You are on page 1of 12

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/319653443

Estimating Value at Risk for Interest Rate Risk Using the Variance Covariance
Method

Article · January 2013

CITATIONS READS

2 5,469

2 authors:

Agata Gemzik-Salwach Paweł Perz


University of Information Technology and Management, Poland Rzeszów University of Technology
56 PUBLICATIONS 119 CITATIONS 20 PUBLICATIONS 22 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Przedsiębiorstwa START-UP View project

International Science Exploration Event 2019 Seminar 7th-8th of November, 2019 in Rzeszow, Poland View project

All content following this page was uploaded by Agata Gemzik-Salwach on 12 September 2017.

The user has requested enhancement of the downloaded file.


ESTIMATING VALUE AT RISK FOR INTEREST
RATE RISK USING THE VARIANCE
COVARIANCE METHOD
Agata Gemzik-Salwach1
Paweł Perz2

Summary
This article presents the way of measuring interest rate risk by means of Value at Risk method. This method is
likely to occupy a special place among methods of quantifying interest rate risk, as, unlike its traditional
counterparts, it allows to take into account the risk of a situation in which yield curves do not shift parallel This,
however, requires applying a number of modifications to the methods of calculating VaR for other types of risk.
The article presents one of possible ways of solving this problem..

JEL classification: D8, C13, C22


Key words: risk, interest rate risk, Value at Risk method

Introduction
The issue of interest rate risk dates back to the mid-1970s, when the prevailing doctrine of
Keynesian economics was replaced with Monetarism and the supply of money was controlled
by means of changing interest rates. In addition, at that time, global financial markets have
witnessed a number of other new trends, such as globalization, liberalization, deregulation,
digitization, etc. Analyzing the changes that took place at that time, one cannot fail to mention
the collapse of the Bretton Woods currency system and the introduction of liquid currency
rates. All these phenomena account for rapid increase in the level of risk related to economic
activities, including interest rate risk. This tendency has been maintained and deepened till
present times.
In the above-described situation, the appropriate method of managing interest rate risk has
become a major issue. Risk management usually consists of three basic stages: identification,
quantification and risk steering. One of the methods that could be used for quantifying interest
rate risk is the Value at Risk (VaR) method. It was developed by banks, where interest rate
risk is a major problem, as operations based on interest rates are the foundation of banks’
activities.

1. Value at Risk as a measure of interest rate risk


Value at Risk is defined as the maximum loss of market value that a particular institution risks
in a certain period of time and with determined confidence (J. P. Morgan Bank, 1999). We
can sometimes find definitions of this measure as the highest loss which may be incurred in a
given period of time and at determined level of significance.3 One of the greatest advantages

1
Dr Agata Gemzik-Salwach, University of Information Technology and Management in Rzeszów, Chair of
Macroeconomics, agemzik@wsiz.rzeszow.pl
2
Dr Paweł Perz, Rzeszow University of Technology, Department of Banking and Finance, pperz@prz.edu.pl
3
This term contains certain internal contradiction as the definition of maximum loss does not function in
statistics, from which the reference to the level of confidence was adopted.
of this method is its universal nature. It allows us to calculate the level of various types of risk
and then to combine them so that the final number is calculated, reflecting the whole level of
risk that in a given period of time and at a determined level of likelihood a particular entity
faces.
Another area of research into Value at Risk refers to the possibility of using this method at a
further stage of risk management. There are concepts in which the VaR is a factor in making
investment decisions (Kiani 2011; Palomba, Riccetti 2012; Ślepaczuk, Zakrzewski, Sakowski
2011). The measures like RAROC or RORAC that use VaR to measure the risk-adjusted
profitability (see, e.g., Buch, Dorfleitner, Wimmmer, 2011) function within the conceptions
discussed above. These ideas could easily be extended to other similar measures, such as bank
interest margins, thereby increasing the set of risk management tools.
The Value at Risk method may occupy a special position among other methods of evaluating
interest rate risk, as it enables us to quantify this risk more precisely than other methods in a
situation when the shift of yield curve is not parallel. Traditional methods of measuring
interest rate risk are often based on an assumption that the correlation coefficient between
short- and long-term interest rates equals one, which leads to the conclusion that yield curves
move only parallel. This assumption, however, is not confirmed in reality, as instruments with
longer periods of maturity usually bear more risk. The VaR method allows us to abolish this
assumption and, as a result, to identify other than parallel shifts of the interest rate time curve.
The use of the Value at Risk method to measure interest rate risk, though, calls for application
of specific behavior, differing from that when quantifying other types of risk by means of the
Value at Risk method.

