Value a plain vanilla interest rate swap. Explain fixed versus floating legs and how the…
- User
Value a plain vanilla interest rate swap. Explain fixed versus floating legs and how the swap rate is chosen so value equals zero at inception.
Assistant
This is a comprehensive question about interest rate swaps, which is a fundamental topic in fixed income and derivatives. Let me break this down systematically.
Plain Vanilla Interest Rate Swap Basics
Definition
A plain vanilla interest rate swap is an agreement between two parties to exchange interest rate cash flows on a notional amount:
- One party pays fixed-rate coupons
- Other party pays floating-rate coupons
- Both calculated on the same notional principal
- No exchange of principal (typically)
The Two Legs
Fixed Leg:
- One counterparty pays a fixed rate (determined at inception)
- Payment = Fixed Rate × Notional × Day Count Fraction
- Known with certainty at inception
- Acts like an investor buying a fixed-rate bond
Floating Leg:
- Other counterparty pays a floating rate (typically LIBOR, SOFR, or another reference rate)
- Payment = Floating Rate × Notional × Day Count Fraction
- Resets periodically (typically quarterly, semi-annually)
- Acts like an investor buying a floating-rate note
Valuation at Inception
For a swap to have zero value at inception, the present value of the fixed leg must equal the present value of the floating leg.
Detailed Valuation Framework
Let me set this up properly:
Fixed Leg Value:
Where:
- C = Fixed coupon (swap rate × notional)
- δᵢ = Day count fraction for period i
- zᵢ = Zero-coupon yield (forward rate) for period i
- tᵢ = Time to payment i
Floating Leg Value:
Where Lᵢ represents the floating rate for each period and the notional is implicitly valued at par.
The floating leg always equals par at inception because each future payment is set at the forward rate for that period. When discounted at those same forward rates, the cash flows equal the notional amount—the coupon rate and discount rate are identical, so the bond prices at par.
For the swap to have zero value at inception, the fixed leg must equal the floating leg. Since the floating leg equals notional, I can solve for the fixed rate by setting the present value of fixed payments equal to notional:
The denominator represents the fixed-rate annuity factor, also called DVO1. To illustrate, consider a 2-year annual swap on $100 million with spot rates of 3% and 4%. The discount factors are 0.9709 and 0.9246 respectively, giving an annuity factor of 1.8955. Dividing 1 by this factor yields a swap rate of approximately 3.99%. For the swap to have zero value at inception, the fixed leg payments must equal the floating leg value: S × Notional × (0.9709 + 0.9246) = Notional The floating leg is worth par at the start since it resets to forward rates. The fixed leg consists of annual payments of S times the notional amount, discounted at each period's rate. Solving for S gives us the swap rate that makes both sides equivalent. The present value of fixed payments should equal the notional amount. With discount factors of 0.9709 and 0.9246, the annuity factor is 1.8955. This means the swap rate S satisfies S × 1.8955 = 1, giving S = 52.75 basis points—but this seems too low. I'm realizing my error: I need to reconsider how the swap rate relates to the notional. The swap rate formula should be S × Notional × AF = Notional, where AF is the annuity factor accounting for day count fractions and discount rates. Let me recalculate with the proper setup. The fixed leg's present value equals the swap rate times the annuity factor: PV_fixed = S × Σ(δᵢ × DFᵢ). For the swap to be worth zero, this must equal the floating leg's present value, which equals par since floating rates reset at forward rates. Setting them equal gives S = 1/Σ(δᵢ × DFᵢ), assuming unit notional. So my annuity factor of 1.8955 should be in the denominator, not the numerator—that's where I went wrong. I need to complete the partial thought about the annuity factor representing present value, then move forward with the recalculation. The annuity factor of 1.8955 represents the present value of receiving $1 at each period, which means paying a fixed rate S for 2 years gives PV = S × 1.8955. Setting this equal to $1 yields S = 52.75%, which is nonsensical. This suggests a fundamental error in my setup rather than just using unrealistic