Form NSAR-U Files Dataset

The Form NSAR-U Files Dataset is the complete EDGAR archive of annual reports filed on Form N-SAR (the NSAR-U variant) by unit investment trusts (UITs) registered under the Investment Company Act of 1940. Each record is a single EDGAR accession — one full NSAR-U or NSAR-U/A submission filed under Section 30 of the Investment Company Act of 1940 and Sections 13 and 15(d) of the Securities Exchange Act of 1934 by a UIT registrant, with the depositor, sponsor, or trustee executing on the trust's behalf. The dataset spans March 1995 through June 1, 2018, when the SEC rescinded Form N-SAR under the Investment Company Reporting Modernization rulemaking and replaced it with Form N-CEN. Filings are distributed as monthly ZIP containers organized by year (YYYY/YYYY-MM.zip), each holding per-accession folders that combine a metadata.json sidecar with the original text-bearing EDGAR documents.

Dataset APIs

Dataset Files

What This Dataset Contains

The dataset contains every Form NSAR-U and NSAR-U/A submission accepted by EDGAR from March 1995 through the form's discontinuation on June 1, 2018. Form N-SAR was a long-running periodic reporting form for registered investment companies; the -U suffix designates the annual report variant for unit investment trusts, as distinct from the management-company variants (-A semi-annual, -B annual, and the -Bf/-Af flavors for SBICs and other special cases). Substantively, the form gathered standardized item-numbered disclosures about portfolio composition, sales and redemption activity, income and expense data, portfolio turnover, the identities of trustees, sponsors, depositors, underwriters and auditors, the use of credit enhancement and insurance on portfolio securities, default and rating history, and a roster of series filed under the trust umbrella.

Initial filings (NSAR-U) and amendments (NSAR-U/A) appear as full, equally-shaped records: an amendment is not a delta against a prior filing but a self-contained replacement that re-states the entire form. Mechanically, Form N-SAR remained a fixed-format, item-driven document throughout its entire lifespan — it was never migrated to HTML or to inline XBRL. From the earliest 1995 records through the final 2018 records, the primary submission is an ASCII text file structured as a sequence of numbered items, with answers placed immediately under each item heading and tabular sections rendered using SGML <TABLE>, <S> and <C> markup rather than HTML tables. The document sits inside EDGAR's legacy <DOCUMENT> / <TYPE> / <SEQUENCE> / <FILENAME> / <DESCRIPTION> / <TEXT> SGML container, with <PAGE> markers separating the logical pages of the printed form.

The dataset is distributed as ZIP containers; the file types inside each archive are TXT (the SGML-wrapped Form N-SAR documents), JSON (the per-record metadata sidecars), and HTML (where EDGAR rendered an HTML index for the submission).

Content Structure of a Single NSAR-U Record

What one record represents

One record in the Form NSAR-U Files Dataset is a single EDGAR accession — one complete Form NSAR-U or Form NSAR-U/A submission filed by a registered unit investment trust. On disk, the record is materialized as a per-accession folder whose name is the 18-digit EDGAR accession number with the dashes stripped. The folder holds two kinds of artifacts: a metadata.json sidecar that re-encodes the EDGAR submission header and adds dataset-level indexing fields, and the original text-bearing documents carried over from the EDGAR submission. Image attachments, if any, are dropped; every text-bearing document in the original SGML wrapper is preserved.

Two stacked structural layers

A single record has two layers:

1. Dataset packaging layer — a folder named for the accession number, containing metadata.json plus one or more EDGAR document files. Only the constituent text-bearing documents are stored locally; the EDGAR-generated "Complete submission text file" wrapper is referenced via URL in metadata but not duplicated on disk, and image attachments are omitted.
2. Filing-content layer — inside the primary .txt document, an SGML-wrapped Form N-SAR consisting of a centered header banner, a gating section (items 1–6) that routes the filer into a sub-form, the UIT-specific item block (items 111–132) that carries the substantive content for NSAR-U, and a final signature page.

The metadata.json sidecar

metadata.json is a flat JSON object that re-encodes the EDGAR submission header and adds dataset-wide indexing fields. The notable top-level fields are:

• formType — "NSAR-U" or "NSAR-U/A".
• accessionNo — canonical dashed accession number (NNNNNNNNNN-YY-NNNNNN).
• filedAt — EDGAR acceptance datetime with timezone offset.
• periodOfReport — fiscal year-end of the reporting period. Because NSAR-U is the annual variant, this is always a fiscal year-end, never a six-month period, regardless of the boilerplate "SEMI-ANNUAL REPORT" wording inside the form body.
• description — human-readable form description.
• linkToFilingDetails, linkToTxt, linkToHtml, linkToXbrl — direct EDGAR URLs to the filing index page, the combined submission text file, the primary HTML rendering (where one exists), and any XBRL exhibit. For NSAR-U, linkToXbrl is always empty because the form never carried XBRL.
• id — stable internal record identifier.
• documentFormatFiles[] — one element per document in the original EDGAR submission.
• dataFiles[] — empty array for NSAR-U; the form never adopted structured-data attachments.
• entities[] — one element per filer/issuer recorded by EDGAR for the submission.

