0:42

Brad and Terry had access to a broad sample.

A brokerage account from a large discount brokerage over the period 1991 to 1996.

So by discount brokerage, we mean this is people are placing trades on their own.

There's no advice given to them, they're just making their trades and

this is really kind of the gold standard in terms of the database to look

at individual investors performance and it covers a broad sample of households across

this US over 66,000 households seeing how they trade over the 6-year period.

So let's kind of think about individuals and

trading stocks on their own, Le Penseur.

So, let's do a pause, think and answer.

The question that I want you to think about is what are the advantages and

disadvantages of trading stocks on your own?

So, you may have your own portfolio where you buy and sell stocks.

So, you're probably implicitly already making this calculation in your head.

Well, let's be explicit.

What are the advantages and

what are the disadvantages of trading stocks on your own?

1:52

So what are the advantages and

disadvantages of trading stocks on your own?

Well, the advantage of trading stock on your own if you think you're well

informed, then you can kind of put your money where your mouth is and

buy the stock that you think are going to go up.

Sell the stock that you think are going to go down.

Also, by buying and selling stocks on your own, you're not

paying any mutual fund fees, you're not paying any fees to financial advisers.

Now, what are the disadvantages of trading stocks on your own?

Well, sometimes these mutual fund fees, particularly for

index funds are very low like five basis points per year and

it enables you to get broad diversification at a very, very low cost.

2:34

Also, when you're selling stocks, buying stocks a lot,

you have to pay transactions costs.

There's commissions you pay to a brokerage.

There's also costs, what we'll talk about later,

that you pay in terms of the bid ask-spread.

So when you're buying and selling stocks,

you incur these two types of transactions cost.

The first what I just mentioned are the commissions and

fees that you pay to the brokerage for trading.

Now, these fees have gone down over time.

You maybe hear about $7 portrayed, but they're still there when you buy and

sell stocks and then another transaction's cost that you basically pay

is what's known as the bid-ask spread.

So you see a stock price, but when you're buying the stock, you pay this ask price,

which is a little higher than the current market price.

And when you sell a stock, you pay this bid price,

which is a little lower than the current market price.

So if you're doing a round trip transaction, you buy it at the ask,

you sell at the bid, you're losing a little bit of return.

That's a little bid transactions cost that you suffer.

3:46

So, who does better when we're looking at individuals that trade stocks?

Is it the active traders here?

And you love this kind of circa 1963 photo here or

is it the kind of laid back, buy and hold investor?

Who does better?

Well, to classify individual investors in their training activity,

we're going to use this thing called portfolio turnover.

It measures how much an investor trades, you can calculate both buy turnover,

what's your buying activity and sell turnover.

What's your selling activity?

So, how do you calculate this?

By turnovers, it's just simply the dollar value of buys divided by the beginning

period dollar value of the portfolio and sell turnovers defined analogous way.

So typically, portfolio turnover is defined as simply the average of the buy

turnover and the sell turnover.

So you can think, if your portfolio turnover is zero,

it means you're a buy and hold investor.

5:57

So let's look at this kind of famous Barber and Odean 2000 trading

is hazardous to your wealth paper where they look at how does the performance of

individual investors that are trading stocks vary by how often they trade,

and it's first interesting just to look at this kind of figure here where

they actually break out Individuals by trading activity.

So, let's look here at group one.

Group one are people that have the lowest turnover.

You can see the white bar here is giving the gross return.

So, what's the return that these households earn before any

transactions costs are accounted for?

Group five on the other hand are those that have the highest turnover.

So if you look, you see these white bars all basically have the same height.

Those that trade the lowest.

Group one, the low turnover.

Those that trade the most, group five,

they all have basically the same return before trading costs.

So right off the bat, this is a bad sign for those who trade a lot,

because those who trade a lot are going to be paying more commission,

are going to be suffering more from the bid-ask, bid-ask spread.

So it's also interesting to look at the portfolio turnover,

that's the grey bar here.

So this is representing on a monthly basis,

how much of your portfolio are you turning over.

For the low turnover people, this is basically zero.

For the high turnover people, this is on the order of 21% per month.

