BANK CAPITAL REQUIREMENTS COMPLIANCE AND THE ENRON FRAUD
By
SALLY RAMAGE
Operations risk modelling is becoming more important.
Different models allow correlations and applications for finance. Modelling has
value for finance, especially institutions and INSURANCE COMPANIES.
Where is operation risk modelling in this area of finance
and insurance? It is used in banks to control Capital Requirement.
The challenges are these:-
1.
Mathematics has methodologies
for calculating data, but not with real data. Banks now collect data but the
challenge is – trying to model real
data.
2.
Diversification is also an
issue.
3.
How to mix internal data with
external data.
The model used by banks at present will change in the next
few years. There are issues of input data, enough data, quality of data, and
how to mix objective data (eg. internal) with quality data. There are different
approaches being used. Basel’s Capital Requirement data must be quality data,
objective data and weights of these elements. There is still not a lot of
internal bank data because banks are still not able to produce objective data. Quality
data exercises can be challenged by insurance institutions. One must haggle
from businesses what the frequency/qualitative data is. Qualitative data uses
triggers such as quality budget. So after a few months of collecting bank data,
they can analyse to see if it is working.
Certain types of risks are increasing. One can examine the
frequency of how a risk might occur. Banks can use scenarios to train
management. - scenarios such as the
Capital scenario. This is operation risk at the beginning of the process,
beginning to understand what the real operation risks are. There are two ideas
at present on this. One is using only high quality special data; the process is
the Pareto analysis.
Limitations of this type of analysis are that conversion to
Pareto distribution seems not to be the definitive solution because it needs a
picture of the global severity distribution. The other problem is – what do you
do when there is not enough data?
The Bank Capital can be calculated but the risk profile
needs to be known. Also the risk in the particular bank. Data used now is joint
external and scenario data to extract intelligence, but these carry big
problems of operation risk.
Regulators are
cautious about correlations. How can future modelling assuage regulators fear?
In the market situation, mostly mathematical calculation is used but it should
be used with great caution because when real data is used in the model, the
results are surprising. For example, use of diversification causes capital
required to be calculated as half of that when another mathematical method is
used. To calculate losses, you need to use cell correlations and a 36- cell
regulatory matrix. Credit risk and operation risk are fundamentally different.
If fundamental data is something about causes, then the problem is what data
should be used in modelling for capital calculation?
For example, foregone revenue cannot be input into the
model and theoretically, this is 25% of the total calculation. Also since events
have multiple consequences, the model is not yet developed to match causes and
consequences. The solution might be to include back-to-book-transactions as per
calculating credit risk because consequences and causes are very well linked.
Operation risk is not a transaction. It is different to credit risk There is
also a difference between operation risk and business risk. Corporate culture
is one driver of operation risk. Operation risk is causal- it is based on correlative frequencies, eg.
Information Technology disaster recovery event, fraud and general disaster are
all items that can be stress-tested. For example flu pandemic is a human,
physical and economic risk and will have different impact on market, credit and
other areas.
Operation risk is basically “process management risk”.
People understand
it better than the concept of BASEL ACCORD.
To determine what
will happen in a bank, internal data is not enough to give a complete picture.
You need external data also. Mathematical theories are not useful here. The
model has to be bespoke to the individual bank. A model for losses in a big
bank cannot be used for a small bank because of scale. So you have to scale the
data. This is the most robust way of using external data. Experiments with
synthetic data converge better with the model. Moreover external data has
issues of quality. The method used must ensure that data is complete and is
classified correctly. Using a mix of internal and external data, needs the
external data to be scaled first to the profile of the particular financial
institution. The method must be able to trigger a review of “severity
parameter”.
The Revco case an
example.
Scenario analysis is difficult to perform because risk
factors are identified by several sources. So the solution in the future might
be, instead of a single scenario, a rectangle frequency from 5 to 20 years as
the range of variation is more realistic.
Some institutions have quality controls but the cost of
this and the complications might make it difficult to introduce this into the
operation risk model. Is capital an approximate proxy for risk assessment? A
lot is down to scenario analysis. Are
what are being measured the right items to be measured? Capital Requirement is
one of the proxies, because of regulations. But other measures are risk factor
and repetition capital model takes care of scenario analysis but capital model
does not include losses of the future. Liquidity risk on the other hand is
stress-tested.
