Form NT-NSAR Files Dataset

The Form NT-NSAR Files Dataset is a closed historical archive of every Form NT-NSAR and Form NT-NSAR/A submission filed on EDGAR between January 1994 and June 2018, when the SEC retired the underlying Form N-SAR and replaced it with Form N-CEN. Each record corresponds to a single Rule 12b-25 notification of late filing of Form N-SAR — the semiannual or annual operational report historically required of registered management investment companies and unit investment trusts under Section 30 of the Investment Company Act of 1940 — or to an amendment of such a notification (NT-NSAR/A). Filings are made by the registered investment company itself, executed by an authorized officer such as the principal financial officer, treasurer, or fund administrator. A record is delivered as an EDGAR accession folder inside a monthly ZIP container and bundles a normalized metadata.json descriptor together with the original EDGAR-submitted documents (TXT, occasionally HTML or PDF; image files excluded). The dataset is static: no new NT-NSAR filings have been generated since the June 1, 2018 cutover to Form NT-NCEN.

Update Frequency
Daily
Updated at
2026-04-15
Earliest Sample Date
1994-01-01
Total Size
8.0 MB
Total Records
2,760
Container Format
ZIP
Content Types
TXT, JSON, HTML, PDF
Form Types
NT-NSAR, NT-NSAR/A

Dataset APIs

Programmatically retrieve the full list of dataset archive files, download URLs and dataset metadata.

Dataset Index JSON API

Download the entire dataset as a single archive file.

Download Entire Dataset:

Download a single container file (e.g. monthly archive) from the dataset.

Download Single Container:

Dataset Files

232 files · 8.0 MB
Download All
2018-05.zip5.1 KB1 records
2018-03.zip8.0 KB2 records
2018-01.zip11.2 KB3 records
2017-12.zip4.0 KB1 records
2017-11.zip16.8 KB5 records
2017-10.zip15.2 KB3 records
2017-08.zip15.4 KB4 records
2017-05.zip23.6 KB5 records
2017-03.zip2.7 KB1 records
2017-02.zip4.0 KB1 records
2016-12.zip2.6 KB1 records
2016-11.zip15.8 KB4 records
2016-08.zip16.3 KB4 records
2016-06.zip18.2 KB5 records
2016-05.zip10.8 KB3 records
2016-02.zip12.2 KB3 records
2015-12.zip8.5 KB2 records
2015-11.zip14.3 KB4 records
2015-10.zip6.1 KB1 records
2015-07.zip5.6 KB2 records
2015-06.zip12.8 KB4 records
2015-04.zip6.0 KB2 records
2015-03.zip20.3 KB4 records
2015-02.zip3.7 KB1 records
2015-01.zip8.7 KB2 records
2014-12.zip2.5 KB1 records
2014-11.zip4.7 KB1 records
2014-08.zip3.6 KB1 records
2014-06.zip6.1 KB1 records
2014-05.zip7.9 KB3 records
2014-04.zip11.9 KB4 records
2014-03.zip9.4 KB2 records
2014-02.zip6.7 KB3 records
2014-01.zip2.8 KB1 records
2013-12.zip3.2 KB1 records
2013-11.zip20.5 KB5 records
2013-08.zip2.6 KB1 records
2013-05.zip22.6 KB6 records
2013-04.zip3.0 KB1 records
2013-03.zip12.0 KB4 records
2013-02.zip16.1 KB2 records
2013-01.zip13.2 KB5 records
2012-12.zip3.4 KB1 records
2012-11.zip6.3 KB2 records
2012-09.zip15.5 KB4 records
2012-08.zip35.2 KB11 records
2012-07.zip3.1 KB1 records
2012-05.zip8.6 KB2 records
2012-04.zip13.5 KB3 records
2012-03.zip3.6 KB1 records
2012-02.zip21.2 KB4 records
2012-01.zip14.4 KB4 records
2011-11.zip20.3 KB6 records
2011-08.zip11.9 KB3 records
2011-06.zip4.1 KB1 records
2011-05.zip8.7 KB2 records
2011-04.zip11.0 KB2 records
2011-03.zip25.9 KB6 records
2011-01.zip6.5 KB3 records
2010-11.zip17.1 KB5 records
2010-09.zip7.5 KB3 records
2010-08.zip33.0 KB8 records
2010-07.zip3.5 KB1 records
2010-06.zip10.8 KB2 records
2010-03.zip32.4 KB7 records
2010-02.zip10.4 KB2 records
2009-12.zip5.4 KB1 records
2009-11.zip6.9 KB2 records
2009-10.zip2.4 KB1 records
2009-09.zip7.2 KB2 records
2009-06.zip12.8 KB3 records
2009-05.zip3.4 KB1 records
2009-04.zip20.4 KB4 records
2009-03.zip65.2 KB17 records
2009-02.zip20.0 KB5 records
2009-01.zip42.8 KB10 records
2008-12.zip77.1 KB19 records
2008-11.zip8.8 KB4 records
2008-10.zip2.1 KB1 records
2008-09.zip10.9 KB3 records
2008-08.zip15.5 KB4 records
2008-07.zip3.6 KB1 records
2008-05.zip3.4 KB1 records
2008-04.zip4.6 KB1 records
2008-03.zip84.9 KB17 records
2008-02.zip28.6 KB9 records
2008-01.zip11.6 KB3 records
2007-11.zip11.0 KB3 records
2007-10.zip27.8 KB7 records
2007-08.zip23.2 KB7 records
2007-05.zip25.2 KB7 records
2007-04.zip10.7 KB3 records
2007-03.zip30.4 KB10 records
2007-02.zip10.1 KB4 records
2007-01.zip47.1 KB12 records
2006-12.zip138.0 KB44 records
2006-11.zip41.2 KB11 records
2006-10.zip31.8 KB7 records
2006-08.zip20.8 KB6 records
2006-06.zip23.7 KB7 records
2006-05.zip11.8 KB4 records
2006-04.zip8.5 KB2 records
2006-03.zip90.0 KB30 records
2006-01.zip6.0 KB2 records
2005-12.zip23.6 KB9 records
2005-11.zip76.3 KB26 records
2005-10.zip3.8 KB1 records
2005-09.zip9.9 KB3 records
2005-08.zip120.7 KB46 records
2005-06.zip37.1 KB10 records
2005-05.zip41.9 KB13 records
2005-04.zip10.1 KB3 records
2005-03.zip135.1 KB46 records
2005-02.zip7.4 KB3 records
2005-01.zip15.6 KB7 records
2004-12.zip5.6 KB2 records
2004-11.zip64.2 KB25 records
2004-10.zip5.4 KB2 records
2004-09.zip35.8 KB15 records
2004-08.zip55.8 KB20 records
2004-07.zip2.5 KB1 records
2004-06.zip12.8 KB5 records
2004-05.zip6.7 KB2 records
2004-04.zip9.0 KB3 records
2004-03.zip79.1 KB29 records
2004-02.zip21.8 KB7 records
2004-01.zip30.7 KB9 records
2003-12.zip100.0 KB38 records
2003-11.zip24.4 KB9 records
2003-10.zip8.3 KB3 records
2003-09.zip20.5 KB8 records
2003-08.zip40.4 KB13 records
2003-07.zip11.5 KB5 records
2003-06.zip25.6 KB9 records
2003-05.zip30.9 KB11 records
2003-04.zip36.7 KB15 records
2003-03.zip13.3 KB5 records
2003-02.zip30.5 KB10 records
2003-01.zip30.3 KB10 records
2002-12.zip68.0 KB25 records
2002-11.zip113.8 KB16 records
2002-10.zip26.5 KB11 records
2002-09.zip3.0 KB1 records
2002-08.zip23.4 KB8 records
2002-07.zip24.3 KB9 records
2002-06.zip10.4 KB4 records
2002-05.zip13.0 KB5 records
2002-04.zip56.7 KB23 records
2002-03.zip90.1 KB30 records
2002-02.zip75.6 KB25 records
2002-01.zip15.7 KB6 records
2001-12.zip53.9 KB20 records
2001-11.zip53.7 KB19 records
2001-10.zip65.4 KB24 records
2001-09.zip37.8 KB15 records
2001-08.zip124.7 KB46 records
2001-07.zip74.3 KB30 records
2001-06.zip39.9 KB15 records
2001-05.zip48.2 KB18 records
2001-04.zip52.3 KB19 records
2001-03.zip173.0 KB58 records
2001-02.zip167.2 KB60 records
2001-01.zip54.8 KB22 records
2000-12.zip99.8 KB39 records
2000-11.zip66.3 KB25 records
2000-10.zip61.2 KB23 records
2000-09.zip24.2 KB9 records
2000-08.zip147.1 KB51 records
2000-07.zip64.6 KB26 records
2000-06.zip86.2 KB35 records
2000-05.zip55.4 KB23 records
2000-04.zip59.3 KB24 records
2000-03.zip50.1 KB20 records
2000-02.zip235.4 KB91 records
2000-01.zip101.0 KB40 records
1999-12.zip61.3 KB27 records
1999-11.zip136.4 KB55 records
1999-10.zip11.5 KB5 records
1999-09.zip11.9 KB5 records
1999-08.zip118.3 KB43 records
1999-07.zip19.7 KB8 records
1999-06.zip68.0 KB24 records
1999-05.zip59.2 KB25 records
1999-04.zip56.4 KB21 records
1999-03.zip88.5 KB32 records
1999-02.zip186.3 KB74 records
1999-01.zip28.3 KB12 records
1998-12.zip92.3 KB36 records
1998-11.zip47.8 KB19 records
1998-10.zip29.5 KB12 records
1998-09.zip3.6 KB1 records
1998-08.zip148.1 KB57 records
1998-07.zip6.4 KB2 records
1998-06.zip46.9 KB18 records
1998-05.zip63.7 KB26 records
1998-04.zip11.2 KB5 records
1998-03.zip48.4 KB19 records
1998-02.zip124.8 KB50 records
1998-01.zip17.1 KB7 records
1997-12.zip38.8 KB16 records
1997-11.zip172.2 KB33 records
1997-10.zip11.5 KB5 records
1997-09.zip11.0 KB5 records
1997-08.zip105.8 KB44 records
1997-07.zip59.3 KB20 records
1997-06.zip51.4 KB15 records
1997-05.zip46.7 KB20 records
1997-04.zip42.3 KB15 records
1997-02.zip102.2 KB43 records
1997-01.zip13.6 KB4 records
1996-12.zip48.3 KB21 records
1996-11.zip45.8 KB18 records
1996-10.zip5.0 KB2 records
1996-09.zip3.5 KB1 records
1996-08.zip34.0 KB15 records
1996-07.zip32.9 KB13 records
1996-06.zip7.0 KB2 records
1996-05.zip5.0 KB2 records
1996-04.zip13.6 KB5 records
1996-03.zip23.1 KB9 records
1996-02.zip58.7 KB22 records
1996-01.zip18.6 KB5 records
1995-12.zip14.4 KB4 records
1995-11.zip11.4 KB4 records
1995-09.zip2.6 KB1 records
1995-08.zip22.3 KB6 records
1995-07.zip6.3 KB2 records
1995-06.zip1.7 KB1 records
1995-05.zip24.8 KB8 records
1995-04.zip6.8 KB2 records
1994-11.zip11.6 KB4 records
1994-01.zip8.7 KB3 records

What This Dataset Contains

The dataset captures the complete population of EDGAR submissions carrying form type NT-NSAR (original notifications) and NT-NSAR/A (amendments) across the entire active lifetime of the form. The underlying filing is a Form 12b-25 notification of late filing — the universal Rule 12b-25 cover-sheet under the Securities Exchange Act of 1934 that is also used to notify late filing of 10-K, 10-Q, 20-F, and other periodic reports — restricted in this dataset to the variant that targets Form N-SAR. The substantive content of an NT-NSAR filing is not a financial report. It is a short administrative document, typically one to three printed pages of plain English, whose only legal purpose is to invoke the Rule 12b-25 extension and disclose the cause of the delay.