So that would mean you change about 20% of your stock portfolio every month or

every four to five months, you have a new portfolio of stocks.

So you see this dramatic increase,

those that trade a lot have a very high portfolio turnover.

Those who trade a little bit have a very, very small,

so then this manifests itself in lower net returns.

So the white bars here were the gross returns before transactions cost over this

1991, 96 period, this averaged about 18 or 19% for all the households.

But once you look at the black bars, the net returns,

then you see, people who trade more, they suffer, right?

Their performance here, their annual returns may be 12% compared

to about 18 or 19 for those who trade a little bit.

Thus the title, Trading Is Hazardous To Your Wealth, which really summarizes well,

the data, the analysis but maybe isn't the preferred title of the anonymous

discount brokerage house that gave Barber and Odean this data, okay?

So let's do, it's always nice to have a good figure to summarize the performance,

but let's get into some regression analysis here.

So, you can think of classifying households or

doing the analysis in two ways.

In panel A, we're grouping households, and

we're looking at the return before trading cost, so gross return on a monthly basis.

And we're looking at households in aggregate.

So on panel A, Bill Gates has a much higher weight than I do,

because he has a much larger portfolio than I do, assuming both Bill Gates and

I are in the sample.

In panel B, we're looking at, again, returns before trading cost, and

we're just taking the average across households.

So, simple average.

So in panel B, Bill Gates and I would have the same weight.

I kind of like Bill Gates and I having the same weight.

So, let's focus on the panel B analyses.

But you'll see generally you'll have a consistent pattern across both panel a and

panel B.

So one thing that Barber and Odean do is they look at how do households

perform in their stock portfolios relative to this own benchmark?

So the idea here is, look and

see what households have as their portfolio at the beginning of the year.

See how that portfolio would of performed if it was a buy and hold situation?

Then use that as a benchmark and see how the household actually did.

You see here the household under performs

the simple beginning of the year buy-in hold strategy.

This is in percentage points, so this is five basis point, .05 percentage points,

or five basis points under performance per month, or about .6% on an annual basis.

So, by buying and selling stocks,

households underperform if they had simply bought, if they had simply held what

they had at the beginning of the year by five basis points per month.

10:41

Now let's look at a simple CAPM analysis, so

let's just look at the average household.

See what their returns are a monthly basis,

put that in a capital asset pricing model regression, and

you can see here what we get for

the alpha, minus 0.014 percentage points per month, and then this beta of 1.087.

The numbers in parentheses are P values here, where for

the beta this is a test, is this 1.087 statistically different from one or not?

So here you see the alpha, here you see the beta, so let's ask the question, okay?

At this point, we're very familiar about the CAPM, so

what do these CAPM analysis, you know kind of imply about households?

So, a few questions here.

How does the average stock

portfolio perform relative to its CAPM benchmark?

And we're looking at returns before transaction costs, gross returns.

How does the average household do?

Okay, is your answer to this first question consistent with the notion

of efficient markets, okay?

And does the average household have a strong tilt towards high-beta or

low-beta stocks?

Okay, so let's think about this.

Now, look at that CAPM regression results to guide you when answering these three

questions.

12:08

Okay, so three questions here to look at the performance of

the average individual investors on the CAPM.

Let's go back to the table.

So, when you look at average investor,

this is a very small amount that isn't statistically different from zero.

You see this P value is extremely high.

It's basically the alpha 0, and the betas essentially won.

You can see this P value indicates it's not statistically different from one.

So, when we average across households, they have a beta,

their stock portfolio one.

It's kind of they're investing in the market on average, and

their alpha is zero.

So, this is consistent with efficient markets, right?

If people are randomly throwing darts, we'd expect to get a very similar

result when we did this CAPM regression, that we get a beta of about one.

We get an alpha of zero.

So, if you think markets are efficient,

you should be able to predict stocks are going to go up on a risk adjusted basis.

This is exactly what you would find, and that's what we find for

the typical household.

And we see when we look at average across households,

they don't have a tilt toward high beta stocks or low beta stocks, right?

This beta's almost exactly, exactly one.

