Index Breakdown and Methodology
Argentina Corporate Credit Spreads — Index Constituents
Daily constituents (holdings) of the Argentine corporate USD z-spread indices: one row per bond per day, for every bond that enters the index calculation, with the index weight, the quotation line used and the day's metrics. Key columns:
symbol— ticker of the quotation line actually used that day. Falls back to the ISIN (alphacast_id) for the ~1% of bonds without a match in the ticker master.linea— line used:D(MEP dollars) /O(pesos) /C(cable).weight_general_pct/weight_sector_pct— the bond's weight (%) in the General index and in its sector index. Market-value weighted with the index caps (40% per bond, 50% per issuer); each sums to 100% per index per day. Blank when the bond has no fresh price/market value that day (it still belongs to the index for the median and equal-issuer series).contrib_general_bps— the bond's contribution to the General market-value spread (z_mvtrue); the daily sum equalsz_mvtrue. It uses the index's winsorized (5/95-clipped) spread, so for extreme-spread (distressed) namescontribis notweight × z_bps— the spread is capped at the 95th percentile, exactly as in the published headline. This is why a distressed bond's contribution looks smaller than its raw spread would imply.z_bps— the bond's raw z-spread, before winsorization.is_new— 1 if the bond entered at the current month's rebalance (was not a constituent the previous calendar month).
Weights and contributions reproduce the published index exactly. Recomputed daily.
Methodology Appendix
Z-Spread Indices for Argentine Corporate Bonds (USD)
This appendix describes the methodology behind the Argentine hard-dollar corporate credit indices that feed the dashboard. It is intended for the data consumer: it explains what each series measures, how it is built, and how to read it.
1. What this product measures
These are daily indices of the z-spread of Argentine corporate bonds (obligaciones negociables, ONs) issued and paid in US dollars at a fixed rate.
The z-spread is the constant credit spread that, added to the entire US Treasury zero-rate curve, makes the present value of a bond's cash flows equal its market valuation. It is the standard, internationally comparable measure of a bond's credit risk: how much it yields above a risk-free asset of equivalent time profile. It is expressed in basis points (bps).
We publish four indices, which form a partition of the market (General = Energía + Bancos + Otros):
| Index | What it represents | Approx. composition (Jun-2026) |
|---|---|---|
| General | The total corporate market | ~150 bonds / ~58 issuers |
| Energía | Oil, gas, power, petrochemicals | ~100 bonds / ~31 issuers |
| Bancos | Banks and financial institutions | ~15 bonds / ~8 issuers |
| Otros | Non-energy, non-financial corporates (industrials, food, telecom, agribusiness, etc.) | ~36 bonds / ~19 issuers |
The number of bonds and issuers changes daily and at each monthly rebalance; the n and n_issuers columns reflect the live composition on each date.
The series begin on January 31, 2024 for General, Energía and Otros, and on January 1, 2025 for Bancos. There is no earlier history because the price input starts in January 2024.
2. Which series to use
Each index publishes several spread measures, because they answer different questions:
| Series | What it is | When to use it |
|---|---|---|
z_mvtrue_bps | Spread weighted by market value (market capitalization at dirty price), the standard of Bloomberg fixed-income indices. | Market headline. "What does the market yield per dollar invested?" |
z_med_bps | Median spread of the day. Not size-weighted. | Robust cross-check and headline for deep history (2024). The most stable series; reflects the typical bond. |
z_iss_bps | Equal-per-issuer average (each issuer carries the same weight). | The "typical issuer". More sensitive to stress than the median. |
z_chain_idx | Chained index, base 100: accumulates only the daily change of bonds present on consecutive days. | To measure moves ("how much did spreads move this month?") free of the noise from constituents entering and leaving. It is not a spread level. |
Note on
z_mvtrueand early history. The market-value weight requires knowing each bond's issued amount. That coverage built up over time and is complete from 2025; in 2024 it was partial. Therefore, for 2024 history we recommend thez_medheadline (which does not depend on the amount and is unaffected), andz_mvtrueas the market headline from 2025 onward.
Supplementary series
| Series | What it is |
|---|---|
tr_idx | Total return index, base 100. Measures how much corporate credit returned (price change + coupons), not just the spread. Reference: General +36% cumulative / ~13.9% annualized in USD since January 2024, with daily volatility ~0.28%. |
dur_mod_yrs | Average modified duration (years): sensitivity to rates. |
sdur_yrs | Average spread duration (years): sensitivity to spreads. |
dts_pct | Duration times spread: the modern measure of credit risk; estimated loss given a 100% relative widening of spreads. |
vol_chain_bps | Spread volatility, in bps/day. 21-business-day rolling standard deviation of the daily chained spread change (winsorized 5/95, common set of bonds day to day). Built from the chained change, so it is free of the composition jump at the monthly rebalance — unlike a volatility computed on a level series, which absorbs that jump. Comparable across sectors. Reference (Jun-2026, calm market): General ~7, Energía ~7, Otros ~10, Bancos ~24 bps/day; historical medians ~15 / 16 / 24 / 27. |
avg_price | Average dirty price (per 100 of face value). |
ttm_yrs | Average remaining life (years). |
n / n_issuers | Number of bonds / issuers in the index that day. |
distress_pct | % of the market yielding above 35% — a credit-stress thermometer. |
top_issuer_pct / top_bond_pct | % of the index concentrated in the largest issuer / largest bond. Diversification control. |
carried_pct | % of the index whose value for the day is carried (no fresh quote). A data-recency signal. |
quality_flag | Quality stamp for the day: OK, or a warning where applicable (e.g. high_conc = high concentration, typical of Bancos). |
3. Universe and basket construction
- Eligibility: corporates, in USD, fixed rate (dollar-linked bonds are excluded), with more than one year of remaining life and verified cash flows.
