Bubbles in Cryptoasset Markets: A Review of the Literature

The very existence of bubbles in financial markets is somewhat controversial. For instance, the eminent financial economist Eugene F. Fama defines a “bubble” as an “irrational strong price increase that implies a predictable strong decline” (Fama 2014, p.1475). Fama’s argument, in essence, is that if one looks at stocks or portfolios of stocks that exhibit a substantial price increase, the average returns going forward are not unusually low. By looking at returns on different US and international industry portfolios, Greenwood et al. (2019) essentially confirm Fama’s argument. However, such conclusion is fundamentally against a large chunk of literature which claims the opposite, that is, asset returns indeed exhibit significant price increases which are followed by even larger market crashes (see, e.g., Shiller 2000). In particular, Shiller (2000) argues that when asset prices significantly depart from their fundamental value, a sudden and large drop in value often occurs, that is, the bubble bursts.

When it comes to cryptoasset markets, things are a bit more complicated. This because of four reasons. First, most cryptoassets are hard to value given current valuing methodologies and their short price history, with no offer of a final promised payment or any dividends. Second, at least for now, cryptoassets do not broadly serve as legal tender or common official means of exchange. Third, perceived value is driven largely on speculative future adoption and use cases. Fourth, these markets are largely driven by retail investors, as opposed to experienced professional investors. These four aspects make Bitcoin, and cryptoassets as a whole, more prone to speculative bubbles, similar to other nascent technologies. In this report we discuss the existing evidence of financial bubbles in Bitcoin and cryptoassets more broadly.

Introduction

Financial bubbles have been heavily covered in the academic literature across a variety of asset classes, with research spanning contagion effects, to the role of algorithmic trading and speculation. The main reason why it is important to understand the unfolding of price bubbles lies in the fact that a burst often coincides with large losses and extreme downside risk. Such price drops are not only a concern for investors and market participants, but also for policymakers and regulators.

Unsurprisingly, astonishing price appreciation over the last few years has spurred substantial interest in this area within cryptoasset markets. A wide spectrum of methods and perspectives have been used to understand price formation in Bitcoin and other cryptoassets. Perhaps the most used perspective is the asset-pricing based approach which consider Bitcoin and altcoins as an investment tool. With this in mind, a bubble is often defined as a significant departure of market price from the fundamental value, which is generally lower and based on production costs. Continuous increases in the multiplicity through which nominal prices exceed fundamental values lead to explosive behaviour and the formation of bubbles. Such deviations from fundamental prices are often due to trading demand from highly optimistic investor sentiment that thereby lead to an increased level of aggregate demand for assets. That is, with supply stable or decreasing, such as for Bitcoin, such sharp demand increase may generate sudden price changes.

Defining a Bubble

Flood ad Hodrick (1990) defined a rational bubble as a price increase due to investors’ beliefs that there will be the possibility to sell an already overvalued asset at a higher price in the future. In other words, as prices keep going up investors require even higher compensation as the risk of a sudden price drop becomes more concrete. The constant requirement for a higher compensation leads to overgrowing prices until finally the bubble bursts. Rational bubbles can be typically characterised as “intrinsic” or “extrinsic” bubbles (see, e.g., Dale et al. 2005). An intrinsic rational bubble is formed when investors wrongly estimate the fundamental value of an asset. A case in point is advanced technologies, whose fundamental values are typically more difficult to estimate. In this respect, crashes are then the result of information updates on growth prospects after long-lasting price run ups. Extrinsic rational bubbles, often called “sunspots” typically occur when investors face high uncertainty about the general economic environment the asset is operating in.

As far as cryptoassets are concerned, one can think about the nature of price run ups as being mostly due to an intrinsic-like rational bubble. That is, given the new and disruptive nature of distributed ledger technologies, it may be particularly difficult for rational investors to fully evaluate the economic fundamentals of the asset class.

