{
  "data_version": "seed-0.7 · marketcap-1 · 2026-08-30",
  "source": "https://finance.foundation",
  "license": {
    "finance_foundation_data": "CC BY 4.0 — attribution: Finance Foundation (https://finance.foundation)",
    "upstream_values": "Fields sourced from a provider carry their own source field and remain under that provider's terms — see https://finance.foundation/api/v1/sources"
  },
  "collection": "research",
  "count": 9,
  "data": [
    {
      "id": "open-financial-data-stack",
      "title": "The Open Financial Data Stack, 2026",
      "category": "Infrastructure",
      "date": "2026-08",
      "summary": "Where open financial data actually stands: what is genuinely open, what is licensed, and where the gaps are for builders and AI agents.",
      "body": [
        "Financial data looks abundant but is structurally closed. Prices, fundamentals, filings, and reference data are dominated by a handful of vendors — Bloomberg, LSEG (Refinitiv), S&P Global, FactSet, Morningstar — whose terminals and feeds are licensed per seat or per use. The result is a paradox: the numbers that describe public markets are mostly not public in any machine-usable way.",
        "The genuinely open layer is thinner than it appears. Regulatory filings (EDGAR in the US, similar registries elsewhere) are open. LEI — the global legal-entity identifier — is open by mandate. OpenFIGI provides open instrument identifiers. Central banks and statistical agencies publish macro series (FRED being the gold standard of usable open data). Crypto is the outlier: its market data, on-chain activity and reference data are radically open compared to traditional finance.",
        "The gap is in the middle: clean, linked, entity-level data. Anyone can download a filing; almost no one can cheaply answer 'show me all listed subsidiaries of this group, with their tickers, exchanges and reporting currencies.' Entity resolution — knowing that Apple Inc., AAPL, US0378331005 and LEI HWUPKR0MPOU8FGXBT394 are the same thing — is where open data still loses to the terminals.",
        "This is the layer Finance Foundation is built around: not another price feed, but the connective tissue — entities, identifiers, relationships and definitions — published in a form that people, applications and AI agents can consume directly."
      ]
    },
    {
      "id": "global-exchange-landscape",
      "title": "Mapping the Global Exchange Landscape",
      "category": "Markets",
      "date": "2026-08",
      "summary": "A structural map of the world's stock exchanges: who owns them, how they concentrate, and why listing venues and trading venues are diverging.",
      "body": [
        "The world's equity markets are more concentrated than most investors realize. The two US exchanges — NYSE and Nasdaq — account for roughly half of global listed market capitalization. Add Shanghai, Shenzhen, Tokyo, Hong Kong, Euronext, LSE and India's NSE, and nine venue groups cover the overwhelming majority of world equity value.",
        "Exchanges themselves have consolidated into listed, for-profit groups: ICE owns NYSE; Nasdaq Inc. is itself Nasdaq-listed; LSEG earns more from data (Refinitiv) than from listings; Deutsche Börse and Euronext roll up national venues across Europe. The exchange business quietly became a data and clearing business with a trading venue attached.",
        "Meanwhile listing and trading have diverged. A company lists in one place but trades everywhere: ADRs in New York, fungible lines in Hong Kong, systematic internalizers and dark pools in Europe, 24-hour tokenized wrappers emerging offshore. 'Where does this stock trade?' now has a dozen correct answers — which is exactly why machine-readable market identifiers (MIC codes) and listing relationships matter more every year.",
        "India's NSE is the structural story of the decade: the world's largest derivatives exchange by contracts, riding a domestic retail investing boom. The Gulf exchanges (Tadawul above all) are the capital-raising story, recycling energy surpluses into listed national champions."
      ]
    },
    {
      "id": "stablecoins-as-market-infrastructure",
      "title": "Stablecoins as Market Infrastructure",
      "category": "Crypto",
      "date": "2026-08",
      "summary": "Stablecoins stopped being a crypto niche and became a parallel dollar-settlement system. What that means for banks, payments and data.",
      "body": [
        "Stablecoins now settle trillions of dollars a year — volumes in the range of the major card networks — with a combined supply above a quarter of a trillion dollars, almost all of it USD-denominated. Functionally, they are offshore dollar accounts with instant, programmable settlement.",
        "Two design families dominate. Reserve-backed coins (USDT, USDC) hold treasury bills and repo against issued tokens — making their issuers, in aggregate, one of the largest holders of US Treasury bills in the world. The issuers' business model is simply the interest on those reserves, which is why stablecoins became one of the most profitable business models per employee in financial history.",
        "The infrastructure consequence is underrated: a stablecoin transfer is a full settlement — no correspondent chain, no cut-off times, no T+1. For cross-border payments, emerging-market savings and machine-to-machine (AI agent) payments, stablecoins are becoming the default rail, with regulation (MiCA in the EU, the GENIUS Act in the US) converting them from grey zone to licensed instrument.",
        "For a data layer, stablecoins are also the most transparent money in existence: supply, flows and holder distribution are on-chain, queryable in real time — a preview of what all financial reference data could look like."
