% of the time, you'll be ruined.
00:24:20 Dr Genevieve Hayes
The implication of that is that it's like a lottery and one ping pong ball is drawn out that says ruin or success completely at random. But.
00:24:30 Dr Genevieve Hayes
What you're saying is that there are circumstances associated with those 5% and circumstances associated with those 95% and you can have a good idea as to what's more likely a success ping pong ball or a ruin ping pong ball.
00:24:47 Todd Tresidder
Harshly correct. There's a nuance to what you said, though. It's even worse than that, because Monte Carlo is randomising actually data.
00:24:54 Todd Tresidder
Usually, sometimes they'll build random data based on statistical probabilistic outcomes and statistical volatility and things like that. They'll actually build random data. A lot of them just randomise existing data if it's randomising existing data.
00:25:08 Todd Tresidder
Then yeah, pretty much. But again, the data isn't actually random and so it's sort of a fundamental flaw in what's being done.
00:25:17 Todd Tresidder
So there's nuance to what you're saying, but it's really close. It's really close to accurate what you're saying.
00:25:23 Dr Genevieve Hayes
And when Rowan does occur, how bad is it? Are we talking just out by a couple of 1000?
00:25:28 Dr Genevieve Hayes
And or people having to go back to work.
00:25:30 Dr Genevieve Hayes
In their old age.
00:25:31 Todd Tresidder
Yeah, ruin is defined as you run out of money before you run out of life, right? That's that's what failure would be in that.
00:25:38 Todd Tresidder
And so the way it's worked that I've seen it in practise is usually once people are down and this is another flaw in like the whole modelling thing is these these models are run out in perpetuity, usually 30 years. The problem is, that's not how it works in real life.
00:25:53 Todd Tresidder
Because if you, let's say, you retire at 65 and five years in, you're down 50% or you're down 10 years in, you're down 50%. I don't know anybody that's not going to change their spending and change their life.
00:26:05 Todd Tresidder
If you're relying on that nest egg and half of it's gone in the first five or ten years, you're going to cut back, you're going to change how you live, you're going to change everything.
00:26:14 Todd Tresidder
And so risk of ruin has sort of in this amorphous thing, like they define risk of ruin mathematically running out the full 30 years. But the way it works in practise is a person's entire retirement.
00:26:25 Todd Tresidder
This dog gets changed because they have to adapt to try to protect what they have left and they'll change asset allocations.
00:26:32 Todd Tresidder
They'll cut off the risky side of the asset allocation. They'll make all kinds of adjustments.
00:26:36 Todd Tresidder
Some of them smart, some of them.
00:26:37 Todd Tresidder
Really dumb because it's frightening if you're completely dependent on the money and you have no additional income coming in and you lose a big chunk of it. I don't care what the data tells you. You're gonna act a specific way reliably.
00:26:50 Dr Genevieve Hayes
And that's another financial truth.
00:26:52 Todd Tresidder
Yeah. See, that's what I said is like you'll get them out of me as we talk, right? You'll just keep hearing them over and over again as we go through it, where the modelling doesn't match financial reality or human reality. You're really dealing with a couple of different things here you're dealing with.
00:27:05 Todd Tresidder
The reality of humans and how they behave with finance. You're also dealing with financial markets and how they behave statistically and how humans work within the financial markets.
00:27:17 Todd Tresidder
There's, like, there's kind of layers of just experiential knowledge that's in there and then you bring in the models and you understand what the modelling can and cannot do.
00:27:27 Todd Tresidder
And again, I'm not trying to put it. I I I mean I am a model junkie. You can see it in my books.
00:27:32 Todd Tresidder
I'm a total model junkie. They're great. I'm a quant. My whole background is Quant is applied to finance.
00:27:39 Todd Tresidder
Right. But you have to be careful. And again where this all came from was when I ran in the quantitative hedge fund, it was called market math.
00:27:49 Todd Tresidder
That was the name of the fund market math. The whole thing was done mathematically. And so if you're wrong, if you have an assumption and you do something wrong.
00:28:00 Todd Tresidder
The markets punish you by loss.
00:28:01 Todd Tresidder
Using money and so you learn very quickly what works, what doesn't, what the limits of knowledge are.
00:28:07 Dr Genevieve Hayes
So you've just gone through how to do it wrong or how to do it less well than you could. How do you do retirement financial modelling, right?
00:28:18
Ohh boy.
00:28:20 Todd Tresidder
So I don't wanna pitch my course but like I have a course called expectancy. Well plan.
00:28:26 Todd Tresidder
OK, so when you understand financial modelling done right, it starts with two equations. OK, I called it expectancy. Well, planning for a reason.
00:28:35 Todd Tresidder
And I had another course. I haven't built it yet, called expectancy investing. The reason they're both started with expectancy is your wealth compounds according to mathematical expectancy, which is probability times payoff.
00:28:46 Todd Tresidder
So it's really two.
00:28:47 Todd Tresidder
Equations that govern your wealth growth and govern your financial outcome in life. It's mathematical expectancy, which is your growth rate, and then future value, which is your growth rate times time.
00:28:56 Todd Tresidder
So within that you've got a set of variables, right? So you've got, you've got time, you've got the capital or how much you save and then you've got your compound rate of growth.
