Quantitative Methods for Finance

Chapter 14
Quantitative Valuation Applications

Post Graduate Program in Finance
Quantitative Methods for Finance • Post Graduate Diploma

Quantitative Valuation Applications

Introduction

Quantitative valuation converts expected cash flows, market prices, interest rates, volatility and time into estimates of value and risk. In finance, valuation is rarely a matter of applying one formula mechanically. A sound valuation requires a coherent set of assumptions about discount rates, compounding, default, liquidity, exercise behaviour, market completeness and numerical approximation.

This chapter develops applications for fixed-income securities, term structures, equities and derivatives. It emphasizes duration and convexity, yield-curve construction and modelling, binomial trees, risk-neutral valuation and option sensitivities. It also examines numerical accuracy and model risk, because a mathematically precise answer may still be economically unreliable when its assumptions are inappropriate.

Learning Objectives

By the end of this chapter, the learner should be able to:

Value coupon bonds and zero-coupon instruments using discount factors.
Explain spot rates, forward rates and the term structure of interest rates.
Calculate and interpret Macaulay duration, modified duration and convexity.
Apply yield-curve bootstrapping and critically evaluate curve models.
Value equity and derivative claims using one-period and multiperiod binomial trees.
Explain risk-neutral valuation and its relationship with no-arbitrage pricing.
Interpret delta, gamma, theta, vega and rho.
Identify numerical, modelling, liquidity and governance risks in valuation practice.

1. Discounting and Fixed-Income Valuation

A security is valued as the present value of its future cash flows. If cash flow Cₜ is received at time t and the appropriate discount factor is D(0,t), then:

V₀ = ∑ₜ₌₁ⁿ CₜD(0,t)

Here, V₀ is current value, n is the number of payment dates, and D(0,t) is the price today of one unit of currency paid at time t. The formula assumes that cash flows are known and that the discount factors are appropriate for their timing and risk.

For a zero-coupon bond with face value F and maturity T, if the continuously compounded spot rate is s(0,T), then:

P(0,T) = Fe^(−s(0,T)T)

If annual effective compounding is used instead:

P(0,T) = F ÷ (1 + s(0,T))^T

The compounding convention must be stated. A nominal annual rate convertible m times per year gives:

P(0,T) = F ÷ (1 + yₘ ÷ m)^(mT)

where yₘ is the nominal annual yield and T is measured in years.

Coupon Bond Valuation

For a bond with coupon C paid annually, face value F, maturity n years and a single annual yield y, the price is:

P = ∑ₜ₌₁ⁿ C ÷ (1 + y)^t + F ÷ (1 + y)ⁿ

This expression assumes a flat yield applicable to all cash-flow dates. In practice, each cash flow should generally be discounted using its maturity-specific spot rate:

P = ∑ₜ₌₁ⁿ CₜD(0,t)

A bond trades at par when its coupon rate equals the relevant yield under the same compounding and credit assumptions. It trades at a premium when the coupon rate exceeds the required yield and at a discount when it is lower.

Worked Example 1: Coupon Bond

Consider a three-year bond with face value 1,000, annual coupon rate 6%, and annual yield 5%. The annual coupon is 60. Its value is:

P = 60 ÷ 1.05 + 60 ÷ 1.05² + 1,060 ÷ 1.05³

The components are approximately 57.14, 54.42 and 914.29. Therefore:

P ≈ 1,025.85

The premium reflects the coupon rate of 6% exceeding the required yield of 5%.

2. Term Structure of Interest Rates

The term structure describes the relationship between interest rates and maturities for instruments with comparable credit and liquidity characteristics. A yield curve may be upward sloping, downward sloping, flat or humped. It is not merely a forecasting device: it is also an essential input into arbitrage-free valuation.

Spot and Forward Rates

The spot rate sₜ is the annualized return associated with a zero-coupon investment from today to time t. A one-year forward rate beginning at time 1 and ending at time 2, under annual compounding, satisfies:

(1 + s₂)² = (1 + s₁)(1 + f₁,₂)

Thus:

f₁,₂ = (1 + s₂)² ÷ (1 + s₁) − 1

Forward rates are implied by current prices; they are not necessarily unbiased forecasts of future short rates. They may include risk premia, liquidity effects and convexity effects.

