Bonds historical simulation value at risk

J. Beleza Sousa, M. L. Esquı́vel, R. M. Gaspar, P. C. Real

February 29, 2012

Abstract
Bonds historical returns can not be used directly to compute VaR
by historical simulation because the maturities of the interest rates
implied by the historical prices are not the relevant maturities at time
VaR is computed.
In this paper we adjust bonds historical returns so that the ad-
justed returns can be used directly to compute VaR by historical
simulation. The adjustment is based on the prices, implied by the
historical prices, at the times to maturity relevant for the VaR com-
putation.
Besides other features, we show that the obtained VaR values agree
with the usual market trend of smaller times to maturity being traded
with smaller interest rates, hence, carrying smaller risk and conse-
quently having a smaller VaR.

1 Introduction
Despite all criticisms [5], historical simulation is by far the most popular VaR
method [4].
It is well known that VaR computation, by historical simulation, of bond
portfolios differs in important ways from VaR computation of stock portfolios
[2]. Essentially, this is because the market historical prices of bonds imply
a historical term structure of interest rates with maturities that are not the
relevant maturities, at time VaR is computed. They are greater than the
relevant maturities at time VaR is computed because they correspond to
past times when the bond maturity was further away then it is when VaR is

1
computed. This moves away the possibility of using market bonds historical
returns, directly in VaR computation.
The popular method to overcome this issue of cash flow mapping in risk
factors, besides ignoring the portfolio specific VaR, being subjective, com-
plex [1] and using lots of information sources, ruins the objectivity and the
simplicity of the historical simulation method.
In this paper we develop a method of adjusting bonds historical returns
so that they can be used directly in VaR computations. The method is based
on computing the returns of the prices implied, by the historical prices, at
the times to maturity relevant for the VaR computation.
We show that the developed method provides results consistent with
the usual market observed trend, in which smaller times to maturity imply
smaller yields, carrying smaller risk and consequently having smaller VaR.
We also show that the developed method strongly preserves the market im-
plicit correlations between the instruments in the portfolio.

2 Time to maturity adjusted bond returns

Consider the VaR computation at day nV aR , with time horizon N days,
and confidence level α percent, of a portfolio with a zero coupon bond with
maturity T > nV aR + N and principal P . See the time line in Figure 1 for
a graphical representation of these instants. Clearly, the relevant maturities
for this VaR computation are T − nV aR and T − (nV aR + N ).
Following the general historical simulation1 method, the bond’s N days
market observed historical returns should be used to compute VaR. Denoting
by p(n), the historical price of the zero coupon bond, at day N < n ≤ nV aR ,
and denoting by HR(n, N ) the N days historical return at day n, defined as
in [3], the N days possibly overlapping historical returns are given by:

p(n)
HR(n, N ) = , n = N + 1, · · · , nV aR . (1)
p(n − N )
These market observed historical returns should be applied to the bond
market value at day nV aR as it follows:

p(n)
p(nV aR )HR(n, N ) = p(nV aR ) , n = N + 1, · · · , nV aR . (2)
p(n − N )
1
VaR historical simulation method is referred by some authors as non-parametric VaR.

2
 h istorica l
p rice s

0 n VaR n VaR + N T

tim e Hin d a ysL

Figure 1: VaR computation time line.

The resulting returns define an empirical distribution of possible N days

bond returns at time nV aR . The VaR should be the potential loss of the
1 − α/100 quantile of this empirical distribution.
The problem with this general approach is that the historical price se-
quence p(n) for 1 < n < nV aR , used in Equation (2), implies a sequence of
term structure daily compounded interest rates2 r(n, T − n), given by:
à ! 1
P T −n
r(n, T − n) = − 1. (3)
p(n)
But, the relevant maturities for this VaR computation do not occur in
the implied historical interest rates. This is why the bond’s historical returns
can not be used directly in VaR computation.
Nevertheless, the implied historical interest rates, for times 1 < n ≤ nV aR ,
provide future bond valuation at times nV aR and nV aR + N .
The future value at time nV aR < m < T of the bond, bought at time
1 < n < nV aR , at historical price p(n), is given by the valuation of the future
cash flow at maturity time, at daily compounded interest rate r(n, T − n),
2
Daily compounded interest rates are used because typical VaR time horizons are spec-
ified in days.

