Simple forecasts of bitcoin prices
One of the most difficult thing for any economist is forecasting. As I deal with bitcoin for a couple of years, here below is one graph showing simple predictions. Simple, informative, speculative…
Recently, I browsed my predictions of bitcoin prices which I did a couple of years ago. I was surprised by my forecasts made in April 2014, which showed that the price of bitcoin can reach up to:
- $9923 if we assume that bitcoin would be as big as the Madoff's Ponzi-scheme,
- $6450 if bitcoin market cap would be 1% of joint capitalization of US dollar (M0 aggregate) and euro (M1),
- $4500 if bitcoin reaches 1% of share of gold market.
At that time bitcoin cost $450 only. I was not brave enough to publish the results, as they seemed for me to be too optimistic. And as a scientist I have to be careful about assumptions and results. That attitude unfortunately has biased my investment decisions...
Bitcoin vs dollar, euro and yen
Now, I think it may be safe to assume slightly higher shares, having in mind growing bitcoin adaptation. For instance, according to different studies 8-11% of Poles already have cryptos (mostly bitcoins). That is a very big number but it is growing.
- In two years the bitcoin supply will be about 18.5 million (halving estimated date is and of May 2020). I also note that some of bitcoins are lost (following Chainalysis 3.8 million).
- I took value of M0 of USD, M1 of EUR, M0 of JPY together. That makes 13.85 billion USD.
- I assume that the share of bitcoin will be 5% of the sum of these fiat currencies. I pessimistically assume that all crypto market will have not more than 10% of total money supply in leading currency areas. I also note that cryptos becoming more specialised and in the future most of them will not play the role of means of payments compared to cash or e-money, but at the moment it is hard to do any estimates of the blockchain industry.
Hence simple math gives that using the assumptions above
It is much pessimistic than speculations of many bitcoin 'whales'. The price may be higher if there is another 'bubble', like the ones in December 2013 or in December 2017 (see below).
In 2014 I also did some simple forecasting based on cycles found within time series, and based on trends. They however failed, as probably MtGox (or ‘Willy Bot’) highly influenced prices and the sample of data I was able to retrieve was too short. I also took it into consideration and partly neglected this period below and collected more historical data.
I took data from 1 January 2011 (I eliminated earlier data) up to 31 July 2018 (7.5 years) what justifies forecasts for 2.5–3.5 years. To be more cautious, I only forecast for two years (up to mid-2020). I have plotted two lines showing the interval based on the maxima and price minima in historical data, and also showed the exponential trend line on the logarithmic scale.
This graph suggests that
It is necessary to pay attention to a very large range fluctuation range. The results received in previous chapter fit to the price range presented on the graph.
I have prepared thousands of forecasting models for insurance companies and banks and I am aware that the above approach is very simple. My students, when they apply econometrics in their MA's do great job and some of their prediction of bitcoin prices were surprisingly correct. But with the ever-growing market, with ever-changing factors influencing the phenomenon under investigation, sometimes such a simple approach can be justified.
Disclaimer: I am bitcoin hodler.
The post should not be considered investment advice.
Investment in cryptocurrencies is very risky and their prices changes substantially.