Each documentFormatFiles[] element carries sequence (1-based ordering inside the SGML wrapper), documentUrl (direct EDGAR URL), description (EDGAR-supplied label such as "FORM N-SAR" or "Complete submission text file"), type (SEC document type, typically "NSAR-U" for the form itself), and size (bytes, as a string). The "Complete submission text file" entry is informational only; its bytes are not copied into the per-accession folder.

Each entities[] element carries legacy EDGAR header fields for one filer/issuer: companyName (with the EDGAR role suffix appended, e.g. "... (Filer)"), cik (CIK, 10-digit, zero-padded), fileNo (typically the 811- Investment Company Act registration number identifying the trust), irsNo, fiscalYearEnd (MMDD), act ("40" for the Investment Company Act of 1940), type (form-type echo), and filmNo (EDGAR film number).

The primary document — the NSAR-U text file

The primary document is a single fixed-width ASCII file enclosed in an SGML <DOCUMENT> block whose opening tags state <TYPE>NSAR-U, <SEQUENCE>1, the filer-chosen <FILENAME>, <DESCRIPTION>FORM N-SAR, and a <TEXT> body. Inside <TEXT> the layout reproduces the printed N-SAR form: a centered banner reading FORM N-SAR / SEMI-ANNUAL REPORT / FOR REGISTERED INVESTMENT COMPANIES, followed by a numbered item sequence broken into logical pages by <PAGE> markers.

Header and gating items (1–6)

The opening items establish identity and route the filer into the correct sub-form. They are completed in every NSAR-U record:

• Item 1 — Registrant name, SEC file number, telephone number.
• Item 2 — Mailing address of the principal office.
• Item 3 — First filing on this form? (Y/N).
• Item 4 — Last filing on this form? (Y/N). In late-cycle filings (2017 and the first half of 2018), this is frequently answered Y because filers anticipated the June 1, 2018 N-SAR rescission.
• Item 5 — Small Business Investment Company? If Y, only items 89–110 are required.
• Item 6 — Unit Investment Trust? If Y, only items 111–132 are required.

Because NSAR-U is by definition the UIT annual variant, item 6 is answered Y in every record, and the form effectively jumps from the gating page directly to the UIT block. Items 7 through 110 — which carry disclosures applicable to open-end and closed-end management investment companies — are left blank by design rather than omitted as missing data. Extractors should treat empty items in this range as structurally "not applicable."

UIT-specific items (111–132) — the substantive body

The substantive content of an NSAR-U record lives in items 111–132. These items collectively describe who runs the trust, what it holds, how its units moved during the fiscal year, and how its portfolio is composed:

• Item 111 — Depositor name, file number, and address.
• Item 112 — Sponsor name, file number, and address.
• Item 113 — Trustee name and address.
• Item 114 — Principal underwriter name and 8- broker-dealer file number.
• Item 115 — Independent public accountant name and address.
• Item 116 — Family of investment companies identifier (a 10-letter code grouping related trusts under a sponsor's umbrella).
• Item 117 — Separate account of an insurance company indicator, with sub-items disaggregating variable annuity, scheduled-premium variable life, flexible-premium variable life, and other 1933 Act products.
• Items 118–122 — Counts and dollar values for the number of series in the trust, new registrations during the period, prospectuses delivered, and additional units sold. Dollar amounts are reported in thousands of dollars ($000 omitted).
• Items 123–126 — Dollar values describing unit movements between series and sales-load collections, again in thousands.
• Item 127 — Portfolio composition table, the densest tabular block in the form. Rows partition the portfolio by security type — U.S. Treasury securities, U.S. agency securities, state and municipal tax-free obligations, public utility debt, broker/dealer debt, other corporate intermediate/long-term debt, other corporate short-term debt, broker/dealer equities, investment-company equities, other equities, and an "other" catch-all. Columns report the number of series invested in each category, total assets attributable to each category (in thousands), and total income distributions associated with each category (in thousands). The table is rendered with SGML <TABLE><S><C> markup.
• Items 128–130 — Credit enhancement, insurance, and default questions for portfolio securities (presence of bond insurance, occurrence of defaults, ratings).
• Item 131 — Total expenses incurred by all series (in thousands).
• Item 132 — Roster of 811- Investment Company Act file numbers for every series included in the filing.