      ]
    },
    {
      "id": "ai-agents-and-financial-data",
      "title": "AI Agents and Financial Data Access",
      "category": "AI",
      "date": "2026-08",
      "summary": "AI agents are becoming first-class consumers of financial data. Most financial websites are unreadable to them. What machine-readable finance looks like.",
      "body": [
        "The fastest-growing consumer of financial information is no longer a person — it is an AI agent doing research, monitoring portfolios, or executing workflows on someone's behalf. These agents don't read charts or navigate menus. They need structured entities, stable identifiers, clean JSON, and honest metadata about freshness and provenance.",
        "Most financial websites fail this test completely: data lives in rendered charts, paywalled terminals, or PDFs. The sites that win agent traffic are the ones that expose the same knowledge as clean endpoints and machine-readable indexes (llms.txt, OpenAPI schemas, MCP servers) — because an agent that can consume you becomes a distribution channel, not a scraper.",
        "The emerging stack is recognizable: a knowledge layer (entities and relationships), an API layer (typed, versioned, cache-friendly), and an agent interface layer (tool definitions the agent can call directly). Provenance is not optional — an agent that cannot tell how old a number is will propagate stale data into decisions.",
        "Finance Foundation is built agent-first on purpose: every page here has an API twin, the whole dataset is one GET away, and /llms.txt tells an agent exactly what exists and how to fetch it."
      ]
    },
    {
      "id": "entity-resolution-in-finance",
      "title": "The Entity Resolution Problem",
      "category": "Data",
      "date": "2026-08",
      "summary": "Why 'which company is this?' is the hardest cheap-sounding question in finance, and how identifiers (LEI, ISIN, FIGI, MIC) actually fit together.",
      "body": [
        "Every financial dataset eventually hits the same wall: the same real-world thing appears under different names, tickers, and codes. Apple is 'Apple Inc.', AAPL on Nasdaq, ISIN US0378331005, CUSIP 037833100, LEI HWUPKR0MPOU8FGXBT394 — and AAPL means something else entirely on another venue's options feed. Merging two datasets without resolving entities silently corrupts both.",
        "The identifier landscape is layered, not redundant. LEI identifies the legal entity (who). ISIN identifies the security (what). The MIC identifies the venue (where). Tickers are venue-local shorthand — convenient and dangerously ambiguous. FIGI, an open standard, identifies an instrument at venue level and is free to use, unlike CUSIP and SEDOL, which are licensed.",
        "The expensive insight of the data industry is that relationships matter more than attributes: issuer → security → listing → venue → jurisdiction. Once those edges exist, everything else (prices, filings, news) can hang off them. This is why 'knowledge graph' is not a buzzword in finance — it is literally the shape of the domain.",
        "A practical rule for builders: store every identifier you encounter, pick one internal id you control, and treat all external ids as claims with provenance, not truth. Identifier mappings drift — companies merge, relist and rename — so resolution is a process, not a table."
      ]
    },
    {
      "id": "financial-knowledge-graphs",
      "title": "Financial Knowledge Graphs: From Tables to Relationships",
      "category": "Infrastructure",
      "date": "2026-08",
      "summary": "Why the future of financial data is a graph, not a table — and how to build one incrementally without boiling the ocean.",
      "body": [
        "Tabular data answers 'what is the value of X'. A knowledge graph answers 'how is X connected to everything else' — which is the question analysts, risk managers and AI agents actually ask. Contagion, supply chains, ownership webs, fund exposures: all of these are graph queries wearing spreadsheet costumes.",
        "A financial knowledge graph does not need exotic technology to start. It needs typed entities (company, security, market, fund, country), typed edges (lists_on, domiciled_in, managed_by, issues, regulates) and stable identifiers on the nodes. A relational database with a well-designed relationships table is a knowledge graph in the way that matters.",
        "The compounding effect is what makes graphs worth it: each new entity type multiplies the value of existing ones. Add funds to a company graph and you can suddenly answer exposure questions; add countries and you get jurisdiction risk; add banks and you get counterparty maps. The marginal dataset gets cheaper to add and more valuable, the opposite of siloed tables.",
        "For AI, the graph is the difference between retrieval and reasoning. A language model given linked entities can traverse — 'this fund is managed by this company, listed on this exchange, in this jurisdiction, whose central bank just moved rates' — instead of pattern-matching over disconnected text. Structured context is the highest-leverage input an agent can be given."
