September 4, 1999
Corporate Performance Measures: An Integrated Risk Management Approach for Financial Institutions
By
Thomas S. Y. Ho PhD
President
Thomas Ho Company
Tom.ho@thomasho.com
A. Challenges to the Financial Industry in Japan
Japanese banking community, with financial safety-net provided by The Financial Securities Agency
including 7.5 trillion yen capital injection of tax-payers’ money to the top 15 banks, seems recovering from
near collapse of financial system last year. Moreover, as three major banks announced their intention for
integration, the oversupply in Japanese banking sector may move toward rationalization.
However, the Japanese financial community, which includes not only banks but other financial sectors like
general and life insurance industries, still faces are several and serious challenges. Many of the problems
are balance-sheet related where low profitability is one of the most critical issues.
On the investment (asset) side, interest rates remain historically low, with depressed real estate value for
the past decade, and poor equity performance over a ten year period. On the liability side, competition is
intense without significant growth in markets. Furthermore, the industry is facing more regulatory changes
and requirements. For example, BIS new proposal requiring new risk management initiatives and
adequacy of capital.
In this environment, financial institutions must take remedial actions. For example, we may seek higher
yielding assets, often with higher risks. On the corporate level, we may seek additional capital infusion.
We may focus more on the profitability of the products and not only on market shares. Life/ general
companies and regional/city commercial banks are different in their experiences; each company has its own
unique challenges and responses. But, by and large, we must explore new approaches to succeed in this
adverse environment.
In this paper, I propose a new approach. This approach is to implement an integrated corporate
performance measure. The proposed solution is not a tactical method that seeks immediate, but often
temporary, positive results. Nor is the solution that would seek relief from government agencies. The
proposed methodology addresses the core issue in risk management of financial industry.
Today, there is a sense of urgency to introduce Value at Risk and other risk measures to monitor our
financial health. Concurrently, new accounting initiatives are introduced to ensure further reporting
transparency. These are positive developments. But are we ready for the challenges described above?
I think not. International banking risk measure standards are designed for transparency of the banking
solvency and not for internal management of financial institutions. Value-at-Risk, Earning-at-Risk, Cash
flow-, Liquidity-, and many other at-Risk measures are after all statistical measures of uncertainties of
financial institutions’ solvency. These measures can at best partially capture the risks and evaluate the
performance of the financial institutions. Often they are not applicable for the internal management of a
financial institution. If we do not improve our internal management, there is no externally imposed risk
measures that can assist us to overcome the challenges.
I propose we must first recognize that financial services companies are in the business of “risk transform”.
A well-designed process has to be in placed to utilize investors’ capital to transform risks from borrowers
to lenders. Regional banks transform saving deposits to loans and mortgages. Commercial banks
transform certificate of deposits to corporate loans. Life insurance companies transform life policies and
annuities to capital market investments. General insurance companies transform property and casualty
policies to capital markets. These transformations must be implemented by well-designed processes. To
ignore the risks of these processes in managing the risk of a financial institution is as meaningless as to
evaluating the Toyotas using all the standardized risk measures, ignoring the process of building a car.
Risk management of a financial services company should begin with evaluating the process of the risk
transform and then measure the risk and performance of each phase of the process. Performance of a
financial institution should then be measured by consolidating the performances of all the major
departments of the process adjusted for their risks.
The new approach is to exploit the latest financial technologies that have experienced tremendous growth
over the past twenty years and apply them to financial services industry. More specifically, I propose to use
this technology to develop a set of accurate, relevant, consistent measures of performance of all the major
departments. These measures should be able to hold each department accountable for their actions. And
all their benchmarked actions will be consistent to a common firmwide objective. In so doing, financial
institutions will be able to perform in an economically rational fashion. This is particularly important
because our industry is rapidly becoming more driven by global capital markets.
The purpose of this paper is to describe the proposed corporate performance measures. I will discuss how
such a set of measures can be implemented, and why other performance measures used to date fail to fulfill
our needs. In Section B, I will first describe the proposed general approach using “process engineering” as
a framework. In Section C, I will describe the assumptions underlying the model of the business process of
an insurance company or a bank. In Section D, I will describe the modeling of an insurer or a commercial
bank showing how financial technologies should be used for corporate management. In Section E, I will
show why VaR may not be an appropriate measure of risk for a financial institution and will define Earning
at Risk as a natural extension of VaR. Section F will then describe the performance measures of the major
departments. Section G is the conclusion, suggesting what actions that we can take in light of the progress
of recent research.