00:29:07 Todd Tresidder
And then within compound rate of growth, you've got basically what is probability times payoff. And again, I'm simplifying it for the audience. The math is more complex, but to.
00:29:16 Todd Tresidder
We get it in verbal form so we can work with it. Essentially, expectancy is probability times payoff. Now the problem for most people is they think in terms of probability.
00:29:26 Todd Tresidder
What are the odds that something's gonna happen? What is the likelihood I should say that something is going to occur? Is the market going to be up or is the market gonna be down?
00:29:34 Todd Tresidder
So to do it right, you have to weigh in the thing that's not intuitive, which is the payoff equation.
00:29:39 Todd Tresidder
And that's the thing that most people miss. And as it turns out, the payoff equation is absolutely critical to how all this works. So conventional finance tells you to ignore volatility. They tell you that interim volatility is not important.
00:29:54 Todd Tresidder
That's absolutely false. It's provable mathematically, and then it's also provable if you look at the distribution of returns of both systems and the markets, what happens is large volatility literally determines the distribution of returns.
00:30:10 Todd Tresidder
And the distribution of returns determines proper strategy and So what you want to do is you want to understand the payoff equations.
00:30:18 Todd Tresidder
Let me give you a little bit of background piece. I I got to fill in. I define investing as putting capital at risk into an unknowable future, really critical that you get in your head that the future is unknowable.
00:30:29 Todd Tresidder
All the financial media is gonna be filled with pundits telling you what's gonna happen in the future. It's all nonsense.
00:30:35 Todd Tresidder
Ignore it. Nobody knows what's gonna happen in the future. You can be right. A lot of the times, but you can't be right off enough to put capital at risk.
00:30:41 Todd Tresidder
And all those things are provable, you know, and I have examples in my writing of things like I used to trade commodity.
00:30:46 Todd Tresidder
Pictures and the Chernobyl blow up occurred. Nobody knows the future of a nuclear meltdown. If they did, it wouldn't happen, right?
00:30:54 Todd Tresidder
A nuclear meltdown is a a noble currents. All the markets went lock limit against me one day overnight. That's an unknowable future.
00:31:00 Todd Tresidder
You have to know how to manage risk in those inevitable outcomes. You know, Ronald Reagan got shot while I was managing.
00:31:06 Todd Tresidder
Money presidential assassination attempts. Not a nullable future, and the list goes on and on and on. If you're going to be in this game long enough, the unknowable future is going to bite you in the rear, so you have to manage the extreme tail of pay offs.
00:31:19 Todd Tresidder
And you have.
00:31:19 Todd Tresidder
To know how to do that, the thing about putting capital at risk in an unknowable future is the probability is not known.
00:31:26 Todd Tresidder
If your growth rate is determined by probability times payoff, then you're really dealing with the payoff equation as the noble, manageable outcome, and that's contrary to everything taught in finance.
00:31:38 Todd Tresidder
Everything taught in finance is about how to find a winning investment, right? Everybody wants to know the the 10 best investments for 2024 or whatever it is.
00:31:46 Todd Tresidder
Think of all the headlines you see in the magazines and in the.
00:31:49 Todd Tresidder
Media and when?
00:31:50 Todd Tresidder
Really, the game is defence. The game is defence.
00:31:55 Todd Tresidder
It's in the.
00:31:55 Todd Tresidder
Payoff equation because that's the manageable part where you can tilt the growth rate by controlling the losses and maximising the gains.
00:32:03 Todd Tresidder
And so it's completely opposite of what's commonly taught.
00:32:07 Dr Genevieve Hayes
How do you even know what the payoff equation is?
00:32:11 Todd Tresidder
You can only manage for the loss. You never know, right? You can manage the loss, which tilts the outcome of the payoff equation.
00:32:19 Dr Genevieve Hayes
So you can basically put in some form of insurance that limits your loss is that?
00:32:25 Dr Genevieve Hayes
What you're saying?
00:32:25 Todd Tresidder
That would be one example. So basically the way I teach it is risk management.
00:32:29 Todd Tresidder
The game. So let's do a sports analogy just to make this intuitive right cause it we get into data and we get into finance and people get lost, whereas sports everybody has a favourite sport, a favourite team sport.
00:32:39 Todd Tresidder
And so in your favourite team sport, what team wins the national championship? Is it the team with the world's greatest offence or the team with the world's greatest defence?
00:32:50 Todd Tresidder
As it turns out, it's neither.
00:32:52 Todd Tresidder
Usually the championship team is in the top 1020% in both categories. It's usually not the top of either offence or defence, but it's really great at both.
00:33:03 Todd Tresidder
As it turns out, that's the game you gotta play and investing to. It's true for any competitive endeavour. If you've got a really great defence, what that means is they don't allow many points to the opponent.
00:33:13 Todd Tresidder
So the opponent doesn't put many points on the board. That means your offence doesn't have to score as many points to create a positive out.
00:33:20 Todd Tresidder
Not only that, if the defence is really good, that means the offence gets they control the ball, the defence controls the ball by being really good.
00:33:29 Todd Tresidder
That means the offence gets more attempts at goal. They have more time and offence and they have more tries on goal to score and so the combination creates A consistent win. It's a reliable win.
