Friday, October 10, 2014

Evidence is Optional for Finance Cynics, a continuing series: Cash Flows to Owners

While I'm complaining about the fact that everyone seems to be wrong about everything, let's talk about cash flows to owners - dividends and share buybacks.

Everyone from Matthew Yglesias at Vox to the Economist are worried about massive buybacks and how this means that corporations are just pocketing your cash instead of investing in America.
Yglesias:  "It means that the basic link between healthy corporate profits and a healthy middle class is broken."
The Economist: "Share buy-backs - Corporate cocaine"
I don't want to even get into the conspiratorial tone with which buybacks are usually treated.  It's simply a return of capital to owners, no different than dividends.  The main adjustment it does demand from us is that historically dividends were the primary method for returning capital to owners, so that indexes, like the S&P500 or the Dow Jones Industrials, which tracked share prices over time also served as decent proxies for the growth of capital.  But, buybacks cause return of capital to be treated, mathematically, like capital gains.  Buybacks have gained favor in recent decades.  So, in order to compare capital growth to other measures of wealth or income, equity indexes need to be reduced by the amount of buybacks.  At this point, return of capital is causing index returns to be overstated by more than 2% per year as a measure of retained capital, which really adds up over time.

I repeat, this is simply a mathematical adjustment that needs to be made for analysis.  Total returns are unchanged, except for this arbitrary change in accounting for them.

But, forgetting all the misconceptions surrounding buybacks, the more basic point is:  There is NOTHING unusual about the current levels of payouts to equity holders.

Here is an excellent article by the inestimable Aswath Damodaran going over the basics of share buybacks.  Here is one of his graphs, showing the total level of corporate payouts to owners over the past 30 years.  The level of payouts is slightly below the levels of the 1980s.

I think arguments can be made for either looking at net payouts or gross payouts.  But, with net payouts between 3%-4% and gross between 4%-5%, both are moderate.

What if we look at a longer time frame?  Here is 150 years of dividend history, from Robert Shiller's data.  Keep in mind that this is only dividends.  Buybacks have only become popular since the 1980's, so we would compare these payouts to the gross payouts from Damodaran's graph, which were around 6% in the 1980's, dipped to 2% to 3% in the 1990's, and are back up to around 4%-5%.  Historically, payout ratios have rarely fallen much below 4%.  If anything, there is a long term downward trend in payouts to owners, following a range that current levels fall squarely within.

How can there be so many topics where so many people are so confident about things that just obviously aren't so?  Would it be a shocking coincidence if the bad guys in these stories were predictable?

How much human misery has been caused by people who got in the habit of believing things, not based on whether they were true or false, but based on whether they had a predetermined bad guy?  It's kind of the story of human history, isn't it?  I'm excited about biotech, AI, robotics, nanotech, and everything else, but, man, the real breakthrough would be if someone invented a way to point this bias out to each of us in a way that would make us say, "Oh, geez.  You're totally right.  That is what I'm doing.  I will now update my beliefs with this in mind.  Thank you."  Human progress would explode.  The main reason the singularity will happen is because, frankly, the bar is set very low - and I don't exclude myself.  We all purposefully develop strong convictions through gross avoidance of evidence.  It may be the core identifying feature of humanity.


  1. Can't we assume that anyone who owns the stock believes the intrinsic value is higher than the share price?

  2. The history seems to be that share buybacks are highest at the peak of the market. I would much prefer that companies declare extraordinary dividends if they have surplus cash burning a hole in their pocket. If companies insist on buybacks perhaps there should be a rule restricting them so they cannot pay a price above the 200 day moving average.

    When the current bubble bursts there can and should be class action law suits against directors who authorized share buy backs at or near the market peak.

    1. Yes. Let's simultaneously demand superhuman abilities from them while depriving them of universal legal rights. That'll learn em.

    2. Kevin: There is no universal legal right to strip cash out of a corporation. The history seems to be that buy-backs are commonly done at the wrong time and contribute to widening the swings in the market. Restricting buy backs to prices below the 200 day moving average would make buy backs a stabilizing mechanism.

    3. "There is no universal legal right to strip cash out of a corporation."

      This is indeed a direct measure of liberty. Countries may be considered repressive or uncivilized specifically based on whether they use financial repression and capital controls. In any case "to strip cash out" is certainly a loaded way to describe receiving cash repayments for an investment.