Signature page

After item 132 the file closes with a SIGNATURE PAGE block carrying the city and state of signing, the date of signing, the registrant name, the sponsor agent's printed name and title, and a typographic signature line rendered as /s/ Name. The signature page is plain text within the same <TEXT> body — it is not a separate document and not a separate SGML element.

Formatting conventions inside the body

Three conventions matter for accurate extraction:

• Dollar scaling. All dollar values are stated with three zeros omitted; figures must be multiplied by 1,000 to recover dollars. This is a global convention of the form, not a per-item annotation.
• Change-only checkboxes. Many items carry a [/] checkbox prefix indicating that the answer should be supplied only when it has changed from the previous filing. A blank [/]-flagged item is a positive assertion of "unchanged from prior filing," not absence of data.
• SGML tables, not HTML tables. Tabular blocks use <TABLE><S><C> markup. HTML-oriented table parsers will see the content as opaque text; column boundaries must be reconstructed from the <S> (stub-column start) and <C> (column) tag positions or from fixed-width whitespace alignment.

What the record includes

Each per-accession folder reliably includes:

• A metadata.json file capturing the EDGAR submission header, document inventory, and entity roster.
• The primary Form N-SAR document as an ASCII text file inside its SGML <DOCUMENT> wrapper. The filename is filer-chosen — it commonly embeds the trust name, the 811- file number, and the fiscal year — and carries a .txt extension.
• Any other text-bearing documents that were part of the original EDGAR submission (cover letters, correspondence, supplemental schedules), each retaining its EDGAR-assigned sequence and description.

Because the substantive body sits entirely inside the primary document, an NSAR-U accession folder is typically very small — often metadata.json plus a single .txt file totaling a few kilobytes to a few tens of kilobytes.

What the record excludes

Three categories of content are not present in the per-accession folder, even when they existed in the original submission:

• Image attachments are deliberately dropped. N-SAR documents rarely carried images, but when they did (scanned signatures, sponsor logos), those images are not stored.
• The EDGAR-generated "Complete submission text file" that bundles the entire submission's SGML wrapper into a single artifact is referenced through linkToTxt but not copied locally; only the constituent documents are stored.
• XBRL or other structured-data exhibits are not present because Form N-SAR never adopted structured-data reporting. dataFiles[] is empty and linkToXbrl is blank for every record.

Evolution of content and structure, 1995–2018

The internal item structure of Form N-SAR — and therefore of NSAR-U — was remarkably stable across the dataset's 23-year span. The item numbering, the gating logic at items 5 and 6, and the 111–132 UIT block remained materially unchanged from the mid-1990s through the form's final months in 2018. Refinements over the years occurred mainly inside sub-items — expanded breakouts of variable insurance product types in item 117, refinements to the credit-enhancement and default questions in items 128–130 — rather than wholesale restructuring. A 1995 NSAR-U record and a 2018 NSAR-U record share the same skeleton.

Two episodes shaped the dataset's distribution rather than its anatomy. First, the SEC adopted Form N-SAR in its modern numbered-item form well before the dataset begins, so the earliest 1995 records already follow the final structure. Second, on October 13, 2016, the SEC adopted the Investment Company Reporting Modernization rules, which created Form N-CEN as a structured (XML) replacement for N-SAR and set a rescission date of June 1, 2018. This produced a visible behavioral shift in late-cycle records: from mid-2017 onward, item 4 ("Last filing on this form?") is frequently answered Y, the population of NSAR-U filings tapers sharply through the first half of 2018, and the form then disappears from the filing stream.

Evolution of data format

Form NSAR-U is one of the SEC form types that never left EDGAR's legacy ASCII/SGML era. From 1995 through 2018 the primary document is a fixed-width ASCII text file inside the same SGML <DOCUMENT> envelope, with <PAGE> markers and <TABLE><S><C> tabular markup. HTML rendering was never introduced for N-SAR, and inline XBRL was never applicable — when the SEC modernized investment-company reporting, it replaced N-SAR with the XML-based N-CEN rather than retrofitting structured data into N-SAR. Every record in this dataset, regardless of vintage, therefore looks structurally identical at the document-format level: SGML-wrapped ASCII for the body, with a JSON sidecar produced by the dataset for indexing.

