Skip to Content

Deep dive into Quantitative Bond Investing

We all invest in bonds differently.


  • Invest in bonds as a portfolio of bonds in a unit trust or an exchange-traded fund (ETF)
  • Buy a $250,000-denominated bond direct and wishes to hold till maturity
  • Buy it through an income fund, which can be holding high-yield bonds, equity and bonds, or a mixture.
  • Buying a triple-leveraged bond ETF as part of a constantly rebalancing strategy with a triple-leveraged equity ETF
  • Leveraging up individual bonds, at high-quality investment grade, low-quality investment grade or junk status

If you know what you are doing, then it is all good.

But often, I realize there are some that think they know what they are doing but deep down I worry for them.

There are different degrees to the part about “knowing what you are doing”.

There will be some who think they know what they’re doing but if they run the same bond investment for a few iterations, shit is going to happen.

It is always good to hear from another perspective.

Corey Hoffstein, CIO of Newfound Research, interviewed Greg Obenshain on his experience at Verdad using quantitative models on the credit and debt markets.

Greg worked as a high-yield portfolio manager at Apollo Global Management and Stone Tower Capital but started out as an equity analyst at GE’s energy financial group.

Verdad is rather data-focused so it is immensely helpful to find out what does the bond data tell us.

In short:

  1. Excess returns come from identifying bonds that are going to be upgraded or downgraded and actioning upon them.
  2. Viewed as an aggregate, rating agencies do a pretty good job in terms of their processes.
  3. The biggest opportunity lies in the bonds that are about to be upgraded to BBB.
  4. The biggest mistake is those folks using only coupons, yield as a metric.
  5. The driver of bond returns is the bond business cycle and not spreads.
  6. Value and momentum work well in fixed income space as equity, it is just that it is difficult to normalize.
  7. Private lending is a high risk, high reward and seldom is there an asymmetry reward.

Greg’s background as a fundamental analyst.

He started off as an energy debt analyst in a firm called Stone Tower. Stone Tower was acquired by famed alternative manager Apollo.

Analyzing energy debt is tough:

  1. Energy by itself is volatile.
  2. Bond by itself is not supposed to be volatile.

The bond world is different from the equity world in that you are dealing with contracts that have to be repaid by a certain due date.

There are a lot more companies with debt than equity. The universe is wider.

You have to understand:

  1. Debt structure
  2. Treasury market and how it’s doing

Greg realizes that he has to understand these things well and there is no easy way to analyze this with just an Excel spreadsheet.

So he bit the bullet and taught himself R and build the database he needs to do the work.

Is there any truth that credit is not an asset class? The distinct asset class is Equity and Treasuries and a combination of these two.

This train of taught came from David Swensen over at Yale Endowment. Swensen wrote in his book Portfolio Management that says credit is an “impure fixed income”.

Credit is just a hybrid asset class. You are just combining equities and government bonds in different amounts.

In credit or corporate bond, you are paid a default spread to take on the risk of default. You are not paid for the equity upside.

If you want to make an extra return, but don’t want to hit a default, you will get something with a lower drawdown, safer but you get to pay a little extra. That is the gold.

In practical application, you cannot make a credit spread by combining equity and treasury is just isn’t true.

So why do we have credit?

The truth is that most people do not want to be in 100% equities. But they also don’t want to be in too much treasury because treasury is low returning.

Credit is the middle ground because it gives you higher returns with much lower drawdowns.

Investors get it very wrong when they think that they are supposed to invest in credit so that they can make a lot of returns. They can do that but this is not the right approach.

You want to use credit the right way, which is to protect you from drawdowns. This means that you are going to have drawdowns but they are not going to be so severe. But it will give you some steady return and dramatically reduce the drawdown.

Should you only allocate to high yield bonds only when the yield spread is very high? That is when the juice is great.

Let us frame properly how the credit market work.

  1. You can have a AAA bond like Microsoft. You make 2%
  2. A CCC bond like WeWork. You make 10%

Most debt will be between the not going to default (AAA) and the very risky (CCC).

The belief to make money is to go down to the CCC and pick the super risky stuff with juicy yield.

The truth is that this is not the way to make money.

The most lucrative segment is not from the super risky stuff. The most lucrative is from the middle segment, the BB, BBB. You are getting an extra return but not all the bad stuff.

One of the ways to invest quantitatively with bonds is to move beyond thinking about yield and coupon, which everyone looks at.