      All investments commonly contribute to widening swings in the market. This is mathematically unavoidable, really. So why stop at buybacks? Why not prevent all capital allocation decisions until your algorithm says it's ok? If the corporation can't buy their own shares, why are you letting other people buy their shares?

    4. Also, the universal legal right I was referring to in my earlier comment was the right not to be prosecuted for market fluctuations that happen after the stock purchase.

      And, regarding controls on capital, it is textbook finance that those kinds of things require higher returns on equities. Thus, financial analysts add risk premiums in foreign markets that don't have solid property rights, and so labor captures much lower shares of national income in those markets because capital requires higher profits on the same asset base to make up for the limits on capital dispensation.

  3. Does it matter what they *believe*? It would only impact whether they *believe* that they will benefit, not whether they will *actually* benefit.

    The reality is that the stock is either undervalued, appropriately valued, or overvalued independent of what they believe. Look at the Nasdaq in 2000. Investors' beliefs did not save them! It is this relative valuation that determines whether a buy back is beneficial or not. My point is that you don't have this issue with dividends (although either way you have the issue of whether it is *optimal* or not).

    Anyway, I don't mean to be argumentative. Love your blog!

  4. Thanks Mark. Beliefs are definitely what matter. Buybacks are basically similar to dividends with an opt-out reinvestment program. My point is that nobody would own the stock if they thought it was overvalued. So, logically, I don't think we can say that buybacks ever represent an allocation of capital which the owners consider to be suboptimal because of price. If they thought they were, they should have sold their own shares.
    Their belief is what matters because the price of the equity is a product of many uncertain future cash flows in the numerator and many individualized factors in the discount rate in the denominator.
    There is no rational way for managers to second guess the denominator for each shareholder, and thus there is no way for them to derive a company-wide "intrinsic value". Every owner has their own intrinsic value, which is what makes a market. Except for points in time where managers have insider information about catastrophic developments, I don't think you can argue against buybacks based on price. If share prices are objectively too high because of unwarranted shareholder optimism about the firm's prospective ROI, then returns of capital that bely that expectation will lower the share price regardless of whether they are dividends or buybacks.
    You can see this dilemma with some of the firms that were overpriced in 2000. There were firms flush in cash, which were priced as if they had 30% ROI that, in hindsight, they didn't have. So, what does management do if they can't justify the share price with their available internal investments? Any return of capital signals that the market has overvalued them. A buyback means that more skeptical shareholders received a higher payout than shareholders who held their shares as the market realized that the firm didn't have a way to meet shareholder expectations. A dividend (assuming no reinvestments) would have created shared losses more broadly, as all shareholders may have received an equal cash payout before they experienced capital losses. In those extreme cases, I don't know if there is a clearly optimal method. But, in any case, I don't think these rare cases can inform us in most contexts, where valuations are within the realm of reasonable differences of opinion about future cash flows and required returns.
    Buybacks involve a voluntary transaction with a subset of shareholders who want to sell at the current price and a voluntary lack of a transaction with most shareholders who want exposure to the firm's internal investments. This should usually be better than a broad dividend that is a sort of compromise method that doesn't account for any individual owners' preferences. (Of course, there are some types of firms for which the dividend is the point, and there the heuristic might be different.)

  5. I raised very similar questions to Damodaran at his blog with respect to his initial post. I guess he would have to answer that question since the graph came from that original blog post. When I first raised the issue, he clearly stated his data represented gross buybacks. After my comment, the blog post went down for a while and then reappeared. I'm not sure, but I suspect the graph on "net" numbers was then posted in response. I raised the question again on October 8 in a comment to which he has not replied. The numbers are pretty confusing, at least to me.

  6. TravisV here.

    Christopher Mahoney just posted an update of his analysis (which I think is great)........

    1. Definitely seems like what we've been talking about.

      I'm not sure about his stock allocation, though. It's right if we just muddle along and the economy has enough juice to overcome tight monetary policy. But, if we're headed for a demand shock, I don't think we want to be in stocks, tactically.

  7. Thanks for your response, Steve. I do agree that there could be problems with measuring payouts compared to market cap. Over time, total required returns to corporate assets (at market value) have been pretty stable, so I don't know if it's that much of a problem. But, using an adjusted payout ratio may be better than an adjusted cash yield, as you suggest.

