SlideShare a Scribd company logo
1 of 42
Download to read offline
The use of Bitcoin for
portfolio optimization
Federico Tenga
federico@chainside.net
Introduction
For many people Bitcoin is considered the ideal store of value,
but most investor still lack to see the value of this asset due to
its technical complexity.
The scope of this study is to analyze Bitcoin strictly under a
financial point of view and show the benefits it brings to
optimize any investment portfolio and have more awareness of
the risk-reward profile of this new asset.
Index
β€’ Bitcoin Supply
β€’ Volatility
β€’ Correlation
β€’ Expected Return
β€’ Daily returns distribution analysis
β€’ Bitcoin for Portfolio Optimization
Bitcoin Supply
Figure 1 Bitcoin supply and inflation
Gold Supply
Figure 2 Gold supply. Source: Number Sleuth ("All The World's Gold Facts")
Volatility
The volatility, measured by the standard deviation, in finance is the degree of variation of
price of an asset, and it can be derived using historical market price data.
To find the volatility of bitcoin, we will compute the standard deviation of daily returns
using the following formula:
𝜎 =
1
𝑛
෍
𝑖=1
𝑛
(π‘₯𝑖 βˆ’ πœ‡)2
Where:
πœ‡ = π‘šπ‘’π‘Žπ‘› π‘£π‘Žπ‘™π‘’π‘’ π‘œπ‘“ π‘‘β„Žπ‘’ π‘‘π‘Žπ‘–π‘™π‘¦ π‘Ÿπ‘’π‘‘π‘’π‘Ÿπ‘›π‘ 
From April 2016 to May 2017 the average daily volatility of bitcoin is 2.67%
Annualized Volatility
To compute the annualized standard deviation, we have to multiply the daily
standard deviation for the square root of the number of trading days.
For traditional financial markets the number of trading days in a year is usually
about 250, but since bitcoin is traded 24/7, every day of the year, we have about
365 trading days, consequentially the formula to compute the annualised
standard deviation will be:
π‘Žπ‘›π‘›π‘’π‘Žπ‘™π‘–π‘§π‘’π‘‘ 𝜎 = 𝜎 365
Resulting in an annualized standard deviation of 51.03%.
Volatility over time
Correlation
The correlation measures the dependence between two variables. The Pearson
correlation, considered to be the β€œtraditional” correlation, is calculated with the
following formula:
𝜌 π‘₯,𝑦=
πΆπ‘œπ‘£(π‘₯, 𝑦)
𝜎 π‘₯ 𝜎 𝑦
Where:
πΆπ‘œπ‘£ π‘₯, 𝑦 = 𝐸( π‘₯ βˆ’ 𝐸 π‘₯ 𝑦 βˆ’ 𝐸 𝑦 )
𝜎 π‘₯ = 𝐸[π‘₯]2βˆ’ (𝐸 π‘₯ )2
If we try to compute the Pearson correlation of bitcoin returns with some major asset
classes we can find some very interesting results
p-value
When you perform a hypothesis test in statistics, a p-value helps you determine the significance of
your results. he p-value is a number between 0 and 1 and interpreted in the following way: A small
p-value (typically ≀ 0.05) indicates strong evidence against the null hypothesis, so you reject the
null hypothesis.
Bitcoin vs S&P 500 Correlation
The Pearson correlation of bitcoin with S&P 500, over the timespan analysed, results to
be 1.57%, which is extremely low. Moreover, the p-value of the Pearson correlation is a
very high 0.582, meaning that we cannot even easily assume that the correlation is
different from zero.
Figure 5 Bitcoin vs S&P 500 daily returns
Bitcoin vs MSCI Emerging Markets Index
The Pearson correlation of bitcoin with MSCI Emerging Markets Index, over the timespan
analysed, results to be 2.69%. Still a very low value, and similarly to what we have seen
with S&P 500 the high p-value of 0.345 suggests that the result is not significant so we can
even be sure that the correlation is not actually zero.
Figure 6Bitcoin and MSCI Emerging Markets Index Daily Returns
Bitcoin vs Oil Correlation
The Pearson correlation of bitcoin with WTI Crude Oil prices, over the timespan analysed,
results to be 0.8%, and just like in the cases seen above the data cannot be considered
significant due to the high p-value of 0.789.
Bitcoin vs Gold Correlation
The Pearson correlation of bitcoin with gold, over the timespan analysed, results to be
just 1.7%, and just like we have seen before the p-value of 0.558 suggests that the results
cannot be considerate significant to assume that the correlation is different from zero.
Correlation and portfolio volatility
The non-existent correlation means that bitcoin does not share systematic risk with other
asset classes, making it a great tool for diversification of a portfolio. Indeed, this is
evident looking at the formula of a portfolio volatility:
π‘ƒπ‘œπ‘Ÿπ‘‘π‘“π‘œπ‘™π‘–π‘œ π‘‰π‘œπ‘™π‘Žπ‘‘π‘–π‘™π‘–π‘‘π‘¦ = 𝑆𝐷 π‘Ž
2
βˆ— π‘Šπ‘Ž
2
+ 𝑆𝐷 𝑏
2
βˆ— π‘Šπ‘
2
+ πΆπ‘œπ‘Ÿπ‘Ÿπ‘’π‘™π‘Žπ‘‘π‘–π‘œπ‘› 𝐸𝑓𝑓𝑒𝑐𝑑
Where:
SD = standard deviation of the asset
W = weight of the asset in the portfolio
Expected Returns (1/2)
A common method to attempt to calculate the assets appropriate return is the Capital
Asset Pricing Model (CAPM), which describes the relationship between systematic risk and
expected return of an asset. The CAPM formula for calculating the expected return of an
asset given its risk is as follows:
ra = rf + Ξ²a (rm – rf)
Where:
rf = risk-free rate
Ξ²a = beta of the asset
rm = expected return of the market
Unfortunately, due to the lack of correlation with a benchmark portfolio, it is not possible
to calculate the Ξ² of bitcoin
We can use instead historical data to estimate future trends, but it is not easy to decide
which timespan can be useful for our analysis. Old trading data are less significant due to
the high market manipulation of the early days, a good option to use post-MtGox data.
We can find the mean daily return of the period April 2014 – May 2017 using the following
formula:
πœ‡ =
1
𝑛
෍
𝑖=1
𝑛
𝑋𝑖
Which gives us a mean daily return of 0.13%
Expected Returns (2/2)
Annualized expected returns
We can than derive the annualized expected return raising πœ‡ to the power of number of
trading days in a year, using the following formula:
πœ‡ π‘Žπ‘›π‘›π‘’π‘Žπ‘™π‘–π‘§π‘’π‘‘ = (1 + πœ‡ )365βˆ’1
Once again since bitcoin is traded 24/7, differently from any other asset we will 365
trading days. Which gives us an annualized expected return of 60.35%
Daily Returns Distribution
During the period from April 2014 to June 2017, we can see an average daily return of
0.2% with a standard deviation on 3.06%
However, it is necessary to proceed with a normality test to discover if the standard
returns can be described by a Gaussian distribution.
Shapiro–Wilk test
In statistics, the Shapiro–Wilk test tests the null hypothesis that a sample x1,...,xn originated
from a normally distributed population. The test is:
π‘Š =
σ𝑖=1
𝑛
aixi 2
σ𝑖=1
𝑛
(xi βˆ’ Η‰π‘₯)
2
π‘Ž1, … , π‘Ž 𝑛 =
π‘šT π‘‰βˆ’1
(π‘š 𝑑 π‘‰βˆ’1 π‘‰βˆ’1 π‘š)1/2
Where:
π‘š = (π‘š1, … , π‘š 𝑛)T
and m1, ..., mn are the expected values of the order statistics of independent and identically
distributed random variables sampled from the standard normal distribution, and V is the
covariance matrix of those order statistics.
Shapiro–Wilk test
Doing the Shapiro-Wink test on the bitcoin daily returns for the period April 2014 – April 2017
we find the following results:
As the computed p-value is lower than the significance level alpha=0.05, one should reject the
null hypothesis (the sample follows a Normal distribution), and accept the alternative
hypothesis (the sample does not follow a Normal distribution).
The risk to reject the null hypothesis while it is true is lower than 0.01%
W 0.868
p-value (Two-tailed) < 0.0001
alpha 0.05
Skewness analysis
The Skewness is a measure to analyse the symmetry of the distribution. If the coefficient of
skewness is positive the distribution is skewed right, if it is negative the distribution is skewed
left.
The coefficient of skewness of a data set is:
skewness: 𝑔1 = π‘š3/π‘š2
3/2
Where:
π‘š3 =
Οƒ x βˆ’ ΰ΄€x 3
𝑛
π‘Žπ‘›π‘‘ π‘š2 =
Οƒ x βˆ’ ΰ΄€x 2
𝑛
A normal distribution has a skewness coefficient of zero (perfect symmetry). The distribution
of bitcoin daily returns from April 2014 to April 2017 has skewness coefficient of 0.024, which
means that the distribution is almost perfectly symmetrical, indicating a good market
efficiency.
Kurtosis Analysis
Another common measure of shape is the Kurtosis which provides information on how the
data are spread among the peak and the tails of the distribution.
Higher kurtosis also means that more of the variance is the result of infrequent extreme
deviations, as opposed to frequent modestly sized deviations. The reference standard is a
