Credit Ratings & Spreads Reference

How creditworthiness is measured, how it translates into yield spreads, and how the three fixed-income universes used in this toolkit — Treasuries, corporate bonds, and private credit — compare.

Credit Ratings

A credit rating is an opinion on the likelihood that a borrower will meet its debt obligations. The two major scales are:

Quality tier	S&P / Fitch	Moody's	This toolkit
Highest quality	AAA	Aaa	AAA
High quality	AA+/AA/AA−	Aa1/Aa2/Aa3	AA
Upper medium	A+/A/A−	A1/A2/A3	A
Medium	BBB+/BBB/BBB−	Baa1/Baa2/Baa3	BBB
Speculative	BB+/BB/BB−	Ba1/Ba2/Ba3	BB
Highly speculative	B+/B/B−	B1/B2/B3	B

Everything BBB / Baa3 and above is investment grade (IG). Everything below is high yield (HY), sometimes called "junk."

The toolkit collapses the notch-level detail (e.g. AA+, AA, AA−) into a single letter grade for simplicity.

Credit Spread

What it measures: The extra yield an investor demands over the risk-free rate to compensate for credit risk (and, in some markets, liquidity risk).

\[ \text{Spread} = y_{\text{bond}} - y_{\text{Treasury}} \]

Spreads are quoted in basis points (bps), where 1 bp = 0.01%.

Key drivers of spread width:

Factor	Effect
Lower credit rating	Wider spread
Longer maturity	Wider spread (more time for default)
Economic stress	Wider spreads across all ratings
Illiquidity	Wider spread (private markets)

The Spread Curve

Spreads vary by both rating and maturity. The toolkit stores a full term structure in assumptions/credit_spreads.csv (values in bps):

Rating	1Y	2Y	3Y	5Y	7Y	10Y	20Y	30Y
AAA	23	35	46	70	93	116	139	151
AA	22	32	42	63	82	100	120	131
A	18	27	36	52	64	76	91	97
BBB	48	68	87	119	143	167	207	223
BB	106	141	177	226	262	297	354	375
B	212	264	312	374	418	460	531	552

When a bond's maturity falls between table tenors, get_spread() linearly interpolates.

Code: get_spread() and get_credit_spreads() in src/alm/read.py

Updating Spreads from Market Data

The spread table can be refreshed from four FRED series that anchor the 10-year point for specific ratings:

Rating	FRED anchor	Series ID	Description
AAA	AAA10Y	AAA10Y	Moody's Aaa yield − 10Y Treasury
A	IG_OAS	BAMLC0A0CM	ICE BofA US Corporate IG OAS
BBB	BAA10Y	BAA10Y	Moody's Baa yield − 10Y Treasury
BB	HY_OAS	BAMLH0A0HYM2	ICE BofA US High Yield OAS

The update algorithm:

Fetch the latest 10Y anchor spread for AAA, A, BBB, BB from FRED
Compute a scale factor for each: new_10Y / old_10Y
Multiply the entire tenor curve for that rating by the scale factor
AA is interpolated between AAA and A, preserving its original relative position between the two
B is extrapolated from BB, preserving the original BB-to-B ratio at each tenor
Round to whole basis points and write back to CSV

Code: update_credit_spreads() in src/alm/read.py

Comparing the Three Asset Classes

The toolkit models three fixed-income universes with distinct risk and return characteristics:

Treasuries (Government Bonds)

Attribute	Detail
Issuer	U.S. government (sovereign)
Credit risk	Effectively zero (AAA / AA rated)
Liquidity	Highest — deep, transparent markets
Spread over risk-free	Minimal (the Treasury is the risk-free benchmark)
Typical ratings in toolkit	70% AAA, 30% AA
Role in portfolio	Duration anchor, safe-haven allocation

Treasuries define the risk-free yield curve. In this toolkit, the "discount rate" generally represents a Treasury-level rate, and all other instruments earn a spread above it.