      ]
    },
    {
      "id": "the-concentration-graph",
      "title": "The Concentration Graph",
      "category": "Data",
      "date": "2026-08",
      "summary": "What our own 237-node map shows: the covered universe is one country, ten companies and seven names deep — and even a supermajor can be unresolvable in open identifier registries.",
      "body": [
        "This note is computed from Finance Foundation's own dataset (seed-0.4): 121 systemically significant companies worth roughly $47tn together, mapped to markets, indices, funds and jurisdictions. The covered set is curated by systemic importance, so every figure below describes this universe, not the whole world — but the set was chosen to represent where the world's market value actually sits, and the shape it reveals is hard to dismiss.",
        "Concentration by country: the United States accounts for 45% of covered companies but 72% of covered market value. Concentration by company: the top ten names — NVIDIA, Microsoft, Apple, Alphabet, Amazon, Meta, Saudi Aramco, Broadcom, Tesla, Berkshire Hathaway — are 46% of the total. A map of 'global' equity value is, to a first approximation, a map of about ten balance sheets.",
        "The fund layer makes the concentration self-reinforcing. Of the ten equity funds in our graph with holdings data, eight list the same seven companies — NVIDIA, Microsoft, Apple, Amazon, Meta, Alphabet, Tesla — among their top holdings; Broadcom appears in seven. An investor diversifying across an S&P 500 tracker, a world index fund and a Nasdaq fund holds the same seven stories three times. In graph terms: the 'holds' edges from different funds converge on the same handful of nodes.",
        "The index layer shows the same convergence from another angle: 25 covered companies sit in two major indices at once (S&P 500 + Nasdaq-100 in the US, DAX + EURO STOXX 50 in Europe, Hang Seng + CSI 300 for the Chinese dual-listings). Index membership multiplies passive demand for exactly the names that are already the largest — the mechanism behind the fund-overlap pattern above.",
        "One more finding surfaced by building the identifier layer rather than by analyzing it: even at this size, open entity resolution has holes. We could verify LEIs for 119 of 121 companies via GLEIF — but Chevron, a $300bn supermajor, and Kweichow Moutai, long China's most valuable listed company, could not be confidently resolved by legal name, country or ISIN mapping. If the connective tissue of open financial data fails on names this large, the long tail is far worse — which is, in one sentence, why this site exists."
      ]
    },
    {
      "id": "the-shape-of-live-data",
      "title": "The Shape of \"Live\" Data",
      "category": "Data",
      "date": "2026-09",
      "summary": "Computed from Finance Foundation's own /api/v1/live feed: the 29 series behind one 'live' label range from same-day crypto prices to a policy rate last set in March 2024, because 'live' means when we last checked, not when the number last changed.",
      "body": [
        "Finance Foundation's /api/v1/live endpoint serves 29 series — cryptocurrency prices, FX crosses, sovereign bond yields, central-bank policy rates, and macro indicators like GDP growth and unemployment — under a single 'live' label, polled on a fixed three-hour cycle. But each series also carries its own value date, its own retrieved_at timestamp, its own source and provider, and its own from_fallback flag, and reading across all 29 at once shows that a fixed polling cadence has almost nothing to do with how current the underlying number actually is.",
        "The freshest tier is priced continuously: all nine crypto series (bitcoin, ether, BNB, XRP, Cardano's ADA, Solana, and the stablecoins DAI, USDC and USDT) carry a value dated the same calendar day they were retrieved. Nine FX crosses sourced from the ECB's daily reference rates, plus the US 10-year Treasury yield from the Treasury's daily par curve, sit one tier back — dated the prior business day, because both publishers fix a rate once per day rather than continuously. Already, at this first split, 'live' quietly covers two different things: a continuously-traded price versus a once-daily official fixing.",
        "Below that sits the statistical-release tier, where the lag runs to weeks and months by construction, not by any fault of the feed: at one reading behind this note, Taiwan's M2 growth was dated eight days before retrieval, euro-area M3 growth and US CPI and unemployment a little over a month behind, euro-area unemployment nearly two months behind. The dataset dates each series to its actual reference period rather than backfilling a fresher-looking date. GDP growth for the US and the euro area, and the US current account, are quarterly — dated to Q1 or Q2 2026 at that same reading, a full quarter or two old by the time anyone reads them, because a quarterly figure cannot exist before its quarter closes.",
        "Policy rates are the extreme case, and the dataset documents why directly in the series' own note: 'the date is when the rate last changed, not when we read it — a policy rate is a level that holds until the bank moves it.' The ECB's main refinancing rate is dated to its last decision; Taiwan's discount rate was dated 2024-03-22 at that reading, over two years earlier, because the Central Bank of the Republic of China had not moved it since and — sitting outside the World Bank, IMF and OECD statistical systems — publishes it only as a figure on its own web page rather than through a data API. Polling that series every three hours for two years returns the identical number every time; the three-hour cadence describes how often the site checks, not how often the value changes.",
        "Provenance is layered the same way freshness is. At the reading behind this note, all nine crypto series were served through their listed fallback provider (Coinbase spot, plus Bybit for BNB) rather than the CoinGecko source named on the record, while every FX, yield and macro series came through its primary source undisturbed — a reminder that 'live' also hides which upstream actually answered. Collapsing a same-day market price, a month-old statistical release and a rate unchanged since 2024 into one undifferentiated 'live' timestamp would misrepresent all three. Carrying date, retrieved_at, source and from_fallback per series, instead of one badge for the whole feed, is what keeps 'live' honest at 29 different speeds."
