Bad Dividend Math

By James Kwak

While working on a new Atlantic column, I came across this article by Donald Luskin (hat tip Felix Salmon/Ben Walsh) arguing that “Taxmageddon” (the expiration of the Bush tax cuts at the end of the year) will cause the stock market to fall by 30 percent.* His argument is basically this: if the marginal tax rate on dividends increases from 15 percent to 43.4 percent, the after-tax yield falls by 33.4 percent, so stock prices should fall by about the same amount.

Ordinarily I don’t bother with faulty claims like this—there are only so many hours in the day—but it bothered me so much it cost me some sleep last night.

The first problem is the only one that Luskin acknowledges: lots of investors don’t pay taxes on dividends. He mentions pension funds; there are also non-profits and anyone with a 401(k) or IRA. According to Luskin, only about one-quarter of dividends are received by people who will pay the top rate. Maybe they are the marginal investors who set prices, he speculates. Well, maybe. But an increase in the tax rate will make dividend-paying stocks more expensive for them but the same price as before for non-taxpaying investors—so as long as we’re going to stick to theory, the former should sell their stocks to the latter for some price between the two.

More important, the price of a stock (in theory, again) is the discounted present value of its future dividend stream aggregated over an infinite horizon. So we need to know what the tax rates will be in all future years. That’s clearly unknowable. If the tax rate goes up on January 1, 2013, that will give us no information about the tax rate in 2113. On the other hand, it will give us very good information about the tax rate in 2013. And it will give us a little bit of information about the tax rate in 2023. In other words, the informational value of a change in tax rates only affects a small part of the summation you have to do if you want to value a stock by its dividend stream. If a company is going to shut down in 2013, liquidate its assets, and return one massive dividend to shareholders, it affects most of the value. If a company is Facebook and is unlikely to pay dividends for a long time, it affects very little of the value. So the impact of such a change on stock prices will be a lot less than the theoretical 33.4 percent that Luskin calculates.

Then there’s the little matter of markets. Luskin’s article chides the “stock market” for ignoring the upcoming change in tax rates on dividends. How does he know? Did he ask the market? More likely, the market is pricing in the possibility of a change in tax policy. In theory, market prices today should reflect the expected future tax level, which is somewhere between 15 percent and 43.4 percent—closer to which one, we don’t know. This is another reason why the actual impact of a tax increase will be smaller than 33.4 percent; the latter assumes that every single investor today is blindly assuming that the tax rate will remain at 15 percent.  (Actually, since the Medicare surtax is already law, every single investor knows that the tax rate will be at least 18.8 percent, not 15 percent.)

But this is all theory. There is actually a way to test these things. To the extent that a change in the dividend tax rate affects stock prices, it should affect high-dividend stocks more than low-dividend stocks. Even on the theory that the value of a stock is the discounted value of its future dividend stream, for a high-dividend stock, much of that value comes from dividends in the next decade, which are likely to be affected by a change in the tax rate. By contrast, for a company that doesn’t pay dividends, the value of its dividend stream is located far out in the future, where a change in today’s tax rate has little expected impact. So if Luskin is right, the 2003 tax cut (which established the 15 percent rate for dividends) should have caused not only a sharp increase in stock prices but also a sharp increase in the price of value stocks relative to growth stocks.

So, courtesy of Yahoo! Finance, here are the closing prices of the Vanguard Value Index (red), which includes high-dividend stocks, and the Vanguard Growth Index (blue), which includes low-dividend stocks, for November 2002 through May 23 2003, the day the final bill was passed by both houses. The question, though, is when the 2003 tax cut would have affected stock prices. There’s no separation between value and growth stocks around November 5, the day the Republicans won the midterm elections.  (Remember, the Democrats had a Senate majority in 2002.) There’s none around January 28, when President Bush called for tax preferences for dividends in his State of the Union address. There’s no reaction around February 27, when the bill that would cut taxes on dividends was introduced.