00:33:41 Todd Tresidder
It's the combination of a viable offence, in other words, an offence that has a proven positive mathematical expectancy, doesn't have to be the best winning strategy, it just has to be reliable through all economic regime.
00:33:53 Todd Tresidder
Teams. And then you have to have a bulletproof defence so you never have large points scored against you. In other words, you never go through a drawdown that violates the asymmetric math of how wealth comes down.
00:34:08 Dr Genevieve Hayes
So basically, you invest sensibly so that you're gonna get a reasonable return that's gonna beat inflation. And OK, no.
00:34:16 Todd Tresidder
No, you have to manage risk. You have to have a viable offence. It beats inflation. So you're close there.
00:34:22 Todd Tresidder
I just got hung up on the sensible cause. Most people would think of sensible as like a cliche diversified portfolio.
00:34:28
Blah blah, blah blah blah.
00:34:29 Todd Tresidder
Right. It's going to vary for every economic regime. This is another financial truth. There's decades where you can get decades long where bonds outperform stocks.
00:34:39 Todd Tresidder
You can get decades long where stocks outperform bonds. You can get decades long where gold outperforms both. You've got really a handful of major asset classes that you're working with you're dealing with.
00:34:50 Todd Tresidder
Bonds, gold, commodities, equities, both international and domestic, both for bonds and equities and real estate and cash. Cash is the missing one. A lot of people don't understand. So like cash can be in a bull market.
00:35:06 Todd Tresidder
Because cash is a relative purchasing power equation, people say, oh, cash is trash, right? But if everything's correlating to the downside, then implicitly cash is in a bull market because it's purchasing power is increasing relative to all other assets.
00:35:22 Todd Tresidder
So you have to understand like there's a financial principle that I'm sharing with you again. You asked for all the financial principle I said I couldn't do them off the top of my head.
00:35:28 Todd Tresidder
But they'd come out as we talk. And so here's another one.
00:35:31 Todd Tresidder
One which is most assets will correlate to the downside during a big bad bear market. Usually one will not and it depends on the type of bear market and a deflationary bear market bonds will be the rising asset because the government will cut interest rates which will allow bonds to get into a bull market because they're trying to.
00:35:51 Todd Tresidder
Trying to fight the decline in an inflationary bull market, it's usually either commodities or gold.
00:35:57 Todd Tresidder
That is the rising asset bonds will get started. That's what we saw in the 2022 decline. Bonds got slaughtered along with stocks and everybody was ohh so surprised.
00:36:06 Todd Tresidder
Well, it's happened in history before. There's no surprise. They just haven't seen an inflationary bull market for a long, long time.
00:36:13 Todd Tresidder
It's been decades. It's a fundamental regime change, like the way I teach it at my website for my community.
00:36:18 Todd Tresidder
I call it epical. Change the investment. Epic changed. We're in an inflationary environment and so the rules of the game change. The financial truths are different.
00:36:27 Todd Tresidder
So everybody's betting on a bond stock diversification. Suddenly they correlate to the downside and they're shocked they.
00:36:32 Todd Tresidder
Shouldn't be. That's what happens in inflationary bear market. And there's times when neither goes down or neither holds up.
00:36:39 Todd Tresidder
And that's when cash is king. So cash is increasing in relative purchasing power of the other assets even though cash itself is only treading water. And so once you understand the major asset classes and how they move.
00:36:53 Todd Tresidder
Then you can structure a portfolio for regime change and it redone statistically with the.
00:36:57 Todd Tresidder
Quality. It doesn't have to have forecasting ability or market insight or anything. It can all be done with statistical validity.
00:37:06 Dr Genevieve Hayes
So basically what a lot of the financial literature says is buy a diversified portfolio, young leave your money in it and then wait 30-40 years and everything will be fine. But it sounds like you're saying.
00:37:21 Todd Tresidder
It sounds a lot like my cynical statement about give the money to your financial planner, say savings grant. Give them money.
00:37:28 Todd Tresidder
Your financial planner and his magical asset education will.
00:37:30 Todd Tresidder
Give you a.
00:37:30 Todd Tresidder
Pot of gold at the end of the rainbow.
00:37:32 Todd Tresidder
Right. You just said the exact same thing.
00:37:34 Dr Genevieve Hayes
And you're saying that that doesn't work and that you basically have to adjust for the circumstances, OK.
00:37:39 Todd Tresidder
That's not true. I'm not saying it doesn't work. It's valid in the sense that it has a positive mathematical expectancy, so it's valid.
00:37:49 Todd Tresidder
However, it takes an entire lifetime of saving and compounding over full market cycles for you to realise the benefit. It's not an efficient path to the goal.
00:38:01 Todd Tresidder
Where they're correct is that they're citing A valid strategy that anyone can implement where they're wrong is they're saying it's the one and only solution that's valid.
00:38:11 Todd Tresidder
It's not the one and only solution. There are other solutions that are more efficient and.
00:38:16 Todd Tresidder
More effective but.
00:38:18 Todd Tresidder
Are they wrong? No, they're not wrong.
00:38:20 Todd Tresidder
If you do.
00:38:20 Todd Tresidder
What they say over an entire lifetime.
