June 27, 2024 • 43 mins
This week on Risky Business, Nate and Maria unveil Nate’s 2024 election model. (The forecast is bad news for Maria.) Also, they explain what insurance industry troubles tell us about the market for risk. And they discuss some key mistakes people make when thinking about risk.
Further Reading:
“Silver Bulletin 2024 presidential election forecast” by Nate Silver and Eli McKown-Dawson
“Climate Change Has Hit Home Insurance. Is Health Insurance Next?” from the Wall Street Journal
“The Home Insurance Crunch: See What’s Happening in Your State” from the New York Times
For more from Nate and Maria, subscribe to their newsletters:
“The Leap” from Maria Konnikova
“Silver Bulletin” from Nate Silver
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:15):
Pushkin. Welcome back to Risky Business, a show about making
better decisions. I'm Maria Kanakova.
Speaker 2 (00:28):
And I'm Nate Silvery. Today on the show, we'll talk
about my new presidential election forecast.
Speaker 1 (00:34):
Well also talk about insurance. Although Nate, I'm really excited
to hear about your election forecasts.
Speaker 2 (00:40):
You might want to ensure against that feeling, Marie, you
may not like what it has to say. I'm talking
about our top three ways that people get risk wrong.
Speaker 1 (00:54):
Nate, let's talk about your election model for the twenty
twenty four elections, which is out this week. You've been
working on this election model, tweaking it, developing it for
sixteen years, and it's been remarkably accurate. So I'm actually
really excited to talk about it and get into it
with you, even you told me that I really shouldn't
be so, I mean, I'm a little scared, too.
Speaker 2 (01:12):
Careful, careful what you wish for. I want you to
guess what number it has in the chance that trumpel
with the election.
Speaker 1 (01:20):
Oh night, this is are you gonna break my heart?
Is it okay? Fifty eight percent?
Speaker 2 (01:27):
Higher?
Speaker 1 (01:28):
Fuck?
Speaker 2 (01:29):
I don't feel the exact number because we're we chake
a couple days in advance. By the time the model
is released, there could be new polls and so forth.
I mean, we'll see what it says when it comes out.
But we're in the We're in the load of mid
sixties is kind of where the where the forecast is landing.
Speaker 1 (01:45):
Fucking you fuck fuck fuck is my is my response
to that night. You're usually the one dropping the f
phones on the show, but I think it's my turn. Yeah.
Speaker 2 (01:53):
Look, my job here is not to tell you who
you should vote for. It's to forecast the election as
best I can. And I think it's disingenuous too to
call the race a toss up. You can kind of
squint and say, oh, if you look at national polls
and Biden's maybe only down by half a pointer or so.
(02:14):
But the problem in the US is that we don't
determine elections by the popular vote. Right, if it was
a popular vote, the model things it would be a
toss up, or maybe Biden slightly favored. In the electoral college.
There has been for the past two cycles with Trump
on the ballot, a gap that favors a GOP. In
twenty and twenty, Biden won the popular vote by four
(02:35):
and a half points, which is a pretty good margin
against an incumbent. But all these key states Georgia and
Wisconsin and Pennsylvania and so forth, the only win by
about one point. So if we're now in twenty twenty
four and Biden's trailing by half a point instead of
winning by four and a half points in the popular vote,
that's a shift of five points. If you have a
(02:57):
state like Pennsylvania that you win by one point and
you shift it by five points, that's not very good. Now.
What's happening in practice is that there's a bit of
a bifurcation between the Midwestern swing states is kind of
trio of rust Belt states Pennsylvania, maybe even maybe not
at the Midwest per se, but Pennsylvania, Wisconsin, Michigan, where
BIA's polling is closer usually within the point maybe two
(03:20):
point Trump lead, versus the sun Belt states, the new
fangled Georgia, Nevada, Arizona, where Biden's polling is very bad.
He might be down five or six points. Now. If
he wins Michigan and Pennsylvania and Wisconsin and holds all
the other states so like New Hampshire, Virginia and so forth,
(03:42):
then he would finish with exactly two hundred and seventy
electoral votes when you need to sixty nine to win.
So there is a path, but not a whole lot
of margin. For aer.
Speaker 1 (03:55):
Nate, this is not the way that I wanted to
start off my week. So I think I think we're
done with the podcast, goodbye, good bye luck as I
go and hide in a cave. But no, in all honesty,
this is sobering news. I so don't want Trump to
win because I know I think it's going to have
incredibly dire consequences for democracy. So I want to kind
(04:20):
of squint and say, yeah, but there's this path and
look at these states, and we're going to do it.
But I think that you need to be realistic and
try to figure out, Okay, given this, there's still it's
not like the election is tomorrow, right, So what can
we do? What shifts in your election model would actually
change that outcome and bring that percentage back down? Are
there certain lovers? Are there certain things that are actually
(04:43):
going to be kind of the most meaningful as you
play around with the inputs and the numbers.
Speaker 2 (04:49):
Yeah, look, Biden is struggling with groups that are traditionally
strong Democratic groups like younger black voters, younger Hispanic voters.
That's why you see his numbers fading in states like Georgia,
for instance, whereas he's actually doing okay with these older
white voters. That seems to be the constituency. It's like
holding up for him the best. There is a debate today,
(05:10):
This episode's coming out on Thursday tonight. That's an opportunity
for Biden to turn things around. You know, the conventions
have get to happen. The concern for Biden is that
the race has been extremely steady so far, the steadiest
you've ever seen Poland. I think in a presidential election
Biden game maybe half a point to a point in
(05:32):
the polls following Trump's conviction on thirty However, many it
was felony counts, you know. But the fact that like
Trump was convicted of paying hush money payments to like
at porn star adult film actress, excuse me. And that's
like a real sentence I can say, And that like
(05:53):
had half a point worth of difference. I mean, that's
that's not great for Biden. I think there have been
a lot of false hopes from Democrats. People say, oh,
this is going to move the poles. It's going to
move the polls when the economy has gotten a fair
bit better over the past six months. Inflation has abate,
the stock market is way up actually, but like the
(06:13):
numbers haven't really changed very much. And I don't know.
I'm sure I'll be called a fascist, biased, republican hack
by every democratic strategist in the book. But the whole
point of having a model is that you're applying structured
thinking where where you don't go in and tweak the
results once you see what it spits out. Right, The
(06:35):
whole point of committing to a model is that like
it does take the emotional component, which I have like
everybody else, out of the process or reduces it. Let's say, right,
I'm someone who thinks professionally, it would be great to
just say, oh, it's a toss of fifty to fifty,
who knows can't get in triple that way. Well, that's
dishonest and it's not what the number says, and it's
(06:55):
not what the process I followed for sixteen year now
years now is two thousand and eight says right, this
is not a year that looks terribly uncertain. You don't
have COVID. You don't have these polls volleying up and
down like you did in twenty sixteen with Clinton and Trump.
They're very well known candidates, and so Trump's advantage is narrow,
but it's been persistent, and he has the more robust paths.
(07:20):
The fact that Biden has to win all three of
those Michigan, Wisconsin, Pennsylvania trio, all where he probably trails
by a point or so, but he has to win
all three and hold these other states. The fact that
like the next best wing states like Arizona for Biden,
where he's down by like four or five points, that's
a real deficit. I don't think you can sit here
(07:41):
with a straight face and say you'd rather be by
I think you can say you'd rather be Trump.
Speaker 1 (07:47):
You started off by saying that it seems that Biden
is slipping amongst some of those voters who are kind
of traditionally democratic, right, like the younger black voters. Are
those voters actually switching to Trump or what's happening to them?
So where's that loss for Biden evaporating? Two?
Speaker 2 (08:06):
I think there might be, you know, some permissions ructures
to vote for Trump that didn't exist last time, but no,
a lot of them are going to undecided, to RFK
junior or out of the electorate. You know, our model
projects turn out of something like one hundred and fifty
million this year, which is which would be down from
(08:28):
twenty to twenty in all the polls, the indications that
there's less enthusiasm for this election than there had been.
Now to Biden this credit, he actually does well. Democrats
in general do well with these very reliable regular voters.
So like, lower turnout might actually be beneficial for Biden.
Speaker 1 (08:47):
So Nate, Okay, since we since we are we are here,
what are kind of if you are at the Biden camp,
where should they focus their efforts between now and election time?
Like give us like top top two, top three priorities,
Like what should the campaign be doing to try to
get those numbers back to where in more winnable territory?
Speaker 2 (09:13):
Maybe fire some people I don't know, certainly, fire anybody
in the campaig who's telling you not to believe the polls, right,
fire them and don't get them a new job. I mean, seriously,
if you're down, if you know, again, if you're at
a mid thirties, maybe maybe they'll got up to the
low forties by the time we published the model, but
probably mid thirties. If you're in that territory, then you
know that's the point when again, when a baseball team
(09:35):
is winning thirty some five six seven percent of its
ballgames and it tends to fire the manager and there's
been very little turnover, and like, their theory of the
case is wrong. They thought that Trump would be relatively
easy to defeat and they could run back the twenty
twenty playbook. You're talking about democracy, democracy, democracy, over and
over and over again, and that this would work. Their
theory of the case is wrong. They did not expect
(09:57):
to be down at this point, right. This is why
they're still doing the RFK junior stuff, because they thought
they'd be winning, and so you want to reduce variants.
