Skip to content →

Bitcoin volatility will match major fiat currencies by 2019

Bitcoin was created in 2009 by a mysterious character who claimed it to be a payments network. But unlike most other payment networks like PayPal and Visa, it screwed with our minds by having its own token. A token that had a price that floated against other currencies. In basic terms, this means if you fund a bitcoin wallet to buy something, it may be worth less (or more) by the time you come around to spending it.

This is the story of bitcoin volatility, we’ll be studying its personality over its short and notorious history. It was partly inspired by Vinny Lingham who calculated Bitcoin will achieve the necessary price stability to be a store of value at $3000 per coin (~$50b market cap), and estimated that to be two year away. We shall see if the data backs this up.

We’ll start this journey with a bit of eye candy. Let me plot for you the volatility of 600 cryptocurrencies against their market caps and 24 hour traded volume (i.e. liquidity). Volatility is represented by the size of their circle. Okay circles, I want you to be small and towards the top, got it? (This equals low volatility and high liquidity).

The Volatility of 600 cryptocurrencies

As it turns out it was a weak correlation between market cap and volatility. Apart from looking real nice, it showed just how far ahead bitcoin is over the other coins. Pundits should know I used log scales and exponentially scaled circles to reduce the exaggerated differences here.

Bitcoin’s volatility over time

Onwards to the real focus of this study… Dearest Bitcoin, how long will it take for you to be as stable as fiat currencies?

To answer this, I collated 5 years of bitcoin price data against USD (BTCUSD) and compared it to EURUSD and NZDUSD currency pairs to produce this very telling graph.

Bitcoin Volatility vs Fiat Currencies

[Live Chart]

First some details about this graph:

  • Circles are proportional to bitcoin’s 24 hour traded volume, I wanted a visual representation of bitcoin’s growth as a traded currency
  • I chose EURUSD volatility as these two are the ballers in the room with M1 money supplies of 7.5 and 3.3 trillion respectively. The big guys should be the most stable
  • NZDUSD volatility is the small guy in the room, it exhibits the highest volatility of the commonly traded FOREX pairs
  • Trendlines have been calculated using exponential regression analysis (i.e. it was done with numbers, I just didn’t eyeball it)
  • 60-day volatility are used in these plots.

What we are seeing is bitcoin’s peak volatility is reducing steadily and will enter the realms of fiat currency (below 5.5%) by around July 2019. I found this surprising.

For a currency with a tiny market cap of $10b, compared to say the Euro’s M1 money supply of $7.5t, bitcoin’s price stability is ridiculously good.

Investor conclusions

The economic properties needed for bitcoin to go mainstream are developing quickly. If we take fiat-level FOREX volatility as a level in which the public find acceptable (this is not necessarily true), we are less than 2 years out for the start of prime time “bitcoin as payments” heaven. For now these are my investor related conclusions:

  • Price stability gives Bitcoin even more bullishness
    Price stability is needed before we get consumers buying and holding bitcoins for short and medium term spending. When this happens, it’ll have a large and positive impact on price. The next 2 years of bitcoin will be bullish in this regard. Buying and holding bitcoin for native payments has a very different economic fingerprint than say via bitcoin as remittances, where the receiver of funds immediately sells to move back to fiat currency which is more stable. Remittances use bitcoin as a pass through token, or in other words it’s price neutral as remittance adoption increases.
  • Payment ventures/projects are too early
    Ventures that focus on bitcoin as a form of payment are premature – examples include payment gateways like BitPay and cryptocurrency point-of-sales ventures Plutus and BlockPay. We are probably at least 2 years out from the necessary price volatility necessary for this sector to be ready. For now bitcoin HODLers are speculative investors. If I was a venture capitalist or an alt-coin ICO investor, I’d be steering clear for a couple more years.
  • Keep away from alternative payment coins
    Given bitcoin has by far the largest network effects and an exponential head start on stability and liquidity, I would say any other payment alt-coin is going to have a hard time competing, especially in 2 years and we get a critical mass of sorts on the bitcoin network. Examples include the privacy coins and payment coins like ZCash, Monero, Dash, and other less known coins too many to list.


FOOTNOTE: As part of this study, I found the bitcoin volatility information most people refer to, namely btcvol.info looks wrong. My volatility calculations put volatility around ~5x higher than this site. To double check my calculations my curves have been confirmed by comparing to the more official tradingview.com charts for FOREX.

I have a lot of brain-farts on crypto-currency. For more farting, follow me on Twitter @dangermouse117

BTC Tips: 1KrYzd8y6gnJekibZpb4ixif4nJYbaaZct

Published in Uncategorized

13 Comments

  1. Antonio Antonio

    Hello Willy,

    First of all congrats for the reports!.

    I am a Bitcoin maximalist. I would like you to propose another topic to research on, but first we have to agree on the main assumptions in order to have a correct and realistic investment tool.