00:38:22 Todd Tresidder
And you're willing to wait till you're old age to have your pot of gold, then fine. But if you would like to retire earlier than old and you'd like to have your lifetime to enjoy your money.
00:38:32 Todd Tresidder
And you'd like to have greater consistency in your financial performance and you'd like to have a higher safe withdrawal rate, roughly double what standard financial advice offers.
00:38:42 Todd Tresidder
And you know all these things, I mean, they make so much more efficient path to your goal and so much more reliable.
00:38:49 Todd Tresidder
So they're not wrong. I've never said buying holds wrong, and I've never said traditional financial advice is wrong.
00:38:55 Todd Tresidder
You can still get the outcome as they promise you, it's just not smart.
00:39:00 Dr Genevieve Hayes
But it's easy to market to large quantities of people and to write books about.
00:39:07 Todd Tresidder
So what I've explained so like you know my community and niche community, you know, it's never gonna have broad appeal because it requires learning this stuff and understanding it.
00:39:17 Todd Tresidder
But it way represents reality, and I mean people are becoming financially independent. My community regularly and they're getting the results despite the economic difficulties because they understand what they're doing.
00:39:27 Todd Tresidder
So it takes effort to learn.
00:39:29 Todd Tresidder
It, but if you're interested, it's well worth the effort. You get paid for that learning. It's knowledge that will actually pay you.
00:39:36 Todd Tresidder
It's not for everyone what I teach, I teach an idea called level one versus level 2 knowledge, and you probably run across this in in your profession. So level one knowledge is characterised as a simple narrative that everyone can understand.
00:39:49 Todd Tresidder
That is closest to the truth that it's supported by a decent amount of data and gets broad appeal because it's simple and so there you go. There's buy and hold, there's conventional financial advice in a nutshell.
00:40:01 Todd Tresidder
OK, that's level one knowledge. Then there's level 2 knowledge which takes into account all the outlier data, all the data that doesn't match the conventional wisdom and is largely ignored or discounted by the conventional wisdom. Things like ignore volatility, ignore the drawdowns you have to buy and hold for the long term.
00:40:22 Todd Tresidder
The reason they say long term is it's baked into the cake of the design of the model and how financial markets work. It requires 12 to.
00:40:29 Todd Tresidder
20 years based on whatever your portfolio construction is, it varies with the portfolio construction to realise a positive mathematical expectancy from a conventional model and that's because of the source of returns of the model.
00:40:41 Todd Tresidder
Again, it's just a financial truth takes that many years in order for the change in market valuation to mean revert, while either the interest on bonds.
00:40:50 Todd Tresidder
And or the dividends and economic growth of stocks, compounds in the background at little tiny incremental numbers.
00:40:57 Todd Tresidder
Consistently creating that upward tilt of the equation to eventually overcome the change in market valuation. That's why it takes 12 to 20 years to realise the positive expectancy.
00:41:07 Todd Tresidder
That's why they say you have to. You have to have a long term outcome, whereas you can take other models that have a positive expectancy in two to three years reliably through all economic regimes. So there's other ways of doing it that are more efficient and more effective.
00:41:20 Todd Tresidder
It just depends on what you're willing to commit. If you want a super simple passive approach that anyone can understand, then you've got a level one understanding. I mean, a great example is physics, right? We had Newtonian physics.
00:41:34 Todd Tresidder
But then there was all this outlier data at the extreme ends right, the extreme small, the extreme large, and then out of that arose relativity theory and quantum theory.
00:41:45 Todd Tresidder
We had to develop more theories in order to account for it and could take it to a level 2 understanding.
00:41:50 Todd Tresidder
So go ahead and try to fly a rocket ship to Mars on Newtonian physics.
00:41:55 Todd Tresidder
You won't get there.
00:41:56 Todd Tresidder
It requires a more advanced understanding. That's because the advanced understanding matches reality. It's more accurate. That's level 2 understand.
00:42:03 Dr Genevieve Hayes
What I was thinking when you're saying this is with financial knowledge, if you go to the airport and walk around the book shop, there are so many personal finance books in the airport bookshop and the sort of finance knowledge you're going to get in a book that is sold at the airport book shop by definition has to be level one knowledge.
00:42:23 Todd Tresidder
Exactly. You you have to get level 1 because it's the only thing popular enough to sell at the airport bookstore.
00:42:29 Todd Tresidder
They will never sell my books there unless something very strange happens.
00:42:32 Dr Genevieve Hayes
And that's why, you know, when I came across your books, it was like, hang on. This is some guy is saying something that's not in these airport bookshop books and.
00:42:41 Dr Genevieve Hayes
I was really excited and now I get it. There's basically two different markets. There's the popular market and then the more niche market for people who want to go deeper.
00:42:52 Todd Tresidder
And I'm not saying the popular markets wrong. What it is is a narrow case. Reality, truth that works in most circumstances.
00:43:02 Todd Tresidder
It's the extreme small and extreme large where it blows up and you're a sample size of 1. Your life is a sample size.
00:43:08 Todd Tresidder
One and the odds of you hitting a blow up period are pretty high. You just have to decide what you want and most of them will just kind of overlook it and say ohh yeah, they told me I have to invest for the long term.
00:43:21 Todd Tresidder
The fact that I'm down thirty 4050% My Portfolio, they told me that was the risk, but I'm in it for the long term.
