Markets

1The Question

The Balance Sheet/Income Statement work answered the question of "what does the world hold." Layer 3 answers the question of "how much trading turns over in world markets during the year" - a completely different order of magnitude (trillions of dollars daily), and completely different sources (not SNA/OECD, but market-statistics institutions like BIS/WFE/SIFMA/ISDA/Swiss Re).

2What It Is Built From

3Method

• Each source is taken in its OWN natural unit, not forced into a common mold. FX and Interest Rate Derivatives are published as a DAILY average in the BIS Triennial Survey - these are converted to an annual equivalent using a 250-business-day assumption (the standard financial-market calendar). Equities, Bonds, CDS, and Insurance are already published as ANNUAL flow at their own sources.
• Repo was NOT annualized because it's a DIFFERENT type of metric from the others. The figure OFR provides isn't a "daily new transaction volume," but the average OUTSTANDING POSITION (exposure) for that day - since repo is mostly overnight, the stock and the daily new-transaction volume become intertwined. That's why Repo is shown separately and excluded from the total.
• The OTC + exchange-traded split is preserved, then summed. For FX and Interest Rate Derivatives, BIS provides two separate sources (Triennial Survey = OTC, Quarterly Review = exchange-traded) - the two are summed to get the category total, and shown separately in the item breakdown.
• Crypto is annualized with 365 days, not 250. Crypto exchanges trade nonstop 24/7 including weekends - so using 365 days gives a more accurate annual total.
• Equity Derivatives are based on PREMIUM TURNOVER, not notional turnover. For options, "notional" (number of contracts × index value × multiplier) is a leveraged/theoretical figure that greatly overstates the money actually paid - the same problem as the misleading "$600T notional" headline figure for OTC derivatives. That's why NSE's (India) own reported PREMIUM TURNOVER was used; for futures there's already a single "turnover" concept (price × lot × contract).
• Hedge Funds, like Repo, are shown separately as a quarter-end OPEN POSITION. OFR's Gross Notional Exposure (GNE) figure is the sum of the absolute values of long+short positions - a leveraged/gross exposure, not an actual cash flow. That's why it wasn't annualized or included in the total.

4What Came Out

5How It Was Tested

• A pre-existing claim from the site itself was verified for the first time. The narrative text on the Layer 3 Overview page (formerly part of the Layer 2 Overview) had stated that "the total of transactions occurring in the world financial system is on the order of roughly 40 T USD daily" - a claim not tied to any source, likely a rough estimate. The FX+Interest Rate Derivatives daily total ($34,99T) came out at the SAME ORDER OF MAGNITUDE as that claim.
• The CDS total reconciles to the penny with its own internal components. ISDA's report states Index CDS ($39,0T) + Single-Name CDS ($2,8T) = $41,8T - the summed sub-items also reach exactly this total.
• Treating Repo differently on purpose is a deliberate methodology decision, not an error. In the first draft, Repo was also annualized ×250 like the others - but this created a double-counting risk by presenting the STOCK of short-term repo positions as if it were a FLOW.

6Where the limits are

7Historical series (Historical Development page)

This page builds a snapshot of 2025 (about 9,700 T$ a year). A separate Historical Development page (/layer3/tarihsel) tracks annual market flow / turnover from 1900 CE to today. Before 1900 this measure did not exist: organized securities and derivatives markets are largely a 19th-20th century phenomenon, so earlier points are empty (no concept). The 1900-2000 figures are rough estimates (low confidence) - to be updated when a better long-run series (BIS / WFE historical) is found. Today's figure is Layer 3's own total (§2-§6).

Derived series (same build, no new sourcing): annual turnover per person, the flow relative to annual world GDP, the flow relative to the same year's physical material output (thousand $/tonne - decoupling from physical production, from 1900), and a growth decomposition + rate. Population and GDP are shared with the Layer 1 historical series. Full method: docs/layer_history_methodology.md §6 in the project repository.

8Perspective page, Leverage and Physical Contact

The Leverage and Physical Contact page (/layer3/detay/kaldirac-temas) in the "Perspective" group reads the 20 market categories from §2-§6 on two new axes: the degree of contact with physical Layer 0 (ownership transfer / liquid exchange / derivative / purely financial) and the typical leverage range. Volumes are aligned to layer3_market_flows.json; the two new axes are a separate compilation (layer3_leverage_contact.json).

9Perspective page, Speed and Latency

The Speed and Latency page (/layer3/detay/hiz-ve-gecikme) added to the "Perspective" group reads the abstraction of value to the speed of light through three measures: the New York-Chicago latency hierarchy, HFT's share of volume per market, and the ratio of traded paper barrels of oil to physical production. A separate dataset outside the main pipeline (layer3_latency.json).

10Perspective page, The Feedback Loop (cross-layer)

The Feedback Loop page (/layer3/detay/geri-besleme) compiles evidence for the coupling between Layer 3 and the physical base in both directions: top-down (L3 decisions → the real world) and bottom-up (physical shocks → the markets). A separate cross-layer dataset (layer_feedback_loop.json). For some channels the direction or size of the effect is contested in the peer-reviewed literature; those rows are given with both views.

11Perspective page, Horizontal Cross-Section (cross-layer matrix)

The Horizontal Cross-Section page (/buyuk-resim/yatay-kesit) places the four layers side by side across six dimensions: workforce, land footprint, energy, time scale, fragility, ontological status. A separate cross-layer dataset (layer_horizontal_matrix.json). The workforce and latency rows reuse existing sourced datasets (layer_workforce.json, layer3_latency.json); land and energy are compiled from primary sources. Each cell is marked “measured / estimate / qualitative”.