Limits of Token Price Futarchy

Governance Designs
1

This is the first post in a series on the challenges and open research of futarchy and decision markets applied to governance.

Token price futarchy is the earliest and most well-known version of futarchy in the crypto landscape.
Examples mentioning this:

We’ll categorize below some of the different kinds of limits:

breadth of confounding factors

As well understood, token price has a number of factors driving it. Futarchic decision markets, like Conditional Funding Markets, make use of comparison between \text{Yes} and \text{No} markets so as to neutralize some of these factors. For example, it appears such a comparison will do a good job at canceling out the effect of crypto market cycles, which would affect both \text{Yes} and \text{No} markets in the same way.

But specific factors like a short period of increased momentum on specific networks or technologies that directly relate to a proposal submitted to decision markets could heavily skew the \text{Yes} price up without impacting the \text{No} price, all the while being potentially a bad predictor of long-term value increase.

A key solution is to limit the scope of confounders by focusing on more precise metrics.

speculation

As in a Keynesian beauty contest, decision market traders will focus on whichever proposal will produce most positive trader sentiment, as sentiment swings affect both decision markets and token price.

This calls for usage of metrics rooted in fundamental value, resistant to such swings, and to measure price changes over timeframes long enough to account for real value creation.

there are better proxies for value creation

Metrics tracking fundamental network health and growth might do a better job at correlating with long-term success, e.g., TVL growth in a DeFi protocol or gas fees in an L2.

As mentioned by Bo Waggoner (see the link shared above):

From what I’ve seen on ethresearch, proposals for DAO futarchy often skip the “vote values” stage entirely and assume the metric is the price of the DAO’s token. But as I understand it, this severely limits the kinds of decisions that futarchy can make. For example, it seems hard for the DAO to decide via futarchy to allocate resources toward charity or something non-profitable.

the token price can be manipulated, especially when governance isn’t entirely futarchic

When governance isn’t relying entirely on futarchy as a mechanism to take decisions, but on a complementary governance mechanism like token voting, not only can whales influence token price via markets but actors with large governance power (either via large token holdings or other means, depending on the governance structure) can do so as well by passing proposals that, e.g., increase fees or reduce supplies (or the contrary).

Any level of control over future token price creates a confounding factor on decision markets. In particular, knowing that whales might forcibly decrease token price down the line if a proposal passes that they don’t like, decision market traders will try only to predict what whales might do rather than the correlation between the proposal itself and the token price.

1 Like
2

I think a token price decision market is, in some sense, a “discounted future project revenue” KPI prediction market, given that this is where most tokens derive their long term value. So in principle the token price therefore is an example of a KPI metric rooted in fundamental value.

The issue, though, seems to be that the token price is a very noisy estimate of the expected discounted future cash flows of the project, due to its high potential for reflexive dynamics and its tendency to significantly deviate from fundamentals in the short term.

The primary reason for this tendency seems to be that predicting future revenue/success is hard, leaving room for a significant uncertainty/narrative/hype to influence trader perceptions without a strong short-term feedback mechanism to correct mispricings, given that narratives often take >1y to be falsified by poor project performance. This seems like it will also apply to KPI decision markets using long term fundamental-value-oriented metrics.

I agree though that token price futarchy amplifies the uncertainty:

  • Token price is a problematic proxy for discounted cash flow, due to token prices kind of being “uncapped”. So the inability for KPI decision markets to provide 1000x returns might also help make their prices less reflexive than token price futarchy, where such returns are possible. This seems to be one of the most important differences between kpi and token price futarchy.
  • I.e. potential for 1000x returns probably makes speculation on “positive trader sentiment” much more attractive than if the returns from doing so are capped at like 3x. It would still be necessary in KPI futarchy to ensure prices can reflect the full range of plausible KPI outcomes. My hunch though is that in most cases, this range is more constrained than the range of potential token prices.

One other way I can see KPI futarchy reducing the impact of beauty contests relative to token price futarchy, is by being more short-term oriented by enforcing that the market resolves according to the value of a metric at a specified date in the future.

This way it kind of forces the market to ignore the uncertainty which results from trying to predict the longer term impact of an action, as would be necessary in a token futarchy. This can also reduce reflexivity/memecoin/speculative cycle dynamics as well, given these all “feed off” the lack of a strong corrective signal when predicting a token’s long term value.

Doesn’t this also potentially apply to KPI-futarchy, given that the whales could similarly attempt to use their governance influence to manipulate the values of the KPIs conditional on a proposal they like/dislike being passed? E.g. a daily active user KPI could be manipulated by creating a temporary usage incentive program.

3

This is correct in most cases when a token represents a potential claim of a share of future revenue, and your analysis is sound in that context: limiting the “discounted future” element will limit reflexivity and sentiment trading. Other than that, in such a context, token price is a just a particular case of a metric objective.

But, precisely, I believe we can say that token price isn’t always related to future revenue claims in context of cryptocurrencies, where memetic value can drive price. Bitcoin is the prime example of this. This is where this analysis gets more interesting.

This is correct. But more importantly, in this context, token-price is suffering from being a single objective whereas opening up to different metrics can leave the futarchy designer with more options to counteract manipulation.

1 Like
4

I am interested in ideas re: how to prevent prediction market shares from attaining similar memetic value/reflexivity/keynesian beauty contest properties. I suppose preferring markets with shorter rather than longer resolution dates is one approach.

Cutting off the long tail of extreme outcomes by capping the maximum KPI value would probably also have the effect of reducing the propensity for the market to approximate a keynesian beauty contest, however this comes at the cost of potentially biasing the outcome downwards due to restricting the outcome space.

Maybe using a log rather than linear mapping between KPI and share resolution value is another approach?