normal distribution, which has a kurtosis of 3.
The coefficient of kurtosis of a dataset is:
kurtosis: π‘Ž4 = π‘š4/π‘š2
2
Where:
π‘š4 =
Οƒ x βˆ’ ΰ΄€x 4
𝑛
π‘Žπ‘›π‘‘ π‘š2 =
Οƒ x βˆ’ ΰ΄€x 2
𝑛
The distribution of bitcoin daily returns from April 2014 to April 2017 has kurtosis
coefficient of 12.439, meaning that most of the variance is caused by infrequent extreme
deviations. This can also be seen as volatility on volatility.
Autocorrelation Analysis
In order to understand if the bitcoin market is efficient, it is useful to test the autocorrelation
of the daily returns. If the returns result to be autocorrelated, it indicates that the price
movements can be predicted and the market is not efficient.
To test the autocorrelation, we will use the Durbin-Watson statistic, which always has a value
d between 0 and 4, where 2 means there is no autocorrelation, 0 means that there is positive
correlation and 4 means that there is negative correlation.
The value d of the Durbin-Watson statistic is calculated with the following formula:
d =
Οƒ 𝑑=2
𝑇
(𝑒𝑑 βˆ’ π‘’π‘‘βˆ’1)2
Οƒ 𝑑=1
𝑇
𝑒𝑑
2
Where:
et = residual associated with the observation at time t
Running the Durbin-Watson test on the bitcoin daily returns from April 2014 to April 2017 we
obtain a d value of 2.094, meaning that there is almost no autocorrelation and the market is
efficient.
Bitcoin Price Resilience
In the history of Bitcoin there have been many big events that caused high volatility periods and
the collapse of the price. It is interesting to notice that even if big events that can potentially
damage the Bitcoin ecosystem keep happening, the impact is becoming less significant.
Figure 13 Bitcoin price change comparison after exchanges' losses
Portfolio Optimization
Portfolio Optimization
A general solution to portfolio selection problem was proposed by Markowitz in
1952, building the foundation of the Modern Portfolio Theory (MPT) in the coming
decades. The MPT is based on the utility maximization concept, emphasizing the
trade-off between risk and return.
One of the assumption of the MPT is that investors are risk-adverse, giving two
portfolios with similar returns a rational investor will always choose the one with
lower risk, and any extra risk will have to be compensated.
With the Modern Portfolio Theory, it is possible to keep a certain level of risk while
maximising the returns, or keeping the returns fixed while minimising the risk.
An investor can then generate a so called efficient frontier where the optimal asset
allocation is achieved.
Portfolio Returns
An asset return can be easily defined as the price difference over a time period divided
the price of the asset at the beginning of the period, as described by the following
formula:
𝑅𝑖,𝑑 =
𝑃𝑖,𝑑 βˆ’ 𝑃𝑖,π‘‘βˆ’1
𝑃𝑖,π‘‘βˆ’1
Where:
𝑃𝑖,𝑑 = π‘π‘Ÿπ‘–π‘π‘’ π‘Žπ‘‘ π‘‘β„Žπ‘’ 𝑒𝑛𝑑 π‘œπ‘“ π‘‘β„Žπ‘’ π‘π‘’π‘Ÿπ‘–π‘œπ‘‘
𝑃𝑖,π‘‘βˆ’1 = π‘π‘Ÿπ‘–π‘π‘’ π‘Žπ‘‘ π‘‘β„Žπ‘’ 𝑏𝑒𝑔𝑖𝑛𝑛𝑖𝑛𝑔 π‘œπ‘“ π‘‘β„Žπ‘’ π‘π‘’π‘Ÿπ‘–π‘œπ‘‘
In a portfolio, the return is simply the weighted average of the return of the assets in the
portfolio, as described by the following formula:
𝑅 𝑝,𝑑 = ෍
𝑛=1
𝑁
𝑀𝑖 𝑅𝑖,𝑑
Where:
𝑀𝑖 = π‘π‘Ÿπ‘œπ‘π‘œπ‘Ÿπ‘‘π‘–π‘œπ‘› π‘œπ‘“ π‘‘β„Žπ‘’ π‘π‘œπ‘Ÿπ‘‘π‘“π‘œπ‘™π‘–π‘œ 𝑖𝑛𝑣𝑒𝑠𝑑𝑒𝑑 𝑖𝑛 π‘‘β„Žπ‘’ π‘Žπ‘ π‘ π‘’π‘‘ 𝑖
Portfolio Risk
We can measure the risk of the portfolio using its standard deviation. For an asset,
the standard deviation can be found, as we have previously seen, using the
following formula:
𝜎 =
1
𝑛
෍
𝑖=1
𝑛
(π‘₯𝑖 βˆ’ πœ‡)2
Portfolio Risk
In a portfolio, we have to consider the standard deviation and the correlation coefficient of all
the assets that compose the portfolio, we can then compute the standard deviation of the
portfolio using the following general formula:
𝜎 𝑝 = ෍
𝑖=1
𝑁
෍
𝑗=1
𝑁
𝑀𝑖 𝑀𝑗 πœŽπ‘– πœŽπ‘—Οπ‘–π‘—
Where:
w = weight of the asset
Οƒ = standard deviation of the asset
ρ = correlation coefficients of the assets
Considering a portfolio with two assets, the resulting standard deviation will be:
𝜎 𝑝 = 𝑀1
2
𝜎1
2
+ 𝑀2
2
𝜎2
2
+ 2𝑀1 𝜎2ρ1,2
The Sharpe Ratio
The Sharpe ratio is a tool to examine the performance of an investment, by measuring
how well the return of an asset compensates for its risk, making also possible to compare
assets with very different levels of risk.
The Sharpe ratio is calculated as follows:
𝑆 π‘Ž =
𝐸[𝑅 π‘Ž βˆ’ 𝑅 𝑏]
𝜎 π‘Ž
Where:
Ra = return of asset
Rb = return of benchmark asset, usually the risk-free rate
Οƒa = standard deviation of the asset
The Sharpe ratio can be used to identify the optimal asset allocation in a portfolio, the
one that provides the highest return per unit of deviation.
Bitcoin – S&P 500 portfolio
To efficiently build a two assets portfolio using S&P 500 and bitcoin, it is first needed to
quickly analyse the two separate assets:
We can already notice that bitcoin has a higher Sharpe ratio than S&P 500, considering
that we have also previously seen that the two assets have no correlation we can already
deduce that bitcoin will significantly improve the returns of the portfolio per unit of
deviation.
S&P 500 bitcoin
Daily mean SD 0.80% 2.67%
Annualized SD 12.66% 51.03%
Daily mean return 0.04% 0.13%
Annualized return 9.46% 60.35%
Sharpe ratio 0.68 1.15
Bitcoin – S&P 500 portfolio
Btc% S&P 500% SD Return Sharpe
0% 100% 12.66% 9.46% 0.6838
30% 70% 17.69% 24.73% 1.3532
100% 0% 51.03% 60.35% 1.1673
In the table below we can see the return and the risk for each possible allocation of
the two assets:
As we can see, the portfolio with the highest Sharpe ratio would be the one with
30% allocated in bitcoin an 70% allocated in S&P 500.
Gold – S&P 500 portfolio
The fact that bitcoin could improve the performance of the S&P 500 portfolio was easily
predictable due to its low correlation coefficient, so it is more interesting to see how the
performance improvment provided by bitcoin is compared to the improvement provided
by gold, an asset that bitcoin has the ambition to replace, at least partially, in its safe
heaven role for investors.
Gold seems to be very good for diversification purposes due to its low correlation with
S&P 500 (only 0.03), but as we can see in the table below, its poor performances in the
past years greatly affect its Sharpe ratio, which is significantly lower compared to
bitcoin’s ratio.
S&P 500 Gold
Daily mean SD 0.80% 0.79%
Annualized SD 12.66% 12.57%
Daily mean return 0.04% 0.013%
Annualized return 9.46% 3.37%
Sharpe ratio 0.68 0.20
Gold – S&P 500 portfolio
After trying to plot every possible portfolio combing the two assets we can find that the
allocation that maximise the Sharpe ratio has 21% of gold. However, the contribution to
improve the performance of the portfolio is modest, the Sharpe ratio of the optimal
portfolio is just slightly higher compared to the S&P ratio of 0.68, while bitcoin provided
a much more significant improvement, almost doubling it.
Gold% S&P 500% SD Return Sharpe
0% 100% 12.66% 9.46% 0.6838
21% 79% 10.42% 8.18% 0.7081
100% 0% 12.57% 3.37% 0.2044
Too much risk?
Choosing an asset allocation that maximise the Sharpe ratio does not allow an investor to
set its desired level of risk. This problem can be solved introducing a risk-free asset,
which let an investor maximise the Sharpe ratio and set the portfolio risk level.
Introducing the risk-free asset, the expected return and the standard deviation of a
portfolio are:
𝐸(𝑅 𝑝) = 𝑀1 𝐸(𝑅1) + (1 βˆ’ 𝑀1) 𝑅 𝑓
Οƒ 𝑝 = 𝑀1
2
Οƒ1
2
Where:
𝑅1= return of the portfolio with only risky assets
𝑅 𝑓= return of the risk-free asset
𝑀1= weight of the risky assets
𝜎1=standard deviation of the portfolio with only risky assets
Capital Allocation Line
After having found the portfolio that maximise the Sharpe ratio, an investor
with a lower risk preference can allocate part of his capital in a risk-free asset,
reducing his exposure, and an investor with a higher risk preference can borrow
at the risk-free rate and invest his capital with leverage in the optimal
portfolio. These practices generate the Capital Allocation Line (CAL), defined
by the following formula:
𝐸 𝑅 𝑝 = 𝑅𝑓 +
𝐸(𝑅1) βˆ’ 𝑅𝑓
Οƒ1
Οƒ 𝑝
Conclusion
Whatever is your risk preference, Bitcoin can help you
to optimize your portfolio
Federico Tenga
COO at Chainside
federico@chainside.net

