Podcast: Richard Martin on EM debt, copulas, machine learning

Quant sceptical of machine learning algos and black boxes

Richard Martin podcast 210618
Richard Martin (right) in conversation with Nazneen Sherif and Mauro Cesa
Image: Monika Ghose

Quant sceptical of machine learning algos and black boxes

Richard Martin, principal in the emerging markets desk (EM) at Apollo Global Management, visited our studio to discuss his new paper: Emerging market corporate bonds as first-to-default baskets.

Emerging economy debt markets have grown significantly over the past decade, due to improved fundamentals and low interest rates in the developed markets that attracted inflows of capital into the asset class.

As in any investment process, the investor takes a decision on what security to invest in and when to do it. In the case of EM corporate bonds, the exposure is to both country risk and the industry the corporate belongs to. Martin’s paper investigates the relationship between these two risk factors and proposes a model based on the famous Merton model for corporate liabilities, modifying its first-to-default basket approach and incorporating jumps into the process to simulate a realistic credit default swap curve.

Martin also shares his views on copula models – the infamous mathematical tools that took a good share of the blame for the collapse of portions of the credit derivatives market during the credit crisis. “Copulas have got a bad name, and I think deservedly so,” says Martin, who believes they have done very little to help the structured credit market.

Before explaining his future research plans, he didn’t hide his scepticism on machine learning, especially when it comes to carrying out optimal execution strategies in trading: “I’m not quite sure of what the machine actually learns,” he says.

“Anecdotally, what I have heard is anything a machine comes up with is likely to be arbed out very quickly.”

 

Index:

00:00. Intro and discussion of corporate bonds in emerging markets

02:30. What is your paper proposing?

09:54. Drawbacks of first-to-default framework

11:25. How does your modified first-to-default model work?

15:20. Copula models. Still around? Why?

22:12. VAR, ES and alternative risk measures

24:50. Machine learning. Overhyped?

30:35. Future research projects: Fokker-Plank equations

To hear the full interview, listen in the player above, or download. Future podcasts in our Quantcast series will be uploaded to Risk.net. You can also visit the main page here to access all tracks or go to the iTunes store to listen and subscribe.

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact info@risk.net or view our subscription options here: http://subscriptions.risk.net/subscribe

You are currently unable to copy this content. Please contact info@risk.net to find out more.

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here