This is a research piece on the bitcoin arbitrage opportunity between South African exchanges and global exchanges. It’s not relevant for non-South African readers.
Bitcoin (BTC) is purchased through BTCZAR markets in South Africa, which is different to BTCUSD markets globally. When converting the BTCZAR price into USD, we tend to see a spread between the BTCZAR and BTCUSD. I.E. BTC trades at a higher price in South Africa than globally. This spread is normal in Emerging Markets. The limited number of bitcoin miners in certain Emerging Markets, including South Africa, is the primary reason for a spread (low electricity prices and or excess electricity supply are the two primary factors motivating bitcoin miner operations).
Miners generate bitcoin supply. Without miners, new bitcoin needs to be sourced off-shore and bought onto South African exchanges, which is costly (time, fees and induces volatility). Usually institutional investors would arbitrage financial market spread away but the lack of institutional activity in this market allows the spread to exist for extended periods of time.
I use Luno weekly close BTCZAR data since 2017 to analyse the historical drivers of this spread, comparing it to Bloomberg for global USDZAR and BTCZAR prices. Thanks to Marius Reitz at Luno for the data.
- There is a clear but volatile bitcoin spread between South African and global exchanges, ranging between -10% and 34%, averaging 7%.
- The spread is falling over time but there is reason to believe that it could increase again, if another bitcoin bull-market takes place.
- The spread widens during BTCUSD volatility, particularly BTCUSD strength
- The spread widens during ZAR strength and reduces during ZAR weakness
- The spread presents an arbitrage opportunity that must be weighed against the opportunity cost of investing elsewhere
The graph below shows the spread between the weekly closing price on Luno’s exchange vs. Bloomberg’s BTCZAR price. The spread is displayed in USD prices.
The second chart shows the spread as a percentage of the bitcoin price.
Intuition tells us that the spread should decrease over time as more players enter the market. Greater regulation would almost certainly narrow the spread because this would allow traditional financial institutions, like banks, to implement their sophisticated trading platforms on the bitcoin market.
The 52-week rolling average of the spread has fallen over time but this is influenced by the last bitcoin bull-market which caused a large spike in the spread in 2018. That data has now fallen out of the 52-week average. While I expect the spread to narrow, it has held up surprisingly well for a surprisingly long period of time.
Drivers: BTC and ZAR
The spread is positively correlated with the bitcoin price, 10% correlation on weekly returns. There is a stronger correlation with the absolute value of weekly returns (20%), which is an indication of volatility. I.E. spread tends to be higher during bitcoin volatility, particularly upside volatility. This makes sense, the quicker the bitcoin price increases, the greater the FOMO amongst South Africans, which encourages them to blindly purchase on South African exchanges irrespective of the price and pushes the spread higher.
There is a negative correlation correlation between the spread and the USD-ZAR. I.E. if the ZAR weakens sharply, the spread tends to narrow. Bitcoin is obviously a currency hedge for South African’s but this data indicates that marginal bitcoin purchases are more sensitive to the bitcoin price than the USDZAR.
Quantifying the opportunity for those who’d like to try and implement an arbitrage trade: Conservatively, one could estimate that the spread might be 5% on any given day. These are strong returns given the weak return potential on many traditional asset classes. Additioally, the trade can be implemented in relatively short time frame. One just needs the capital to clear in a off-shore bank account, which varies depending on the bank and relationship. For South African’s with a $1mn off-shore travel allowance this could imply roughly ZAR50K worth of annual returns. Those with special $10mn allowances, could potentially earn ZAR500K worth of annual returns. These are interesting numbers for sure!
That being said, the opportunity should be weighed against bitcoin itself, which I suspect has strong return potential over the coming year. The spread might look attractive during a bitcoin bull market but one may prefer to be exposed to the underlying and implement the arbitrage during pull-backs in the bitcoin price, if this is possible. Hopefully the information presented makes the investigation of these opportunities a little clearer.
Bottom-line: opportunity exists to capitalise off spread, but must be weighed vs. potential of holding underlying (BTC). Opportunity is falling over time but remains reasonable. BTC upside volatility should result in a higher spread, but ZAR weakness presents an opposing risk that would reduce potential returns.