Regulatory capital vs Economic capital
- Thread starter Imad
- Start date
Hi Imad,
I don't think Basel II says that, if you can point me to the context, maybe I can help parse it?
Basel II expects them to be different, for one thing the EC is a broader measure, by design meant to catch risks that "fall through the regulatory cracks."
In practice, you'd expect EC = max[reg capital, internal EC], but even this misleads and is not necessary as Basel Pillar one is, by definition, a RC requirement and Basel generally (incl P2 and P3) is concerned, not with imposing an fixed, internal EC threshold, but giving supervisors the job of ensuring the banks internal EC is sound/comprehensive/disclosed. EC is a big topic and hard to grasp as it's not available capital but rather a risk measure (it is on a different yardstick, so to speak. A key reason is the, unlike RC which is somewhat in common to all banks, EC is internal and varies by bank), but I can say: in my opinion, that statement is not true.
Although, both RC and EC do cover UL, they are similar in this respect. But the UL is just a quantile (VaR) based on a distribution. RC imposes a distribution; EC aggregates its own distribution based on whatever internal method, so there is a superficial sense in which RC and EC ~ UL but it's superficial b/c all of the work and differences are about "how do we figure UL?"
This FRM assigned reading may be helpful: http://db.tt/nFGKzmdr
(it contains some concrete points that show why in practice EC is different than RC).
Thanks, David
I don't think Basel II says that, if you can point me to the context, maybe I can help parse it?
Basel II expects them to be different, for one thing the EC is a broader measure, by design meant to catch risks that "fall through the regulatory cracks."
In practice, you'd expect EC = max[reg capital, internal EC], but even this misleads and is not necessary as Basel Pillar one is, by definition, a RC requirement and Basel generally (incl P2 and P3) is concerned, not with imposing an fixed, internal EC threshold, but giving supervisors the job of ensuring the banks internal EC is sound/comprehensive/disclosed. EC is a big topic and hard to grasp as it's not available capital but rather a risk measure (it is on a different yardstick, so to speak. A key reason is the, unlike RC which is somewhat in common to all banks, EC is internal and varies by bank), but I can say: in my opinion, that statement is not true.
Although, both RC and EC do cover UL, they are similar in this respect. But the UL is just a quantile (VaR) based on a distribution. RC imposes a distribution; EC aggregates its own distribution based on whatever internal method, so there is a superficial sense in which RC and EC ~ UL but it's superficial b/c all of the work and differences are about "how do we figure UL?"
This FRM assigned reading may be helpful: http://db.tt/nFGKzmdr
(it contains some concrete points that show why in practice EC is different than RC).
Thanks, David
David,
Thankyou for your reply. In schweser (book 3), it states that: "Economic capital/regulatory capital is used for unxepected variations from expected losses called unexpected losses (UL)".
I can understand from the above that both capitals are to cover UL (which is not right?).
Thanks
Imad
Thankyou for your reply. In schweser (book 3), it states that: "Economic capital/regulatory capital is used for unxepected variations from expected losses called unexpected losses (UL)".
I can understand from the above that both capitals are to cover UL (which is not right?).
Thanks
Imad
Imad,
That is a TRUE statement (but please note that statement in no way implies that EC = RC).
If we consider credit risk (eg), UL = CVaR(confidence) - EL. Both regulatory capital and, generally, economic capital will "cover" this UL, but the risks included and methodology will differ between the two, so the capital levels will be different. And this same idea applies to market risk (although here the drift/EL is positive) and operational risk. In this respect, your quote above is absolutely true about this similarity. Thanks, David
That is a TRUE statement (but please note that statement in no way implies that EC = RC).
If we consider credit risk (eg), UL = CVaR(confidence) - EL. Both regulatory capital and, generally, economic capital will "cover" this UL, but the risks included and methodology will differ between the two, so the capital levels will be different. And this same idea applies to market risk (although here the drift/EL is positive) and operational risk. In this respect, your quote above is absolutely true about this similarity. Thanks, David
Hi David,
Sorry for hammering you with questions! Topic is "Portfolio Risk: Analytical methods". I know that to get an optimal portfolio, Return over MVAR should be constant (E/MVAR=cte). Let's say we have two assets X and Y, E/MVAR (X) is less than E/MVAR (Y). To move toward the optimal portfolio, will the manager increase the allocation of X or Y? Appreciate your feedback on this.
Thanks
Imad
Sorry for hammering you with questions! Topic is "Portfolio Risk: Analytical methods". I know that to get an optimal portfolio, Return over MVAR should be constant (E/MVAR=cte). Let's say we have two assets X and Y, E/MVAR (X) is less than E/MVAR (Y). To move toward the optimal portfolio, will the manager increase the allocation of X or Y? Appreciate your feedback on this.
Thanks
Imad
Hi Imad,
If E/MVAR(Y) is higher, we shift (increase) allocation to asset Y; as it outperforms on the Treynor ratio. So this is like,"let's put more of the portfolio into our best risk-adjusted returns."
In doing so, we will increase the beta (Y, portfolio) and decrease the beta (X, portfolio); e.g., as we increase Y, it looks more an more like the portfolio, tending toward becoming the portfolio.
This, in turn, causes the higher E/MVAR(Y) = E/beta to decrease and the lower E/MVAR(X) to increase (denominator falling).
So just shifting allocation into the higher E/MVAR(Y) sends both ratios nearer to each other.
Thanks, David
If E/MVAR(Y) is higher, we shift (increase) allocation to asset Y; as it outperforms on the Treynor ratio. So this is like,"let's put more of the portfolio into our best risk-adjusted returns."
In doing so, we will increase the beta (Y, portfolio) and decrease the beta (X, portfolio); e.g., as we increase Y, it looks more an more like the portfolio, tending toward becoming the portfolio.
This, in turn, causes the higher E/MVAR(Y) = E/beta to decrease and the lower E/MVAR(X) to increase (denominator falling).
So just shifting allocation into the higher E/MVAR(Y) sends both ratios nearer to each other.
Thanks, David
Imad, That Dowd's GPD formula (p 203 of the assignment), the VaR answer is similarly given in percentage terms; i.e., the output VaR (e.g., 3.952) is at least consistent with the input (2.00).
The formula is replicated in our learning XLS @ http://www.bionicturtle.com/how-to/spreadsheet/2011.t5.d.3.-evt
Thanks, David
The formula is replicated in our learning XLS @ http://www.bionicturtle.com/how-to/spreadsheet/2011.t5.d.3.-evt
Thanks, David
on the topic of econ capital vs reg capital.. do you agree with the following statements:
"economic capital can be used to validate a firm's regulatory capital requirement against its own assessment of the risks it's running"
and
"since regulatory capital models and economic capital models have different objectives, econ capital models cannot help regulators in setting regulatory capital requirements."
If you can explain the reasoning behind that'd be much appreciated (as always
Thanks
"economic capital can be used to validate a firm's regulatory capital requirement against its own assessment of the risks it's running"
and
"since regulatory capital models and economic capital models have different objectives, econ capital models cannot help regulators in setting regulatory capital requirements."
If you can explain the reasoning behind that'd be much appreciated (as always
Thanks
Similar threads
