Is BitCoin Arbitrage Feasible?

BitCoin has been in the news lately with its rapid rise in exchange value, its huge fluctuations in intra-day value, and the susceptibility of services using it to hacking attacks.

It should be obvious to any observer that a position (in the investment term, not opinion) in BTC is speculative in nature, and carries any number of risks that are hard to evaluate.

There may be a way for investors to make money on BTC through arbitrage though – with relatively well-defined and calculable risks.

There are literally dozens of BTC exchanges that allow users to trade BitCoins for fiat currencies. Of them, there are a handful of large ones that are relatively liquid, transparent and secure (although occasionally subject to DDOS attacks), and a number of smaller ones that are infrequently traded and which may or may not be legitimate businesses. Many of them publish their open trade books, so it is possible to make some sort of minimalistic determination of sentiment on them. And here’s the important part – there are often large differences in valuations for short periods of time between exchanges.

Here’s the list of potential issues with arbitrage:

  • The process of moving fiat currencies in and out of these exchanges is slow (in my experience it can take days).
  • There are transactional costs associated with moving money in and out, in addition to exchange fees. There are also built-in transactional costs associated with moving BitCoins around, although these are relatively small.
  • The above implies that an arbitrageur would need to keep fiat currency balances on hand with several exchanges in order to exploit differences in exchange rate.
  • That in turn implies a good knowledge of whether the exchanges in question are legit, sufficiently liquid, secure etc.
  • There are also limits on trade and balance size at most legitimate exchanges, in addition to physical market size limitations (the entire market for BTC is around $1.5 billion right now, which is small from a financial transaction point of view) that likely place an upwards cap on the amount of BTC that can be traded at any given time – and thus a cap on potential profits.
  • The BTC exchange market can move extremely rapidly, so each trade has some degree of risk involved as well.
  • There are other hidden risks – potential for government intervention etc.

It should be possible to put together a database model that appropriately defines risks and necessary spreads, given the above.

To illustrate why somebody might want to do this though, this morning the following numbers applied (within ranges) for several hours:

  • MtGox traded at between $135 and $144 CAD
  • LibertyBit traded around $149 to $150
  • Virtex was in a range between $130 and $140

Careful timing could have resulted in an arbitrage gross difference (excluding transactional costs) of about $15 CAD (or roughly 10% on a single trade).

Please note: I’m not advocating any of the above exchanges, and therefore have not included links (google them yourself if needed, and use at your own risk!)

Quick Clarification: I’ve seen plenty of posts on bulletin boards from opportunistic traders claiming to have “done this last week, and made lots of money”. That isn’t what this post is about. My question (which I don’t have an answer for) is whether a company (or skilled individual) can do this on a regular basis, profitably, while still keeping track of VAR (value at risk) and other risk models.

Update (14 April 2013):

I’ve been playing around for a couple of weeks trying to find practical arbitrage strategies. The following points may be of use to somebody attempting to do this:

  • Moving money between exchanges is very slow (up to a week in some cases), and moving BTC is relatively slow (order of hours). If you are relying on a particular spread, and you are moving currency between exchanges, there’s a very high likelihood of things moving against you in the process, particularly given recent volatility.
  • The way to do this is to keep balances of both money and BTC at a variety of exchanges, and to occasionally perform “balancing” operations.
  • Many exchanges have relatively small maximum allowed balances due to “know your customer” regulations. Virtex, for example restricts you to $40k CAD. MtGox has a maximum balance of $1 million USD, but then the question is to what extent you trust them with that amount of money – reliability (i.e. frequent downtime), hacking attacks, and slowness in withdrawing large balances have all been reported in the past. Similar issues apply to other exchanges as well.
  • What this means in practice is that very small and very large customers won’t be able to do this to best effect (at least until large and reliable exchanges come online, which may in turn reduce the profitability of arbitrage). My best guess is that the lower limit, practically, is a few thousand dollars (split between 3-5 exchanges), and an upper limit of around a million.
  • The good news is that profitable arb opportunities are very common (many times per day), so even if one can only use a small percentage of their total balance each time, a dedicated trader can perform transactions on a regular basis.
  • I found a chart that looks for arbitrage opportunities automatically. I haven’t tested its hypotheses, so use with caution. It also only have a few exchanges listed. You can find it here:
  • One last point – don’t get too greedy trying to maximize profit on a transaction. Since the price of BTC can move quickly, if you try to get the absolute best rate at each end, there’s a good chance that your transaction won’t complete as the rate moves away from you. The point here is to trade with the minimum risk possible, in a very risky market.

One response on “Is BitCoin Arbitrage Feasible?