Bootstrapping

Bootstrapping derives discount factors from observed prices of coupon-bearing instruments. Suppose a one-year zero-coupon bond with face value 100 has price 96.15. Its discount factor is:

D(0,1) = 96.15 ÷ 100 = 0.9615

If a two-year bond pays 5 at the end of year 1 and 105 at the end of year 2 and has price 98.00, then:

98.00 = 5(0.9615) + 105D(0,2)

Therefore:

D(0,2) = [98.00 − 4.8075] ÷ 105 ≈ 0.88755

The two-year annually compounded spot rate is:

s₂ = (1 ÷ D(0,2))^(1 ÷ 2) − 1 ≈ 6.15%

Bootstrapping requires consistent payment dates, clean treatment of accrued interest, reliable market prices and a specified interpolation method between observed maturities.

Yield-Curve Models

Curve construction and curve modelling serve different purposes. Bootstrapping creates discount factors that reproduce selected market instruments. Parametric models impose a smooth functional form and are useful for interpolation, scenario generation and risk management.

Common approaches include:

Polynomial or spline methods: flexible, but potentially unstable outside observed maturities.
Nelson–Siegel models: represent level, slope and curvature factors.
Svensson models: extend Nelson–Siegel with an additional curvature component.
Short-rate models: model the evolution of an instantaneous short rate, as in Vasicek or Cox–Ingersoll–Ross frameworks.
Forward-rate models: model rates associated with future periods directly, including Heath–Jarrow–Morton-type approaches.

A useful curve model should fit market prices, avoid economically implausible behaviour, permit stable sensitivities and remain usable under stress. Perfect historical fit is not sufficient: overfitting can make risk estimates unstable.

3. Duration and Convexity

Bond prices generally move inversely to yields. Duration approximates the sensitivity of price to a change in yield, while convexity improves the approximation for larger changes.

Macaulay Duration

For annual cash flows, Macaulay duration is the weighted average time to receipt of the bond’s cash flows:

D_M = [∑ₜ₌₁ⁿ tCₜ ÷ (1 + y)^t] ÷ P

where t is measured in years, Cₜ is the cash flow at t, y is the yield per period and P is the bond price. The numerator has units of currency-years and division by price leaves years.

Modified Duration

For annual compounding:

D_mod = D_M ÷ (1 + y)

For a small change ∆y in yield, the percentage price change is approximately:

∆P ÷ P ≈ −D_mod∆y

Here ∆y is expressed as a decimal, not as a percentage-point label. A 50-basis-point increase is ∆y = 0.005.

For a yield compounded m times per year, a common form is:

D_mod = D_M ÷ (1 + y ÷ m)

where duration is usually expressed in years after appropriate conversion of period timing.

Convexity

For annual compounding, price convexity is:

Convexity = [1 ÷ P]∑ₜ₌₁ⁿ [t(t + 1)Cₜ ÷ (1 + y)^(t + 2)]

The second-order approximation is:

∆P ÷ P ≈ −D_mod∆y + ½Convexity(∆y)²

Convexity is normally positive for a plain option-free bond. It means that the price increase caused by a yield fall is larger than the price decrease caused by an equal yield rise. Callable bonds and other securities containing embedded options may exhibit negative or changing convexity.

Worked Example 2: Duration Approximation

Suppose a bond has price 980, modified duration 4.2 years and convexity 28 years². If its yield rises by 50 basis points, ∆y = 0.005. The duration-only estimate is:

∆P ÷ P ≈ −4.2(0.005) = −0.021

Thus the estimated price change is −2.1%, or approximately −20.58 currency units. Including convexity:

∆P ÷ P ≈ −0.021 + ½(28)(0.005)²

∆P ÷ P ≈ −0.02065

The estimated price is approximately 980 × 0.97935 = 959.76. The convexity adjustment reduces the estimated loss because the bond’s price-yield relationship is curved.