When you're ahead and getting RFK junior, who's a wild card?
I guess you can say, off the b it is
something to do that. You know. Look, I think some
of this stuff that you've done tactically suggests improvement, but
(10:18):
but I don't know. I mean, I mean, you know,
for all the fucking talk about how important the election is,
and I think some of that talk is we'll scrutinize
that more. At some point. I think some of it's
a little precious. But like if you really think that
like this is the you know, the end all be
all of elections, then I'm not sure really acting like
(10:38):
it very much. You know, they're acting like people who
have incentive. We're called principal agent problems. Right, you don't
want to be the polster that gives the president bad news.
You definitely don't want to be the polster or the
advisor who tells the president, hey, maybe you should stand
down to think about having someone else run, because then
you lose your job. But you know, and then there's
(10:59):
a whole Kamala Harris issue too, where I mean, I
think if she were more popular, then there's some chance
that Biden would have stood down and Democrats to be
in better shape. So I don't know. Look, all you
can say is like, if Biden loses, and there's still
a pretty some chance he wins, it's not like you
can't see this coming. He has been behind in the
(11:20):
polls consistently for you know, since Midsummer last year. Voters
consistently have said you are too old, you should not
run again. The first term was fine, you should not
run again. You know they have ignored that message at
every turn, and that's because the data is ambiguous. Right again,
you can squint in the right day and look at
the right poles and see not just a path to
(11:43):
it Biden victory, but like a case where his chances
look pretty decent. Right, you can't look at things like
the twenty twenty two midterms where Democrats did relatively well,
and you can hope that the polls are biased in
your favor because biased against you, rather to outperform them,
because sometimes they are. But that's usually very hard to predict,
(12:04):
and you don't need many reminders of cases where the
polls were biased against Trump in stead.
Speaker 1 (12:10):
All right, so shake things up. I think we see
the old status quo bias that we've talked about week
after week after week, and firesome people focus on those
key states and like, just get your ass and gear
and realize that there's a very very good chance you'll lose.
Speaker 2 (12:29):
We'll be right back. Nay.
Speaker 1 (12:43):
Let's talk about some huge shifts in the home insurance industry.
There has been a lot of news recently about homeowners
insurance because insurers are actually getting out of multiple states
multiple areas and refusing just flat out and refusing to
ensure homes, which is obviously affecting homeowners, home buyers, people
(13:04):
who need to get you know, their their home mortgage,
anything like that, because they like physically cannot get their
home insured. And Nate, last week, you and I talked
a lot about climate change and all of those effects,
and this is directly related to that. So I think
that this is one way that consumers are feeling what
it means to be living in a world where the
(13:26):
risks of one off events are no longer one off events,
where the storm of the century is no longer the
storm of the century, where you can actually see these
people who use models to try to model out the
risks losing money and saying, oh shit, we don't want
to ensure your home anymore. We no longer want to
take this risk.
Speaker 2 (13:47):
So let me ask, Yeah, what might seem like a
naive question, right, why can't you just price these risks
in that you're in California and you know the risk
of wildfires, you're in Florida, you know the risk of hurricanes.
You can build a model and then it costs more
than it might what's a safe state, you know, Wyoming there,
(14:08):
like Bolcanos there and that anyway, but why can't you
charge more instead of not offering the risk at all?
Speaker 1 (14:16):
Yeah, I mean, I think that that's a really really
good point, and I think that some insurance are trying
to do that and trying to charge more. But I
think that some are using old data instead of new
data for the models. And the problem is there are
no good data, right. How do you model something that's
changing so rapidly in an environment that's changing rapidly, where
there's a lot of volatility, and volatility in the actual
(14:39):
sense of the word, which means that you know, ups
and downs are very extreme in a short period of time.
How do I factor in the uncertainty of these risks
given that we don't have any historical comps, right, given
that we don't know how it's going to play out
in the next ten years, because the last ten years
are not representative of the years before that are not representative,
(15:00):
and we no longer have this steady model where where
we can go back and say, Okay, you know, risk
of this x percent, risk of this experson. I mean,
I know that one way of handling volatility is to
price it even higher, but at some point you know
you're pricing it so high because of the uncertainty that
it might not make sense anymore.
Speaker 2 (15:23):
Yeah, if you go to the New York Times, they
have a interactive graphic called the Home Insurance Crunch, see
What's happening in your state, where it just shows every
state whether home insurance has been profitable or losing for
the insurers. And what you quickly discover is that you
have kind of limited upside where nothing goes wrong you
(15:46):
do make a profit, but kind of uncapped downside where
one year with crazy hurricanes or flooding or wildfires or
things like that can wipe out ten marginally profitable years.
And in states like California and Colorado a lot of
fires Florida, of course, with hurricanes like that happens pretty often,
(16:10):
so it's hard to measure tail risks, especially if the
tails are expanding. We talked on the last week's episode
about how if there's even a small shift in the distribution,
the part of the curve under the tail can really
get shifted out where it becomes substantially more probable. Yeah,
(16:31):
and there's also you know, there's also something called adverse
selection where maybe the people who want to buy insurance
are people that are worse risks for you to take.
Maybe they know something about their home that it's vulnerable
to landslides, for instance, And therefore you kind of have
what economists we call like a market for lemons, where
you don't you have information asymmetries between buyer and seller
(16:52):
that reduce the overall size of the market. And in
those cases it might be irrational to might be rational
rather to not offer a product. There's not a kind
of market clearing price.
Speaker 1 (17:03):
Wait Nate, one second, let's clarify what exactly a market
for lemons means.
Speaker 2 (17:08):
Yeah, the term was coined, I believe by the economist
George Akerlove, who I think is now at Berkeley won
a Nobel Price. He talks about, I mean, Elemon is
a defective used car, and it's hard to know everything
wrong with the car if you take a test drive
for a few minutes or something like that. But basically,
(17:29):
it's like when buyers don't have any way to trust sellers.
There's no third party vouching for the reliability of these
used cars. What happens is just the volume of transactions
goes way down. Right, You have dead weight loss when
there's no credible way to achieve trust. So the theory
here would be, like, you know, if we think that
(17:51):
insurance is a win win transaction, right, someone's paying some
expected value on average, you lose money if you buy insurance,
but they protect their downside risk. Right, that's a win
win on a risk adjusted basis. But if there's not
reliable information you're not sure you can trust the insurer,
or if you're insured, you're not sure you can trust
the homeowner, then you have transactions not happening that would
(18:15):
create utility for both sides and for society.
Speaker 1 (18:19):
Yeah, I think that something that I know about investing
is if you're trying to kind of look at an
opportunity and you're comparing upside and downside risk. Right, when
you have unlimited downside risk and your upside risk is
your upside is capped, that's not an investment you want
to be making, right. You want it to be asymmetric
(18:41):
the other way. You want your downside risk to be
capped and your upside to actually be able to go
up further, especially if you're leveraged. Right, If you're thinking
about it that way, and if you're thinking about the
fact that this is incredibly asymmetric risk, and this is
something that they've you know, this is something they've been
doing for years where it's been asymmetric in their favor,
(19:02):
where it has been basically a capped downside because these
kind of once in a lifetime events didn't happen, or
they did happen once in a lifetime, and so all
of your other years where you were making money paid
for it. But now if you're making less money, because
there's always something happening, right if you look at the
news every year, there's something happening. There's you know, home
(19:25):
destruction because of all sorts of different things, tornadoes, hurricanes, wildfires,
you know, you name it, you know, we got it.
And then if you have that, so in your up years,
up years, you're not nearly as up and then you
have these catastrophic events that are happening over and over
and over. Then to me, like, as if you're a
rational investors, it's a bad investment. So you end up
(19:47):
pulling out.
Speaker 2 (19:48):
What's not neglect the role of dorm regulations. According to
the Washington Post, in California's case, insurance companies must use
historical data rather than forward looking models when they price
insurance plans that means is supporting them. Their policies may
not reflect the actual risk you're supposed to hedge against.
That's stupid. I mean, there's no serious put laws in California,
(20:12):
and like I happen to agree with that. Night if
you can't actually like adjust for changes in the climate,
which I think most people in California would care about,
then you're just making the market less efficient. You know,
how you optimize this function where on the one hand,
you probably don't want insurers to have literally unlimited liability
(20:34):
because they can't. No one has infinity infinity dollars, right,
they won't be able to pay it. On the other hand,
you know, you don't want insurance not to make people
relatively whole when there is a disaster. Otherwise what's the point.
But I would gather that one that like these state
regulations are not always that well designed.
Speaker 1 (20:51):
Yeah, I think that that's absolutely right because and states
are trying to step in, by the way, not just
in California, right, because there are these gaps where they say,
oh shit, you know, what do we do if people
in our state can't buy houses because they can't get insurance,
Like this is actually a big, big issue, Nate. I
know that at some point point in the in the past,
(21:11):
you were considering buying a place in Miami. Is you know,
is that something where like, let's talk about this on
a personal level, Like, would you do that right now,
given what's happening with insurance, given what we know about
the risks there, given that you probably would not be
able to get insurance on the you know, on an
apartment that you wanted.
