Eventus
Software


View User Guide

EventStream

Eventus Features

Eventus and
SAS


Who Uses
Eventus?

Order

Examples for Users

Upgrade
Information

FAQ

Support

Request a Quote

Fama-French Factors

Volume Indexes

Patches

View Documentation

Ordering
Information

New Features

New versions of Eventus software become available periodically. Current one-time-fee (perpetual) licensees of Eventus 8 are eligible for a discount on Eventus 9.0. Annual subscribers receive free upgrades during the paid subscription period. For a specific upgrade price quote for your organization, please use our quote request form. We also have a price list for first-time licensees.

New Features in Eventus 9.0
    Please see our 22 page summary of new and enhanced features

What was new in Eventus 8.0 (and still in Eventus 9.0)

  • Fama-French three-factor model can be used as a general stock-by-stock benchmark model, not just with calendar-time portfolio regression and Ibbotson RATS approaches, using the FAMAFRENCH TWOSTEP option combination on the Evtstudy statement.
  • The momentum factor suggested by Carhart (1997) can be added to the Fama-French three-factor model using the MOMENTUM option (in addition to the FAMAFRENCH option, and other desired options such as TWOSTEP or IRATS) on the Evtstudy statement.
  • Event studies can use an arbitrary number of user-supplied factors in a linear model framework similar to the traditional market model (not just in the event parameter approach as in previous versions) in a simple, straightforward way. CRSP-based event studies do not need to be ported to a non-CRSP mode to make use of user-supplied factors.
  • Event studies can incorporate user-supplied �companion portfolios,� or characteristic-dependent indexes in lieu of a market-wide index,  in a simple, straightforward way. CRSP-based event studies do not need to be ported to a non-CRSP mode to make use of companion portfolios.
  • Event studies using CRSP can incorporate a user-supplied, non-CRSP market index without the inconvenience of moving the whole study to SASNONCRSP mode. This new feature supports a single market-wide index; also please see our new companion portfolio facility above. The user-supplied index takes the place of the CRSP value-weighted index.
  • The Evtstudy statement can output �window� cumulative or compounded abnormal returns directly to a SAS data set or ASCII file, avoiding the need to run the separate Extract statement.
  • Time-series graphs of return, abnormal return, cumulative and compounded return, cumulative and compounded abnormal return are now available using the PLOT option on the Evtstudy statement.
  • When the OUTWIN= option is specified on the Evtstudy statement, and the Patell (default), Standardized Cross-Sectional, EGLS or Collins-Dent test is used in the event study, the output data set will include WLS weights and standardized CARs by default. One can still use WPREFIX= to specify a variable name prefix for the weights, but it is no longer necessary; the default prefix is CARWeight. The prefix for the standardized CARs is SCAR. To suppress weights and standardized CARs when using OUTWIN= with eligible tests, specify WEIGHTS=0 on the Evtstudy statement.
  • New features are available for the Extract statement, which is no longer needed but still available for compatibility with earlier versions. (Extract�s functionality is now available through the new OUTWIN= and FILEWIN= options of the Evtstudy statement.) The old syntax still works, or specify Evtstudy statement-like options to select benchmarks, e.g., Raw, CP, MAR, MM, FF or FFM. More than one at a time may be specified. The BOTH, VALUE or EQUAL options may be used, when applicable, to select between results using value-weighted and equal-weighted indexes. When the OUTSAS= option is specified and the Patell (default), Standardized Cross-Sectional, EGLS or Collins-Dent test was used in the original event study, specifying the WEIGHTS option on the Extract statement will add WLS weights and standardized CARs to the output data set. One can still use WPREFIX= as in earlier versions to select weights and specify a variable name prefix, but it is no longer necessary; if WEIGHTS is specified, the default prefix is CARWeight. The prefix for the standardized CARs is SCAR.
  • When the Request statement option CUSIPERM is specified, most listing output and output data sets now include the original CUSIP from the request file. In data sets, the variable name is Request_File_CUSIP.
  • Trading-cost adjustment using Nasdaq bid and ask data, introduced in Eventus 7.0, now can be reversed (establish position at bid, liquidate at ask), or made conditional in conjunction with the SHORT option (open long positions at the ask, open short positions at the bid). The TRADINGCOST option of the Evtstudy statement, introduced in Eventus 7.0, now accepts the arguments =REVERSE and =LONGSHORT to implement the above functions.
  • The RUNNING option introduced in Eventus 7.0 now creates running compounded abnormal returns when the BUYHOLD option is in effect. Previously running cumulative abnormal returns were created regardless of BUYHOLD.
  • The SIZEINDX=OWNSYSTEM, BETAINDEX=OWNSYSTEM and STDINDEX=OWNSYSTEM options select size decile, beta decile and standard deviation decile portfolios, respectively, as market index replacements. NYSE-AMEX decile portfolios are selected used for NYSE and AMEX stocks and Nasdaq portfolios for Nasdaq stocks. For SIZEINDX, the OWNSYTEM differs from OWNMARKET in that the latter selects NYSE decile portfolios for NYSE stocks and AMEX decile portfolios for AMEX stocks.