00:43:28 Todd Tresidder
And they don't really think about it. They don't think. Wow. Asymmetric compounding of wealth. I just incurred a 50% drawdown and that took three years to occur.
00:43:39 Todd Tresidder
And now I'm gonna have to double my money in order just to.
00:43:43 Todd Tresidder
Get back to.
00:43:43 Todd Tresidder
Even so, let's take the 2000 example, the 2000 top. That was 12 years for the passive index to get back to.
00:43:51 Todd Tresidder
Even now, let's say that you're retired and you've got a conventional financial planning assumption of 4% rule. So you're pulling 4% of your portfolio.
00:44:00 Todd Tresidder
Per year, that means 12 years later, not accounting for inflation. By the way, we still haven't thrown that in, right?
00:44:06 Todd Tresidder
We're just trying to break even from the drawdown 12 years later. The S&P 500 magically goes back to break even because of the inexorable compounding of the source of return behind it. Right as we talked about.
00:44:17 Todd Tresidder
Earlier that goes back to even, but through volatility effects and through the drawdown as a retiree, your portfolio is down over 50%.
00:44:27 Todd Tresidder
So the market went back to even exactly as the pundits would tell you. The market climbed to new highs. Congratulations. Meanwhile, you're sitting on a poor foot. It's down more than 50% because you're a retiree.
00:44:38 Todd Tresidder
The 4% safe withdrawal rate. And So what are you gonna do? You're gonna change your life. And I didn't even account for inflation in there.
00:44:45 Todd Tresidder
Let's say you're throwing inflation rate at another 3%. You're down on your purchasing power. What are we doing 7 * 12 years?
00:44:52 Todd Tresidder
So 84%.
00:44:53 Todd Tresidder
.84% loss of purchasing power during that drawdown go through that once or twice and you tell me you feel safe.
00:45:02 Todd Tresidder
And I'm just playing loose with the numbers, right? I mean, it's actually worse in reality because you've got expenses, you've got compound effects, you got the volatility effects, there's all kinds of stuff going on in there.
00:45:13 Todd Tresidder
So are they wrong? No, they're not wrong. The market did come back 12 years later. The market went to new highs, exactly as they promised. And there's reasons why that has occurred in the past and will continue to occur in the future.
00:45:25 Todd Tresidder
Exactly as promised. There's a reason why it works that.
00:45:27 Todd Tresidder
Way. OK, that doesn't mean it's good. That doesn't mean it's smart.
00:45:32 Dr Genevieve Hayes
And if you're 75, when it occurs, you're probably stuffed.
00:45:36 Todd Tresidder
Yeah, exactly. See, here's the thing. You're taught that you have that. Here's another truth. You're asking for truths, right?
00:45:44 Todd Tresidder
So you're taught that you have a long term horizon of 20 to 30 years. It has to do with the math of the source of returns, right? Which is dividends plus economic growth plus or minus change in market valuation.
00:45:56 Todd Tresidder
That's the source of return for stock market. So what happens is the dividends plus the economic growth has this inexorable small compounding and the size of it depends on when your starting point is what the valuations of the markets were at the starting point. Again, notice I'm bringing the concept of valuations.
00:46:11 Todd Tresidder
That's the inexorable compounding then you've got this huge tail that wags the dog, which is plus or minus change in market valuations.
00:46:18 Todd Tresidder
That's why the markets are so volatile. They'll swing up 30% one year, down 50% the next year, up 25% the next year.
00:46:25 Todd Tresidder
They're going all over the place, but then 1015 years later, you've got this compound effect of seven percent, 8%, whatever the number is.
00:46:32 Todd Tresidder
But you had this wild ride to.
00:46:34 Todd Tresidder
Up there. And so you're told you have to have this long term time horizon for your expectations for investing.
00:46:41 Todd Tresidder
But the reality is different. OK, the reality is you have a series of 10 and 15 year periods throughout your lifetime, emphasised at the 10 to 15 years before you retire and the 10 to 15 years after you retire.
00:46:55 Todd Tresidder
The 10 to 15 years after you retire is called sequence of returns risk, and there's actually a name for.
00:47:00 Todd Tresidder
It and it shows that if you have adverse returns during the 10 to 15 years after you retire, that it does such a mess to the math of your distribution that you actually have a lower safe withdrawal rates.
00:47:13 Todd Tresidder
Safe withdrawal rates will vary anywhere from 3% to 12%, depending on your sequence of returns. Risk immediately after you retire.
00:47:20 Todd Tresidder
So everybody is actually a market timer by birthright.
00:47:23 Todd Tresidder
They just don't know it.
00:47:25 Todd Tresidder
Well, as it turns out, the same thing is true for the 10 to 15 years before you retire. It's just the inverse of the picture, because if you have a adverse sequence of returns for the 10 to 15 years before you retire, you can retire with 1/2 or as quarter as much money as you assumed because in your retirement planning you assume this nice steady compound.
00:47:45 Todd Tresidder
Growth of seven or 8%. You're assuming that's what's going to occur in your modelling, but what will actually occur is a conventional passive portfolio will grow in fits and spurts. So it looks more like a stair.
