Litecoin – Nick – Medium

Litecoin – Nick – Medium

The Cult of the Drunks’ Dog

This article explores if there is a stock-to-flow relationship to Litecoin value. A log-log model is tested for statistical validity against the least squares assumptions, for stationarity in each variable and for potential spurious relationships. A Vector Error Correction Model (VECM) is built and tested against the stock-to-flow model. An interaction with the Bitcoin price is included in the terms and explored. These models provide sufficient evidence to reject the hypothesis that stock-to-flow is at all related to the value of Litecoin, but insufficient evidence to reject that the value of Litecoin is cointegrated with the value of Litecoin.

This is not financial advice.

We have already established that there is a non-spurious relationship between stock-to-flow (SF) and Bitcoin value [1, 2]. In [2] we show that SF is a non-random variable, akin to a road that the value of Bitcoin wanders upon. In [2] we explored the concept of cointegration. We found there was a cointegrating relationship between SF and Bitcoin value.

An old saying is “birds of a feather, flock together”, usually with some sort of undercurrent of social disilussionment. In this saying though we find an ultimate truth-similarities in consumers lead to similarities in behaviour. People buying Bitcoin are more able to buy other coins, simply because they are already part of the gang. One of the crew. They have been indoctrinated into the cult of cryptocurrency.

This false indoctrination is fought with vigor by those whom understand the true value of Bitcoin; that is — that it is ultimately not only driven by its’ use case, not only impelled from the fact that is digitally scarce, and nor is it induced solely by the truth that each Bitcoin is mined with unforgeable costliness. No, it is the interaction of these very special ingredients that gives Bitcoin its’ value.

Many separate patches of a beautiful Persian rug, does not a beautiful Persian rug make. Without all of the intricate complexities and interactions woven together as one, the patches would fray and fall apart at the slightest hint of any traffic.

Now, if Bitcoin is the whole damn rug, Litecoin is a cheap Aldi knock off. It attempts to replicate the whole of Bitcoin, if not for altering a few (essentially inconsequential) variables here and there. One of those occultists we described earlier might see this knock off version as a “cheaper” version of the same product. Of course we know, the product (Bitcoin) is a protocol and expensive or cheap isn’t really relevant in these terms. But let’s chase our indoctrinated friend through the cloud of opiate smoke and see if we can’t catch the shitcoin dragon.

Once again we build a model of the form established in [2]:

where β is a 2 x 1 matrix (consisting of the constant and coefficient).

Above we have estimated β as:

so our equation is:

where MC is the Litecoin market cap and SF is the Litecoin stock to flow ratio.

How do we interpret this? Well, this is something that I think is not well communicated in other loglog models.

Focusing on the effect of SF — take two values of SF, SF1 and SF2, where SF1 is the “current” SF, and SF2 is the SF at which you want to evaluate the effect, then the equation above yields:

Which can be simplified to:

leading to:

This tells us that as long as the ratio of the two SF values, SF2/SF1 stays the same, the expected ratio of the outcome variable MC stays the same. For example, we can say that for any 10% increase in SF, the expected ratio of the market cap will be (1.10)^β=(1.10)^1.6716273=1.1727168.

In other words, we expect about a 17.3% increase in market cap when stock to flow increases by 10%. IF the relationship was valid. However, due to the faulty assumptions identified below, it would be relatively unwise to use this estimate of the relationship.

Figure 1: Model performance — there is a mild correlation
Figure 2: Leverage v residuals — not too much to see here. Basically no points are asserting undue influence in the regression. That said, the residual seems a little high.

Published at Sat, 26 Oct 2019 02:10:29 +0000

{flickr|100|campaign}

Spread the word. Share this post!