B. What are the Corporate Performance Measures?
The process for financial institutions focuses on the specification of roles and responsibilities to all the
departments participating in transforming the financial risks. The precise design of the process of course
should depend on the particular institutions’ business mission, culture and other management issues. In
general, the process should have four cycle phases: specification, design, test and implementation phases.
Senior management is responsible for the specification phase. They must specify the limits of the risk
exposure and the expected returns of the shareholders’ equity or the embedded value (the asset economic
value net of the liability economic value) of the financial institution given the supply and demand of funds
in their market place. Asset and liability Committee is responsible for the design phase. They are
responsible for the planning of the growth and structure of the balance sheets that would meet the
requirements specified by senior management. The portfolio management of an insurer or the treasury of a
bank is responsible for the test phase. In this phase, current market prices and liquidity are used to ensure
that the target balance sheets and profitability can be realized and that the design is consistent with market
reality, the realizable transaction and transaction prices. Finally, the line businesses and trading are
responsible of implementations, where loans are made, deposits are collected and securities are traded.
Performances are then reported to the senior management for the feedback control, and thus completing the
process cycle. A more detail discussion of such a process is beyond the scope of this paper. Ho (1995,
1999) discusses this process in more detail.
The central function of process engineering is to design and specify in more detail the above mentioned
process. Financial modeling is used to describe this process and this modeling provides the following
performance measures:
 Identify the balance sheets and the income statements that incorporate of all the phases ( a specific
example is given below.)
 Measure risk and performance of each phase against its benchmarks.
 Attribute the institutions’ performances and risks to each phase of the process.
While performances are in general additive, risks are not. Indeed, not only risks are diversifiable (and hence
not additive), risks are often cross hedged across different business lines. Therefore risk attribution must
take these issues into account to assure coherence in the analysis.
I propose that the role of risk management is the design, monitoring, and management of this process.
Risks are monitored by the performance measures along the entire process using quality assurance
techniques, like Pareto charts, Ishikawa diagrams and control charts.This approach enables us to identify
the sources of risks, before the risks lead to large losses. The main success of process engineering is to
identify the potential problem by measuring the problems when they are small. For example, in the Leeson
case at Barings. If we can identify the market exposure and measure the losses earlier, then maybe we
would stop the process before the really large loss comes about. You can do what you like with respect to
deleting or explaining more.
In sum, this process approach provides an integrated framework of risk management to the institution. By
way of comparison to using risk measures like value-at-risk, earning-at-risk, and others, the proposed
approach of this article has three important attributes:
(1) Bottom up
The methodology ties information from the transaction level to the corporate goals, from trading decisions
to corporate strategic decisions. In contrast, to date, there is no consistent framework to relate value-at –
risk and other risk measures to the corporate goals and management actions. For example, some proposed
that the VaR be measured in economic values while the Earning at Risk (EaR) be measured in book values.
These measures are therefore not consistent with each other.
(2) Process Oriented
The methodology focuses on the process of risk transform in the financial institution. To date, value-at-risk
and other measures focuses on the risks of shareholders’ equity and returns only. Without identifying the
process within the institution, such risk measure fail to identify the sources of the risks. The sources of risks
are not confined to market risk versus credit risks, but the sources of risks in the risk transformation
process.
They includes: line of business risks (mortality/morbidity risks, disaster risks, depositors withdrawal risk
and many others.), liquidity risks, the timing of inflow and outflow of funds for the institution, asset and
liability mismatch on the structural level for the balance sheet. Senior management, asset liability
committee, treasury (portfolio management), lines are departments that are identified to manage part of
these risks.
(3) Consistency
The methodology ensures consistency of risk measures. In contrast, Value at Risk and other measures may
not be consistent with each other. For example, Value at Risk is economic value based while Earning at
Risk is book value based. They do not readily tie to a set of consistent corporate actions.
In comparing with GAAP accounting approach, this proposed methodology is forward looking, based on
future stochastic cash flows, capturing future uncertainties in the measure We use capital market
observation to provide a systematic and consistent present value calculation. In contrast, GAAP accounting
must be backward looking using deferred acquisition cost, amortization of good will and other accrual
basis. Further, the measures of the risks tend to be rule based and not continually calibrated to the market
realities.