Duration is reliable mainly for small, approximately parallel yield shifts. For non-parallel shifts, key-rate duration is more informative. Key-rate duration measures sensitivity to a change at a particular maturity point while holding other curve points constant under a specified methodology.

4. Equity Valuation

Equity valuation links expected shareholder cash flows to required returns. The dividend discount model for a finite stream is:

P₀ = ∑ₜ₌₁ⁿ Dₜ ÷ (1 + kₑ)^t + Pₙ ÷ (1 + kₑ)ⁿ

where Dₜ is the dividend at time t, kₑ is the required return on equity and Pₙ is the terminal value at time n.

Under a constant-growth perpetuity, if dividends grow forever at rate g and kₑ > g:

P₀ = D₁ ÷ (kₑ − g)

where D₁ is the dividend expected one period from now. The condition kₑ > g is essential; otherwise the perpetuity does not converge.

The required return is often estimated using the capital asset pricing model:

kₑ = r_f + βₑ[E(R_m) − r_f]

where r_f is the risk-free rate, βₑ is equity beta and E(Rₘ) − r_f is the expected market risk premium. This model is useful as a framework but is sensitive to beta estimation period, benchmark choice and assumptions about the risk-free rate and market premium.

5. Binomial Trees for Derivative Valuation

A binomial tree approximates the evolution of an underlying asset through discrete up and down movements. It is especially useful for American options, whose exercise may occur before maturity.

Let S₀ be the current underlying price. In a one-period model:

S_u = S₀u, S_d = S₀d

where u > 1 is the up factor and 0 < d < 1 is the down factor.

For a non-dividend-paying asset, the risk-neutral probability is:

q = [(1 + r) − d] ÷ (u − d)

for one period when r is the risk-free rate per period. With continuous dividend yield δ and continuously compounded risk-free rate r, a common form is:

q = [e^((r − δ)∆t) − d] ÷ (u − d)

The no-arbitrage condition is:

d < e^((r − δ)∆t) < u

This ensures 0 < q < 1. If the condition fails, the assumed tree permits arbitrage.

For a European derivative with terminal payoffs Vᵤ and V_d, its current value is:

V₀ = e^(−r∆t)[qVᵤ + (1 − q)V_d]

With annual effective discounting, e^(−r∆t) is replaced by 1 ÷ (1+r).

Replicating Portfolio Interpretation

Risk-neutral valuation is equivalent to constructing a portfolio of ∆ units of the underlying and borrowing or lending an amount B such that:

∆S_u + B(1 + r) = V_u

∆S_d + B(1 + r) = V_d

Solving the two equations gives:

∆ = (V_u − V_d) ÷ (S_u − S_d)

and the portfolio value ∆S₀ + B equals the derivative value. This demonstrates that risk-neutral probabilities are a pricing device derived from no-arbitrage, not a claim that investors are literally risk-neutral.

Worked Example 3: One-Period European Call

Assume S₀ = 100, u = 1.2, d = 0.8, strike K = 100 and one-period risk-free rate r = 5%. The terminal stock prices are 120 and 80. The call payoffs are:

C_u = max(120 − 100, 0) = 20

C_d = max(80 − 100, 0) = 0

The risk-neutral probability is:

q = (1.05 − 0.8) ÷ (1.2 − 0.8) = 0.625

Therefore:

C₀ = [0.625(20) + 0.375(0)] ÷ 1.05 ≈ 11.90

The replicating delta is:

∆ = (20 − 0) ÷ (120 − 80) = 0.5

The call therefore has a hedge ratio of 0.5 shares per option in this one-period model.

American Options

For an American option, the value at each node is the greater of immediate exercise value and continuation value:

V = max[exercise value, discounted expected continuation value]

For a call on a non-dividend-paying stock, early exercise is generally not optimal under standard assumptions. For puts and calls on dividend-paying stocks, early exercise may be rational. A binomial tree handles this feature naturally by comparing the two values at every node.

