JLM Simulator Documentation

JLMSim: Trace File Output Description


One of the output files produced by the run with input to be found now in

JLMSimInput for Case M23a NrS16NrIT8_1000_50_200NL -1.txt

is the Trace File

Trace file for Case M23a NrS16NrIT8_1000_50_200NL -1.txt


PART I: INITIAL CONDITIONS

The Trace File has two parts. The first part provides details about conditions at the start of the run. The second provides details concerning the course of the run. We describe the first part in this section and the second part in the next.

The part of the Trace File begins with parameters used in historical return generation, such as the following:

 

*********************************************************************

The first output items are the means and standard deviations of the factors used in the factor model for generating initial returns. The second is a table showing the alphas, betas and standard deviations of idiosyncratic risk used in the factor model. The average volume of the security, before simulation, is also shown for each security. (This an input; see Input File Documentation.)

The next output gives each statistician's initial estimates of the means, variances and covariances of each security, as illustrated below for Statistician 0:


Statistics after initialization

Statistician Nr. 0 

The next output in the Trace File shows the expected return vectors initially used by each portfolio analyst, including the expected returns to short positions inclusive of rebates. In the headings, a long position in Security 0 is denoted L 0, a short position S 0. "Slack" is a variable that appears in the optimization and represents the difference between allowed and used L + |S|.  (See Input File Documentation.)


Expected return vectors used by Portfolio Analysts

Portfolio SecNr: 

The next two tables present material concerning the initial mean-variance efficient frontiers computed by each of the portfolio analysts. A mean-variance efficient frontier is piece-wise linear (in portfolio space). Each of these pieces is called an efficient segment. The end points of efficient segments are called corner portfolios.

The first table (shown, in part, below) gives for each portfolio analyst in turn information about each of the efficient segments in the portfolio analyst's frontier. (Only the "lowest" five segments, i.e., the segments with lowest portfolio mean and variance, for the Portfolio Analyst 0 are reproduced below.) The third through sixth items of information shown are the highest and lowest means and standard deviations of portfolios on each segment. The first and second items of information are the lowest and highest "LambdaE"s on each segment, where LambdaE equals the slope of a curve with 1/2 times variance on the vertical axis and expected return on the horizontal axis. This varies linearly along any efficient segment. It is possible to have LambdaE (and some other Lagrangian multipliers) vary on an efficient segment, while the portfolio remains unchanged. This appears as a kink in the set of efficient mean-standard deviation combinations. The last column of the first table indicates whether there is kink in the latter curve at this point.


Initial efficient frontiers 

Portfolio Analyst 0 

The next table of the two describing initial efficient portfolios shows, for each portfolio analyst in turn, the holdings in each corner portfolio on the analyst's frontier, listed from most cautious to most aggressive. Holdings include long positions, short positions, cash, debt and slack, as illustrated below for Portfolio Analyst 0.


Initial corner portfolios

Holdings
Portfolio Analyst 0 

II. REPORTS PRODUCED DURING THE SIMULATION RUN

Once the simulation run begins, two events produce trace output. One is the end-of-day event; the other is the time for reoptimization for investors who have been selected for trace output. (See Input File Documentation for a description of how individual investors are selected to be traced.)

The following is an example of the report produced for an investor-to-be-traced at reoptimization time if the investor is OK (not currently defunct or subject to a margin call):


Reoptimization for Investor with Template 7 Seq Nr. 37 at 0.032108 

**************************************************************************************

The heading for the report notes the investor's template, sequence number, and the time of the report. Additional heading information includes the investor's starting wealth, deposits received, withdrawals paid, withdrawals requested but not yet paid, the difference between deposits received and withdrawals requested (whether paid or due), the current wealth of the investor, and account status. Account status may be:

0 for OK,
1 for currently subject to margin call, or
2 for defunct.

The body of the report provides information concerning the investor's holdings and intentions for each security. For each of the sixteen user-defined securities of this run, numbered 0 through 15, the report displays security number, number of shares (x_units) currently held, the current value of these shares, the fraction this represents of the investor's portfolio, the fraction that the optimization says is ideal, the fraction of the portfolio that the investor will attempt to buy (+) or sell (-) after considering turnover dampening rules, the current price of the security, and the number of shares that the investor will attempt to buy or sell. In this output, short positions are represented as negative x_units.

Securities 16 and 17 of this run are cash and debt, respectively. The latter figures reflect the fact that random deposits or withdrawals have been decided by this point in the code, but no transactions have yet taken place. Cash and debt always have a price of 1.0 and -1.0 respectively. No d_fraction or d_units (buys or sells) are noted for cash. There is an inconsistency in this report in that x_value and x_fraction for Debt is reported with a minus sign, whereas x_ideal is reported (as it appears in the optimization calculation) with a plus sign. Thus the last line of the table should be interpreted as follows: investor 7-37 currently has debt equal to 1.452 percent of the equity in the account but, according to the optimization calculation, would like to increase this to 50 percent of the equity in the account. The increase in debt sought this time is not listed in the Security 17 row under d_units. Rather it is implicit in the fact that the value of the purchase sought this time exceeds the sum of the values of the sales sought.

In the present case, none of the traced investors had a margin call in effect at the time a reoptimization would normally occur for the investor; nor did any become defunct. The tables below (taken from Case M23a NrS16NrIT8_1000_500_600 -1) give illustrations of the cases for a margin call at reoptimization and a defunct investor.

EXAMPLE 1:

Usual reoptimization time for Investor with Template 0 Seq Nr. 204 at 310.424862

*********************************************************************************************

Security 5's price exceeds $10,000 at the time of the above report. Thus the investor is short over $100,000,000 worth of Security 5. The account includes an equal amount of "cash" as collateral against the short position (not shown as part of the over $17 million held in cash, represented by Security 16). The investor still has a net worth of over $23 million; but this is not sufficient to avoid a margin call with total positions, long + |short|, well over $100million.

EXAMPLE 2:

Usual reoptimization time for Investor with Template 1 Seq Nr. 50 at 364.203056

The above investor is defunct. At the time the last of its assets were sold (or short positions covered) it still owed $3,488,440.This amount has been transferred from debt outstanding to bad debt in the macro statistics.



In addition to reports at reoptimization time for each investor marked for tracing, the Trace File includes a System Summary Report for the end of each simulated day. The following is an example from Case M23a NrS16NrIT8_1000_50_200NL -1:


End of day 407.000000 TotalBorrowing= 578212457 TotalLending= 712276978 TotalBadDebt= 0
TotalWithDrawals_due= 65422

This end-of-day report includes total borrowing, total lending, total bad debt, and total withdrawals due for the entire system. It also shows, for each security, the number of limit orders on the Buy Book, the highest and lowest bids among these, and similar information for the Sell Book. It also shows, for each security, the price at which it last traded and its current price. The latter may differ from the former when the last trade price is below the best bid or above the best offer. Further information for each security is the (day and) time it last traded, and whether it is currently on an up or down tick. The "s" following the best offer for security 11 indicates that this offer is a short sale. No trade takes place even though the best bid equals the best offer, since the trade would have to take place at $154.30, the last trade price, and the security is currently on a down tick.
 

Back to Top

© 2009 Jacobs Levy Equity Management. All rights reserved.
For problems or questions regarding this website contact JLM_Sim@jacobslevy.com.
Last updated: June 30, 2005