Measuring a Bubble

There is no apparent consensus on how to measure and predict price bubbles, and in particular bubble bursts. Rational bubbles could appear in the form of prices continuously departing from fundamentals, explosive autoregressive (i.e., AR(1)) price processes, deterministic time trends, or even more complex stochastic dynamics.

The first view, i.e. prices departing from fundamentals, is somewhat more traditional. Typically, the fundamental value of an asset is defined based on the present value of future payoffs and / or cash flows. Thereby, a bubble is defined when the discounted value of future payoffs does not coincide with the current market value of the same asset. Another approach is based on more, say, reduced-form econometric approaches; for instance, Phillips et al. (2014) and Phillips et al. (2015) propose a test for bubbles based on extensions of otherwise standard tests for unit roots in the price dynamics.

Both approaches have pros and cons; for instance, if a reliable estimate of the future payoffs of an asset is available then a more fundamental-based approach may be suitable as it grounds on pure accounting identities. On the other hand, if reliable data on future payoffs are not available, then a simple test based on the price trajectory may be informative enough.

Bubbles and Innovation

The idea of intrinsic rational bubbles is inherently linked with the idea that bubbles might be recurrent phenomena, endogenous to the market system of advanced technologies. They result from the way technological revolutions are assimilated. That is, major price run ups and collapses signal the need for a structural shift in the forces guiding growth and innovation within the cryptoasset space. If history is a guide, the current high volatility and sequence of all time highs can provide some guideline on what one can expect over the next few years.

The idea that bubbles and technological innovation may be interlinked has been advocated by Perez (2003). The paradigm shift brought by new, disruptive, technologies can translate into massive changes in all aspects of business, from trading to wealth management. The fundamental value of a fast-growing technology, with potentially far-reaching implications for the financial system is, by definition, rather difficult to assess. Cycles of price run ups and collapses can then be thought of as being a natural part of the “installation period”, as in the words of Perez (2003), of cryptoassets. In this respect, bubbles are viewed as somewhat instrumental and a physiological component of the growth process of Bitcoin and altcoins.

The Case of Bitcoin

There have been an increasing number of empirical papers that investigate the bubble price dynamics in cryptoasset markets. Being the first and most famous cryptoasset, the majority of studies focus on Bitcoin prices. For instance, Cheung et al. (2015) use daily Bitcoin data from 2010 to 2014 and adopt the Phillips et al. (2015) methodology to examine whether price bubbles exist in Bitcoins traded on what was previously the largest exchange, Mt. Gox. Estimates from the Supremum Augmented Dickey-Fuller (SADF) test reveal that most of the bubble-like periods over that time span did not last for long, and indeed only a few days. Similarly, Corbet et al. (2018) employ daily data on Bitcoin from 2009 to 2017 and provide evidence of Bitcoin bubble behaviour around the turn of the year from 2013 to 2014. Moreover, Ethereum exhibits bubble behaviour in the beginning of 2016 and in the mid-2017.

Geuder et al. (2019) also look at the existence of bubble-like dynamics in the price of Bitcoin. They look at the period from 2016 to 2019 and investigate the presence of bubbles based on two alternative methodologies: the Phillips et al. (2015) method, which is based on a supremum augmented Dickey-Fuller (ADF) test, and log-periodic power law (LPPL) approach by Filimonov and Sornette (2013), which is based on the alternative idea that prices in a bubble period do not follow an exponential dynamic, but rather a faster-than-exponential price growth. Empirically, they show that there are indeed bubble periods in Bitcoin prices, regardless of the testing methodology implemented. For instance, after December 6, 2017, Bitcoin’s price growth rate changes and bubble deflates as part of the latest bubble period ending in January 2018. Before December 6, 2017, the LPPL test show that the Bitcoin price is characterized by faster-than-exponential growth.

