[MUSIC] All right, let's start our discussion of the effects of loss aversion with looking at how loss aversion affects the decision to sell a stock. And how kind of there's a tug of war, like loss aversion's saying realize gains, hold onto losses. But taxes is saying the opposite, so which of these two kind of wins out and do we see any evidence of both may have some impact on how individuals are trade, okay? So right off the bat, let's start with a question, it's always kind of a good sign when you see. Get immediate, it's time to actually think about the material and all these stuffs very relevant. Maybe it'll get you thinking about your own financial decisions. So, let's think about loss aversion and stock trading, okay. So first, what's a natural benchmark for an investor in a stock or any asset? They don't have to say an investor, yourself, okay? If you buy a stock or an asset, what's a natural benchmark for you going forward, okay. Given this benchmark, how should loss aversion affect the types of stocks that are sold? So, what benchmark do you have in mind after you buy a stock? Given this benchmark you have, how should loss aversion affect the type of stocks that are sold? What's a natural benchmark for an investor in a stock? Given this benchmark, how should loss aversion affect the type of stocks that are sold? Well, Terry Odean did some work on this and found that individual investors are more likely to sell stock investments that have gone up than those that have gone down. So this is consistent certainly with the predictions from a loss aversion. So let's think. Based on this result, that people seem to kind of sell stocks that have gone up on average, hold those that have gone down on average. Let's think about what this means. Okay, will this loss aversion induced trading be good or bad for investors? Okay, why do you think so? All right so, what did you think, is this loss aversion trading going to be good for investors, bad for investors, I guess I didn't give it as an option but you could say well doesn't matter. One way or the other, and why? Okay well, really, there's a couple things to think about, one this loss aversion, this trading is actually trading against short term momentum effects. So they talked about in the US, historically stocks have gone up the past year. Continue to drift up, stocks have gone down the past year, continue to drift down on average. Loss averse trading is having you sell past winners, hold on to past losers. So you're actually trading against this profitable momentum strategy, so that's probably on average not a good thing. And then also, tax costs could be important here as well. Not an issue in a retirement plan, where you're not paying capital gains taxes on your assets you sell, but potentially very important if you're trading in a taxable account, why? Taxes would say hold on to gains, you don't have to pay the tax when you realize the gain. Sell losers right away, so you can get the tax deduction right away. Loss aversion is kind of forcing you, or causing you, to do exactly the opposite. So this loss aversion, induce trading, could potentially be bad for your financial well being because of these two reasons. So, there's different motivations for the sale of stock, Odean finds that loss aversion appears to be a key determinant of stock sales but doesn't mean it's the only motivation. Do taxes also play some role in the sale decision of individual investors? And it doesn't have to be only one thing matters, it could be the case that loss aversion matters some and taxes matters some as well, okay. One way to actually test if taxes matter is to look at a household and look at the trades of households that have both taxable accounts as well as individual retirement accounts. So, in your taxable accounts if you sell a stock with a gain, you'll be liable to pay tax on that gain. You can get the benefit from selling a stock with a loss, so you can get the benefit of the tax deduction. While an individual retirement accounts, these are kind of saving for retirement in a tax-deferred manner. You don't have to pay capital gains taxes following sales. Or you don't get deductions following the sale of losers, you don't have to pay capital gains taxes following the sale of winners. So this actually provides a nice laboratory to compare how people trade in their taxable account. How people trade in their individual retirement account. Both accounts should be subject to loss aversion if that's present, but only tax motivated behavior would affect stocks or trading in your taxable account. So, you would expect the net loss aversion effect to be smaller in taxable accounts. So loss aversion say, sell your winning investments, hold onto your losers. Taxes is mitigating against that, because that's putting pressure on you not to sell your winner. So the loss aversion effect should be mitigated or lower in taxable accounts than in the tax-deferred or individual retirement account setting, okay. It sounds like I thought about this a lot. Well, actually I have because I wrote a paper on this with my coauthors Zoran Ivkovic and Jim Poterba where we actually test for evidence of tax-motivated trading by individual investors by comparing the trades of a given individuals that hold both taxable and tax-deferred accounts. So we can look at