More Related Content

What's hot

Comparative study of various approaches for transaction Fraud Detection using...
Comparative study of various approaches for transaction Fraud Detection using...Comparative study of various approaches for transaction Fraud Detection using...
Comparative study of various approaches for transaction Fraud Detection using...Pratibha Singh
Β 
Retirement Portfolio Financial Analysis - Graduate Project
Retirement Portfolio Financial Analysis - Graduate ProjectRetirement Portfolio Financial Analysis - Graduate Project
Retirement Portfolio Financial Analysis - Graduate ProjectMedicishi Taylor
Β 
Inference with big data: SCECR 2012 Presentation
Inference with big data: SCECR 2012 PresentationInference with big data: SCECR 2012 Presentation
Inference with big data: SCECR 2012 PresentationGalit Shmueli
Β 
Assignment #3 10.19.14
Assignment #3 10.19.14Assignment #3 10.19.14
Assignment #3 10.19.14Lourdes Greenwood
Β 
Credit card fraud detection
Credit card fraud detectionCredit card fraud detection
Credit card fraud detectionvineeta vineeta
Β 
FACTOR analysis (July 2014 updated)
FACTOR analysis (July 2014 updated)FACTOR analysis (July 2014 updated)
FACTOR analysis (July 2014 updated)Michael Ling
Β 
Crime Analysis using Regression and ANOVA
Crime Analysis using Regression and ANOVACrime Analysis using Regression and ANOVA
Crime Analysis using Regression and ANOVATom Donoghue
Β 
Frequentist inference only seems easy By John Mount
Frequentist inference only seems easy By John MountFrequentist inference only seems easy By John Mount
Frequentist inference only seems easy By John MountChester Chen
Β 
Credit card fraud detection using machine learning Algorithms
Credit card fraud detection using machine learning AlgorithmsCredit card fraud detection using machine learning Algorithms
Credit card fraud detection using machine learning Algorithmsankit panigrahy
Β 
Statistical-Process-Control-Analysis-Unraveled_updated210
Statistical-Process-Control-Analysis-Unraveled_updated210Statistical-Process-Control-Analysis-Unraveled_updated210
Statistical-Process-Control-Analysis-Unraveled_updated210pbaxter
Β 

What's hot (12)

Comparative study of various approaches for transaction Fraud Detection using...
Comparative study of various approaches for transaction Fraud Detection using...Comparative study of various approaches for transaction Fraud Detection using...
Comparative study of various approaches for transaction Fraud Detection using...
Β 
Retirement Portfolio Financial Analysis - Graduate Project
Retirement Portfolio Financial Analysis - Graduate ProjectRetirement Portfolio Financial Analysis - Graduate Project
Retirement Portfolio Financial Analysis - Graduate Project
Β 
Inference with big data: SCECR 2012 Presentation
Inference with big data: SCECR 2012 PresentationInference with big data: SCECR 2012 Presentation
Inference with big data: SCECR 2012 Presentation
Β 
Assignment #3 10.19.14
Assignment #3 10.19.14Assignment #3 10.19.14
Assignment #3 10.19.14
Β 
PyGotham 2016
PyGotham 2016PyGotham 2016
PyGotham 2016
Β 
Credit card fraud detection
Credit card fraud detectionCredit card fraud detection
Credit card fraud detection
Β 
FACTOR analysis (July 2014 updated)
FACTOR analysis (July 2014 updated)FACTOR analysis (July 2014 updated)
FACTOR analysis (July 2014 updated)
Β 
Crime Analysis using Regression and ANOVA
Crime Analysis using Regression and ANOVACrime Analysis using Regression and ANOVA
Crime Analysis using Regression and ANOVA
Β 
Frequentist inference only seems easy By John Mount
Frequentist inference only seems easy By John MountFrequentist inference only seems easy By John Mount
Frequentist inference only seems easy By John Mount
Β 
Credit card fraud detection using machine learning Algorithms
Credit card fraud detection using machine learning AlgorithmsCredit card fraud detection using machine learning Algorithms
Credit card fraud detection using machine learning Algorithms
Β 
IQ versus RQ
IQ versus RQIQ versus RQ
IQ versus RQ
Β 
Statistical-Process-Control-Analysis-Unraveled_updated210
Statistical-Process-Control-Analysis-Unraveled_updated210Statistical-Process-Control-Analysis-Unraveled_updated210
Statistical-Process-Control-Analysis-Unraveled_updated210
Β 

