Mark Gritter (markgritter) wrote,
Mark Gritter

"Can't be taxed" and other stupid notions

Launch ran a breathlessly enthusiastic blog entry on Bitcoin today. A snippet:

The benefits of a currency like this:

a) Your coins can’t be frozen (like a Paypal account can be)
b) Your coins can’t be tracked
c) Your coins can’t be taxed
d) Transaction costs are extremely low (sorry credit card companies)

(C) is a category error (although not one promulgated by Bitcoin itself.) We don't pay sales and income taxes due to some inherent property of the USD. Law enforcement can go after unreported income no matter what form it takes. Cash is just as anonymous, yet paying cash to your employees and neglecting to withhold taxes gets you in just as much trouble. "Can't be taxed" is nonsense. If the transaction can be defined in law, a tax can be levied on it. A government could literally tax breathing. Once you admit that, you are just arguing about the visibility to tax authorities of monetary flows, and the scale of disobedience. And the authorities have had a lot of experience finding things.

(B) is less blatantly bad, but is by no means an absolute guarantee. Your transactions can't be tracked if you use a different keypair for each pair of transactions, and these keypairs are not recoverable, and your client securely erases all transaction information, and the public announcement protocol does not leak any information about your identity, and your counterparty does not keep track of any of this same information.

It's not clear to me that (D) is the case either, if the system actually scales to billions of dollars and Visa-like levels of transactions/sec. There is some handwaving on the Bitcoin wiki which tries to take a swag at the numbers. The transaction-fee description has some clever ideas, but there is not really a conclusive demonstration that the economics works out to a lower transaction fee. Let's see if it's reasonable:

In particular, in order to collect transaction fees you must solve a block and have the block accepted as part of the longest chain. You may decide not to include transactions that don't offer a transaction fee. Using the Bitcoin wiki's assumptions you would only need about 12 cores to keep up with the transaction flow, and some larger number if you actually wanted to win the race to create blocks.

Using Amazon EC3 as a rough guide to Capex/Opex, one double extra large reserved instance = $4000/3 year term plus $0.34 per hour, or $0.49/hour over the term. So it takes about $6/hour just to keep up in CPU cost. 1KB transaction * 2000/hour = 2MB/s, so network costs are trivial unless you have to make many copies or win a block and have to transmit a GB of data out quickly. Each outbound GB of data costs you about $0.15 on Amazon. So let's say about $24/hour to both handle transactions and have a decent chance of solving some blocks first.

It's very unlikely that a hypothetical VISA scale BitCoin network would have nodes which win a block every single day, at least, a healthy network would hopefully not have hash power concentrated so tightly into single miner nodes.

So, anybody who is competitive is paying $576/day for the privilege. If they want to net $100K per year they need to have gross transaction fees about $850/day. If they win just once per day then their income is based on a single block which has 2000 * 60 * 10 = 1,200,000 transactions. This is only $0.0007/transaction, which is indeed much lower than current costs.

Looks reasonable, but the costs are extremely suspect IMO and would bear a closer look. I suspect a Visa-scale data center requires something within an order of magnitude as Visa spends (even if you're not recording transactions persistently) which ends up more like $100K or $10K/day rather than $1K/day.
Tags: finance, geek, internet, rant
  • Post a new comment


    default userpic

    Your reply will be screened

    Your IP address will be recorded 

    When you submit the form an invisible reCAPTCHA check will be performed.
    You must follow the Privacy Policy and Google Terms of use.
  • 1 comment