Session 25: Divided Policy - The Trade Off

In this session 25 of the 36 session corporate finance class, I'd like to talk about the trade-offs on dividend policy. I'd like to start off with some really bad reasons why companies pay dividends, as well as some good reasons why companies may start paying dividends and then increase them. In the last session, we introduced basic facts about dividends, that they were sticking, that companies increasingly are turning to stock buybacks, and they're affected by tax loss. In this session, I'd like to talk about the trade-off that animates the dividend decision. And in particular, we're going to talk about three different schools of thought in dividends that kind of span the spectrum. So here are the three different schools of thought in dividends. The first school of thought, here's what it says. If investors are treated exactly the same way on dividends and price appreciation, treated by the tax guy, essentially the same way, they're taxed exactly the same. And companies can go out and raise new equity if they need to, with no transaction, no flotation, no issuance cost. Then it does not matter how much you pay in dividends. Well, guess who came up with this school of thought? It's a Milamudugliani argument that dividends do not matter. In a world where taxes treat dividends and price appreciation the same, and equity issues basically cost companies nothing, it doesn't matter how much you pay in dividends. The second school of thought is built on what we talked about the tax law looked like pre-2003. From 1913 to 2003, at least in the US, dividends were taxed more heavily than capital gains. So this school of thought said if dividends are being taxed more heavily than capital gains, dividends are bad, and companies should not pay them. So dividends don't matter, that's a Milamudugliani school of thought. Dividends are bad, why? Because they get taxed more highly. And the third school of thought rounds out the three different views on dividends. If your stockholders like dividends, and that's what the third school of thought is based on, for whatever reason, rational or irrational, then it says, look, pay them more in dividends, because dividends are good. So dividends can be good, they can be bad, or they should not matter. Now that's, of course, difficult to reconcile, because they're pushing in three different directions. It's, of course, very convenient for a CFO, because no matter what you do, you've got an entire school of thought backing you up. So if you increase dividends, you can point to the dividends of good school. You cut dividends, you point to the dividends of bad school. If you have no idea what you're doing on dividends, you say you belong to the Milamudugliani school. But all three schools, though, have an element of truth to them. And here's how I'd like to take those three schools of thought apart. Let's start with the Milamudugliani school. Here's the intuition that drives why dividends don't matter. Let's say you're a company and you pay out too much in dividends. In other words, you have only $100 coming and you choose to pay out $200. You have a cash deficit, right? But in the Milamudugliani school, that's not a problem, because here's what you do. You go out and you raise $100 of new equity to cover the extra dividends you pay. What this will mean, of course, is there are more shares outstanding. Your stockholders will get less price appreciation, but they'll be exactly compensated for that less price appreciation for higher dividends. So in the Milamudugliani argument, what happens is the total return to stockholders remains the same, no matter what you pay in dividends. If you pay no dividends, all of the returns come in the form of price appreciation. If you pay very large dividends, most of it might take the form of a dividend yield and less from price appreciation, but the total returns remain the same. For this argument to work, though, the projects that companies take has to remain invariant, no matter what their dividend policy. If companies start rejecting investments because they've paid out too much in dividends, then this school of thought will break down. So in a world where a company's investment policy is set based on whether the projects are good or bad, then paying out too much in dividends or too little in dividends will have no consequences, because investors will get the same total return, no matter what you do as a company. So that's my starting point. It is possible that in some companies, dividends either don't matter or matter very little. But for that to be true, the tax law has to cooperate, and you have to be in a market where raising new equity costs you almost nothing in terms of transactions costs. Now let's talk about the dividends are bad school. This graph, I actually have the tax rates on dividends and capital gains going back over US history. And remember the story I said for much of US corporate history. The tax rate on dividends has been much higher than the tax rate on capital gains. In fact, there are periods in the 1950s the tax rate on dividends was as high as 80% to 90%, and the tax rate on capital gains was only 30%. That's a pretty big difference in tax rates. So if you as an investor received dividends, a big chunk of the dividend would have gone out as taxes. You receive the same return as price appreciation, then in a sense, you are paying very little in taxes. Of course, as a company, you might ask, how do I know whether my investors care about dividends or worry about taxes on dividends and capital gains? What are they paying as taxes? There might be a way to find out by looking at what happens on the ex-dividend day for your company. What's the ex-dividend day? That's the day by which you have