Some insurers begin to use cash flow testing (simulations of future free cash flows of the insurer under
different scenarios to enable the senior management to simulate future events. But we must note that cash
flow testing, while a useful tool, is limited in its application. The method enables us to identify our future
scenarios on a run off basis under our own, often, subjective set of assumptions. Since the results are
somewhat removed from the process of internal management, the optimal solution to enhance the financial
institution’s value is often less apparent.
C. The Assumptions of the Basic Model
This section will present the basic model of the business process for a life company (commercial bank).
The model is also applicable to other financial services company. There are many possible extensions to
this basic model. These extensions will be left for future research. The purpose of this paper is to propose
the basic model that may capture the salient features of the corporate performance measures viewing
financial services as a process.
We will make the following assumptions about the life insurer (commercial bank) and the capital market:
1.
Asset valuation
For clarity of exposition, we here assume that the asset portfolio consists of only marketable securities,
although the approach does allow loans and other non tradable assets in general. All assets are tradable
and have market prices. This is a simplifying assumption enabling us to focus our discussion on the
corporate measures without being distracted by complicated issues of valuing non-marketable
securities.
2.
Liability valuation
We assume that the insurer (bank) sells guaranteed investment contracts (GIC)(time deposits). There is
no other type of liabilities such as annuities (savings deposits), corporate short term and long term
debts. For transfer pricing purposes, the GIC (time deposits) are discounted by the government bond
rates to determine the internal economic value of the liability payments.
3.
Organizational structure of the insurer (bank)
The insurer (bank) has five departments: senior management, asset and liability management (ALM
Committee), portfolio management (Treasury), and line business, and risk management. The
responsibilities of each department is given below:
(1)Senior management: for the operations of the insurer and setting the insurer’s (bank’s) performance
targets; Senior management would include the management committee that represents the
stakeholders’ interests.
(2)ALM: for determining the asset and liability structure. Asset and liability management include also
the risk management, for the purpose of this paper (This seems conflicting since now you have risk
management as an independent department)
(3)Portfolio management (Treasury): for investments. Investments includes asset allocation, sector
rotation and securities evaluation and trading. For most insurers, portfolio management is separated
into trading and other functions. The proposed methodology can be used for drilling down to such
disaggregated levels.
(4)Line business: for the sale of GICs (time deposits and purchase of loans). These are term products
with a single payment. This product is investment-like and is used for this paper for simplicity without
being distracted by mortality, morbidity and other product risks.
(5)Risk management: as mentioned above is responsible for the management of the process.
4.
Insurer’s (bank’s) objective function
The insurer’s (bank’s) objective is to increase the shareholders’ equity value or the embedded value
( asset economic value net of the liability economic value). Also, this assumption is made to simplify
the analysis and exposition. All insurers (banks) have to be concerned with regulatory solvency (or
capital adequacy) tests, financial reporting, and franchise value (present value of all future earnings)
beyond (One can argue that in theory present value of all future earnings should be the asset economic
value minus liability economic value. Can you show me examples of this “beyond” net economic
value.) the company’s embedded value. The paper will show that the model can be extended to deal
with these issues. Incorporating all the other objectives and constraints to the model, while important
to the implementation, will add much complexity to the exposition of this paper.
5.
Reporting Period
The model can be used in a multi-period context. For clarity of exposition, we will present the model
as a one period model. The period is the reporting period which maybe one month or three months.
The model will be used in a prospective basis when the model is used for risk management. At the
same time we will also use the model in a retrospective basis when the model is used for performance
measures.
While these assumptions are unrealistic for an insurer (a bank), I believe such simplifications will
enable us to study the process of insurance (commercial banking business) more effectively.
D. Financial Modeling of the Life Company (commercial bank)
We will present the model using the approach of the financial statements. As opposed to the financial
statements of GAAP, these financial statements are determined on an economic valuation basis. To
date, financial research enables us to report these economic-value based (present value based) financial
statements. This model provides a set of internally consistent measures, using a double entry
approach, and the model captures the salient features of measuring risk and performance. Further, the
model can depict the business process.
