Accounting firms receive brokerage statements from dozens or hundreds of clients each quarter, and again in bulk at tax season. Each statement runs several pages, uses a different layout depending on the custodian, and mixes account summaries, transaction tables, income detail, and fee schedules in ways that resist standard data extraction.
This article explains what a brokerage statement contains, why extracting data from one is harder than it looks, and how firms automate the conversion from PDF to structured data without writing code.
What Is a Brokerage Statement?
A brokerage statement is a periodic report issued by a financial institution summarizing the activity and current state of an investment account. Statements are typically issued monthly for active accounts and quarterly for less active ones. Accounting firms receive them from clients as source documents for financial review, tax preparation, and portfolio reconciliation.
Most brokerage statements share a common structure, though the exact layout, terminology, and level of detail vary considerably across custodians:
Account summary. The opening section shows the account value at the start and end of the period, net change, and a breakdown by asset class or account type.
Holdings. A table of current positions listing each security, quantity held, cost basis (where the broker maintains it), current market price, and market value. This is often the largest section of the statement.
Transaction history. A chronological log of all activity during the period: buys, sells, dividends received, interest credited, transfers, and fee charges. Each line typically includes the date, security name, ticker or CUSIP, quantity, price, and net amount.
Income summary. Aggregated income from dividends, interest, and capital gain distributions received during the period. Some statements break this out by security. Others show only totals.
Realized gains and losses. For taxable accounts, many brokers include a realized gains section showing proceeds, cost basis, holding period, and short-term or long-term gain or loss for each security sold.
Fees and charges. Advisory fees, management fees, transaction costs, and any other charges applied to the account during the period.
Cash and money market. The opening and closing cash balance, sweep account details, and any money market fund positions held as the cash equivalent.
How a Brokerage Statement Differs From Related Documents
Accounting teams work with several types of investment documents that look similar but serve different purposes:
A trade confirmation is issued immediately after a single order executes. It confirms the security, quantity, price, commission, and settlement date for that specific transaction. It is a point-in-time document, not a summary.
A 1099 composite is a tax document issued annually summarizing reportable income and proceeds. It is derived from the brokerage statement data but follows IRS format requirements and includes only reportable items.
A custodian report or portfolio appraisal is a valuation snapshot at a specific date, typically produced on demand. It shows holdings and market values but does not include transaction history or income detail.
Brokerage statements are the most comprehensive of the three and the most commonly processed for ongoing accounting work.
Why Brokerage Statement Data Extraction Is Difficult
Brokerage statements present extraction challenges that simpler documents do not:
No single standard layout. Each custodian and brokerage firm designs its own statement format. Fidelity, Schwab, Vanguard, and smaller custodians all produce statements that look different, use different column headers, and organize sections in different orders.
Multi-page tables. The holdings table or transaction history often runs across multiple pages, with column headers that may not repeat on every page. Page breaks can fall in the middle of a table row.
Mixed content types. A single statement combines summary paragraphs, summary tables, detailed transaction tables, footnotes, and legal disclosures in a document that can run 20 or more pages for an active account.
Scanned and image-based PDFs. Statements downloaded from a portal are usually text-based PDFs. Statements that arrive as paper and are scanned by a client are image PDFs that require optical character recognition before any data can be read.
Layout changes. Custodians periodically update their statement designs. A format that worked reliably for months can shift after a broker redesigns its template, requiring review of the extraction configuration.
Ticker symbols and financial codes. Brokerage statements use CUSIP numbers, ticker symbols, and asset class codes that may not appear in recognizable formats. Extraction needs to capture these codes accurately because they drive downstream reconciliation.
What Data Can You Extract From a Brokerage Statement?
The fields that accounting firms most commonly extract from brokerage statements include:
From the account summary: account number, account owner name, account type (individual, joint, IRA, trust), statement period start and end dates, opening value, closing value, net change.