6. Option Sensitivities

Option sensitivities, commonly called Greeks, quantify how a derivative value changes when a model input changes. They are local measures and depend on the valuation model.

Sensitivity	Meaning	Typical interpretation for a long call
Delta, ∆	∂V ÷ ∂S	Value change for a small change in underlying price
Gamma, Γ	∂²V ÷ ∂S²	Change in delta for a change in underlying price
Vega	∂V ÷ ∂σ	Value change for a one-unit change in volatility convention
Theta, Θ	∂V ÷ ∂t	Time-passage effect, subject to sign convention
Rho, ρ	∂V ÷ ∂r	Value change for a change in interest rate

For a small movement in several inputs, a first-order approximation is:

∆V ≈ ∆∆S + ρ∆r + Vega∆σ + Θ∆t

A more accurate approximation may include gamma:

∆V ≈ ∆∆S + ½Γ(∆S)² + ρ∆r + Vega∆σ + Θ∆t

The units must be clear. For example, if vega is quoted per 1 percentage-point change in implied volatility, ∆σ = 1 means a one-percentage-point movement in that quotation convention, not a unit increase from 20% to 120%.

Delta is used for directional hedging, gamma for monitoring hedge instability, vega for volatility exposure, theta for time decay and rho for interest-rate exposure. A delta-neutral portfolio can still have substantial gamma or vega risk.

7. Numerical Accuracy and Model Risk

Quantitative valuation involves several layers of approximation:

Market-data error: prices, curves, volatilities and dividends may be stale, thinly traded or inconsistent.
Calibration error: model parameters are estimated rather than known.
Specification error: the model may omit stochastic volatility, jumps, liquidity or credit dependence.
Discretization error: a finite tree or numerical grid approximates a continuous process.
Interpolation error: values between quoted maturities or strikes depend on the chosen method.
Implementation error: coding, compounding, day-count or sign mistakes can materially alter value.

Good practice includes independent price verification, sensitivity checks, convergence testing, stress testing and documented model validation. For a tree, increase the number of steps and observe whether the estimated value converges. If a ten-step and a thousand-step tree produce materially different results, the coarse approximation should not be used without investigation.

Numerical precision should not be confused with valuation certainty. Reporting a price to six decimal places does not imply that the economic value is known to that precision.

8. Illustrative Case: Hedging a Corporate Bond Portfolio

Background

A treasury desk holds a portfolio of option-free corporate bonds valued at 50 million currency units. Its effective modified duration is 6.4 years. The desk expects a parallel upward shift of 25 basis points in the relevant government and credit curves.

Problem

Estimate the immediate market-value impact and determine whether duration alone is adequate for decision-making.

Analysis

Using the duration approximation:

∆P ÷ P ≈ −D_mod∆y = −6.4(0.0025) = −0.016

The estimated loss is:

50,000,000 × 0.016 = 800,000

This estimate assumes a parallel shift, unchanged credit spreads, unchanged cash flows and a sufficiently small rate movement. In reality, corporate bonds have both interest-rate duration and spread duration. A government curve shift may differ from a credit-spread shift, and the portfolio may contain bonds with different maturities, coupons and embedded options.

Approach and Solution

The desk should supplement effective duration with:

key-rate durations across the curve;
spread duration and spread-stress scenarios;
convexity estimates;
liquidity and bid–ask assumptions;
issuer concentration analysis; and
hedge-instrument basis risk.

If the portfolio has positive convexity, the actual loss for a 25-basis-point rise may be slightly smaller than the duration-only estimate. However, callable bonds can have lower or negative effective convexity, so the adjustment cannot be assumed positive.

Outcome

The 800,000 estimate is a useful first-order risk measure, not a complete forecast of realized loss. A professional hedge decision should compare the portfolio’s key-rate and spread exposures with those of proposed futures, swaps or credit instruments.