There is cross-risk diversification- identifying which
portion of the operation risk is coming from the other risk, say credit risk.
There is also economic impact calculation for high-impact events, such as the
New York Trade Centre bombings. .There is post-operative risk, risk that has
never happened before. but can the model take into account an event
that has not yet occurred? This
can be done by producing stress analysis, using internal data and stress data,
and then get an expert opinion and by introducing possible losses and using
Monte Carlo simulation. These risks can be put into the economic capital
calculation and be Basel 2 compliant.
The European
Union Banks
At present the European Union banks produce Pareto
distribution quarterly and every five years. But using Pareto technique can lead
to very unstable distribution. There is
big over-estimation of capital requirement, sometimes by more than a country’s
Gross National Product. With the Pareto distribution technique, even a small
variation produces big changes in capital requirement. The solution may be to
merge different banks’ databases where the size of the bank and the operation
control environments are similar. But in practice this is unrealistic because
control environments may not be similar. Alternatively it might be possible to merge
directly, scaling down to similar size using gross margin or to calculate an
individual bank’s own scaling factor, applying different weights. Possibly,
then “capital at risk” might be calculated.
For instance, capital requirement today for Spanish banks is 5% of GDP,
but regulators say it should be 8% of Gross Domestic Product. At present the
calculation models used by banks, when they are used, can circumvent Basle
Accord regulations, if accepted by regulators, because they have no Audit
Trail.
A summary of Risk measurement and systemic risk
Research
on risk measurement and systemic risk-related issues, has progressed
substantially since 1995 by the value-at-risk (VaR) methodology. How risk could be quantitatively measured
and what the meaning of such measures are explained in several academic
papers.. In 1997, the Asian crisis erupted, triggered by and itself triggering
events that were beyond the bounds envisioned by standard Value at Risk
methodology. Systemic banking crises,
contagion and monitoring were issue to be addressed. The series of
banking and currency crises that emerged in various parts of the world during
the past two decades suggests that financial stability is not to be taken for
granted. The term “systemic risk” simply means the danger that problems in a
single financial institution might spread and such contagion could disrupt the
normal functioning of the entire financial system.
Banking stability and systemic crises
The
theory of banking based on liquidity risk sharing, with banks emerging as
providers of the required liquidity insurance showed how, under asymmetric
information, bank runs can emerge in such a fractional reserve banking system.
Allowing for the possibility of bank runs, the Diamond/Dybvig [1]
model is not able to explain the causes of banking crises. Extensions of the
model have introduced uncertainty about asset returns to proxy for the impact
of the business cycle on the valuation of bank assets. In these models with
aggregate shocks to asset returns, financial crises are driven by fundamentals.
Shocks to asset returns by reducing the value of bank assets. Despite its
widespread use in theoretically analysing financial instability, the model and
its various extensions do not provide a completely plausible description of
actual patterns of banking crises. In the latest version of the Allen and Gale
model a market for long-term assets is introduced into the analysis enabling
banks to liquidate these assets. Liquidation costs are therefore endogenous. As
a result, asset markets provide a transmission mechanism that serves to channel
the effect from the liquidation of assets by some banks to other banks in the
economy. If a sufficient number of banks are forced to liquidate their assets
and the demand for liquidity rises above a certain level, asset prices will
move sharply. This may, in turn, force other banks into insolvency and
exacerbate the original crisis.
As
a result, the model, compared with earlier theories, provides a more realistic
explanation of how and why financial crises may develop. It also highlights the
importance of asset market liquidity for the avoidance of financial crises. One
model is about the failure to recognise the role of interbank credit. In such a
model, banks compete in the loan market, while the interbank market serves as
an insurance mechanism against deposit withdrawals due to liquidity shocks.
Mergers affect bank balance sheets via increased concentration and potentially
enhanced cost efficiency, while also altering the structure of liquidity
shocks. The model highlights the importance of functioning interbank markets
for financial stability and sheds some light on potential trade-offs between
antitrust and supervisory policies.
Contagion across markets
Dungey
et al identified contagion by looking at daily movements in bond spreads in an
effort to quantify the effects of unanticipated regional shocks across borders.