to buy a stock to collect the dividend. So here's one very simple way to visualize the ex-dividend day. Let's say we're all stockholders in a company, and let's say we all bought stock five years ago, and that the price at which we bought the stock was given. So we all bought at the same price five years ago. Tomorrow is going to be an ex-dividend day. So if you hold the stock through the ex-dividend day, you get the dividend. If you sell it, then you lose the dividend. So here's the choice you face. You can sell before the ex-dividend day. And let's say if you sell before the ex-dividend day, the price you get is PB. And if you do that, of course, you lose the dividend. Or you can wait through the ex-dividend day, get the dividend, and sell it at the end of the ex-dividend day, and let's call that PA. As an investor then, here's what your cash flows would look like from selling before the ex-dividend day versus after the ex-dividend day. If you sell before the ex-dividend day, you're going to collect the PB, which is the price before the ex-dividend day. You're going to have to pay capital gains taxes, because remember, you bought the stock way back in time. And you're going to pay the capital gains tax rate on the capital gain. There are no dividends because you sold before the ex-dividend day. So that is your cash flow from selling before the ex-dividend day. Now, let's contrast that with the cash flow you'd get if you waited that extra day. The price you'd get would be PA. You'd still have to pay capital gains taxes on the difference between that price and the price you originally paid. But now you'd also get a dividend on which you pay an ordinary tax rate, or the dividend tax rate. Now, to keep this world stable, what's got to be true for most investors is that that cash flow from selling before has to be roughly equal to the cash flow from selling after. Why? Well, if that were not true, you'd get a wave of selling either before, if the cash flow from selling before were higher, or a wave of selling from selling after. So to keep this market at a stable state, the cash flows from selling before and the cash flow from selling after have to equate. That's what I've done in this. And if you work with that equation, it turns out that you end up with a very simple measure of the differential tax rate. You look at the change in price on the ex-dividend day, as a percentage of the dividend. That should give you a rough sense of what investors in your company think about the ordinary tax rate, or the dividend tax rate, versus the capital gains tax rate. Or put more generally, if investors face exactly the same taxes on dividends and capital gains, then you should expect to see the price change on the ex-dividend day be roughly equal to the dividend. If investors pay a much higher tax on dividends and on capital gains, you should expect to see the price change on the ex-dividend day be far smaller than the dividend. If, in contrast, they pay a much higher tax rate on capital gains than they do on dividends, then you should see the price change on the ex-dividend day actually exceed the amount of the dividend. So this is the opening we're going to use. We're essentially going to look at the change on the ex-dividend day. So the price change. We're going to relate that price change to the dividend, and we're going to look at that ratio. If that ratio is well below 1, the price change is well below the dividend, we're going to argue that investors in your company face a tax disadvantage in dividends. If the price change is roughly equal to the dividend, we're going to argue that investors in your company pay roughly the same taxes on dividends and capital gains. If the price change is much greater than the dividend, then in a sense we're going to argue that in your company at least, your investors actually pay less in taxes on dividends than they do on capital gains. To do this, though, you want to focus not on one ex-dividend day, because strange things can happen on any given day, but across a bunch of ex-dividend days. 16, 20, 24 different ex-dividend days. In fact, if you can do it across lots of stocks, even better, because you get the law of large numbers working in your favor. And there are studies that have actually done this. In fact, on this page, I actually summarized findings that studies have brought to the table on what the change on the ex-dividend day is on the price relative to the dividend. The earliest of these studies happened in the 1960s, and what it found, on average, was that the price change was roughly 78% of the dividend. And that shouldn't surprise you. In the 1960s, the tax rate on dividends was much higher than the tax rate on capital gains. If you look at the subsequent two time periods, which are later in time, you notice that the price change has become a smaller percentage of the dividend. One reason for that is the tax disadvantage of dividends has shrunk, especially in the 1980s, and the kinds of investors entering the market, institutional investors, might not bear as much a burden of paying higher taxes on dividends than on capital gains. In effect, if you look across time, the price change as a percent of dividends seems to at least reflect an overall trend of a smaller tax disadvantage to dividends as opposed to capital gains. I haven't seen one of these studies post-2003. It'd be interesting to see, now that the tax rates are the same, whether the price change has roughly turned out to become equal to the dividend. I don't think it will have, and here's the reason. Even though the tax rates remain the same, or are the same for dividends and capital gains now, dividends, I think, still have a tax disadvantage relative to capital gains, and it's a timing disadvantage. You