2. The variance covariance method


The most popular method of estimating value at risk is the parametric method of variance
covariance. It is based on two underlying assumptions: it is assumed that the growth in the
value of all risk factors of a bank portfolio is a normal distribution and that the changes in
portfolio value, with the exception of option instruments, depend linearly on the changes of
these risk factors. The combination of these two assumptions lead to the limit theorem of this
method saying that the variation distribution of a diversified bank portfolio is a Gaussian
distribution. Taking the assumption of the normal distribution for financial time series is not
always confirmed in practice, nevertheless, this method is widely applied and even
recommended to banks by their supervisors (Miłoś, 2011).
The process of estimating value at risk by means of the variance covariance method when
assessing interest rate risk is presented in figure 1.
Figure 1: VaR estimation for interest rate risk by means of variance covariance method

Determining
composition of
interest rate
portfolio

Identifying Decomposition of
knots on yield portfolio into cash Transformation into Determining VaR Calculating
curves and flows and yield curve knots in knots VaR for
calculating risk determining their portfolio
factors current value

Calculating
return on risk Assessing the level
factors of variation and
correlation

Introductory actions Proper calculation of value at risk

The processes of determining the size of risk exposure has been market by means of an ellipse, the rhombus
represents the stage of assessing variation, while activities joining these two parts have been presented in
rectangles.

Source: own elaboration on the basis of: Jorion P. (2003). Financial Risk Manager.
Handbook. Second Edition. New Jersey: John Wiley & Sons, pp. 374-376; Szafarczyk, E.
(2001). Metoda Value at Risk, Materiały i Studia, Zeszyt nr 132, NBP, Warszawa, p. 19.

The calculation of value at risk must always be preceded by determination of portfolio


composition and identification of risk factors it faces. Risk factor is defined as a parameter
whose value changes in financial markets and the consequence of these changes is the change
of portfolio value (Best, 2000). In case of estimating VaR concerning interest rate risk, we
usually have a higher number of variables that in other types of risk, as it requires modeling
changes of the whole time structure of interest rates.
Theoretically, when calculating maximum expected portfolio loss, we should assess the
variation of each particular item, which requires thorough knowledge of market interest rates
in particular, determined by dates of interest payments and instrument maturity, points of the
curve of interest rate time structure, variation of discounted factors determined on the basis of
appropriate length of time series of these parameters, and then taking into account all these
correlations between particular risk factors. In practice, such procedure often turns out to be
impossible or too labor-intense, due to too many risk factors generated by interest items,
which causes technical problems related to gathering, keeping and processing such a huge
amount of data. Moreover, we do not always have the possibility of obtaining relevant
information, as, for example, in case of new financial instruments there are no sufficiently
long series of historic prices. In view of the above, the application of the value at risk method
to measure the interest rate risk requires finding a method which will allow us to group
various elements of the portfolio and, in this way, reduce their number.
The yield curve may be represented by a relatively small number of points, called knots.
Unfortunately, there is no method which allows us to unambiguously identify these points,
apart from maturity dates of portfolio instruments, the basic premise may be characteristics of
variation and correlation ratios. In order to do so, we must study the structure and behavior of
yield curves and choose the places in which their properties change. Other premises allowing
us to make this choice are portfolio complexity and required precision of representation
(Szafarczyk, 2001). The more risk factors we distinguish, the more precise the result will be.
The number of vertexes of the yield curve should therefore be the function of the essence and
degree of interest rate risk an institution faces. In market practice, there is a tendency to
emphasize the significance of short-term interest rates over long-term ones. This is due to the
fact that the number of long-term operations is considerably smaller that the number of short-
term ones. Moreover, short-term operations are easier to manage.
However, the most frequently used criterion of choosing knot points on the yield curve is the
availability of data. Most databases, including the best-known and offered free of charge in
the Internet by investment bank J. P. Morgan, base of the RiskMetrics system, distinguishes
14 representative buy=out periods, such as: 1,3,6 and 12 months as well as 2, 3, 4, 5, 7, 9, 10
15, 20 and 30 years, and for these periods information is provided on a daily basis concerning
variation of return rates and prices of financial parameters and correlations between them (J.
P. Morgan Bank, 1996). The Basel Committee in its proposals on statutory capital related to
market risk, recommends breaking down the yield curve into at least six time periods, each
with different volatility. The Committee, moreover, states that banks with a high number of
items with different periods of maturity or involved in complex strategies of arbitration need
to determine a greater number of risk factors (Basel Committee, 1995). Moreover, each
currency has its own unique variation of interest rate4.
The next step following the establishment of knot points of the yield curve is to determine
their corresponding risk factors, which are the value of discounted factors, calculated on the
basis of market prices and rates, and then conducting calculation of changes of market risk
factors, defined usually as:

 Pi (t ) 
ri  ln   , (1)
 Pi (t  l ) 

where ri – logarithmic l-day return on i factor,


Pi(t) – price of i factor in t period,
Pi(t-l) – price of i factor in the period of l days before t.

The return on risk factor thus is a natural logarithm of the quotient of prices of factors noted
in two subsequent periods, while for calculating value at risk we usually use a day’s delay
(l=1).
The series of returns on interest rate risk factors obtained in this way is the basis of calculating
variation and correlation ratios. All transactions or positions of the bank which depend on
interest rates, that is bonds, bills, deposits, futures transactions, FRA, percentage and
percentage-currency swap contracts, bond options, cap, floor and related instruments and their
exotic combinations (for example swaptions FRA options, futures, etc.) must be presented as
a flow or set of cash flows, subordinated to proper knots of their corresponding yield curve.5
VaR models, assigning cash flows to vertexes of the yield curve, apply the principle of
preserving neutrality to risk, using the prior calculated measures of variation and
interdependence.

4
In case of the Polish zloty field curie, due to lack of treasury bonds with appropriately long time horizon, we
analyze a smaller number of knot points than for other currencies, usually these are periods of: 1 and 7 days, 1, 3
and 6 months, 1, 2 , 3 , 5 , 7 and 10 years.
5
The process of assigning cash flows to knot points of the yield curve is often called mapping.
Current values of flows, mapped to particular knots are ultimately aggregated in them and
constitute an element of exposure to particular risk factors (Gup, 1997).6 Combining this data
with variation and correlation parameters we may determine the size of value at risk in
particular knots on the yield curve and the total maximum expected loss of the bank due to
interest rate risk exposure.

3. Decomposition of interest instruments


Decomposition of financial instruments into basic assets allows us to concentrate on market
variables, not on positions occupied by particular instruments. The method assumes that the
whole portfolio may be presented as a synthetic set of basic instruments, representing
exposure to basic risk factors (Szafarczyk, 2001). Generally, the base instrument on which we
can construct nearly all debt positions is the zero-coupon bond.
In order to do so, we must distribute all instruments sensitive to interest rate changes which
are at the bank’s disposal, into cash flows which will be generated by them in the future. Each
position generates at least one cash flow whose size depends on the market interest rate. The
next step is to discount the actual size of distinguished future cash flows to the current value,
while the discounting factor is either yield curve used for assessing a given value or a new
yield curve is created, which is the outcome of listings of various interest rates and we refer
cash flows from all instruments to it7.
The basic and the simplest interest instrument is a fixed interest rate bond. It is determined by
the nominal value N, coupon rate k and n consecutive periods in which coupon interests are
paid out8. Figure 2 shows the presentation of such instrument by means of cash flows 2.

Figure 2: Cash flow in fixed interest rate bonds and annual periods of coupon payments
N(k+1)
Nk Nk Nk

0 t 1 2 3 n

time
Source: own elaboration
A fixed interest bond is an instrument which, during its “life” generates the cycle of coupon
payments in shape of the product of its nominal value and coupon rate (kN), while on the date
of its maturity a purchaser will also obtain additional return on invested capital. Each of the
distinguished cash flows can be realized separately and as a result a set of zero-coupon bonds
is created, each of which has the period before buyout equal to the period of interest payment,
while in case of nominal value, the period before buyout ends on the day of issuing bonds

6
This value corresponds to the net present value (NPV), determined for particular knots of the yield curve, thus
it is the sum of values of present flows assigned to particular knots, of which some or all may be negative.
7
This second approach greatly simplifies the reality, nevertheless it is quite popular as banks try to evaluating all
instruments by means of the same reference rates, such as zero-coupon rates.
8
With ordinary bonds it is assumed that the periods between payment of interest are of the same length, and the
year has 360 days. Alternatively, in some markets, the actual length of periods between coupon flows is used at
the same length of the year, for example, some participants of the British market, in order to achieve accuracy
taking into account the leap year, use the ACT/365 ¼ convention.
(Fabozzi, 2004). The value of bonds equals the sum of values of all its zero-coupon elements,
therefore each part of the notations below is a formula for the value of a particular flow9:

n j
Nk N Nk
PV 
(1  yn)
n

(1  yn) n
 
i  j (1  yi )
i
. (2)