The dataset covers the full perimeter of the form: every registered investment company filing a Rule 12b-25 notification against Form N-SAR between January 1994 and June 2018. Records are organized into monthly ZIP containers named YYYY-MM.zip. Inside each container, every record is a self-contained accession folder holding a metadata.json descriptor plus the original EDGAR documents. The dataset terminates organically in mid-2018: the SEC's adoption of Form N-CEN under the October 2016 investment company reporting modernization rules rescinded Form N-SAR effective June 1, 2018, after which late-filing notifications for the new annual census report take the form type NT-NCEN rather than NT-NSAR.

Content Structure of a Single Record

What one record represents

One record in the Form NT-NSAR Files Dataset is a single EDGAR submission of either Form NT-NSAR or Form NT-NSAR/A, identified by its 18-character SEC accession number and delivered as a self-contained folder inside a monthly ZIP container. The folder bundles a normalized JSON descriptor of the EDGAR submission header together with the documents that the filer originally transmitted to EDGAR, image files excluded. A record therefore corresponds one-to-one to a single Rule 12b-25 notification of late filing of Form N-SAR by a registered investment company, or to an amendment of such a notification when the form type is NT-NSAR/A.

Within a monthly ZIP archive the record manifests as a directory whose name is the EDGAR accession number with hyphens stripped (for example 000158064216012479 for accession number 0001580642-16-012479). Inside that directory sit two kinds of artifact: a single metadata.json file that always exists, and one or more EDGAR-submitted documents — most commonly a single short plain-text file named ntnsar.txt, occasionally accompanied by HTML or PDF exhibits.

Content layers of a single record

A record has three nested layers, and an extraction pipeline traverses all three:

  1. Container layer — a monthly ZIP archive named YYYY-MM.zip whose root contains a single subdirectory YYYY-MM/, under which sit the accession folders for that month.
  2. Accession folder layer — one directory per filing, named with the 18-digit hyphen-stripped accession number, containing the metadata descriptor and the original EDGAR documents.
  3. Document layer — the EDGAR submission file or files. For NT-NSAR this is almost always one plain-text notification document wrapped in EDGAR's SGML <DOCUMENT> envelope; the substantive Form 12b-25 prose sits inside the <TEXT> block of that envelope.

The hyphenated accession number is preserved in the JSON descriptor (accessionNo field) while the folder name strips hyphens, so the folder name and the metadata field are the canonical join key between the on-disk layout and the structured header.

The metadata.json descriptor

metadata.json is the structured mirror of the EDGAR submission header. Its fields are produced from the SGML header EDGAR generates at the moment of filing acceptance, normalized into JSON. The fields that carry meaningful content for an NT-NSAR record are:

  • formType — either "NT-NSAR" or "NT-NSAR/A". This is the only place in the record where the amendment distinction is captured structurally; the folder layout itself does not separate amendments from originals.
  • accessionNo — the hyphenated 18-character EDGAR accession number, e.g. "0001580642-16-012479".
  • filedAt — ISO-8601 timestamp with timezone offset (Eastern Time), recording the exact moment EDGAR accepted the submission.
  • periodOfReport — the fiscal period-end date of the underlying Form N-SAR whose filing is being delayed, formatted YYYY-MM-DD. Because N-SAR was semiannual, this date is overwhelmingly a March 31, June 30, September 30, or December 31.
  • description — a human-readable label such as "Form NT-NSAR/A - Notification of inability to timely file Form N-SAR: [Amend]".
  • linkToFilingDetails — URL to the primary document (the ntnsar.txt) on EDGAR.
  • linkToTxt — URL to the complete SGML submission text file on EDGAR.
  • linkToHtml — URL to the EDGAR filing-index HTML page for the accession.
  • linkToXbrl — empty string on every record; the form does not carry XBRL.
  • documentFormatFiles — array of every document in the EDGAR submission, each entry carrying sequence, size (byte count, as a string), documentUrl, description, and type. The primary notification is always sequence "1" with type matching the form type and a filer-supplied description. A second entry, with blank sequence and type and description "Complete submission text file", points to the consolidated SGML wrapper EDGAR builds around the whole submission.
  • dataFiles — an empty array for every NT-NSAR record. The field exists because the metadata schema is shared with form types that carry XBRL data exhibits; it should be treated as optional and consistently empty for this form family.
  • seriesAndClassesContractsInformation — usually an empty array. Fund filers that elect to register at the series/class level can populate it with series IDs and class IDs, but the form imposes no requirement to do so on a notification of late filing.
  • entities — an array of filer entities. For each entity the descriptor records companyName (with a parenthesized role suffix such as (Filer)), cik (zero-padded 10-digit central index key), irsNo (often "000000000" for fund shells), fileNo (SEC file number, typically prefixed 811- for registered management investment companies, 814- for business development companies), act (almost always "40", marking the Investment Company Act of 1940), stateOfIncorporation, filmNo (EDGAR film/control number), type (the entity-level relationship to this filing, mirroring the form type), and tickers (an array, frequently multi-valued because a single fund commonly issues several share classes under distinct tickers).
  • id — an internal content hash of the record (MD5-like hex string).

The combination of cik, fileNo, periodOfReport, and formType uniquely places each record in the fund's filing timeline; the 811-/814- file-number prefix together with act="40" confirms that the filer is a registered management investment company governed by the Investment Company Act of 1940, the population to which Form N-SAR (and therefore Form NT-NSAR) applied.