have to pay taxes when you receive dividends. You don't have a choice. You decide when to pay taxes on price appreciation because you don't have to pay taxes till you sell the stock. That timing option might still give a tax advantage to capital gains as opposed to dividends, and that might still tilt the scales, but only the facts will. We have to look at the data to see if that, in fact, is true. So that's the second school of thought. It says dividends are bad because of a tax disadvantage, which brings us to the third school of thought about dividends being good, and I'm going to dispense with two arguments I've heard for dividends that are not very good arguments. The first is what's called the bird-in-the-hand fallacy, and this is the argument that goes with this fallacy. It says dividends are more certain than capital gains, and the argument at the outset, at least, looks reasonable, and here's why. If you receive a dividend, you get the check in the mail today, right? Whereas if you have to wait for price appreciation, who knows what that price appreciation will be two or five years out. So if you compare dividends today to price appreciation in the future, it's true. Dividends today are more certain than price appreciation in the future, but remember what we said happens in the ex-dividend day? The price drops by roughly the amount of the dividend or close to it. So if you're an investor, you actually have to pick between dividends today and price change today. There's not that much of a difference. So if you're paying dividends expecting your stockholders to reward you because they're more certain than capital gains, you might be doing the wrong thing. So if that's your reason for paying dividends, don't do it. It's not a good reason. Here's a second argument for paying dividends. We had a good year. Our earnings were high this year. Why don't we pay it out as dividends? Not a bad argument if your earnings are high for the long term, but if you're a company whose earnings are volatile, remember, a good year can be followed by a bad year, and if you expect to go back to the market in future years, either because you expect bad years or because you might need new financing in the future, then my suggestion to you is don't pay the dividends today because if you pay dividends today and have to issue stock in the future, it's expensive to raise new equity, and that expense is going to overwhelm whatever benefits you get from paying dividends. The issuance cost of equity is large enough that you don't want to go down this road. So if you're thinking about paying dividends, at least have a good reason for paying dividends, and here are three potentially good reasons, though one of them is at the margin. The first is called a clientele effect, and what it effectively means is if your stockholders, for whatever reason, are stockholders who like dividends, then paying out more in dividends will help you as a company, will push up your stock price. The second is called a signaling story. If you're a company with great investment prospects, one way you can signal that to the market is by increasing your dividends, and we'll talk about why this story holds up and when it might fall apart. And the third reason, and this is the one at the margin, is if you've borrowed money, it might turn out that by paying dividends, you're able to transfer wealth from the lenders to the equity investors. A marginal benefit, but a benefit that you might want to take advantage of in some companies. So let's take each of these arguments for dividends and let's see how well they hold up. Let's start with the clientele effect. The argument here is over time, you start to collect investors who like your dividend policy. What I mean by that is if you're a company that does not pay dividends, over time you accumulate investors who like the fact that you don't like to pay dividends. If you're a company that pays out big dividends, you accumulate investors who like big dividends. And because of that, over time, you end up with an investment clientele that likes your dividend policy. And in fact, there's some evidence to back it up. If you look at companies that pay much more in dividends, you tend to find that investors in those companies tend to be poorer and they tend to need dividends. They tend to be older and poorer investors who value those cash dividends more. So if you look at dividend policy across companies, you tend to find that that dividend policy matches up to the clientele that those companies have accumulated. So that's the first argument for dividends is if you have a clientele that likes dividends, you might want to continue with that dividend policy because changing that policy is going to upset them. Let's look at the second argument for dividends. There's an argument that dividends can be used to signal the future. What I mean by this is if you're a company with great projects, it's actually very difficult to convince markets that you have great projects. You can, of course, call a press conference and say you have great projects, but nobody believes you. One way you can back up that contention that you have great projects is by going on and increasing dividends. You're saying, why would that convince markets? Remember that dividends are sticky? So by increasing dividends, think of the message you're sending to markets. You're saying, look, we think our projects are so good that we can sustain these dividends. So the argument goes, and this is the signaling story, that only companies with better investment projects will go out and increase dividends. And when you see an increase in dividends, then you should expect to see stock prices go up. There is some overall evidence that backs this up where you see stock prices do