The bitcoin network will take exactly the same amount of energy mining an empty or full block. So a transaction does not use much energy (except for propagating it etc).
By holding bitcoin, you reduce the supply, thus increasing the price of bitcoin. By increasing the price, you increase the value of the block reward, thus making it more attractive to mine bitcoin. So actually holding it, not transacting, will increase energy consumption.
I've been doing a lot of reading & thinking on this lately as I just finished building https://carbon.fyi
Your take is 100% correct, but we still opted to use Ethereum Gas units as our "emissions factor" because it represents useage of and relative benefit from the Ethereum network.
We are interested in raising money for carbon offsets, so if we ask the question of "who pays?" it wouldn't be fair to place the burden on miners when all users benefit from the security and stability that comes from mining.
I'd eventually like to develop a model that allocates emissions based on "unit of time held" which is a more accurate allegory for 'benefit' for speculative assets like Bitcoin (if we are honest about Bitcoin's purpose), and because of the macroeconomic reasons you mentioned.
It's an interesting problem and discussion because unlike driving a car or eating meat, simply reducing consumption won't fix the problem, and yet we would all agree heavy users or benefactors of the network should bear more of the burden.
My understanding is that there is basically a fixed amount of energy being used (fluctuating mostly due to mining profitability) for the ~fixed block rate. How does adding transactions increase the energy use?
My understanding is that you pay money to the miners to get your transaction included, and the money you pay (on average) goes to pay for electricity, so the marginal difference of making a transaction is that your transaction pays more than whatever transaction would have otherwise been included, and ~100% of that extra money becomes burned electricity.
Energy use is proportional to the revenues gained from mining.
Revenues are based on the block coinbase reward and transaction fees from transactions (txns) included in the block. The coinbase reward is currently a large majority of mining revenue, but this will decrease over time according to the coinbase halving schedule every ~4 years.
Since block capacity is arbitrarily limited to 1MB, txn throughput is limited. Therefore, all txns contribute to txn fee competition. There is a transaction fee market where users must outbid each other to ensure their txns are incorporated into the blockchain in a timely manner.
The net effect is that txn demand does contribute to mining rewards and ultimately the energy used for mining. But it is a very small effect at present:
* Txn fee competition only rarely kicks in. The network operates most of the time without substantial mempool backlog. https://jochen-hoenicke.de/queue/#0,6m This does change with market activity, and we are certainly looking at more txn fees overall in the future.
* Coinbase rewards dominate mining revenues, so the effect is proportionally small. As these rewards decrease in the future, txn effects will have bigger effects.
The upshot of all of this is a shift into a 'mature' paradigm of blockchain security funding. It's a pretty big shift and easily the biggest unknown in the future of Bitcoin. More and more, the focus will be on scaling and whether the opposing goals of affordable transactions and miner revenue (and therefore network security) can be balanced. There are also interesting technical issues such as whether txn-fee dominated revenue will present new malicious actions. Namely, whether miners can withhold blocks to gain advantage, plaguing the network with reorgs and further centralizing mining.