The notification document (ntnsar.txt)

The primary document is a plain-text file wrapped in EDGAR's SGML <DOCUMENT> envelope. The envelope carries five header tags — <TYPE>, <SEQUENCE>, <FILENAME>, <DESCRIPTION>, and <TEXT> — and the entire substantive content sits between <TEXT> and </TEXT>. Inside that block lies the literal Form 12b-25 text, organized into the four numbered parts of the SEC paper form, with a literal <PAGE> token marking page breaks from the original paginated layout.

The opening of the body reproduces the SEC heading and the checkbox grid that selects which report is being delayed:

1 U.S SECURITIES AND EXCHANGE COMMISSION
2 WASHINGTON, D.C. 20549
3
4 FORM 12b-25
5 NOTIFICATION OF LATE FILING

On the checkbox grid the box for Form N-SAR is marked, and a For the Period Ended: line carries the fiscal period whose filing is being delayed. The remainder of the document follows the four standard parts.

Part I — Registration Information

Part I supplies the registrant's legal identifying block: full company name and the address of the principal executive office (street, city, state, ZIP). Layout is plain-text prose with the same line breaks as the SEC paper form. SEC file number and IRS employer identification number sometimes appear here as well, although most NT-NSAR registrants leave the IRS line zeroed because the filer is an investment-company shell rather than an operating entity.

Part II — Rules 12b-25(b) and (c)

Part II contains the formal Rule 12b-25 representations. The filer affirms, by checking the corresponding boxes, that (a) the report could not be filed on the prescribed date without unreasonable effort or expense, (b) the report will be filed within the 15-day extension window permitted by Rule 12b-25(b)(2)(ii) for annual reports (or the shorter window for quarterly-style reports), and (c) any required accountant's statement is attached as Exhibit 1 to the notification. For an N-SAR notification the (c) representation is generally not invoked because N-SAR does not require an auditor's statement, but the boilerplate boxes remain on the form.

Part III — Narrative

Part III is the free-text core of the filing: a short prose explanation of why the underlying N-SAR could not be filed on time without unreasonable effort or expense. Typical content includes late receipt of audited financial statements from the fund's accountant or sub-administrator, changes in fund administrator or accounting agent, delays in obtaining sub-adviser confirmations, system or staffing disruptions, internal-controls remediation, mergers and reorganizations affecting the fund complex, and back-testing of fair-value valuations. This is the only free-form narrative in the entire filing and the only place where the substantive cause of the late filing is disclosed.

Part IV — Other Information

Part IV captures three small pieces of structured information:

  1. The name, telephone number, and area code of an individual at the registrant who can be contacted regarding the notification — typically the treasurer, chief compliance officer, or principal financial officer of the fund.
  2. Two yes/no representations: whether the registrant has filed all reports required during the preceding 12 months (or such shorter period as the registrant was required to file), and whether any significant change in results of operations from the corresponding period of the prior year is anticipated and, if so, an explanation of that change.
  3. The signature block: the typed name and title of the signer, the date of signing, and an /s/-prefixed electronic signature line.

The signature date frequently precedes filedAt by a day or two, since filers sign internally and then transmit to EDGAR shortly afterwards.

What the record includes

A record includes the complete header metadata from the EDGAR submission (in metadata.json), the primary NT-NSAR notification document in its original SGML-wrapped plain-text form, and any HTML or PDF exhibits the filer attached to the original submission. The file-types present in the dataset are TXT, JSON, HTML, and PDF; in practice the overwhelming majority of records consist of a metadata.json plus a single ntnsar.txt, because NT-NSAR notifications rarely have exhibits. When HTML or PDF documents do appear they are preserved verbatim from the EDGAR submission and listed in documentFormatFiles with their original sequence numbers, byte sizes, descriptions, and URLs.

What is excluded

Image files (GIF, JPG, and similar binary image formats) are deliberately stripped from each accession folder. These are almost never substantive for NT-NSAR — typically scanned letterheads or signature images on the rare PDF exhibits — and removing them keeps containers small. Beyond images, no other EDGAR content is removed: the SGML-wrapped notification, the complete-submission text file, and any HTML or PDF exhibits are all retained.

The dataset does not include the underlying Form N-SAR itself. Form N-SAR was a separate filing with its own EDGAR submissions and its own structured exhibits, and is outside the scope of this dataset. Nor does the dataset include other 12b-25 variants such as Form NT 10-K, Form NT 10-Q, Form NT 20-F, or NT-NCEN, which target different periodic reports and live in their own datasets.

Structural evolution over time

The form's required content stayed remarkably stable across the 1994 to 2018 window. The Rule 12b-25 framework, the four-part layout of Form 12b-25, and the checkbox grid identifying which periodic report is being delayed have remained essentially unchanged for the entirety of the dataset's date range. What changed is the surrounding regulatory context rather than the body of the form:

  • Form N-SAR retirement (June 2018). The most consequential structural event for this dataset is the SEC's adoption of Form N-CEN under the modernization rules adopted in October 2016, which rescinded Form N-SAR effective June 1, 2018. After that date, late-filing notifications for the new annual census report take the form type NT-NCEN rather than NT-NSAR. The Form NT-NSAR Files Dataset therefore terminates organically in mid-2018; later notifications by fund filers do not appear here.
  • Filer population. Throughout the lifetime of the form, the filer population was confined to registrants filing under Section 30 of the Investment Company Act of 1940 — registered management investment companies, unit investment trusts, and business development companies — which is reflected in the near-universal act="40" and the 811- (open- and closed-end management companies, unit investment trusts) or 814- (business development companies) file-number prefix in the metadata's entities block.
  • Stability of the narrative requirement. The Part III free-text obligation to explain why the underlying report cannot be filed without unreasonable effort or expense was not amended in any way that materially changes the form's content during the dataset window.