Learning Points

Valuation is inseparable from risk measurement. A single duration number can conceal curve-shape risk, spread risk, liquidity risk and optionality. The appropriate level of model complexity depends on the decision, materiality and exposure structure.

9. Governance, Ethics and Contemporary Practice

Valuation outputs influence investment decisions, capital allocation, performance measurement, collateral calls and financial reporting. Deliberately selecting favourable inputs, ignoring stale prices or concealing model limitations can mislead users even when the calculation is technically correct.

Important governance practices include clear valuation policies, segregation of front-office and independent valuation functions, audit trails, model approval, exception reporting and periodic back-testing. Fair-value frameworks generally require an assessment of the observability and reliability of inputs; where market inputs are unavailable, judgement and valuation adjustments should be documented.

Technology improves valuation through automated curve construction, Monte Carlo simulation, algorithmic calibration and machine-learning-assisted data cleaning. However, artificial intelligence does not remove the need for economic reasoning. A machine-learning model may detect patterns in historical prices but can fail under regime change, illiquidity or unprecedented stress. Use of AI should therefore include explainability, data lineage, access controls, bias testing and human oversight.

Best-Practice Checklist

Match discount rates to cash-flow timing, currency, risk and liquidity.
State compounding, day-count and settlement conventions.
Use spot discount factors rather than an unjustified single yield when appropriate.
Check no-arbitrage conditions in trees and curves.
Distinguish market-implied forward rates from forecasts.
Use duration for small parallel shifts and key-rate measures for curve movements.
Include convexity when rate changes are material.
Test numerical convergence and independent implementations.
Reconcile model prices to observable instruments.
Document assumptions, limitations and valuation adjustments.
Treat risk-neutral probabilities as pricing constructs, not empirical probabilities.
Report sensitivities with clear units and quotation conventions.

Chapter Summary

Quantitative valuation expresses financial value as the discounted value of future cash flows or as the no-arbitrage cost of replicating a payoff. Bond valuation relies on discount factors, spot rates and consistent compounding. The term structure can be constructed by bootstrapping and represented through parametric or short-rate models, but interpolation and model choice create material risk.

Macaulay duration measures the weighted timing of cash flows, modified duration approximates percentage price sensitivity to yield and convexity improves the approximation for larger changes. Equity valuation uses expected dividends or cash flows and requires careful treatment of growth and required return assumptions.

Binomial trees provide an accessible framework for valuing European and American derivatives. Risk-neutral valuation is equivalent to replication under no-arbitrage and does not require investors to be risk-neutral. Option Greeks describe local exposures to price, volatility, time and interest rates, but they should be supplemented by stress tests and higher-order analysis.

Finally, valuation quality depends not only on formulas but also on market data, calibration, numerical convergence, governance and professional judgement. A precise model output is useful only when its assumptions are economically defensible and transparently communicated.

Key Terms

Discount factor: Present value of one unit of currency received at a specified future date.

Spot rate: Rate associated with a zero-coupon investment to a particular maturity.

Forward rate: Rate implied today for borrowing or lending during a future period.

Term structure: Relationship between interest rates and maturities.

Bootstrapping: Sequential extraction of discount factors or spot rates from market instruments.

Yield to maturity: Single rate that equates a bond’s price with the present value of its promised cash flows.

Macaulay duration: Present-value-weighted average time of a bond’s cash flows.

Modified duration: Approximate percentage price sensitivity to yield.

Convexity: Curvature of the price-yield relationship.

Key-rate duration: Sensitivity to a change at a particular maturity on the yield curve.

Risk-neutral probability: Probability used for no-arbitrage valuation after adjusting expected growth to the risk-free rate.

Replication: Construction of a portfolio with the same payoff as a target claim.

Binomial tree: Discrete up-and-down model for an underlying asset or interest rate.

American option: Option exercisable at any time up to expiration.

European option: Option exercisable only at expiration.

Delta: First-order sensitivity of derivative value to the underlying price.

Gamma: Sensitivity of delta to the underlying price.

Vega: Sensitivity of derivative value to volatility.

