Excess Volatility in Bitcoin

–IIM Kozhikode Society & Management Review

The increasing popularity of crypto-currencies as a worldwide acceptable asset class for trading, speculation and investment purpose has drawn the attention of practitioners and researchers.

Bitcoin is the most popular among all crypto-currencies available with a share of 55% (as of December 2018) crypto-currencies market capitalization worldwide. Although academic literature on Bitcoin has received a considerable amount of impetus, very little is known about the price behaviour of the same.

Despite the growing popularity, the basic fundamental features of Bitcoin (or any cryptocurrencies) are considered to be a ‘black box’. Bitcoin’s characteristics as a financial asset, such as fiat money, gold and equity, are unclear. Academic studies have compared bitcoin with other assets such as currency and gold (Baur, Lee, & Hong, 2018; Klein, 2018).

Bitcoin has few characteristics that allow it to function as money, such as its easy transferability and utility as a useful payment method (not always). Its value is determined by demand–supply. However, the biggest shortfall of Bitcoin is the absence of a guarantor unlike other fiat money (Central Banks play the role of a guarantor).

Due to the prohibition put forth by most of the sovereign states in using cryptocurrencies, Bitcoin lacks a reliable proponent to sustain and expand its usage as a medium of exchange. Further, high volatile characteristic of Bitcoin makes it a risky instrument to be considered as a store of value as in the times of falling prices, people would prefer assets with stable value.

Moreover, no economic fundamentals such as inflation and trade balance are associated with the price of Bitcoin. Holders of Bitcoin do not receive dividends or interest payments. All these aspects make it extremely difficult to derive a fair price for the Bitcoin. This uncertainty about its price makes it more volatile.

There has been a reasonable attempt to understand the different characteristics of Bitcoin prices such as Bitcoin price volatility (Baek & Elbeck, 2015; Blau, 2017; Katsiampa, 2017). Bitcoin prices (as well as other cryptocurrencies) are known to be highly volatile (Dwyer, 2015).

However, studies on the volatility of Bitcoins are mostly limited to –

(a)    finding best asymptotic Generalized Autoregressive Conditional Heteroscedasticity (GARCH) models (Chu et al., 2017; Dyhrberg, 2016a; Katsiampa, 2017; Tiwari et al., 2019; Wang et al., 2019),

(b)    comparing Bitcoin volatilities with other assets (Baur, Lee, & Hong, 2018; Klein, 2018) or examining Bitcoin volatility spillover on other assets (Bouri et al., 2018)

(c)     predicting Bitcoin volatility (Balcilar et al., 2017; McNally, 2018) and

(d)    hedging opportunity (Bouri et al., 2017a, 2017b, 2017c; Dyhrberg, 2016b; Feng et al., 2018).

There is a growing literature to understand volatility patterns of Bitcoins (see Aalborg et al., 2018; Bouri et al., 2018; Ji et al., 2018). A true understanding of Bitcoin volatility patterns could reveal more information on market efficiency, overreactions (Chevapatrakul & Mascia, 2018) and also other aspects of Bitcoin price movements.

However, the volatility of an asset depends on the rate at which information is flowing into the market (Balcilar et al., 2017). The continuous arrival of new information helps in price movements. Major news in the market could lead to sharp reactions and could push the market temporarily unstable before the market prices adjust and reach a relatively stable price range.

The sharp movements in the prices reveal more information about price changes (change in fundamentals and overreactions) and can be captured by extreme value (EV) prices such as daily high and low (Chou et al., 2010).

This study aims to capture the overreaction or excess volatility in Bitcoin prices. However, most of the studies on Bitcoin volatility use daily closing prices for the volatility estimation. In this study, we estimate the volatility of Bitcoin prices using Open, High, Low and Close (OHLC) prices. Our estimation is model-free and based on squared returns of EVs. Volatility estimation based on squared returns is more robust and reliable as argued by Balcilar et al. (2017) and Balcilar et al. (2018).

Sudden changes in prices occur due to the overreactions by the market participants and therefore prices bounce back after sometimes. However, these sharp movements in Bitcoin prices are not necessarily overreactions all the time. It could be consistent with Bitcoin fundamentals. In that case, prices will not bounce back so quickly. We capture the overreactions in Bitcoin prices by comparing two different measures of volatility.

First, we compute high-to-low volatility which reflects jumps and price-reversal in Bitcoin prices. Then we compare it with open-to-close volatility which reflects the price movements over the full trading hours. In the absence of overreactions and instant price adjustment, both measures of volatility will not differ and their ratio should be unity under perfect condition (Kayal & Maheswaran, 2018).

If the ratio goes above unity, then the volatility reflects a significant sign of overreaction in price changes of Bitcoin. We adopt the definition of Maheswaran et al. (2011) and call it excess volatility. For any asset, if sharp price changes are not supported by its fundamentals then the asset prices will exhibit excess volatility. We assess the same for Bitcoin prices.

.

.

.