Interpretation notes

Several nuances matter for downstream use:

• The banner inside the form body reads FORM N-SAR / SEMI-ANNUAL REPORT even in NSAR-U (annual) filings. The actual period type is determined by the Y/N answers immediately below the banner ("Report for six month period ending" vs. "or fiscal year ending"), not by the banner text. The authoritative period-of-report value is periodOfReport in metadata.json, which always reflects a fiscal year-end for NSAR-U.
• Items 7–110 are blank by design because item 6 routes UIT filers into the 111–132 block. Blank ranges in this region should be treated as "not applicable," not as missing data.
• The [/] checkbox convention means a blank item is sometimes a positive assertion of "unchanged from the previous filing." Reconstructing a current value may require walking back through prior-period filings for the same registrant.
• Dollar amounts are reported with three zeros omitted throughout the form.
• Tabular sections use SGML markup that is visually similar to but mechanically distinct from HTML tables. Generic HTML table parsers will fail on them; column extraction requires SGML-aware parsing or whitespace-based reconstruction.
• An NSAR-U/A record is a full restated filing with the same anatomy as the original; the only structural differences are the formType value and the answers to the amendment-flag items. There is no diff format and no pointer to the prior accession inside the form body itself — the relationship between an amendment and the filing it amends must be reconstructed from CIK, period of report, and filing date.
• A single Form NSAR-U filing may cover a UIT family with many series (item 132 enumerates each series' 811- file number), but the dataset's record granularity is always one EDGAR accession, never one series.
• The entities[] array can contain more than one entity when EDGAR records co-filers or related filers, but a single filer entity is the norm for NSAR-U.

Who Files or Publishes This Dataset, and When

Who files the record

Each record is an annual report on Form N-SAR (NSAR-U variant) filed by a unit investment trust (UIT) registered under the Investment Company Act of 1940. The legal filer is the registered investment company itself; the depositor, sponsor, or trustee executes and submits the form on the trust's behalf.

NSAR-U/A records are amendments to a previously filed NSAR-U. They are submitted by the same UIT registrant under the same CIK, with the "/A" suffix marking the filing as a correction, supplement, or restatement rather than an original report.

The filing population

A UIT is one of the three statutory classes of investment company under Section 4 of the 1940 Act, alongside face-amount certificate companies and management investment companies. A UIT:

• is organized under a trust indenture or similar instrument
• has no board of directors
• issues only redeemable securities representing undivided interests in a fixed, essentially unmanaged portfolio
• does not actively trade its holdings

Typical NSAR-U filers include sponsor-organized defined-portfolio equity trusts, municipal and corporate bond trusts, and insurance separate accounts registered as UITs that fund variable annuity and variable life contracts. Mutual funds, closed-end funds, and ETFs are management investment companies, not UITs, and did not file NSAR-U.

Statutory and regulatory basis

Form N-SAR was adopted under Section 30 of the Investment Company Act of 1940, which requires registered investment companies to file periodic reports prescribed by Commission rule. Two implementing rules split the population:

• Rule 30b1-1 required UITs to file an annual report on Form N-SAR (the NSAR-U variant).
• Rule 30a-1 required management investment companies to file Form N-SAR semi-annually (NSAR-A) and at fiscal year end (NSAR-B).

To the extent a UIT registrant also had Exchange Act reporting obligations tied to its registered securities, Section 13 and Section 15(d) of the Securities Exchange Act of 1934 provided overlapping authority for the filing.

Trigger and cadence

NSAR-U is periodic and fiscal-year-end-driven, not event-driven. The trigger is the close of the UIT's fiscal year, and the report was due within 60 days after fiscal year end.

UITs had no semi-annual report N-SAR obligation. NSAR-U was the only N-SAR-series filing required of them; there is no UIT analogue to NSAR-A or NSAR-B.

NSAR-U/A amendments follow no schedule. They are filed whenever the registrant determines the original NSAR-U requires correction.

Date range and rescission

EDGAR records in this dataset begin in March 1995, tracking the phase-in of mandatory electronic filing for N-SAR in the mid-1990s.

The Commission rescinded Form N-SAR, including NSAR-U, as part of the Investment Company Reporting Modernization rulemaking (Release No. IC-32314, adopted October 13, 2016). The rescission took effect on June 1, 2018, after which UITs and other registered investment companies report census-type information on Form N-CEN instead. The dataset has a fixed terminal boundary at that date; only NSAR-U/A amendments to pre-rescission filings could appear afterward.

Important distinctions

NSAR-U vs. NSAR-A vs. NSAR-B. The three live variants were not interchangeable:

• NSAR-U annual report by a UIT under Rule 30b1-1.
• NSAR-A semi-annual report by a management investment company under Rule 30a-1.
• NSAR-B annual report by a management investment company under Rule 30a-1.

A registrant filed either the UIT track (NSAR-U only) or the management-company track (NSAR-A plus NSAR-B), determined by statutory classification under Section 4, not by asset class or strategy.

Insurance separate accounts. When a variable annuity or variable life separate account is registered as a UIT, the separate account is the filer, not the sponsoring insurer (which is the depositor).

Filer vs. named service providers. The trustee, sponsor, depositor, principal underwriter, custodian, and independent accountant are identified inside the form but are not the filer. The Section 30 obligation runs to the registered investment company.

