The market for Initial Coin Offerings (ICO) is still largely unexplored. The sources of systematic risk which matter for the dynamics of risk premia are by and large an unresolved puzzle. In particular, given the limited amount of data often available and the volatility associated with ICO markets, precise measures of risk proxies, such as value, size, and momentum, can be hardly obtained.
In this report, we aim to start filling this gap by analysing the dynamics of returns for ICO markets both at the individual level and for less granular industry classifications. The industry classifications are of particular interest as exploiting both time series and cross-sectional information can help obtain better estimates of CAPM betas and the corresponding risk-adjusted returns (i.e. the alphas).
We analyse 380 ICOs over 50 different industries over the period Sept-15 to Jan-19 on a weekly basis, for a total of 20,368 ICO-week observations. The panel is unbalanced, meaning that the number of ICOs at the beginning of the sample is not the same of the ICOs at the end of the sample. The average funds raised for an ICO at inception is $35 million. The cross-sectional average market capitalization is $35 million over the period.
Figure 1: The cross-sectional distribution of Sharpe ratios
The left panel of Figure 1 reports the distribution of Sharpe ratios for each ICO in the sample. The average Sharpe ratio over the sample is -0.12 on a weekly basis. The range of possible Sharpe ratios span from a minimum of -0.78 to a maximum of 0.55, again on a weekly basis. The vast majority of ICO deals turned out to generate negative Sharpe ratios.
The right panel of Figure 1 shows the value-weighted index calculated based on the weekly close prices and the relative market capitalization of each ICO over the sample. The dynamics are quite consistent with the conventional wisdom of a boom-bust cycle of ICOs, and more generally cryptocurrencies, between the end of 2017 and the beginning of 2018. The ICO market has traded laterally since mid-2018 until now.
Aggregate ICO Market
We now run a simple CAPM-style analysis in which we regress the returns of each ICO on a market factor, calculated as the log-returns over the value-weighted market index as shown in Figure 1. In particular, we estimate a regression of the form:
where rit is the log-returns for a given ICO at time t, αi is the so-called Jensen's alpha, βi represents the “market beta", and xt represents the returns on the value-weighted market portfolio. The estimates of the parameters are corrected for the presence of outliers, as well as autocorrelation and heteroskedasticity in the residuals.
Figure 2: Cross-sectional distribution CAPM alphas (left panel) and market betas (right panel)
The left panel of Figure 2 shows the cross-sectional distribution of the CAPM alphas for each ICO. Consistent with Figure 1, the risk-adjusted returns of ICOs are negative for most of the sample, although to a lower extent than by simply looking at raw Sharpe ratios. The right panel shows the exposure to “market risk", that is, the cross-sectional distribution of the estimates beta, i = 1, ..., N. The exposure to the aggregate market factor is mostly positive in the cross section.
We now delve further into the dynamics of ICO returns and investigate the industry-specific raw and risk-adjusted returns. We classify industries as per the data provider indication, from Agriculture to Education, passing by more sensible industries such as Fintech and Payments, for a total of 50 different industries. We compute industry portfolios as value-weighted averages of the ICO returns within each industry.
Figure 3 shows the dynamics of the value-weighted portfolio for selected industry classifications. The top left (right) panel shows the time series of the Fintech (Artificial Intelligence) industry portfolio. The bottom left (right) panel shows the time series of the Payment (Blockchain) industry portfolio. It is clear there is a substantial heterogeneity in the dynamics of wealth across different ICO industries.
Figure 3: Value-weighted industry portfolios for selected industry classifications
The left panel of Figure 4 shows the Sharpe ratios for each of the industry portfolios. The labelling of each industry is consistent with the indication of the data provider. The figure shows that not all industries turned out to generate positive Sharpe ratios over the sample.
As a matter of fact, only 32 out of 50 industries generated positive Sharpe ratios. The right panel shows those industries. Nevertheless, it is worth noting that industries such as Web Services, Payments, Energy and Blockchain generated massively positively Sharpe ratios, up to 0.2 on a weekly basis.
The fact that Sharpe ratios are positive does not necessarily implies that such industries represent portable investment opportunities once adjusted for sources of systematic risks. Mimicking the analysis in Section 2, we estimate a set a CAPM regressions for each industry and report both the risk-adjusted excess returns (i.e. the alphas), and the exposures to industry risk (i.e. the industry betas).
Figure 4: Sharpe ratios for different sectors
Figure 5 reports the Jensen's alphas for each industry in our sample.
Figure 5: Jensen's alphas for different industries
The left panel reports the whole cross section of alphas, without discriminating in terms of statistical significance. Consistent with the cross-sectional distribution of industry-specific Sharpe ratios, a good chunk of risk-adjusted returns turned out to be negative. The right panel of Figure 5 focuses uniquely on those industries which generate positive and significant risk-adjusted returns. There are a handful of industries, such as Cybersecurity, Banking, Financial Services, and E-Commerce, which generated substantial and significant positive alphas over the sample.
Turning the attention to the industry betas, Figure 6 shows the market betas for each industry. Some industries, such as Cybersecurity, Web Services, AI, and Security, are highly exposed to market risk. On the other hand, a good fraction of industries show betas which are considerably lower than 1. As a matter of fact, some industries, such as Gambling, do not show significant exposure to market risk at all. The right panel of Figure 6 shows those industries for which market betas are significant. Interestingly, only 28 industries have significant exposure to market risk.
Figure 6: Market betas by industry
Distributed Ledger Technologies (DLT) are a relatively recent arrival in the world of finance. Nevertheless, they are already driving new forms of financial innovation and creating new processes and platforms. Initial Coin Offerings (ICOs) are one of the most prominent applications of blockchain for finance, allowing for an innovative and inclusive way of financing for small and medium-sized companies (SMEs). Over the past two years, the use of ICOs has gone from ‘too small to care’ to ‘too big to ignore’ for markets and regulators alike.
This report takes the analysis further, discussing the risk-return trade-off of the ICO market across different industries. It analyses issues around industry returns and the risks faced by investors, accounting for the heterogeneity due to different industry factors. In this respect, this report highlights the importance of industry network effects as a source of returns in ICO offerings and the limits of looking at ICOs at the aggregate level.