Page 1 of 1

Credit curves for illiquid bonds

Posted: October 11th, 2016, 4:44 pm
by Samsaveel
any papers on how to build credit curves for illiquid bonds, looking here for frameworks to estimate spreads, illiquidity premium,country premium,etc.

Re: Credit curves for illiquid bonds

Posted: October 11th, 2016, 9:12 pm
by list1
any papers on how to build credit curves for illiquid bonds, looking here for frameworks to estimate spreads, illiquidity premium,country premium,etc.

Re: Credit curves for illiquid bonds

Posted: October 11th, 2016, 9:30 pm
by outrun
Interesting new development list1.

It depends a bit on your application. Is it for pricing, risk management?

A common method is to look a the spread between the illiquid bond and a comparable liquid reference bond (or curve) you pick. But only look a the spread at times when the illiquid bond traded because it can be stale for a long time while the liquid reference bond gets constant price updates.

When the illiquid bond trades it'll have a wide bid ask spread, so you can also think a bit about that: was side was traded: was an offer lifted, or a bid hit? You might be able to estimate the bid-ask spread based on the few trades you observe, but it won't be robust. That wide bid ask spread is also something to think about in the context of valuation, risk, liquidity cost etc.

Finally when you estimate the spread between your illiquid bond and the liquid reference one you need to take care of the fact (for some applications) that the spread is dynamic, it'll change over time,.. there is risk in that too.

Re: Credit curves for illiquid bonds

Posted: October 12th, 2016, 6:39 pm
by Samsaveel
it is for risk management purposes, mainly corporate bonds issued in domestic currency (GCC /currency pegged to USD) . when priced off interbank curve, the price  does not reflect the illiquid nature of the underlying credit. My take on this is to start with the US treasury curve as base curve and move from there to decompose the spread into country + illiquidity premium, not sure if this is the right step 

Re: Credit curves for illiquid bonds

Posted: October 12th, 2016, 7:49 pm
by outrun
Sound like a good plan. Maybe also add a credit rating curve. 

Depending on the type of analysis, a remaining shortcoming is that this doesn't capture the illiquidity premium risk factors. 

E.g. if you look at two different illiquid bond tied to the same country curve, and you have a long position in one, and short in the other, then this setup will report it to be risk free whereas it's actually double risky (from the illiquidity point of view). The issue is the same with only long bonds: the illiquidity premium will change unpredictable in time, premium changes will correlate between different bonds, and they will correlate with the underlying curve movements.

I don't think there is a simple solution and it's probably best to manage the illiquidity risk separate with different methods than analysing the impact of curve movements. Monitor and enforing caps on the exposure, think about how much it would cost to liquidate them if you had to.

Re: Credit curves for illiquid bonds

Posted: October 13th, 2016, 3:00 pm
by list1
it is for risk management purposes, mainly corporate bonds issued in domestic currency (GCC /currency pegged to USD) . when priced off interbank curve, the price  does not reflect the illiquid nature of the underlying credit. My take on this is to start with the US treasury curve as base curve and move from there to decompose the spread into country + illiquidity premium, not sure if this is the right step 
I wold say that current economy does not so classic to use classical models of default. Low interest rates in top economies is stipulated by excessive supply of cash and not of economy good conditions. Therefore the assumption that US t rate is risk free is too strong. The indirect effect on credit rating of the bonds makes also stock market which based ob excessive supply of cash does not show a visible growth. Reduced form should be extended taking into account currencies values and bond effect on indexes.

Re: Credit curves for illiquid bonds

Posted: October 24th, 2016, 3:40 am
by Samsaveel
Sound like a good plan. Maybe also add a credit rating curve. 

Depending on the type of analysis, a remaining shortcoming is that this doesn't capture the illiquidity premium risk factors. 

E.g. if you look at two different illiquid bond tied to the same country curve, and you have a long position in one, and short in the other, then this setup will report it to be risk free whereas it's actually double risky (from the illiquidity point of view). The issue is the same with only long bonds: the illiquidity premium will change unpredictable in time, premium changes will correlate between different bonds, and they will correlate with the underlying curve movements.

I don't think there is a simple solution and it's probably best to manage the illiquidity risk separate with different methods than analysing the impact of curve movements. Monitor and enforing caps on the exposure, think about how much it would cost to liquidate them if you had to.
Agree, the illiquidity premium will be a tough one  to quantify in less developed markets. what are the major risks that lead to higher illiquidity premiums beside Economic & financial risks?

