Good morning, Glenn, thank you for your diligent analysis as always. Bloomberg has modeled out implied high yield spreads based on VIX and the Move Index. According to their model HY spreads should be 628 bps based on current market vol vs. the OAS of the Bloomberg HY index which is 335 bps as of this AM ET. Do you think there is any credence to the modeled vol adjusted spread? i.e. where is reality?
Sorry about length below. A good thought exercise...excuse any typos.
On quant models generally…
When we were ready to start CreditSights way back in 2000 as the TMT credit cycle was turning, one of our priorities was to have a default model. It is nice to have a model and people like them as a gut check with objectives metrics.
My own version of talking about models is captured in something I rolled out at a conference once:
Models are great. But look at models as being in a hardware store. Shop for what tools you need at any given time. Just remember, it is a hardware store and not a church, so don’t worship. The models are just one more tool.
I got a laugh, but there was a triple jointed quant in the front row who scowled. Consider what any given model might have or not have and the limited time horizon you will be using it for.
To your question: Do you think there is any credence to the modeled vol adjusted spread? i.e. where is reality?
My answer: No. that +600 bps handle level is not reality to me. That takes you comfortably north of 10% YTW and expected return. You buy that all day long, but that is a dream for another day. That is a long-term equity return in a BB heavy index. The major bank system is sound (some regional bank question marks) and you have a booming market for alternative credit sources (private credit).
I personally think the level of bad news that you would need to get to 600+ include massive trade wars, rising stagflation fear, major FOMC mistakes, and an oil spike to get to that >600 level any time soon. Then you need to put it all in portfolio context vs. UST, credit, equities, etc. across the more typical asset allocation alternatives. Who would get crushed even more? If you believe outcome, you are in cash waiting for better pricing.
The model is sending a message that the market is materially tight, and it says you are not getting paid for the mix of scenarios. It is a wakeup call or maybe or even a tornado warning if widespread trade wars do in fact break out. If HY OAS widens by 100 bps, you will still be inside the long-term median. This gap from today vs. that model is closer to an additional 300 bps in OAS. That backdrop would send equities into a downward flush and swirl also.
The HY bond index does have a radically different mix by ratings tier and deal quality in HY bonds than 1997-2008 averages (higher quality now). The migration of the “crazy late cycle deal tradition” and usual excess risk-taking has been migrating to the private credit sector and not HY bonds. So that is a mix issue also on the low end.
A +600 bps handle OAS derived from a model is worth listening to as telling you the risk-reward symmetry sucks today. You can figure that out with a flip of the coin right now (heads too tight, tails too tight), but the question is which alternatives do you have for your chosen investment time horizon. Models often tell you the risk symmetry is poor, but that is pretty clear when you have been sitting within June 2007 OAS levels until this week’s widening.
If the focus is just spread (vs. relative yields, minimum expected yields, or cash income mix in total return), you are still triple digits away from fair value in spreads. A few questions I always ask myself are below on total return expectations using credit assets (I personally am waiting for the inevitable HY bond sell-off. I am not a credit bear per se or risk averse. I own some BDCs and related ETFs as long-term core holdings in my “old man income portfolio”).
Questions to ponder:
• What about the curve? Bull flattener view? Another bear steepener like 2024? Stagflation signals in 2H25? (that stagflation scenario would be a disaster).
• We saw FOMC easing in 2019 after a brutal end to 2018. Can we now expect bad outcomes from trade war stupidity, so the FOMC to the rescue in later 2025 like it did in 2019?
• Do we sell into relative strength? Ok where do we go? (I always get a kick of “underweight HY” recommendations without a destination recommendation, Cash? UST? The reality of asset class constraints means some HY portfolios need to be more defensive in quality Bs and BBs for a while. Income needs matter and a coupon matters subject to time horizons.
To wrap the model side, I have had more than my share of debates on models over the years with PhDs way out of my gray matter weight class. From my stint as Head of Credit Risk at Lehman in the mid-1990s debating “theoretical credit lines” on derivative exposures to our own models at CreditSights, there is always a good mental exercise to second guess inputs and recognize limitations (again, what is not in the model in the cause and effects bucket and relative importance of what is not there). Too many war stories there in risk limits and measurement of counterparty credit lines.
The best thing a model guy can do is admit limitations of inputs and not let extrapolation run amok. I usually beat them up on what is not in the model from macro variables, market structure issues (regulatory framework differences across the timeline, Glass Steagall, credit contraction risks, liquidity backstops, relative secondary liquidity mishaps and gap risk, mismarking of select assets including some that are only audited once a year, etc.). What about health of bank system and new rules in prop in a record sized OTC market? Depends on the asset class.
Fair value pricing has different ingredients from bond to loans and actively traded loans to SME exposures. Someone is always seeking the perfect model menu. Fair value models are food for thought. The whole pro-cyclical topic has been picked over by people way smarter than me. The volatility matters, but mapping it to pricing has an element of fake science. It is just one more item in the hardware store.
Anyway, that is my too long answer with some thoughts.