Speaker 2 (21:32):
Honest answer, they probably wouldn't affect my decision that much
because look, home insurance is negative expective value on a
non risk adjusted basis, right, Otherwise it wouldn't be offered.
By definition, insurance means that someone who has more capital
and more risk tolerance than you makes money on average,
(21:54):
and homeowners lose money on average. I mean, it wouldn't
be great. I'll probably take the negative evy like safe bet.
I mean, I'm probably not going to buy a place
in Mimi for other reasons. But no, I mean, but
I'm also conscious about how much of my net worth
I'm tied it being real estate, right. I think it's
(22:16):
generally a mistake, for many reasons to have a great
deal of your net worth tied up in real estate
as opposed to a broader portfolio of stuff. So I
think I think I'm hedged enough in diverse fi it
enough that, like, although not ideal, I I you know,
it wouldn't be a deal breaker.
Speaker 1 (22:33):
So you would basically write it off, right, Like, as
this is an investment that might go to zero, I
don't care that. I'm not going to get insurance on it,
and I understand that it might get destroyed, and I'm
okay with that. I'm just going to make it a
small enough investment that if it goes to zero, I'll
be okay.
Speaker 2 (22:50):
That's right.
Speaker 1 (22:51):
Yeah, So this, I think. So this actually kind of
illustrates that it's affecting different people differently. Where if you
are in a position where you can do that and
you can say I will buy this place and I'm
just going to write it off, then you can buy it.
But if you're someone who can't do that, then the
calculus changes.
Speaker 2 (23:08):
Like with those things, people who are you know, better
off financially or who have more robust ways of looking
at risk benefit from it, and the people who are
disadvantaged are harmed by him more because yeah, I mean, look,
if you're like a multi billionaire in Silicon Valley or something, right,
(23:30):
and you lose your beautiful seven million dollar home in
a wildfire or an earthquake, you'll still be okay, Right,
if most of your net worth is tied up in
your home, then you won't be.
Speaker 1 (23:42):
So it seems like when we're talking about this, we
basically have the absolute like shit show of factors, if
I may be, if I may use a psychological term here,
So we have huge correlations, right, we have no good
historic data, we have a lot of things changing, and
(24:03):
so it's incredibly difficult to model. And so it's basically like,
you know, this is your worst nightmare when you're trying
to build a model that predicts risks and tries to
figure out what in the world do we do? And
so I think that at least in the short term,
what I'm taking away from this is that a lot
of people, especially kind of the people who cannot afford
(24:23):
to just write off these these homes, are kind of
fucked and cannot buy homes, have to move out of
homes that they've owned where the insurers have stepped out,
and there's really no clear solution. And as you know,
as someone who is making these decisions on a personal level.
Where do I move, you know, where where do I
buy a house, Where do I rent? Where do I
look at real estate? Where do I build a community.
(24:45):
I think these are all things that people need to
keep in mind as they're thinking about the future. And
sometimes that sucks because sometimes you know, you really want
to move to beautiful wildfire wildfire California or lovely coastal Florida,
and that's just kind of the emotional elements of that
(25:06):
are not going to be worth it for you in
the long term. We'll be right back after the break.
Speaker 2 (25:24):
All right, Maria, Let's talk about our top three ways
and people get risk wrong.
Speaker 1 (25:29):
So on the show we often talk about kind of
these meta risk concepts, you know, elections and those types
of big questions. But right now I'd love to talk
bring it back to kind of individual psychology and talk
about how people, how individuals get risk assessments wrong, the
biggest problems that they have when they're thinking about it
(25:52):
on an individual level. And I think you and I
both have some thoughts as to what the most important
ones are. I can lead us off if you'd like.
Speaker 2 (26:00):
I think that one of the biggest a draft of
like cognitive biases.
Speaker 1 (26:05):
Yeah, draft of cognitive biases, but specifically.
Speaker 2 (26:09):
The amplast fallacy. You're looking pretty good and picking them
at two.
Speaker 1 (26:13):
Exactly exactly.
Speaker 2 (26:15):
It's rookie out of Providence College.
Speaker 1 (26:19):
You and I. You and I can rate each other
on our on our propensity to uh to exhibit some
of these risky tendencies. But uh, but but it's on
a serious note. I think that the one of the
things that people get wrong, and I'm not going to
give it a name or not. It's part of a
lot of different fallacies, but incorrectly waiting small percentages, both
(26:40):
overweighting them and underwaiting them depending on their personal experience
and their situation. By small percentages, I mean like you know,
one percent risk, two percent risk, less than one percent risk.
I think that people are really really horrible at those
at those tiny percents. Is this one that you actually
had in your in your top mistakes as well or no?
Speaker 2 (27:03):
No, but I have something related And by the way,
I'm not sure somebody these kind of like the risk
to oh one mistakes, right is are not things like
the Gambler's fallacy. I mean people think that if you
get you know, if you get the coin heads four
times in a row, therefore you must get tails next time.
That's not true if the coin's fair and independent. There's
a little bit more advanced. No, what I do on
my list is not understanding the notion of calibration or forecast.
(27:27):
Maybe it's kind of like related to the gambler's fallacy. Yeah,
I think if you have and if you have a forecast.
We talked about the forecast model this week. Right, if
Biden has whatever, like a thirty five percent chance, he's
actually supposed to win the election thirty five percent of
the time. If the model is right, you are supposed
to lose your flushtraw hands and poker thirty five you know,
(27:48):
against a flush straw thirty five percent of the time,
or the deck is rigged, or you're like the luckiest
sun runner on Earth. I asked chat Shept. This question
is a way I used to like test large language models.
Maybe version four or something has gotten it right. But
you say, yeah, I know a forecast, are supposed to
every forecast right, it's not who for a probabilistic forecast.
And it's a long stand, long standing pet peeve of mine,
(28:12):
I think is yeah.
Speaker 1 (28:14):
Well I actually have that on my list too. So
I have this as two separate things. I have the
small percentages, which I think we can talk about separately,
but then what I have is rounding to absolutes, right
zero on one hundred, And I think that that's kind
of what this is on a fundamental level, where if
where if anything is like above seventy percent, or you're like, Okay,
this is going to happen, and if it's like, you know,
(28:36):
below below twenty percent, you're like, this ain't going to happen,
or you know that that's a little bit of a
rough estimate. But I think that that's often how human
minds work, right. We gravitate toward absolutes because we gravitate
towards certainty and away from uncertainty. And so if you're
talking about you know, you just mentioned the weather, if
you're talking about weather forecasters, this is like the quintessential
(28:58):
example of when people get it wrong. When you say,
you know, seventy five percent chance of rain, you bring
your umbrella and it doesn't rain, You're like, fuck, you know,
I've been lugging this umbrella around all day, liked it
or like ninety percent chance of sun and it starts raining.
When you get to the beach, You're like, what the hell,
Like it wasn't supposed to rain. Ten percent is really
not zero, But your brain just like goes to the
(29:20):
absolutes and the weather is something we've all experienced, and
yet we still, you know, we still have that notion.
So I do think that this rounding thing is incredibly,
incredibly important. But I think that we should talk separately
about the small percentages because that's also very important, and
I think that's a slightly different issue, especially when we're
talking about risks that are incredibly important, and we've talked
(29:44):
about some of them, Like we talked about p doom, right,
the risk of destruction in the past. We've talked about
the risks of one off storms events when it comes
to climate change. I mean, we have to deal with
small percentages all the time, and the brain just breaks down.
Speaker 2 (30:00):
Yeah. I was narrating the Silicon Valley chapter in my
book just last week, so was rethinking about some of
the stuff. You know, Silicon Valley understands the importance of
large payoffs. They live in fear of a Marcus Zuckerberg
or Elon Musk or whatever exiting their office on sand
(30:22):
Hill Road, turning down their offer or not having been
given an offer right when they could result in the
one hundred X or one thousand X or ten thousand
X payoff. They understand that there's a whole culture around
Silicon Valley that leads them to be very strange in
certain ways to encourage that type of risk taking because
it's like not normal for people. There is a second
(30:44):
I guess this is my number two, which is also
inspired by Silicon Valley though, which is understanding the value
of a portfolio of risks that if you have a
lot of these high upside, high risk bets, magically, if
you have enough of them, then your risk can be
actually quite low. The top decile Silicon Valley firms actually
(31:05):
are almost guaranteed a profit based on the research I've
done for my book, because they first of all, they
have like selection of xs, they get the best talent
in the door. But like you know, if you're making
a new fund every year and the fund has twenty
five companies, then you're hedging a lot. You might have
two hundred investments. If you have several years worth of
(31:27):
funds at any given time. Right, So so all of
a sudden, what seems risky is not if you can
aggregate a portfolio, which is the different options. And there
are issues about correlation, like we talked about in the
last segment. If you're only invested in crypto or AI
companies for example, those companies are gonna be correlated in
(31:48):
how well they do. But still it's a pretty good
business to be able to take enough high upside bets that, like,
you know, if you could enter Maria, for example, if
ten thousand copies of yourself could enter the World Series
of Poker main event, which is a profitable tournament, so
half the fucking players are clones of you. I mean
(32:08):
that we get a little weird, right, But yeah, but
you do it, you know, if you have the capital
for it.