Re: Credit curves for illiquid bonds

Posted: November 6th, 2016, 10:04 pm
by pimpel
it is for risk management purposes, mainly corporate bonds issued in domestic currency (GCC /currency pegged to USD) . when priced off interbank curve, the price  does not reflect the illiquid nature of the underlying credit. My take on this is to start with the US treasury curve as base curve and move from there to decompose the spread into country + illiquidity premium, not sure if this is the right step 
If you start from US treasury and are concerned about the spread in domestic currency, then definitely you need to take into account the way you fund your position so Cross Currency Basis Swap spread will be crucial. Then on top of that, in some of the countries you may find some non-negligible tax impact on bond prices, which should also be taken into account.
I would rather first built a liquid credit curve for a peer group in that currency and then just marked the specific spread to this curve as it was already indicated in this thread. By doing so, you would address at least systemic component of your risk, with specific part left to rare updates.

Re: Credit curves for illiquid bonds

Posted: November 8th, 2016, 4:52 pm
by Samsaveel
it is for risk management purposes, mainly corporate bonds issued in domestic currency (GCC /currency pegged to USD) . when priced off interbank curve, the price  does not reflect the illiquid nature of the underlying credit. My take on this is to start with the US treasury curve as base curve and move from there to decompose the spread into country + illiquidity premium, not sure if this is the right step 
If you start from US treasury and are concerned about the spread in domestic currency, then definitely you need to take into account the way you fund your position so Cross Currency Basis Swap spread will be crucial. Then on top of that, in some of the countries you may find some non-negligible tax impact on bond prices, which should also be taken into account.
I would rather first built a liquid credit curve for a peer group in that currency and then just marked the specific spread to this curve as it was already indicated in this thread. By doing so, you would address at least systemic component of your risk, with specific part left to rare updates.
is the funding part a uniform spread  and over what reference curve ? does the capital gain tax play a major or minor role in the setup?

Re: Credit curves for illiquid bonds

Posted: December 8th, 2016, 9:06 pm
by pimpel
Regarding the funding spread, you may find evidence for such add-on in recent FRTB standardized method for market risk capital requirement, where the regulator requires to hold capital based on bond's sensitivity if the bond is denominated in foreign currency. If the regulator see the problem of the cross currency basis swap spread on foreign bond, the market must see this issue too.
Regarding the tax issues, it is very country/region specific. There must be a reason, why a lot of Eurobonds is issued by SPVs registered in Stockholm, just to start with. The tax issues are out of my horizon, therefore for me it is hard to comment on that in detail.

Re: Credit curves for illiquid bonds

Posted: December 8th, 2016, 11:01 pm
by list1
Regarding the funding spread, you may find evidence for such add-on in recent FRTB standardized method for market risk capital requirement, where the regulator requires to hold capital based on bond's sensitivity if the bond is denominated in foreign currency. If the regulator see the problem of the cross currency basis swap spread on foreign bond, the market must see this issue too.
Regarding the tax issues, it is very country/region specific. There must be a reason, why a lot of Eurobonds is issued by SPVs registered in Stockholm, just to start with. The tax issues are out of my horizon, therefore for me it is hard to comment on that in detail.
Can we interpret the tax payments issue as floating coupon payments occurred at the same dates as tax payments for tax payers. 

Re: Credit curves for illiquid bonds

Posted: December 9th, 2016, 3:33 am
by Samsaveel
Regarding the funding spread, you may find evidence for such add-on in recent FRTB standardized method for market risk capital requirement, where the regulator requires to hold capital based on bond's sensitivity if the bond is denominated in foreign currency. If the regulator see the problem of the cross currency basis swap spread on foreign bond, the market must see this issue too.
Regarding the tax issues, it is very country/region specific. There must be a reason, why a lot of Eurobonds is issued by SPVs registered in Stockholm, just to start with. The tax issues are out of my horizon, therefore for me it is hard to comment on that in detail.
are you referring to the credit spread risk (credit delta) capital charge or the GIRR (IR Delta)  under FRTB-SBA? have to dig more on the SPV part:)

Re: Credit curves for illiquid bonds

Posted: December 9th, 2016, 8:41 am
by pimpel
Girr risk.not sure if it was in Basel dokument.for sure it is in European Commission draft CRR amendment.

Re: Credit curves for illiquid bonds

Posted: December 9th, 2016, 9:30 am
by list1
in two pages http://www.marketsrisks.com/ one can find an introduction to credit problems