    That said, I don't understand where the Slack Wire's numbers could have come from. S&P500 dividend payouts were historically around 5% - frequently more. And net margins tended to be 5-7%. Recently margins have been notoriously high, and total payouts have not been higher than they have been historically.
    Off the top of my head, I wonder if he is using domestic profits and comparing that to total corporate payouts (which would reflect global profits). His graph contradicts known quantities too much to be accepted without some detailed explanation. I suspect he's partly just measuring the increasingly international character of corporate profits so he's operating with inappropriately mixed data.

    As to the rest of your and his analysis, I think there are an awful lot of presumptions being made about informed investors vs. passive investors and about the motives of capital owners. Even with dividends, there is active trading around ex-dividend dates and arbitrage involving the relationship between share price and dividend payments. These are very liquid markets, with a lot of large players who frequently adjust positions, so I think your description of the differences need to rise above mere assertion. Some of your assertions about buybacks are embedded in pretty strong and falsifiable anti-EMH presumptions.

    Further, your implication that there is some problem with market activities that favor active traders who are on the right side of the trade relative to intrinsic value is perplexing. And, the Slack Wire post seems polemical. This quote seems loaded: Shareholders "take pretty much the same view of the corporation as a praying mantis does of her mate." He begins the post complaining that firms are paying out too much of their cash flows to shareholders. And he ends by acting like Apple owners are inappropriate for even having a discussion about asking for some of the largest pile of cash ever collected to be distributed. This isn't a discussion based on data or even on theory. It's just based on affiliations. Is he on the record supporting any return of capital, or whatever the return is, is there always a preferred beneficiary instead of the owners?

    You bring up very interesting ideas at your blog, but there sure looks like a lot of motivated reasoning going on here.

    Maybe these presumptions about capital owners are as accurate as the presumptions about praying mantises.

  8. Vivian, I think much of it is from stock options. At first glance, I agree with you that option awards are much more than an insignificant part of the picture. But, he might have a deeper understanding in mind. There are a lot of subtle things going on with stock option issuance. I haven't thought a lot about it, but it seems like the extensive use of stock options could inflate the stated gross payout level, since the exercise of options brings in capital, that, presumably, the firm will basically redistribute if they are targeting a given capital balance.

  9. Steve, here is the Wikipedia page on payout ratios. This doesn't seem to match with The Slack Wire chart.

  10. Kevin,

    Let's not get too snarky about mantisses and motivated reasoning. We all engage in motivated reasoning. We're human. We communicate to help us test our reasoning, whatever its motivations. I'll be the first to defend polemics in blog posts, makes them more fun to write and more fun to read. They add less, IMHO, one-on-one(-ish) conversations.

    We're interested in the total payout ratio, not the dividend payout ratio. Much of this conversation depends upon upon the increasing role of buyouts as a means of returning cash to shareholders. One way to resolve some of our questions would be to ping Josh Mason for his sources (which ought to have been better documented in his posts) and/or information on his computations.

    Re "your implication that there is some problem with market activities that favor active traders who are on the right side of the trade relative to intrinsic value is perplexing". I have a certain sympathy to your implication. Investing, in a view I used to hold dearer than I do, is information work, a competition in which those who are better informed punish those who are worse informed, in order to drive share prices quickly to fundamental value. Positive returns are returns to quality of research, prediction, etc. Poor results are condign punishment for low quality information work.

    However, even under that view, the dynamic I allege, speculatively, cannot be defended. If there is any truth in those speculations, a purpose of share buybacks is precisely to prevent rapid and efficient price adjustment by giving informed investors a counterparty at the overvaluations where they wish to sell. It's one thing to reward the informed over the uninformed in support of efficient pricing. It's quite another thing to prevent efficient pricing to reward the informed.