the same individual and say, hey, are you trading differently in response to, are you trading stocks that have gone up or gone down differently depending on whether they're in your taxable account or in your tax-deferred individual retirement account, okay. We do this using the same data on the thousands of individual investors from an anonymous discount brokerage that was used by Barber and Odean. This is really the gold standard, as you can probably tell by now, for studding individual investor behavior here. The key thing is are there differences in the sensitivity of stock sales to past performance since purchase across stocks in taxable and tax-deferred account. Loss aversion should affect your stock trade in both the taxable and the tax-deferred account or IRA account. But tax effect should only affect your stocks and the taxable account. That's the key to how we kind of do the study, and that's key to the setup, okay. So let's get into the specifics of this design here. So we're going to simply do some simple regressions. We're looking at stock purchases of at least $10,000, so we're looking at kind of bigger investors here. And we're going to look at simply what's the likelihood you sell a stock related to the amount of the gain of that stock or the amount of the loss of that stock. So if the stock has gone up 20% gain will be 0.2, if the stock has fallen 20% loss will be -0.2, okay. We'll report the coefficients from that regression, and then we'll report the standard errors of those estimates in parenthesis. If you see stars behind the coefficient, as usual that means view this as statistically significant result. Okay, so let's look here at a lot of kind of regression results here. Let's just focus on a few of them to give you the picture. So first, we're looking at, what's the likelihood you sell a stock in the second month since purchase and we're relating it to the gain or loss you had over that first month. So what are these regression coefficients telling us? First, the constants from the regression of 11%. So that means if there's zero return there's an 11% chance you sell the stock holding in month two. The coefficient a gain of 21, well if the gain happens to be at 10% return, 0.1, 0.1 times 20 is 2, so that adds 2 percentage points onto the likelihood of sale. If it's a 10% gain, your likelihood of selling it in month 2 is 11 plus 2, 13 percentage points, okay. How about instead of a 10% gain it was a 10% loss? Well now loss takes on the value of minus 0.1, minus 0.1 times 10 is minus 1, so the likelihood of selling a stock that is loss value in the first month and month 2, is 11 minus 1, 10 percentage points, okay. And this is for stocks in the taxable account, okay? So you can see here that hey, even though it's a taxable account you're more likely to sell stocks that have gone up in value than stocks that have gone down. Remember, loss takes on negative value, so a positive coefficient on the loss variable means you're less likely to sell, losing investments a positive coefficient on the gain variable means you're more likely to sell winning investments, okay. What else do you notice here? This baseline of sale, this what's the likelihood of selling the stock in month 3, or month 4, or month 5, all the way down to month 12. You can see that this is basically monotonically or almost monotonically declining. So in month 2, this likelihood of sale is 11 percentage points. In month 12, the likelihood you sell month 12 is about 3 percentage points. So the longer you hold the stock the longer you're likely to hold it going forward. This likelihood of sale declines month after month, after month. A lot of stock sales happen one, two, three, four months after purchase, okay. Also you can look at these coefficients on gain, you can see they're generally positive, particularly over the first six months. So this positive coefficient on gain, the stars are indicating statistical significance means over these horizons, you're more likely to sell stocks that have gone up in value. The higher is the gain, the higher is the likelihood of sale. Okay, and then in the third column, the coefficients on loss are also positive, and a lot of them are statistically significant. Remember, the loss takes on negative value like a minus 10% return would be a loss of minus 0.1. So that means the more losses you have, the less likely you are to sell that stock in any given months, okay. So that's fine, what's key is to do our comparison of how do people trade stocks in taxable accounts relative to how they trade stocks in individual retirement accounts. So for that let's pick a particular month, let's look at what's the likelihood you sell the stock in the seventh month after purchase. Relate that to what were the returns the gain and loss over the prior six months, okay. So look at this regression, you see if there's no return if the stock is trading at the price you bought it at, there's a 4.6% likelihood you sell the stock in month seven. This gain coefficient of 1.3 if it's a 10% gain, 0.1 times 1.3 means here 0.13 percentage points more likely to sell a stock that's gone up 10%. The loss coefficient is 2.7, if the loss is 10% minus 0.1 times 2.7 being you're minus 0.27 percentage points less likely to sell a stock that has gone down in value. So again, you have this relationship