      ]
    },
    {
      "id": "bank-balance-sheets-vs-gdp",
      "title": "Balance Sheets Bigger Than Nations",
      "category": "Data",
      "date": "2026-10",
      "summary": "Computed from Finance Foundation's own banks and countries tables: summing covered bank total assets by headquarters country and dividing by that country's GDP shows five of eleven banking jurisdictions carry bank balance sheets larger than their whole economy.",
      "body": [
        "This note joins two tables already in the dataset: 20 banks, each carrying a total_assets_usd_tn figure, and 37 countries, each carrying a gdp_usd_tn figure. Grouping the banks by their country field and summing total assets, then dividing by that country's GDP, turns two unrelated-looking reference tables into a systemic-size ratio for each of the 11 countries that have at least one covered bank. Seventeen of the 20 banks are on the Financial Stability Board's global systemically important bank (G-SIB) list; the other three — DBS, Rabobank and Intesa Sanpaolo — are domestically systemic rather than globally so. Either way, this is a curated set of large, named banks, not a census of every bank chartered in each country.",
        "Five countries clear a ratio of 1.0. France tops the table: its three covered banks — BNP Paribas, Crédit Agricole and Société Générale — carry a combined $7.5tn in assets against a $3.2tn economy, a ratio of about 2.3. Switzerland is next at roughly 1.8, and on a single name: UBS alone, at $1.7tn, against Swiss GDP of $0.94tn — the exact shape of the concern Swiss regulators raised when UBS absorbed Credit Suisse in 2023 and became, on its own, a balance sheet bigger than the country that charters it. Spain (Banco Santander, 1.11), Singapore (DBS, 1.09) and the United Kingdom (HSBC plus Standard Chartered, 1.07) round out the group of five.",
        "A ratio above 1 is not by itself a solvency statement — it mostly reflects what these particular banks are. HSBC, Santander, BNP Paribas and UBS are international banks whose balance sheets are built from deposits, loans and trading books gathered far outside their home market, not a claim sized to the domestic economy alone. But the ratio is exactly the quantity regulators and the IMF reach for when asking a harder question: if this balance sheet needed a sovereign backstop, could the sovereign behind it actually provide one? A central bank can print the funding currency; a government's taxable base cannot be resized to match a bank's book. That gap — not the balance sheet's quality — is what the France and Switzerland rows are actually showing.",
        "The ratio also cuts the other way, and the largest absolute numbers in the dataset sit in the lowest ratios for exactly that reason. The United States' five covered banks sum to $13tn against a $29tn economy (0.45); China's three sum to $16.9tn against $19tn (0.89, the second-highest absolute exposure in the table despite a ratio under 1). Germany sits at the opposite end from France at 0.32 — one covered bank, Deutsche Bank, against a $4.7tn economy — a reminder that Germany's own banking system (regional Landesbanken and savings banks, none of them in this dataset) is far larger than its single G-SIB entry implies.",
        "That last point is the honest limit of this note: the ratio is a property of which banks happen to be curated here, not a complete national balance-sheet-to-GDP figure for any of the 11 countries, and the 26 countries in the dataset with zero covered banks are uncovered, not bankless. What the computation does show cleanly is why 'bank assets ÷ GDP' is a ratio worth being able to compute at all — it is one of the plainest early-warning numbers in financial stability, built from two fields that are each individually mundane, and it only exists here because total_assets_usd_tn and gdp_usd_tn live in the same queryable graph rather than two different providers' spreadsheets."
      ]
    }
  ]
}