Similar to The Use of Bitcoin for Portfolio Optimization

Sensitivity &amp; Scenario Analysis
Sensitivity &amp; Scenario AnalysisSensitivity &amp; Scenario Analysis
Sensitivity &amp; Scenario AnalysisDr. Rana Singh
Β 
A high level overview of all that is Analytics
A high level overview of all that is AnalyticsA high level overview of all that is Analytics
A high level overview of all that is AnalyticsRamkumar Ravichandran
Β 
Data Analysison Regression
Data Analysison RegressionData Analysison Regression
Data Analysison Regressionjamuga gitulho
Β 
Project 3(234A)
Project 3(234A)Project 3(234A)
Project 3(234A)Kaishi Wang
Β 
Study on Evaluation of Venture Capital Based onInteractive Projection Algorithm
	Study on Evaluation of Venture Capital Based onInteractive Projection Algorithm	Study on Evaluation of Venture Capital Based onInteractive Projection Algorithm
Study on Evaluation of Venture Capital Based onInteractive Projection Algorithminventionjournals
Β 
Risk Ana
Risk AnaRisk Ana
Risk Ananeetu goel
Β 
Portfolio risk and retun project
Portfolio risk and retun projectPortfolio risk and retun project
Portfolio risk and retun projectRohit Sethi
Β 
Fin415 Week 2 Slides
Fin415 Week 2 SlidesFin415 Week 2 Slides
Fin415 Week 2 Slidessmarkbarnes
Β 
Quantitative Methods for Lawyers - Class #20 - Regression Analysis - Part 3
Quantitative Methods for Lawyers - Class #20 - Regression Analysis - Part 3Quantitative Methods for Lawyers - Class #20 - Regression Analysis - Part 3
Quantitative Methods for Lawyers - Class #20 - Regression Analysis - Part 3Daniel Katz
Β 
Summer 07-mfin7011-tang1922
Summer 07-mfin7011-tang1922Summer 07-mfin7011-tang1922
Summer 07-mfin7011-tang1922stone55
Β 
NBS8001-Nikolaos Sfoungaros-Three Stars Analysts report
NBS8001-Nikolaos Sfoungaros-Three Stars Analysts reportNBS8001-Nikolaos Sfoungaros-Three Stars Analysts report
NBS8001-Nikolaos Sfoungaros-Three Stars Analysts reportNikolaos Sfoungaros
Β 
MomentumFinal
MomentumFinalMomentumFinal
MomentumFinalTom Wilson
Β 
Machine learning algorithms and business use cases
Machine learning algorithms and business use casesMachine learning algorithms and business use cases
Machine learning algorithms and business use casesSridhar Ratakonda
Β 
Combining Economic Fundamentals to Predict Exchange Rates
Combining Economic Fundamentals to Predict Exchange RatesCombining Economic Fundamentals to Predict Exchange Rates
Combining Economic Fundamentals to Predict Exchange RatesBrant Munro
Β 

Similar to The Use of Bitcoin for Portfolio Optimization (20)

Bivariate Regression
Bivariate RegressionBivariate Regression
Bivariate Regression
Β 
Sensitivity &amp; Scenario Analysis
Sensitivity &amp; Scenario AnalysisSensitivity &amp; Scenario Analysis
Sensitivity &amp; Scenario Analysis
Β 
A high level overview of all that is Analytics
A high level overview of all that is AnalyticsA high level overview of all that is Analytics
A high level overview of all that is Analytics
Β 
Data Analysison Regression
Data Analysison RegressionData Analysison Regression
Data Analysison Regression
Β 
Project 3(234A)
Project 3(234A)Project 3(234A)
Project 3(234A)
Β 
Fm5
Fm5Fm5
Fm5
Β 
Study on Evaluation of Venture Capital Based onInteractive Projection Algorithm
	Study on Evaluation of Venture Capital Based onInteractive Projection Algorithm	Study on Evaluation of Venture Capital Based onInteractive Projection Algorithm
Study on Evaluation of Venture Capital Based onInteractive Projection Algorithm
Β 
Risk Ana
Risk AnaRisk Ana
Risk Ana
Β 
Decision theory
Decision theoryDecision theory
Decision theory
Β 
Portfolio risk and retun project
Portfolio risk and retun projectPortfolio risk and retun project
Portfolio risk and retun project
Β 
Graduate RP
Graduate RPGraduate RP
Graduate RP
Β 
Graduate RP
Graduate RPGraduate RP
Graduate RP
Β 
Fin415 Week 2 Slides
Fin415 Week 2 SlidesFin415 Week 2 Slides
Fin415 Week 2 Slides
Β 
Risk notes ch12
Risk notes ch12Risk notes ch12
Risk notes ch12
Β 
Quantitative Methods for Lawyers - Class #20 - Regression Analysis - Part 3
Quantitative Methods for Lawyers - Class #20 - Regression Analysis - Part 3Quantitative Methods for Lawyers - Class #20 - Regression Analysis - Part 3
Quantitative Methods for Lawyers - Class #20 - Regression Analysis - Part 3
Β 
Summer 07-mfin7011-tang1922
Summer 07-mfin7011-tang1922Summer 07-mfin7011-tang1922
Summer 07-mfin7011-tang1922
Β 
NBS8001-Nikolaos Sfoungaros-Three Stars Analysts report
NBS8001-Nikolaos Sfoungaros-Three Stars Analysts reportNBS8001-Nikolaos Sfoungaros-Three Stars Analysts report
NBS8001-Nikolaos Sfoungaros-Three Stars Analysts report
Β 
MomentumFinal
MomentumFinalMomentumFinal
MomentumFinal
Β 
Machine learning algorithms and business use cases
Machine learning algorithms and business use casesMachine learning algorithms and business use cases
Machine learning algorithms and business use cases
Β 
Combining Economic Fundamentals to Predict Exchange Rates
Combining Economic Fundamentals to Predict Exchange RatesCombining Economic Fundamentals to Predict Exchange Rates
Combining Economic Fundamentals to Predict Exchange Rates
Β 

More from Federico Tenga

Decentralised Applications on Bitcoin
Decentralised Applications on BitcoinDecentralised Applications on Bitcoin
Decentralised Applications on BitcoinFederico Tenga
Β 
Smart Contracts Technical Overview - Meetup Roma - 17/09/19
Smart Contracts Technical Overview - Meetup Roma - 17/09/19Smart Contracts Technical Overview - Meetup Roma - 17/09/19
Smart Contracts Technical Overview - Meetup Roma - 17/09/19Federico Tenga
Β 
Sustainability of a multi blockchain ecosystem
Sustainability of a multi blockchain ecosystemSustainability of a multi blockchain ecosystem
Sustainability of a multi blockchain ecosystemFederico Tenga
Β 
State Smart Contract Technologies
State Smart Contract TechnologiesState Smart Contract Technologies
State Smart Contract TechnologiesFederico Tenga
Β 
Blockchain for IoT
Blockchain for IoTBlockchain for IoT
Blockchain for IoTFederico Tenga
Β 
Bitcoin Revolution
Bitcoin RevolutionBitcoin Revolution
Bitcoin RevolutionFederico Tenga
Β 
Bitcoin Fork Wars: from xt to 2 x
Bitcoin Fork Wars: from xt to 2 xBitcoin Fork Wars: from xt to 2 x
Bitcoin Fork Wars: from xt to 2 xFederico Tenga
Β 
Bitcoin and blockchain talk - Pavia
Bitcoin and blockchain talk - PaviaBitcoin and blockchain talk - Pavia
Bitcoin and blockchain talk - PaviaFederico Tenga
Β 
Slide Federico Tenga - Conferenza Blockchain Roma 17-06-16
Slide Federico Tenga - Conferenza Blockchain Roma 17-06-16Slide Federico Tenga - Conferenza Blockchain Roma 17-06-16
Slide Federico Tenga - Conferenza Blockchain Roma 17-06-16Federico Tenga
Β 

More from Federico Tenga (9)

Decentralised Applications on Bitcoin
Decentralised Applications on BitcoinDecentralised Applications on Bitcoin
Decentralised Applications on Bitcoin
Β 
Smart Contracts Technical Overview - Meetup Roma - 17/09/19
Smart Contracts Technical Overview - Meetup Roma - 17/09/19Smart Contracts Technical Overview - Meetup Roma - 17/09/19
Smart Contracts Technical Overview - Meetup Roma - 17/09/19
Β 
Sustainability of a multi blockchain ecosystem
Sustainability of a multi blockchain ecosystemSustainability of a multi blockchain ecosystem
Sustainability of a multi blockchain ecosystem
Β 
State Smart Contract Technologies
State Smart Contract TechnologiesState Smart Contract Technologies
State Smart Contract Technologies
Β 
Blockchain for IoT
Blockchain for IoTBlockchain for IoT
Blockchain for IoT
Β 
Bitcoin Revolution
Bitcoin RevolutionBitcoin Revolution
Bitcoin Revolution
Β 
Bitcoin Fork Wars: from xt to 2 x
Bitcoin Fork Wars: from xt to 2 xBitcoin Fork Wars: from xt to 2 x
Bitcoin Fork Wars: from xt to 2 x
Β 
Bitcoin and blockchain talk - Pavia
Bitcoin and blockchain talk - PaviaBitcoin and blockchain talk - Pavia
Bitcoin and blockchain talk - Pavia
Β 
Slide Federico Tenga - Conferenza Blockchain Roma 17-06-16
Slide Federico Tenga - Conferenza Blockchain Roma 17-06-16Slide Federico Tenga - Conferenza Blockchain Roma 17-06-16
Slide Federico Tenga - Conferenza Blockchain Roma 17-06-16
Β 