3
fixed by price p(n). The possibility of a default event is assumed to be
implicitly incorporated in the price p(n) itself. Denoting by f (m, n) the
future value of the bond at time nV aR < m < T fixed by the price p(n) at
time 1 < n < nV aR , f (m, n) is given by:
P P
f (m, n) = =³ ´ T −m . (4)
(1 + r(n, T − n))T −m P T −n
p(n)

In this paper we define the nV aR time to maturity adjusted N days histor-

ical return at day n, denoted by AHR(n, N, nV ar ), as the quotient between
f (nV aR + N, n) and f (nV aR , n − N ):

f (nV aR + N, n)
AHR(n, N, nV ar ) = , n = N + 1, · · · , nV aR . (5)
f (nV aR , n − N )

Substituting Equation (4) in Equation (5), the defined adjusted historical

return is given by:

³ ´ T −nV aR
P T −(n−N )
p(n−N )
AHR(n, N, nV ar ) = ³ ´ T −(nV aR +N ) , n = N + 1, · · · , nV aR . (6)
P T −n
p(n)

Note that the AHR(n, N, nV ar ) value if fixed by historical market prices

p(n) and p(n − N ), thus capturing the market changes between n − N and n,
while being adjusted to the VaR computation relevant maturities, namely,
T − nV aR and T − (nV aR + N ).
Our proposal is to replace the original historical returns of Equation (1)
by those of Equation (6).
Using this adjusted historical returns directly in the VaR computation,
the VaR is the potential loss of the 1 − α/100 quantile of the following time
to maturity adjusted empirical distribution:
³ ´ T −nV aR
P T −(n−N )
p(n−N )
p(nV aR ) ³ ´ T −(nV aR +N ) , n = N + 1, · · · , nV aR . (7)
P T −n
p(n)

4
3 Extensions
In this section we discuss the extension of the bonds historical simulation VaR
method developed in the previous section to other scenarios beside computing
the VaR for zero coupon bonds at the time the historical sequence of prices
end.

3.1 Coupon bonds

The extension to coupon bonds is straight forward. In order to compute the
future value of a coupon bond at time m, based on the market price of the
bond at time n < m, two differences from the zero coupon bond case arise:
1. the yield to maturity at time n is computed using the bond’s dirty price
and accounting for all future cash flows after time n;
2. the value of the bond at time m accounts for all future cash flows after
time m.
Then, the adjusted returns are defined by Equation (6) as in the case of
a zero coupon bond and the VaR is computing as the loss corresponding to
the quantile of the empirical distribution of Equation (7).

3.2 Adjusting for past values

Suppose a bond B has already expired and the issuer of the bond issues a
new bond, B1 , equal to bond B, i.e., with the same type, principal, maturity,
number of coupons, coupon rate (if applicable), etc.
Consider a portfolio that contains the bond B1 . The portfolio VaR with
time horizon N days is to be computed by historical simulation at day 1.
The only historical prices available from bond’s B1 issuer are those of bond
B.
The adjustment method proposed in this paper can still be applied to
adjust the historical prices of bond B to past times, times before the historical
prices were observed, namely, for day 1. The process is the same as for
future values: for each historical price compute the daily yield to maturity
implied by the historical price; than compute the bond’s value at a previous
time valuing the bond’s cash flows following the time considered, with the
implied daily yield to maturity; finally use the previous times values to get
the adjusted returns of Equation (5), and compute the VaR.

