The market value of the non-voting shares should theoretically be the value of the voting shares minus what the market believes the value of voting to be. If a stock buyback only buys voting shares and the price of shares in that class rises, such that the difference in prices of the share classes is greater than the value of voting rights, then an arbitrage opportunity exists. Thus the market will bring the price of the non-voting shares up, and the shareholders of that class will benefit too.
> If a stock buyback only buys voting shares and the price of shares in that class rises, such that the difference in prices of the share classes is greater than the value of voting rights, then an arbitrage opportunity exists.
I'd say the opposite is true. It's more likely non-voting share prices will crash because it's now clear to the market that voting rights are worth quite a lot given that they can be used to steal money from non-voting shares.
I think this assumes that the value of the voting rights is constant. In OP's example, the stock buyback only happens for voting shares, which should cause the market to update its estimate of the value of voting rights.
Good point! I don't really have enough (or any) economics knowledge to know how to model the value of voting shares in this case, but I think that there would also be downward pressure, for example as the voting power of company founders increases due to the reduction in number of voting shares.
