Trading and Arbitrage in Cryptocurrency Markets

Igor Makarov is Associate Professor of Finance at the London School of Economics and Political Science and Antoinette Schoar is the Michael M. Koerner Professor of Entrepreneurship at the MIT Sloan School of Management. This post is based on their recent article, forthcoming in the Journal of Financial Economics.

Cryptocurrencies such as bitcoin or ethereum have rocketed to public attention over the past few years. These are digital currencies built on blockchain technology that allows verification of payments and other transactions in the absence of a centralized custodian. While significant attention has been paid to the dramatic ups and downs in the volume and price of cryptocurrencies, there has not been a systematic analysis of the trading and efficiency of cryptocurrencies markets. Cryptocurrencies are traded on many nonintegrated exchanges that are independently owned and exist in parallel across countries. But in contrast to traditional, regulated equity markets, the cryptocurrency markets lack any provisions to ensure that investors receive the best price when executing trades. As a result, it is centrally important to understand how arbitrageurs trade across different markets; and if there are any constraints to the flow of arbitrage capital which can result in market segmentation. In our article Trading and Arbitrage in Cryptocurrency Markets (forthcoming in the Journal of Financial Economics), we attempt to fill this gap using trade level data for 34 exchanges across 19 countries.

We first show that there are large and recurring deviations in bitcoin prices that open up across different exchanges and often persist for several hours, and, in some instances, days and weeks. These price deviations are much larger across countries (or regions) than within the same country and exist even between countries with the most liquid exchanges, such as the US, Japan, or Korea. For example, the daily average price ratio between the US and Korea between December 2017 to February 2018 reached 40% for several days, which was noted in the popular press as the “Kimchi premium”. We estimate that during this period a minimum of $2 billion of potential total arbitrage profits were left on the table. In contrast, the price deviations between exchanges in the same country typically do not exceed 1%, on average.

Second, our analysis shows that price deviations occur during periods of particularly quick appreciation of bitcoin prices. These periods coincide with times when there is a particularly strong increase in demand for bitcoin worldwide. We construct a measure of “buying pressure” and show that countries which have a higher premium over the US bitcoin price are also those who show more appreciation in local cryptocurrency prices when buying pressure is high. In other words, these countries respond more strongly in widening arbitrage deviations in times when buying pressure goes up in the US. In addition, deviations in bitcoin prices across countries are highly asymmetric. In countries outside the US and Europe, bitcoin typically trades at a premium relative to the US and almost never at a price below the US. And arbitrage spreads open up and close at the same time across countries.

These results show that the marginal investor outside the US is willing to pay more for bitcoin in response to positive news or sentiment. How can one explain this difference in valuation? We conjecture that they might reflect tighter capital controls or weaker financial institutions in the countries outside the US and Europe. The marginal investor in a country with poorly functioning financial institutions or tighter capital controls is willing to pay more for bitcoin since they would benefit more from the adoption of cryptocurrencies. Therefore, any news about the potential adoption of bitcoin technology would increase the price in these countries more.

Of course, these price differences could only persist if capital markets are segmented or capital is slow to flow across borders. And any constraints to the flow of capital between regions reduce the efficient use of arbitrage capital. It is important to note that capital controls are most relevant for the price of crypto to fiat currencies. In contrast, transactions between two cryptocurrencies should not be affected since they are explicitly designed to circumvent restriction to capital flows.

To test the importance of capital controls on fiat currencies, we analyze whether the positive correlation in arbitrage spreads between the countries we show above is explained by the level of openness of a country. If countries that are relatively closed have a higher convenience yield for bitcoin, then we should see their arbitrage spreads (relative to the world market price) move more closely together. Countries that are more open should not be correlated since any price deviation will be immediately arbitraged away. Indeed, we find that there is a significantly positive relation between the correlation of arbitrage spreads and capital controls. In other words, two countries that are both relatively closed to capital flows have a higher correlation in arbitrage spreads.

In further support of the idea that capital controls play an important role, we find that arbitrage spreads are an order of magnitude smaller between cryptocurrencies (say bitcoin to ethereum or to ripple) on the exact same exchanges where we see big and persistent arbitrage spreads relative to fiat currencies. Since the main difference between fiat and cryptocurrencies is the inability to enforce capital controls, our findings suggest that such controls contribute to the large arbitrage spreads we find across regions.

Our findings suggest that there are significant barriers to arbitrage between regions and, to a lesser extent, even between exchanges in the same country. But the magnitude of the arbitrage spreads we document is still surprising. While regulations in some countries make cross-border transfers in fiat currencies difficult for retail investors, industry reports suggest that large institutions should be able to avoid these constraints. Overall, it appears that there is insufficient arbitrage capital relative to the size of the price deviations we observe. So when large arbitrage opportunities open up in some countries, scarce arbitrage capital seems to be diverted from countries like the US and Europe towards the out-sized opportunities in other places. But these outflows are not immediately replaced by other market participants.

Finally, we conduct a number of robustness tests to show that mere transaction costs cannot explain the size of arbitrage spreads across exchanges since their magnitudes are small in comparison to the arbitrage spreads we show. Similarly, the governance risk of cryptocurrency exchanges being hacked or misappropriating client funds is also unlikely to explain these arbitrage spreads, since they cannot explain the direction of the arbitrage spreads. However, it is possible that ex ante concerns about the safety of cryptocurrency exchanges might make some arbitrageurs stay out of the market all together, which might ultimately explain why arbitrage capital is limited. Of course, increasing professionalization, better governance of the cryptocurrency market, and innovations over time might reduce these constraints to arbitrage.

The complete article is available here.

Both comments and trackbacks are currently closed.