Absolute time synchronizes data based on actual calendar time. Time periods are specified as the actual calendar dates when the observations occurred, unlike event time where time periods are relative to specified events.
adjdt (Base Date)
Data on the base date are unadjusted, and other data are converted based on the split events between the base date and the time of that data. /djYYYYMMDDn
A binary code is a numeric flag based on the bitwise “or” of several possible values. It is also known as a bit flag or mask. A flag that is the sum of multiple pieces is valid for all those pieces.
A CRSP calendar is a set of time periods with header information about those time periods. The calendar time periods are chosen as points of interest rather than all calendar days, and therefore a daily calendar contains only the dates where trading was conducted on a major US exchange. Data are never provided on weekends or trading holidays. The standard identification of a time period is the date, an integer in YYYYMMDD format, at the end of the period. There are currently five calendars provided with CRSPAccess databases: daily, monthly, weekly, quarterly, and annual. The daily calendar is used to derive the others so that the last trading date in each month, week, quarter, or year is used to build those respective calendars. Time series data are always associated with one of these calendars. The list of time series observations is synchronized with a calendar so that the nth time series observation is associated with the nth calendar period.
Calendar mapping is the operation of converting data to be reported at the frequency of a given calendar. The result data reported for each target calendar period depend on the type of input data and conventions used. Header and event data can be mapped by finding the applicable event or making a composite value from all events relevant to the calendar period. Time series data can also be mapped to different calendars by summarizing observations in a time series with a more frequent calendar, or by reporting an average value in a time series with a less frequent calendar.
CRSPAccess is the name given to the format of CRSP stock and index data files introduced in 1996. The CRSPAccess format is a binary format with data utilities and C random access data libraries supported on multiple platforms. The data specifications of the data are not available; access to the data must use the data utilities or application program interfaces provided. A CRSPAccess database contains a set of binary computer files in a single directory and is identified by the path of this directory.
CRSPDB is another name for a database in CRSPAccess format
cusip (Header CUSIP)
The latest 8 character CUSIP identifier for a security through the end of the file. /ky cusip
A data object is a set of data organized in a common structure. Data in CRSP files use four basic data objects. These are header data, event arrays, time series, and calendars.
daterange (Date Range)
Dates can be entered in YYYYMMDD, YYYYMM, or YYYY formats. Enter a single date or a range of dates separated by a dash.
In CRSP documentation and utility programs, an entity refers to a single security, portfolio, or index. CRSP databases are organized by entity. Stock databases are organized by security, and Index databases are organized by index.
An event array contains a list of unscheduled transactions, or observation or status changes. There is a count of the number of events. The fields in each event are dependent on the specific data item. The time of the event and relevant information are stored for each observation. Unscheduled transactions record events as they occur. The status observations and changes usually contain information that is in effect until modified by another similar event. Examples of event arrays in CRSP Stock files are the name history and the distribution history. The name history has a new name observation recorded when any name information changes. The distribution history records information about all distributions made to shareholders of a security. CRSP provides utilities and programming tools to retrieve event data, or to convert relevant event data into time series.
An event study synchronizes the time series history of securities relative to a selected event in order to measure the effects of that event.
Event time is the time relative to a defined event. Each event contains a security and event date supplied by a user based on CRSP data or outside information. Event time is the number of calendar periods in a time series before or after the event.
Exchange screening is used to restrict data to issues listed on one or more specified exchanges.
factype (Adj Flag)
Toggles whether to use all price factors to adjust data, or only stock splits and stock dividends. /djYYYYMMDDn
grouptype (Group Types)
Historical membership in a predefined group, such as the S&P 500 Universe /hr
hcusip (Historical CUSIP)
A list of any CUSIPs in a security's name history, plus the header CUSIP if no names exist in the name history. /ky hcusip
Header data is information relevant to the entire history of an entity. It usually includes identifiers, summary data, and data ranges. All entities have some kind of header data. The components of the header are specific to the header type. The CRSP stock file header data include the PERMNO, the header CUSIP, and date ranges of a security. CRSP index file header data include index identifier and flags describing the index methodology.
Header ID and Date Ranges
Include date ranges in the report. /hr
A set of variables that identifies an issue and summarizes its classification. Contains the most current information on the issue maintained in the file. /hh
Index group refers to a set of related index series, where each member in the set is one portfolio made up of different subsets of the total universe. The set of all series in a portfolio index is an index group. Index groups are convenient for example, when comparing a security whose membership may fluctuate from portfolio to portfolio over time. Index groups are available only in the CRSP US Index Database and Security Portfolio Assignment Module.