The yi symbol in the above formula denotes yield rate, in an annual presentation, of an i-
period zero-coupon bond and is a key parameter, vital for determining the present value of
particular financial flows. If the calculation of the current value of a bond is not performed at
the moment of issuing it or just after payment of the coupon but, as it usually happens,
between coupon payments, the discount rate used must correspond with the time left before
the nearest payment:

n j
Nk N Nk
PV 
(1  yn  t )
n t

(1  yn  t ) n t
 
i  j (1  yi  t )
i t
, (3)

where t – the moment in which the calculation is made, belonging to the time range (0,1).

The bond price obtained by means of the above equations is known as gross price or
transaction price, and differs from its value calculated in time of bond issuance or coupon
payment, which gives us net price. Gross price allows us to take into account the size of
interest due but not paid out yet.
In case of variable interest bonds, we have an instrument whose coupon payments are
determined by the amount of capital N and coupon rate whose size depends on how the prior
established parameter is shaped. Due to this, at the moment of conducting an analysis we only
know the amount of the nearest coupon payment, and in order to determine the size of further
flows we use implied forward rates. The flow of cash payments generated by variable interest
bonds, using ordinary capitalization, discounted by means of zero-coupon bond rates can be
presented as follows:

n j
Nyj N Nfj , i
PV 
(1  yj ) j

(1  yn) n
 
i  j (1  yi  j )
i j
, (4)

where yj – j-period immediate yield rate of zero-coupon bond,


yi+j - i+j-period immediate yield rate,
fj- j-period term yield rate for investment which will start in a period of i years.

The value of cash flows due to a variable interest bond may be presented as nominal value of
this bond in a period equal to the nearest coupon payment. This principle, however, does not
work in case of calculating gross price, where we have the following formula:

N (1  y1)
PV  (5)
(1  y1  t )1t
The value of a variable interest bond may then, from the perspective of t time, be expressed as
t no coupon bond with maturity date equal to the payment of the nearest coupon and nominal

9
It can be noticed that the total price of bonds equals present net value (NPV) of future cash flows.
value equal the value of variable coupon bond increased by the value of the nearest coupon.
Similarly to bonds, other interest instruments may be treated. For example a swap transaction
may be presented as a combination of two bonds with fixed and variable interest rate, in one
of these instruments a short position is occupied, in the other – a long one. Both components
of the equation are calculated applying appropriate formulas for bonds.

4. Transforming cash flows into the map of risk factors


All cash flows generated by instruments sensitive to interest rate changes must be
subordinated to appropriate points representing the yield curve. In case when the maturity of a
single cash flow coincides with the maturity of a vertex of the yield curve, such flow should
be fully assigned to a particular risk factor. Most frequently though, we witness a situation
when the cash flow appears between maturity points which were selected as risk factors, and
the problem arises of how to allocate these monetary means. In this case we should divide the
present value of the cash flow between two nearest vertexes. As a result of the mapping
process, a single risk factor may represent hundreds or even thousands of particular elements.
There are a number of methods for establishing weights of divided cash flows. RiskMetrics,
when dividing cash flows, observes three principles (J. P. Morgan Bank, 1996):
1. To preserve unchanged market value – market values of the present sum of monetary
means before and after allocation must be equal.
2. To preserve unchanged market risk – portfolio variations of both items created after
dividing the original value into two new periods must equal variation of the previous
instrument.
3. Preserving the signs of cash flows – when assigning a cash flow to two different periods,
there are usually a number of possibilities of dividing it. The rule says that we should use
the solution which maintains the conformity of signs of transposed cash flows with the
sign of the cash flow before its division.
In practice the method recommended by RiskMetrics boils down to determining present value
of the original cash flow using interpolated zero-coupon yield rate and then determining
weights of separated cash flows on the basis of quadratic equation. Determining interpolated
zero-coupon yield rate is conducted on the basis of known amounts of zero-coupon yield rates
of the whole cash flow for two periods – the closest lower one and the closest upper one:

yp = âyd +(1-â)yg, 0â1 (6)

where yp – zero-coupon yield rate of a single cash flow,


yd – zero-coupon yield rate of the lower asset (corresponding with the knot preceding
the date of the original flow),
yg – zero-coupon yield rate of the upper asset (corresponding with the knot directly
following the original date of flow),
â – part of the distance between vertexes.