- Monthly rebalance: the composition is fixed on the last business day of each month and takes effect on the first business day of the following month. No bond enters or leaves intra-month; new issues are incorporated at the month-end rebalance following their appearance.
- Quotation line: each bond may quote in pesos, MEP dollars and cable, with differing yields. A single line per bond per day is used, prioritizing the MEP line (the highest USD-volume and lowest-noise line).
- Data validity: a bond's yield is considered current for up to 5 business days; thereafter its spread is carried for up to 10 more business days before dropping out of the calculation. This prevents a bond that stops trading from disappearing abruptly.
4. Calculation methodology
- Base curve. The US Treasury yield curve is transformed into semiannual zero rates (discount-factor interpolation).
- Valuation. The present value of the bond's cash flows, discounted at its market yield, defines its valuation. This approach cancels any scale factor (face size, residual value, exchange rate): the z-spread does not depend on the nominal price or the FX rate.
- Z-spread. We solve for the
zsuch that the present value of the cash flows at (zero curve + z) equals the present value at the bond's market yield, with semiannual discounting (Bloomberg's YAS convention). - Weighting and caps. The market-value average applies an iterative capping algorithm with limits of 40% per bond and 50% per issuer, redistributing the excess among the uncapped names and normalizing exactly to 100%. These limits hold on the final weights actually used.
- Winsorization. Each day, the cross-section values are clipped to the 5th and 95th percentiles before averaging, so that an isolated extreme value does not distort the aggregate. The median is insensitive by construction.
- Total return (
tr_idx). Each bond's return combines its dirty-price change and the coupons and principal collected, computed only between days with a fresh price (no returns over trading gaps). The index uses market-value weights fixed at the start of each month (not contemporaneous), so the weight does not incorporate the very price move being measured — standard practice for total-return indices. - Publication minimum. A sector index is published only with at least 5 bonds on the day.
5. Treatment of distressed bonds and credit events
- Credit deterioration is included: bonds whose yield widens up to 100% are part of the index at full weight (bounded by winsorization). The index reflects stress as it happens.
- A yield above 100% is treated as unreliable data (data hygiene), not as a credit event: that observation is excluded, but the bond's last valid spread is carried before it drops out. It is not a threshold that erases real economic events.
- A default or exchange is processed as a membership removal at the month-end rebalance, at the last observed price. The spread widening the bond experienced remains preserved in the history.
- These indices measure performing to distressed credit; they are not indices of already-defaulted debt (they do not mark recovery prices).
A drop in n together with an elevated distress_pct is the signal of an ongoing credit event.
6. Usage considerations
- History stability and the settling window. Recent values settle over a short window: the current and previous calendar month remain provisional and may adjust slightly as the underlying (often illiquid) prices firm up. Every date older than that is final and is never revised. In practice, a given date becomes definitive between one and two months after it occurs (sooner for end-of-month dates, later for early-month ones). Any correction to already-final history would be applied only as an announced restatement, never silently — so the deep history a client analyzes is stable, while the most recent ~2 months should be read as still consolidating.
- Price quality. Yields come from market prices of bonds that are in many cases illiquid. Daily moves of ±10 to 30 bps in the aggregates are normal microstructure noise for this market; read trends rather than single points.
- Bancos is a concentrated indicator. With ~5 issuers, the diversification caps bind on almost every day, but the sector remains structurally concentrated. Use it as a thermometer of bank risk, not as a diversified index comparable to General, Energía or Otros.
- Levels vs. chained. A spread level (
z_med,z_mvtrue) can fall because new bonds entered at lower spreads, without any bond having tightened. The chained index (z_chain_idx) isolates that composition effect. - Duration-based metrics move in monthly steps — that is the rebalance, not the market.
sdur_yrs,dur_mod_yrs,dts_pct(and any estimated-loss figure derived from them, e.g. the price impact of a +100 bps spread shock ≈−sdur × 1%) are properties of the frozen monthly basket: they barely move intra-month (most days the daily change is ~0) and jump on the first business day of each month, when the composition rolls over. These steps are rebalances, not market regime shifts or credit events — expected behaviour of a monthly frozen-basket index (Bloomberg-style). - Measuring volatility. To gauge day-to-day spread turbulence, use
vol_chain_bps(rolling volatility built from the chained daily change, in bps/day). A volatility computed instead on the daily change of a level series (z_mvtrue,z_med) absorbs a spurious composition jump at each monthly rebalance (~2× a normal day onz_mvtrue), which a 1-month rolling window then carries as a plateau;vol_chain_bpsis free of that artefact.
7. Data sources
| Input | Use |
|---|---|
| Market prices and yields of the bonds (BYMA and associated sources) | Daily yield, maturity, issuer, industry |
| Each bond's cash-flow schedule | Coupons and amortizations |
| US Treasury yield curve | Risk-free base rate |
| Issuance data (outstanding amount) | Market-value weight |
The indices are recalculated daily. Methodological questions can be directed to the Alphacast team.