In a portfolio, the weight of each entity is based on the rules of that portfolio. The weight indicates the relative holdings of the security within the portfolio. Reweighting refers to the rules for changing the weights of the existing portfolio components over time.
An index series refers to data and results of a single portfolio of securities. A single market index, a standard selection of securities, or one decile from a set of decile portfolio indexes is an index series. For example, Portfolio 10 of the CRSP Cap-Based Portfolios for NYSE is an index series, whereas all the portfolios of the CRSP Cap-Based Portfolios for NYSE comprise an index group.
In CRSPAccess databases, all indexes are assigned an INDNO. See “Chapter 3: CRSP Index Methodologies” on page 25 of the Data Description Guide for a full list of available CRSP indexes and information on the composition and methodologies for calculating the different indexes.
Market Segment Index
A market segment index is a type of index where the market of eligible issues is divided into a fixed number of portfolios at different rebalancing intervals based on some rule or statistic. The breakpoint function is continuous so that all eligible issues are in exactly one portfolio during each period. The partitioning rules and index calculations are dependent on the specific index methodology. The partitions can be used to define composite portfolios combining the membership of portfolios. A market segment index is also commonly called a Decile Index if there are ten portfolios, and can also be called a Fractile Index in the general case when there are more or less than ten portfolios.
Identification variables effective at specified times during the history of a security. /n
nasdin (NASDAQ Info)
History of an issue’s trading status and classification on The NASDAQ Stock Market (SM). /q
A unique permanent company identification number assigned by CRSP to all companies with issues on a CRSP File. This number is permanent for all securities issued by a company regardless of name changes. /ky permco
A unique permanent identification number assigned by CRSP to each security. You can track a security through its entire trading history in CRSP’s files with one PERMNO, regardless of name or capital structure changes. /ky permno
Price and Index Levels
Price, Ask/High, Bid/Low, Volume, Return
Price, Returns, Returns w/o Divs
Price, Shares, Returns
Set identifier is a predefined subset of a set type in a CRSPAccess database. Data for two different set identifiers with the same set type use the same data variables, but have different characteristics within those structures. For example, daily and monthly stock sets have different set identifiers since the time series are associated with different calendars and different available portfolio types. Multiple set identifiers of the same set type can be present in one CRSPAccess database. The predefined set identifiers in CRSPAccess stock and index files are:
|Data||Set Type||Set Identifiers|
|CRSP Stock Data||STK||
|CRSP Index Data||IND||
400 Monthly Index Groups
420 Monthly Index Series
440 Daily Index Groups
460 Daily Index Series
Set type is a predefined type of financial data supported in a CRSPAccess database. Stock databases support stock (STK), index (IND), and calendar (CAL) set types. Data for each set type have the same data variables, identifiers, and CRSP programming library access functions.
Share Code Screening
Share code screening is used to restrict data to issues with selected share type characteristics. The share types are based on the CRSP share code. Restrictions can be based on the primary type or secondary type of share and company classification included in the share code.
A stock subset is the selected history of data that meets criteria based on time range and identifying information. CRSP allows subsetting by date range, exchange code, share type, NASDAQ National Market status, and when-issued status. Data in the history that do not meet the selected restrictions are excluded before any extractions or calculations are done.
A time series is a list of observations synchronized with a specific calendar of time periods. There is a beginning and ending of valid data and a link to a calendar defining the time periods. Each entity has exactly one observation for each period within its valid range. The observations can be simple values or contain multiple components, depending on the time series. The values of different variables are generated by checking a value at a consistent time in each time period or by summarizing events occurring during the time period. The primary data variables in the CRSP files such as prices and returns are time series. In daily databases there are observations for each trading day, and in monthly databases there are observations for each month. In both databases there are portfolio time series with statistics and assignments generated when portfolios are rebalanced, usually annually. Components of a time series include the type of data, the array of data, the associated calendar, and the beginning and ending range of data for an entity. It is possible to manipulate event and header data into time series, or to convert between different time series items, or to convert between different frequencies of observations. CRSP provides utilities and programming options to manipulate the CRSP variables into a wide range of time series data types.