Now, there is a separation around May 15, when the Senate version initially  passed. (Passage in the House was assured because of the Republican majority there.) This implies that there was significant uncertainty about whether the bill would pass; when the uncertainty cleared, high-dividend stocks gained relative to low-dividend stocks. Score one for Luskin!

But if there was uncertainty that cleared on May 15, and Luskin is right, then two things should have happened: high-dividend stocks should have gained relative to low-dividend stocks, and all stock prices should have shot up. But that’s not what happened. High-dividend stocks went up; low-dividend stocks went down. Investors’ overall appetite for U.S. stocks didn’t change; at the margin, some realized that after-tax dividend yields had just gone up, so they switched from low-dividend to high-dividend stocks.

By May 23, the last date on that chart, passage was a certainty, so the impact of the tax change should have been 100 percent priced in. Do you see a 30 percent increase? I don’t.

Want more evidence? Here are the same two index funds for December 1 through December 17, 2010, when the dividend tax cut was extended for two years. The extension was in serious doubt until December 6, when Democrats and Republicans reached a compromise agreement. Again, you can see an increase in the price of high-dividend stocks relative to the price of low-dividend stocks, starting around December 6. This indicates that the market was reacting to a significant change in the probability of an extension. But there’s no sharp, 30 percent increase in the overall level of stock prices.

So the tax rate on dividends does seem to have a small but visible impact on the relative price of high- and low-dividend stocks. And it may have a small impact on the overall price level, which would make sense. But 30 percent, or anything close to it, is pure fantasy.

So why all this hysteria about a collapse in the stock market on January 1? Well, here’s one hint, from Luskin’s article:

“If there’s a bargaining failure and the scheduled tax hikes on dividends aren’t stopped, we’ll be sorry we’re spending so much political energy now debating about the ’1%’ and their supposed privileges. It’s the 30% down in the stock market we ought be worrying about.”

This is just another attempt to mask the blatant unfairness of the Bush tax cuts by arguing by arguing that that they are good for all of us (well, at least all of us who own stocks, but that’s a matter for another post). They’re not.

* Luskin also talks about “trillions more in new tax hikes under ObamaCare.” Huh? The revenue provisions of the Affordable Care Act are projected to bring in $520 billion over the next decade; even if you include the revenue-increasing coverage provisions (like the excise tax on high-cost health plans), you only get up to $813 billion. That’s not “trillions,” unless you’re talking about an undiscounted infinite horizon.

13 responses to “Bad Dividend Math

  1. Desi Girl - Doing God's Work

    We have a huge problem in that academic elites like yourself and simon are responsible for educating the youth of this nation.

    What do you do? Respond to a useless article with your own useless crap. Any taxi driver will tell you that dividend paying stocks do not accurately reflect the tax rate. There are other things like the risk free rate and risk premia…but i know that you know these things.

    More Importantly, when refuting a pointless article, why don’t you do some very basic research to look like you have half a brain – this is after all, your chance to make a good argument:

    VIGAX top holdings include Appple, Google, that do not yet pay dividends and IBM and Coke that DO PAY dividends.

    VVIAX top holdings includes Wells Fargo and JP Morgan that were not allowed to pay dividends in 2011.

    Now, VIGAX dividend yield is 1.15% and a 3Yr. dividend growth rate of 10.15%

    VVIAX dividend yield is 2.5% and a 3Yr. dividend growth rate of -5.3%

    It would have been simpler for you to respond to Luskin with a simple “Ha Ha Ha”.

    To me, you are exactly like the mortgage broker selling a subprime security to a bunch of folks who do not know what they are buying. You sit in a university instead of a bank.

    No wonder these kids end up squatting on the streets outside my office in NYC.

