Monday, June 26, 2017

More non-voting shares

Tim Kroencke at the University of Basel wrote a nice follow up on non-voting shares (previous posts here and here) , which I share with permission. Some of the controversy was whether companies would issue shares and whether investors would by them. It turns out, yes, and he sends a gorgeous example in which control rights and cash flow rights are priced differently and react to different events:


In Germany, it is quite common for large companies to issue voting shares (Stammaktien) and non-voting shares (Vorzugsaktien) and one can make nice case studies. Here is one I did a while ago,  that I have updated today, and I want to share with you:

In 2005, Porsche started to buy Volkswagen shares. In 2008, it became obvious that Porsche tried to overtake Volkswagen and the price of voting shares, and only the voting shares, skyrocked. Volkswagen became the world’s biggest company…  well, for a couple of days.

Some figures to give perspective: first, the share price of non-voting vs voting Volkswagen shares traded in Frankfurt:

The dividend yield:

 And here is how prices and yields add up to total returns:

Some observations:

First, the voting component of a share price can diverge substantially from its cash-flow related value. What is small most of the time does not need to be small all of the time. This should really worry any passive investor who simply wants to earn a factor premium. The typical broad index investor wants to earn the market premium. I really doubt that such an investor wants to be involved in the house of cards of the Wiedekings and Piëchs. There is a reason for hedge fund investors being around.

Second, voting and non-voting shares nicely move together - in the long-run. After all, they pay out a very similar cash-flow stream, as you write and as one would expect if the law is set up in a sensible way.

Third, non-voting shares outperformed the voting shares by roughly 100% over 18 years. This is the long-run picture, difference in cumulated returns come from differences in dividend yields. Cumulated over time, the premium for voting can be quite big!  Sure, this does not have to be the case in general. (For example, in 2011, die dividend yield of voting shares was slightly higher.) After all, we are looking at a failed takeover and this is just an example. However, I think one can safely make the point that non-voting shares are likely to better track the value of the cash-flow value of a company and are indeed well-suited for passive long-run investors.


I was initially puzzled by the rate of return difference. If you start and end at the same price, but pay the same dividend, how do you achieve a different return? I think the answer is reinvestment. Dividends paid to the voting shares during the spike are reinvested at a time of terribly high prices, and so lose.


  1. This series of three "non-voting" posts shows how wonderful this blog is in spreading the information and learning new things together with the world!

  2. Yes, Wonderful (W.) is the right word here.

  3. There is something odd about this comparison between the two types of shares. Perhaps voting shares perform relatively poorly only if we refuse to sell them during the takeover bid, when there is a large premium for the voting right. But we expect fund managers to make such decisions optimally on our behalf.

  4. If the charter of an index fund is to track an index, wouldn't it be obligated to buy whichever shares, voting or non-voting, are in the index?

    Do German companies that are not controlled by a family or dominant owner issue both classes of shares? How does this practice differ from the multi-class shares in companies like Facebook and Google?

  5. Here in Brazil we have a lot of stock classes. In general with and without voting rights.

    Those without voting rights we call "ações preferenciais" (preferred shares, but it is not like in US - because it is not mandatory to pay dividends in a regular basis).

    Thoses without voting rights has, additionally, some different kinds.

    It is a little tricky.

  6. I think you are all focusing too much on the Asset Pricing side of the issue. The voting vs. non voting problem is mainly a Corporate Governance problem that is well known since Grossman & Hart (1980) and (1986). The reason why non voting shares are extremely dangerous is because they make possible for a raider not to internalize completely the effect that he's going to have on the value of the firm, implying that value decreasing acquisitions where the raider can "steal" from the firm end up being incentivized by a dual-class system. Adams & Ferreira (2008) have a great survey on the empirical evidence and they conclude that the dual class system seems to generate more harm than good:

  7. John,

    "In 2005, Porsche started to buy Volkswagen shares. In 2008, it became obvious that Porsche tried to overtake Volkswagen and the price of voting shares, and only the voting shares, skyrocked. Volkswagen became the world’s biggest company… well, for a couple of days."

    What should also be apparent from your graph is that the price of non-voting shares plummeted. By itself, Porcshe running out and buying up voting shares should have had no effect or perhaps a slight positive effect on the price of non-voting shares.

    Instead, it looks like is that Volkswagen invoked a clause or performed a market operation that converted its outstanding voting shares to non-voting shares.

    Of course, looking at both price and quantity for both types of shares would tell you the whole story. My hunch is that the market cap for both types of shares (voting and nonvoting) was relatively unchanged from 2007 to 2010.

    And I would also venture a bet that it was the voting shareholders (or at least a majority of them) that approved that measure.

  8. John,

    Has it occurred to you that the voting shareholders of Volkswagen could have voted to convert a portion of their voting shares into nonvoting shares?

    That would explain both the run up in price of voting shares AND the fall in price of non-voting shares starting in 2008.

    And it would also pour cold water over your "Total Return Index". Your total return index assumes that stock ownership is limited to only owning voting shares or only owning non-voting shares.

    But suppose that during the period in question (2008-2009) - Volkswagen voting share holders agree to accept 4 non-voting shares in exchange for 1 voting share - even if the market prices them as 2 non-voting shares per 1 voting share circa late 2007? That would explain both price spikes (rise in value of voting shares, fall in value of non-voting shares).

    And so, how would non-voting shareholders do in such a situation? They would be getting hosed there Timmy. They would be absorbing a stock dilution at non-market prices with the benefactors being the voting shareholders.

    I don't have any concrete information that tells me that is exactly what happened. And going by the chart provided, it wasn't long after that the prices of the two share types re-converged in late 2009. And so even if a share conversion took place in 2008, it was likely reversed in 2009.

  9. I wonder if there are any lessons looking at convertible bonds, or preferred shares?


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.