From the holdings table: security name, ticker symbol, CUSIP, asset class, quantity, cost basis per share (where available), current price, market value, unrealized gain or loss, percentage of portfolio.
From the transaction history: transaction date, settlement date, transaction type (buy, sell, dividend, interest, fee), security name, ticker, quantity, price per unit, gross amount, commission or fee, net amount.
From the income summary: dividend income by security, interest income, capital gain distributions, total income for the period.
From the realized gains section: security name, acquisition date, sale date, holding period (short-term or long-term), cost basis, proceeds, gain or loss.
From fees: advisory fee amount, management fee, fee basis (percentage or flat), billing period.
How Accounting Firms Typically Handle Brokerage Statements
Most firms use one of four approaches, each with different trade-offs on accuracy, speed, and setup effort:
| Method | Speed | Accuracy | Setup | Scales? |
|---|---|---|---|---|
| Manual keying | Slow | Human error | None | No |
| PDF-to-Excel converter | Fast | Layout-dependent | Minimal | Partly |
| Template-based OCR | Medium | High for known formats | Per-custodian setup | Partly |
| AI document extraction (Parseur) | Fast | High across layouts | One-time configuration | Yes |
Manual keying is the baseline most firms start with. A staff member opens the PDF, reads each value, and types it into a spreadsheet or accounting system. For high-volume tax seasons this approach does not scale, and keying errors compound across hundreds of documents.
PDF-to-Excel converters extract all text from a PDF into a spreadsheet. They work well for simple, text-based statements but produce unpredictable output for complex layouts with merged cells, multi-page tables, and footnotes mixed into table rows. The result is raw text that still requires manual cleanup before it can be used.
Template-based OCR tools use visual templates tied to specific document layouts. They perform well once a template is built and the format is stable. The limitation is that each custodian requires its own template, and any layout change from the broker breaks the template.
AI document extraction reads the document semantically rather than by position. It learns what a field means across different layouts without requiring a new template for each custodian. It handles multi-page tables, scanned PDFs, and layout variations more reliably than position-based tools.
How Parseur Extracts Data From Brokerage Statements
Parseur uses an AI Vision engine to read PDFs and an AI Text engine for text-based documents including spreadsheet exports and email-delivered statements. The extraction workflow has four steps:
Step 1: Create a mailbox or upload point. Each client, custodian, or document type gets its own Parseur mailbox. Statements forwarded to the mailbox email address are ingested automatically. You can also upload files via drag-and-drop or API.
Step 2: Define the fields to extract. Describe the fields you need in plain language. The AI engine reads the document and identifies where each field appears, handling variation across pages and layouts without positional templates.
Step 3: Review the extracted data. Parseur's review interface shows the extracted values alongside the source document. Flag exceptions, add validation rules for numeric ranges or required fields, and mark documents that need a second look before they export.
Step 4: Export to your system. Extracted data exports to Google Sheets, Excel, your accounting software, or any downstream platform via webhook, Zapier, Make, Power Automate, or API. Each transaction row in the statement becomes a row in the output, ready for import.
Seven Checks to Run on Extracted Brokerage Data
Automation reduces manual work, but extracted financial data still requires a review step before it enters accounting records. These seven checks catch the most common issues:
1. Period totals reconcile. The sum of extracted transaction amounts should match the account's net change for the period. If it does not, a transaction was missed, duplicated, or misread.
2. Holdings count matches. The number of position rows extracted should equal the number of positions on the statement. Missing rows indicate a table that was not fully captured.
3. Ticker and CUSIP integrity. Verify that tickers and CUSIPs are correctly separated from security names. These codes drive downstream reconciliation. A merged or truncated value causes lookup failures.
4. Transaction type classification. Confirm that buys, sells, dividends, and fees are classified correctly. Misclassified transaction types affect income reporting and cost basis tracking.
5. Date format consistency. Statements from different custodians use different date formats. Confirm that all dates exported to your system follow a consistent format before import.