  2. Time to recalibrate you lithium, Desi Girl.

  3. Desi, I think he trying to make the point of the interest in the blog -vs- the future taxes incured from operating such a blog, yes, we have educational issues, and its either the gvts fault, the parents fault, or the persons fault. The world today needs to much documented proof to get things right. And when they want proof that a God exists, the timer of demise starts within them, and it curtains right around the corner. He does need to better sleep though, if he is losing it over this.

  4. The obvious question to ask Luskin is whether stock prices increased by 30% overnight when the qualified dividend tax rate was lowered in 2003.

  5. I guess Desi Girl’s having a PMS moment…

  6. Luskin stands everything on its head. Honestly, if you can’t make money why bother investing to begin with. There is really no rational reason to believe that expiration of Bush tax cuts will make any difference on how people will invest. You would have to move your money somewhere. I guess Luskin is suggesting that come January 1st everybody will have a mattress full of money — unless they take their money and actually spend it? Anyways, the whole Bush Tax Cut was a reaction to the argument of double taxation of corporations and the dividends paid to their shareholders. Please if you are going to make an argument. Please make the right one, Mr Luskin. It is not like multi-millionaire or billionaires are really hard up for this tax break anyways.

  7. Moses Herzog

    Hey, what’s a mathematics error??? When a TBTF bank can lose $2Billion from the CIO unit, which is money they use from DEPOSIT ACCOUNTS, while simultaneously telling us how “expensive” and “costly” the Volcker Rule would be, See here: http://www.bloomberg.com/news/2012-05-10/jpmorgan-chase-says-cio-unit-suffered-significant-loss.html
    …. obviously math is something silly that only “old-fashioned” people like Paul Volcker or bank-welfare providing taxpayers would care about (you know, those people lacking “innovative” minds, and excessively “old-fashioned” about people playing games with their savings deposits).

    Thank God, that we have Jamie Dimon, CNBC anchors, Lloyd Blankfein, Mitt Romney, James A. Johnson (of Goldman Sach’s compensation department) and others to fight this evil propaganda of the Volcker Rule. Otherwise this “innovative” new idea to lose $2Billion from deposit money in JP Morgan’s CIO unit never could have been “invented”. Only the sharpest of minds work for the TBTF banks. Maybe JP Morgan can ask Bruce Berkowitz for advice on how to get out of this. Bruce Berkowitz is a man known to be especially creative, imaginative, and even INNOVATIVE on investor public conference calls with TBTF banks. Berkowitz is a selfless man obviously “deeply concerned” that retail investors not be led down the primrose path.

    Super Berkowitz to the rescue!!!!!

  8. End Math Abuse

    Yet another problem with the calculation—it assumes that the risk premium for stocks (net of taxes) as well as interest rates will remain unchanged. An investor suddenly finding his future net-dividends reduced by higher taxes would only sell his stock if he could find a more suitable usage of the money. He either would spend it or buy other investments.

    Even if people assume taxes will never change in the future and the upcoming change is unexpected, stock prices would fall only to the extent that investors moved their money into investments unaffected by the higher taxes–bonds. Interest rates would fall as would stock prices until a new equilibrium is reached.

    Luskin’s piece is an embarrassment.

  9. In fact, all qualified pension plans are federal tax-exempt, and so an increase in dividend tax rates will have no effect upon them, or upon the plan beneficiaries who will receive the plan’s income in their distributions, which are taxable at ordinary income rates. Luskin’s piece is inflammatory and completely false as it applies to pension plans.

  10. All qualified pension plans are federal tax-exempt, and so an increase in dividend tax rates will have no effect upon them or their participants. Pension distributions, which include participant contributions as well as the investment income, are taxable at ordinary income rates. Likewise, distributions will be unaffected by an increase in dividend tax rate. Luskin’s article is a red-herring, completely false and non-applicable to pension plans.