    Introduction:

    The model I would like to build is a investment ratio that clarifies if the bitcoin price is “expensive” o “cheap”. It would be like the PER -price earnings ratio-, namely, the ratio of a company’s stock price to the company’s earnings per share or, as I like to think, the number of years it will take the company to get my investment back assuming the interest remains constant.

    There is no such a ratio in Bitcoin, or at least, I did not find any accurate ratio that represents the price against the current inflation, there is just an approach through the miners revenue, which is quite enlighten, but we have no figure and I cannot model the average and compare it with another events easily.

    Assumptions:

    * Bitcoin maximalist approach -there is no true competitor-.
    * Short term market irrationality vs long term rational value -as per Austrian School-
    * Highly deflationary system.

    Ratio

    PER represents the number the number of years it will take the company to get my investment back assuming the interest remains constant. Thinking in this way helps to allocate investments; when the PER is high discourages from buying the stock as there are peers with a much better return on investment and in the long term there is a stable PER quantity -historically between 10 and 20 and it should growth when the productivity is increased-.

    In the other hand, I believe there is an equilibrium in the bitcoin price based on the returns of the bitcoin miners: if the price is high more companies are going to start mining resulting in lower return in the long term and finally making a part of them unprofitable; if the price is low means there are just a few profitable companies mining that will get most of the earnings resulting in a selfish control of the supply that would finally drive to higher prices here, at this point, you’ve got what it is the economic bitcoin cycle.

    There are of course many other factors that we should take into account, like development of the ecosystem, increase of demand, outflow of capital caused by events, etc.. but at the end, I believe there is a kind of equilibria like in the stock market.

    When I look at the chart of miners revenue, somehow you can intuit the equilibrium and can perfectly guess the miners thoughts: bitcoin price maximization, namely, cover costs, get rid of a couple of extra bitcoins in order to diversify the risk and hoard the rest till the price is high enough).

    https://blockchain.info/en/charts/miners-revenue?timespan=all

    The ratio should be something like the [price] divided by [money earned per block plus transaction fees].

    After that, we could research on the evolution of miners in order to check if the bitcoin price had actually influenced the bitcoin mining sector according to our theory, but this relation should no be also considered as the causal relationship. Not yet.

    What do you think? Does it make sense at all? Do I forget something important that makes the model inaccurate?
    I really appreciate your feedback and comments!

    • It’s probably worth plotting to see if miners payout / market cap revolves around a steady baseline %. I’ll do that out of sheer interest. Thanks.

      • Antonio Redondas Antonio Redondas

        great! It is going to be interesting. We’ll see if there is any correlation between number of miners, miners revenue and bitcoin price!.

        • Actually it’s a really boring graph, I should have seen it coming. It’s just a stepwise plot of bitcoin inflation rate which halves every 4 years. Miners revenue is predominantly block reward as fees are pretty small as a proportion. So essentially it’s a ratio of 50BTC divided by total BTC in existence (regardless of USD price), then 25, then 12.5. There’s a downward slope in the first step as the BTC in existence is changing in proportionally significant amounts. So it will continue to trend downwards until fees become a significant factor. This is more of a plot of Bitcoin security, the cost of attacking the network vs the size of market cap that can be compromised.

          • Antonio Antonio

            Noted. Thanks for the feedback

  2. Really interesting analysis! Just a couple of points for further consideration:
    1. You are expressing volatility as a function of the size of the money supply. Algebra reveals that these two variables are strongly correlated. This auto correlation is a key reason your first graph exhibits classic heteroskedacity. You can remove some of the heteroskedacity by changing the log scale to a natural log, square root, or a few other y-variable transformations.
    2. There are other major variables that contribute to currency-pair volitality such as fiscal policy, geo/political/social shocks, change in the price of other asset classes, change in GDP, etc. It would be interesting to do correlation analysis on volitality for each of these factors to determine the relative importance of each of those variables.

    Finally, you may find it interesting to review the Quantity Theory of Money. https://youtu.be/uZH6EWclmYM It shows that the changes in price (volatility) are directly proportional to the changes in the money supply. That assumes that the velocity of money is constant (something that I tend to question).

    • 1. I found a very low correlation between volatility and market cap, so didn’t publish that graph as it doesn’t say much. So that first graph really shows the correlation between traded volume vs market cap. So this graph says liquidity tends to hover around a standard value per market. And taken together it shows is how big each respective market is on the small (bottom left) to big (top right) spectrum from which you can gauge the volatility with bubble size.