go up on average when companies increase dividends and go down on average when they decrease dividends. There is a strange asymmetry, though. An increase in dividends sees your stock price go up only 1%, but a cut in dividends, you might see a stock price drop by 4%, 5%. That, of course, reflects the fact that dividend decreases are much less common than dividend increases. But overall, at least, there is some backing for the dividends are a good signal argument. But be careful because that dividend signaling argument can very quickly be turned against you. And this is especially true when companies initiate dividends. Think about it this way. You're an investor in a growth company. You're investing in the company because you expect it to have great growth prospects, right? Let's say this company announces it's going to start paying dividends. You should have mixed feelings. Mixed feelings, why? Well, the dividends are good. That means you're going to start getting cash back from the company. But the fact that they initiate dividends can be viewed as evidence that their growth prospects are starting to ease, that they're not as good as they used to be. In fact, there's some indirect evidence, at least, that this is true. This is a graph that looks at earnings growth around dividend initiations. Notice that the growth actually peaks in the quarter in which companies initiate dividends. And if you look at the quarters after, you see declining earnings growth. So it might turn out that when a company initiates dividends, it's actually signaling that future growth is going to be lower. So signaling stories are tricky. They can be made to cut either way. They can be good news or bad news depending on the company you're looking at. In fact, whatever signaling story you're telling, there seems to be some evidence that dividend signals are getting less effective. This is actually a graph that compares the effect of increases and decreases in dividends across time periods. And now, notice how in the most recent time periods, an increase in dividends has a much smaller impact on stock prices, and a decrease in dividends has a much smaller impact. Here's a story I tell to see why this might be happening. I think there are far better ways in which you can get information out to markets now than cutting or increasing dividends, and markets might be reflecting it. The richer your information is about a company, the less you need dividends to send news to the market. So maybe what you're seeing here is a reflection of the fact that investors have access to more information, and if they have access to more information, they trust dividend changes, or they need dividend changes less to make judgments about companies. Overall, though, the signaling stories sometimes used to increase dividends can sometimes be used to cut dividends, but it's a story that can cut both ways. Which brings me to my third and final argument for paying dividends. Remember this graph I showed you that on average when dividends or increased stock prices go up? Well, if you look at that same dividend increase data, what happens to bond prices, you see the opposite effect. In other words, what's good news for stockholders tends to be bad news for lenders and bondholders, and here's why. If you're a banker or a bondholder, you've lent money to a company that has a big cash balance. You'd much rather that the company keep the cash balance than return it to stockholders, right? Because that cash can be used to back up your interest payments and principal payments. So when a company pays out big dividends, it's hurting its bondholders. That sounds like a terrible thing to do, but if lenders and bondholders have built in the presumption that you will pay dividends, and you don't pay dividends, that's actually bad for your stockholders because you're denying them something that's due to them. So if you have lenders or bondholders who have built in the presumption that you're going to pay or increase dividends, it might actually make sense for you to give more cash back to your stockholders. So three good reasons for paying dividends. You have investors who like dividends. You might be using dividends as a way of signaling to markets that your prospects in the future are better than they thought they would be. Or you're using dividends as a way of getting back to some steady state between lenders and equity investors. Do managers think of all of these things when they change dividends? Not necessarily. This table actually is the result of a survey where managers were asked about what they consider to be the most important criteria on dividends. And some of the answers might or might not surprise you, but the thing to remember here is these are not necessarily the right factors, but this is what managers think about dividend policy. For instance, they do think that dividend payout ratios affect stock price. Whether that's true or not is irrelevant. That's what they think. And at the other extreme, they don't seem to care very much what their stockholder preferences are when they set dividends. So take a look at these answers, because in a sense, they cover the gamut of the trade-off. And some might be good reasons, some might be bad reasons, but dividend policy is often set by inertia. In other words, most companies pay out what they do in dividends because that's what they've always done. So changing dividends is actually the most difficult thing to do. And in future sessions, we'll talk about why companies might be forced to change dividends. But in this session, I just wanted to set up the trade-off so we have a good basis on which to build on when we talk about the right ways of thinking about dividends in future sessions. Thank you very much for listening.