Sure. I have seen former CPAs grapple with this. I am not a CPA and have no desire to become one but here is my understanding as an investor. I'll do this in outline because nobody is paying me to write a dissertation.
The GAAP treatment of a Model S sale with a RVG (as I understand it to be).
Background: Tesla transfers full title of a vehicle to a customer. Customer arranges full cash to be received by Tesla. Tesla promises a conditional repurchase of nominally 50% of the purchase price between 36-39 months thereafter if the customer, without obligation, wants to sell it, and if the customer wants to sell it to Tesla and not to a third party. At the end of 39 months, the offer to repurchase expires worthless. Tesla retains no right or title to the vehicle at any time including no right to demand that the vehicle be returned to Tesla.
GAAP places an overwhelming priority on the Residual Value Guarantee and treats 50% of the sale as a 36 month lease and 50% of the sale as never having occurred.
Because the customer has paid 100% Tesla sets up a GAAP lease account for 50% which is then recognized on a straight line bases over 36 months. As for the RVG (of nominally 50% of the original purchase price). Tesla sets up a liability to the customer for the other 50% as though the customer had lent Tesla the money for something he as not yet fully agreed to buy.
At 36 months, the straight line revenue recognition eventually adds up to 50% of the purchase price and then the question comes (under GAAP) of whether or not the *customer* will buy the other 50%, completely reversing the sense of the actual transaction in prospect. If the customer does not decide to sell the car to Tesla then GAAP assumes the customer just bought the 50%. The 50% RVG liability is extinguished and recognized as sales revenue. On the other hand if the customer decides to sell the car to Tesla then the 50% liability is extinguished in cash to the customer (with first creditor being his account at his lending institution) and the Operating Lease asset representing the car is transferred to Inventory.
While all of this is going on, at the point of "sale", 100% of the car is effectively bought by Tesla at full retail price (according to GAAP) and not the customer i.e. no wonder it's hard to make a GAAP profit. These phantom purchases end up on the Operating Lease ledger in the Asset column. This is despite the fact that Tesla does not legally own any of these assets, 1.36 $billion of them at last count, the customers legally own these. As for COGS. 50% of the COGS is metered out in a straight line over 36 months alongside the phantom lease payments. The other 50% of COGS relates to the RVG half of the car which according to GAAP Tesla never sold. This is why GAAP Gross Margins are approximately the same as non-GAAP. The big difference is that GAAP pretends that Tesla bought the car at full retail to place in a lease fleet, instead of the customer buying the car (which actually happens, as customers will attest). The fractional difference in GM between non-GAAP and GAAP relates to the treatment of the warranty reserve which is fully recognized for the full sale under non-GAAP and only relates to half the sale under GAAP and that metered out over 36 months.
"how [does] this affect[] the quarterly income statement, balance sheet and statement of cash flows"
Depending on you frame of reference. From a non-GAAP / management accounts frame of reference (the view from the business and the logical Investor's perspective)
A. GAAP treatement of RVGs trashes the quarterly income statement by pretending Tesla is buying its own cars for a lease fleet instead of selling them to customers, which except for a relatively trivial quantity of real leases it isn't doing.
B. GAAP strongly misrepresents the balance sheet by pretending that Tesla has a huge chunk of Operating Lease Vehicles in the Assets column that it absolutely does not own.
C. It has no effect on the statement of cash flows.
From a blind acceptance of GAAP frame of reference (bearish Tesla non-GAAP equals bogus overstatement of performance, Enron blah blah etc):
A. GAAP strongly and misleadingly understates quarterly revenue and profits i.e. the performance of the business luring shorts and competitors to systematic underestimation of the health of the Tesla business.
B. GAAP overestimates Assets relating to non-existent Operating Lease Vehicles which is prone to be conflated with rising inventories, a bearish red flag in the traditional automotive industry sector.
C. It has no effect on cash flows, excepting widespread layman's confusion that GAAP profits and positive cash flows are one and the same thing.
Footnote for the sake of completeness. Tesla's non-GAAP reverses out non-cash stock compensation. This is relatively standard practice for non-GAAP accounts and I personally believe it is appropriate because as a shareholder I have already paid for non-cash stock compensation in the dilution of the shares I hold. I don't hold a diluted position in a company that has also been impaired by an equal amount. It's one or the other. GAAP lease accounting for RVGs is the major concern for the integrity of GAAP in the case of Tesla. This is the thing that can lead people to imagine that Tesla loses more money the more it sells which at 25% GM is obvious nonsense - but in terms of GAAP revenues, it is equally obviously true that the more Tesla sells but is obliged to account for as buying a lease fleet rather than selling cars, then of course on face value GAAP revenues and profits will fall the more it sells.