Skip to content
All library documents

Replicating Portfolios and the Arbitrage Pricing Argument

Article Quant Q&A · Author: BCLC

Summary

The document explains the no-arbitrage argument that a claim with a replicating portfolio must have the same initial price as that portfolio. If the claim is overpriced, short the claim, buy the replicating portfolio, and invest the surplus at the risk-free rate. At maturity, sell the portfolio to fund the payoff owed on the short claim; the initial surplus has grown into a risk-free profit.

If the claim is underpriced, the reverse trade is to buy the claim and short the portfolio, retaining the difference. The answer addresses a confusion about the portfolio’s terminal value: it is used to meet the claim liability, because replication makes the two payoffs equal. Borrowing against an unknown terminal portfolio value at the initial date is unnecessary. The argument assumes the portfolio really replicates the claim, that the trades can be executed at the stated prices, and that borrowing, shorting, and investing are available under the model’s assumptions.

Key ideas

  • A replicating portfolio and its claim must have equal initial values in an arbitrage-free market.
  • When the claim is overpriced, short it and buy the replicating portfolio.
  • Invest the initial price difference to earn a risk-free return.
  • At maturity, the portfolio payoff covers the liability on the short claim.
  • The reverse trade applies when the claim is underpriced.

Tags

Full text
# Arbitrage Strategy Proof in Bjork


# Arbitrage Strategy Proof in Bjork












In Tomas Bjork's Arbitrage Theory in Continuous Time (or here), $\exists$ this proposition

Proposition 2.9 Suppose that a claim X is reachable with replicating portfolio h. Then any price at t=0 of the claim X, other than $V_0^{h}$ will lead to an arbitrage possibility.

My prof uses $V_0({\phi})$ instead of $V_0^{h}$, but $\phi$ still refers to the portfolio.

Let $\Pi(t;x)$ be the price of the contingent claim at time t. Then, $\Pi(0;x)$ must = $V_0({\phi})$.

This is the proof written on the board:

Suppose $\Pi(0;x)$ > $V_0({\phi})$.

The arbitrage strategy is:

Sell (or short sell) the claim for $\Pi(0;x)$, and obtain the portfolio $\phi$ worth $V_0({\phi})$.

Left over amount is $\Pi(0;x)$ - $V_0({\phi})$.

At t = 1, the payoff for the claim X w/c you will be liable for will be covered the value of the portfolio $V_1({\phi})$ at t=1.

Suppose $\Pi(0;x)$ < $V_0({\phi})$.

The arbitrage strategy is:

Sell (or short sell) the portfolio worth $V_0({\phi})$. Use that amount to buy claim worth $\Pi(0;x)$.

Left over amount is $V_0({\phi}) - \Pi(0;x)$.

At t = 1, you will get payoff X, w/c is equal to $V_1({\phi})$.

Soooo I tried constructing the arbitrage strategy for the first part to see the exact profit, but I seem to be missing a step.

Suppose $\Pi(0;x)$ > $V_0({\phi})$.

At t = 0,

transaction $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ cash flow

1 short/sell claim for $\Pi(0;x)$ $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ +$\Pi(0;x)$

2 Buy $\phi$ $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ -$V_0({\phi})$

3 Invest $\Pi(0;x)$ - $V_0({\phi})$ at R until t=1 $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ -($\Pi(0;x)$ - $V_0({\phi})$)

$\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ 0

At t = 1,

transaction $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ cash flow

1 collect investment $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ +($\Pi(0;x)$ - $V_0({\phi})$)(1+R)

2 Portfolio grows in value $\phi$ $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ +$V_1({\phi})$

3 [...]? $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ -$V_1({\phi})$

4 Close short position if needed $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ 0

Profit $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ +($\Pi(0;x)$ - $V_0({\phi})$)(1+R)

Where does the $V_1({\phi})$ go? I was thinking that we were supposed to borrow $\frac{V_1({\phi})}{1+R}$ at t=0 so it would look something like:

Suppose $\Pi(0;x)$ > $V_0({\phi})$.

At t = 0,

transaction $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ cash flow

1 short/sell claim for $\Pi(0;x)$ $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ +$\Pi(0;x)$

2 Buy $\phi$ $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ -$V_0({\phi})$

3 Borrow $\frac{V_1({\phi})}{1+R}$ at R until t=1 $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ +$\frac{V_1({\phi})}{1+R}$

4 Invest $\Pi(0;x)$ - $V_0({\phi})$ + $\frac{V_1({\phi})}{1+R}$ at R until t=1 $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $$-(\Pi(0;x)$ - $V_0({\phi})$ + $\frac{V_1({\phi})}{1+R})$

$\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ 0

At t = 1,

transaction $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ cash flow

1 collect investment $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $$(\Pi(0;x)$ - $V_0({\phi})$ + $\frac{V_1({\phi})}{1+R})(1+R)$

2 Pay debt $\phi$ $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ -$V_1({\phi})$

3 Close short position if needed $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ 0

Profit $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $ +($\Pi(0;x)$ - $V_0({\phi})$)(1+R)

Or is there some action for the missing step that makes both of them equivalent?

## Answer by pincopallino (score 1, accepted)

https://quant.stackexchange.com/a/12743

If $V_0(\phi) < \Pi(0,x)$

at $t=0$

- You sell short the claim and collect $\Pi(0,x)$

- You buy the portfolio $\phi$ for $V_0(\phi)$

- You put the money $\Pi(0,x) - V_0(\phi)$ in your risk-free instruments

at $t=1$

- At $t=1$ you'll be liable the payoff of the claim you have shorted. The money you owe the counterpart long the claim is $\Pi(1,x)$.

- $\phi$ is the replicating portfolio hence $V_1(\phi) = \Pi(1,x)$. You can sell the portfolio $\phi$ and get $V_1(\phi)$

- Therefore, you are left with $(\Pi(0,x) - V_0(\phi)) (1+R)$ in your bank account.

I don't think you'll need to borrow $V_1(\phi) / (1+R)$ at $t=0$: you can already finance the long position in the portfolio with part of the proceeds from the short sale of the claim.

In any case, you don't know the value $V_1(\phi)$ at $t=0$. All you know is that $V_1(\phi) = \Pi(1,x)$ at $t=1$.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.