Recently uploaded

Stock Market Brief Deck for 4/24/24 .pdf
Stock Market Brief Deck for 4/24/24 .pdfStock Market Brief Deck for 4/24/24 .pdf
Stock Market Brief Deck for 4/24/24 .pdfMichael Silva
Β 
02_Fabio Colombo_Accenture_MeetupDora&Cybersecurity.pptx
02_Fabio Colombo_Accenture_MeetupDora&Cybersecurity.pptx02_Fabio Colombo_Accenture_MeetupDora&Cybersecurity.pptx
02_Fabio Colombo_Accenture_MeetupDora&Cybersecurity.pptxFinTech Belgium
Β 
Monthly Market Risk Update: April 2024 [SlideShare]
Monthly Market Risk Update: April 2024 [SlideShare]Monthly Market Risk Update: April 2024 [SlideShare]
Monthly Market Risk Update: April 2024 [SlideShare]Commonwealth
Β 
How Automation is Driving Efficiency Through the Last Mile of Reporting
How Automation is Driving Efficiency Through the Last Mile of ReportingHow Automation is Driving Efficiency Through the Last Mile of Reporting
How Automation is Driving Efficiency Through the Last Mile of ReportingAggregage
Β 
VIP Call Girls Service Dilsukhnagar Hyderabad Call +91-8250192130
VIP Call Girls Service Dilsukhnagar Hyderabad Call +91-8250192130VIP Call Girls Service Dilsukhnagar Hyderabad Call +91-8250192130
VIP Call Girls Service Dilsukhnagar Hyderabad Call +91-8250192130Suhani Kapoor
Β 
06_Joeri Van Speybroek_Dell_MeetupDora&Cybersecurity.pdf
06_Joeri Van Speybroek_Dell_MeetupDora&Cybersecurity.pdf06_Joeri Van Speybroek_Dell_MeetupDora&Cybersecurity.pdf
06_Joeri Van Speybroek_Dell_MeetupDora&Cybersecurity.pdfFinTech Belgium
Β 
Instant Issue Debit Cards - School Designs
Instant Issue Debit Cards - School DesignsInstant Issue Debit Cards - School Designs
Instant Issue Debit Cards - School Designsegoetzinger
Β 
Dividend Policy and Dividend Decision Theories.pptx
Dividend Policy and Dividend Decision Theories.pptxDividend Policy and Dividend Decision Theories.pptx
Dividend Policy and Dividend Decision Theories.pptxanshikagoel52
Β 
Solution Manual for Financial Accounting, 11th Edition by Robert Libby, Patri...
Solution Manual for Financial Accounting, 11th Edition by Robert Libby, Patri...Solution Manual for Financial Accounting, 11th Edition by Robert Libby, Patri...
Solution Manual for Financial Accounting, 11th Edition by Robert Libby, Patri...ssifa0344
Β 
OAT_RI_Ep19 WeighingTheRisks_Apr24_TheYellowMetal.pptx
OAT_RI_Ep19 WeighingTheRisks_Apr24_TheYellowMetal.pptxOAT_RI_Ep19 WeighingTheRisks_Apr24_TheYellowMetal.pptx
OAT_RI_Ep19 WeighingTheRisks_Apr24_TheYellowMetal.pptxhiddenlevers
Β 
The Economic History of the U.S. Lecture 17.pdf
The Economic History of the U.S. Lecture 17.pdfThe Economic History of the U.S. Lecture 17.pdf
The Economic History of the U.S. Lecture 17.pdfGale Pooley
Β 
Log your LOA pain with Pension Lab's brilliant campaign
Log your LOA pain with Pension Lab's brilliant campaignLog your LOA pain with Pension Lab's brilliant campaign
Log your LOA pain with Pension Lab's brilliant campaignHenry Tapper
Β 
Call US πŸ“ž 9892124323 βœ… Kurla Call Girls In Kurla ( Mumbai ) secure service
Call US πŸ“ž 9892124323 βœ… Kurla Call Girls In Kurla ( Mumbai ) secure serviceCall US πŸ“ž 9892124323 βœ… Kurla Call Girls In Kurla ( Mumbai ) secure service
Call US πŸ“ž 9892124323 βœ… Kurla Call Girls In Kurla ( Mumbai ) secure servicePooja Nehwal
Β 
Lundin Gold April 2024 Corporate Presentation v4.pdf
Lundin Gold April 2024 Corporate Presentation v4.pdfLundin Gold April 2024 Corporate Presentation v4.pdf
Lundin Gold April 2024 Corporate Presentation v4.pdfAdnet Communications
Β 
Independent Lucknow Call Girls 8923113531WhatsApp Lucknow Call Girls make you...
Independent Lucknow Call Girls 8923113531WhatsApp Lucknow Call Girls make you...Independent Lucknow Call Girls 8923113531WhatsApp Lucknow Call Girls make you...
Independent Lucknow Call Girls 8923113531WhatsApp Lucknow Call Girls make you...makika9823
Β 
Interimreport1 January–31 March2024 Elo Mutual Pension Insurance Company
Interimreport1 January–31 March2024 Elo Mutual Pension Insurance CompanyInterimreport1 January–31 March2024 Elo Mutual Pension Insurance Company
Interimreport1 January–31 March2024 Elo Mutual Pension Insurance CompanyTyΓΆelΓ€keyhtiΓΆ Elo
Β 
Best VIP Call Girls Noida Sector 18 Call Me: 8448380779
Best VIP Call Girls Noida Sector 18 Call Me: 8448380779Best VIP Call Girls Noida Sector 18 Call Me: 8448380779
Best VIP Call Girls Noida Sector 18 Call Me: 8448380779Delhi Call girls
Β 
High Class Call Girls Nagpur Grishma Call 7001035870 Meet With Nagpur Escorts
High Class Call Girls Nagpur Grishma Call 7001035870 Meet With Nagpur EscortsHigh Class Call Girls Nagpur Grishma Call 7001035870 Meet With Nagpur Escorts
High Class Call Girls Nagpur Grishma Call 7001035870 Meet With Nagpur Escortsranjana rawat
Β 
letter-from-the-chair-to-the-fca-relating-to-british-steel-pensions-scheme-15...
letter-from-the-chair-to-the-fca-relating-to-british-steel-pensions-scheme-15...letter-from-the-chair-to-the-fca-relating-to-british-steel-pensions-scheme-15...
letter-from-the-chair-to-the-fca-relating-to-british-steel-pensions-scheme-15...Henry Tapper
Β 
(DIYA) Bhumkar Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(DIYA) Bhumkar Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(DIYA) Bhumkar Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(DIYA) Bhumkar Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...ranjana rawat
Β 

Recently uploaded (20)

