The Joke’s On Us: Crypto just repeated one of the mistakes it was created to fix.
Both crypto in 2018 and real estate in 2008 were victims of positive feedback loops.
Last year, we all witnessed the extreme volatility of cryptocurrency prices. Obviously, this volatility encompassed more than last year, but people seem to be a lot more upset about volatility when it manifests itself in one particular direction. So, what we really saw was the expansion and then painful deflation of a speculative bubble. Such a bubble is not unexpected as a new technology is introduced, but did it have to be as big as it was? After all, the price of Ether peaked at 1422.47 in early 2018, before declining to $82.08, a decline of 94.2%. Markets have seen 94% declines before, but this was in less than a year. (And, as of this writing, Ether has bounced significantly off those lows, but who knows which way next.) Bitcoin’s drop was not as large, at close to 85%.
Was there something aside from the usual excitement and subsequent disappointment of a new market driving this volatility? I would argue that there was (otherwise, you know, short article). That additional element was feedback, specifically positive feedback.
The ironic part is that cryptocurrencies grew, at least in part, as a response to the chaos and carnage of the 2008 banking crisis. And the problem underlying that banking crisis was positive feedback.
Little thought appears to have been given to feedback, positive and negative, in the design of many coins and tokens. Perhaps, it is time to remedy this.
Just to clarify, some define positive feedback as forces driving prices upward and negative feedback as those forces that drive prices lower. This is incorrect. Positive feedback is properly defined as “the enhancement or amplification of an effect by its own influence on the process that gives rise to it.” Negative feedback is the opposite. It is the reduction of an effect in response to that effect.
Extrinsic Positive Feedback in Ethereum
As previously and rather obviously mentioned, the price of Ether has been crazy. From an “ICO” price of approximately $0.30, the price rallied 474,000%, before its 94% plunge.
During the rally, which was doing well enough on its own, companies began to raise through ICOs. The simplest method for raising was to send from an Ethereum address to an Ethereum smart contract address. The company raising would then send back a token to the Ethereum address of the sender.
There was nothing inherently wrong in this process, but it picked up steam in the midst of a boom in prices. As ICOs raised funds, especially if they raised early enough in the cycle, any Ethereum retained increased the value of the funds raised. So, companies kept large proportions of the Ether raised. In fact, the evidence available suggests that many ICO companies have maintained the majority of their funds in Ether until at least a few months ago.
How does this affect the market? First, there is an increased demand for Ether to invest in ICOs. ICO investors purchase the ETH to exchange for tokens, then transfer those to the ICO company. Then, in many, and probably most, cases, the ICO company does not sell any significant portion of its ETH back into the market. The overall effect is that just as ETH was heading toward its all-time highs, the tradable supply was decreasing.
Later, the market for ICOs began to cool. Over the course of 2018, the amount of funds raised per month has declined (measured in dollars). So, as the market was declining the amount of ETH leaving the market declined. Eventually, there was a huge drop in ETH, which I suspect but cannot prove was some portion of the ICO companies throwing in the towel on their ETH holdings. If this is correct, it is supply being added to the market as the market hits its lows.
None of this is to suggest flaws in the design of Ether itself. In fact, these positive feedbacks developed over time and outside the control of the designers. Nor am I suggesting there is anything wrong with ICO/STOs. That is a subject for a different article (by someone else). But this funding mechanism is pro-cyclical, adding to buying in rallies and adding to selling in slumps. The price will continue to be volatile in both directions until and unless ICO/STO efforts shift to a different funding mechanism (or cease altogether).
BTC Feedback
Bitcoin has its own positive feedbacks with its hash rate. While there is no proven valuation metric for the dollar price of Bitcoin, many models use the hash rate as a metric. But as the price of Bitcoin declines below the electrical cost to produce Bitcoin, miners begin dropping off. This reduces the total hash power. While the hash rate and the price are not necessarily related, there are dangers to the falling hash rate. For instance, recently, the total mining power removed was sufficient to re-enter and launch a 51% attack. Whether this is a realistic concern or not is beside the point. It is a concern that can affect the price in a negative way as the price reaches new lows. The low price itself is what is driving miners out of the equation. This is a positive feedback mechanism that increases volatility. It is not as extreme as the one operating in Ether, and that can be seen in the price action of Bitcoin compared to Ether, but it is pro-cyclical nonetheless.
Real Estate Feedback and the Banking Crisis
In the case of real estate and banking, the feedbacks that fueled the housing and debt bubble were positive.
While many like to blame “greedy bankers”, this fails as explanation, not because bankers aren’t greedy but because bankers have always been greedy. Greed is and always has been part of the equation where humans are involved.