5
4 Application
In this section we illustrate the the usage of the bond adjusted historical
return of Equation (6), by computing the VaR of the simplest possible port-
folio, namely, a portfolio with a unique real zero coupon bond. We use a
sequence of real historical prices of an alive zero coupon bond and compute
the VaR at time the historical prices sequence ends.
We first present the used portfolio, than we detail the adjustment of single
historical return, and then we adjust all the available historical returns and
compute the VaR.
Additionally we illustrate the adjustment for past values by computing
the VaR at the time the historical prices sequence begins.

4.1 Portfolio
Consider the zero coupon bond, B, with principal P = 1000, maturing at
day T = 731 whose real market historical prices are in Figure 2. The prices
were obtained from a quote service that delivers market prices aggregated
from different dealers responsible for trading (market makers) this particular
bond.
The prices in Figure 2 imply the market observed term structure interest
rates represented in Figure 3. Recall from Figure 3 that each day corresponds
to a different time to maturity.
Figure 3 clearly shows the usual trend observed in the market, in which
smaller time to maturities are traded with smaller implied interest rates.

4.2 Adjustment of a single return

Consider the N = 10 days historical return at day n = 190 of bond B,
detailed in Table 1(a). The prices that determine this historical return are
highlighted in Figure 4 with the black circles. The maturities underlying this
historical return are 731 − (190 − 10) = 551 and 731 − 190 = 541 days to
maturity.
Now, consider the VaR, computed by historical simulation, at day nV aR =
372, with a time horizon of N = 10 days, of the portfolio containing the bond
B. The empirical distribution of the possible 10 days returns at 731 − 372 =
359 days to maturity must be used. In order to obtain this distribution we
adjust each of the historical returns with Equation (6).

6
 98

96

94
Price H% L

92

90

88
0 50 100 150 200 250 300 350
Da y

Figure 2: Real historical prices of a zero coupon bond with principal P =

1000 and maturing at day T = 731, as a percentage of the principal.

Da ys to Ma tu rity

731 681 631 581 531 481 431 381

4
Yie ld H% L

Yie ld H% L
3

0
0 50 100 150 200 250 300 350
Da y

Figure 3: Term structure annualized daily compounded interest rates, im-

plied by the historical prices of Figure 2, as a function of both time and time
to maturity.

7
 N = 10 days historical return at day n = 190
p(190)
p(190 − 10) p(190) p(180)
94.25 95.03 1.00828
(a)

Annualized daily YTM (%)

r(180, 551) r(190, 541)
4.001 3.499
(b)

Adjusted return for nV aR = 372

f (372, 190 − 10) f (372 + 10, 190) ff (382,190)
(372,180)
96.215 96.765 1.00571
(c)

Table 1: (a) N = 10 days market observed historical return at day n = 190.

(b) Days n = 180 and n = 190 annualized daily yields to maturity. (c)
Adjusted N = 10 days return at day n = 190, for VaR computed at day
nV aR = 372.

8
 98

96

94
Price H% L

92

Historica l p rice s

90 Da y 190 re tu rn p rice s

Fu tu re p rice s a t d a ys 372 a n d 382

88

0 50 100 150 200 250 300 350

Da y

Figure 4: The prices that determine the N = 10 days historical return at

time n = 190, the corresponding future prices at times nV aR = 372 and
nV aR + N = 372 + 10 = 382, along with the historical prices sequence. The
arrows represent future values.

9
 In the particular case of the prices highlighted in Figure 4 with the black
circles, the corresponding adjusted prices, with Equation (4), are highlighted
with the black squares.
Table 1 shows the market observed historical return at day n = 190, as
well as the corresponding adjusted return for VaR computed at day nV aR =
372. The annualized daily yields to maturity and future prices used to com-
pute the adjusted return are also detailed.
As it can be observed from table 1 the adjusted return is closer to one
than the historical return. This is in accordance with the trend observed in
Figure 3. Once the interest rates of smaller maturities tend to be smaller,
the returns at time to maturity 369 should be closer than one than those at
time to maturity 541.