Stock Market Brief Deck for 4/24/24 .pdf
Stock Market Brief Deck for 4/24/24 .pdfStock Market Brief Deck for 4/24/24 .pdf
Stock Market Brief Deck for 4/24/24 .pdf
Β 
02_Fabio Colombo_Accenture_MeetupDora&Cybersecurity.pptx
02_Fabio Colombo_Accenture_MeetupDora&Cybersecurity.pptx02_Fabio Colombo_Accenture_MeetupDora&Cybersecurity.pptx
02_Fabio Colombo_Accenture_MeetupDora&Cybersecurity.pptx
Β 
Monthly Market Risk Update: April 2024 [SlideShare]
Monthly Market Risk Update: April 2024 [SlideShare]Monthly Market Risk Update: April 2024 [SlideShare]
Monthly Market Risk Update: April 2024 [SlideShare]
Β 
How Automation is Driving Efficiency Through the Last Mile of Reporting
How Automation is Driving Efficiency Through the Last Mile of ReportingHow Automation is Driving Efficiency Through the Last Mile of Reporting
How Automation is Driving Efficiency Through the Last Mile of Reporting
Β 
VIP Call Girls Service Dilsukhnagar Hyderabad Call +91-8250192130
VIP Call Girls Service Dilsukhnagar Hyderabad Call +91-8250192130VIP Call Girls Service Dilsukhnagar Hyderabad Call +91-8250192130
VIP Call Girls Service Dilsukhnagar Hyderabad Call +91-8250192130
Β 
06_Joeri Van Speybroek_Dell_MeetupDora&Cybersecurity.pdf
06_Joeri Van Speybroek_Dell_MeetupDora&Cybersecurity.pdf06_Joeri Van Speybroek_Dell_MeetupDora&Cybersecurity.pdf
06_Joeri Van Speybroek_Dell_MeetupDora&Cybersecurity.pdf
Β 
Instant Issue Debit Cards - School Designs
Instant Issue Debit Cards - School DesignsInstant Issue Debit Cards - School Designs
Instant Issue Debit Cards - School Designs
Β 
Dividend Policy and Dividend Decision Theories.pptx
Dividend Policy and Dividend Decision Theories.pptxDividend Policy and Dividend Decision Theories.pptx
Dividend Policy and Dividend Decision Theories.pptx
Β 
Solution Manual for Financial Accounting, 11th Edition by Robert Libby, Patri...
Solution Manual for Financial Accounting, 11th Edition by Robert Libby, Patri...Solution Manual for Financial Accounting, 11th Edition by Robert Libby, Patri...
Solution Manual for Financial Accounting, 11th Edition by Robert Libby, Patri...
Β 
OAT_RI_Ep19 WeighingTheRisks_Apr24_TheYellowMetal.pptx
OAT_RI_Ep19 WeighingTheRisks_Apr24_TheYellowMetal.pptxOAT_RI_Ep19 WeighingTheRisks_Apr24_TheYellowMetal.pptx
OAT_RI_Ep19 WeighingTheRisks_Apr24_TheYellowMetal.pptx
Β 
The Economic History of the U.S. Lecture 17.pdf
The Economic History of the U.S. Lecture 17.pdfThe Economic History of the U.S. Lecture 17.pdf
The Economic History of the U.S. Lecture 17.pdf
Β 
Log your LOA pain with Pension Lab's brilliant campaign
Log your LOA pain with Pension Lab's brilliant campaignLog your LOA pain with Pension Lab's brilliant campaign
Log your LOA pain with Pension Lab's brilliant campaign
Β 
Call US πŸ“ž 9892124323 βœ… Kurla Call Girls In Kurla ( Mumbai ) secure service
Call US πŸ“ž 9892124323 βœ… Kurla Call Girls In Kurla ( Mumbai ) secure serviceCall US πŸ“ž 9892124323 βœ… Kurla Call Girls In Kurla ( Mumbai ) secure service
Call US πŸ“ž 9892124323 βœ… Kurla Call Girls In Kurla ( Mumbai ) secure service
Β 
Lundin Gold April 2024 Corporate Presentation v4.pdf
Lundin Gold April 2024 Corporate Presentation v4.pdfLundin Gold April 2024 Corporate Presentation v4.pdf
Lundin Gold April 2024 Corporate Presentation v4.pdf
Β 
Independent Lucknow Call Girls 8923113531WhatsApp Lucknow Call Girls make you...
Independent Lucknow Call Girls 8923113531WhatsApp Lucknow Call Girls make you...Independent Lucknow Call Girls 8923113531WhatsApp Lucknow Call Girls make you...
Independent Lucknow Call Girls 8923113531WhatsApp Lucknow Call Girls make you...
Β 
Interimreport1 January–31 March2024 Elo Mutual Pension Insurance Company
Interimreport1 January–31 March2024 Elo Mutual Pension Insurance CompanyInterimreport1 January–31 March2024 Elo Mutual Pension Insurance Company
Interimreport1 January–31 March2024 Elo Mutual Pension Insurance Company
Β 
Best VIP Call Girls Noida Sector 18 Call Me: 8448380779
Best VIP Call Girls Noida Sector 18 Call Me: 8448380779Best VIP Call Girls Noida Sector 18 Call Me: 8448380779
Best VIP Call Girls Noida Sector 18 Call Me: 8448380779
Β 
High Class Call Girls Nagpur Grishma Call 7001035870 Meet With Nagpur Escorts
High Class Call Girls Nagpur Grishma Call 7001035870 Meet With Nagpur EscortsHigh Class Call Girls Nagpur Grishma Call 7001035870 Meet With Nagpur Escorts
High Class Call Girls Nagpur Grishma Call 7001035870 Meet With Nagpur Escorts
Β 
letter-from-the-chair-to-the-fca-relating-to-british-steel-pensions-scheme-15...
letter-from-the-chair-to-the-fca-relating-to-british-steel-pensions-scheme-15...letter-from-the-chair-to-the-fca-relating-to-british-steel-pensions-scheme-15...
letter-from-the-chair-to-the-fca-relating-to-british-steel-pensions-scheme-15...
Β 
(DIYA) Bhumkar Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(DIYA) Bhumkar Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(DIYA) Bhumkar Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(DIYA) Bhumkar Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
Β 