6. Negative signs on credits and debits. Check that debits carry the correct sign. Some statements show all amounts as positive with a separate debit or credit indicator column. Extraction should preserve the sign convention your accounting system expects.
7. Pages processed count. Confirm that Parseur processed all pages of a multi-page statement. For long statements, verify the page count in the extracted output against the source PDF.
Brokerage Statement Terminology and Codes
Accounting teams working with extracted brokerage data encounter several terms and codes that are worth understanding:
CUSIP. A nine-character alphanumeric identifier assigned to each US and Canadian security. CUSIPs are the primary key for matching holdings across custodians and in accounting systems.
ISIN. The International Securities Identification Number, a twelve-character code used in international markets. Some custodians report both CUSIP and ISIN for the same security.
Lot. A specific purchase of a security at a specific date and price. Cost basis tracking requires lot-level data because different lots of the same security may have different holding periods and tax treatment.
Short-term vs long-term. The holding period determines whether a gain or loss is taxed at ordinary income rates (held one year or less) or long-term capital gains rates (held more than one year). Extraction should preserve the classification from the statement.
Wash sale. A sale of a security at a loss followed by a purchase of the same or substantially identical security within 30 days before or after. Brokers are required to report wash sale adjustments on 1099 forms. The brokerage statement may or may not flag individual transactions as wash sales.
Sweep account. The cash holding vehicle in a brokerage account, typically a money market fund. Cash that is not invested sits in the sweep. Statement extraction should distinguish sweep balances from invested positions.
Advisory fee vs transaction fee. Advisory fees are charged by the investment advisor as a percentage of assets under management. Transaction fees are commissions charged per trade. Both appear in the fees section of the statement but are categorized differently for accounting purposes.
Security and Compliance Considerations
Brokerage statements contain non-public personal financial information. Firms processing client statements need to address data security at every step of the workflow.
FTC Safeguards Rule. The FTC Safeguards Rule requires financial institutions, including accounting firms that prepare tax returns or provide financial planning services, to implement a written information security program. Any tool in the document processing workflow, including the extraction tool, is part of that program and should be evaluated for access controls, encryption, and incident response.
IRS Written Information Security Plan (WISP). The IRS requires tax preparers to maintain a WISP documenting how client data is protected. Document extraction tools that handle client financial data should be included in the WISP scope.
Data minimization. Extract only the fields your workflow requires. Avoid storing full statement PDFs longer than necessary. Define retention policies for both the source documents and the extracted data.
Parseur security. Parseur is SOC 2 Type II compliant and GDPR-compliant. Data in transit uses TLS encryption. Data at rest is encrypted. Access controls allow you to restrict which users can view or export data from each mailbox. A Data Processing Agreement is available for firms that require one.
What Parseur Does Not Do
Being clear about scope prevents workflow design errors:
Parseur extracts data from documents. It does not calculate portfolio performance, compute cost basis, determine tax liability, or identify wash sales. Those calculations happen in your accounting, tax, or portfolio management software using the data Parseur extracts.
Parseur does not connect directly to brokerage or custodian APIs. It processes documents: PDFs, emails, and spreadsheet files. If a custodian offers a data feed, that is a separate integration from document extraction.
Parseur does not replace a review step for financial data. Extraction is accurate, but accounting work that carries regulatory or legal responsibility requires human review before the data enters official records.
Parseur does not provide investment advice or portfolio analysis. It is a document processing tool, not a financial platform.
Getting Started
Brokerage statement extraction with Parseur starts with a single custodian and a batch of sample statements. Define the fields that matter most to your workflow, review the extracted output against the source documents, then expand to additional custodians and higher volume. Most firms have a working extraction pipeline within a day of starting.
The volume advantage becomes clear at tax season. Instead of staff manually keying data from hundreds of client statements, documents are forwarded to Parseur as they arrive and the extracted data is available for review without the bottleneck.
Last updated on