1. Balance sheet equation
S = A - L
(1)
S: the shareholders’ equity or the surplus (or often called the embedded value )
A: asset market value
L: liability economic value, where the value is based on the run-off business. Similar to the cash
flow testing methodology (simulations of future cashflows), we will use only the run-off business.
The reason for this assumption of ignoring the future sales growth is that we can use “option”
valuation approach.
There is an extensive literature in the area of valuation of liabilities using this option valuation
approach. The approach this paper proposes has been explained elsewhere, see Reitano (1997),
Girard (1996) and more recently Grosen and Jorgensen (1999). In essence, this approach provides
a systematic methodology to determining the present value of all the simulated cash flows over a
set of scenarios. One aspect of much discussion of the valuation of liabilities is the spread
required in the discounting. This model assumes that the insurer (bank) establishes a transfer
pricing methodology (see Wallace (1997) for a discussion of transfer pricing) and the model
assumes no spread off this transfer pricing curve. In essence, we assume a gain-on-sale
methodology (recognizing the all gains at the time of sale and not deferring the gains over time.)
However, the model does not have to use the gain-on-sale method in general.
2. Income statement equation
y = pv + rAA - rLL - e - t
(2)
y: after tax income to the insurer (bank) on a total return basis, taking the capital gain (realized and
unrealized) and income into account . This income is not the realized income on the accounting
basis. This income is the added value to the firms embedded value over the reporting period.
p: profit margin of the GIC (time deposits), where the profit margin is defined as the expected
proceed over the reporting period net of the economic value (of time deposits) divided by the
current economic value (of time deposits). The amount received is ( 1 + p ) v. v is the sale
volume over the reporting period ( a month or a quarter), measured by the economic value. Unlike
the balance sheet equation (equation (1) ) where the liability is considered only to be the run off
business, for the income equation, we incorporate the growth of new sales. Therefore, the income
equation brings in the franchise value of the insurer.
rA: total return of the assets, where the realized and unrealized capital gain and all the cash flow
received during the reported period are taken into account. Total return is a standard performance
measure for many investment styles. This model seeks to maintain consistency of measures
between asset management and liability management. Total return approach has been discussed in
more detail in Ho, Scheitlin and Tam (1995) . One important aspect that this paper incorporates is
the sales growth model.
rL: total return of the liabilities, where the liabilities are marked to market and the total return is the
change in market value together with all the liability payouts during the reporting period. Payouts
would include for life insurance the benefits and the cash value withdrawals. For time deposits, it
will be all the interests payout and withdrawals. Note that the modeling is based on cash flows.
Therefore when an incident (or early withdrawal) occurs during the reporting period, the incident
will increase the total return of the liability. The increase is the present value of the future
insurance (time deposits) payments plus expenses. This approach does not make the distinction
between earned and unearned premium. The model only concerns with the expected cash flows
and current portfolio holdings. Premiums (or time deposits) that have been received as inflows
have no effect on the total returns of the liabilities.
e: the operating expenses for the period. This paper focuses on the financial theory of risk
transform and not on the operational aspects of the insurance (commercial banking) business.
While expenses are central to the performance of an insurer, life or general (a commercial bank),
the model will assume that these expenses are certain and predictable, in order not to distract the
purpose of the paper.
t: the tax payments for the period. Taxes are often based on accounting financial statements.
Therefore the tax payments amount is not the product of the marginal tax rate and the income
based on total returns. We assume that there is a tax model that derives the tax amount.
3.
Retained earnings statement equation
S = y - d
(3)
S: the change of the surplus (embedded value)
d: the dividend payout. This model does not solve for the optimal dividend policy. We will
assume a certain policy that the insurance company (commercial bank) follows.
4.
Sources of funds equation
L = rLL + f
(4)
f: net funding, where
f = b - l + v
(5)
b: new borrowing, and this includes loan or bond issuance
l: paydown of liabilities, benefit payments (early withdrawal or deposit withdrawal) that leads to a
fall in the market value of the liability by the same amount.
v: new volume sales
The increase of liability is viewed as a source of external funding. When the expected total return
of the liability is high, the model interprets that the cost of funding is high. When the cost of
insurance (or deposit) increases unexpectedly, the liability value increases. This is not interpreted
as an increase in sources of funds but an unexpected increase in the cost of funding.