Speaker 1 (32:13):
Yeah, oh absolutely, I would for sure. No, I think
that that's actually that's a great point. And I think
that diversification is something that we really do not think
about enough. And it's so funny because I'm just as
you're talking, I'm like, we have so many cliches in
like popular culture about this, like don't put all your
eggs in one basket, and what do we do over
and over and over. We take all our fucking eggs
(32:34):
and we put them in one basket, and we're like,
I'm just gonna go all out. I'm gonna go both
to the wall. I'm gonna like go all out on
this bet. That's not the correct way of taking risk.
Speaker 2 (32:45):
And we should take popular cliches and parables and idiom
to like rate them. We're like, oh, that's pretty smart,
right for sure, familiar eggs in one baskets of top
fucking five idiom. It absolutely difference.
Speaker 1 (32:59):
I mean it's saying yeah, I don't know, yeah, but
but what you know, I was just thinking about what
we talked about a few weeks ago about roaring kids,
like that that is the opposite of diversification as far
as we know, right, Like, that's just like putting all
of your bets onto this one thing, and like if
it ends up going south, like your millions disappear, right like,
(33:22):
you're fucked. And I think that that's really really difficult
for people to understand, especially when things are going well.
So I think that's psychologically it's really important to distinguish
moments when things are going well and moments when things
aren't going well because our ability to make good. Risky
decisions are really affected both ways. Right, So like if
you made this huge bet and it's going up, you're like,
(33:43):
I am a genius, right, Like I'm brilliant. This is amazing,
And even though you should probably exit at some point,
then you're not going to. And if things are going poorly,
then sometimes you actually end up doubling down. You don't
sell when you should because kind of you're emotionally invested,
and so you end up compounding your losses and not
(34:05):
thinking clearly once again. And so this is just such
a you know, such a broken way of human thinking
that then kind of compounds your mistakes. And this happens,
by the way to professional traders as well. This isn't
just you're You're just like someone sitting down on your couch.
(34:27):
So what's your what's your number three? Or do you
want me to do my number three?
Speaker 2 (34:32):
I can go all right. I'm trying to figure out
like a cute way to frame this. We need like
some something's law or something. I think people kind of
conflate risky decision making and ill informed decision making. I
don't think people understand that, like, oftentimes having more information
(34:55):
enables you to take more risk or more intelligent risks. Right.
So I was at dinner last night with my partner
some Italian place in the West Village which we had
not been to before, and he caught me looking at
Yelp to decide what I wanted to order. He's like,
why do you care about Carmen and dubukee Iowa? We
(35:18):
always make this woman from Iowa? What does she know
about dining in New York City?
Speaker 1 (35:25):
That's not what your person sounds like. It's a really
bad impression.
Speaker 2 (35:30):
I only could do an impression without him being self
conscious about it. He no, I'm not gonna I'm not gonna. Yeah,
I'm not gonna do it. But anyway, but he's like,
if you're such a risk taker, Nate, how can you
won't even order if fucking played a pasta without looking
at yelp reviews? And I'm like, I'm not a big
fan of octopus, don't love, don't hate, but you usually
(35:51):
wouldn't order an a menu with fifteen items or something.
Right if the review says this is the best octopus
you've ever tried, even if you're not an octopus like her,
you gotta try this octopus right then, I might get
the octopus. It can enable you to like to deviate
from your own plans. I mean, I think oftentimes on
restaurant used by Tyler Cowan, the economist on this, oftentimes
(36:14):
the least appetizing item on a menu is the thing
you should actually order, right, but it's only there because
like it must be good if it seems unappealing. Is
kind of the kind of Tyler's an economist, it's kind
of his theory. But yeah, the idea that like, uh,
you know, I know, the idea that like you should
(36:34):
just fly blind and it's like a smart risk to
take to just kind of wing it. I think that's
the opposite you know, preparation. If you talk to mountain
climbers I literally talk to in my book, they prepare
very extensively. It does not mean that every risk you
take it twenty nine thousand feet can be mitigated to zero,
but like, but they are the most prepared people in
the world pretty much now.
Speaker 1 (36:55):
I think when it comes to risk, like we do
want to be prepared and we do want to do
our homework. Like you're not going to just randomly fly
on an airline that you've never heard of before, Like
at least I'm not like I'm going to do some research.
I'm going to look at some reviews. I'm going to
see what safety record is. I'm going to see what
kind of planes safely And I think preparation is really
important in a lot of different moments. So so I'm
(37:18):
with you on this one, and I think that you know,
this is as I always say, and as as I
say when it comes to poker, when it comes to
all decision making, information is power, right, you want to
have the information advantage. And so that's that's a really
good way of looking at it. Now for my final thing,
I'm going to bring us more back to kind of
the economics realm and the realm of delay discounting, because
(37:39):
I think that that's something that we haven't talked about
and something that people have problems with. So delay discounting
is kind of your humans inability to kind of correctly
decide between smaller sooner rewards and larger later rewards. Right,
So kind of thinking about future and trying to figure out, Okay,
(38:01):
you know, what do I want now versus if I
can delay and take that risk and kind of get
this later, how do I want to think about it,
and I think fundamentally one of the reasons that people
are bad at it is we're really really horrible at
picturing ourselves in the future and knowing what our future
self is going to be like and what our future
(38:23):
self is going to want. And so this leads to
a lot of issues where like we just project our
current self right, like the Maria of today, and we
just imagine that the Maria in ten years from now
is going to be identical. And that's not the way
that we should be thinking about things. That's not the
way we should be thinking about risk. That's not the
way we should be planning for a risk, that's not
the way we should be just thinking about these things
(38:45):
in general. And that's kind of that's a big issue.
And one of the reasons why it's so difficult to
fix it is because it's not always irrational, because life
is unpredictable, and you know, there are some people who
would argue, like, take the reward right now, because you
might be dead in ten years, and I think that's
something that we also have to consider.
Speaker 2 (39:08):
If the AI doom for gusts or correct.
Speaker 1 (39:11):
Yeah, p do is correct, then take the reward now.
Speaker 2 (39:14):
I think for every ten people who err on that side, right,
for every ten people who err on the side of
being too impulsive in the moment or not necessarily impulse
or something wrong with impulse per se. But like, yeah,
applying to steep a discount to maximize their overall expected happiness.
(39:37):
To put it in a very clinical sounding way, there
have to be ten people who do that versus everyone
who does the opposite. And by the way, it's also
like the triad. Along with understanding expected value and understanding
portfolio theory, the thirteen silicon valley does well is they
have long time horizons. You know, even the best investments
(39:58):
often aren't profitable for ten thirteen years something like that.
It's just a huge advantage. I mean, even just investing
in the stock market, if you're in your twenty thirty
to forties, if your fifties early right us returns are
so much higher over the long run, there are enough
time for business cycles and there will be crashes eventually
to even out long time horizons are are are you know?
(40:20):
Evolutionarily you're probably not wired to do that, right, I mean,
you used to live shorter lives, Humans used to live
closer for this subsistence level, meaning that we didn't really
have any time of access at all. You protect what
you have, right, But you know a bird in the
hand is worth two in the bush? Is that like
another what's name mean? Yeah? Yeah, I love it?
Speaker 1 (40:42):
Yes, I got though, right it is. Yeah, it is
a bad one. It is a bad one. So we
have one good one. So so as our running ranking
of idioms of sayings, we we like that don't put
all your eggs in one basket. We don't like the
bird in the hand nearly as much.
Speaker 2 (41:00):
But it's it's a scarcity mindset to write. Like, Actually,
one thing I'll get annoyed by you've heard of, like
the marshmallow experiment, where like, of.
Speaker 1 (41:07):
Course Nate was my thesis advisor. I was this final
bad student.
Speaker 2 (41:12):
Well, I don't like the marshmallow experiment because I think
it kind of reads some privilege, where like if some
weird fucking graduate student was like, I'll give you two
marshmallows and I was like a poor kid, I'd be like,
I don't fucking trust you, right.
Speaker 1 (41:21):
No, no, See that's a very very silly criticism. We're not
going to get into it. But all this was controlled for.
Speaker 2 (41:26):
All segment coming up on the marshmallow problem. Tune in
next week.
Speaker 1 (41:33):
Yes, seriously, I will, I will come in. I know
this literature better better than probably better than most people alive,
and today so I will, I will, I will come
out guns of blazing. It's a very good study and
it was absolutely controlled for with the poor kids. That's
actually how it originated in in uh Island communities of
(41:56):
kids who had no money. So so quick recap my
top three incorrectly waiting small percentages both over and underwaiting
them rounding two absolutes to zero or one hundred, and
problems with delayed discounting so preferring smaller sooner two larger
later rewards, and how that might not always be irrational.
Speaker 2 (42:19):
And three. The first is the failure to understand the
nature of probabilistic forecasts. I will admit that's a pet
peeve number steen with three I think can be more
useful in your everyday life. Two is the failure to
understand the value of portfolio theory and diversification. And three
is the failure to understand that more information can actually
make you more risk taking and smarter atrist taking.
Speaker 1 (42:50):
Risky Business is hosted by me Maria Kanakova and me
Mate Silver. The show is a co production of Pushkin
Industries and iHeartMedia. This episode was produced by Isabel Carter.
