There was an interesting question raised by @lkngtn here about how to handle AGPs that re-allocate or re-balance assets in the AA treasury. Although perhaps not applicable to the AGP idea proposed by the OP on that thread, since in my opinion his proposal could be handled without a Finance track AGP, other AGPs in the future might not share similar circumstances.
In the OP’s proposal linked above, the AA itself is proposing the portfolio re-allocation. Since the AA has the power and the explicit responsibility for treasury management without having to go through the AGP process for each decision, they can already make these decisions without an AGP. But what if someone outside of the AA is proposing the re-allocation, and they cannot convince the AA to make the re-allocation outside of the AGP process and so must pass an AGP to “force” the AA to do it?
The simple way I would interpret this is that a portfolio re-allocation has no net impact on the budget, because presumably the Finance AGP in question would have the AA exchange one asset for another of equal value. However this presumption isn’t always true, and this is where things can get complicated. There are questions that could be raised about whether the asset being exchanged for actually is equal in value to the original asset, whether it is equally liquid, etc.
The OP’s AGP, if it were actually presented as a Finance track AGP, presents just such a complicated case because there are good reasons to argue (some of which have been mentioned already on the OP’s thread) that the asset being exchanged for (ANJ) is not actually equal in value to the original asset (ANT). Even if the initial exchange rate is 1 ANT = 1 ANJ, the same is not true in reverse, and there are risks to holding activated ANJ that ANT does not have, the liquidity of the ANJ market will be different than that for ANT, ANJ carries different technology and adoption risks than ANT, etc.
First there is the challenge of valuing the ANJ at the time the Finance track AGP approving the exchange is processed. If this proposal were passed in ANV-5, the AA or a third party assessor would have a difficult time assigning a fair market value to a token that does not exist yet, that serves a market that barely exists, in a world with few if any precedents for this type of product.
Second there is the challenge of valuing ANJ once the curve is live and there is an active market. This will probably be much easier than the pre-launch valuation but still is not entirely straightforward due to the factors mentioned above (liquidity, technology risks, etc). Once the next Finance track AGP is passed in ANV-6 or beyond, the AA will have to assign a fair market value to their ANJ holdings in order to calculate the budget. Then the ANJ would be added to the total market value of the treasury converted to DAI and the budget would be calculated accordingly as usual.
Of course, the ANJ proposal is just used as an example here. You can substitute “ANT” and “ANJ” for any possible pairing or combination of assets being exchanged in a hypothetical portfolio re-balancing or re-allocation AGP. The question then becomes, is this kind of execution “good enough” to stay within the spirit of AGP-103? Or should we create another AGP specifically to handle portfolio allocation decisions by ANT holders? Personally, I think it’s good enough. It doesn’t perfectly handle some edge cases like the ANJ example above, but then I don’t think we can possibly cover every edge case with an AGP anyways other than to give the AA some flexibility, as they already do have, to handle edge cases ad hoc and in the best way they can at the time.