In the interest of time, this will be a vastly oversimplified and incomplete discussion of a complex process. I focus only on the feedback mechanisms.
By 2003 the Fed had lowered overnight rates to 1% and kept them there for over a year. This had two effects relevant to housing. First, it left investors searching for yield all over the world. Second, it began to fuel an increase in housing prices.
As real estate prices began to increase, loans and mortgage instruments had lower delinquency rates, raising returns and increasing the perceived reliability of the instruments. At the same time, the search for yield brought new buyers into the mortgage market. This, in turn, brought more money into the housing market, pushing prices higher. In such a strong market, even the most aggressive loans were profitable, so lending standards began to decline. As lending standards declined, more buyers qualified, driving prices higher, which in turn made the loans backing the purchases look good. Even regulators and credit rating agencies saw only low failure rates. They did not serve as an effective negative feedback.
We all know the end of the story. As prices finally leveled off, and adjustable rate mortgages began to reset to higher rates and payments en masse, delinquencies finally began to rise. This led to tighter lending standards and less money available to lend. Finally, banks had to liquidate loan portfolios at the lows. Foreclosures flooded the market with supply at the lows.
Brian Wesbury, of First Trust strongly suggests that these positive feedbacks only got out of control because of an additional positive feedback — the FASB ruling requiring banks to mark loans to market, even when cash flows were solid and loans were performing. This rule went into place in 2007, just as the mortgage market was weakening. His argument makes sense and, when those FASB rules were loosened in March 2009, the stock market bottomed and housing prices stopped falling in most markets.
Whether it was the overall feedbacks in the system or the specific feedbacks in the FASB standards, positive feedback loops were a major source of the crisis that rocked the economy in 2007–2009.
The joke is on us. We did this to ourselves. We’ve responded to the collapse of a positive feedback system by building more positive feedback loops.
Stablecoins — too much negative feedback?
Stablecoins clearly have negative feedbacks. Basis, which recently returned capital to investors, was comprised of a series of mechanisms to keep the price of the Stablecoin at or close to $1.00 — negative feedbacks all. On the other hand, tokens like USDC or, at least purportedly, Tether hold $1 for every $1 of token in circulation.
Stablecoins are great. They serve many useful functions. But they do not solve the problems cryptocurrency was made to solve.
What happens when the currency against which a coin is stable, isn’t stable itself? The backing is, directly or indirectly the U.S. dollar. But there are lot of dollars in the world, and there are going to be more. While the dollar is not yet the Venezuelan Bolivar, it could go that direction. The U.S. debt is on a path to unsustainability. The U.S. Social Security and Medicare systems are still demographically unsupportable. The dollar might survive these crises, but it might not. Bitcoin or other cryptos may give you some protection against this, but Stablecoins clearly cannot.
Possible Solutions
With respect to Ether, the simplest solution is for ICOs and STOs to offset, eliminate or reduce their ETH/USD exposure. They could do this by demanding Stablecoins as payment, selling ETH immediately, or by hedging the ETH. The hedging instruments available are not great at the moment, but better than nothing. But hanging on to Ether at the high and hoping is not a great strategy. Of course, there is no way to force any company raising funds through Ethereum to do any of this.
The issues with Bitcoin are based on its design. These risks may dissipate as and if Bitcoin reaches far higher market capitalization.
With regard to real estate and banking, the banking sector should really have hedging for its real estate exposure. The tool available now is generally securitization, which is expensive in terms of lost yield. Also, securitization does not provide informational feedback the way hedging markets would. If you lend a lot against real estate, and so does everyone else, the hedge begins to get more and more expensive, telling you, the lender, that perhaps too much lending is going on. You may not care about the information, but you will care about the cost and take the signal.
Future token designs or token economics designs should take these feedbacks into account. Looking back to the Internet bubble and its deflation, we can observe one of the winners that emerged from that period. As the price of Amazon’s stock dropped from Dec. 1999 through Dec. 2001, revenues doubled. There was no feedback between the price and the activity of the company as measured in total receipts.
At GSD we have been developing a token economic model that insulates, or largely insulates, the activity the design intends to incentivize from the price of the token itself. To be clear, this cannot and will not remove all price volatility from the token price. Instead, without this one particular positive feedback, volatility should be lower than it would otherwise be and network activity should not suffer.
Other systems may incorporate elements of negative feedback, along the lines of those used in stablecoins, to reduce volatility. The challenge in those cases is to preserve decentralization, but it seems possible.
Feedback is not the only source volatility in cryptocurrency prices. The primary source of volatility is humans. Looking back at history, we do this a lot. But feedback effects do not seem to have been considered. We can minimize these feedback loops and since they build additional volatility into already volatile assets, we should.