The Use of Bitcoin for Portfolio Optimization

  • 1. The use of Bitcoin for portfolio optimization Federico Tenga federico@chainside.net
  • 2. Introduction For many people Bitcoin is considered the ideal store of value, but most investor still lack to see the value of this asset due to its technical complexity. The scope of this study is to analyze Bitcoin strictly under a financial point of view and show the benefits it brings to optimize any investment portfolio and have more awareness of the risk-reward profile of this new asset.
  • 3. Index β€’ Bitcoin Supply β€’ Volatility β€’ Correlation β€’ Expected Return β€’ Daily returns distribution analysis β€’ Bitcoin for Portfolio Optimization
  • 4. Bitcoin Supply Figure 1 Bitcoin supply and inflation
  • 5. Gold Supply Figure 2 Gold supply. Source: Number Sleuth ("All The World's Gold Facts")
  • 6. Volatility The volatility, measured by the standard deviation, in finance is the degree of variation of price of an asset, and it can be derived using historical market price data. To find the volatility of bitcoin, we will compute the standard deviation of daily returns using the following formula: 𝜎 = 1 𝑛 ෍ 𝑖=1 𝑛 (π‘₯𝑖 βˆ’ πœ‡)2 Where: πœ‡ = π‘šπ‘’π‘Žπ‘› π‘£π‘Žπ‘™π‘’π‘’ π‘œπ‘“ π‘‘β„Žπ‘’ π‘‘π‘Žπ‘–π‘™π‘¦ π‘Ÿπ‘’π‘‘π‘’π‘Ÿπ‘›π‘  From April 2016 to May 2017 the average daily volatility of bitcoin is 2.67%
  • 7. Annualized Volatility To compute the annualized standard deviation, we have to multiply the daily standard deviation for the square root of the number of trading days. For traditional financial markets the number of trading days in a year is usually about 250, but since bitcoin is traded 24/7, every day of the year, we have about 365 trading days, consequentially the formula to compute the annualised standard deviation will be: π‘Žπ‘›π‘›π‘’π‘Žπ‘™π‘–π‘§π‘’π‘‘ 𝜎 = 𝜎 365 Resulting in an annualized standard deviation of 51.03%.
  • 9. Correlation The correlation measures the dependence between two variables. The Pearson correlation, considered to be the β€œtraditional” correlation, is calculated with the following formula: 𝜌 π‘₯,𝑦= πΆπ‘œπ‘£(π‘₯, 𝑦) 𝜎 π‘₯ 𝜎 𝑦 Where: πΆπ‘œπ‘£ π‘₯, 𝑦 = 𝐸( π‘₯ βˆ’ 𝐸 π‘₯ 𝑦 βˆ’ 𝐸 𝑦 ) 𝜎 π‘₯ = 𝐸[π‘₯]2βˆ’ (𝐸 π‘₯ )2 If we try to compute the Pearson correlation of bitcoin returns with some major asset classes we can find some very interesting results
  • 10. p-value When you perform a hypothesis test in statistics, a p-value helps you determine the significance of your results. he p-value is a number between 0 and 1 and interpreted in the following way: A small p-value (typically ≀ 0.05) indicates strong evidence against the null hypothesis, so you reject the null hypothesis.
  • 11. Bitcoin vs S&P 500 Correlation The Pearson correlation of bitcoin with S&P 500, over the timespan analysed, results to be 1.57%, which is extremely low. Moreover, the p-value of the Pearson correlation is a very high 0.582, meaning that we cannot even easily assume that the correlation is different from zero. Figure 5 Bitcoin vs S&P 500 daily returns
  • 12. Bitcoin vs MSCI Emerging Markets Index The Pearson correlation of bitcoin with MSCI Emerging Markets Index, over the timespan analysed, results to be 2.69%. Still a very low value, and similarly to what we have seen with S&P 500 the high p-value of 0.345 suggests that the result is not significant so we can even be sure that the correlation is not actually zero. Figure 6Bitcoin and MSCI Emerging Markets Index Daily Returns
  • 13. Bitcoin vs Oil Correlation The Pearson correlation of bitcoin with WTI Crude Oil prices, over the timespan analysed, results to be 0.8%, and just like in the cases seen above the data cannot be considered significant due to the high p-value of 0.789.
  • 14. Bitcoin vs Gold Correlation The Pearson correlation of bitcoin with gold, over the timespan analysed, results to be just 1.7%, and just like we have seen before the p-value of 0.558 suggests that the results cannot be considerate significant to assume that the correlation is different from zero.
  • 15. Correlation and portfolio volatility The non-existent correlation means that bitcoin does not share systematic risk with other asset classes, making it a great tool for diversification of a portfolio. Indeed, this is evident looking at the formula of a portfolio volatility: π‘ƒπ‘œπ‘Ÿπ‘‘π‘“π‘œπ‘™π‘–π‘œ π‘‰π‘œπ‘™π‘Žπ‘‘π‘–π‘™π‘–π‘‘π‘¦ = 𝑆𝐷 π‘Ž 2 βˆ— π‘Šπ‘Ž 2 + 𝑆𝐷 𝑏 2 βˆ— π‘Šπ‘ 2 + πΆπ‘œπ‘Ÿπ‘Ÿπ‘’π‘™π‘Žπ‘‘π‘–π‘œπ‘› 𝐸𝑓𝑓𝑒𝑐𝑑 Where: SD = standard deviation of the asset W = weight of the asset in the portfolio
  • 16. Expected Returns (1/2) A common method to attempt to calculate the assets appropriate return is the Capital Asset Pricing Model (CAPM), which describes the relationship between systematic risk and expected return of an asset. The CAPM formula for calculating the expected return of an asset given its risk is as follows: ra = rf + Ξ²a (rm – rf) Where: rf = risk-free rate Ξ²a = beta of the asset rm = expected return of the market Unfortunately, due to the lack of correlation with a benchmark portfolio, it is not possible to calculate the Ξ² of bitcoin
  • 17. We can use instead historical data to estimate future trends, but it is not easy to decide which timespan can be useful for our analysis. Old trading data are less significant due to the high market manipulation of the early days, a good option to use post-MtGox data. We can find the mean daily return of the period April 2014 – May 2017 using the following formula: πœ‡ = 1 𝑛 ෍ 𝑖=1 𝑛 𝑋𝑖 Which gives us a mean daily return of 0.13% Expected Returns (2/2)
  • 18. Annualized expected returns We can than derive the annualized expected return raising πœ‡ to the power of number of trading days in a year, using the following formula: πœ‡ π‘Žπ‘›π‘›π‘’π‘Žπ‘™π‘–π‘§π‘’π‘‘ = (1 + πœ‡ )365βˆ’1 Once again since bitcoin is traded 24/7, differently from any other asset we will 365 trading days. Which gives us an annualized expected return of 60.35%
  • 19. Daily Returns Distribution During the period from April 2014 to June 2017, we can see an average daily return of 0.2% with a standard deviation on 3.06% However, it is necessary to proceed with a normality test to discover if the standard returns can be described by a Gaussian distribution.
  • 20. Shapiro–Wilk test In statistics, the Shapiro–Wilk test tests the null hypothesis that a sample x1,...,xn originated from a normally distributed population. The test is: π‘Š = σ𝑖=1 𝑛 aixi 2 σ𝑖=1 𝑛 (xi βˆ’ Η‰π‘₯) 2 π‘Ž1, … , π‘Ž 𝑛 = π‘šT π‘‰βˆ’1 (π‘š 𝑑 π‘‰βˆ’1 π‘‰βˆ’1 π‘š)1/2 Where: π‘š = (π‘š1, … , π‘š 𝑛)T and m1, ..., mn are the expected values of the order statistics of independent and identically distributed random variables sampled from the standard normal distribution, and V is the covariance matrix of those order statistics.
  • 21. Shapiro–Wilk test Doing the Shapiro-Wink test on the bitcoin daily returns for the period April 2014 – April 2017 we find the following results: As the computed p-value is lower than the significance level alpha=0.05, one should reject the null hypothesis (the sample follows a Normal distribution), and accept the alternative hypothesis (the sample does not follow a Normal distribution). The risk to reject the null hypothesis while it is true is lower than 0.01% W 0.868 p-value (Two-tailed) < 0.0001 alpha 0.05
  • 22. Skewness analysis The Skewness is a measure to analyse the symmetry of the distribution. If the coefficient of skewness is positive the distribution is skewed right, if it is negative the distribution is skewed left. The coefficient of skewness of a data set is: skewness: 𝑔1 = π‘š3/π‘š2 3/2 Where: π‘š3 = Οƒ x βˆ’ ΰ΄€x 3 𝑛 π‘Žπ‘›π‘‘ π‘š2 = Οƒ x βˆ’ ΰ΄€x 2 𝑛 A normal distribution has a skewness coefficient of zero (perfect symmetry). The distribution of bitcoin daily returns from April 2014 to April 2017 has skewness coefficient of 0.024, which means that the distribution is almost perfectly symmetrical, indicating a good market efficiency.