13:22

Okay, so now we can also now subject the look of the performance of these

individual stock portfolio in this data set, or the period 1991 to 1996

in a Fama-French three-factor model setting where then we're going to see

are there any investment tilts on the size dimension or the value dimension?

And once we control for

those portfolio tilt, do we see any out performance or underperformance?

So, let's look at this three-factor model.

Result, here's the alpha with the P value of that estimate in parentheses.

Here's the beta, the sensitivity to the market

with the P value indicating is this statistically different from one or not?

Then we have, in this column here,

is we have the coefficient on the value factor, value minus growth.

And then the coefficient here, finally, on the size factor, small minus big.

So, let's look at this three-factor model regression result, and

then think what that means about the typical average household stock investor.

So another pause, think, and answer.

Le Penseur, kind of a good intuition be that this question's going to be

looking at the three-factor model regression results.

So, does the average household have a tilt toward value stocks, or growth stocks?

Does the average household have a tilt towards small stocks or large stocks?

And then finally, does the average household, how does it perform relative to

its 3-Factor model benchmark in gross returns before transactions cost?

Controlling for the tilt of the portfolio towards small,

toward value, toward market risk.

Does the average household beat its benchmark, or

does it underperform its benchmark in the 3-Factor model?

15:13

Okay, so to answer these questions,

why don't we go back to the regression results?

We're looking here at the three-factor model, so you can see there's a positive

coefficient here on the value factor, so there's a slight tilt toward value stocks.

You can see there's a pretty big coefficient here on the size factor, so

individuals have a pretty big tilt in their portfolio to smaller stocks,

as opposed to large stocks.

So, that kind of makes sense, if you think of institutional investors,

they're generally investing in large stocks.

So, another way of saying that is smaller investors are relatively

investing more in small stocks, and that's exactly what we see here.

When we look at this alpha here, we see it's minus 0.154%, on a monthly basis.

So, on an annual basis,

this would be underperforming the benchmark, by about two percentage points.

But this high P value here suggests it's not.

Statistically different from zero.

So again it's not really strong evidence that individuals trading their stock

portfolio are under performing given the risk of their portfolio but

they're certainly not evidence aggregate their out performing either.

Which again, would be consistent with efficient markets.

If you're randomly throwing darts,

if there really isn't good information on average behind the stock picks,

you'd expect you don't outperform or under perform on average.

So let's look at some of the analysis here.

And again, we're focusing on that returns on a monthly basis here and

we have various you know, coded benchmark that we're looking at.

The first one, was we just compared how a household actually is doing

during the course of the year, relative to how they'd be doing if they simply.

Held the stocks they had at the beginning of the year.

So it's like household actually performance relative to if they simply did

a buy and hold strategy with the stock they held at the beginning of the year.

When we were looking at gross returns this difference wasn't this much,

it was only like five basis points or 0.6% on an annual basis.

Once you take into account the trade-in costs,

households are under performing their own kind of simple buy and

hold, what you had at the beginning of the year benchmark

by about 20 bases point per month or about two and half percentage points per year.

This reflecting the costs of trades and theses trades on average eon't seem to

be based on good information, because they don't out-perform the benchmark.

Then if we look at the Fama-French three-factor benchmark here,

we see this negative alpha of about 0.31 basis points per month, or

about 3.5% on an annual basis are households under performing a benchmark.

And remember, this benchmark has implicit assumed in it a 0% expense ratio.

So given all the trades, they're underperforming that by about three and

a half percentage points.

So let's think about breaking households like we did in that earlier

figure into five categories here based on how much they trade.

Okay, so you see group one.

18:24

They're monthly portfolio turnover is basically zero.

Group five, it's 21% on a monthly basis.

So that means every four to five months they

totally have a new stock portfolio because of all their buying and selling activity.

So let's break households into these five categories and

see how the performance varies by those who trade a lot.

Group 5 versus those who trade a little bit, group 1, okay?

And one of the things they'll do is they'll do a three factor model analysis

on each of these households.

One thing that kind of is striking when you do this analysis

is those who trade a lot have a much higher loading, or

a much bigger coefficient on this SMB factor.