Data-format characteristics

Because Form NT-NSAR is a short administrative notification, it did not participate in the structured-data modernization that reshaped operating-company filings:

  • SGML envelope throughout. From the start of EDGAR filing in 1994 to the form's retirement in 2018, the primary document was distributed as plain text wrapped in the SGML <DOCUMENT> envelope. The four-part Form 12b-25 body is uniformly ASCII, with <PAGE> tokens between parts.
  • HTML and PDF as exhibits, not bodies. When HTML or PDF documents appear in a record they are almost always supporting exhibits rather than the primary notification. The primary ntnsar.txt retains the plain-text format throughout the dataset's history; HTML or PDF were never widely adopted for the primary document of this form, in contrast to operating-company periodic reports where HTML supplanted ASCII as the dominant primary-document format during the early 2000s.
  • Compact submissions. NT-NSAR documents are typically one to three pages of ASCII with no exhibits, financial statements, or structured data, which makes the per-record on-disk footprint very small relative to operating-company filings.

Interpretation notes

Several practical considerations matter for downstream use of a record:

  • Amendments. NT-NSAR/A filings are amendments to prior NT-NSAR notifications. The amendment relationship is not encoded as a parent-accession reference inside metadata.json; reconstructing the amendment chain requires linking on cik and periodOfReport across the record set. Amendments typically re-state the entire Part III narrative rather than diffing against the prior filing.
  • Free-text Part III. Because Part III is unstructured prose, any analysis of why funds were filing late requires natural-language processing of the <TEXT> body. Common cue phrases (unreasonable effort or expense, auditor, subadviser, fair value, internal control) cluster in identifiable patterns but are never tagged.
  • <PAGE> tokens. Page-break markers embedded in the body should be treated as soft separators when reconstructing the document into a linear text; they do not delimit semantically meaningful sections.
  • Multiple tickers per entity. The tickers array on a single filer entity is routinely multi-valued because closed-end funds and open-end funds with multiple share classes register a ticker per class. Reading the ticker as the single identifier of the filer is incorrect; the canonical filer identifier is cik, with fileNo providing the SEC file-number anchor.
  • periodOfReport semantics. periodOfReport refers to the period of the delayed Form N-SAR, not the date the notification itself was filed. The two can differ by several months when the filer is significantly late.
  • Optional/empty schema fields. dataFiles, linkToXbrl, and seriesAndClassesContractsInformation are inherited from the broader EDGAR metadata schema and are usually empty on this form family; they should be parsed as optional rather than treated as missing data.
  • Join key. The canonical join between the on-disk folder name and the metadata descriptor is the hyphen-stripped accession number versus the hyphenated accessionNo field. Any pipeline that walks the folder tree must normalize one form to the other to align them.
  • Signature dating. The signature date inside Part IV and the filedAt timestamp in metadata.json are independent; downstream timelines should distinguish "date signed" from "date accepted by EDGAR".

Who Files or Publishes This Dataset, and When

Who files the record

The filer is always the registered investment company itself (the registrant), identified by its EDGAR CIK and its 1940 Act file number (typically the 811- prefix). The form is executed by an authorized officer of the registrant, most often the principal financial officer, treasurer, or fund administrator.

The filing population is limited to entities subject to Form N-SAR under Section 30 of the Investment Company Act of 1940:

For series trusts, the NT-NSAR is filed at the registrant level even where the underlying delay relates to specific series or classes.

Entities outside this population do not appear in the dataset. Operating-company Exchange Act issuers used Form NT 10-K, NT 10-Q, or NT 20-F. Business development companies (BDCs), although investment companies, file Exchange Act periodic reports and used the NT 10-K / NT 10-Q regime rather than NT-NSAR.

When the record is created or required

The triggering event is the registrant's determination that it cannot timely file Form N-SAR for a given reporting period without unreasonable effort or expense. The underlying N-SAR deadlines are:

  • Management companies: within 60 days after the close of each semi-annual fiscal period (Rule 30b1-1).
  • UITs: within 60 days after the close of each calendar year (Rule 30a-1).

When the registrant cannot meet that 60-day deadline, Rule 12b-25 under the Securities Exchange Act of 1934 requires it to file a notification of late filing on the prescribed form. For N-SAR filers, that form is Form NT-NSAR.

Deadline mechanics:

  • Form NT-NSAR must be filed no later than one business day after the original N-SAR due date.
  • A properly executed NT-NSAR containing the representations required by Rule 12b-25(b) grants an automatic 15-calendar-day extension to file the underlying N-SAR (the extension applicable to annual reports; the 5-day quarterly extension does not apply because N-SAR is not a quarterly report).
  • If the underlying N-SAR is filed within the 15-day window, it is deemed timely. If not, the registrant loses the extension and is treated as a delinquent filer.

Form NT-NSAR/A is an amendment to a previously filed NT-NSAR, used to correct disclosure errors, revise the anticipated N-SAR filing date, or update identifying information. It is not an independent late-filing notification.

Historical scope

The dataset is a closed historical archive covering EDGAR submissions from January 1994 through June 2018. The earliest records reflect the phase-in of electronic filing for investment company reports; the latest records correspond to the final N-SAR cycle before the SEC retired Form N-SAR and replaced it with Form N-CEN as part of the 2016 investment company reporting modernization (effective June 1, 2018). No new NT-NSAR filings are accepted after the cutover, and the dataset is static.

Important distinctions

  • NT-NSAR vs. NT-NCEN. After June 2018, the analogous notification for the successor Form N-CEN is Form NT-NCEN. NT-NCEN filings are not included in this dataset.
  • NT-NSAR vs. operating-company NT forms. Exchange Act issuers used NT 10-K, NT 10-Q, or NT 20-F. NT-NSAR was tied specifically to the N-SAR obligation under the 1940 Act.
  • BDCs. Treated as Exchange Act periodic filers despite being investment companies; they did not file NT-NSAR.
  • Filer vs. affected funds. The legal filer is the registrant (often a series trust), even when the delay affects only certain series or classes.
  • NT-NSAR does not satisfy the N-SAR obligation. It only provides the extension and notifies the Commission of the delay; the underlying N-SAR (with required financials and exhibits) must still be filed within the extension window to be deemed timely.