Perhaps on a different note, more recently, Perepelitsa and Timofeyev (2020) argue that the fact that Bitcoin can be infinitely divisible in smaller units makes possible the emergence of self-sustained price bubbles. That is, a group of investors can “pump” the price up at a positive rate, without ever running out of cash. Although the idea of never-ending bubbles can be appealing to any investor, the idea is questionable at the very least. More precisely, the idea of Perepelitsa and Timofeyev (2020) is that, given BTC is infinitely divisible, thus everyone can afford a fraction of it, prices will keep increasing as far as the fraction of optimist investors dominates the pessimistic ones. Although it is reasonable to think that investors acting adaptively can generate long-lasting, positive price dynamics, it is perhaps less likely that optimists will continue to dominate pessimists in the long run. Nevertheless, Perepelitsa and Timofeyev (2020) pointed out a key aspect of the price dynamics of BTC, that is, it can be thought of as primarily driven by investors’ sentiment, an aspect that is constantly referred to in existing empirical research.

All in all, the evidence on bubble periods in Bitcoin prices is somewhat robust across methods and studies. However, neither the timing nor the magnitude of these bubbles is consistent across studies, which raise some question about the actual reliability from a purely market timing perspective. In addition, a natural question that arises from the existing empirical results is what eventually caused these episodes of bubble behaviour. Apart from the rational bubble view, there are many further reasons and determinants that may help to explain Bitcoin price behaviour. Given the global monetary environment for fiat currencies, institutional changes such as the introduction of Bitcoin futures, the use of Bitcoin for web-based transactions, governments restricting the use of cryptoassets and other regulatory changes may have significantly influenced Bitcoin price dynamics. A reduced-form econometric framework is unlikely to provide any meaningful answer in this respect.

The Evidence on Cryptoasset Markets

In a more recent strand of research, there is increasing evidence of bubble-like price dynamics across a variety of alternative cryptoassets. For instance, Chen and Hafner (2019) investigate whether sentiment-induced bubbles exist in cryptoasset markets based on daily data on a broad value-weighted market index sampled from 2014 to 2018. They test for bubbles using a smooth transition autoregressive model (STAR) with regime switching. Estimations indicate multiple bubble periods from May 2017 to April 2018. Moreover, the empirical evidence seems to show that returns volatility is higher during bubble periods, suggesting that bubble bursts could be associated with increasing market uncertainty. Similarly, Cagli (2019) investigate explosive behaviour in the market values of Bitcoin, Ethereum, Ripple, Litecoin, Stellar, NEM, Dash and Monero by employing daily data spanning from September 2015 to January 2018. The empirical evidence indicates that the vast majority of cryptoassets exhibit significant explosive price dynamics.

Again, the academic evidence reveals a clearer bubble character in major cryptoassets, especially Bitcoin, whereas others present price increases at a more modest level. Note however that the evidence is far from being conclusive and often based on ad-hoc modelling choices. That is, further research needs to be done on the topic.

Concluding Remarks

The substantial body of evidence that seeks to test for the existence and measurement of price bubbles in cryptoasset markets has been growing over the last decade. The possibly speculative trading dynamics in cryptoasset markets has fuelled further interest by researchers and policy makers alike, with some interesting debate especially since the recent sequence of all time highs in prices. Although the evidence of price bubbles in Bitcoin and cryptoassets seem somewhat convincing, few words of caution must be spelled.

First, most, if not all, existing empirical research is based on reduced-form econometric approaches which, although have merit in being quite flexible, do not explicitly consider the evolution of the underlying fundamental values of cryptoassets. Therefore, the detection of bubble formation processes may be misleading by looking uniquely at the price dynamics. The difficulty of using, for instance, present value models, is that the vast majority of cryptoassets do not pay dividends and / or regular cash flows. This impairs the possibility of using standard present value models to capture significant and persistent divergences between market prices and fundamental values.

