ibes_ltg

Retrieve and process data on EPS Long-Term Growth (LTG) forecasts from WRDS IBES database.

The Forecast Period End Date (fpedats) is the ending month and year of the fiscal period to which the estimate applies (unless you’re asking for a long-term-growth estimate, in which case the horizon is 3-5 years, so no explicity fiscal period is applicable).

The Activation Date (actdats) is the date that the forecast/actual was recorded by Thomson Reuters.

The Announce Date (anndats) is the date that the forecast/actual was reported.

The Review Date (revdats) is most recent date that an estimate was confirmed as accurate.

The Forecast Period Indicator (fpi) contains information about the horizon (how far into the future we are estimating). Key values: ‘0’ for LTG, ‘1’-‘5’ for 1 to 5 years in the future, ‘6’-‘9’ for 1 to 4 quarters in the future. Farther horizons are available but they are extremely poorly populated.

Note that for Long Term Growth (LTG) estimates (fpi='0'), you must NOT select “Forecast Period End Date” as the Date Variable or the query will not return any estimates.

TICKER is the IBES ticker, which is not necessarily the same as the offical ticker of the firm.

It is possible for a contributing broker to provide multiple revisions to an estimate on the same day. In this scenario, all estimates are available in the Detail history files and only the most current estimate is included in the mean.

The Brokers (estimator) and Analysts (analys) are provided under numeric codes.

Estimate Revisions

There are estimates which are dated “after” the announcement date. We have no explanation other than the entry is in error.

Announcement of earnings will increment the FPI variable by 1 in all IBES records for which review date (REVDATS)> report date (ANNDATS_ACT)

If at the time of the next review date the analyst at the same brokerage changes her forecast for the same (TICKER, ANNDATS, FPEDATS, FPI, MEASURE, USFIRM) combination, IBES will add a new observation. If the forecast remains unchanged, IBES will not add new observations, but will adjust the review date accordingly (REVDATS)

from __future__ import annotations
from typing import List

import pandas as pd

import pandasmore as pdm
from finsets.wrds import wrds_api

PROVIDER = 'Refinitiv via WRDS'
URL = 'https://wrds-www.wharton.upenn.edu/pages/get-data/ibes-thomson-reuters/ibes-academic/unadjusted-detail/history/'
LIBRARY = 'ibes'
TABLE = 'detu_epsus'
LINK_LIBRARY = 'wrdsapps_link_crsp_ibes'
LINK_TABLE = 'ibcrsphist'
FREQ = 'M'
MIN_YEAR = 1925
MAX_YEAR = None
ENTITY_ID_IN_RAW_DSET = 'permno'
ENTITY_ID_IN_CLEAN_DSET = 'permno'
TIME_VAR_IN_RAW_DSET = 'date'
TIME_VAR_IN_CLEAN_DSET = f'{FREQ}date'

source

list_all_vars

 list_all_vars ()

Collects names of all available variables from WRDS f{LIBRARY}.{TABLE}

all_vars = list_all_vars()

all_vars.name.count()

source

default_raw_vars

 default_raw_vars ()

source

parse_varlist

 parse_varlist (vars:Union[List[str],str]=None,
                required_vars:List[str]=[], prefix:str='a.')

Adds required variables to requested variables, validates them, and builds the SQL string with their names

	Type	Default	Details
vars	List[str] \| str	None	list of variables requested by user
required_vars	List[str]	[]	list of variables that will get downloaded, even if not in `vars`
prefix	str	a.	string to add in front of each variable name when we build the SQL string of variable names
Returns	str

parse_varlist(['value','fpi'])

'a.value,a.fpi'

source

get_raw_data

 get_raw_data (vars:List[str]=None, required_vars:List[str]=['ticker',
               'anndats'], nrows:int=None, start_date:str=None,
               end_date:str=None, permno_match_score:tuple=(1,))

Downloads vars from start_date to end_date from WRDS ibes.detu_epsus library and adds PERMNO from CRSP

	Type	Default	Details
vars	List[str]	None	If None, downloads `default_raw_vars`; `permno`, `ticker`, and `anndats` added by default
required_vars	List[str]	[‘ticker’, ‘anndats’]	list of variables that will get downloaded, even if not in `vars`
nrows	int	None	Number of rows to download. If None, full dataset will be downloaded
start_date	str	None	Start date in MM/DD/YYYY format
end_date	str	None	End date in MM/DD/YYYY format; if None, defaults to current date
permno_match_score	tuple	(1,)	accuracy of permno-ibes link. 1-6. 1 is best. use >1 with caution.
Returns	pd.DataFrame

ltg = get_raw_data(start_date='01/01/2019', end_date='01/01/2022', nrows=1)
ltg.head()

	ticker	anndats	value	fpi	fpedats	revdats	actdats	estimator	analys	pdf	permno
0	0001	2019-01-04	-174.4	0	None	2019-01-04	2019-01-04	183.0	48368.0	D	14392.0