00:47:58 Todd Tresidder
Days where you'll go through 1015 year periods where there's very low returns, then the 1015 year period where it has high returns and it averages out long term over this seven, 8% compounded those staircase effects are due to valuation effects where you've got a sudden ramp up in valuations then the markets work off the valuations, then you get a ramp up in valuations again.
00:48:20 Todd Tresidder
And those periods or epics as they change if one occurs immediately before or after you retire, it can mess up all your planning and so.
00:48:28 Todd Tresidder
So they tell you, you have to have a long term time horizon, but in fact what you're really dealing with is you have to have positive expectancy and reliable turns over all 1015 year time Windows, particularly those that immediately preceding and immediately falling retiring.
00:48:43 Dr Genevieve Hayes
And the biggest factor that influences that is the year in which you were born, which is something that you have no control over whatsoever.
00:48:51 Dr Genevieve Hayes
What this is making me think is if there are all these problems with financial planning and financial modelling, these same problems have got to occur in so many other disciplines where people are just going straight to the data building models, not taking into account the truths of whatever the domains their models are in.
00:49:11 Dr Genevieve Hayes
And it seems like if you're saying that for finance, people need to take a step back and work out what the truths are first.
00:49:20 Dr Genevieve Hayes
I would say that data scientists in any discipline need to take a step back and work out what the truths of whatever the discipline they're working in, are fit.
00:49:28 Dr Genevieve Hayes
Just and then build their models based on those truths.
00:49:33 Todd Tresidder
Yeah it can.
00:49:33 Todd Tresidder
Go both ways. You just have to be careful. Right? So like one of things AI's doing right now is it's just deep diving into data and trying to discern what possible truths may exist so we can go back to.
00:49:45 Todd Tresidder
We talked about biotech recently, AI's diving into these massive bio databases like UK has one because they have a public health system.
00:49:54 Todd Tresidder
So they have these large databases from the public health system and they're trying to train AI on them and figure out what truths can be figured out from.
00:50:03 Todd Tresidder
But but ultimately you have to be able to go backwards on it if you will. So yeah, you can use the AI to ferret out what may be something valid, but ultimately you have to run it by somebody truly knows the field to understand if it's actually telling you a nuance of something else before you actually put it into practise. So.
00:50:23 Todd Tresidder
Now ultimately I think you need somebody with deep expertise in the field to philtre what the math tells you and give you the nuanced truth of it.
00:50:33 Todd Tresidder
Kind of like I was trying to do here in this interview I was trying to show a lot of the nuances.
00:50:39 Todd Tresidder
Behind a lot of this.
00:50:40 Dr Genevieve Hayes
Stuff because what? The AI is at the end of the day is it's just pattern recognition and you can have a pattern happening over a short period of time. But as you said, those patterns aren't necessarily going to continue into the future.
00:50:53 Todd Tresidder
Yeah. Again, you're always dealing with data limitations. So like the AI, I'm most familiar with is the language models, because I'm a writer, and that's the ones most popular.
00:51:03 Todd Tresidder
Right now and one of the most developed at least that I'm aware of, you know, in my mind it's just, you know, it's computerised plagiarism and it's limited to the information it's trained on.
00:51:13 Todd Tresidder
So there's no genuine expertise on it. All it's doing is parenting back from the patterns in what the language shows about it. But it's limited to what's there. There's no generation of actual knowledge in it.
00:51:27 Todd Tresidder
You know, so you're never gonna be better off with it than you with with the genuine knowledge.
00:51:33 Dr Genevieve Hayes
That's what I tell my students. Do not use ChatGPT unless you understand what you're asking it, because otherwise it'll give you a wrong answer and you won't realise it. You are better off just doing the research yourself.
00:51:47 Todd Tresidder
Yeah, and it can be a great research tool. There's all kinds of ways to use it wisely. I'm not against it.
00:51:51 Todd Tresidder
I do think a really interesting area is gonna be copyright violation, at least with large language models. That's gotta be figured out.
00:51:58 Todd Tresidder
I mean, you're seeing it with, you know, the Hollywood strikes and everything else going on is, is there's a real problem with intellectual property with this now because it is not citing.
00:52:07 Todd Tresidder
Source and it is.
00:52:08 Todd Tresidder
Not paying to source based on the value derived from that source, and so it's just basically a big rip.
00:52:15 Dr Genevieve Hayes
Off and people are getting very angry and suing, and rightly so.
00:52:20 Todd Tresidder
Yeah. I mean, as an author of books, I don't want it going in and pattern recognising my material and parenting it back out.
00:52:27 Todd Tresidder
You can pay for that access. That's only fair. That's what intellectual property laws for. That's why it was created.
00:52:33 Todd Tresidder
Otherwise, why would I create it? It's an extraordinary amount of work to create this stuff.
00:52:38 Dr Genevieve Hayes
And yeah, I fully support that.
00:52:40 Dr Genevieve Hayes
Is there anything on your radar in the AI data and analytics space that you think is going to become important in the next three to five?
00:52:48 Todd Tresidder
Well, we just hit it. Copyright law with applied to large language models. I think it's going to be a game changer, right?
00:52:55 Todd Tresidder
Because how can you apply AI and defend copyright? I don't know how you do that and what are we going to do?