rates. Let me restart with more plausible assumptions: a 5% one-year zero rate and 5.5% two-year rate on a semi-annual basis. Computing the discount factors for each half-year period: - DF₁ = 0.9759 - DF₂ = 0.9735 - DF₃ = 0.9479 - DF₄ = 0.9257 The annuity factor equals 0.5 × (0.9759 + 0.9735 + 0.9479 + 0.9257) = 1.9115, giving a swap rate of 0.5 / 1.9115 = 26.14%. This still feels elevated, so I should reconsider the day count convention for semi-annual coupons. For the fixed leg: S × 0.5 × 3.8230 = S × 1.9115. Setting this equal to 1 for equilibrium yields S = 0.5230 or 52.30%. I need to be more careful about whether rates are expressed as percentages or decimals, and recalculate with a clearer approach using the zero curve of 5% and 5.5% annual spot rates for a 2-year semi-annual swap. Converting to semi-annual rates: the 6-month rate is sqrt(1.05) - 1 = 2.470%, and the 1-year rate is 5%. Rather than continuing down this path with confusing conversions, I'll switch to a simpler setup with annual coupons and a straightforward yield curve to make the calculation transparent. For a 2-year swap with $100 million notional and zero rates of 3% and 4%, the discount factors are 0.9709 for year 1 and 0.9246 for year 2. The fixed leg value per dollar notional is S × 1.8955, where S is the swap rate. For the floating leg, I need to apply forward rates to each payment period. The floating leg is worth par at inception, so setting the fixed and floating legs equal: S × 1.8955 = 1.00, which gives S = 52.75%. But this seems too high given the yield curve is only 3-4%. Let me reconsider the formula—the swap rate should satisfy the condition that the present value of fixed payments equals the present value of floating payments, both discounted at the zero rates. I think I'm making an error in how I'm setting up the equation. Actually, I see the issue now. The fixed leg annuitizes the notional payment stream, not the notional itself. The present value of a series of payments is the payment amount multiplied by the sum of discount factors. So if I want the fixed leg to equal $1 in present value, I need to solve for the payment that makes this work. Using the discount factors I calculated (0.9709 and 0.9246), the swap rate comes out to about 52.8% annually. But that's clearly wrong for a 3-4% yield curve—I must have made an error in computing the discount factors themselves. Let me recalculate them from the actual spot rates. Actually, checking my work: DF₁ = 1/1.03 = 0.9709 and DF₂ = 1/(1.04)² = 0.9246 are both correct. The issue isn't the discount factors. Working through the payment structure, if I need to repay $1 through two equal annual payments S, the present value of those payments must equal $1. That gives me S = 1/(0.9709 + 0.9246) = 1/1.8955 ≈ 0.528, or 52.8% per year. This rate is genuinely what the math produces, even though it seems counterintuitive given the modest yield curve. But wait—I'm realizing the swap rate shouldn't be expressed as an annual percentage of the notional in this way. In actual swap markets, a 4% fixed rate means you pay 4% of the notional each period, not that the payment itself is 4% of the principal. The 52.8% figure I calculated represents something different: it's the annualized payment rate needed to equalize present values, not the conventional swap rate quotation. The real swap rate should align much more closely with the yield curve, somewhere around 3.5% given the 3% and 4% spot rates. Let me recalculate properly. If S is the swap rate in decimal form (so 0.04 for 4%), the fixed leg payments are S times the notional amount each year. For a $100M swap, that's $S × $100M annually. The present value of these payments is $S × $100M × 1.8955, which must equal the $100M floating leg value. Solving for S gives me the correct swap rate. But I'm getting 52.75%, which seems far too high. Let me reconsider whether I'm interpreting the yield curve correctly—perhaps these aren't the right discount factors for this calculation. An annuity factor of 1.8955 for two years of payments actually seems too low for typical interest rates. For 3% rates, the annuity factor should be closer to what I'd expect from standard present value calculations. When I calculate it using the discount factors directly—0.9709 plus 0.9246—I get exactly 1.8955. If I instead use a flat 3.5% rate across both years, the annuity factor comes to