Don’t look at yield and coupon.

If you chase after them, you are chasing after losses.

The market is efficient.

If you chase after a 15% yielding bond, you will get a 4% return and realize 11% in losses.

Investors need to think about credit in terms of what is getting better and what is getting worse.

It is better to own that 4% bond that is going to trade to 3% than the 8% bond that is going to trade to 9%.

It is true that we should only buy credit when credit spreads are wide. During Covid, the spreads widen and Greg’s team put out a piece talking about what to buy and ranked highest is high yield bonds. The credit spreads are wide and the risk of default is much lesser.

But in truth, what investors should be buying is smallcap value equity because it is the riskiest and highest return.

In their team’s view of the world, buying credit is more closer to the preservation of capital.

The right time to buy credit is when the credit spreads are very tight. This is when yields are really low and you don’t like what is going on in the rest of the market.

You are able to earn a return, limit your portfolio drawdown, and when the risky situation comes, sell your credit and safer bonds and go out and buy the riskier stuff.

The credit spread opportunity statement is true if we look narrowly at the asset class but less true if we look holistically.

Why didn’t Quantitative Fixed Income take off in a bigger way?

Firstly, in order to start a quant fixed income fund, you will need people with credit expertise.

There is a limited amount of people. Most people have equity expertise.

Secondly, you need data. It is exceptionally difficult to get the data.

There are thousands of ways to get equity data but it is difficult to get fixed income data. At Verdad, Greg’s team have to tediously build their own database by linking equity data with fixed income data.

Then, they are not allowed to trade half of those bonds unless they are an institutional bond buyer.

One of the big problems Greg sees is that those who try to break into quant fixed income approach it from the distressed side because the yields are higher, and you will get situations where the returns are out-sized.

That distressed side is also much smaller. Fewer bonds, poorer selection. The risk of blowing yourself up is also higher.

What is the main driver of Cross-sectional Excess Credit returns?

Cross-sectional means comparing one kind of bond to another kind of bond. Excess credit returns mean whether we are able to earn excess returns above the risk of default we are taking on.


In general, Greg reiterates that model:

  1. Move to bonds that are getting better in credit quality
  2. Move away from bonds where the credit quality is going to get worse

Usually, an analyst will look at this from a fundamental basis but quant models can work pretty well with this.

Philosophically, the rating agencies such Moody and Fitch are the early quants, doing the re-rating higher and lower.

We can don’t trust Moody or Fitch’s data, but we can make use of that structure.

If we know that a BBB bond is going to get better and upgraded to a BB bond, we are able to figure out what is the value it will have at BB because we are able to see the value of the BB bond today.

Verdad built their model based on this what is getting better and what is getting the worse concept.

What are the important characteristics to identify those rating upgrades and downgrade in bonds?

If we peel away from debt and look at the bigger picture, the businesses that issue the bonds are in the business of borrowing money and making money from the money borrowed.

Knowing the existing rating does not tell us much, and what is more important is what goes into how the rating is derived.

Greg thinks there is a simpler way to think about this.

If we are borrowing debt at 7% and our return on investment is 4%, our company is in liquidation mode (not good). This is often rated as single B credit.

If our return on investment is 7%-10% and we are borrowing at 3%, or that our business is producing so much cash flow that we do not even need to borrow in the first place, then this company will have a great credit rating.

In high yield credit, the key thing is whether a company can survive. And whether a company can survive is related to whether a company can make money.

A high yield company dies when they take on a higher cost of debt/capital versus their return on investment. They will be in slow liquidation mode.

There are good proxies for sensing that.

Revenue to Total Assets may be better than the traditional metric Debt to EBITDA. Debt to EBITDA is useless to assess oil and gas companies because if EBITDA gets cut in half, you will have a problem assessing it.

In general, we want to sense how much the business is spending versus how much return we are getting for the business. How profitable we are versus our cost of capital.

If a business has a high-interest expense, it impedes them from reinvesting in their business. If you cannot re-invest, you cannot grow your way out of the problem.

Leverage kills a business because it distracts a company from its ability to reinvest.

The company cannot make mistakes.

The more you mess up, the lower your credit, the higher your cost of capital.

If a company is adding on more debt, that is also not healthy.

The company should be seen as paying down debt. Debt paydown is the highest return strategy you can have. It is far better than paying a dividend to shareholders.