00:53:02 Todd Tresidder
Just throw right copyright protection. How? How do you do that and not destroy the creation of viable material that when cause you're destroying the incentive?
00:53:11 Todd Tresidder
To create it, you know it's like, are we gonna throw away patent rights on drugs? I mean, there's so much invested upfront, you have to protect that intellectual property so that there's a payout down the road to justify the effort to create it.
00:53:24 Todd Tresidder
And so you can't just have a I trained on the latest, greatest information out there and violate all the copyrights.
00:53:31 Todd Tresidder
There's no precedent for that, and there's a reason why those copyrights exist, so there's got to be something done for that.
00:53:37 Todd Tresidder
Otherwise there'll be a fundamental change in society around publishing of information to keep it away from large language models. I mean, I would literally have to put everything behind paywalls.
00:53:49 Todd Tresidder
And keep it out. You can't allow public domain access to information because there's no value proposition. Once a I can take.
00:53:57 Todd Tresidder
It I mean it literally changes the creators incentive package.
00:54:01 Dr Genevieve Hayes
Or everything that is out there just becomes a mash up of everything that existed prior to these models coming out.
00:54:07 Todd Tresidder
Well, that's another thing that's really fascinating is we've gotten really close to it already, right. So in my space in the financial space, you've got a really interesting, phenomenal Google results, right? So Google is a very early user of AI and its rankings.
00:54:22 Todd Tresidder
But you've also got a thing of they call it eat criteria, right? Which is expertise authority, and I can't remember the thing for tea, but it's it's an acronym EAT. It's used specifically for the medical and financial space because it's considered a risk to humans to have bad information. And so they have human screeners that screen.
00:54:41 Todd Tresidder
How sites rank based on the expertise and authority of the authors.
00:54:46 Todd Tresidder
So they've actually had to bring humans into there for the top ranks because you know, otherwise there's no way to know what's genuinely authorative versus just popular.
00:54:53 Todd Tresidder
And so now all of a sudden the cost to produce authoritative sites goes to 0 because anybody can train an AI bot based on models around search rankings.
00:55:06 Todd Tresidder
To mimic, but then you already got it right now. If you look at the search rankings for financial content for Google, they're terrible.
00:55:12 Todd Tresidder
And I'm not trying to insult Google. It used to be really good, but in the past several years you look at the top 10 rankings and it's very generic and they're very consistent because what happens is everybody models whatever the top is using what's known about the algorithm and all the content starts looking similar because they're all trying to compete for the same top three rankings, which is where all the traffic is.
00:55:36 Todd Tresidder
The only time you find anything interesting is when you go past two through 10. You have to start diving deep to find anything unique or different, because all the top rankings are games.
00:55:46 Todd Tresidder
And so now it's going to get multiplied worse, because now you're going to have the AI bots producing the content way more intelligently than freelance riders ever did.
00:55:56 Todd Tresidder
And what happened? Now you've got all the large corporate sites are all dominating the rankings because they had the business model and the money to fund large freelance rider Staffs.
00:56:06 Todd Tresidder
To produce professional quality content that's completely generic, with no actual expertise. Because they're journalism majors, they're not finance experts.
00:56:14 Todd Tresidder
And so they would produce just reams and mountains of this content, and they build these huge authoritative sites for these large corporations, which then dominate the rankings.
00:56:24 Todd Tresidder
And you end up with this. All this generic crap content that actually has no genuine expertise in it. It's all written for the rankings, and Google is playing right into it. And now you're going to have AI bots.
00:56:35 Todd Tresidder
That are messing the whole thing up even worse, and I think ultimately my personal opinion, Google tried to do it before and they backed off from it is they did actual author algorithms and I think they're ultimately gonna have to.
00:56:50 Todd Tresidder
They're going to have to go back and identify authors, identify expertise, and then rank based on author expertise. Ultimately, they're gonna have to.
00:56:58 Todd Tresidder
Otherwise, the AI bots just game the whole thing and anybody can throw up an expertise, a site, a massive complex side of great expertise with nothing more than an editor and an AI bot.
00:57:09 Todd Tresidder
The cost and value of content is going to go down to 0.
00:57:12 Dr Genevieve Hayes
You see it already being advertised these services that will actually basically produce massive quantities of content for your site that from what I can gather, has absolutely no value whatsoever.
00:57:25 Todd Tresidder
But then you look.
00:57:25 Todd Tresidder
At somebody like me that has the genuine expertise. So I go to create the content.
00:57:29 Todd Tresidder
It takes me a long time to produce a quality article. Am I going to publish it in its entirety on the web for free so the AI bots can then harvest?
00:57:37 Todd Tresidder
Of course not. That makes no sense as long as they're allowed to plagiarise, I'm gonna have to put a teaser piece of the article up to establish the interest, to establish the expertise and put the rest of the article behind some sort of membership closure in order to keep it from the AI bots getting to it and plagiarising it.
00:57:56 Todd Tresidder
It's going to completely change the structure of the Internet, otherwise why would I publish it? The I bots are just going to steal it and plagiarise it and republish it as.
00:58:04 Dr Genevieve Hayes
Their own? Yeah. And so if we run, ends up suffering as a result.
00:58:09 Dr Genevieve Hayes
So what final advice would you give to data scientists looking to create business value from data?