Entities outside the NSAR-U path. Closed-end funds, ETFs, master-feeder mutual fund structures, and face-amount certificate companies are not UITs for 1940 Act purposes and did not file NSAR-U, despite occasional informal "trust" labeling.

Post-2018 successor. For comparable UIT disclosure after June 1, 2018, look to Form N-CEN. The Section 30 reporting obligation continues; the form has changed.

How This Dataset Differs From Similar Datasets or Filings

Form NSAR-U occupies a narrow slot in SEC investment company reporting: annual operational filings by UITs under Section 30 of the 1940 Act, covering March 1995 through June 2018. The most useful comparison set falls into three groups: other members of the rescinded N-SAR family, the post-2018 modernization forms that replaced it, and adjacent UIT-related disclosures.

Within the N-SAR family

Form NSAR-A — semi-annual report for management investment companies. Same N-SAR data architecture, different filer and cadence. NSAR-A is filed semi-annually by open-end and closed-end management companies, not UITs. Much of its content (board, advisory contracts, 12b-1 plans) has no UIT analog, so NSAR-A cannot substitute for NSAR-U.

Form NSAR-B — annual report for management investment companies. The closest functional analog to NSAR-U: same annual cadence, same Section 30 regime, same rescission date (June 1, 2018). It differs solely in filer population — NSAR-B covers management companies and exposes items on advisers, sub-advisers, underwriters, directors, custodians, and 12b-1 plans, while NSAR-U covers UITs and exposes items on trustee/sponsor identity, unit sales and redemptions, and trust-specific operations.

Form NSAR-U/A — amendments to NSAR-U. Not a separate corpus. NSAR-U/A submissions are included in this dataset and should be joined to their originals by CIK and reporting period rather than treated as independent filings.

Post-2018 replacements

Form N-CEN — annual census report. Direct successor to the entire N-SAR family effective June 1, 2018, covering both UITs and management companies. Differences: structured XML versus N-SAR's legacy text format, and a reorganized data model with non-equivalent item definitions. Continuous time series across the 2018 boundary require a manual crosswalk; NSAR-U remains the only source for UIT operational data before that date.

Form N-PORT — monthly portfolio holdings. Filed by management investment companies and ETFs structured as UITs. Granular position-level holdings, derivatives, and risk metrics — none of which appear in NSAR-U. Conversely, N-PORT carries no trustee, sponsor, or aggregate sales/redemption data. Different content type, not a substitute.

Form N-Q — rescinded quarterly portfolio holdings (2004–2019). Holdings-only, quarterly, and largely management-company focused. Useful as a portfolio complement to NSAR-U's operational data for pre-2018 research, but disjoint in content.

Adjacent disclosures

Forms Form N-CSR / Form N-CSRS — certified shareholder reports. Annual and semi-annual shareholder-facing reports for management companies under Rule 30b2-1, containing audited GAAP financials and SOX certifications. NSAR-U is UIT-only, regulator-facing, and statistical rather than narrative or audited. No content overlap beyond the broad 1940 Act context.

Form 24F-2 — annual notice of securities sold. Annual fee-payment notice for open-end funds and UITs registering units under Rule 24f-2. Same annual cadence and partial filer overlap, but content is limited to sales totals for fee computation, not the full Section 30 operational schedule. Useful for cross-checking NSAR-U sales items only.

Forms Form 485APOS / Form 485BPOS — post-effective prospectus amendments. Forward-looking, investor-facing prospectus updates to Form N-1A and analogous registration statements. Narrative and disclosure-oriented; NSAR-U is backward-looking and statistical. Complementary, not interchangeable.

Form N-8B-2 — UIT registration statement. One-time, point-of-creation filing that establishes trust structure, sponsor, trustee, depositor, and inception portfolio. NSAR-U is the recurring annual counterpart. The natural pair for longitudinal UIT profiles: N-8B-2 for inception structure, NSAR-U for year-by-year operations through fiscal 2018.

Key differences at a glance

• Filer population. UITs only — distinguishing NSAR-U from NSAR-A, NSAR-B, N-CSR/N-CSRS, and most N-PORT filings.
• Cadence. Annual only — distinguishing it from semi-annual NSAR-A, quarterly N-Q, and monthly N-PORT.
• Content. Operational and statistical items prescribed by Section 30 of the 1940 Act — distinguishing it from prospectus (485APOS/485BPOS), registration (N-8B-2), shareholder (N-CSR), holdings (N-PORT, N-Q), and fee-notice (24F-2) filings.
• Historical window. March 1995 through June 2018, then rescinded — distinguishing it from its successor N-CEN, which begins where NSAR-U ends.