more likely to sell stocks that have gone up, less likely to sell stocks that have gone down. That seems like that's consistent with loss aversion, right? And we're focusing on stocks that are in your taxable account, so it seems like this is evidence against tax-motivated trading. But key is to compare how do people trade stocks in their taxable account with those in their individual retirement account. So let's look at these coefficients of 1.3 for gain, 2.7 for loss, and let's compare them to the coefficients here on the right side of the table. And these coefficients here are telling us, what's the difference? Okay, what's the probability of selling stock in your taxable account relative to a near tax-deferred account? So what's this coefficient on gain being negative? It means that gains have a smaller effect on your sale and taxable accounts than they do on tax-deferred accounts. Or another way of saying this is the relationship between selling winning stocks is much stronger, is more positive in tax-deferred accounts, is relatively less in taxable accounts. So for example, instead of the coefficient being 1.3 like it is in the taxable account sale regression, it would be 1.3 plus 4.5, 5.8 in the individual retirement account regression. So it's saying this likelihood to sell stocks that have increased in value, it's a much stronger effect in the individual retirement account than it is in the taxable account, okay. So, the way to say this, this loss aversion effect, stronger in the individual retirement accounts is 4.5 coefficient lower In the taxable account which would be consistent with loss aversion affects people's trades. But when you have a taxable transactions, you have a less loss aversion effect than if you don't have this taxable account setting. For losses it's the same, so you can see this -2.8, it means the coefficient are loss in the taxable account is -2.8 smaller in magnitude than it is for the tax-deferred account. Okay, so again this relationship you're more likely to sell gains than losses. It's much stronger in the tax-deferred setting where there isn't a tax consequence. It's much smaller in the taxable accounts setting which suggests that taxes matter and influence how people trade. If we look at these difference, this is taxable coefficient minus tax-deferred. If we just simply had a third group of columns here which showed what's the effect in the tax-deferred setting, the coefficient on gain would be 1.3 plus 4.5, would be 5.8. The coefficient on loss would be 2.7 plus 2.8, 5.5. So you'd see in the tax-deferred setting people are much more likely to sell gains, they're much less likely to sell losses than they are on taxable accounts, which means that taxes must be muting this loss aversion affect. So both matter, both taxes and loss aversion. Another way to kind of simply show this same information is in two figures. So were looking here at the light added likelihood of sale for a $10,000 purchase in a taxable account. So here were looking at, you have a gain of 25%. The black bar or the gray bar is a loss of 25%. How does that move up or down the likelihood of sale? So if you have a gain of 25% entering the month going from month 2 to 3 to 4, all to 7 and 8, it increases the likelihood you're going to sell the stock. And if you have a loss entering the month, the gray bars are negative. It means a reduced likelihood of selling the stock, okay. So this would seem like, this is evidence the against taxes affecting people's financial choices, because taxes say you hold onto gains, you sell losses, but this regression result gives the exact opposite pattern, okay. But what we want to do is, we want to say well let's do a comparison. Let's look at how people trade in their taxable accounts relative to how they trade in their tax-deferred accounts, and that's showed in this next picture. So this next picture is saying how does a 25% gain influence your likelihood of selling the investment in a taxable account relative to what it is in a tax-deferred account. Like an individual retirement account, the black bar here is showing for a 25% gain, and this is showing that the 25% gain is less likely to be sold in the taxable account than it is in the tax-deferred account. The black bars are now negative. The gray bars on the other hand are positive. So it means for a 25% loss you're more likely to sell that losing investment in a taxable account than a tax-deferred account. So loss aversion certainly matters, but there's also evidence that taxes influence people's choices as well, by doing this simple comparison, how does a given person trade stocks in their taxable account versus their individual retirement account. So the bottom line is, there's multiple motivations for why people sell stock. So certainly loss aversion is a big factor, okay. No doubt about that, but there's also evidence for a tax lock-in effect. And this evidence is uncovered by comparing the trades of an individual in both their taxable account and their tax-deferred account. Loss aversion should affect trades in both accounts. But by looking at differences in how people trade across them, you can identify tax lock-in effect. So bottom line, loss aversion, the psychology influences people's financial decisions, but some evidence that Uncle Sam and taxes matter as well. [SOUND]