The resulting contagion measure controls for common global shocks, country
specific shocks and regional factors. They find contagion originating from the
Russian default, with the measured level of the effect larger for emerging
economies. However, the proportion of total volatility attributable to
contagion varies widely across countries and is not always more substantial for
developing countries. Contagion effects are found to be widely distributed
across both developed and developing markets, making contagion a phenomenon reserved not only for developing
countries.
As
to contagion across markets and countries, despite its importance for financial
market stability, it remains less than completely understood. Contagion is at
the heart of any analysis of financial crises, because it is contagion that
makes the initial shock a truly systemic event. To understand financial sector
risks, one has to deal with the origin of these risks as well as the channels
of propagation. The increasing use of complex risk transfer instruments and
speed of financial market transactions add to the complexity and rapidity of
the potential propagation of shocks, making these risks difficult to gauge. Contagion can therefore be viewed as the
propagation mechanism that causes small idiosyncratic or systematic shocks to
have systemic consequences.
Systemic monitoring
Systemic
events can impose substantial social costs on the affected economies, as bank
runs; for example, will disrupt credit relations and efficiency, leading to non-trivial
direct and indirect effects in the form of output losses. Systemic monitoring
and the analysis of systemic risks are high on the policy agendas of central
banks Can market-based indicators be usefully employed to predict banking
fragility? Useful and well-behaved indicators can be derived from stock market
data but to date the focus has been much on subordinated debt spreads.
Market-based indicators are useful in predicting banking fragility.
Market-based indicators can be used in supervisors’ early warning models.
Exposures to domestic interbank
market
To
monitor counterparty exposures in the domestic interbank market, a model known
as the Riksbank model uses data detailing the largest un-collateralised
exposures of the four major players in the Swedish banking system by exposing a
proxy for the Swedish banking system, the four biggest banks, to solvency
shocks originating from outside the interbank market and assessing how the
system is affected. The results were that Swedish domestic direct contagion
effects are less than what might have been expected in the Swedish banking
system. Where one of the large banks is assumed to fail, other banks are found
not to suffer direct losses that would reduce their capital ratio significantly
below the regulatory level. Similar results were found for the risk of direct
contagion from abroad, which mainly arises from foreign exchange settlement
exposures.
Market liquidity
Banks
are providers of insurance for liquidity risk. They serve this function by following
a liquidity immunisation strategy, implemented via individual asset markets and
interbank credit markets, to guard themselves against the possible effects of
forced asset liquidation.
Study of the Treasury market
Individual
market participants are likely to underestimate potential price movements
resulting from shocks to markets and predictably underestimate the risk of
their own exposures. This is based on standard market microstructure models.
Basel 2001 indicators are constructed to evaluate how liquidity has evolved
since the 1997 Asian banking crisis and
to examine the determinants of changes in liquidity. The analysis showed that,
having deteriorated during the Asian and Russian financial crises, market
liquidity has broadly recovered to pre-crisis levels. Basel 2001 estimators are now commonly used to assess the
tails of return distributions.
The Enron fraud
The
Enron fraud indicates that the pendulum may not be very far from swinging back
between the two extremes of very pervasive public intervention and complete
laissez-faire. And public intervention is needed to support the orderly
functioning of financial markets.
The
increasing complexity of the financial system renders it more and more
difficult to identify the origin of
risks. Managing the risks associated
with the uncertainty and risk of the real sector is at the core of financial
intermediation. The way in which risk is spread within the financial system
varies over time in relation to several factors, including market and
regulatory developments. Enron fraud
highlight the importance of two aspects
that characterise risk propagation today, the growing use of complex financial
instruments to assume and transfer risks and the abrupt changes in international
capital movements. Evidence suggests that the markets for credit risk
transfer instruments are concentrated in terms of dealers and ultimate
risk-takers. As to abrupt changes in international capital movements, lack of
data on many important players in the global financial system is the black hole
on the possible sources of destabilising capital movements.
Strong public intervention is necessary on those occasions when
markets fail to work properly. This should not be confused with a wide and pervasive
intervention in the markets as public authorities used to do in the past.
.Those responsible for the oversight of markets should signal their commitment
to well-defined and effective intervention, when needed, and so contribute to
the stability of the financial system.
ENDS
click below to
[1]
D Diamond
and P Dybvig, “Bank runs, deposit insurance, and liquidity”, Journal of Political Economy 91, 1983, pp 401-19.