How This Dataset Differs From Similar Datasets or Filings

Form NT-NSAR sits at the intersection of two regulatory regimes: the Rule 12b-25 late-filing framework under the Securities Exchange Act of 1934, and the investment company periodic reporting obligations under the Investment Company Act of 1940. The closest comparison targets fall into three groups: the underlying report whose lateness is being noticed (N-SAR), the successor regime that replaced it (N-CEN and NT-NCEN), and the parallel Rule 12b-25 notifications filed for other report types (NT-NCSR, NT-NCSRS, NT 10-K, NT 10-Q, NT 20-F).

Form N-SAR and Form N-SAR/A

N-SAR was the semi-annual report (annual for unit investment trusts) required of registered management investment companies and UITs under Section 30 of the Investment Company Act and Rule 30b1-1. It is the substantive report; NT-NSAR is the procedural notification that the substantive report will arrive late.

The two datasets are tightly coupled but share almost no content. N-SAR carried structured operational and financial data in a fixed answer-format: portfolio turnover, sales loads, advisory fees, custodian and transfer agent identity, distribution arrangements, securities lending activity, and similar fields. NT-NSAR contains only the 12b-25 cover-sheet narrative — a brief explanation of the delay, the expected filing date, and a representation about whether the lateness signals a material change in results. Both forms were discontinued together in June 2018.

Use N-SAR for fund operations and economics; use NT-NSAR for compliance behavior, late-filing patterns, and audit-cycle disruption.

Form N-CEN

N-CEN replaced N-SAR effective June 1, 2018 and is the modern annual census-style report for registered investment companies, filed in structured XML under Rule 30a-1. N-CEN is the successor to N-SAR, not to NT-NSAR — it is substantive, not a notification. The relevance is temporal: NT-NSAR is a closed historical series (1994 to mid-2018), and any continuous time series of investment-company late-filing behavior must stitch NT-NSAR to NT-NCEN. Note that the underlying cadence also shifts from semi-annual (and annual for UITs) under N-SAR to annual under N-CEN, which changes the denominator of any lateness rate.

Form NT-NCEN

NT-NCEN is the direct functional successor to NT-NSAR — the same Rule 12b-25 notification mechanism, redirected at the new annual N-CEN report. Structure and narrative content are nearly identical. The differences are temporal and quantitative: NT-NCEN covers a single annual filing per registrant rather than semi-annual N-SAR filings, and it remains an active series while NT-NSAR is closed. NT-NCEN is the natural forward extension of this dataset for continuity research; it is out of scope for pure historical N-SAR-era analysis.

Form NT-NCSR and Form NT-NCSRS

Form NT-NCSR and NT-NCSRS are Rule 12b-25 notifications filed by the same investment-company population as NT-NSAR, but for different underlying reports: Form N-CSR (certified annual shareholder reports) and Form N-CSRS (certified semi-annual shareholder reports) under Rule 30b2-1. The content of those underlying reports — audited or unaudited shareholder financial statements, management discussion, and Sarbanes-Oxley certifications — differs fundamentally from the operational census data on N-SAR.

A fund late on its shareholder report files NT-NCSR or NT-NCSRS; a fund late on its operational census report filed NT-NSAR. The two notification streams are independent and can co-occur for the same fund in the same period. Unlike NT-NSAR, NT-NCSR and NT-NCSRS remain active.

Form NT 10-K and Form NT 10-Q

NT 10-K and NT 10-Q are the operating-company analogs of NT-NSAR. All three use the identical Rule 12b-25 cover-sheet structure and Part III narrative explaining inability to file timely. The boundary is the filer population and underlying report: NT 10-K and NT 10-Q come from operating companies filing 1934 Act annual or quarterly reports; NT-NSAR came from registered investment companies filing a 1940 Act report.

A pooled Rule 12b-25 dataset can be useful for studying notification mechanics, but mixing operating-company and fund-complex late filings without segmentation will conflate very different drivers — audit timing, restatements, and M&A on the operating-company side versus fund mergers, adviser changes, and transfer agent transitions on the fund side.

Form NT 20-F

NT 20-F is the foreign private issuer analog within the same Rule 12b-25 framework, covering late filings of the annual Form 20-F. It shares the cover-sheet design with NT-NSAR but applies to a disjoint filer population — FPIs reporting under the 1934 Act, not registered investment companies. It is relevant only for mapping the perimeter of the 12b-25 notification family, not as a substitute or complement for NT-NSAR research.

Boundary summary

The Form NT-NSAR dataset captures a narrow conjunction: a registered investment company or UIT, invoking Rule 12b-25, to delay Form N-SAR, between January 1994 and June 2018. No other dataset isolates this slice. N-SAR holds the substantive report but not the lateness signal. NT-NCEN carries the lateness signal forward but onto a different report on a different cadence. NT-NCSR and NT-NCSRS share the filer population but cover shareholder reports rather than census reports. NT 10-K, NT 10-Q, and NT 20-F share the 12b-25 architecture but apply to entirely different filer regimes. NT-NSAR is best treated as a compliance-behavior series, complementary to N-SAR rather than substitutable for it, and sharply bounded by the 2018 retirement of the N-SAR regime.

Who Uses This Dataset

Because the dataset is a closed historical archive, users work retrospectively: reconstructing fund-complex reporting histories, studying the N-SAR regime, or completing Rule 12b-25 archives. The fields that matter most across user types are CIK and registrant identity, periodOfReport, the 12b-25(b)/(c) elections, and the Part III narrative explanation.