By paying down debt, the shareholders benefit by expanding equity and the company has a greater capacity to reinvest.

Is the catalyst for the excess return the upgrade and downgrade of the rating by the rating agency or is it from the market automatically pricing in the risk into an expected upgrade or downgrade?

The rating agency is not in the business of being very responsive to downgrading businesses.

Greg observes that if we look at individual bond ratings, the rating agencies can get the rating wrong often.

But if we view the rating as a cohort, the rating generally gets things well at force ranking.

The market is right most of the time, but in general, we also cannot ignore the rating agencies.

The most often observed example is when the rating agency rate that this is a B3 bond (which is a step above CCC). They will observe a lot of investors buying such bonds left and right due to the high yield.

These are the worse kind of bonds you could absolutely buy. This is because the downgrade of ratings will kill the bonds. The actual returns are horrific after factoring in the capital losses.

We can use the case study of the oil and gas crisis.

Greg observed that during the period leading up to the crisis, the general rating for the bonds is low. The rating agencies deserve a lot of credit for getting that right. While he will disagree on individual credit ratings, Greg has a lot of respect for their process.

Which are the characteristics that rating agencies overemphasize and which they underemphasize?

Quality and stability tend to be oversold. The companies you are rating will tend to pitch you rosy growth prospects.

Nobody has a good way to judge the quality of a management team, other than looking at the numbers.

Anyone who knows fundamental analysis can write a good story about it.

Seldom the rating agency will find a company that has really good prospects but for some scary fundamentals and rate them correctly.

Does the data debunked some of Greg’s pre-conceived notions about the debt market?

Debt-to-EBITDA, which is often used, does not test that well.

Equity market valuations get things correct quite a lot. We can use equity market value over debt.

There are metrics that may be totally stupid but it works.

One of them is total assets. If two companies trade at the same spread and yield, pick the one with more assets. It has more leeway when bad things happen.

Total asset is a strong metric that Greg previously did not see it coming.

There is a thought out there that private debt and lending outperform public debt. The data actually says otherwise. We are just taking on more risks for more yield with private debt.

Private debt is lending to highly leveraged private equity firms.

There is also a prevalent thought that if I buy senior, secured, floating-rate debt, I am buying the best debt.

Here is the reason.

If you are a strong issuer, you would be issuing a senior, secured, FIXED RATE debt with a long duration.

But if you can only issue floating-rate, with debt covenants, it is not a signal that the debt is good, it is a signal the debt is weak.

What are some of the factors that work in the equity world that also work in the fixed income world?

Value and momentum also work in the fixed-income world.

We can also overlay a quality screen overvalue (don’t buy a cheap company that is going to be bankrupt)

The trouble in the quant fixed-income world is how do you normalize these variables. How do we measure value in fixed income?

The metrics that need to be created are the probability of default or the relative ranking of the bond.

Leveraging on the different dimensionality of bonds issued by the same company.

The debt issued by a company can have different maturity dates, different yields, different spreads, different security levels.

You can simplify by averaging them out.

The numerous maturity dates can work in your favour in a way it will be difficult for equity.

For example, if we think that a company is going to be upgraded, we can buy a bond with a longer duration, because the bond with a longer duration is going to be more volatile.

If we are concerned about a certain aspect of the company, we can buy a shorter duration bond.

These are done without changing the dollar amount invested.

We can do the same with another dimension such as security level.

If we think a company is going to get upgraded, we should get the debt with the lowest seniority, because that is the most reactive to good news in a positive way.


How much machine learning has Verdad incorporated?

Machine learning is important to assess the quality of the bonds or the different rating categories.

Bonds that are going to go default act very differently from others.

If a bond that is going to go into default trades up, it tends to continue to go up. An investment-grade bond will go up but typically mean reverts quite quickly.

This is difficult to model and so what they did was to define about 98 different variables and built an upgrade and downgrade model (how to better predict price upgrade and downgrades).

The machine learning model is especially effective for identifying the bonds that are going to get downgraded.

The Fallen Angels Trade – Are there similar trades like this?

Fallen Angel trade is a well-known trade based upon the structural edge that emerges based on downgrades due to institutional constraints in the credit space.

This is also a trade that got Greg curious about why it works and whether there are similar trades like the Fallen Angel trade.

The best opportunity in credit: The category of debt slightly below investment grade is the area where the highest return you could get (the debt that is higher than CCC).

