Nakamoto Portfolio Theory
-----------------------------
Bitcoin Valuation based on Fixed Income
Inspired by @FossGregfoss's work, we'll explore how to estimate the value of Bitcoin using credit default swaps (CDS) as fiat insurance.
You know... "for the kids". 1/n
In line with our other research pieces, this one comes with a web app so you can also run the numbers.
In this one you can change pretty much all assumptions (numbers in blue can be edited) including CDS levels, debt levels and other assumptions. nakamotoportfolio.com/apps/foss
2/n
Just like in our previous pieces, the idea is to give a framework to understand potential scenarios. Although I keep referring to them as models, these are frameworks. We provide some initial assumptions and you can create your own.
3/n
First, what is a CDS?
A Credit Default Swap (CDS) is a financial derivative that acts as insurance against the risk of a borrower defaulting on a loan. The buyer of the CDS pays a premium to the seller, who agrees to pay the face value of the loan if the borrower defaults.
4/n
CDS can be used to estimate the probability of default by analyzing the CDS spread, which reflects the annual amount the buyer must pay the seller. A wider spread indicates a higher perceived risk of default. #itsjustmath 5/n
Now, if we know the probability of a country defaulting, we can use that to calculate an insurance premium.
For example, if you drive a $10,000 car and you know that the probability of totaling that car is 1% / yr , you would be willing to pay $100 / yr in insurance.
6/n
It works the same way here. If these countries default, their debt is "gone"... So we estimate the total value of this dollar premium by multiplying each country's probability of default by its debt plus unfunded liabilities,
estimated by OpenAI's GPT-4 (verify).
7/n
Part 3.5 of @FossGregfoss paper provides an explanation of how to use CDS to value bitcoin. We made major adjustments that are listed at the app. Some are approximations. But the goal, again, is not to get to an exact price but rather understand potential outcomes.
8/n
Not surprising, the US is by far the one to watch. We got an estimate for US unfunded liabilities at 157tri (social security, medicaid, ...). Outside of the US the estimates are hard to find (I doubt China is only 500bi)
9/n
Now, these probabilities only price for a scenario of a HARD default. One where debt is just not paid.
But remember, many of these countries have large debts in their own currency. And you know what that means... brrrrrr...
10/n
Money printer equals soft default by inflation. A highly probable scenario not priced here. There's no default and yet, fiat dies a slow death (or not so slow).
11/n
In summary, this bitcoin Valuation model based on Default Probabilities (CDS) serves as a framework to understand potential outcomes and impacts on Bitcoin's valuation. Shoutout to @FossGregfoss for inspiring this analysis!
1/ I just recorded @Swan Signal with @samcallah and I briefly highlighted how I think game theory has been shaping my thinking of Bitcoin. But I don't think I did a great job in our discussion so let me try to make it better here. 🧵
2/ Bitcoin is an individual revolution. It starts at the individual level, which is a beautiful thing. Unlike other adoption cycles, where those with broader access to information or resources adopted first, Bitcoin has the most ethical and fair adoption cycle ever seen. Those who truly understand its value proposition are the ones with the advantage. Knowledge is the barrier.
3/ If that's the case, how do we know that individual interests will lead to the broader good? Enter game theory. Game theory studies strategic decision-making, where the outcome for each participant depends on the actions of others. It balances self-interest with the collective good.
Using history as a guide, we can run a simulation to forecast 1-year returns using a Monte Carlo Simulation.
Quick results:
. Average expected price in 1-year = $144k
But the average here is not a great indicator. This is the same as saying you will stick your head in a refrigerator and your feet in the oven and (on average) you are fine.
. 95% of the simulations fall between $30k and $448k.
. The worst simulation ends at $6.5k
. The best one ends at $901k
Conclusion? Prices are close to be impossible to forecast.
Anyone picking prices with certainty did not run the numbers.
The good news?
77% of the simulations are positive
The returns compared to the vol are among some of the best you can find in any asset class.
[1/n]
"But Bitcoin is too volatile."
Would you rather invest in bonds? Fun... Let's run the numbers!
Here are results for TLT (iShares 20yr Treasury ETF).
. Average returns are 3% (compared to 70% on the Bitcoin Monte Carlo)
. 60% of times you get positive returns (compared to 70%)
. 95% of values give you a return between -25% and +43% (yes, bonds have volatility).
14% annualized vol with 3% annualized returns
compared to BTC:
70% annualized vol with 72% annualized returns
5x the vol = 24x the returns
I'll take the vol every day on BTC....
[2/n]
"But bitcoin is not a store of value"
Here are the results for GLD (Gold ETF)
9% annualized return with 17% vol
I would take BTC's 70% vol (4x higher) to get BTC's returns (7.7x higher) all day long.
2/ The concept of liquidity in real estate is tricky.
Unlike stocks or bitcoin, you can't sell a part of your house instantly. The infrequent updating of your property's value can mask its true volatility.
Investors usually analyze returns only when they buy or sell a home.
Nakamoto Portfolio Theory
-----------------------------
How are bitcoin's historical returns distributed?
Turns out there's nothing 'normal' about them.
We've tumble down a rabbit hole, ending up deep in the world of #Bitcoin forecasting.
(1/n)
Why does this matter? Good forecasting helps us understand risk.
To run Monte Carlo simulations (a key forecasting tool), we need a statistical distribution that accurately mirrors Bitcoin's historical returns. But first, we need to see if we can even find one that fits!
(2/n)
After analyzing various statistical distributions, a generalized hyperbolic distribution seemed to fit Bitcoin returns best. It's a flexible model that can handle Bitcoin's unique characteristics better.
(3/n)
The Rat Poison Portfolio 🐀
An equally weighted portfolio of Berkshire Hathaway, Microsoft, JP Morgan and BlackRock
Guess what happens when we add #bitcoin to this portfolio?
Higher volatility?
Higher drawdowns?
Higher risk?
Let's run the numbers
Since 2014, that portfolio returned 16% annualized (assuming quarterly rebalancing). Not bad.
BUT by adding a 2.5% allocation to Bitcoin the Portfolio increases returns to ~20% WITH reduced risk.
Bitcoin would have actually reduced the drawdown of the original portfolio. This may be a bit counterintuitive. How can it be that including such a volatile asset actually reduces risk?
The answer is below. Bitcoin has VERY low correlation to these assets.
Nakamoto Portfolio Theory
-----------------------------
The Nakamoto Portfolio Theory is a set of frameworks to help investors understand the impact an emerging asset like Bitcoin has a portfolio allocation.