  11. ConclusionHopper

    The Luskin argument is flawed, and there are other ways to show that mathematically. First of all, the dividend growth model is not the only show in town. It doesn’t really work for companies paying no dividends, just as DCF models don’t work for companies with negative FCF’s due to investments in operating activities. Using a model like the Abnormal Earnings Growth model would put firms on a more level playing field in terms of valuation, because it anchors value on earnings, as opposed to dividends or free cash flows. It also allows for cum-dividend growth rate calculations, but all of this mumbo-jumbo is irrelevant.

    Do we think firms are managed my machines? No – they’re managed by people who want to avoid pissing off shareholders. There is a behavioral side to corporate finance, and that is being completely overlooked here. If (some) shareholders actually have to pay higher taxes on dividends, then they’ll seek to force a change in payout policy for their firms. Granted, CFOs would do anything to avoid cutting dividends, up to and including selling off their first-born son. But that’s in the current model. If shareholders want payout to shift from paying dividends to repurchasing shares, than that’s what would happen.

    I can’t predict the future, but this seems more realistic than claiming the entire market will drop 30% overnight. Not to mention the arrogance of assuming the world economy is 100% controlled by the US tax laws… but let’s leave that alone. The bottom line is this – shareholders will make firms change undesirable payout behaviors, and investors will still want to invest. Let’s not forget these are people on both sides of the market, some acting more rationally than others, but all wanting to make money. They’re not going to take their ball and go home.

  12. dividendstocksonline

    The market will price this in ahead of time and the impact will not be measured in a day. Its true that fund managers that run investors 401ks will not make changes but many retail investors will. It will also impact future investment and the valuation given to these stocks.

  13. Normally, I would not respond to anything revolving around Donald Luskin because, basically, he has been so wrong about everything for so long it is hard to understand why he is still given a platform in the media. However, as the fairly recently retired Chief Investment Officer of a $100 billion plus asset management firm I feel some points need to be made.

    1) This is an area where there are reams of empirical data because for years the tax rate on dividends diverged from the tax rate on capital gains and was, in general, quite a bit higher and so we know what happens, which is:

    a) the pre-tax expected return of dividend paying stocks will be a bit higher than the low dividend paying stocks, basically by the amount equal to the effective marginal rate in the market.
    b) the result of this is that tax-exempt and tax-deferred investors tend to overweight the dividend paying sector. This is the primary source of the so-called value effect as documented most popularly by Fama and French, but was first hypothesized and implemented by Wells Fargo Investment Advisors (who many corporate transformations later is now part of Blackrock) in the late 1970′s via a mechanism they referred to as the Security Market Plane, effectively a two variable CAPM. You can see this described if you can find a 1970′s version of Bill Sharpe’s “Investments” text book.
    c) companies will respond by cutting dividend payouts and substituting share buyback programs etc. Note that the dividend yield on the US market is and has been meaningful lower for the past 30 years. This also partially reflects the triumph of Modigliani-Miller (dividends don’t matter.)
    d) last I checked we still managed to have robust stock market returns through the 80′s and 90′s irregardless of tax regimes, and since most of Luskin’s personal prescriptions have been implemented we are still waiting for a return to the market highs of over 12 years ago.

    To those who feel they should utilize a dividend discount model for valuation purposes, please note that any modern practitioner will tell you that what is normally used is an estimate of the dividend that could be paid should management so choose since dividends can actually accrue to shareholders in many forms (successful acquisitions, stock buy backs, spin-offs, etc.)

    It is true, however, that one can see that a differential taxation of cap gains and dividends tends to incentivize some undesirable behavior on the part of management, such as holding onto cash even when they have nothing particularly good to do with it. In general, I support higher taxes on capital vis-a-vis labor but big wedges between cap gains and dividend income just incentivize bad management behavior. Something I would observe that many managements are good at all by themselves, so in general, I think a good argument can be made that equal or nearly equal tax treatment of cap gains and dividends is more socially desirable.

    It is important to remember that almost anything Donald Luskin says is nonsense. I believe at Barry Ritholz’s blog you can find an archived post detailing this in exquisite detail.