    There is plenty of active trading around dividends and attempts to arbitrage ex dividend dates. Once upon a time that was a fetish of mine. But for better or for worse (I think mostly for worse, and perhaps we'd agree on that), "we" have decided that equity markets are a place where completely uninformed people should be encouraged to put their savings at risk, via index funds, pensions, and other intermediaries. If that weren't true, we might take the attitude that "the market is a shark tank, if you intend to buy stock, expect to swim with the sharks". But that is not a reasonable attitude given the social function equity markets are actually asked to perform. Most of the money in equities is institutional money managed under mandate that does not engage in any kind of high frequency arbitrage. Unless we remedy that (which would mean a massive reconfiguration of our financial system), we have to structure those markets to be "fair" to such participants. Fairness doesn't mean they shouldn't lose money — when equities are overvalued they must fall. But it does mean that practices which 1) disproportionately advantage "fast" or "influential" market shareholders; and 2) don't contribute to the efficiency of market pricing are problematic and should be opposed. In my view, again admittedly speculative but I've been following this stuff a long time, share buybacks increasingly fall into this category.

  11. (continued from above. i talk too much.)

    (Note the increasingly. Early on, managers really did seem to engage in buybacks when firms were undervalued, evidenced by future performance of buyers-back. That strikes me as unproblematic. Then buybacks were viewed as a signal of undervaluation, which created opportunities to signal strategically or to boost share-linked compensation. This added a lot of noise to the signal. Recently, I don't think there's much information at all in buyback announcements with respect to valuations, as buybacks are motivated by a mix of real undervaluation, strategic attempts to signal undervaluation, and influential-shareholder attempts to enable exits with less price action.)

    Note that dividends, while frictions surrounding their issue and valuation may create some opportunities for arbitrageurs, don't create differential opportunities for informed vs uninformed shareholders. Everyone gets the dividend, which presuming the expected price decrease, doesn't alter under/over-valuation at all. Informed investors who believe shares undervalued can reinvest (intermediated investors tend to reinvest in a diverse portfolio). Informed investors who believe shares overvalued have been given no new means of exit, no way of putting future losses to less informed shareholders. They have to go to market to sell, and endure their share of the repricing. Buybacks are just not the same as dividends in this way.

  12. (Oh wait! I was too mean to Mason. He cites his source. It's the Fed's Z.1 Flow of Funds report, which does seem a reasonable place to document aggregate fund flows.)

  13. Thanks for the replies, Steve. I apologize if I seemed snarky. I agree that it can be unfair to go after motives. It's awful easy these days to besmirch owners as a class. I just think the good-guy-vs-bad-guy stuff undermines intelligent analysis. I'd like to find people who can disagree with me without using it. But, maybe I'm just a bad guy, and I'm asking for the impossible. :-) Has Mason ever seen a distribution to owners that he liked?

    Here's a link to a Fred chart with corporate dividends. I'm sure it's not exactly what he's using, but all the combinations of Z.1 data I find do show the same pattern of dividends rising after the 1970's. It's strange that it differs so much from the data on the S&P 500. He shows a 100% payout, just from dividends, from 2000 to 2004. And, he has the drop in 2005, which I see at Fred also, which doesn't show up in S&P 500 dividend yield graphs, which, again, makes me wonder if some of it has to do with comparing foreign vs. domestic profits to dividends, since this came after the 2004 repatriated profits tax holiday.

    I've used Z.1 data to come to counterintuitive conclusions myself, so I can't rightly throw stones. But, I'd be curious to hear more details.

  14. This comment has been removed by the author.

  15. Hmm. The dip in 2005 doesn't show up in the graph I linked to above, which is for all corporations. But, it does show up in the graph for domestic non-financial corporations.

  16. Some observations on the effects of tax rules:

    1. Prior to 2003, dividends were taxed at ordinary tax rates. The marginal tax rate being, generally, 39.6 percent. Prior to 2003, LTCG were favorably taxed at marginal rates of 20 percent. From 2003, marginal rates were generally equalized at a rate approximately one-half of marginal ordinary rates. Given the tax consequences, all else equal, one should have seen a significant relative shift from distributions via stock buybacks to dividends after 2003. The data does not seem to bear that out. Why? Are corporate finance experts oblivious to the tax consequences to their shareholders? Or, is something else going on?

    2. While rates on both LTCG and dividends both decreased (particularly the latter) during this period of time, corporate federal tax rates remained constant (35 percent). Thus, the relative tax burdens should have resulted in higher corporate payouts after 2003. The data does seem to (slightly) bear this out. Conversely, one would expect a lower distribution rate when and if corporate (effective) tax rates are ever reduced.

