Analysis of ETF correlations for the first half of 2020

44 ETFs and BPIFs are currently traded on the Moscow Exchange. These are exchange-traded funds that hold ready-made specialized portfolios and packaged in a stand-alone stock (like a microservice in a docker container). This share can be bought by receiving a share in the total portfolio.



To identify relationships between funds, mathematical methods of time series analysis were used: correlation based on price changes with subsequent clustering by distance. As you know, mathematics can be illogical and find something that does not exist in reality. Math is blind to the news, the coronavirus and Tesla's red shorts.



Correlation matrix

There is an interactive version of the matrix (be careful, traffic)



Let's start right away with the results of the analysis of exchange-traded funds. In the matrix above, there are correlations of all ETFs and BPIFs that appeared before January 2020. Until the 20th year, there were too few BPIFs, there was nothing to analyze there.



Correlation in the diversification of the investment portfolio



The modern portfolio theory invented by Markowitz uses only two factors: the expected return on an asset and its correlation with other assets. By using a set of assets between which the correlation tends to -1 (red), the investment portfolio becomes immune to external factors.



This insensitivity is the main point of diversification. Correlation is just a formal criterion for the selection of assets in a diversified portfolio.



The correlation matrix is ​​beautiful, but even if sorted by color blocks with a large number of funds, it is very difficult to assimilate. Therefore, let's call machine learning for help, so that the computer itself finds similar funds for us in the correlation matrix. Yes, ML, which you hear about from every iron, has finally solved a simple, but very necessary applied problem for us here and now.



Dendrogram of connections



, . β€” , , . :



  • FXUS SBSP ( ) , - .
  • FXRB FXTB ( ) , .


3 , 7 . 3 , . 7. β€” . .



- "" . , " ".



. ? ? ? .



ETF



http://beta-stock.site/correlations/corr_Cons.png



FXRU
RUSB
SBCB
FXMM
FXTB




. ( ), .



: " , ".





TUSD
TEUR
FXGD




. , FXGD , , . .





http://beta-stock.site/correlations/corr_USA.png



ETF FXCN, . .



, VTBE. , VTBE , FXCN. .



, FXCN . .



-



AKNX
AKSP
VTBH
AKEU
FXKZ




, S&P IT, . . , ( -) .



. , FXUS AKSP, AKSP . , - .



!http://beta-stock.site/correlations/corFXUSvsAKSP2.png](http://beta-stock.site/correlations/corFXUSvsAKSP2.png)

AKSP () , FXUS ().



- 2 : VTBH FXKZ.



, FXKZ AKSP. VTBH. , .



-



FXWO
VTBA
VTBE
FXIT
FXUS
SBSP




, " ".



. S&P β€” . FXIT β€” 91 S&P. IT S&P, .



VTBE, , . , VTBE S&P , QE. VTBE S&P? VTBE 2600 . , S&P. VTBE FXCN...





http://beta-stock.site/correlations/corr_RU.png





SBRB
SBGB
TRUR




, . , . , .



TRUR : , . TRUR SBGB.





FXRL
SBMX
RUSE
VTBB
FXRW
FXDE




. (RUSE, FXRL, SBMX) (FXDE, FXRW). FXDE FXRW , .



IMOEX VTBB. , VTBB SBCB. , . , VTBB IMOEX .





FXRB
GPBM
GPBS
MTEK




. , .



FXRB , FXTB.



GPBM GPBS , . , .



MTEK . IT MTEK 10 10% . IT .





.



  • , . , VTBH , VTBB .
  • ( , ) . , .
  • :

    • IT (FXIT, AKNX, MTEK). , . 2020 IT (SCIP, TECH). , .
    • (FXCN , FXKZ , VTBE ). , .
  • … , ?!



All Articles