Second, the evidence on the origins of price bubbles in cryptoasset markets is scattered, at best. Most of the explanations for bubble-like price behaviour grounds on market sentiment and investor attention. Leaving aside the measurement issues of both the market hype and beliefs, one may argue that again, without conditioning on economic fundamentals we may exaggerate the role of demand from public excitement on the price formation process. Related to that, it is worth asking as to whether the bubble characteristics of cryptoassets will perpetuate in the future. Predicting a bubble burst is the typical “one-million dollar” question. To the extent that elevated investor optimism continues, and irrational behaviour dominates investing strategies, prices can remain in an upward trajectory far more than what any reduced-form econometric model can predict.

Bibliography

  • Cagli, Efe Caglar. “Explosive behaviour in the prices of Bitcoin and altcoins.” Finance Research Letters 29 (2019): 398-403.
  • Chen, Cathy Yi-Hsuan, and Christian M. Hafner. “Sentiment-induced bubbles in the cryptocurrency market.” Journal of Risk and Financial Management 12.2 (2019): 53.
  • Cheung, Adrian, Eduardo Roca, and Jen-Je Su. “Crypto-currency bubbles: an application of the Phillips–Shi–Yu (2013) methodology on Mt. Gox bitcoin prices.” Applied Economics 47.23 (2015): 2348-2358.
  • Corbet, Shaen, Brian Lucey, and Larisa Yarovaya. “Datestamping the Bitcoin and Ethereum bubbles.” Finance Research Letters 26 (2018): 81-88.
  • Cretarola, Alessandra, and Gianna Figà-Talamanca. “Detecting bubbles in Bitcoin price dynamics via market exuberance.” Annals of Operations Research (2019): 1-21.
  • Dale, R. S., Johnson, J. E., & Tang, L. (2005). Financial markets can go mad: evidence of irrational behaviour during the South Sea Bubble 1. The Economic history review, 58(2), 233-271.
  • Fama, Eugene F., 2014, “Two Pillars of Asset Pricing,” American Economic Review 104, 1467-1485.
  • Filimonov, Vladimir, and Didier Sornette. “A stable and robust calibration scheme of the log-periodic power law model.” Physica A: Statistical Mechanics and its Applications 392.17 (2013): 3698-3707.
  • Flood, Robert P., and Robert J. Hodrick. “On testing for speculative bubbles.” Journal of economic perspectives 4.2 (1990): 85-101.
  • Geuder, Julian, Harald Kinateder, and Niklas F. Wagner. “Cryptocurrencies as financial bubbles: The case of Bitcoin.” Finance Research Letters 31 (2019).
  • Greenwood, R., Shleifer, A., & You, Y. (2019). Bubbles for Fama. Journal of Financial Economics, 131(1), 20-43.
  • Perepelitsa, Misha, and Ilya Timofeyev. “Self-sustained price bubbles driven by Bitcoin innovations and adaptive behavior.” arXiv preprint arXiv:2012.14860 (2020).
  • Perez, Carlota. Technological revolutions and financial capital. Edward Elgar Publishing, 2003.
  • Phillips, Peter CB, Shuping Shi, and Jun Yu. “Specification sensitivity in right‐tailed unit root testing for explosive behaviour.” Oxford Bulletin of Economics and Statistics 76.3 (2014): 315-333.
Share the love ❤️
Share on facebook
Facebook
Share on twitter
Twitter
Share on linkedin
LinkedIn

All of the content published on this site is strictly for informational and educational purposes and neither does it constitute investment advice nor solicitation to buy or sell blockchain-based tokens or securities. No investments whatsoever should, therefore, be made based upon information provided or discussed by any guests or hosts appearing within this video or audio content.

Any appearance by any company or individual does not infer any kind of endorsement by Scrumline Ltd (trading as Coinscrum) of either their products and/or services or in any related investment opportunity that they may be pursuing from time to time.

Crypto-tokens and equity investments are high risk in nature and you should always take the advice of a professional financial advisor from within the jurisdiction in which you reside before making any investment decisions.

About Author
About Author
OUR
Sponsors
cortex-logo
Dynamic Crypto Data, Essential Analysis
7 Day Giveaway: Unlimited Streaming Crypto Data for 12 Months