      2. The second graph is obviously just a numerical plot over time while the market was subject to bubbles and pops, market uncertainties and so forth. So it was good to see what the peak volatility was with natural forces at play. I do think it would be telling to measure all fiat currencies and Bitcoin against a basket of global currencies to provide a better baseline. e.g. when BTCUSD volatility gets really low in a few years, it could be measuring the movement in USD

      Fundamentally, stability of a currency is based on the depth and velocity of orders at the exchanges. Your variables (e.g. political and fiscal policies) will impact trader behaviour at the exchanges. Taken together, since Bitcoin is a global currency which will hit the exchanges with an order whenever a merchant makes a sale and converts back to their base fiat currency, it becomes clear the order books for Bitcoin in due time will be orders of magnitude larger than FOREX which only reflects international trade, and doesn’t pick up domestic trade. This is the reason why I think BTC stability is so high given it’s tiny market cap. This line of thought suggests BTC has a shot of being the most stable currency the world has yet seen.

      • Let me take a step backwards as I can use some enlightenment.

        I’m confused why the Bitcoin community uses the term “market cap”. A currency is fundamentally different than a stock which is purchased with a single currency.

        When measuring the size of the Bitcoin market, why doesn’t the community multiply the Bitcoin M2 by the velocity (average number of times a Bitcoin unit changes hands per year). Bitcoin’s M2 is known and its velocity can be calculated from the blockchain. The resulting number would then be expressed in Bitcoins not in US dollars.

        Related to this, I understand how liquidity factors into volitility, but I cannot see how “market cap” does…other than the simple fact that that a big market is generally more liquid than a tiny market. If you measure volitility using market cap, then you have the problem in which the change of price of a Bitcoin affects the market capitalization by proportionally the same amount. Thus both “independent variables” are closely linked. This results in what’s called auto-correlation and is the reason you have such strong heteroskedacity in your first graph.

        Third, currency volatility is a measure (in standard deviations) of the changes in price with respect to another currency. This is very different from a stock price that, by definition, is expressed in a single currency. There is no such thing as “the USD volitility”. There’s USD-GBP volitility, USD-EUR volitility, etc.. Why doesn’t the Bitcoin community follow this convention?

        • It’s an interesting direction you’re raising Ian. Coming from an economist’s viewpoint, I do think measuring the left side of the Quantity Theory of Money equation could be more useful in future as Bitcoin becomes adopted into real economies. But for now, in this stage of the game, I think market cap is probably the most useful expression as just about everyone uses Bitcoin (99% at a guess) for trading/price speculation, investment and currency hedging, very similar to stocks.

          Also measuring the velocity of Bitcoin may be impossible to ascertain from blockchain analysis as there are many issues that cannot be easily resolved:

          1) Accounting for opaquely conducted off-blockchain transactions, such as Coinbase to Coinbase wallet payments

          2) Bitcoin mixers will skew the data. This is where many txs get generated just for the purpose of obfuscating the money trial.

          3) If I assume velocity should only measure transactions conducted for a purchase of goods and services (right side of the equation), then consider that a large proportion of transactions are movements between a single owner’s addresses. e.g. Hot wallet to hardware wallet.

          4) Digging into what a transaction is, there’s input fragments and output fragments. Some of those outputs go back to sender as change, in HD wallets which are very common, those addresses are different from the originating address so it’s impossible to determine with accuracy what the real transaction value is.

          On your last point on measuring volatility, for the purposes of this study, I expressed BTCUSD volatility, but as mentioned, I think it would be more useful to measure each currency and BTC against an index of currencies. Do you think SDRs would suffice?

          • Here’s my analysis: http://asymmetric.info/wp-content/uploads/2016/11/Bitcoin-Volatility.pdf

            It forecasts volatility as a function of transaction volume from data I got on blockchain.info. As I stated above, I don’t think it makes sense to use Market Cap as a forecasting variable for volatility because Market Cap is computed using price…and that same price is also used to compute volatility. In other words volatility and market cap are both dependent on price, and this leads to high autocorrelation.

            • Hey Ian, if you read the article carefully, no correlation to market cap is used in my analysis, I found no correlation in that data with market cap, but Vinny Lingham used BTC price in his model which he alluded to but never published. My study was just done with raw volatility trend. The results were 2019 being where peak volatility of BTC drops below peak fiat pairs, and you can eyeball where average BTC volatility matches average fiat pair volatility, from which you can deduce it is around 2021 compared to your 2024. The discrepancy can be explained by my analysis being contiguous 60-day volatility calculated daily using computer processes (2200 data points) while your analysis was a yearly snapshot based resulting in 6 data points (one per year) to ascertain the trend. I like how you correlated volatility to transaction volume, which is a great graph to plot. Unfortunately tx volume for the last year has been skewed as the Bitcoin network has topped out at full capacity.

              • Ian Berman Ian Berman

                Actually, I used 6 years of daily data but computed the daily volatility ℅ myself instead of using 60-day continuous. The 6 histograms were made using all 2100 data points.

                • LOL. Not to nitpick but yes I saw that. Each one of my 2200 data points consisted of 60 underlying data points to generate 60-day volatility, similar to your 365 data points to generate each of your 6 yearly volatility figures. The higher resolution should reduce sampling errors in the trend line.

Leave a Reply

Your email address will not be published. Required fields are marked *