Our associate producer is Gabriel Hunter Chang. Our executive producer
is Jacob Goldstein.
Speaker 2 (43:06):
And if you want to listen to an add free version,
sign up for Puchkin Plus. For six thirty nine a month,
you get access to ad free listening. Thanks for tuning in.
Pushkin. Welcome back to Risky Business, a show about making
better decisions. I'm Maria Kanakova.
Speaker 2 (00:28):
And I'm Nate Silvery. Today on the show, we'll talk
about my new presidential election forecast.
Speaker 1 (00:34):
Well also talk about insurance. Although Nate, I'm really excited
to hear about your election forecasts.
Speaker 2 (00:40):
You might want to ensure against that feeling, Marie, you
may not like what it has to say. I'm talking
about our top three ways that people get risk wrong.
Speaker 1 (00:54):
Nate, let's talk about your election model for the twenty
twenty four elections, which is out this week. You've been
working on this election model, tweaking it, developing it for
sixteen years, and it's been remarkably accurate. So I'm actually
really excited to talk about it and get into it
with you, even you told me that I really shouldn't
be so, I mean, I'm a little scared, too.
Speaker 2 (01:12):
Careful, careful what you wish for. I want you to
guess what number it has in the chance that trumpel
with the election.
Speaker 1 (01:20):
Oh night, this is are you gonna break my heart?
Is it okay? Fifty eight percent?
Speaker 2 (01:27):
Higher?
Speaker 1 (01:28):
Fuck?
Speaker 2 (01:29):
I don't feel the exact number because we're we chake
a couple days in advance. By the time the model
is released, there could be new polls and so forth.
I mean, we'll see what it says when it comes out.
But we're in the We're in the load of mid
sixties is kind of where the where the forecast is landing.
Speaker 1 (01:45):
Fucking you fuck fuck fuck is my is my response
to that night. You're usually the one dropping the f
phones on the show, but I think it's my turn. Yeah.
Speaker 2 (01:53):
Look, my job here is not to tell you who
you should vote for. It's to forecast the election as
best I can. And I think it's disingenuous too to
call the race a toss up. You can kind of
squint and say, oh, if you look at national polls
and Biden's maybe only down by half a pointer or so.
(02:14):
But the problem in the US is that we don't
determine elections by the popular vote. Right, if it was
a popular vote, the model things it would be a
toss up, or maybe Biden slightly favored. In the electoral college.
There has been for the past two cycles with Trump
on the ballot, a gap that favors a GOP. In
twenty and twenty, Biden won the popular vote by four
(02:35):
and a half points, which is a pretty good margin
against an incumbent. But all these key states Georgia and
Wisconsin and Pennsylvania and so forth, the only win by
about one point. So if we're now in twenty twenty
four and Biden's trailing by half a point instead of
winning by four and a half points in the popular vote,
that's a shift of five points. If you have a
(02:57):
state like Pennsylvania that you win by one point and
you shift it by five points, that's not very good. Now.
What's happening in practice is that there's a bit of
a bifurcation between the Midwestern swing states is kind of
trio of rust Belt states Pennsylvania, maybe even maybe not
at the Midwest per se, but Pennsylvania, Wisconsin, Michigan, where
BIA's polling is closer usually within the point maybe two
(03:20):
point Trump lead, versus the sun Belt states, the new
fangled Georgia, Nevada, Arizona, where Biden's polling is very bad.
He might be down five or six points. Now. If
he wins Michigan and Pennsylvania and Wisconsin and holds all
the other states so like New Hampshire, Virginia and so forth,
(03:42):
then he would finish with exactly two hundred and seventy
electoral votes when you need to sixty nine to win.
So there is a path, but not a whole lot
of margin. For aer.
Speaker 1 (03:55):
Nate, this is not the way that I wanted to
start off my week. So I think I think we're
done with the podcast, goodbye, good bye luck as I
go and hide in a cave. But no, in all honesty,
this is sobering news. I so don't want Trump to
win because I know I think it's going to have
incredibly dire consequences for democracy. So I want to kind
(04:20):
of squint and say, yeah, but there's this path and
look at these states, and we're going to do it.
But I think that you need to be realistic and
try to figure out, Okay, given this, there's still it's
not like the election is tomorrow, right, So what can
we do? What shifts in your election model would actually
change that outcome and bring that percentage back down? Are
there certain lovers? Are there certain things that are actually
(04:43):
going to be kind of the most meaningful as you
play around with the inputs and the numbers.
Speaker 2 (04:49):
Yeah, look, Biden is struggling with groups that are traditionally
strong Democratic groups like younger black voters, younger Hispanic voters.
That's why you see his numbers fading in states like Georgia,
for instance, whereas he's actually doing okay with these older
white voters. That seems to be the constituency. It's like
holding up for him the best. There is a debate today,
(05:10):
This episode's coming out on Thursday tonight. That's an opportunity
for Biden to turn things around. You know, the conventions
have get to happen. The concern for Biden is that
the race has been extremely steady so far, the steadiest
you've ever seen Poland. I think in a presidential election
Biden game maybe half a point to a point in
(05:32):
the polls following Trump's conviction on thirty However, many it
was felony counts, you know. But the fact that like
Trump was convicted of paying hush money payments to like
at porn star adult film actress, excuse me. And that's
like a real sentence I can say, And that like
(05:53):
had half a point worth of difference. I mean, that's
that's not great for Biden. I think there have been
a lot of false hopes from Democrats. People say, oh,
this is going to move the poles. It's going to
move the polls when the economy has gotten a fair
bit better over the past six months. Inflation has abate,
the stock market is way up actually, but like the
(06:13):
numbers haven't really changed very much. And I don't know.
I'm sure I'll be called a fascist, biased, republican hack
by every democratic strategist in the book. But the whole
point of having a model is that you're applying structured
thinking where where you don't go in and tweak the
results once you see what it spits out. Right, The
(06:35):
whole point of committing to a model is that like
it does take the emotional component, which I have like
everybody else, out of the process or reduces it. Let's say, right,
I'm someone who thinks professionally, it would be great to
just say, oh, it's a toss of fifty to fifty,
who knows can't get in triple that way. Well, that's
dishonest and it's not what the number says, and it's
(06:55):
not what the process I followed for sixteen year now
years now is two thousand and eight says right, this
is not a year that looks terribly uncertain. You don't
have COVID. You don't have these polls volleying up and
down like you did in twenty sixteen with Clinton and Trump.
They're very well known candidates, and so Trump's advantage is narrow,
but it's been persistent, and he has the more robust paths.
(07:20):
The fact that Biden has to win all three of
those Michigan, Wisconsin, Pennsylvania trio, all where he probably trails
by a point or so, but he has to win
all three and hold these other states. The fact that
like the next best wing states like Arizona for Biden,
where he's down by like four or five points, that's
a real deficit. I don't think you can sit here
(07:41):
with a straight face and say you'd rather be by
I think you can say you'd rather be Trump.
Speaker 1 (07:47):
You started off by saying that it seems that Biden
is slipping amongst some of those voters who are kind
of traditionally democratic, right, like the younger black voters. Are
those voters actually switching to Trump or what's happening to them?
So where's that loss for Biden evaporating? Two?
Speaker 2 (08:06):
I think there might be, you know, some permissions ructures
to vote for Trump that didn't exist last time, but no,
a lot of them are going to undecided, to RFK
junior or out of the electorate. You know, our model
projects turn out of something like one hundred and fifty
million this year, which is which would be down from
(08:28):
twenty to twenty in all the polls, the indications that
there's less enthusiasm for this election than there had been.
Now to Biden this credit, he actually does well. Democrats
in general do well with these very reliable regular voters.
So like, lower turnout might actually be beneficial for Biden.
Speaker 1 (08:47):
So Nate, Okay, since we since we are we are here,
what are kind of if you are at the Biden camp,
where should they focus their efforts between now and election time?
Like give us like top top two, top three priorities,
Like what should the campaign be doing to try to
get those numbers back to where in more winnable territory?
Speaker 2 (09:13):
Maybe fire some people I don't know, certainly, fire anybody
in the campaig who's telling you not to believe the polls, right,
fire them and don't get them a new job. I mean, seriously,
if you're down, if you know, again, if you're at
a mid thirties, maybe maybe they'll got up to the
low forties by the time we published the model, but
probably mid thirties. If you're in that territory, then you
know that's the point when again, when a baseball team
(09:35):
is winning thirty some five six seven percent of its
ballgames and it tends to fire the manager and there's
been very little turnover, and like, their theory of the
case is wrong. They thought that Trump would be relatively
easy to defeat and they could run back the twenty
twenty playbook. You're talking about democracy, democracy, democracy, over and
over and over again, and that this would work. Their
theory of the case is wrong. They did not expect
(09:57):
to be down at this point, right. This is why
they're still doing the RFK junior stuff, because they thought
they'd be winning, and so you want to reduce variants.
When you're ahead and getting RFK junior, who's a wild card?
