Global stress, positioning and the world cycle
How the bundles under macro/global/stress/, macro/global/positioning/ and macro/global/cycle/ are built. Code: pipeline/tipsheet/compute/global_stress.py, compute/global_cycle.py and publish/global_backdrop.py; plug-in step steps/data_global.py.
Everything here is descriptive. A percentile tells you how unusual a reading is against its own history. It does not forecast anything.
Is something breaking? (stress)
| Bundle | What it shows | Source and licence |
|---|---|---|
stress/ofr_fsi_weekly, ofr_fsi_daily_2y | The OFR Financial Stress Index (0 = average stress) and its five category contributions (credit, equity valuation, safe assets, funding, volatility) and three regional ones | Office of Financial Research; public domain |
stress/cmdi_weekly | The New York Fed’s Corporate Bond Market Distress Index: market, investment grade and high yield, from 0 (calm) to 1 | NY Fed; reuse with attribution |
stress/funding_weekly, funding_daily_1y | SOFR, TGCR, BGCR, EFFR and OBFR; the Fed’s target range; stress spreads | NY Fed reference rates |
stress/ust_auctions_* | US Treasury 2-, 10- and 30-year auctions: bid-to-cover and bidder shares | US Treasury Fiscal Data; public domain |
Funding spreads
sofr_minus_target_top_bp: SOFR minus the top of the FOMC target range. A positive value means overnight repo cleared above the Fed’s ceiling, which signals scarce reserves. It peaked at +300 bp on 17 September 2019.sofr_p99_minus_p1_bp: the spread between the 99th and 1st percentile of SOFR trades. It widens when some borrowers pay up.sofr_minus_effr_bp: secured overnight rates against unsecured ones.
Auctions
- Bidder shares are of competitive bids accepted.
- A high primary-dealer share means dealers had to take more of the issue, which signals weak end-demand.
- Each auction’s bid-to-cover and dealer share is ranked against the previous five years of auctions of the same term (trailing only).
- TIPS and floating-rate notes are excluded.
- Tails can’t be computed, because Fiscal Data has no when-issued yield.
Checks (2026-10-02 run)
- The OFR categories sum to the headline within 0.002 on every day from 2000.
- The regions also sum to the headline, except on 2018-11-01 (a gap of 0.108, a source defect the Data bank has reviewed).
- The 2-, 10- and 30-year terms have 311, 279 and 219 auctions.
Who is positioned where? (CFTC)
positioning/cftc_weekly and positioning/cftc_latest: futures-only Commitments of Traders, weekly from June 2006.
Measure
- The speculative group is leveraged funds in the Traders in Financial Futures report: Treasuries, equity index futures and the dollar index.
- For commodities it is managed money in the disaggregated report: gold, silver, copper, WTI and natural gas.
- Net positioning is (long − short) as a percentage of open interest, which makes it comparable across years as contract sizes and open interest change.
- For the equity index futures the asset-manager net share is also shown. Asset managers are the structural longs that leveraged funds often hedge against, so a leveraged-fund short there is not necessarily a bearish bet.
pctile_3yranks the latest reading within the trailing 156 weeks;pctile_fullranks it within the whole history.
Timing: positions are as of Tuesday and published on Friday.
Missing: there is no CFTC report for the Indian rupee.
Where is the world in its cycle?
| Bundle | Measure |
|---|---|
cycle/cli_monthly, cli_phases | OECD composite leading indicators (amplitude adjusted, trend = 100) for India, the US, China, G7, G20, the euro area and nine others |
cycle/confidence_monthly | OECD business (BCI) and consumer (CCI) confidence |
cycle/gscpi_monthly | NY Fed Global Supply Chain Pressure Index, latest vintage |
cycle/fao_food_monthly | FAO Food Price Index and sub-indices, with % y/y |
cycle/commodity_*_monthly | World Bank Pink Sheet: indices (energy, food, fertilisers, metals, precious metals) and the prices that matter for India’s import bill: Brent, Dubai, Australian coal, LNG, urea, DAP, potash, palm, soy and sunflower oil, wheat, rice, sugar, gold, silver, copper, aluminium and iron ore |
cycle/india_ports_weekly | IMF PortWatch: 28-day cargo (imports and exports, tonnes; tanker and container imports) and port calls across 37 Indian ports, with year-on-year change |
cycle/chokepoints_weekly | Daily transits (7-day average) through Suez, Bab el-Mandeb, Hormuz and Malacca, against the 2019 average |
cycle/india_oil_monthly | JODI: India’s crude production, imports, refinery intake and product demand (thousand barrels a day) |
cycle/cofer_quarterly | IMF COFER: currency shares of allocated FX reserves |
Cycle phase follows OECD’s convention:
- above 100 and rising: expansion
- above 100 and falling: downturn
- below 100 and falling: slowdown
- below 100 and rising: recovery
The latest CLI months are revised most, so a phase flip in the last month or two is tentative.
Year-on-year changes compare with the exact same month (or the same 28 days) a year earlier. If that observation is missing, the change is missing too.
Caveats found while building
- Hormuz. Transits collapsed after 28 February 2026, when US and Israeli strikes on Iran led the major carriers to suspend transits. The 7-day average was about 4 a day in September 2026, against about 69 in 2019. This is a real event, not a data fault (IMF PortWatch’s own event page and Lloyd’s List report the same).
- Indian port calls rose by about half from April 2026 at nearly every port, far more than cargo tonnes did.
- Cause: we have not reconciled this with official port traffic (Indian Ports Association). It may be diversions and smaller vessels, or a change in PortWatch’s AIS model.
- The site leads with cargo tonnes and flags the call counts.
- JODI barrels for India run about four months behind.
import_share_of_intake_pctcompares imports with refinery intake in the same month, so stock changes can push a single month above 100%.
- GSCPI history is revised with every release; only the latest vintage is used.
Licences
| Source | Terms |
|---|---|
| OFR, CFTC, US Treasury | US government works, public domain |
| New York Fed | Reuse with attribution |
| OECD, FAO, World Bank | CC BY 4.0 |
| IMF (COFER, PortWatch) | Published with IMF attribution under the IMF’s terms |
| JODI | Free with attribution |
All bundles here hold either index values from official sources or prices under CC BY 4.0, so no third-party licensed levels are published. The licence gate is clean.