Conduct Stress Test on Portfolio

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

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

• Use trading costs to screen stocks in a portfolio and estimate the cost to liquidate or purchase a specified number of shares.

• Analyze trading costs during volatile periods of time such as a financial crisis, flash crash, or debt crisis.

To access the example code, enter edit KRGStressTestingExample.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);

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

load KRGExampleData TradeDataStressTest

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

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);

Prepare Data for Stress Testing

Specify the date range from May 1, 2015, through July 31, 2015.

startDate = '5/1/2015';
endDate = '7/31/2015';

Determine the number of stocks numStocks in the portfolio. Create a date range dateRange from the specified dates. Find the number of days numDates in the date range.

numStocks = length(TradeDataStressTest.Symbol);
dateRange = (datenum(startDate):datenum(endDate))';
numDates = length(dateRange);

Preallocate the output data table o.

outLength = numStocks*numDates;
symbols = TradeDataStressTest.Symbol(:,ones(1,numDates));
sides = TradeDataStressTest.Side(:,ones(1,numDates));
dates = dateRange(:,ones(1,numStocks))';

o = table(symbols(:),sides(:),dates(:),NaN(outLength,1),NaN(outLength,1), ...
    'VariableNames',{'Symbol','Side','Date','MI','MIDollar'});

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

TradeDataStressTest.Shares = abs(TradeDataStressTest.Shares);

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

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

Conduct Stress Test by Estimating Historical Market-Impact Costs

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

kk = 1;
for ii = dateRange(1):dateRange(end)
  
    for jj = 1:numStocks
    
        k.MiCode = TradeDataStressTest.MICode(jj);
        k.MiDate = ii;

        o.MI(kk) = marketImpact(k,TradeDataStressTest(jj,:));
        o.MIDollar(kk) = (TradeDataStressTest.Shares(jj) ...
            * TradeDataStressTest.Price(jj)) ...
            * o.MI(kk) /10000 * TradeDataStressTest.FXRate(jj);
 
        kk = kk + 1;
    
    end
    
end

Display the first three rows of output data.

o(1:3,:)

ans = 

    Symbol    Side      Date        MI      MIDollar
    ______    ____    _________    _____    ________

    'A'       1.00    736085.00     3.84      384.31
    'B'       1.00    736085.00    11.43    14292.24
    'C'       1.00    736085.00    32.69    20430.65

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

Retrieve the daily market-impact cost dailyCost. Determine the number of days numDays in the output data. Loop through the data and sum the market-impact costs for individual stocks for each day.

numDays = length(o.Date)/numStocks;

idx = 1;
for i = 1:numDays

  dailyCost.Date(i) = o.Date(idx);
  dailyCost.DailyMiCost(i) = sum(o.MI(idx:idx+(numStocks-1)));
  idx = idx+numStocks;

end

Display the daily market-impact cost in the specified date range. This figure demonstrates how market-impact costs change over time.

plot(b.Date,b.DailyMiCost)
ylabel({'Daily Cost','(bps)'})
title('Daily Market-Impact Cost Stress Test')
xlabel('Date')
grid on
xData = linspace(b.Date(1),b.Date(92),11);
a = gca;
a.XAxis.TickLabels = datestr(xData,'mm/dd/yyyy');
a.XTickLabelRotation = 45;

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.