I guess you can say, off the b it is
something to do that. You know. Look, I think some
of this stuff that you've done tactically suggests improvement, but
(10:18):
but I don't know. I mean, I mean, you know,
for all the fucking talk about how important the election is,
and I think some of that talk is we'll scrutinize
that more. At some point. I think some of it's
a little precious. But like if you really think that
like this is the you know, the end all be
all of elections, then I'm not sure really acting like
(10:38):
it very much. You know, they're acting like people who
have incentive. We're called principal agent problems. Right, you don't
want to be the polster that gives the president bad news.
You definitely don't want to be the polster or the
advisor who tells the president, hey, maybe you should stand
down to think about having someone else run, because then
you lose your job. But you know, and then there's
(10:59):
a whole Kamala Harris issue too, where I mean, I
think if she were more popular, then there's some chance
that Biden would have stood down and Democrats to be
in better shape. So I don't know. Look, all you
can say is like, if Biden loses, and there's still
a pretty some chance he wins, it's not like you
can't see this coming. He has been behind in the
(11:20):
polls consistently for you know, since Midsummer last year. Voters
consistently have said you are too old, you should not
run again. The first term was fine, you should not
run again. You know they have ignored that message at
every turn, and that's because the data is ambiguous. Right again,
you can squint in the right day and look at
the right poles and see not just a path to
(11:43):
it Biden victory, but like a case where his chances
look pretty decent. Right, you can't look at things like
the twenty twenty two midterms where Democrats did relatively well,
and you can hope that the polls are biased in
your favor because biased against you, rather to outperform them,
because sometimes they are. But that's usually very hard to predict,
(12:04):
and you don't need many reminders of cases where the
polls were biased against Trump in stead.
Speaker 1 (12:10):
All right, so shake things up. I think we see
the old status quo bias that we've talked about week
after week after week, and firesome people focus on those
key states and like, just get your ass and gear
and realize that there's a very very good chance you'll lose.
Speaker 2 (12:29):
We'll be right back. Nay.
Speaker 1 (12:43):
Let's talk about some huge shifts in the home insurance industry.
There has been a lot of news recently about homeowners
insurance because insurers are actually getting out of multiple states
multiple areas and refusing just flat out and refusing to
ensure homes, which is obviously affecting homeowners, home buyers, people
(13:04):
who need to get you know, their their home mortgage,
anything like that, because they like physically cannot get their
home insured. And Nate, last week, you and I talked
a lot about climate change and all of those effects,
and this is directly related to that. So I think
that this is one way that consumers are feeling what
it means to be living in a world where the
(13:26):
risks of one off events are no longer one off events,
where the storm of the century is no longer the
storm of the century, where you can actually see these
people who use models to try to model out the
risks losing money and saying, oh shit, we don't want
to ensure your home anymore. We no longer want to
take this risk.
Speaker 2 (13:47):
So let me ask, Yeah, what might seem like a
naive question, right, why can't you just price these risks
in that you're in California and you know the risk
of wildfires, you're in Florida, you know the risk of hurricanes.
You can build a model and then it costs more
than it might what's a safe state, you know, Wyoming there,
(14:08):
like Bolcanos there and that anyway, but why can't you
charge more instead of not offering the risk at all?
Speaker 1 (14:16):
Yeah, I mean, I think that that's a really really
good point, and I think that some insurance are trying
to do that and trying to charge more. But I
think that some are using old data instead of new
data for the models. And the problem is there are
no good data, right. How do you model something that's
changing so rapidly in an environment that's changing rapidly, where
there's a lot of volatility, and volatility in the actual
(14:39):
sense of the word, which means that you know, ups
and downs are very extreme in a short period of time.
How do I factor in the uncertainty of these risks
given that we don't have any historical comps, right, given
that we don't know how it's going to play out
in the next ten years, because the last ten years
are not representative of the years before that are not representative,
(15:00):
and we no longer have this steady model where where
we can go back and say, Okay, you know, risk
of this x percent, risk of this experson. I mean,
I know that one way of handling volatility is to
price it even higher, but at some point you know
you're pricing it so high because of the uncertainty that
it might not make sense anymore.
Speaker 2 (15:23):
Yeah, if you go to the New York Times, they
have a interactive graphic called the Home Insurance Crunch, see
What's happening in your state, where it just shows every
state whether home insurance has been profitable or losing for
the insurers. And what you quickly discover is that you
have kind of limited upside where nothing goes wrong you
(15:46):
do make a profit, but kind of uncapped downside where
one year with crazy hurricanes or flooding or wildfires or
things like that can wipe out ten marginally profitable years.
And in states like California and Colorado a lot of
fires Florida, of course, with hurricanes like that happens pretty often,
(16:10):
so it's hard to measure tail risks, especially if the
tails are expanding. We talked on the last week's episode
about how if there's even a small shift in the distribution,
the part of the curve under the tail can really
get shifted out where it becomes substantially more probable. Yeah,
(16:31):
and there's also you know, there's also something called adverse
selection where maybe the people who want to buy insurance
are people that are worse risks for you to take.
Maybe they know something about their home that it's vulnerable
to landslides, for instance, And therefore you kind of have
what economists we call like a market for lemons, where
you don't you have information asymmetries between buyer and seller
(16:52):
that reduce the overall size of the market. And in
those cases it might be irrational to might be rational
rather to not offer a product. There's not a kind
of market clearing price.
Speaker 1 (17:03):
Wait Nate, one second, let's clarify what exactly a market
for lemons means.
Speaker 2 (17:08):
Yeah, the term was coined, I believe by the economist
George Akerlove, who I think is now at Berkeley won
a Nobel Price. He talks about, I mean, Elemon is
a defective used car, and it's hard to know everything
wrong with the car if you take a test drive
for a few minutes or something like that. But basically,
(17:29):
it's like when buyers don't have any way to trust sellers.
There's no third party vouching for the reliability of these
used cars. What happens is just the volume of transactions
goes way down. Right, You have dead weight loss when
there's no credible way to achieve trust. So the theory
here would be, like, you know, if we think that
(17:51):
insurance is a win win transaction, right, someone's paying some
expected value on average, you lose money if you buy insurance,
but they protect their downside risk. Right, that's a win
win on a risk adjusted basis. But if there's not
reliable information you're not sure you can trust the insurer,
or if you're insured, you're not sure you can trust
the homeowner, then you have transactions not happening that would
(18:15):
create utility for both sides and for society.
Speaker 1 (18:19):
Yeah, I think that something that I know about investing
is if you're trying to kind of look at an
opportunity and you're comparing upside and downside risk. Right, when
you have unlimited downside risk and your upside risk is
your upside is capped, that's not an investment you want
to be making, right. You want it to be asymmetric
(18:41):
the other way. You want your downside risk to be
capped and your upside to actually be able to go
up further, especially if you're leveraged. Right, If you're thinking
about it that way, and if you're thinking about the
fact that this is incredibly asymmetric risk, and this is
something that they've you know, this is something they've been
doing for years where it's been asymmetric in their favor,
(19:02):
where it has been basically a capped downside because these
kind of once in a lifetime events didn't happen, or
they did happen once in a lifetime, and so all
of your other years where you were making money paid
for it. But now if you're making less money, because
there's always something happening, right if you look at the
news every year, there's something happening. There's you know, home
(19:25):
destruction because of all sorts of different things, tornadoes, hurricanes, wildfires,
you know, you name it, you know, we got it.
And then if you have that, so in your up years,
up years, you're not nearly as up and then you
have these catastrophic events that are happening over and over
and over. Then to me, like, as if you're a
rational investors, it's a bad investment. So you end up
(19:47):
pulling out.
Speaker 2 (19:48):
What's not neglect the role of dorm regulations. According to
the Washington Post, in California's case, insurance companies must use
historical data rather than forward looking models when they price
insurance plans that means is supporting them. Their policies may
not reflect the actual risk you're supposed to hedge against.
That's stupid. I mean, there's no serious put laws in California,
(20:12):
and like I happen to agree with that. Night if
you can't actually like adjust for changes in the climate,
which I think most people in California would care about,
then you're just making the market less efficient. You know,
how you optimize this function where on the one hand,
you probably don't want insurers to have literally unlimited liability
(20:34):
because they can't. No one has infinity infinity dollars, right,
they won't be able to pay it. On the other hand,
you know, you don't want insurance not to make people
relatively whole when there is a disaster. Otherwise what's the point.
But I would gather that one that like these state
regulations are not always that well designed.
Speaker 1 (20:51):
Yeah, I think that that's absolutely right because and states
are trying to step in, by the way, not just
in California, right, because there are these gaps where they say,
oh shit, you know, what do we do if people
in our state can't buy houses because they can't get insurance,
Like this is actually a big, big issue, Nate. I
know that at some point point in the in the past,
(21:11):
you were considering buying a place in Miami. Is you know,
is that something where like, let's talk about this on
a personal level, Like, would you do that right now,
given what's happening with insurance, given what we know about
the risks there, given that you probably would not be
able to get insurance on the you know, on an
apartment that you wanted.
Speaker 2 (21:32):
Honest answer, they probably wouldn't affect my decision that much
because look, home insurance is negative expective value on a
non risk adjusted basis, right, Otherwise it wouldn't be offered.
By definition, insurance means that someone who has more capital
and more risk tolerance than you makes money on average,
(21:54):
and homeowners lose money on average. I mean, it wouldn't
be great. I'll probably take the negative evy like safe bet.