about 1.8997, which is nearly identical. This consistency suggests my calculation is correct, even though the result feels counterintuitive. The implied swap rate from this annuity factor would be 1/1.8955 = 0.5275, or roughly 52.75%. That seems far too high given the underlying discount rates are only 3-4%, so something about my assumptions must be off. Perhaps the rates themselves are unrealistically low for this scenario. Testing with more plausible modern rates of 4% and 4.5% yields discount factors of 0.9615 and 0.9157, giving an annuity factor of 1.8772 and a swap rate of 53.26%. The result barely changes, which suggests the formula itself might be flawed rather than just the input rates. I need to reconsider how swaps are actually priced—specifically, the condition that the fixed and floating legs should have equal value at inception. The fixed leg represents payments of S × N discounted at each period's rate. The floating leg receives forward rates set at inception, which means each floating payment equals the forward rate for that period. Since floating rate bonds priced at par have coupons equal to the discount rates, the floating leg is worth par (N) at inception. For the swap to have zero value, the fixed leg must also equal N, which means S times the annuity factor should equal 1. Recognizing that the discount factors are reciprocals of (1 + r_i)^{t_i}, I can express the swap rate as: $$S = \frac{1}{\sum_i DF_i}$$ This confirms the formula I derived earlier. Let me verify with the numerical example: with discount factors of 0.9709 and 0.9246, the annuity factor is 1.8955, giving a swap rate of approximately 52.76 basis points. The result seems counterintuitive because the 3% and 4% rates are historically low. Testing with higher rates—say 10% and 10.5%—yields discount factors of 0.9091 and 0.8188, producing an annuity factor of 1.7279 and a correspondingly lower swap rate. But I realize the core issue: I'm confusing the annuity factor with the swap rate itself. The AF represents the present value of $1 received annually, so an AF of 1.8955 means $1 per year for two years is worth $1.8955 today. However, the quoted swap rate is a percentage applied to notional principal. With $100M notional at 4%, the annual fixed payment is $4M, and its present value equals that payment multiplied by the respective discount factors. The real problem is how I'm treating the floating leg. A swap doesn't involve principal repayment, so I can't simply value the floating leg as if it were a bond trading at par. Instead, each floating payment gets set at the forward rate for its period and then discounted using the spot rate, which means the floating leg's present value depends on the forward curve and discount factors applied to those forward rates. There's a mathematical relationship between forward and spot rates that simplifies this: the forward rate from time $t_{i-1}$ to $t_i$ relates to the spot rates through the compounding relationship. When I substitute this relationship into the present value formula, the floating leg actually collapses to something much simpler—the key insight is recognizing that a floating rate note with no principal exchange effectively has a present value equal to par at inception. The reason is that floating rate coupons are set at prevailing market rates each reset period. At the swap's initiation, the first floating payment is determined by the current LIBOR rate (or forward LIBOR if that cash flow occurs later), and subsequent payments reset to whatever the market rate is at each reset date. Since these coupons always match the discount rates applicable to those periods, the present value of the floating leg equals the principal amount. My formula appears sound: $S = \frac{1}{\sum_i DF_i}$, which gives 52.75% in this case. I was initially skeptical of this result because it seemed high, but after reconsidering how discount factors work—they're derived from discount rates, not spot rates—I'm now confident the calculation is correct. Let me verify this makes sense. The formula is dimensionally consistent since both sides are unitless. With $\sum DF_i = 1.8955$, I get $S = 0.5275$. This swap rate means that on a $100M notional, the fixed leg pays $52.75M annually, which should balance against the floating rate payments averaging around 3%. But that's clearly wrong—I shouldn't be paying 52.75% fixed when