Subject: Re: SIRI
mungofitch:
This description is a bit of exaggeration, as the current strategy of each firm can make sense if the shares are dirt cheap (cheaper than fair price for a piece of a moribund underlying business), and the firms are not dead yet. But it shows the circularity issue. Even giant buybacks won't fix the issues of a ho-hum business.
I just thought of a way that "distributing" FCF by using it to buyback shares might make a material improvement in the earnings the remaining shareholders might achieve by comparison to just directly distribugint FCF to shareholders.
Consider three scenarios
1) Dividend : the entire FCF each year is distributed as a cash dividend to share holders. Berkshire never buys or sells any shares.
2) Buyback : the entire FCF each year is used by SIRI to do share buybacks at price = IV/share. Berkshire never buys or sells any shares but all other owners do so this goes on until Berkshire owns 100% of SIRI.
3) BerkDRIP : the entire FCF is distributed as a cash dividend to share holders. Berkshire DRIPS its dividends each year to buy shares from the other share owners at price = IV. The other owners will sell their shares to Berk at price = IV to enable Berk's DRIP but Berk never sells any shares so this goes on until Berk owns 100% of SIRI.
In Scenario 1), Berk recieves 37% of FCF cash each year forever. The other shareholders split up the remaining 63% of FCF forever.
In Scenario 2), Berk starts out receiving 37% of FCF, but each year Berk owns an increasing share of SIRI and receives an increasing share of each year's FCF, until eventually Berk owns 100% of SIRI. At that point, Berkshire is receiving the remaining FCF payments in cash as they are the only holder of the shares that SIRI will buy as part of its continuing buybacks.
Obviously (?) in 2) Berk recieves significantly more than in 1). The FCF per year generated by SIRI is the same in the two scnearios, but in 1) Berks gets a fixed 37% of it in 2) Berk's share rises above 37% more each year until it tops out at 100% of any and all further FCFs.
In Scenario 3), Berk starts out receiving 37% of FCF but then in each subsequent year receives increasingly more than 37% due to its additional DRIP share purchases each year. The other owners start out receiving 63% of FCF but receive increasingly less as they collectively own fewer shares each year as (some of them each year) have sold their shares to Berkshire. Eventually Berkshire buys the last outstanding share it does not already own and from that year forward Berskshire receives all the FCF in cash each year.
Remarkably, if what we have stated thus far is the only thing going on, I'm pretty sure Scenarios 2) and 3) produce the exact same cash flows to Berk as each other. In each case, BRK retains no cash until it has acquired 100% ownership of SIRI, after it which it receives 100% of SIRI FCF in perpetuity, or until SIRI liquidates, whichever comes first.
So we might say look at that! No difference in total payout to Berk with or without share repurchases! Mungofitch's intuitions about that were correct!
But the thing we have not considered is taxes.
In Scenario 2) Berkshire is taxed on its FCF dividend "income" from SIRI every year forever, even though for the first few decades it doesn't keep any of that income as cash since it has to DRIP it all until it owns 100% of SIRI before it starts keeping cash. even though it doesn't keep that income but instead uses it to increase its stake in SIRI.
In Scenario 3), Berk just owns appreciating stock for the first few decades until finally it is the only owner left of stock, at which point if SIRI continues doing stock buybacks it is buying the shares from Berk, upon which sales Berk will need to pay capital gains taxes.
If, as I am 98% confident is true, the total cash flows, ignoring taxes, from scenarios 2) and 3) to berkshire are identical, but the time-series of tax payments are different. But if in some sense the tax rate on cap gains rates and on dividends were the same, the the total amount of taxes paid in scenarios 2 and 3 would be the same. But in Scenario 2), it was paid years sooner than in Scenario 3).
So I think under the assumptions:
1) all stock purchases are made at a price = IV/share at the time the purchases are made
2) There are no taxes
Then the Dividend with DRIP scenario and the Buyback scenario are equally valuable to Berk, the last standing stockholder in SIRI.
But in the event that
2') Dividends are taxed at the same rate as capital gains, with the taxes due in the year the Dividend is received or the cap gain is realized
then the Dividen with DRIP scenario is inferior in that, while the same amount of taxes are ultimately paid in each, the taxes are paid years earlier in the Dividend with DRIP scenario than in the buyback scenario.
So in Summary:
Buybacks result in greater deferral of tax liability than do Dividends with DRIP, even though both produce the same cashflow to the ultimate owner (Berk) and both produce the same (not discounted for time) tax liability.
R:)
******
ChatGPT 5 comments on this post:
You’ve basically built a very clean thought experiment here that strips away most of the usual noise in the buyback-vs-dividend debate, and it’s elegant because:
You hold all other factors constant — FCF is the same, purchases are always at IV/share, no change in business quality.
You clearly isolate the only differences as:
Who sells and who doesn’t.
The timing of tax liabilities.
Why scenarios 2 and 3 are the same before taxes
If every purchase is at IV/share and Berkshire never sells until it owns 100%:
Buyback scenario (2): SIRI retires shares from others; Berkshire’s ownership percentage rises automatically.
Dividend + DRIP scenario (3): Berkshire takes its pro-rata dividend and uses it to buy shares from others; Berkshire’s ownership percentage rises identically.
Mathematically, the share count shrinkage and Berkshire’s fractional ownership path are identical, so the stream of pre-tax cash flows to Berkshire is identical — just all of it gets retained in the form of a growing % stake until the very end, when it’s 100% and cash flows begin.
That’s exactly what Munger/Mungofitch have often pointed out — ignoring taxes, buybacks and “you buying more shares with your own dividend” are the same mechanism.
Where they differ: tax timing
Your assumptions:
Scenario 2 (Buybacks): Berkshire receives no dividends, so no annual dividend tax. Gains accrue inside the appreciated stock. Tax only when selling (or never, if held forever or stepped up at death).
Scenario 3 (Dividend + DRIP): Berkshire is taxed on the dividend every year, even though it doesn’t retain the cash — it just reinvests in stock. The capital gain on the DRIP purchases will eventually be taxed too, but the dividend tax hits immediately each year.
If dividend and capital gains tax rates are the same:
Total taxes paid over the full horizon are the same in nominal terms.
But the present value of taxes paid is worse for the DRIP scenario, because you pay earlier.
Buybacks are a built-in tax deferral mechanism.
Implications in the real world
For long-term holders who don’t need cash now, buybacks are usually more tax-efficient than dividends, even if the two are “economically identical” in pre-tax terms.
If you expect tax rates to fall in the future, deferring recognition (buybacks) becomes even more advantageous.
If you expect tax rates to rise, the opposite could be true — locking in the lower rate earlier could be beneficial.
In reality, buybacks rarely happen exactly at IV/share — so in practice, buybacks can be value-adding or value-destroying depending on price discipline, whereas a dividend at least distributes face-value cash.