Good morning, Glenn, thank you for your diligent analysis as always. Bloomberg has modeled out implied high yield spreads based on VIX and the Move Index. According to their model HY spreads should be 628 bps based on current market vol vs. the OAS of the Bloomberg HY index which is 335 bps as of this AM ET. Do you think there is any credence to the modeled vol adjusted spread? i.e. where is reality?
Sorry about length below. A good thought exercise...excuse any typos.
On quant models generally…
When we were ready to start CreditSights way back in 2000 as the TMT credit cycle was turning, one of our priorities was to have a default model. It is nice to have a model and people like them as a gut check with objectives metrics.
My own version of talking about models is captured in something I rolled out at a conference once:
Models are great. But look at models as being in a hardware store. Shop for what tools you need at any given time. Just remember, it is a hardware store and not a church, so don’t worship. The models are just one more tool.
I got a laugh, but there was a triple jointed quant in the front row who scowled. Consider what any given model might have or not have and the limited time horizon you will be using it for.
To your question: Do you think there is any credence to the modeled vol adjusted spread? i.e. where is reality?
My answer: No. that +600 bps handle level is not reality to me. That takes you comfortably north of 10% YTW and expected return. You buy that all day long, but that is a dream for another day. That is a long-term equity return in a BB heavy index. The major bank system is sound (some regional bank question marks) and you have a booming market for alternative credit sources (private credit).
I personally think the level of bad news that you would need to get to 600+ include massive trade wars, rising stagflation fear, major FOMC mistakes, and an oil spike to get to that >600 level any time soon. Then you need to put it all in portfolio context vs. UST, credit, equities, etc. across the more typical asset allocation alternatives. Who would get crushed even more? If you believe outcome, you are in cash waiting for better pricing.
The model is sending a message that the market is materially tight, and it says you are not getting paid for the mix of scenarios. It is a wakeup call or maybe or even a tornado warning if widespread trade wars do in fact break out. If HY OAS widens by 100 bps, you will still be inside the long-term median. This gap from today vs. that model is closer to an additional 300 bps in OAS. That backdrop would send equities into a downward flush and swirl also.
The HY bond index does have a radically different mix by ratings tier and deal quality in HY bonds than 1997-2008 averages (higher quality now). The migration of the “crazy late cycle deal tradition” and usual excess risk-taking has been migrating to the private credit sector and not HY bonds. So that is a mix issue also on the low end.
A +600 bps handle OAS derived from a model is worth listening to as telling you the risk-reward symmetry sucks today. You can figure that out with a flip of the coin right now (heads too tight, tails too tight), but the question is which alternatives do you have for your chosen investment time horizon. Models often tell you the risk symmetry is poor, but that is pretty clear when you have been sitting within June 2007 OAS levels until this week’s widening.
If the focus is just spread (vs. relative yields, minimum expected yields, or cash income mix in total return), you are still triple digits away from fair value in spreads. A few questions I always ask myself are below on total return expectations using credit assets (I personally am waiting for the inevitable HY bond sell-off. I am not a credit bear per se or risk averse. I own some BDCs and related ETFs as long-term core holdings in my “old man income portfolio”).
Questions to ponder:
• What about the curve? Bull flattener view? Another bear steepener like 2024? Stagflation signals in 2H25? (that stagflation scenario would be a disaster).
• We saw FOMC easing in 2019 after a brutal end to 2018. Can we now expect bad outcomes from trade war stupidity, so the FOMC to the rescue in later 2025 like it did in 2019?
• Do we sell into relative strength? Ok where do we go? (I always get a kick of “underweight HY” recommendations without a destination recommendation, Cash? UST? The reality of asset class constraints means some HY portfolios need to be more defensive in quality Bs and BBs for a while. Income needs matter and a coupon matters subject to time horizons.
To wrap the model side, I have had more than my share of debates on models over the years with PhDs way out of my gray matter weight class. From my stint as Head of Credit Risk at Lehman in the mid-1990s debating “theoretical credit lines” on derivative exposures to our own models at CreditSights, there is always a good mental exercise to second guess inputs and recognize limitations (again, what is not in the model in the cause and effects bucket and relative importance of what is not there). Too many war stories there in risk limits and measurement of counterparty credit lines.
The best thing a model guy can do is admit limitations of inputs and not let extrapolation run amok. I usually beat them up on what is not in the model from macro variables, market structure issues (regulatory framework differences across the timeline, Glass Steagall, credit contraction risks, liquidity backstops, relative secondary liquidity mishaps and gap risk, mismarking of select assets including some that are only audited once a year, etc.). What about health of bank system and new rules in prop in a record sized OTC market? Depends on the asset class.
Fair value pricing has different ingredients from bond to loans and actively traded loans to SME exposures. Someone is always seeking the perfect model menu. Fair value models are food for thought. The whole pro-cyclical topic has been picked over by people way smarter than me. The volatility matters, but mapping it to pricing has an element of fake science. It is just one more item in the hardware store.
Anyway, that is my too long answer with some thoughts.
Thank you! I greatly appreciate the depth of response and candor.