The â coefficient required in the above formula is calculated in the following way:

tg  tp
â= , (7)
tg  td

where tp – maturity period of a single cash flow expressed in years,


td – maturity period of the closest lower vertex expressed in years,
tg – maturity period of the closest upper vertex expressed in years.
The interpolated yield rate calculated in this way must be put into a proper equation
determining the present value of a cash flow appearing at that time.
The next step is to determine the variation of a risk factor of a single cash flow. This figure
can be obtained, just like the amount of Yield rate of zero-coupon bond, as a result of linear
interpolation of price variation measures for two vertexes lying in the closest vicinity of cash
flow:

σp = âσd + (1- â)σg, 0â1 (8)

where σp – variation of a single cash flow,


σg – variation of a lower vertex,
σg – variation of an upper vertex,
â – part of the distance between vertexes.

Finally, the proportions in which a single cash flow should be divided between two
neighboring risk factors, denoted by α and (1-α) can be determined using the variance formula
(Butler, 2001):

σp2 = α2σd2 + (1-α)2σg2 + 2α(1-α)σdσgpd,g (9)

where α – share (weight) of the cash flow assigned to the closest lower vertex,
1-α – share (weight) of the cash flow assigned to the closest upper vertex,
pd,g – correlation coefficient between upper and lower risk factors, corresponding to
the knot point of the yield curve.

This notation can be transformed so as to obtain the quadratic equation with α variable:

α2(σd2 + σg2 - 2δσdσgpd,g) + α(2σdσgpd,g - 2σg2) + σg2 - σp2 = 0 (10)

 b  b 2  4ac
The solutions of the equation are the roots   , while:
2a
a = σd2 + σg2 – 2σdσgpd,g,
b = 2σdσgpd,g - 2σg2,
c = σg2 - σp2.

When we have two solutions of a quadratic equation, the rule says we should choose the root
belonging to the 0,1 range.

5. Determining Value At Risk


After isolating the vertexes of the yield curve, assessing their variation and correlation,
conducting decomposition of the portfolio and assigning all cash flows to these parameters,
we can commence to determine the level of Value At Risk itself. This process is divided into
two stages: the first step consists in calculating the VaR value for all isolated risk factors and
then we conduct the ultimate calculation of the value at risk due to interest rate risk for the
whole bank.
The variance covariance method uses the parametrical approach, therefore the value at risk,
expressed as percentage value for a single risk factor, is determined on the basis of the
following equation:
VaR  Eksp  k    t (11)

where VaR – Value at Risk in an analyzed period,


Eksp. – value of exposure to risk,
k – the multiple of standard deviation around arithmetic mean, typical of a given
confidence level,
σ – standard deviation around arithmetic mean of a random variable with normal
distribution,
t – number of days in a period.

The method of calculating VaR both on the level of particular portfolios and the whole
institution is the same. It consists in expanding the above quoted formula so that it covers
correlations existing between particular pairs of risk factors and gives vertexes of the yield
curve appropriate weights. In a situation when we deal with interest instruments denominated
in foreign currencies, we should conduct such analysis separately for each interest market and
then examine correlations between all differentiated risk factors.
As the weights of risk factors we can assume the shares of the size of cash flows assigned to
particular risk factors, expressed in present values of the whole cash flow generated by
positions sensitive to interest rate changes which are owned by the bank, which can be
expressed as:

PVi
wi  n
, (12)
 PV
i 1
i

where PVi – present value of cash flows imaged to the i-th risk factor.

Putting the weights calculated in this way into the equation for value at risk, we must assign
appropriate signs to them. If the value of current receivables assigned to a particular risk
factor exceeds the value of current liabilities assigned to it, then the weight of this risk factor
has a positive sign, whereas of the present value of the cash flow of the risk factor is based
more on liabilities, then its weight should be given a negative sign.
VaR presented in absolute values, calculated for the whole bank can be written down by
means of the following equation:

n n n
VaRb  Eksp  k   wi 2i 2  2 wiwjijpij  t ,
i 1 i 1 j i
(13)

where VaRb – value at risk calculated on the level of the bank (portfolio),
i, j – numbers of isolated elements of pairs of risk factors,
σi, j – index of variation (of yield or price) of risk factors,
wi, wj – weights assigned to particular risk factors,
pij – correlation coefficient between two risk factors.