Fund compliance officers and in-house fund counsel

Used during internal audits, exam responses covering legacy periods, and diligence on funds inherited through adviser transactions or sub-advisory changes. They pull every NT-NSAR under a fund family by CIK, map missed N-SAR semiannual periods via periodOfReport, and use Part III narratives to characterize stated causes (system conversions, service-provider transitions, audit delays, fair-valuation issues). Output is a reporting-history memo or representation to a successor adviser or trustee.

Regulatory researchers studying the N-SAR to N-CEN transition

Treat the dataset as a complete population of fund-side Rule 12b-25 notifications across the form's active window. They analyze filedAt and periodOfReport for seasonality and clustering, quantify 12b-25(b) versus (c) elections, and code Part III narratives to track root causes over time. Output is longitudinal analysis of late-filing behavior and historical context for current N-CEN exception-reporting policy.

Fund administrators and accountants conducting service-provider reviews

Join entities and CIK to provider-to-fund maps to test whether NT-NSAR filings cluster around specific administrators, accounting agents, or platform transitions. periodOfReport and Part III explanations drive the analysis. Output is a service-provider reliability review or benchmarking input for RFPs and contract renewals.

Academic researchers in fund governance and operational risk

Merge NT-NSAR records with N-SAR fund-level data, returns, board composition, and adviser identity to test whether late-filing events correlate with subsequent adviser turnover, board changes, or performance. Part III narratives support hand-coded studies of stated causes; the 12b-25(b) representation distinguishes routine extensions from serious filing breakdowns.

Litigation support and forensic accounting teams

Construct authoritative reporting timelines in disputes involving fund advisers, administrators, or auditors. They align every NT-NSAR for a CIK against the eventual N-SAR or N-SAR/A filings to compute actual versus permitted delay, then mine Part III for the registrant's contemporaneous explanation. The 12b-25(c) election is read carefully because it touches anticipated material changes in results of operations. Output is a timeline exhibit or expert report appendix.

Fund adviser M&A diligence teams

Surface historical reporting friction in a target adviser by running CIK searches across every fund it manages, clustering NT-NSAR filings by periodOfReport, and reading Part III for patterns of operational stress around semiannual N-SAR cycles. Output feeds purchase-price adjustments, reps and warranties, or post-closing remediation plans.

Data engineers building historical EDGAR archives

Use this dataset to fill the registered-investment-company portion of the Rule 12b-25 universe (alongside NT 10-K, NT 10-Q, NT 20-F, NT-NCSR) for the 1994 to 2018 window. They consume accessionNo, filedAt, CIK, and entity metadata to normalize records and ingest document payloads for downstream text extraction.

LLM and RAG developers working with fund-disclosure text

Use the Part III narratives as a compact, well-bounded corpus tied to structured metadata for training classifiers that label delay reasons, summarize operational disruptions, or link narrative content to periodOfReport and 12b-25 election flags.

Specific Use Cases

The use cases below anchor in the dataset's specific structure: CIK and file-number identity, periodOfReport, the 12b-25(b)/(c) representations, and the Part III free-text explanation.

1. Reconstructing late-filing histories for adviser-acquisition diligence

A fund-adviser M&A diligence team enumerates every CIK managed by a target adviser, pulls all NT-NSAR and NT-NSAR/A records for those CIKs, and clusters them by periodOfReport. Repeated lateness around the same March 31 or September 30 N-SAR cycles, or recurring Part III references to audit firms or sub-administrator transitions, becomes evidence for purchase-price adjustments, reps and warranties on regulatory compliance, or post-closing remediation conditions.

2. Building a coded taxonomy of stated delay reasons

Regulatory researchers and academics treat Part III as a closed-corpus NLP task. They strip <PAGE> tokens from ntnsar.txt, segment on standard cue phrases (unreasonable effort or expense, auditor, sub-adviser, fair value, internal control, merger), and hand-label or LLM-classify a sample to produce a stated-cause taxonomy. Outputs are time series of delay causes from 1994 to 2018 and inputs to studies of N-SAR-era operational risk and the design rationale for N-CEN exception reporting.

3. Computing actual-versus-permitted delay in litigation timelines

Forensic accounting and litigation-support teams join NT-NSAR records (by cik and periodOfReport) against the eventual N-SAR or N-SAR/A acceptance dates for the same period. The 15-day Rule 12b-25(b) extension window from filedAt defines the "permitted" cutoff; the gap between that cutoff and the substantive N-SAR filedAt quantifies the breach. The 12b-25(c) election and Part III narrative supply the registrant's contemporaneous explanation for expert reports and timeline exhibits.

4. Service-provider reliability benchmarking for fund administrator RFPs

A fund administrator or accountant joins the entities block (CIK, fund file numbers under 811- and 814-) against an external map of fund-to-administrator and fund-to-auditor assignments for the relevant period. Counting NT-NSAR filings per provider-year, weighted by fund complex size, produces a reliability scorecard. Part III text explicitly naming administrator transitions, conversions of accounting platforms, or auditor handoffs adds qualitative evidence for RFP responses and contract renewals.

5. Stitching a continuous fund-side 12b-25 series across the N-SAR to N-CEN transition

Compliance teams and historians building a single 1994-to-present view of registered-investment-company late filings concatenate NT-NSAR with NT-NCEN starting June 1, 2018. The dataset's formType, periodOfReport, cik, and fileNo align cleanly with NT-NCEN's schema, but the underlying cadence shifts from semi-annual N-SAR to annual N-CEN, so the analyst must renormalize the lateness denominator at the regime break.

6. Identifying amendment chains and restated narratives

Because metadata.json does not carry a parent-accession reference, reconstructing amendment chains requires linking all NT-NSAR/A records back to the original NT-NSAR on the same cik plus periodOfReport. The output is a per-fund amendment ledger that distinguishes notifications corrected for clerical reasons from those whose Part III narrative was materially re-stated, supporting compliance memos and exam-response packages covering legacy periods.

