Subject: Re: Strategy - covered calls in retirement
My modest proposal: just pay out $100b in dividends, suspending the repurchases, begun when the PE ratio was less than 10.

That would be rational, but of course wildly unpopular.
First, the dividend would be a big tax nuisance for a whole lot of people.
(If I held Apple stock directly, I'd probably sell it before the ex-date and buy it back after)

And of course the rational dividend policy that you outline (and which I endorse) is exactly the opposite of what people want:
they want solid dividends even more when the price is low, as won't feel like selling any shares and are generally feeling poor.

A compromise which might be pretty rational capital allocation AND make people somewhat happy:
When the valuations are high, let the cash stack up.
Next time the valuations are low, blow half of it on buybacks (tender offers are nice) and half of it on a special dividend.

The problem, as always, is nobody ever knows when valuations will get low again, nor how low they might get.
I guess you could at least try a trigger rule like "cheapest in the last 4 years" on a couple of metrics.

As a vicarious shareholder of Apple by virtue of Berkshire, I would much prefer seeing a big dividend than more share repurchases of a company
that is now trading for 33 times earnings with an annual growth rate of 15% over the last 3 years, a growth rate that is, in my opinion, likely to be hard to maintain even at that level.

Indeed.
I am forever trying to get my head around the circularity of the problem caused by the intersection of large buybacks and market valuations.

Imagine a fabulous underlying business wrapped in a corporate shell which is very cash rich, more cash coming in than the great underlying business activity can deploy.
They should do buybacks: it's a fabulous underlying operating business, and they don't need the cash.
But let's say the shares are trading at a high multiple of visible profits and ostensible value.
If the multiples are high, then the very same underlying business doing buybacks is not as high quality at the overall corporation level:
Among other activities, they are allocating lots of capital at high multiples of current earnings (a low ROIIC),
into a company that isn't all that great specifically BECAUSE they are allocating lots of capital at high multiples of earnings in a business that isn't all that great specifically BECAUSE they are allocating lots of capital at high multiples of earnings in a business that...etc.
So the firm doesn't really deserve the high multiples it's getting, and the buybacks aren't all that smart after all.
The more you expect a firm to do big future buybacks at high valuation levels, the dumber buybacks are at any given high valuation level.

Yet if they were trading at low multiples, this same feedback loop makes the buybacks smarter and smarter and the business more and more valuable, and therefore deserving of higher multiples...
until they get them.

Bottom line, I guess a long run policy of large buybacks even through periods of high valuations makes sense (to an extent) only if you don't expect the high valuation levels to be permanent.
The lower the future valuations (because of future buybacks), the more sense it makes to be a buyer at current high valuations, and vice versa.
If I am going to be an indirect holder of a huge Apple position in the next decade, I sure hope the share price is horrible most of the time.

Jim