The basic parameter describing joint appearance of measurable features is covariance, which
can be determined by means of the following equation:
Covi,j = ijpij, (14)

therefore:
n n n
VaRb  Eksp  k  w
i 1
i i 2  2 wiwj cov ij  t ,
2

i 1 j i
(15)

Because k  wi  σi  t is equal to VaR size, for the i risk factor, this equation can be also
presented as:

n n n
VaRb  Eksp  VaRi 1
i
2
 2VaRiVaRjpij
i 1 j i
(16)

where VaRi, VaRj – the size of Value at Risk for all cash flows assigned to i and j risk
factors.

The elements of weight, time root and k parameter corresponding to the accepted level of
confidence, have been neglected in this formula, as they are reflected in the size of particular
values at risk.

Conclusions
Determining Value at Risk for interest rate instruments is not a simple process. The
considerations presented in this article have shown that we can modify the method to make it
useful in measuring interest rate risk. This requires adopting at various stages of calculation
different assumptions which influence the result obtained. In consequence, VaR calculated
even by means of the same method by various computer systems may differ considerably.
Taking advantage of the risk calculated in this way we should not forget the assumptions at
which it was calculated and we should treat these numbers as special black boxes.

Bibliography
Basel Committee (1995). An International Model Based Approach to Market Risk Capital
Requirements. Basle.
Best, P. (2000). Wartość narażona na ryzyko; obliczanie i wdrażanie modelu VaR. Kraków:
Oficyna Ekonomiczna.
Blake, D. (1994). Financial Market Analysis. London: McGraw - Hill Book Company.
Buch A., Dorfleitner G., Wimmmer M. (2011), Risk capital allocation for RORAC
optimization, Journal of Banking & Finanse, 11, 3001-3009.
Butler, C. (2001). Tajniki Value at Risk. Warszawa: K. E. Liber.
Fabozzi, F. J. (2004). Bond Markets, Analysis and Strategies. New Jersey: Prentice Hall.
Gup, B. E., Brooks R. (1997). Zarządzanie ryzykiem stopy procentowej. Warszawa: Związek
Banków Polskich.
Jorion, P. (2003). Financial Risk Manager. Handbook. New Jersey: John Wiley & Sons.
J. P. Morgan Bank (1999). Risk Management: A Practical Guide. New York: RiskMetrics
Group. Retrieved from http://www.riskmetrics.com/.
J. P. Morgan Bank (1996). RiskMetrics. Technical Document. New York: RiskMetrics
Group. Retrieved from http://www.riskmetrics.com/.
Kiani K. M. (2011), Relationship between portfolio diversification and value at risk:
Empirical evidence, Emerging Markets Review, 4, 443-459.
Miłoś, E. (2011). Wykorzystanie metody Value at Risk w estymacji ryzyka inwestycyjnego w
spółki branży metalurgicznej. Internetowy kwartalnik Finansowy e-Finanse, 7 (1), 39-51.
Retrieved from http://www.e-finanse.com/.
Palomba G., Riccetti L. (2012), Portfolio frontiers with restrictions to tacking error volatility
and value at risk, Journal of Banking & Finance,
http://www.sciencedirect.com/science/article/pii/S0378426612001422?v=s5
Szafarczyk, E. (2001). Metoda Value at Risk. Materiały i Studia, Zeszyt nr 132, NBP,
Warszawa.
Ślepaczuk R., Zakrzewski G., Sakowski P (2011), Investment strategies beating the market.
What can we squeeze from the market?, WNE Working Papers WP14(70).

Abstract
Estimation of Value at Risk for Interest Rate Risk by Using Variance Covariance Method
The article presents a way to measure interest rate risk by using Value at Risk method. This method has the
potential to occupy a special place among the methods for the quantification of interest rate risk, because unlike
its traditional counterparts, allow for the calculation of risk when the shift of yield curves aren’t parallei.
However, this requires using an array of variations to the calculation of VaR for the other types of risk. The
article shows one of the possible ways to solve this problem.

JEL classification: D8, C13, C22


Keywords: risk, interest rate risk, Value at Risk

View publication stats

You might also like