  • Eventus internally checks SASNONCRSP mini-database data sets and stops and prints a log message if the data sets do not meet requirements.
  • SASNONCRSP mini-database data sets no longer need to be indexed before use with Eventus. Eventus will index the data sets if necessary.
  • The SASNONCRSP mini-database header data set (HEADDS= specification) is no longer needed. Eventus internally creates a temporary header data set.
  • The ESTINTER option, which allows different data frequencies to be used for the estimation period and event period, now is allowed in more situations than before. The FFF_EST= option, which works like the FFF= option, has been introduced to specify Fama-French factor input data for the estimation period frequency.
  • When the event date from the request file is altered by an AUTODATE, SHIFT1 or SHIFT2 option, or by a non-daily data frequency option such as MONTHLY, the original event date is stored under the variable name Original_name where name is the corresponding event date variable Eventdat, Eventda1, Eventda2, CRSPday, CRSPday1 or CRSPday2 as documented in the User�s Guide.
  • When the user specifies a SIZEINDX, BETAINDEX or STDINDEX option in an event study, the capitalization decile or risk decile number appears in the parameter estimates section of the printed output (unless PAGE=TALL is specified, or the parameter estimates section is suppressed with the NOPLIST option or NOPRINT option).
  • When the user specifies a SIZEINDX, BETAINDEX or STDINDEX option in a Returns statement run with the VSAS option, the capitalization decile number or risk decile number appears in the output SAS data set under the variable (column) name __CAP__. (This has been done for event studies since several releases ago, but is new for the Returns statement output.)
  • The new PVALUE option of the EVTSTUDY statement prints numeric p-values instead of significance symbols.
  • The STACK option of the EVTSTUDY statement has been completely redesigned and now prints medians below means and numeric p-values below test statistics. It is no longer considered experimental.
  • The event study output now flags the positive:negative abnormal return count with symbols like <<< and >>> to indicate the significance and direction of the generalized sign test. This was a feature of Eventus 6.x output that many users requested be brought back.
  • The option NONAMES is now available on the PRICES, RETURNS and VOLUME statements. It has the same effect as the longstanding EVTSTUDY statement option of the same name: it suppresses the listing of names and data availability that appears by default in the Output window or .lst file. This can be useful for large data extractions.
  • The variable _WEIGHT_ in output data sets has been changed from five characters to 14 characters and now provides a more specific description of non-default indices or decile portfolios.
  • The WINDOWS statement is not sensitive to spacing (except in TWIN event studies). As long as the word windows is followed by at least one blank, zero or more blanks may appear inside and between window specifications.
  • Event date labels (�descriptors�) for TWIN event studies can be up to 64 characters (letters, number, underscores) each.
  • In monthly event studies with CRSP data, calendar event dates in a SAS data set request file can be any day of the month. Previously, calendar dates in a text request file could be any day but a SAS data set request file for monthly runs had to contain first-of-month dates.
  • The comparison-period mean and raw returns benchmarks, selected by the Evtstudy statement options CP and RAW, now are available when the VALUE option also is in effect. Previously the CP and RAW options were available only with the BOTH option or the default equal-weighted index.
  • The various types of market capitalization (size) decile portfolios are more clearly distinguished in the output.
  • Full compatibility with SAS 9.x (9.1.2 or greater recommended).
  • Full compatibility with CRSP U.S. Stock Database formats shipped in 2005 and 2006 including both the original and expanded daily database calendars.
  • Several messages displayed in the SAS log have been edited for clarity, and new messages introduced to provide additional information. Several labels and headings in listing output and variable name labels in output data sets have been enhanced also.

What Was New in Eventus 7.0 (And Still in Eventus 8.0 and 9.0, Of Course)

  • Fama-French factor support. (Uses calendar-time and Ibbotson RATS event study methods).
  • Skewness-adjusted test for long-run buy-and-hold return event studies.
  • Ability to value-weight event study samples.
  • Bootstrapped versions of 5 parametric event study tests (including skewness-adjusted).
  • Adjust for bid-ask spread (for Nasdaq samples using CRSP data, or with appropriately structured user-assembled data from any source).
  • Ibbotson RATS event study methods.
  • Calendar time event study methods including Fama-French.
  • Wilcoxon signed-rank test.
  • More complete features for TWIN option (firm-specific length event window) event studies.
  • Volume event studies.
  • More complete and convenient options for pre-extracted data from non-CRSP sources.
  • Extensive support for event studies using Datastream data through the optional EventStream application.
  • And more!

Back Home Next

� 1999-2012 Cowan Research LC                             Site Search     Support     Sales     Home     Privacy    Legal