This is because they stand a better chance of an upgrade to investment grade. When the bonds get upgraded they do really well.

They also have a longer duration and higher quality.

The worse opportunity in credit: The edge between CCC and lower than that rating.

If you are a high-yield manager, usually you will have a CCC basket. There is a limit to how much CCC bond allocation you can load up on. So to get yield, what you will do is to load up on the bonds that are trading like CCC but not rated as CCC.

Greg looks at the returns of what is really really bad, it is this category.

What does the data tell us about private credit?

Greg was very sceptical about private credit because after being in the business for a long time, there wasn’t such a situation where there are so many businesses out there starving for loans.

The most bizarre part is: If this is such a good opportunity, then why are the banks not getting into this space since they are best positioned?

While there are good opportunities, most of the time, you are dealing with a counterparty willing to push their boundaries to a high level.

These counterparties have better lawyers than you and the banks are working for them not you.

As a lender, what you will get is something that is

  1. Overleveraged
  2. Not well documented

They rest in the region of single B credit, not yet CCC, that can get downgraded anytime.

When things go bad, the asymmetry is very terrible.

How does Fiscal Monetary Policy affect debt?

Loose monetary policy affects equity and bond markets.

It takes a long time before companies go bankrupt. Companies that should be dead get refinance.

Greg does not buy the argument that the Federal Reserve is saving the junk bond market.

What they were bailing out was the Fallen Angels and shoring up the investment-grade market. They were trying to make sure the mutual funds that hold the investment-grade bonds that got downgraded could do it and not kill the liquidity in the market.

Verdad has the credit spreads for the past 100 years so they know how the spreads have been historical.

The data says the credit spread is rather wide right now, relative to the long history.

The Fed will find it hard to suppress the credit spread, which indicates something real, which is losses.

The data tells Greg that the spreads have been rather stable and opportunities have not gone away.

Corey mentioned that according to the Bank of America High Yield OAS, the spread has narrow to points unseen since 2007. How would Greg reconcile this?

When Greg say the spread is rather wide, he is referring to the BAA spread which is the investment grade. In April 2021, CCC is very tight right now.

Why the high yield spread used by Bank of America is narrowed is because of a composition element.

If we go back to the first principle, that is how do we make money from credit, it is to preserve capital. And this may not be the ideal time to invest in bonds that have tight spreads.

This might be counter-intuitive, but tight spreads are when we should be investing in debt.

This is because the driver of bond returns is not credit spreads but the business cycle. We have a bond business cycle when the yield curve is inverted.

You have a bond business cycle when the credit spread is widening instead of narrowing.

This is when the FED starts or are lowering rates more.

Currently, none of these things mentioned above is happening.

There is a good argument to long risk (equities)

I invested in a diversified portfolio of exchange-traded funds (ETF) and stocks listed in the US, Hong Kong and London.

My preferred broker to trade and custodize my investments is Interactive Brokers. Interactive Brokers allow you to trade in the US, UK, Europe, Singapore, Hong Kong and many other markets. Options as well. There are no minimum monthly charges, very low forex fees for currency exchange, very low commissions for various markets.

To find out more visit Interactive Brokers today.

Join the Investment Moats Telegram channel here. I will share the materials, research, investment data, deals that I come across that enable me to run Investment Moats.

Do Like Me on Facebook. I share some tidbits that are not on the blog post there often. You can also choose to subscribe to my content via the email below.

I break down my resources according to these topics:

  1. Building Your Wealth Foundation – If you know and apply these simple financial concepts, your long term wealth should be pretty well managed. Find out what they are
  2. Active Investing – For active stock investors. My deeper thoughts from my stock investing experience
  3. Learning about REITs – My Free “Course” on REIT Investing for Beginners and Seasoned Investors
  4. Dividend Stock Tracker – Track all the common 4-10% yielding dividend stocks in SG
  5. Free Stock Portfolio Tracking Google Sheets that many love
  6. Retirement Planning, Financial Independence and Spending down money – My deep dive into how much you need to achieve these, and the different ways you can be financially free
  7. Providend – Where I used to work doing research. Fee-Only Advisory. No Commissions. Financial Independence Advisers and Retirement Specialists. No charge for the first meeting to understand how it works
  8. Havend – Where I currently work. We wish to deliver commission-based insurance advice in a better way.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

This site uses Akismet to reduce spam. Learn how your comment data is processed.