Boundary summary

The NSAR-U dataset is defined by the intersection of UIT filer population, annual cadence, Section 30 operational content, and a closed 1995–2018 window. No other SEC dataset reproduces this combination. For pre-2018 UIT operational research it is the authoritative source; N-CEN is the only forward-compatible successor, but the underlying data model changed enough that the two cannot be concatenated without a crosswalk.

Who Uses This Dataset

Because the NSAR-U archive is a closed historical record, users come to it for retrospective work: reconstruction, litigation support, longitudinal research, and the NSAR-to-N-CEN normalization that is unique to historical UIT analysis.

Investment-company regulatory analysts

Compliance and fund-administration analysts use the archive to reconstruct historical UIT reporting practice. They pull sponsor and depositor identification (Item 7), trustee identity (Item 9), series-level CIK and fiscal year end, sales and redemptions (Items 70-74), and the operating expense sub-items to produce gap analyses between pre-2018 NSAR-U practice and current N-CEN obligations.

SEC enforcement, exam, and economic-analysis staff

Examination and enforcement staff working historical UIT matters use the filings and any NSAR-U/A amendments to support docket reconstruction and to test allegations on fee disclosure, distribution practices, portfolio turnover, and sponsor or trustee misstatements. Economic-analysis staff aggregate Item 127 portfolio composition and Items 70-74 sales and redemption flows to build long-run industry baselines.

Forensic accountants and litigation-support analysts

Forensic accountants engaged in UIT sponsor, depositor, or trustee disputes pull period-end net assets, portfolio turnover, advisory and distribution fee items, and the expense sub-items to underpin damages models. NSAR-U/A amendments are especially valuable because they expose prior-period restatements relevant to misreporting claims. Output is expert reports, deposition exhibits, and reconstructed financial timelines.

Securities counsel handling UIT matters

Counsel advising UIT sponsors, trustees, and successor entities retrieve historical filings for diligence: confirming series existence at a given date, verifying Item 7 sponsor and Item 9 trustee identifications, locating fiscal year ends, and pulling prior expense or turnover disclosures. The data supports opinions on succession, indemnification, and historical compliance posture in legacy-trust transactions.

Academic researchers and financial historians

Researchers studying the 1995-2018 UIT industry extract panel data on units outstanding, gross sales and redemptions (Items 70-74), net assets, expense ratios, and portfolio turnover across series and time. Sponsor and trustee fields support studies of consolidation, product life cycles, and the rise and decline of UIT categories. Output is peer-reviewed papers and book chapters on fund-industry evolution.

Data engineers normalizing NSAR-U to N-CEN

Engineering teams at data vendors and quantitative research shops build field-level crosswalks so the NSAR-U archive can be joined to post-2018 N-CEN records into a continuous panel. Mapping centers on Item 7 sponsor and Item 9 trustee identifiers, Item 127 portfolio composition, Items 70-74 sales and redemptions, portfolio turnover, and expense classifications, with accession number and period-of-report metadata aligning each NSAR-U observation to its N-CEN successor.

Commercial data vendors building historical fund datasets

Vendors assembling historical fund-industry products ingest series-level identifiers, sponsor and trustee names, fiscal year ends, total net assets, and expense and turnover metrics into proprietary databases that feed fund directories, fee benchmarking tools, and UIT competitive intelligence.

UIT and ETF product strategists

Product teams designing new trust or ETF offerings study the competitive history of the UIT segment: how comparable trusts were sized, how quickly units sold or redeemed, and what expense levels prevailed. They segment Items 70-74, total net assets, expense ratios, and portfolio turnover by sponsor or trust type to identify viable structures.

Fund-administration consultants

Consultants to UIT sponsors, trustees, and transfer agents benchmark expense structures, turnover, and unit flow across comparable trusts. They mine operating expense sub-items, advisory and distribution fee disclosures, and trustee fields to support fee studies and operational reviews for trusts still in run-off.

Compliance officers for legacy UIT books

Compliance officers at firms still administering NSAR-U-era trusts retrieve original reporting baselines for audits, regulator inquiries, and successor-trustee transitions. They rely on accession metadata, Item 7 and Item 9 identifications, fiscal-period coverage, and historical expense and turnover disclosures to demonstrate continuity of reporting.

LLM and RAG developers for asset-management retrieval

Teams building domain-specific retrieval systems use the item-numbered NSAR-U text and tables as a schema-aware corpus, expanding coverage of the pre-2018 UIT segment that is otherwise sparsely represented in modern fund datasets.

Business and law librarians

Reference and archival staff use the packaged accession metadata and original EDGAR submissions to deliver complete pre-2018 UIT filing sets to academic, legal, and journalistic researchers without per-filing scraping.