4.3 Portfolio VaR

Consider the VaR, with a time horizon of N = 10 days and confidence level
α = 99%, computed by historical simulation, at day nV aR = 372, of the
portfolio containing the bond B.
The VaR is computed using the empirical distribution of the adjusted
returns of Equation (6). In order to obtain this distribution the adjustment
of the single return detailed in the previous section is repeated for all available
historical returns.
Figure 5 shows the real, market observed historical prices, and also, the
corresponding future prices, f (m, n) of Equation (4), at days m = nV aR =
372 and m = nV aR + N = 372 + 10 = 382. The future prices are plotted
as a function of the day n of the historical price p(n) which fixes the future
value f (m, n). The prices highlighted in Figure 4 with the black circles and
squares are highlighted again in Figure 5, but now plotted as a function of
n.
Figure 6 shows the sequence of the bond’s historical returns along with
the sequence of the corresponding adjusted returns computed from the future
prices of Figure 5. Figure 7 shows the respective histograms.
Finally, Table 2 shows the VaR value computed from the empirical dis-
tribution of the overlapping adjusted returns of Equation (7), along with the
possible loss corresponding to 1−α/100 quantile of the overlapping historical
returns empirical distribution of Equation (2), for comparison purposes. It
also shows the correlation coefficient between the original and the adjusted
returns.

10
 98

96

94
Price H% L

Historica l p rice s

92
Fu tu re p rice s a t tim e 372

Fu tu re p rice s a t tim e 382

90
Da y 190 re tu rn p rice s

Fu tu re p rice s a t d a ys 372 a n d 382

88
0 50 100 150 200 250 300 350
Da y

Figure 5: Historical prices, and the corresponding future prices at times

m = nV aR = 372 and m = nV aR + N = 372 + 10 = 382. The future prices
are plotted as a function of the time n, of the historical price p(n), that fixed
the future price.

1.04
Historica l re tu rn s

Ad ju ste d re tu rn s
1.03

1.02
Re tu rn

1.01

1.00

0.99

0 50 100 150 200 250

Da y

Figure 6: Historical returns and the corresponding adjusted returns for

nV aR = 372.

11
 Historical returns Adjusted returns

20

15

10

0.96 0.98 1.00 1.02 1.04

Figure 7: Historical returns and the corresponding adjusted for nV aR = 372

returns, histograms.

Time horizon Confidence level VaR Quantile Correlation

N = 10 α = 99% -9.576 -9.935 0.984

Table 2: Time horizon N = 10, confidence level α = 99%, bond B VaR, com-
puted at day nV aR = 372 by historical simulation using adjusted historical
returns.

12
 Time horizon Confidence level VaR Quantile Correlation
N = 10 α = 99% -20.291 -9.935 0.985

Table 3: Time horizon N = 10, confidence level α = 99%, bond B VaR,

computed at day nV aR = 1 by historical simulation using adjusted historical
returns.

As it can be observed from Figure 6, the adjusted returns are closer to

one than the historical ones. This can be observed again in Figure 7 where
the adjusted returns histogram is more concentrated towards one than the
historical returns histogram. This results in a VaR value smaller than the
corresponding loss of the 1−α/100 quantile of the possible historical returns.
Again, this result conforms with Figure 3 which shows a clear decreasing
trend in interest rate as time to maturity decreases.