7. Training a delay-reason classifier on a bounded compliance corpus

LLM and RAG developers use the Part III bodies as a small, clean fine-tuning or evaluation corpus for compliance-text classification: each ntnsar.txt is paired with structured labels from metadata.json (form type, period quarter-end, file-number prefix distinguishing 811- management companies from 814- BDCs). The resulting classifier can be redeployed against active NT-NCEN, NT-NCSR, and NT 10-K filings, with NT-NSAR providing a labeled historical baseline.

Dataset Access

Dataset Index JSON API: https://api.sec-api.io/datasets/form-ntnsar-files.json

This endpoint returns the dataset's metadata along with the full list of container files. The metadata includes the dataset name, description, last updated timestamp, earliest sample date (1994-01-01), total record and size counters, the form types covered (NT-NSAR, NT-NSAR/A), the container format (ZIP), and the file types contained inside each container (TXT, JSON, HTML, PDF). The containers array lists every monthly ZIP archive with its key, size, records, updatedAt timestamp, and downloadUrl. Use this endpoint to monitor which containers changed in the latest refresh run so you can selectively re-download only what's new.

This endpoint does not require an API key.

Example response:

Example
1 {
2 "datasetId": "1f13365b-9ae0-6977-b435-8aa67686a569",
3 "datasetDownloadUrl": "https://api.sec-api.io/datasets/form-ntnsar-files.zip",
4 "name": "Form NT-NSAR Files Dataset",
5 "updatedAt": "2026-04-15T08:04:25.558Z",
6 "earliestSampleDate": "1994-01-01",
7 "totalRecords": 2760,
8 "totalSize": 7985619,
9 "formTypes": ["NT-NSAR", "NT-NSAR/A"],
10 "containerFormat": "ZIP",
11 "fileTypes": ["TXT", "JSON", "HTML", "PDF"],
12 "containers": [
13 {
14 "downloadUrl": "https://api.sec-api.io/datasets/form-ntnsar-files/2018/2018-06.zip",
15 "key": "2018/2018-06.zip",
16 "size": 45213,
17 "records": 12,
18 "updatedAt": "2026-04-15T08:04:25.558Z"
19 }
20 ]
21 }

Download Entire Dataset: https://api.sec-api.io/datasets/form-ntnsar-files.zip?token=YOUR_API_KEY

Downloads the complete dataset as a single ZIP archive covering all NT-NSAR and NT-NSAR/A filings from January 1994 through the form's discontinuation in June 2018. The full archive is small, so a bulk download is trivial and usually the simplest way to work with this dataset. This endpoint requires an API key.

Download Single Container: https://api.sec-api.io/datasets/form-ntnsar-files/2018/2018-06.zip?token=YOUR_API_KEY

Each container is a year-month ZIP archive that bundles all filings submitted in that month. Pass the key value from the containers array of the dataset index JSON to download an individual archive instead of the full dataset. This endpoint requires an API key.

Frequently Asked Questions

What form does this dataset cover?

The dataset covers Form NT-NSAR and its amendment variant Form NT-NSAR/A — the Rule 12b-25 notification of late filing submitted by registered investment companies when they cannot timely file Form N-SAR, the periodic operational report historically required under Section 30 of the Investment Company Act of 1940.

What does one record in the Form NT-NSAR Files Dataset represent?

One record is a single EDGAR submission of Form NT-NSAR or Form NT-NSAR/A, identified by its 18-character SEC accession number and delivered as a self-contained accession folder inside a monthly ZIP container. The folder bundles a metadata.json descriptor with the original EDGAR documents — most commonly a single plain-text ntnsar.txt notification.

Who is required to file Form NT-NSAR?

The filer is always the registered investment company itself: open-end and closed-end management investment companies (which filed N-SAR semi-annually under Rule 30b1-1) and registered unit investment trusts (which filed N-SAR annually under Rule 30a-1). The form is executed by an authorized officer such as the principal financial officer, treasurer, or fund administrator. Operating companies and foreign private issuers used different NT forms (NT 10-K, NT 10-Q, NT 20-F) and do not appear in this dataset.

What time period does the dataset cover?

The dataset is a closed historical archive of EDGAR submissions from January 1, 1994 through June 2018. It terminates when the SEC retired Form N-SAR under the 2016 investment company reporting modernization rules, replacing it with Form N-CEN effective June 1, 2018; subsequent late-filing notifications use Form NT-NCEN, which is not included here.

What file formats are in the dataset?

The dataset is distributed as monthly ZIP containers named YYYY-MM.zip. Inside, each accession folder contains a normalized metadata.json plus the original EDGAR documents. File types present are TXT (the SGML-wrapped ntnsar.txt notification), JSON (the metadata descriptor), and occasionally HTML or PDF exhibits. Image files (GIF, JPG, and similar) are deliberately stripped.

How does this dataset differ from Form N-SAR or Form NT-NCEN?

Form N-SAR is the substantive operational report; Form NT-NSAR is only the procedural notification that the N-SAR will arrive late. The two share filer population and time window but carry almost no overlapping content. Form NT-NCEN is the direct functional successor to NT-NSAR after June 2018, using the same Rule 12b-25 mechanism but targeting the new annual Form N-CEN; for any continuous fund-side 12b-25 time series, NT-NSAR must be stitched to NT-NCEN at the June 1, 2018 regime break.

Where is the cause of the late filing disclosed in a record?

The substantive cause is disclosed in Part III of the Form 12b-25 body inside ntnsar.txt — the only free-text narrative section in the filing. Typical disclosures include late receipt of audited financials, changes in fund administrator or accounting agent, sub-adviser confirmation delays, internal-controls remediation, fund mergers, and fair-value back-testing. Because Part III is unstructured prose, automated analysis requires natural-language processing of the SGML <TEXT> body.