5.
Uses of funds equation
A = rAA + f + pv - e - d - t
(6)
Uses of funds equation is the corresponding equation to the sources of funds on the asset side.
Note that this model ensures that the double entry approach is maintained. That is, the change of
the surplus is the change of asset net the change of the liability value.
E. Value at Risk ( VaR ) and Earning at Risk ( EaR )
Recently there is much discussion of VaR as a risk measure for the balance sheet of a financial institution.
In essence, VaR focuses on the balance sheet equation (equation (1)), measuring the surplus value risk.
However, such a risk measure has limited value for the structural balance sheet of a financial institution.
This is because a financial institution does not necessarily focus on the risks exposure over a short term.
On the contrary, the structural balance sheet should be concerned with the longer term, taking the sales
(deposit-taking), investment (loan) returns and other factors into consideration.
For this reason, we can use the income statement equation (equation (2)) to determine the EaR. The
distribution of the income (y) given the parameters to the equation can determine the EaR. To date, EaR
measures are determined by the accounting methods. Therefore, EaR and VaR are not comparable. In
using equation (2) to determine EaR, we have extended the VaR measure to incorporate earnings to the risk
measures
F.
Corporate Performance Measure
The model can be used on a retrospective basis for measuring performances of each department on the
process of the insurance (commercial .banking) business.
This is accomplished by setting up asset benchmark returns ( r A*) and liability benchmark returns ( rL*).
The benchmark returns are the returns of portfolios (loans or deposits) based on the average performance
determined by the senior management. For the assets, this is often accomplished by using some broad
based market index , tilted to reflect the desired risk exposure of that asset and liability management view.
Similarly, the liability benchmarks are determined by the liability modeling (as described by option pricing
modeling, mentioned above) without assuming significant superior in knowledge and information of the
line of business. The performance of the ALM department depends on the views that the ALM department
takes and their views are reflected by the benchmarks that they establish for each reporting period.
Therefore, their performance is measured by the difference of the returns of the asset and liability
benchmarks. Specifically, we have:
y (ALM) = rA*A - rL*L
(7)
Performance of the portfolio management (y(PM)) is measured by the expected return of the asset portfolio
net the expected returns of the benchmark on the prospective basis. For return attribution, on the
retrospective basis, the performance would be the realized returns of the assets net the realized returns of
the liabilities. Specifically, we have:
y (PM) = rAA - rA*A
(8)
Performance of the line business (y(LB)) is measured by the profits they generate from the new sales and
their management of the liabilities in their performance against the benchmarks.
y ( LB) = pv + rL*L - rLL
(9)
Then we can now specify the corporate performance measure by noting that:
y = y (ALM) + y (PM) + y (LB) - e - t
(10)
Senior management’s role is to ensure that the income (y) will enhance shareholders’ value by managing
the process and ensuring the net income (y) meets the shareholders’ expectations.
It is important to note that this paper proposes a set of performance measures. I do not suggest that
management compensations should be directly related to these measures, though, these measures can be
part of the inputs. This paper is also not proposing a management system dealing with the human resource
issues. This paper focuses on the process engineering aspect of the risk transform and control for an
insurer.
G. Conclusions
Process engineering is central to risk management for financial institutions as it is for manufacturing
companies. Meeting international banking risk measurement standards (such as BIS capital adequacy
rules) enables us to satisfy (minimum capital or) solvency requirements, but it does not enable us to
manage the risk and performance of our business. Process engineering offers a process design approach,
with bottom up analysis in a consistent framework
This process engineering approach in determining corporate performance has immediate implications to
actionable decisions. First, the method enables a financial institution to measure economic performance in
a more transparent fashion. This approach can assist financial institutions to secure external funding for
risk capital.
Second, by incorporating future sales to the analysis, unlike cash flow testing method, we can better
identify the insolvency scenarios. Our planning can be more realistic. Also, this model is calibrated to the
market value to returns, which is the normal market returns. For example, it should be a spread off the
market yield. The simulations will be less subjective than what will be accomplished by some cash flow
testing
Third, by establishing a process, a financial institution can better manage risks that cannot be hedged. By
focusing on establishing the process of evaluating risks and controlling risk, the investment or treasury
department can have a procedure to evaluate higher yielding investments (loans or assets).
