Skip to contents

Import raw 990 governance fields via the panel990 package

Source: R/get-data.R
get_governance_data.Rd

Builds the input to get_features from the IRS efile archive using the panel990 package. It assembles a sample frame that restricts the data to full 990 filers (RETURN_TYPE == "990"; the governance index does not apply to 990EZ returns, which omit these fields), then calls panel990::panelize() to download the required tables, keep the needed columns, and merge them into one row per filing.

Usage

get_governance_data(
  years,
  eins = NULL,
  source = NULL,
  cache = c("temporary", "retain"),
  path = "efdata",
  verbose = FALSE
)

Arguments

years

Integer vector of tax years to import.

eins

Optional character vector of formatted EINs (EIN2, e.g. "EIN-43-2031361") to restrict the download to given organizations.

source

A panel990::data_source() configuration. Defaults to the NCCS efile v2.1 archive. Point root at a local directory to read already-downloaded tables.

cache

Passed to panel990::panelize(): "temporary" (default) uses a session temp directory; "retain" keeps a durable cache in path.

path

Cache directory used when cache = "retain".

verbose

Print panel990 progress messages.

Value

A data.frame with one row per full-990 filing: ORG_EIN, the raw governance fields, and identifier columns, ready for get_features.

Details

The governance and management fields come from four tables:

F9-P04-T00-REQUIRED-SCHEDULES

Part IV – required-schedule triggers (audited financials, interested-person transactions, non-cash and art/historical contributions).

F9-P06-T00-GOVERNANCE

Part VI – governing body composition and governance/management policies and disclosures.

F9-P12-T00-FINANCIAL-REPORTING

Part XII – accounting method and audit oversight.

SM-P01-T00-NONCASH-CONTRIBUTIONS

Schedule M – non-cash contribution review process.

The tables are imported raw and handed to get_features, which performs the governance-specific normalization. panel990::panel_normalize() only rewrites blank financial fields and leaves these checkbox and count fields untouched, so it is safe (but unnecessary) to apply beforehand.

Install panel990 with remotes::install_github("Nonprofit-Open-Data-Collective/panel990").

See also

get_features, get_scores, get_governance_scores; panel990::panelize() and panel990::create_sfw() for the underlying sample-frame workflow.

Examples

if (FALSE) { # \dontrun{
dat <- get_governance_data(years = 2022)
features <- get_features(dat)
scores <- get_scores(features)
} # }