I mean, I'm probably not going to buy a place
in Mimi for other reasons. But no, I mean, but
I'm also conscious about how much of my net worth
I'm tied it being real estate, right. I think it's
(22:16):
generally a mistake, for many reasons to have a great
deal of your net worth tied up in real estate
as opposed to a broader portfolio of stuff. So I
think I think I'm hedged enough in diverse fi it
enough that, like, although not ideal, I I you know,
it wouldn't be a deal breaker.
Speaker 1 (22:33):
So you would basically write it off, right, Like, as
this is an investment that might go to zero, I
don't care that. I'm not going to get insurance on it,
and I understand that it might get destroyed, and I'm
okay with that. I'm just going to make it a
small enough investment that if it goes to zero, I'll
be okay.
Speaker 2 (22:50):
That's right.
Speaker 1 (22:51):
Yeah, So this, I think. So this actually kind of
illustrates that it's affecting different people differently. Where if you
are in a position where you can do that and
you can say I will buy this place and I'm
just going to write it off, then you can buy it.
But if you're someone who can't do that, then the
calculus changes.
Speaker 2 (23:08):
Like with those things, people who are you know, better
off financially or who have more robust ways of looking
at risk benefit from it, and the people who are
disadvantaged are harmed by him more because yeah, I mean, look,
if you're like a multi billionaire in Silicon Valley or something, right,
(23:30):
and you lose your beautiful seven million dollar home in
a wildfire or an earthquake, you'll still be okay, Right,
if most of your net worth is tied up in
your home, then you won't be.
Speaker 1 (23:42):
So it seems like when we're talking about this, we
basically have the absolute like shit show of factors, if
I may be, if I may use a psychological term here,
So we have huge correlations, right, we have no good
historic data, we have a lot of things changing, and
(24:03):
so it's incredibly difficult to model. And so it's basically like,
you know, this is your worst nightmare when you're trying
to build a model that predicts risks and tries to
figure out what in the world do we do? And
so I think that at least in the short term,
what I'm taking away from this is that a lot
of people, especially kind of the people who cannot afford
(24:23):
to just write off these these homes, are kind of
fucked and cannot buy homes, have to move out of
homes that they've owned where the insurers have stepped out,
and there's really no clear solution. And as you know,
as someone who is making these decisions on a personal level.
Where do I move, you know, where where do I
buy a house, Where do I rent? Where do I
look at real estate? Where do I build a community.
(24:45):
I think these are all things that people need to
keep in mind as they're thinking about the future. And
sometimes that sucks because sometimes you know, you really want
to move to beautiful wildfire wildfire California or lovely coastal Florida,
and that's just kind of the emotional elements of that
(25:06):
are not going to be worth it for you in
the long term. We'll be right back after the break.
Speaker 2 (25:24):
All right, Maria, Let's talk about our top three ways
and people get risk wrong.
Speaker 1 (25:29):
So on the show we often talk about kind of
these meta risk concepts, you know, elections and those types
of big questions. But right now I'd love to talk
bring it back to kind of individual psychology and talk
about how people, how individuals get risk assessments wrong, the
biggest problems that they have when they're thinking about it
(25:52):
on an individual level. And I think you and I
both have some thoughts as to what the most important
ones are. I can lead us off if you'd like.
Speaker 2 (26:00):
I think that one of the biggest a draft of
like cognitive biases.
Speaker 1 (26:05):
Yeah, draft of cognitive biases, but specifically.
Speaker 2 (26:09):
The amplast fallacy. You're looking pretty good and picking them
at two.
Speaker 1 (26:13):
Exactly exactly.
Speaker 2 (26:15):
It's rookie out of Providence College.
Speaker 1 (26:19):
You and I. You and I can rate each other
on our on our propensity to uh to exhibit some
of these risky tendencies. But uh, but but it's on
a serious note. I think that the one of the
things that people get wrong, and I'm not going to
give it a name or not. It's part of a
lot of different fallacies, but incorrectly waiting small percentages, both
(26:40):
overweighting them and underwaiting them depending on their personal experience
and their situation. By small percentages, I mean like you know,
one percent risk, two percent risk, less than one percent risk.
I think that people are really really horrible at those
at those tiny percents. Is this one that you actually
had in your in your top mistakes as well or no?
Speaker 2 (27:03):
No, but I have something related And by the way,
I'm not sure somebody these kind of like the risk
to oh one mistakes, right is are not things like
the Gambler's fallacy. I mean people think that if you
get you know, if you get the coin heads four
times in a row, therefore you must get tails next time.
That's not true if the coin's fair and independent. There's
a little bit more advanced. No, what I do on
my list is not understanding the notion of calibration or forecast.
(27:27):
Maybe it's kind of like related to the gambler's fallacy. Yeah,
I think if you have and if you have a forecast.
We talked about the forecast model this week. Right, if
Biden has whatever, like a thirty five percent chance, he's
actually supposed to win the election thirty five percent of
the time. If the model is right, you are supposed
to lose your flushtraw hands and poker thirty five you know,
(27:48):
against a flush straw thirty five percent of the time,
or the deck is rigged, or you're like the luckiest
sun runner on Earth. I asked chat Shept. This question
is a way I used to like test large language models.
Maybe version four or something has gotten it right. But
you say, yeah, I know a forecast, are supposed to
every forecast right, it's not who for a probabilistic forecast.
And it's a long stand, long standing pet peeve of mine,
(28:12):
I think is yeah.
Speaker 1 (28:14):
Well I actually have that on my list too. So
I have this as two separate things. I have the
small percentages, which I think we can talk about separately,
but then what I have is rounding to absolutes, right
zero on one hundred, And I think that that's kind
of what this is on a fundamental level, where if
where if anything is like above seventy percent, or you're like, Okay,
this is going to happen, and if it's like, you know,
(28:36):
below below twenty percent, you're like, this ain't going to happen,
or you know that that's a little bit of a
rough estimate. But I think that that's often how human
minds work, right. We gravitate toward absolutes because we gravitate
towards certainty and away from uncertainty. And so if you're
talking about you know, you just mentioned the weather, if
you're talking about weather forecasters, this is like the quintessential
(28:58):
example of when people get it wrong. When you say,
you know, seventy five percent chance of rain, you bring
your umbrella and it doesn't rain, You're like, fuck, you know,
I've been lugging this umbrella around all day, liked it
or like ninety percent chance of sun and it starts raining.
When you get to the beach, You're like, what the hell,
Like it wasn't supposed to rain. Ten percent is really
not zero, But your brain just like goes to the
(29:20):
absolutes and the weather is something we've all experienced, and
yet we still, you know, we still have that notion.
So I do think that this rounding thing is incredibly,
incredibly important. But I think that we should talk separately
about the small percentages because that's also very important, and
I think that's a slightly different issue, especially when we're
talking about risks that are incredibly important, and we've talked
(29:44):
about some of them, Like we talked about p doom, right,
the risk of destruction in the past. We've talked about
the risks of one off storms events when it comes
to climate change. I mean, we have to deal with
small percentages all the time, and the brain just breaks down.
Speaker 2 (30:00):
Yeah. I was narrating the Silicon Valley chapter in my
book just last week, so was rethinking about some of
the stuff. You know, Silicon Valley understands the importance of
large payoffs. They live in fear of a Marcus Zuckerberg
or Elon Musk or whatever exiting their office on sand
(30:22):
Hill Road, turning down their offer or not having been
given an offer right when they could result in the
one hundred X or one thousand X or ten thousand
X payoff. They understand that there's a whole culture around
Silicon Valley that leads them to be very strange in
certain ways to encourage that type of risk taking because
it's like not normal for people. There is a second
(30:44):
I guess this is my number two, which is also
inspired by Silicon Valley though, which is understanding the value
of a portfolio of risks that if you have a
lot of these high upside, high risk bets, magically, if
you have enough of them, then your risk can be
actually quite low. The top decile Silicon Valley firms actually
(31:05):
are almost guaranteed a profit based on the research I've
done for my book, because they first of all, they
have like selection of xs, they get the best talent
in the door. But like you know, if you're making
a new fund every year and the fund has twenty
five companies, then you're hedging a lot. You might have
two hundred investments. If you have several years worth of
(31:27):
funds at any given time. Right, So so all of
a sudden, what seems risky is not if you can
aggregate a portfolio, which is the different options. And there
are issues about correlation, like we talked about in the
last segment. If you're only invested in crypto or AI
companies for example, those companies are gonna be correlated in
(31:48):
how well they do. But still it's a pretty good
business to be able to take enough high upside bets that, like,
you know, if you could enter Maria, for example, if
ten thousand copies of yourself could enter the World Series
of Poker main event, which is a profitable tournament, so
half the fucking players are clones of you. I mean
(32:08):
that we get a little weird, right, But yeah, but
you do it, you know, if you have the capital
for it.
Speaker 1 (32:13):
Yeah, oh absolutely, I would for sure. No, I think
that that's actually that's a great point. And I think
that diversification is something that we really do not think
about enough. And it's so funny because I'm just as
you're talking, I'm like, we have so many cliches in
like popular culture about this, like don't put all your
eggs in one basket, and what do we do over
and over and over. We take all our fucking eggs
(32:34):
and we put them in one basket, and we're like,
I'm just gonna go all out. I'm gonna go both
to the wall. I'm gonna like go all out on
this bet. That's not the correct way of taking risk.