In essence, to date, we should exploit quality control process for financial institutions to manage our capital
in order to meet our current challenges in our industry. As our manufacturing companies used quality
control in the sixties to compete globally, our financial industry should also adapt quality control to
compete globally at the turn of the millennium.
Today, Japanese financial services industry may be facing unprecedented challenges of dramatic changes.
This may also be the great opportunity for the financial services industry to rise up to the challenge as the
manufacturing industry did several decades ago.
References
Grosen, Anders and Peter Jorgensen “ Fair valuation of life insurance liabilities: the impact of interest rate
guarantees, surrender options, and bonus policies” 9th International AFIR colloquium proceedings 1999
Girard, Luke N., “Fair valuation of liabilities – are the appraisal and option pricing methods really
different?” Risk and Rewards, no 25 March 1996
Ho, Thomas “A VaR model of an investment cycle: Attributing returns and performances” North American
Actuarial Journal , January 1999
Ho, Thomas; Marco Stalder and Bernhard Straub “ An integrated approach to managing risk and
performance” Research Paper 1998
Ho, Thomas “ Quality based investment cycle: toward quality assurance in the investment industry” The
Journal of Portfolio Management, volume 22 number 1, Fall 1995
Ho, Thomas, Alex Scheitlin, Kin Tam “ Total return approach to performance measurement” E.T. Altman
and I. T. Vanderhoof, editors, The Financial Dynamics of the Insurance Industry, Burr Ridge, IL, Irwin,
1995
Reitano, Robert R., “ Two paradigms for the market value of liabilities”, North American Actuarial
Journal, Vol 1, Number 4, October, 1997.
Wallace, Marsha “Performance measurements using transfer pricing..” Transamerica Insurance Company
Research Paper 1997
Furthermore, the new risk management and capital adequacy issue, which BIS's new proposal raised in
June, will become the new challenge. Japanese banks have been passively reacting to BIS's capital
adequacy requirements in the past. The new BIS proposal and probably ensuing tighter regulations will
demand international banks to keep abreast to new risk technologies against market volatility, which
sometimes get violent.
(will delete from here to the end of this color)
Unlike in the past, Japanese banks have to actively counteract the BIS regulations, otherwise there is no
guarantee to keep profitability and competitiveness in globalized financial market even after succeeding in
financial recovery.
The new BIS proposal is quite different from the past approach to the 'risk' in two areas;
1. the one is the expansion of 'risk' concept. In addition to credit risk, regulated by the first BIS rule in
1988, and market risk set by BIS in 1997, the new proposal demands capital allocation to other risks such
as liquidity risk, operational risk.
2. the other is swinging back towards qualitative approach of risk nature from quantitative one. This change
reflect somewhat lost confidence of statistical methodology to risk by regulators.
In fact, risk management methodology is now at the juncture. The VaR method, which BIS recommended
particularly for market risk measurement and control, failed to prove its usefulness after Asian and Russian
financial crises. VaR is a probabilistic number which says , for example that in 99 days out of 100 days, the
financial losses a bank suffers will be within such number. However at the crash, the loss a bank sees is
much higher number which could be, from statistical point of view with the probability of less than
0.0001%.
Much worse, if a bank controls its risk limit by VaR number, as traders’ risk position far exceed their VaR
limit, they are forced to reduce their position at the market with little liquidity, which cause the bank much
bigger realized loss. Credit Var is similar to Greenbank (or Drake's) universe equation, which is the
formula to calculate how many intellectuals exist in the universe by multiplying several parameters.
However, each parameter relies on many volatile assumptions and unreliable or insufficient data. Other
risks like operational risk are still at the stage of argument of their definitions.
Due to the current limitation of these quantitative approach to the risk management, regulators including
BIS are balancing back to qualitative risk management by calling for fortification and cooperation of risk
monitoring functions within as well as outside the banks.
Although its effort should be appreciated, it is questionable how effective these efforts will be because the
levels and purpose of these risk monitoring functions are so different, and may cause confusions and
disruption of bank's smooth operation.
Lastly, even the excellent risk management does not guarantee the return, the other side of the coin. What
we require now is the coherent framework for qualitative and quantitative risk management.
The Process Engineering we propose here is one of such methodologies to respond to the current
requirement of risk engineering.(delete)