Long term relationships also strengthen, to some extent, around areas indicated as bubbles. The previously observed long term relationship between Google Trends and Bitcoin price [8] can also be seen here, between late 2012 and 2014 (period band 64–256). With the benefit of extra data it can be observed that the relationship disappears around 2014 (for lower period bands) and 2015 (for higher period bands), before the relationships start occurring more consistently in 2016 and 2017 (a region with a number of bubbles identified). The previously observed relationship between Wikipedia views and Bitcoin observed in 2013 (64–128 band), disappears before again returning in mid-2016 and 2017.
But this time feels different. It feels like a bubble. The fever in the post-Thanksgiving moonshot ran hotter than we’d seen before. We also began to see a robust supply response.

GARY GENSLER: This asset class is rife with fraud, scams and abuses in certain applications. There's a great deal of hype and spin about how crypto assets work.
Interpretation of visual scalograms is subjective so it is desirable to find a more quantifiable way to validate the strengthening of coherence in bubble regimes. Fig 6 shows the wavelet coherence over time for the different period bands, in the case of the “new authors” factor for Ethereum. Coherence values, plotted on the vertical axis, vary between zero and one. Time is plotted on the horizontal axis. The areas of the price time series that are recorded as bubble-like regimes using the GSADF test are shaded red.