Conduct Back Test on Portfolio

This example shows how to conduct a back test on a set of stocks using transaction cost analysis from the Kissell Research Group.

• Analyze the implementation of an investment strategy on a specific day or date range.

• Estimate historical market-impact costs and the corresponding dollar values for the specified historical dates.

• Analyze the trading costs of different orders on various dates.

To access the example code, enter edit KRGBackTestingExample.m at the command line.

Retrieve Market-Impact Parameters and Load Historical Data

Retrieve the market-impact data from the Kissell Research Group FTP site. Connect to the FTP site using the ftp function with a user name and password. Navigate to the MI_Parameters folder and retrieve the market-impact data in the MI_Encrypted_Parameters.csv file. miData contains the encrypted market-impact date, code, and parameters.

f = ftp('ftp.kissellresearch.com','username','pwd');
mget(f,'MI_Encrypted_Parameters.csv');
close(f)

miData = readtable('MI_Encrypted_Parameters.csv','delimiter', ...
    ',','ReadRowNames',false,'ReadVariableNames',true);

Create a Kissell Research Group transaction cost analysis object k. Specify initial settings for the date, market-impact code, and number of trading days.

k = krg(miData,datetime('today'),1,250);

Load the example data TradeDataBackTest from the file KRGExampleData.mat, which is included with the Trading Toolbox™.

load KRGExampleData TradeDataBackTest

For a description of the example data, see Kissell Research Group Data Sets.

Prepare Data for Back Testing

Determine the number of stocks numRecords in the portfolio.

numRecords = length(TradeDataBackTest.Symbol);

Preallocate the output data table o.

o = table(TradeDataBackTest.Symbol,TradeDataBackTest.Side, ...
    TradeDataBackTest.Date,NaN(numRecords,1),NaN(numRecords,1), ...
    'VariableNames',{'Symbol','Side','Date','MI','MIDollar'});

Ensure that the number of shares is a positive value using the abs function.

TradeDataBackTest.Shares = abs(TradeDataBackTest.Shares);

Convert trade time trade strategy to the percentage of volume trade strategy.

TradeDataBackTest.TradeTime = TradeDataBackTest.TradeTime ...
    .* TradeDataBackTest.ADV;
TradeDataBackTest.POV = krg.tradetime2pov(TradeDataBackTest.TradeTime, ...
    TradeDataBackTest.Shares);

Conduct Back Test by Estimating Historical Market-Impact Costs

Estimate the historical market-impact costs for each stock in the portfolio on different dates using marketImpact. Convert market-impact cost from decimal into local dollars. Retrieve the resulting data in the output data table o.

for ii = 1:numRecords
    
    k.MiDate = TradeDataBackTest.Date(ii);
    k.MiCode = TradeDataBackTest.MICode(ii);

    o.MI(ii) = marketImpact(k,TradeDataBackTest(ii,:));
    MIDollars = (TradeDataBackTest.Shares(ii) * TradeDataBackTest.Price(ii)) ...
        * o.MI(ii)/10000 * TradeDataBackTest.FXRate(ii);

    o.MIDollar(ii) = MIDollars;
    
end

Display the first three rows of output data.

o(1:3,:)

ans = 

    Symbol    Side       Date        MI     MIDollar
    ______    ____    __________    ____    ________

    'A'       1.00    '5/1/2015'    1.04     103.91 
    'B'       1.00    '5/1/2015'    3.09    3864.44 
    'C'       1.00    '5/1/2015'    8.54    5335.03 

The output data contains these variables:

• Stock symbol

• Side

• Historical trade date

• Historical market-impact cost in basis points

• Historical market-impact value in local dollars

References

[1] Kissell, Robert. “Creating Dynamic Pre-Trade Models: Beyond the Black Box.” Journal of Trading. Vol. 6, Number 4, Fall 2011, pp. 8–15.

[2] Kissell, Robert. “TCA in the Investment Process: An Overview.” Journal of Index Investing. Vol. 2, Number 1, Summer 2011, pp. 60–64.

[3] Kissell, Robert. The Science of Algorithmic Trading and Portfolio Management. Cambridge, MA: Elsevier/Academic Press, 2013.

[4] Chung, Grace and Robert Kissell. “An Application of Transaction Costs in the Portfolio Optimization Process.” Journal of Trading. Vol. 11, Number 2, Spring 2016, pp. 11–20.