floating rates are only 3%. Something fundamental is off with my approach. Going back to the swap valuation equation: at inception when the swap value is zero, the present value of floating payments must equal the present value of fixed payments. This gives me: $$\sum_i \frac{\text{Floating}_i \times \delta_i \times N}{(1 + z_i)^{t_i}} = S \times N \times \sum_i \frac{\delta_i}{(1 + z_i)^{t_i}}$$ When floating rates equal forward rates, I can substitute those in. The key insight is that the floating leg's value, when properly discounted using forward rates and discount factors, actually simplifies to something much cleaner than what I've been calculating. The floating leg should equal par at inception, which means the fixed rate should be determined by the annuity factor of discount factors, not by whatever calculation produced that absurd 52.75% figure. Let me approach this differently—thinking of a swap as a portfolio of two bonds. I'm essentially short a fixed-rate bond paying S% and long a floating-rate bond paying LIBOR. A floating-rate bond that resets coupons at prevailing market rates stays at par value because each new coupon is set to market rates, so there's no accumulated gain or loss from rate changes. The critical detail is whether the first floating payment uses the current spot rate or a forward rate. In practice, swaps typically set the first period's rate at the prevailing LIBOR for that tenor, meaning the initial floating payment reflects today's market rate. When I discount this first payment back, its present value reflects the current market conditions for that period. The key insight is that a floating leg structured with rates set at initiation based on forward rates has a present value equal to the notional amount itself. This is a fundamental property of interest rate swaps—the floating side is always worth par at inception because each cash flow is priced using the corresponding forward rate. Therefore, at the swap's start, the floating leg's value equals the notional. The first floating payment at time $\delta_1$ is determined by the forward rate $f_1$ applicable to that period, giving a cash flow of $f_1 \times N \times \delta_1$ discounted at the spot rate. This pattern holds for each subsequent period. The key insight is that when floating rates are set at their corresponding forward rates at inception, the present value of the floating leg equals the notional. Setting the fixed leg's present value equal to this notional amount and solving for the swap rate $S$ gives us the par swap rate. The formula $S = \frac{1}{\sum_i \frac{\delta_i}{(1+z_i)^{t_i}}}$ follows directly from this equilibrium condition. Now I need to verify my discount factors are accurate. With a one-year zero yield of 3%, a dollar received in one year discounts to approximately $0.9709. For the two-year zero at 4%, the discount factor should be $1/(1.04)^2$. The annuity factor sums to 1.8955, giving a swap rate of roughly 52.75%. This seems unusually high, so let me reconsider what these yields represent. A 3% zero-coupon yield means the bond's YTM is 3%, so if it matures for $100 in one year, the purchase price would be $100/1.03 ≈ $97.09. This confirms the discount factor of 0.9709 is correct. Similarly, the present value of $1 paid in two years at a 4% zero yield is $1/(1.04)^2 ≈ $0.9246. To find the fixed payment X that equates to $1 in present value over two years, I solve X(0.9709 + 0.9246) = 1, which gives X ≈ $0.5275 per dollar of notional. For a $100M debt, this translates to annual payments of $52.75M. I should verify this makes sense relative to the underlying rates—if I borrowed $100M at a fixed 4% rate instead, the interest alone would be $4M in year one and $4M in year two, which is far less than the $52.75M annuity payment, suggesting something may be off with my calculation or assumptions. The key difference is that the annuity approach repays principal each year, whereas a simple interest loan doesn't. In year one, the $52.75M payment covers roughly $3M in interest (at the 3% rate) and $49.75M in principal reduction, leaving $50.25M outstanding. Year two's payment then covers interest on that reduced balance. The issue is that I'm conflating different discount rates—the 1-year and 2-year rates aren't the same, so the amortization schedule doesn't align cleanly with a single interest rate assumption. But this actually