Corporate Bonds

Attribute	Detail
Issuer	Corporations (public companies)
Credit risk	Moderate — depends on rating
Liquidity	Good for IG; lower for HY
Spread over risk-free	50–550+ bps depending on rating and maturity
Typical ratings in toolkit	30% A, 50% BBB, 15% BB, 5% B
Role in portfolio	Yield enhancement, diversification

Corporate bonds are the largest source of credit spread income in a typical insurance portfolio. The toolkit tilts toward BBB — the sweet spot at the bottom of investment grade, where spreads are meaningfully wider than A-rated debt but default risk remains relatively contained.

Private Credit

Attribute	Detail
Issuer	Middle-market companies, project finance, etc.
Credit risk	Higher — typically BB / B rated
Liquidity	Very low — no secondary market
Spread over risk-free	Wide: credit spread + illiquidity premium + other
Typical ratings in toolkit	40% BB, 60% B
Role in portfolio	Yield pickup, illiquidity premium capture
Regulatory cap	10% of portfolio (SAA constraint)

Private credit earns the widest total yield because it compensates for both credit risk and illiquidity. The toolkit decomposes the total yield into four components:

\[ y_{\text{total}} = y_{\text{risk-free}} + s_{\text{credit}} + s_{\text{illiquidity}} + s_{\text{other}} \]

Component	Toolkit default	Source
Risk-free rate	Varies (discount rate)	Treasury curve
Credit spread	Varies by rating / maturity	`assumptions/credit_spreads.csv`
Illiquidity spread	200 bps (hardcoded)	Industry estimate
Other spread	50 bps (hardcoded)	Catch-all (complexity, structuring)

Valuation note: The PrivateCredit class discounts at the risk-free rate by default (not the total yield). This means the PV exceeds par — the excess represents the economic value of the illiquidity premium to a buy-and-hold investor like an insurer.

Code: PrivateCredit in src/alm/asset.py, PC_RATING_DIST in src/alm/core.py

Side-by-Side Summary

	Treasury	Corporate	Private Credit
Credit quality	AAA / AA	A to B	BB / B
10Y spread (bps)	~0	76–460	297–460 + 250 illiq/other
Liquidity	Excellent	Good (IG) / Fair (HY)	Poor
Valuation method	Discount at yield	Discount at yield	Discount at risk-free rate
Duration behavior	Standard	Standard	Standard (bullet-like)
SAA weight (default)	40%	30%	10%
Maturities (default)	5, 10, 20, 30Y	5, 10, 20, 30Y	3, 5Y

Rating Distributions

Each asset class in the toolkit has a hardcoded rating distribution used when generating block-level portfolios:

Government bonds (GOVT_RATING_DIST):

Rating	Weight
AAA	70%
AA	30%

Corporate bonds (CORP_RATING_DIST):

Rating	Weight
A	30%
BBB	50%
BB	15%
B	5%

Private credit (PC_RATING_DIST):

Rating	Weight
BB	40%
B	60%

Code: GOVT_RATING_DIST, CORP_RATING_DIST, PC_RATING_DIST in src/alm/core.py

Rating	1Y	2Y	3Y	5Y	7Y	10Y	20Y	30Y
AAA	23	35	46	70	93	116	139	151
AA	22	32	42	63	82	100	120	131
A	18	27	36	52	64	76	91	97
BBB	48	68	87	119	143	167	207	223
BB	106	141	177	226	262	297	354	375
B	212	264	312	374	418	460	531	552

Rating	1Y	2Y	3Y	5Y	7Y	10Y	20Y	30Y
AAA	23	35	46	70	93	116	139	151
AA	22	32	42	63	82	100	120	131
A	18	27	36	52	64	76	91	97
BBB	48	68	87	119	143	167	207	223
BB	106	141	177	226	262	297	354	375
B	212	264	312	374	418	460	531	552

Rating	1Y	2Y	3Y	5Y	7Y	10Y	20Y	30Y
AAA	23	35	46	70	93	116	139	151
AA	22	32	42	63	82	100	120	131
A	18	27	36	52	64	76	91	97
BBB	48	68	87	119	143	167	207	223
BB	106	141	177	226	262	297	354	375
B	212	264	312	374	418	460	531	552