The core datasets that drive asset management decisions

# The core datasets that drive asset management decisions

A single wrong character in an ISIN once forced a European fund to restate its Net Asset Value (NAV) and refund investors who had traded at the wrong price. The ISIN (International Securities Identification Number, the 12-character global code for a security) had been keyed with a transposed digit, so the system priced the position against a different bond. Nobody noticed until reconciliation the next morning. By then, thousands of shares had changed hands at a stale price.

This is why asset management is, underneath the glossy performance charts, a data discipline. Portfolios are only as trustworthy as the five feeds that describe them. Let us walk through each one.

The five feeds that describe a portfolio

Every portfolio decision, valuation, and client report is assembled from five interlocking datasets. Get one wrong and the error propagates through everything downstream.

1. Security master

The security master is the reference dictionary. It holds the static and semi-static facts about every instrument: identifiers (ISIN, CUSIP, SEDOL, Bloomberg FIGI, ticker), issuer name, currency, asset class, maturity date, coupon, country of risk, and sector classification.

Think of it as the "who's who" of instruments. When a portfolio system says it holds "US0378331005", the security master translates that into "Apple Inc common stock, USD, US equity."

Concrete failure mode: two vendors classify the same bond differently, one as "corporate" and one as "government-related." Your sector exposure report now disagrees with itself depending on which feed loaded last.

• CUSIP: 9-character identifier for US and Canadian securities.
• SEDOL: 7-character identifier used in the UK and Ireland.
• FIGI: Financial Instrument Global Identifier, an open identifier maintained by Bloomberg and freely available.

Because CUSIP and ISIN carry licensing costs, many firms mapmapUsing software to automate repetitive marketing tasks and campaigns, enabling personalisation at scale across channels like email, web, and social.Voir la définition complète → everything to the open FIGI standard as a neutral internal key.

2. Holdings (positions)

Holdings tell you what you own and how much: quantity, cost basis, trade date, settlement status, and the account or fund the position sits in.

Holdings look simple but hide traps. A common one is the difference between trade-date and settle-date positions. If you bought 10,000 shares today but they settle in two business days (T+2, though US equities moved to T+1 in May 2024), your trade-date holdings and your custodian's settle-date holdings will not match until settlement. Reconciliation must know which basis it is comparing.

3. Pricing

Pricing supplies the market value for each holding. Sources vary by asset class:

• Equities: exchange closing prices, usually straightforward.
• Bonds: often no last trade, so prices come from evaluated pricing vendors (ICE Data Services, Bloomberg BVAL, and others) that model a fair value.
• Private and illiquid assets: periodic manual or model-based marks.

The subtlety: bond "prices" are frequently opinions, not observed trades. Two vendors can quote the same bond several cents apart, which moves NAV.

4. Corporate actions

Corporate actions are events that change a security or its holders' entitlements: dividends, stock splits, mergers, spin-offs, rights issues, coupon payments, and name or identifier changes.

These are the single most error-prone feed because they change the security master AND the holdings AND the pricing at once, often overnight.

Example: a 2-for-1 stock split doubles your share count and halves the price. If the split is applied to quantity but not to your cost basis or your benchmark, your performance and exposure numbers break for that day. Mishandled corporate actions are a leading cause of NAV restatements industry-wide.

5. Benchmark constituents

A benchmark (for example the S&P 500 or the Bloomberg Global Aggregate bond index) is itself a dataset: the list of constituents, their weights, and their identifiers, refreshed on a schedule.

Index providers (MSCI, S&P Dow Jones Indices, FTSE Russell, Bloomberg Index Services) license these files. If your portfolio's ISINs do not match the benchmark's ISINs exactly, your active weight and tracking error calculations are garbage even though every individual number looks plausible.

How one mislabeled ISIN cascades

Follow the chain. Suppose a position is loaded with ISIN US0378331005 but mistyped as US0378331050.

1. Security master: the wrong ISIN either fails to resolve (position shows as "unknown") or resolves to a different instrument.

2. Pricing: the pricing feed keys on ISIN, so it either returns no price (position valued at zero or stale) or the wrong price.

3. NAV: fund NAV per share = (total assets, liabilities) / shares outstanding. A mispriced or zero-valued holding pushes NAV off.

4. Client reporting and dealing: investors buy or sell at that NAV. Trades executed at a wrong NAV may need to be unwound and compensated.

5. Benchmark comparison: the position no longer matches any benchmark constituent, inflating apparent active risk.

One character, five broken systems. This is why identifier integrity is treated as a control, not a convenience.

A simple NAV worked example

Illustrative, not real fund data:

Holdings:
  Apple    10,000 shares  @ $220.00  = $2,200,000
  Bond X    5,000 units   @ $ 98.50  = $  492,500
  Cash                                 $  307,500
  ---------------------------------------------
  Total assets                         $3,000,000
  Liabilities (fees accrued)          -$   10,000
  Net assets                           $2,990,000

Shares outstanding: 100,000
NAV per share = 2,990,000 / 100,000 = $29.90

Now mislabel the bond so it prices at zero:

  Bond X priced at $0 -> assets drop by $492,500
  Net assets = $2,497,500
  NAV per share = $24.975  ~ a 16.5% error

A 16.5 percent NAV error from one bad identifier. That is the concrete cost of poor reference data.

🎬 [VIDEO: "What is a Security Master?" - youtube.com/results?search_query=security+master+data+management - a short primer on why reference data sits at the center of the investment data stack]

Where these feeds come from

The market for this data has a real balance of power worth knowing.

• Reference and pricing data: Bloomberg, LSEG (London Stock Exchange Group, which owns Refinitiv), FactSet, and ICE Data Services dominate. These are estimates of market position, not exact shares, and they shift.
• Index data: MSCI, S&P Dow Jones Indices, and FTSE Russell are the heavyweights in equities; Bloomberg is dominant in fixed income indices after acquiring the Barclays index business.
• Corporate actions: sourced from the same major vendors plus custodian banks (State Street, BNY, Northern Trust, JPMorgan) that hold the assets and process events.

Regulatory context you should name correctly: in Europe, MiFID II (Markets in Financial Instruments Directive II) drove demand for accurate instrument identification and reporting. In the US, the SEC (Securities and Exchange Commission) sets rules on fund NAV accuracy and pricing. Both regimes make clean reference data a compliance obligation, not just good hygiene.

Reconciliation: the daily proof

None of the above matters unless it is checked. Reconciliation is the daily process of comparing your internal records against an independent source, usually the custodian bank that actually holds the securities.

Three reconciliations run in most shops:

• Position reconciliation: your holdings vs custodian holdings, by identifier and quantity.
• Price reconciliation: your prices vs an independent price source, with tolerance thresholds (for example, flag any equity price differing by more than 0.5 percent).
• Cash reconciliation: your expected cash vs the bank balance.

A break is any mismatch. Firms track break counts, break aging (how long a break stays open), and match rates as core quality metrics. A healthy equity book might target a same-day match rate above 99 percent; illiquid books tolerate more.

Simple match-rate calculation:

Positions matched cleanly: 4,955
Total positions:           5,000
Match rate = 4,955 / 5,000 = 99.1%
Breaks to investigate: 45

That 0.9 percent is where operations teams spend their morning, because any one of those 45 breaks could be the next NAV restatement.

Key Takeaways

• Five feeds underpin every portfolio: security master, holdings, pricing, corporate actions, and benchmark constituents. They interlock, so an error in one contaminates the rest.
• Identifiers are controls. ISIN, CUSIP, SEDOL, and the open FIGI standard are the join keys that let feeds talk to each other; one wrong character can produce double-digit NAV errors, as the worked example showed.