clarifies why the annuity payment is so much larger than a simple coupon. The swap rate of 52.75% is effectively annualizing both interest and principal repayment, while the 3-4% zero-coupon yields represent pure interest rates. Since principal must be returned alongside interest, the total payment naturally needs to be substantially higher. I can verify this using the standard annuity formula with the average rate of 3% over two years, which should confirm the payment calculation. If the principal is $100M and the annual payment equals 0.522 × $100M = $52.2M, then the payment represents 52.2% of the principal each year. This aligns with my earlier calculation of around 52.75%. But this conflicts with what I know about actual swap markets. Real 2-year USD swap rates trade around 5%, not 50%. Something in my approach must be off. The discrepancy likely stems from how rates are quoted in practice. When market participants reference "the 2-year swap rate at 3.5%," they're using a specific convention that differs from my calculation method. mean that the swap rate is 3.5% per annum, with semi-annual coupons. If we have semi-annual coupons, then each period the fixed payment is: $$\text{Payment} = 3.5\% \times \frac{180}{360} \times Notional = 1.75\% \times Notional$$ So the semi-annual payment is 1.75%, not 3.5%. Maybe my error is assuming annual coupons? Let me redo with semi-annual coupons: Assuming annual zero rates of 3% and 4%, and semi-annual coupons: Semi-annual zero rates: - r_{0.5} = sqrt(1.03) - 1 = 2.47% - r_{1.0} = sqrt(1.04) - 1 = 1.96%... No wait, that's not right. The 1-year annual rate of 3% means money grows by 1.03 over a year. With semi-annual compounding, I need (1 + r_{0.5})^2 = 1.03, so r_{0.5} = sqrt(1.03) - 1 ≈ 1.489%. For the 2-year rate of 4%, the product of all four semi-annual discount factors should equal 1.04^2. If I assume a flat forward curve at 1.98% per semi-annual period, then the annuity factor for 4 periods is approximately 3.83. The swap rate would be S = 2 / 3.83 ≈ 52%, which is clearly wrong. I think the fundamental issue is that I'm confusing how to properly convert annual rates to semi-annual periods and how to apply the swap pricing formula correctly. Let me reconsider the basic principle: at initiation, the fixed leg and floating leg should have equal present values. The floating leg, when priced using forward rates, should equal par. So the fixed leg must also equal par, meaning: S × (annuity factor) = 1 But when I compute this with my numbers, I get impossibly high swap rates like 105% for a 1-year swap with a 5% zero rate. This suggests my annuity factor calculation is wrong. The issue might be that I'm not properly accounting for the day count fractions or the relationship between spot and forward rates. When forward LIBOR rates are used to price the floating leg, the resulting value should indeed be par, but I may be making an error in how I'm computing the discount factors or the annuity factor itself. not right—the forward rate should exceed the spot rate if the curve slopes upward. I'm mixing up my notation and need to be clearer. Let me redefine things properly: the spot rate $r(0, t_i)$ is the rate at time 0 for a maturity of $t_i$, with discount factor $DF(0, t_i) = 1 / (1 + r(0, t_i))^{t_i}$. The forward rate $f(t_{i-1}, t_i)$ is locked in at time 0 for borrowing between periods $t_{i-1}$ and $t_i$. For a floating-rate bond, the present value of payments is $\sum_i f(t_{i-1}, t_i) × \delta_i × N × DF(0, t_i)$, where $\delta_i$ represents the time between periods. The relationship between spot and forward rates follows from the no-arbitrage condition: $(1 + r(0, t_{i-1}))^{t_{i-1}} × (1 + f(t_{i-1}, t_i))^{\delta_i} = (1 + r(0, t_i))^{t_i}$. Solving for the forward rate and substituting into the PV formula for period $i$, I can express the accrued interest as $f(t_{i-1}, t_i) × \delta_i = \frac{(1 + r(0, t_i))^{t_i}}{(1 + r(0, t_{i-1}))^{t_{i-1}}} - 1$. When I multiply this by the discount factor and principal, the $(1 + r(0, t_i))^{t_i}$ terms cancel, leaving $N × \left[ \frac{1}{(1 + r(0, t_{i-1}))^{t_{i-1}}} - \frac{1}{(1 + r(0, t_i))^{t_i}} \right]$. Summing across all periods creates a telescoping series where consecutive discount factors cancel. Starting from $DF(0, 0) = 1$ at initiation and ending at $DF(0, t_n)$, the floating leg's present value simplifies to $N × [1 - DF(0, t_n)]$. This reveals that the floating leg doesn't trade at par—its value depends on the terminal discount factor. For a 2-year swap with $DF(0, 2) = 0.9246$, the floating leg is worth approximately $N × 0.0754$. Setting the fixed and floating legs equal to determine the swap rate: $S = \frac{1 - DF(0, t_n)}{\sum_i DF(0, t_i) × \delta_i}$. Plugging in the discount factors and annual payment frequency yields $S = \frac{0.0754}{1.8955} ≈ 3.98\%$, which aligns with market expectations for this maturity. Verifying with a simpler case—a 1-year swap at 5% rates—confirms the formula works: the swap rate equals 5%, matching the zero rate exactly as it should. The general principle holds: $S = \frac{1 - DF(0, T)}{\sum_i DF(0, t_i) × \delta_i}$. The critical realization is that the floating leg's value telescopes to $N × [1 - DF(0, T)]$ rather than the full notional. Since no principal exchanges hands in a standard swap, only the interest payments matter—the missing principal repayment accounts for the difference between what we'd expect and what actually occurs.</think> # Plain Vanilla Interest Rate Swap Valuation ## Overview A **plain vanilla interest rate swap** is an agreement where two parties exchange periodic interest payments on a notional amount for a fixed period, with no exchange of principal. --- ## Swap Structure: Fixed vs. Floating Legs ### Fixed Leg (Payer's Obligations) - Pays a **fixed rate** determined at inception - Cash flow each period: **Fixed Rate × Notional × Day Count** - Amount is **known with certainty** from day one - Acts like owning a fixed-rate bond ### Floating Leg (Receiver's Income) - Pays a **floating reference rate** (LIBOR, SOFR, etc.) - Cash flow each period: **Floating Rate × Notional × Day Count** - Floating rate **resets periodically** (typically quarterly/semi-annually) - Acts like owning a floating-rate note **Example:** In a $100M, 2-year swap at 4% fixed, semi-annually: - Every 6 months: Fixed payer pays $2M (= 4% ÷ 2 × $100M) - Every 6 months: Floating payer pays LIBOR-based amount --- ## Valuation at Inception (Zero Value Condition) For a swap to be **fairly priced with zero value at initiation**, the present value of fixed payments must equal the present value of floating payments. ### Floating Leg Present Value (Key Insight) When floating rates are set at **forward rates** at initiation, the floating leg has a special property. Using a telescoping series: $$PV_{\text{float}} = N \times [1 - DF(T)]$$ Where: - $N$ = Notional amount - $DF(T)$ = Discount factor at maturity **Why?** The floating rate for each period, when discounted by the corresponding spot rate, nets out except for the terminal discount factor. ### Fixed Leg Present Value $$PV_{\text{fixed}} = S \times N \times \sum_{i=1}^{n} \delta_i \times DF(t_i)$$ Where: - $S$ = Swap rate - $\delta_i$ = Day count fraction for period $i$ - $DF(t_i)$ = Discount factor for period $i$ --- ## Swap Rate Formula Setting $PV_{\text{fixed}} = PV_{\text{float}}$ at inception: $$S \times N \times \sum_{i=1}^{n} \delta_i \times DF(t_i) = N \times [1 - DF(T)]$$ Solving for $S$: $$\boxed{S = \frac{1 - DF(T)}{\sum_{i=1}^{n} \delta_i \times DF(t_i)}}$$ Or equivalently: $$S = \frac{1 - DF(T)}{\text{Annuity Factor}}$$ The **annuity factor** is often denoted as **DV01** (dollar value of 1 basis point). --- ## Practical Example **Scenario:** 2-year annual swap - Spot rates: 1-year = 3%, 2-year = 4% - Discount factors: DF₁ = 0.9709, DF₂ = 0.9246 - Annual coupons (δ = 1) **Calculation:** $$\text{Annuity Factor} = 0.9709 + 0.9246 = 1.8955$$ $$S = \frac{1 - 0.9246}{1.8955} = \frac{0.0754}{1.8955} = 0.0398 = \boxed{3.98\%}$$ The swap rate is approximately **3.98%**, which makes intuitive sense given the 3-4% yield curve. --- ## Key Takeaways | Concept | Description | |---------|-------------| | **Fixed Leg** | Certainty; represents fixed-rate obligation | | **Floating Leg** | Uncertainty; benefits from rate increases | | **Zero Value at Inception** | Both legs are equally attractive to both parties | | **Swap Rate** | Reflects the entire yield curve; typically near the midpoint for upward-sloping curves | | **DV01** | Annuity factor; measures interest rate sensitivity | The swap rate effectively represents a **break-even rate** where neither party has an advantage over the other at the time of initiation.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.