Synthesis

The dataset serves a narrow but professionally diverse group working on a closed reporting regime: regulators and enforcement staff reconstructing past UIT conduct, forensic accountants and counsel mining operational data for disputes, researchers building long-run panels, engineers stitching NSAR-U to N-CEN for continuous coverage, and product, consulting, and reference users benchmarking or retrieving the historical record. Each role draws on a different slice of Items 7, 9, 70-74, 127, the expense sub-items, and portfolio turnover, which is why the complete 1995-to-2018 archive, rather than a sample, is the version that matters.

Specific Use Cases

The use cases below describe concrete workflows analysts run against the item-level NSAR-U content and the bridge to post-2018 N-CEN.

1. Building a continuous 1995-present UIT operational panel by crosswalking NSAR-U to N-CEN

Engineering teams join the full NSAR-U population (every NSAR-U and NSAR-U/A filing, period-of-report through fiscal year-ends straddling June 1, 2018) to N-CEN observations starting June 2018. The crosswalk anchors on CIK and 811- file numbers from entities[] and item 132, with field-level mappings from item 112 (sponsor), item 113 (trustee), item 127 (portfolio composition), items 118-126 (sales, redemptions, unit movements), item 131 (total expenses), and portfolio turnover into the corresponding N-CEN XML elements. The output is a single time-aligned UIT panel suitable for long-horizon industry analytics that neither corpus supports on its own.

2. Reconstructing the 1995-2018 UIT industry size and flow series

Researchers and economic-analysis staff aggregate items 118-126 (number of series, additional units sold, sales-load collections, movements between series) and item 131 (total expenses) across the full population to produce annual industry-level series for gross sales, redemptions, expense totals, and series counts. Because dollar figures are scaled in thousands, extraction must apply the global x1,000 rule before aggregation. The result is a baseline for tracking the rise of equity-UIT issuance, the contraction of fixed-income UITs, and structural turning points across 23 fiscal years.

3. Sponsor-and-trustee concentration and consolidation studies

Analysts extract item 111 (depositor), item 112 (sponsor), item 113 (trustee), item 114 (principal underwriter), and item 116 (family-of-investment-companies code) from every filing, then compute per-year Herfindahl indices and entity-level market share by series count and net assets. Tracking the same sponsor or trustee across the panel exposes mergers, sponsor exits, and trustee migrations, including the late-cycle pattern in which item 4 is answered Y from mid-2017 onward as trusts wind down ahead of the N-CEN transition.

4. Portfolio-composition benchmarking from item 127

Forensic accountants, product strategists, and consultants pull the item 127 SGML <TABLE><S><C> block — rows partitioning the portfolio across U.S. Treasuries, U.S. agencies, state and municipal tax-free obligations, public utility debt, broker/dealer debt, corporate intermediate/long-term debt, corporate short-term debt, broker/dealer equities, investment-company equities, other equities, and an "other" bucket, with series counts, total assets, and income distributions per row. Output supports fee studies, peer-portfolio benchmarking for litigation damages models, and sponsor-by-sponsor product-mix histories. Extraction requires SGML-aware parsing because HTML table parsers fail on the <S>/<C> markup.

5. Damages reconstruction and amendment-driven restatement detection

Litigation-support analysts join NSAR-U/A amendments to their originating NSAR-U filings by CIK and periodOfReport (the form body carries no pointer to the prior accession). Comparing item 127, item 131, items 118-126, and the credit-enhancement and default fields in items 128-130 between the original and the amendment exposes restated balances, corrected sales totals, and revised default disclosures. This drives expert reports, deposition exhibits, and reconstructed financial timelines in sponsor, trustee, and depositor disputes.

6. Insurance-product-wrapped UIT identification via item 117

Compliance teams and product strategists filter the population on item 117 — the separate-account-of-an-insurance-company indicator with sub-items for variable annuity, scheduled-premium variable life, flexible-premium variable life, and other 1933 Act products — to isolate insurance-dedicated UITs. The filtered subset feeds variable-product fee benchmarking, regulator inquiries into legacy variable-insurance trusts, and successor-trustee due diligence on books still in run-off.

7. Series-level inventory from item 132 for legacy-trust diligence

Securities counsel and compliance officers parse the item 132 roster of 811- file numbers in every filing for a given sponsor to enumerate every series ever included under that trust umbrella, with first-appearance and last-appearance dates derived from periodOfReport. The resulting series-lifecycle table supports succession opinions, indemnification scope reviews, and audit responses for trusts still administered after 2018 even though no new NSAR-U records are produced.

Dataset Access

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

This endpoint returns dataset-level metadata including the name, description, last updated timestamp, earliest sample date, total record count, total size, form types covered (NSAR-U and NSAR-U/A), the container format (ZIP), and the file types contained in each archive (TXT, JSON, HTML). It also provides the download URL for the entire dataset and the full list of individual container files, with per-container size, record count, updated timestamp, and download URL. This endpoint does not require an API key.