    Also, one should expect stock buybacks to increase at market lows if the rationale for stock buybacks is to repurchase intrinsically undervalued stock. But, both the "gross" and "net" figures seem to correlate more with market highs. I strongly suspect this is due to the fact that much of that activity is due to the need to fund employee stock and option programs and not merely that companies have more cash on hand. Stock options are exercised at market highs---not lows. The same is true of other types of equity compensation, such as restricted stock, etc. This leads me, again, to believe that the effect of equity compensation plans on the level of stock buybacks is significantly understated by many observers, including the inestimable Damodaron. In fact, it even draws into question the assertion that the difference between "gross" and "net" in his data really represents the effect of "equity compensation". I'm very suspect of the underlying data.

    1. Thanks for the interesting information, Vivian.

      I think it's worth considering the amount of cyclical movements that are caused by real changes in potential output. Higher distributions at cyclical peaks don't necessarily represent herding or irrational behavior.

  17. Oh, we're none of us bad guys, or good. We're working at it from different perspectives, trying to figure stuff out. My view if the social world is that we're all trapped in it, and we're all most likely fucked. (Sorry. I'm not an optimist, but I'm working at shifting the train to a better track as best I can.) My view of the political world is that the institution of "blame" is a social prerequisite to creating the incentives that define "accountability", and the incentives of the system, not individual goodness or badness, determine whether we go nasty places or good. So that leads to a kind of schizophrenia: you can't be bloodless, it's ineffective, but you're never any better than the people whose blood you are after. In a recent post you disliked, I thought I did a pretty good job of analyzing how the perfectly rational incentives of people with different amounts of wealth relative to expected consumption requirements would lead to divergent interests, even if they were all the same person. So there's no good or bad in that, just a theory (correct or not) of an an agent with a certain utility function under uncertainty that would imply she behaves differently under different endowments. But then there is the sneeriness that you disliked too. I won't apologize for that: I'm in this as an applied thinker. I want to understand correctly, but I also want to change things, and that requires some moralizing and color. But again, in one-on-one conversations, I think that's usually a lot less useful. Which isn't to say I don't do it, in less public conversations. I just usually regret it more, when I fail to stop myself. You have nothing to apologize for. I just didn't want to get into a tit-for-tat kind of exchange.

    (BTW, I think the same is true of Mason. He can speak for himself, but I think he claims that there has been a meaningful derangement of corporate capitalism over the last few decades. That implies that he probably liked the kind of distributions that happened before that derangement okay, or at least better than he likes present disgorgements.)

    It's not obvious to me how Mason would have pulled net share repurchases from the flow of funds data (It's got to be there somewhere, but I don't see it, I think you'd have to tease it out, and it's might not be easy to, say, disentangle flows into corporate equities from flows into mutual funds that are indirect flows into corporate equities, etc.) But I do have a lot of confidence in Mason as a researcher. So I guess we should ask.

  18. So this is a naive graph, dividing series from different sources, just quickly playing around in FRED, I haven't thought the data issues through very much here and it's an area where definitions and discretionary choices in measurement can matter. Caveating aside, the quick-and-dirty FRED look does show a significant increase in divident payout ratios beginning in 1980s, which could only have been exacerbated by the increased prevalence of buybacks over the period. If you adjust profits for changes in the valuation of inventory and capital, the increase looks more modest (suggesting "capital gain" as untraditional profit on corporate inventory and capital), but it is still pretty clearly there (those "capital gains" can't account for all the increase in payouts).

    Anyway, here's that graph, apologies if I've made some obvious or less obvious data gaffe.

  19. I have posted a question at Mason's blog post about the data.

  20. Here's another issue I have on the "gross" and "net" distribution data. How do these data sets deal with corporate mergers? In a straight merger between domestic corporation A into B, A issues stock to B shareholders. B's stock "disappears". Similar things happen in a forward or reverse "triangular merger", B's stock would be indirectly acquired by A (through a merger of B into A's sub), with stock issued to B's shareholders. Does this reduce "net distributions?" How are the people putting this data together sorting that out? What if one of the acquirer or target is a foreign entity? How does this show up if we are relying on US stock market, especially on data of only one market index within the US (if the target and acquirer are, for example, not both on the S&P 500?) I have many, many questions about how this data is put together.

    1. OK, that should be B merges into A but the point remains.

    2. Great questions. I don't have an answer.

  21. Every week-end I used to pay a fast visit this site, because I’d like enjoyment, because this web site conations certainly fussy material.