00:58:15 Todd Tresidder
I think we've.
00:58:16 Todd Tresidder
Covered a lot of it. I think you just.
00:58:17 Todd Tresidder
Have to really understand what's knowable and what's forever unknowable.
00:58:24 Todd Tresidder
To get really business value you need domain expertise as we referred to and you need to develop that. So I would probably do if your background is data science, I would probably.
00:58:35 Todd Tresidder
Try to bring data science to a company that isn't using that resource very well, and then you try to get to where you're working at the top level with the top level executives bringing in a data science.
00:58:46 Todd Tresidder
Term and see if you can really get a strong mentoring in the business principle so you could take it and apply it out as a business of your own, you know so you can get a quick run at the financial truce if you will.
00:58:59 Todd Tresidder
Using my domain expertise as an example, but as you said it could apply to plumbing, it could apply to water.
00:59:06 Todd Tresidder
Science it could.
00:59:08 Todd Tresidder
Electrical. I mean it can apply to any field. Yeah, I think you just really gotta understand that data science is incredibly valuable skill and there's limits.
00:59:16 Todd Tresidder
You're not dealing with absolute engineering truth. You have to understand the nuance of the application. I guess is the piece I'd leave you with. I mean, that's the main thing I've seen in my students that are.
00:59:28 Todd Tresidder
******** scientists and ******** engineers that come to me.
00:59:31 Todd Tresidder
I they follow a.
00:59:33 Todd Tresidder
Really clear pattern. Every time I have to.
00:59:36 Todd Tresidder
Weed them of what I call the engineers fallacy. They'll come in and they'll go on allocate smartly and there's this place where you can build optimised portfolios using pure math and quant, and it's built right in the platform and it's so easy to do and they think they're so smart, right?
00:59:49 Todd Tresidder
And so they build these optimised portfolios and then they get to the next lesson explains why none of that works, why you cannot do that.
00:59:56 Todd Tresidder
And it will not work going forward and now I have enough time period since I wrote those lessons.
01:00:03 Todd Tresidder
That all of it's proven true.
01:00:05 Todd Tresidder
You can look at the actual data since then and had you optimise your portfolios based on the data back in 2021 or 2020 when those lessons were written, what would have worked then and worked best on history? Did not work at all going forward, whereas the stuff that I green lined and there has worked very well going forward, is it optimal? Is it perfect?
01:00:26 Todd Tresidder
No, but I never promised anybody that what I said in the less.
01:00:29 Todd Tresidder
Reasons is you're not shooting for the optimal portfolio of the past, which is what data science, Arcore data science will give you.
01:00:37 Todd Tresidder
That will give you the optimal portfolio, the past what you're actually searching for is the profile that has the highest probability of performing well in the future. Those are very different questions that you ask.
01:00:50 Todd Tresidder
And it's going to be the same in any profession. I'm just applying it to portfolios. That's my expertise.
01:00:55 Dr Genevieve Hayes
On that note, for listeners who want to learn more about you or get in contact, what can they do?
01:01:02 Todd Tresidder
Financialmentor.com so I have social media channels, but I don't do any with them. My assistant runs them for me.
01:01:08 Todd Tresidder
You know, I'm just not a social media guy, so everything's email and everythings on the website and I have several books which you had mentioned a couple of the two most popular leverage equation and how much money do I need to retire? And then I have a course.
01:01:22 Todd Tresidder
My master course is expectancy, wealth planning and then in there is also where I teach all the use of tactical asset allocation as part of your paper assets, but it's much bigger than that because it's also.
01:01:35 Todd Tresidder
About principles in real estate and principles in business, entrepreneurship, which those are the three asset classes you use to build wealth and then it goes through and it integrates all the personal with it about how you design your own personalised wealth plan.
01:01:47 Todd Tresidder
And so that's my master course. And then there's the books, which are the low price entry point. And then there's free stuff too, which is the newsletter, the public facing newsletter and.
01:01:56 Todd Tresidder
All the free content on the site and I have a huge suite of financial calculators which are free too, because again, all wealth is math.
01:02:03 Todd Tresidder
Right.
01:02:04 Todd Tresidder
So we'll come full circle, right. You have to know financial truths, but you also have to know the math truths. So all wealth is math, right? It's the expectancy equation. And it's the future value.
01:02:13 Todd Tresidder
Equation and so how do you work with those? How do you combine them with your personal life situation and your resources, your skills, your timeline, your goals?
01:02:25 Todd Tresidder
How do you integrate that all into a wealth plan that's actually gonna work for you in this lifetime? And that's what the expectancy, wealth, planning, course figures.
01:02:31 Todd Tresidder
Out for you.
01:02:32 Dr Genevieve Hayes
Well, thank you for joining me today.
01:02:35 Todd Tresidder
Yeah, hopefully that wasn't too much of a fire hose I.
01:02:37 Dr Genevieve Hayes
Hope that was enjoyable. I loved it. So yeah, I hope all of our listeners.
01:02:40
OK, good.
01:02:41 Dr Genevieve Hayes
Loved it too.
01:02:45 Dr Genevieve Hayes
And for those in the audience, thank you for listening. I'm doctor Genevieve Hayes, and this has been value driven data science brought to you by Genevieve Hayes Consulting.