So remember that means, this big positive coefficient, those who trade a lot,

seem to trade a lot in small stocks, relative to those who are buy and hold.

That the group five has a much bigger coefficient on this SMB factor,

than group one.

So that's something to take into account.

When we're taking into account trading costs,

because usually smaller stocks will have a higher bid ask spread, which is bad for

you if your buying and selling the stocks.

Smaller stocks have a higher bid ask spread typically than larger stocks.

So let's look at the just simple raw performance cross these groups

before trading costs.

This is what we showed in this earlier figure.

And you can see not controlling for any differences in risk that the returns

are about 1.5 percentage points per month across all five of these groups.

About 18 percentage points per year.

Very small difference and you can see,

this was exactly the figure I showed at the top here.

Where you see everyone, regardless of how much they trade,

has about an 18 percentage point annual return.

On the far right, we show over this period, 1991 to 1996,

was the average return of the S&P 500.

It was very similar, you can see this white bar but

the S&P 500 has essentially no turnover.

So you can see the growth's return, the white bar, and the black bar,

the net return, are basically the same.

So you can think of kind of like a rough benchmark for these households given their

average beta is one, is how are they performing relative to the S&P 500?

On a gross return basis are performing the same, but

obviously once you take into account the trading costs, and you look at the black

bars, the S&P 500 is beating the performance of those who trade a lot.

21:22

No one is having a positive alpha here.

You can see, if anything, all these alphas are slightly negative.

But none of them are statistically significant.

All these P values are greater than 10%, so not surprising.

The average household doesn't beat the market so to speak when you break

households into those who trade a little bit versus those that trade a lot.

You also don't see a positive alpha.

The alpha seems to be slightly negative but not statistically different from zero.

But again we're looking at the gross returns and

this is a result that's very much consistence with market efficiency.

We might find it strange if that on average households were beating the market

by a lot because that would suggests, well, institutional investors

must be kind of losing against the market a lot in their trades.

And if we think in efficient markets, you should be able to

predict the future performance based on the current information.

22:19

Okay, you do see there does seem to be on the margin, those that trade a lot.

Actually seem to under perform those that trade a little bit.

In this three factor model the difference in the alpha is about 30 basis points

per month, or 3.5% per year, this might reflect that those who trade a lot

have the same return as those that trade a little bit,

in terms of raw returns, but remember those that trade a lot specialize or

focus on these small stocks, that have higher returns on average.

So once you kind of increase their benchmark because they're

focusing on small stocks, that makes their kind of risk-adjusted performance

seem worse, okay.

So let's now look at the net performance given the turnover here.

So now we're taking into account trading costs when we're looking at the risk

adjusted performance.

Not surprisingly, when we look at those that just trade a little bit, the gross

returns and the net returns are basically the same because they're basically buy and

hold investors.

They're hardly buying or selling.

So there aren't any transactions costs.

So when you look at all these risk adjusted returns whether it's relative to

the household being a just holding what they had at the beginning of the year,

whether it's a CAPM model alpha or a three factor model alpha.

All these are zero okay so they are not beating their benchmark but

they aren't really under performing their benchmark either.

Now when you look at group 5 those that trade the most

not surprisingly you see big negative alphas here.

So under performing the benchmark, under performing what returns they would have

earned if they had simply held what they had had in the beginning of the year and

did a buy and hold strategy with their initial investments,

that by actually trading, they under performed this hypothetical buy and

hold strategy by 60 basis points per month, almost 7% on an annual basis.

If we look at the CAPM model, they're alpha.

Take into account tradings cost is about 70 basis points per month or

eight percentage points are sold on an annual basis.

And relative to the three factor model,

these guys are underperforming by 80 to 90 basis points per month.

Or almost ten percentage points on an annual basis.

So the average household isn't beating the market.

Okay, by trading a lot they're making a brokerage house rich, but

they're having a substantial deterioration in their own performance,

so this may, you know, kind of give you, might be a hobby,

you may get joy from the buying and selling of stocks, but

that does lead to a pretty substantial deterioration of performance

here that Barbara Nodine will find in this 91 to 96 data set.

Therefore the title of paper, Trading is Hazardous to Your Wealth.