[jeremy_arnold]

OP here. Can give some partial answers.

1. Lots of possible deductions. Charitable giving is a big one here. I guess we'll see from future instalments in the ProPublica series which ones were specifically used.

2. You can roll-over loans by borrowing new money to pay back old debt, up to a practical cap of a max % of collateralized share value. And this is worth it if the expected growth on your shares is higher than interest on the loans. But at some point you or an heir will have to realize gains if you want to exceed that practical cap -- or if the loans get called. Bezos could probably raise $2bn for a sports franchise on loan no problem coz that's only ~1% of his shares. If you're a regular billionaire with say $4-5bn in shares, much harder.

The meta point here though is that the taxes are always going to be paid at some point, and deferring that point to the future is fine if the rate of growth on the gov's share exceeds their borrowing costs (currently between 0-2% depending how you look at it).

[dwighttk]

Haven’t there been tax holidays in the past where gains can be realized at a much lower rate so with this level of wealth it is worth putting it off just in case one of those rolls around again?

(Or is that just an urban legend I’ve been told?)

[jeremy_arnold]

I know they happen for offshore wealth getting repatriated. Possible that they've existed for onshore too? Interesting thing to look into. But even if so, that just seems a vanilla case of "Congress should be held accountable to not give away the farm", as it would be a pretty perverse incentive if the rich felt that such a holiday were likely to arise.

[midasuni]

The rich can pay for such a holiday

[wmf]

Yeah, there's a definite asymmetry between billionaires with 30-50 year horizons and politicians worrying about 4-8 years.

[Tehdasi]

https://www.nytimes.com/2018/01/17/technology/apple-tax-bill...

Not for individuals, but when you are a high net worth individual, it's all the same.

[bradleyjg]

You criticize PP for shallow analysis but your analysis of GRATs, GRUTs, and similar estate tax avoidance strategies is even more superficial and conclusory.

[jeremy_arnold]

Most of those instruments retain the initial cost basis. They're a way of getting around gift taxes, not capital gains.

[bradleyjg]

In grantor trusts, transactions between the grantor and the trust are disregarded. So at the end of the trust term the grantor can buy back the appreciated assets and substitute cash in the trust vehicle. That cash (assuming a zeroed out GRAT) represents the delta between the conservative IRS assumed rate and the actual appreciation, passes to heirs gift tax free, and with no embedded capital gains basis.

Again, if you are going to throw punches you ought to make really sure you’ve done your own homework.

[jeremy_arnold]

We'll start with the meta here. Some of these vehicles have arcane applications that I wouldn't know about. I'm not a CPA, and I pay people for corrections for a reason (which I'm pretty sure I'm about the only person on the internet who does this). If you can demonstrate to me that I've missed something meaningful here, I'm happy to pay off and revise/note my piece accordingly.

As for the concrete, the important thing here is whether those repurchased shares retain their original cost basis or not (to the grantor) from the IRS's POV. If they do, all that's happened is a complicated tax-free cash transfer. If they don't, then cap gains taxes have been avoided.

So far as I understand the rules, it's the former. If I'm wrong, I'm happy to be pointed to a credible source etc.