  • 23. Kurtosis Analysis Another common measure of shape is the Kurtosis which provides information on how the data are spread among the peak and the tails of the distribution. Higher kurtosis also means that more of the variance is the result of infrequent extreme deviations, as opposed to frequent modestly sized deviations. The reference standard is a normal distribution, which has a kurtosis of 3. The coefficient of kurtosis of a dataset is: kurtosis: π‘Ž4 = π‘š4/π‘š2 2 Where: π‘š4 = Οƒ x βˆ’ ΰ΄€x 4 𝑛 π‘Žπ‘›π‘‘ π‘š2 = Οƒ x βˆ’ ΰ΄€x 2 𝑛 The distribution of bitcoin daily returns from April 2014 to April 2017 has kurtosis coefficient of 12.439, meaning that most of the variance is caused by infrequent extreme deviations. This can also be seen as volatility on volatility.
  • 24. Autocorrelation Analysis In order to understand if the bitcoin market is efficient, it is useful to test the autocorrelation of the daily returns. If the returns result to be autocorrelated, it indicates that the price movements can be predicted and the market is not efficient. To test the autocorrelation, we will use the Durbin-Watson statistic, which always has a value d between 0 and 4, where 2 means there is no autocorrelation, 0 means that there is positive correlation and 4 means that there is negative correlation. The value d of the Durbin-Watson statistic is calculated with the following formula: d = Οƒ 𝑑=2 𝑇 (𝑒𝑑 βˆ’ π‘’π‘‘βˆ’1)2 Οƒ 𝑑=1 𝑇 𝑒𝑑 2 Where: et = residual associated with the observation at time t Running the Durbin-Watson test on the bitcoin daily returns from April 2014 to April 2017 we obtain a d value of 2.094, meaning that there is almost no autocorrelation and the market is efficient.
  • 25. Bitcoin Price Resilience In the history of Bitcoin there have been many big events that caused high volatility periods and the collapse of the price. It is interesting to notice that even if big events that can potentially damage the Bitcoin ecosystem keep happening, the impact is becoming less significant. Figure 13 Bitcoin price change comparison after exchanges' losses
  • 27. Portfolio Optimization A general solution to portfolio selection problem was proposed by Markowitz in 1952, building the foundation of the Modern Portfolio Theory (MPT) in the coming decades. The MPT is based on the utility maximization concept, emphasizing the trade-off between risk and return. One of the assumption of the MPT is that investors are risk-adverse, giving two portfolios with similar returns a rational investor will always choose the one with lower risk, and any extra risk will have to be compensated. With the Modern Portfolio Theory, it is possible to keep a certain level of risk while maximising the returns, or keeping the returns fixed while minimising the risk. An investor can then generate a so called efficient frontier where the optimal asset allocation is achieved.
  • 28. Portfolio Returns An asset return can be easily defined as the price difference over a time period divided the price of the asset at the beginning of the period, as described by the following formula: 𝑅𝑖,𝑑 = 𝑃𝑖,𝑑 βˆ’ 𝑃𝑖,π‘‘βˆ’1 𝑃𝑖,π‘‘βˆ’1 Where: 𝑃𝑖,𝑑 = π‘π‘Ÿπ‘–π‘π‘’ π‘Žπ‘‘ π‘‘β„Žπ‘’ 𝑒𝑛𝑑 π‘œπ‘“ π‘‘β„Žπ‘’ π‘π‘’π‘Ÿπ‘–π‘œπ‘‘ 𝑃𝑖,π‘‘βˆ’1 = π‘π‘Ÿπ‘–π‘π‘’ π‘Žπ‘‘ π‘‘β„Žπ‘’ 𝑏𝑒𝑔𝑖𝑛𝑛𝑖𝑛𝑔 π‘œπ‘“ π‘‘β„Žπ‘’ π‘π‘’π‘Ÿπ‘–π‘œπ‘‘ In a portfolio, the return is simply the weighted average of the return of the assets in the portfolio, as described by the following formula: 𝑅 𝑝,𝑑 = ෍ 𝑛=1 𝑁 𝑀𝑖 𝑅𝑖,𝑑 Where: 𝑀𝑖 = π‘π‘Ÿπ‘œπ‘π‘œπ‘Ÿπ‘‘π‘–π‘œπ‘› π‘œπ‘“ π‘‘β„Žπ‘’ π‘π‘œπ‘Ÿπ‘‘π‘“π‘œπ‘™π‘–π‘œ 𝑖𝑛𝑣𝑒𝑠𝑑𝑒𝑑 𝑖𝑛 π‘‘β„Žπ‘’ π‘Žπ‘ π‘ π‘’π‘‘ 𝑖
  • 29. Portfolio Risk We can measure the risk of the portfolio using its standard deviation. For an asset, the standard deviation can be found, as we have previously seen, using the following formula: 𝜎 = 1 𝑛 ෍ 𝑖=1 𝑛 (π‘₯𝑖 βˆ’ πœ‡)2
  • 30. Portfolio Risk In a portfolio, we have to consider the standard deviation and the correlation coefficient of all the assets that compose the portfolio, we can then compute the standard deviation of the portfolio using the following general formula: 𝜎 𝑝 = ෍ 𝑖=1 𝑁 ෍ 𝑗=1 𝑁 𝑀𝑖 𝑀𝑗 πœŽπ‘– πœŽπ‘—Οπ‘–π‘— Where: w = weight of the asset Οƒ = standard deviation of the asset ρ = correlation coefficients of the assets Considering a portfolio with two assets, the resulting standard deviation will be: 𝜎 𝑝 = 𝑀1 2 𝜎1 2 + 𝑀2 2 𝜎2 2 + 2𝑀1 𝜎2ρ1,2
  • 31. The Sharpe Ratio The Sharpe ratio is a tool to examine the performance of an investment, by measuring how well the return of an asset compensates for its risk, making also possible to compare assets with very different levels of risk. The Sharpe ratio is calculated as follows: 𝑆 π‘Ž = 𝐸[𝑅 π‘Ž βˆ’ 𝑅 𝑏] 𝜎 π‘Ž Where: Ra = return of asset Rb = return of benchmark asset, usually the risk-free rate Οƒa = standard deviation of the asset The Sharpe ratio can be used to identify the optimal asset allocation in a portfolio, the one that provides the highest return per unit of deviation.
  • 32. Bitcoin – S&P 500 portfolio To efficiently build a two assets portfolio using S&P 500 and bitcoin, it is first needed to quickly analyse the two separate assets: We can already notice that bitcoin has a higher Sharpe ratio than S&P 500, considering that we have also previously seen that the two assets have no correlation we can already deduce that bitcoin will significantly improve the returns of the portfolio per unit of deviation. S&P 500 bitcoin Daily mean SD 0.80% 2.67% Annualized SD 12.66% 51.03% Daily mean return 0.04% 0.13% Annualized return 9.46% 60.35% Sharpe ratio 0.68 1.15
  • 33. Bitcoin – S&P 500 portfolio Btc% S&P 500% SD Return Sharpe 0% 100% 12.66% 9.46% 0.6838 30% 70% 17.69% 24.73% 1.3532 100% 0% 51.03% 60.35% 1.1673 In the table below we can see the return and the risk for each possible allocation of the two assets: As we can see, the portfolio with the highest Sharpe ratio would be the one with 30% allocated in bitcoin an 70% allocated in S&P 500.
  • 34.
  • 35. Gold – S&P 500 portfolio The fact that bitcoin could improve the performance of the S&P 500 portfolio was easily predictable due to its low correlation coefficient, so it is more interesting to see how the performance improvment provided by bitcoin is compared to the improvement provided by gold, an asset that bitcoin has the ambition to replace, at least partially, in its safe heaven role for investors. Gold seems to be very good for diversification purposes due to its low correlation with S&P 500 (only 0.03), but as we can see in the table below, its poor performances in the past years greatly affect its Sharpe ratio, which is significantly lower compared to bitcoin’s ratio. S&P 500 Gold Daily mean SD 0.80% 0.79% Annualized SD 12.66% 12.57% Daily mean return 0.04% 0.013% Annualized return 9.46% 3.37% Sharpe ratio 0.68 0.20
  • 36. Gold – S&P 500 portfolio After trying to plot every possible portfolio combing the two assets we can find that the allocation that maximise the Sharpe ratio has 21% of gold. However, the contribution to improve the performance of the portfolio is modest, the Sharpe ratio of the optimal portfolio is just slightly higher compared to the S&P ratio of 0.68, while bitcoin provided a much more significant improvement, almost doubling it. Gold% S&P 500% SD Return Sharpe 0% 100% 12.66% 9.46% 0.6838 21% 79% 10.42% 8.18% 0.7081 100% 0% 12.57% 3.37% 0.2044
  • 37.
  • 38. Too much risk? Choosing an asset allocation that maximise the Sharpe ratio does not allow an investor to set its desired level of risk. This problem can be solved introducing a risk-free asset, which let an investor maximise the Sharpe ratio and set the portfolio risk level. Introducing the risk-free asset, the expected return and the standard deviation of a portfolio are: 𝐸(𝑅 𝑝) = 𝑀1 𝐸(𝑅1) + (1 βˆ’ 𝑀1) 𝑅 𝑓 Οƒ 𝑝 = 𝑀1 2 Οƒ1 2 Where: 𝑅1= return of the portfolio with only risky assets 𝑅 𝑓= return of the risk-free asset 𝑀1= weight of the risky assets 𝜎1=standard deviation of the portfolio with only risky assets
  • 39. Capital Allocation Line After having found the portfolio that maximise the Sharpe ratio, an investor with a lower risk preference can allocate part of his capital in a risk-free asset, reducing his exposure, and an investor with a higher risk preference can borrow at the risk-free rate and invest his capital with leverage in the optimal portfolio. These practices generate the Capital Allocation Line (CAL), defined by the following formula: 𝐸 𝑅 𝑝 = 𝑅𝑓 + 𝐸(𝑅1) βˆ’ 𝑅𝑓 Οƒ1 Οƒ 𝑝
  • 40.
  • 41. Conclusion Whatever is your risk preference, Bitcoin can help you to optimize your portfolio
  • 42. Federico Tenga COO at Chainside federico@chainside.net