I’m not normally one to delve too deeply into the world of financial charts, math and indices – oh I can read all variety of charts just fine and make arbitrary speculative guesses with the best of them, but I’ve seldom considered myself clever enough to make any attempt at inventing my own. The standard investment data has been around for ages, after all, and many were thought up by Nobel laureates so what kind of hubris would I need to even make an attempt? I recently had a realization that changes that viewpoint: None of those Nobel laureates had the kind of perfect and transparent view of a market that Bitcoin provides so maybe my observations, no matter how obvious they seem, weren’t implemented because the data wasn’t available, rather than because they’re silly or wrong…
In a recent post, I mused that the primary driver of Bitcoin’s value is the network effect – the more people and businesses we have participating in this thing, the more utility it represents and therefore the higher its inherent value. There’s another thing that can raise Bitcoin’s value, though, and it’s not a good thing: speculation. Now speculation in small doses is fine, we all do a bit of it here and there, it can be fun and profitable and indeed active exchanges and folks to arbitrage between them are important parts of any market – too much speculation, however, and you get what we had last summer.
For those of you who weren’t around or somehow missed the memo, in June of 2011 Bitcoin experienced a massive economic bubble. Rampant speculation combined with reckless optimism ratcheted up the price of a bitcoin to just over $30 USD before the bubble finally popped and the price drifted back down and remained stable around $5 for quite some time. Today, one bitcoin will cost you around $11 USD, more than double that stable post-bubble price point. The run up to $11 has been much slower and stabler than the run to $31 but that’s no guarantee at all since most economic bubbles happen much slower than the one we saw last summer. How can we even attempt to measure whether a change in value is genuine or not?
Well, Bitcoin is trying to be a currency, right? So let’s look at what a healthy currency market looks like. If I look at how the average U.S. Dollar gets spent, I can see that some fraction of them are flying around exchanges being traded for other currencies but I can also see that the vast majority of the dollars that move in a given day are being spent by people, moving around the everyday economy that we all participate in. In other words, the ratio of transaction dollars to trade dollars is quite high.
It’s also important to note that looking at one piece of data without the other would be a bad idea. Speculators have to move bitcoins around too, between exchanges for arbitrage or across wallets or services to get into or out of the market – all mundane-looking transactions that we’ve got no way of filtering out. So let’s stick to that first idea and look at the ratio. Bitcoin, unlike any other currency that’s ever existed, is uniquely able to provide every last snippet of data we need at whatever resolution we’d like. One Python script and a Blockchain.info API later I had comma-separated data ready for Excel-powered analysis.
Above is a chart showing Bitcoin price in U.S. Dollars (blue) and transaction volume divided by trade volume (red). Sadly, the span of time I’m most interested in – the very beginning of the speculative bubble – is a pit of useless data. With daily transaction and trade volumes often under $1,000 USD the ratios were little more than random data. One person spending $100 on an exchange could completely destroy the ratios. I also noted there were a couple of large spikes in transaction volume that, to my knowledge, have never been explained fully but are mostly due to the movements of a single large balance across multiple accounts. Ignoring the early noise and obvious outliers, some useful data finally emerges.
Little more than a glance shows that the tx/trade ratio sat stagnant through the entire bubble – the Bitcoin economy, at least the non-speculative bits of it, didn’t appear to grow at all. Equally interesting, the long slow run-up to $11 we’ve seen in recent days has in fact been accompanied by a slow and steady increase of that same ratio, suggesting that the non-speculative economy has actually grown at a commensurate pace with price and that the current price may well be perfectly reasonable.
So is this simple piece of math, which I’ve decided to call the “Speculation Index” the magical predictor we’ve all been waiting for? Hardly. I think it provides some useful data that we were not only missing before, but which may not even be possible to calculate in any traditional currency market – and that data is certainly worth a glance. Thus far there seems to be a clear pattern emerging from the data, but it could just as easily be that I’ve stumbled upon an arbitrary calculation that fits historical data nicely and is meaningless moving forward. Only time will tell.