-55 Distributions 33. Distribution of trades--specify either (new card, ALPHA) MARKET SHARE USING NORMAL DISTRIBUTION or INVENTORY POSITION OF THE FIRM. Specify one parameter if the former was chosen (new card, REAL)--the standard deviation as a proportion of market share; otherwise, specify zero (new card, REAL). Items 34 to 42 specify probability distributions. In each case the distribution name is specified (new card, ALPHA) from TRIANGULAR NORMAL LOGNORMAL GAMMA Two parameters must follow each distribution name (new card, REAL). The first is the mean (except for the triangular distribution where it is the mode). The second is the standard deviation (except for triangular where it is the maximum). (The mean and standard deviation are computed for the triangular distribution and listed in the input summary.) 37. Cash customer certificate arrival time distribution. 38. 39. 40. 41. 42. Breakdown of customer orders into lot size distribution. Breakdown of institutional orders into lot size distribution. 43. 44. 45. 46. 47. 48. 49. Stock name (new card, ALPHA)- --same as item 13. OTC code (new card, INTEGER)--1 for OTC and 0 if issue is Industry float in 100 share units (INTEGER). This input is not used in the current version of the model except for description. Probability of market recovery from zero volume (REAL fraction). If daily trading volume equals zero, the recursive trading function will not recover. Items 46 and 47 provide a simple method to recover from this case. Average market recovery volume from zero in 100 share units Number of values contained in the change in volume random -57 50. List the cumulative probabilities (REAL fraction) associated with each volume multiplier in the random step variable-same number as in item 48. 51. Proportion of market share per firm (in percent) (INTEGER) specify share of each firm. Sum must be less than or equal to 100. 52. List firm data (INTEGER)--list items a to h by firm. a. Business attributable to cash customers (in percent). b. Business attributable to margin customers (in percent). C. Business attributable to institutional customers (in d. Business attributable to principal customers (in percent). Location of firm--0 if New York based; 1 otherwise. Nonhouse name inventory--at beginning of simulation in House-name inventory--again an initial figure in 100 Segregation requirement in 100 share units--the initial A number of outputs are generated by the TCS program. Some of these are for debugging and/or tracing trading activities on a very detailed level, and their generation is thus an input specification. Other reports are printed on every run. This section traces a representative sample of a simulation run's output. Input Summary Report The Input Summary (Fig. 19) lists most inputs relevant to the simulation. Those inputs associated with suppression of certain reports are excluded. Also certain firm characteristics, such as initial inventories and segregation requirement, are omitted since they appear in the first firm status report after the end of the first simulated trading day. The distribution of percentage changes in volume was described on page 13. -59 SIMULATION PARAMETERS: NUMBER OF FIRMS = 10 NUMBER OF DAYS THE THE FIRST DAY DATA FUR SUMMARY STATISTICS IS GATHERED = 10 .1000 = YEARLY INTEREST RATE STOCK CHARACTERISTICS: IDENTIFICATION = XYZ PROBABILITY OF REMAINING IN A ZERO STATE = .8503 DISTRIBUTION OF PERCENTAGE CHANGES IN VOLUME 5 PROPORTION OF FREE INVENTORY AVAILABLE FOR STUCK LOANS = 85 100 AVERAGE INTERREGIONAL DELIVERY TIME 5/100 = 2 DAYS DURATION FIRMS HULD CERTIFICATES NOT IN THEIR NAME 2 DAYS PROPORTION CASH CUSTOMERS LEAVING CERTIFICATES WITH BROKER = 40/100 5/100 PROPORTION MARGIN BUYS COVERING SHORT POSITION = 5/133 5/100 DAILY NET SORTED AND EQUALS MATCHED AFTER FACH ORDER MADE Fig. 19--Input Sericu |