4.4 Adjusting for past values

Suppose that the bond B has already expired and its issuer issues a new
bond, B1 , equal to bond B.
Consider the portfolio containing the bond B1 and the VaR computed
by historical simulation at day n = 1 with the historical prices of bond B,
showed in Figure 2.
Following section 3.2 the past values of Equation (4), f (m, n) with m =
1 ≤ n are used to compute the adjusted historical returns of Equation (6)
and the vaR is computed from the resulting empirical distribution. In this
section we illustrate this process by repeating the figures and the table of the
previous section, but now, for nV aR = 1.
It can be observed from Figure 11 that the adjusted returns are now
less concentrated towards one than the historical returns. This results in a
VaR value, showed in Table 3, which is now greater than the corresponding
loss of the 1 − α/100 quantile of the possible historical returns. Again,
this observation in accordance with Figure 3 and the fact that the time to
maturity at time VaR is computed, nV aR = 1, is greater that the times to
maturity at following times, namely, when the historical prices were observed.

13
 98

96

94
Price H% L

92

Historica l p rice s

90 Da y 190 re tu rn p rice s

Pa st p rice s a t d a ys 1 a n d 11

88

0 50 100 150 200 250 300 350

Da y

Figure 8: The prices that determine the N = 10 days historical return at time
n = 190, the corresponding past prices at times nV aR = 1 and nV aR + N =
1 + 10 = 11, along with the historical prices sequence. The arrows represent
past values.

14
 98

96

94
Price H% L

Historica l p rice s

92
Pa st p rice s a t tim e 1

Pa st p rice s a t tim e 11

90
Da y 190 re tu rn p rice s

Pa st p rice s a t d a ys 1 a n d 11

88
0 50 100 150 200 250 300 350
Da y

Figure 9: Historical prices, and the corresponding past prices at times m =

nV aR = 1 and m = nV aR + N = 1 + 10 = 11. The past prices are plotted
as a function of the time n, of the historical price p(n), that fixed the future
price.

1.04
Historica l re tu rn s

Ad ju ste d re tu rn s
1.03

1.02
Re tu rn

1.01

1.00

0.99

0.98

0 50 100 150 200 250

Da y

Figure 10: Historical returns and the corresponding adjusted returns for
nV aR = 1.

15
 Historical returns Adjusted returns

14

12

10

0.96 0.98 1.00 1.02 1.04

Figure 11: Historical returns and the corresponding adjusted for nV aR = 1

returns, histograms.

5 Conclusions
Bond historical returns can not be used directly to compute VaR by histor-
ical simulation because the maturities of the interest rates implied by the
historical prices are not the relevant maturities at time VaR is computed.
In this paper we adjust bonds historical returns so that the adjusted
returns can be used directly to compute VaR by historical simulation. The
adjustment is based on the prices, implied by the historical prices, at the
times to maturity relevant for the VaR computation.
The proposed method has the following features:

• The time to maturity adjusted bond returns are used directly in the
VaR historical simulation computation.

• VaR of portfolios with bonds can be computed by historical simulation

keeping the simplicity of the historical simulation method.

• The portfolio specific VaR is obtained.

• The VaR values obtained are consistent with the usual market trend

16
 of smaller times to maturity being traded with smaller interest rates,
therefore carrying smaller risk and having a smaller VaR.

• The only source of information used is the market, through the bonds
historical prices.

• The correlation between each bond return and the returns of the other
instruments in the portfolio is strongly preserved.

• The VaR for the desired time horizon is computed directly with no VaR
time scaling approximations.

We left for future work, the research of the mathematical properties of the
developed method, and backtesting the method with benchmark portfolios.

References
[1] C. Alexander. Market Risk Analysis, Value at Risk Models. John Wiley
& Sons, 2009.

[2] Gangadhar Darbha. Value-at-risk for fixed income portfolios a compar-

ison of alternative models. Technical report, National Stock Exchange,
Mumbai, 2001.

[3] John C. Hull. Options, Futures, and Other Derivatives. Prentice Hall, 7
edition, 2008.

[4] Christophe Pérignon and Daniel R. Smith. The level and quality of value-
at-risk disclosure by commercial banks. Journal of Banking & Finance,
34(2):362 – 377, 2010.

[5] Matthew Pritsker. The hidden dangers of historical simulation. Journal

of Banking & Finance, 30(2):561–582, February 2006.

17