Model risk: Risk of loss arising from inappropriate assumptions, implementation or use of a valuation model.

Review Questions

Short Answer Questions

What is the difference between a spot rate and a forward rate?
Why does a bond with a coupon rate above its required yield trade at a premium?
Define modified duration and explain the sign of its yield sensitivity.
What is the economic interpretation of a risk-neutral probability?
Why can a delta-neutral portfolio still be exposed to significant risk?

Long Answer Questions

Explain how bootstrapping derives discount factors from coupon-bearing securities, including the importance of payment conventions and interpolation.
Derive the duration-plus-convexity approximation and discuss when it may be unreliable.
Compare polynomial, spline, Nelson–Siegel and short-rate approaches to yield-curve modelling.
Explain how a one-period binomial model supports both risk-neutral valuation and replication.
Discuss the main sources of numerical and model risk in derivative valuation.

Analytical / Application Questions

A five-year bond has modified duration 4.6 and convexity 31. Estimate its percentage price change for a 75-basis-point yield increase using both duration alone and duration plus convexity. Interpret the difference.
Construct a one-period binomial valuation for a put with S₀ = 100, u = 1.25, d = 0.80, K = 105 and a risk-free rate of 4%. Show the terminal payoffs, risk-neutral probability and present value.
A portfolio has duration exposure concentrated at two-year and ten-year maturities. Explain why a parallel-shift hedge may fail and design a more appropriate key-rate risk assessment.

Multiple Choice Questions with Answers

Question 1

A three-year bond has annual cash flows of 60, 60 and 1,060 and a yield of 5% with annual compounding. What is its approximate price?

A. 960.00

B. 1,000.00

C. 1,025.85

D. 1,080.00

Correct Answer: C. 1,025.85

Question 2

If the one-year spot rate is 4% and the two-year spot rate is 5%, what is the implied one-year forward rate from year 1 to year 2 under annual compounding?

A. 5.00%

B. 6.01%

C. 6.25%

D. 9.00%

Correct Answer: B. 6.01%

Question 3

A bond has modified duration 5. If its yield rises by 20 basis points, the approximate percentage price change, ignoring convexity, is:

A. −10.0%

B. −1.0%

C. −0.1%

D. +1.0%

Correct Answer: B. −1.0%

Question 4

For a one-period binomial model with u = 1.2, d = 0.8 and a risk-free rate of 5%, the risk-neutral up probability is:

A. 0.375

B. 0.500

C. 0.625

D. 0.800

Correct Answer: C. 0.625

Question 5

In a standard one-period binomial model, the no-arbitrage condition requires:

A. u < 1 + r < d

B. d < 1 + r < u

C. 1 + r < d < u

D. d < u < 1 + r

Correct Answer: B. d < 1 + r < u

Question 6

For a one-period European call with terminal payoffs 20 in the up state and 0 in the down state, and risk-neutral up probability 0.625, with a 5% risk-free rate, the value is approximately:

A. 9.52

B. 11.90

C. 12.50

D. 20.00

Correct Answer: B. 11.90

Question 7

Which measure captures the curvature of a bond’s price-yield relationship?

A. Macaulay duration

B. Modified duration

C. Convexity

D. Yield to maturity

Correct Answer: C. Convexity

Question 8

Under the constant-growth dividend discount model, which condition is necessary for convergence?

A. g > kₑ

B. g = kₑ

C. kₑ > g

D. kₑ = 0

Correct Answer: C. kₑ > g

Question 9

Which Greek measures the sensitivity of an option’s value to volatility?

A. Delta

B. Gamma

C. Vega

D. Rho

Correct Answer: C. Vega

Question 10

A principal limitation of a duration-only estimate for a corporate bond portfolio is that it may omit:

A. Face value completely

B. Credit-spread and curve-shape risk

C. The existence of coupons in every case

D. The bond’s currency by definition

Correct Answer: B. Credit-spread and curve-shape risk

Quantitative Methods for Finance
Chapter 14 of 15

Chapter 14 Quantitative Valuation Applications