Speaker 2 (32:45):
And we should take popular cliches and parables and idiom
to like rate them. We're like, oh, that's pretty smart,
right for sure, familiar eggs in one baskets of top
fucking five idiom. It absolutely difference.
Speaker 1 (32:59):
I mean it's saying yeah, I don't know, yeah, but
but what you know, I was just thinking about what
we talked about a few weeks ago about roaring kids,
like that that is the opposite of diversification as far
as we know, right, Like, that's just like putting all
of your bets onto this one thing, and like if
it ends up going south, like your millions disappear, right like,
(33:22):
you're fucked. And I think that that's really really difficult
for people to understand, especially when things are going well.
So I think that's psychologically it's really important to distinguish
moments when things are going well and moments when things
aren't going well because our ability to make good. Risky
decisions are really affected both ways. Right, So like if
you made this huge bet and it's going up, you're like,
(33:43):
I am a genius, right, Like I'm brilliant. This is amazing,
And even though you should probably exit at some point,
then you're not going to. And if things are going poorly,
then sometimes you actually end up doubling down. You don't
sell when you should because kind of you're emotionally invested,
and so you end up compounding your losses and not
(34:05):
thinking clearly once again. And so this is just such
a you know, such a broken way of human thinking
that then kind of compounds your mistakes. And this happens,
by the way to professional traders as well. This isn't
just you're You're just like someone sitting down on your couch.
(34:27):
So what's your what's your number three? Or do you
want me to do my number three?
Speaker 2 (34:32):
I can go all right. I'm trying to figure out
like a cute way to frame this. We need like
some something's law or something. I think people kind of
conflate risky decision making and ill informed decision making. I
don't think people understand that, like, oftentimes having more information
(34:55):
enables you to take more risk or more intelligent risks. Right.
So I was at dinner last night with my partner
some Italian place in the West Village which we had
not been to before, and he caught me looking at
Yelp to decide what I wanted to order. He's like,
why do you care about Carmen and dubukee Iowa? We
(35:18):
always make this woman from Iowa? What does she know
about dining in New York City?
Speaker 1 (35:25):
That's not what your person sounds like. It's a really
bad impression.
Speaker 2 (35:30):
I only could do an impression without him being self
conscious about it. He no, I'm not gonna I'm not gonna. Yeah,
I'm not gonna do it. But anyway, but he's like,
if you're such a risk taker, Nate, how can you
won't even order if fucking played a pasta without looking
at yelp reviews? And I'm like, I'm not a big
fan of octopus, don't love, don't hate, but you usually
(35:51):
wouldn't order an a menu with fifteen items or something.
Right if the review says this is the best octopus
you've ever tried, even if you're not an octopus like her,
you gotta try this octopus right then, I might get
the octopus. It can enable you to like to deviate
from your own plans. I mean, I think oftentimes on
restaurant used by Tyler Cowan, the economist on this, oftentimes
(36:14):
the least appetizing item on a menu is the thing
you should actually order, right, but it's only there because
like it must be good if it seems unappealing. Is
kind of the kind of Tyler's an economist, it's kind
of his theory. But yeah, the idea that like, uh,
you know, I know, the idea that like you should
(36:34):
just fly blind and it's like a smart risk to
take to just kind of wing it. I think that's
the opposite you know, preparation. If you talk to mountain
climbers I literally talk to in my book, they prepare
very extensively. It does not mean that every risk you
take it twenty nine thousand feet can be mitigated to zero,
but like, but they are the most prepared people in
the world pretty much now.
Speaker 1 (36:55):
I think when it comes to risk, like we do
want to be prepared and we do want to do
our homework. Like you're not going to just randomly fly
on an airline that you've never heard of before, Like
at least I'm not like I'm going to do some research.
I'm going to look at some reviews. I'm going to
see what safety record is. I'm going to see what
kind of planes safely And I think preparation is really
important in a lot of different moments. So so I'm
(37:18):
with you on this one, and I think that you know,
this is as I always say, and as as I
say when it comes to poker, when it comes to
all decision making, information is power, right, you want to
have the information advantage. And so that's that's a really
good way of looking at it. Now for my final thing,
I'm going to bring us more back to kind of
the economics realm and the realm of delay discounting, because
(37:39):
I think that that's something that we haven't talked about
and something that people have problems with. So delay discounting
is kind of your humans inability to kind of correctly
decide between smaller sooner rewards and larger later rewards. Right,
So kind of thinking about future and trying to figure out, Okay,
(38:01):
you know, what do I want now versus if I
can delay and take that risk and kind of get
this later, how do I want to think about it,
and I think fundamentally one of the reasons that people
are bad at it is we're really really horrible at
picturing ourselves in the future and knowing what our future
self is going to be like and what our future
(38:23):
self is going to want. And so this leads to
a lot of issues where like we just project our
current self right, like the Maria of today, and we
just imagine that the Maria in ten years from now
is going to be identical. And that's not the way
that we should be thinking about things. That's not the
way we should be thinking about risk. That's not the
way we should be planning for a risk, that's not
the way we should be just thinking about these things
(38:45):
in general. And that's kind of that's a big issue.
And one of the reasons why it's so difficult to
fix it is because it's not always irrational, because life
is unpredictable, and you know, there are some people who
would argue, like, take the reward right now, because you
might be dead in ten years, and I think that's
something that we also have to consider.
Speaker 2 (39:08):
If the AI doom for gusts or correct.
Speaker 1 (39:11):
Yeah, p do is correct, then take the reward now.
Speaker 2 (39:14):
I think for every ten people who err on that side, right,
for every ten people who err on the side of
being too impulsive in the moment or not necessarily impulse
or something wrong with impulse per se. But like, yeah,
applying to steep a discount to maximize their overall expected happiness.
(39:37):
To put it in a very clinical sounding way, there
have to be ten people who do that versus everyone
who does the opposite. And by the way, it's also
like the triad. Along with understanding expected value and understanding
portfolio theory, the thirteen silicon valley does well is they
have long time horizons. You know, even the best investments
(39:58):
often aren't profitable for ten thirteen years something like that.
It's just a huge advantage. I mean, even just investing
in the stock market, if you're in your twenty thirty
to forties, if your fifties early right us returns are
so much higher over the long run, there are enough
time for business cycles and there will be crashes eventually
to even out long time horizons are are are you know?
(40:20):
Evolutionarily you're probably not wired to do that, right, I mean,
you used to live shorter lives, Humans used to live
closer for this subsistence level, meaning that we didn't really
have any time of access at all. You protect what
you have, right, But you know a bird in the
hand is worth two in the bush? Is that like
another what's name mean? Yeah? Yeah, I love it?
Speaker 1 (40:42):
Yes, I got though, right it is. Yeah, it is
a bad one. It is a bad one. So we
have one good one. So so as our running ranking
of idioms of sayings, we we like that don't put
all your eggs in one basket. We don't like the
bird in the hand nearly as much.
Speaker 2 (41:00):
But it's it's a scarcity mindset to write. Like, Actually,
one thing I'll get annoyed by you've heard of, like
the marshmallow experiment, where like, of.
Speaker 1 (41:07):
Course Nate was my thesis advisor. I was this final
bad student.
Speaker 2 (41:12):
Well, I don't like the marshmallow experiment because I think
it kind of reads some privilege, where like if some
weird fucking graduate student was like, I'll give you two
marshmallows and I was like a poor kid, I'd be like,
I don't fucking trust you, right.
Speaker 1 (41:21):
No, no, See that's a very very silly criticism. We're not
going to get into it. But all this was controlled for.
Speaker 2 (41:26):
All segment coming up on the marshmallow problem. Tune in
next week.
Speaker 1 (41:33):
Yes, seriously, I will, I will come in. I know
this literature better better than probably better than most people alive,
and today so I will, I will, I will come
out guns of blazing. It's a very good study and
it was absolutely controlled for with the poor kids. That's
actually how it originated in in uh Island communities of
(41:56):
kids who had no money. So so quick recap my
top three incorrectly waiting small percentages both over and underwaiting
them rounding two absolutes to zero or one hundred, and
problems with delayed discounting so preferring smaller sooner two larger
later rewards, and how that might not always be irrational.
Speaker 2 (42:19):
And three. The first is the failure to understand the
nature of probabilistic forecasts. I will admit that's a pet
peeve number steen with three I think can be more
useful in your everyday life. Two is the failure to
understand the value of portfolio theory and diversification. And three
is the failure to understand that more information can actually
make you more risk taking and smarter atrist taking.
Speaker 1 (42:50):
Risky Business is hosted by me Maria Kanakova and me
Mate Silver. The show is a co production of Pushkin
Industries and iHeartMedia. This episode was produced by Isabel Carter.
Our associate producer is Gabriel Hunter Chang. Our executive producer
is Jacob Goldstein.
Speaker 2 (43:06):
And if you want to listen to an add free version,
sign up for Puchkin Plus. For six thirty nine a month,
you get access to ad free listening. Thanks for tuning in.
Popular Podcasts
1. Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
2. Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations.
3. Crime Junkie
If you can never get enough true crime... Congratulations, you’ve found your people.