Use this endpoint to monitor which containers were modified in the most recent refresh run and to decide which monthly archives to re-download on a day-by-day basis.

Example response:

          1
          {
        

          2
            "datasetId": "1f13365b-9ae0-691b-9d42-2e0af6f449a8",
        

          3
            "datasetDownloadUrl": "https://api.sec-api.io/datasets/form-nsaru-files.zip",
        

          4
            "name": "Form NSAR-U Files Dataset",
        

          5
            "updatedAt": "2026-04-15T05:52:49.674Z",
        

          6
            "earliestSampleDate": "1995-03-01",
        

          7
            "totalRecords": 17710,
        

          8
            "totalSize": 86298667,
        

          9
            "formTypes": ["NSAR-U", "NSAR-U/A"],
        

          10
            "containerFormat": "ZIP",
        

          11
            "fileTypes": ["TXT", "JSON", "HTML"],
        

          12
            "containers": [
        

          13
              {
        

          14
                "downloadUrl": "https://api.sec-api.io/datasets/form-nsaru-files/2018/2018-05.zip",
        

          15
                "key": "2018/2018-05.zip",
        

          16
                "size": 412874,
        

          17
                "records": 84,
        

          18
                "updatedAt": "2026-04-15T05:52:49.674Z"
        

          19
              }
        

          20
            ]
        

          21
          }
        

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

Downloads the complete archive containing every Form NSAR-U and NSAR-U/A filing from March 1995 through June 2018 in a single ZIP file. This endpoint requires a valid SEC API key passed via the token query parameter.

Download Single Container: https://api.sec-api.io/datasets/form-nsaru-files/2018/2018-05.zip?token=YOUR_API_KEY

Containers are organized by year and month (YYYY/YYYY-MM.zip), with each ZIP holding the accession-level metadata file and original EDGAR submission documents (excluding images) for filings submitted in that month. Use this route to fetch a specific month instead of the full archive. This endpoint requires a valid SEC API key.

Frequently Asked Questions

What form does this dataset cover?

The dataset covers Form N-SAR — specifically the NSAR-U variant, the annual report filed by unit investment trusts under Section 30 of the Investment Company Act of 1940 and Sections 13 and 15(d) of the Securities Exchange Act of 1934. Both initial filings (NSAR-U) and amendments (NSAR-U/A) are included.

What does one record in this dataset represent?

One record is a single EDGAR accession — one complete NSAR-U or NSAR-U/A submission filed by a registered UIT. On disk, the record is a per-accession folder named for the 18-digit EDGAR accession number (dashes stripped), containing a metadata.json sidecar and the original text-bearing EDGAR documents.

Who was required to file Form NSAR-U?

Unit investment trusts registered under the Investment Company Act of 1940 were required to file NSAR-U annually under Rule 30b1-1. The legal filer was the registered investment company itself; the depositor, sponsor, or trustee executed and submitted the form on the trust's behalf. Mutual funds, closed-end funds, and ETFs are management investment companies rather than UITs and filed NSAR-A and NSAR-B instead.

What time period does the dataset cover?

The dataset spans March 1995 through June 1, 2018, when the SEC rescinded Form N-SAR under the Investment Company Reporting Modernization rulemaking (Release No. IC-32314) and replaced it with Form N-CEN. Only NSAR-U/A amendments to pre-rescission filings could appear afterward.

What file format is the dataset distributed in?

The dataset is distributed as ZIP containers organized by year and month (YYYY/YYYY-MM.zip). Inside each container, per-accession folders carry a JSON metadata sidecar plus the original EDGAR documents in TXT (the SGML-wrapped Form N-SAR primary document) and HTML (where EDGAR rendered an HTML index) formats. Image attachments and the EDGAR-generated "Complete submission text file" wrapper are not stored locally.

Does Form NSAR-U include XBRL data?

No. Form N-SAR never adopted structured-data reporting; it remained a fixed-width ASCII text file inside an SGML <DOCUMENT> envelope throughout its lifespan. In every record dataFiles[] is empty and linkToXbrl is blank. When the SEC modernized investment-company reporting it replaced N-SAR with the XML-based N-CEN rather than retrofitting structured data into N-SAR.

How does NSAR-U differ from NSAR-B and Form N-CEN?

NSAR-U is the UIT annual report under Rule 30b1-1; NSAR-B is the annual report by management investment companies under Rule 30a-1, with items on advisers, sub-advisers, directors, and 12b-1 plans that have no UIT analog. Form N-CEN is the post-June-2018 census-style XML successor that replaced the entire N-SAR family for both UITs and management companies; its data model is reorganized and does not align field-for-field with NSAR-U, so building a continuous UIT panel across the 2018 boundary requires a manual crosswalk.