Asian Options: What Happens When the Market Feels Your Trade?

Asian options are contracts that pay out based on the average price of an asset, not just its final value. Think of betting on the average price of a pizza over a week – the payoff depends on all the daily prices, not just the last one.

In an ideal market, buying or selling doesn’t move the price. In reality, large trades can shift a stock’s value, a phenomenon called permanent price impact. Researchers have added this effect to a binomial market model, a step‑by‑step price tree where each move can go up or down.

Using a modified risk‑neutral probability, they derived an exact formula for geometric Asian options and two‑sided bounds for arithmetic ones. The striking finding: "Permanent price impact systematically raises Asian option prices." In other words, the more your trade pushes the market, the pricier the option becomes.

They also identified the no‑arbitrage region – where risk‑free profit isn’t possible – based on hedging volumes. Simulations show that higher impact parameters or larger hedging volumes increase option prices. This helps traders and analysts gauge how much to pay when the market reacts to their moves.

In short, the market isn’t a passive mirror; your actions leave a trace, and Asian options feel that weight.

In the traditional Cox‑Ross‑Rubinstein binomial model, the underlying asset price follows a Markov chain with up‑probability p and down‑probability 1‑p, preserving no‑arbitrage via a risk‑neutral measure. The proposed extension adds a permanent impact term ΔV, proportional to trade volume V, which modifies the price dynamics: S_{t+1}=S_t(1+u)−ΔV for an up‑move and S_{t+1}=S_t(1+d)+ΔV for a down‑move. This introduces a non‑linear path‑dependent component, as ΔV accumulates the effects of all prior trades.

For geometric Asian options, the geometric mean G_T=\big(\prod_{t=1}^T S_t\big)^{1/T} allows an analytical closure thanks to the multiplicative property of log‑prices. A pathwise representation is obtained: V_0=E^{\mathbb{Q}}[e^{-rT}\,\max(G_T-K,0)], where \mathbb{Q} is the modified risk‑neutral measure. Permanent impact translates into an extra drift term in the log‑normal process, shifting the distribution of G_T toward higher values.

For arithmetic Asian options, the arithmetic mean A_T=\frac{1}{T}\sum_{t=1}^T S_t lacks a closed‑form solution. The authors construct two bounds: a lower bound based on a sub‑optimal hedging strategy that ignores impact, and an upper bound obtained via a martingale transformation that overestimates the impact contribution. Both bounds converge to the true price as ΔV → 0.

The no‑arbitrage region is defined by conditions on V: for each node of the binomial tree, the hedging volume must satisfy 0\le V\le V_{max}(S_t,Δ), where V_{max} depends on the current price and the impact parameter Δ. Exceeding this limit creates arbitrage opportunities, as the option price diverges from its replicating portfolio.

Numerical simulations show that, for Δ=0.01 and V=10^5 units, the price of an arithmetic Asian option with maturity T=30 days rises by roughly 12% compared to the classic model. The effect is more pronounced for deep‑in‑the‑money options, where sensitivity to the average price is higher. These findings suggest that in highly liquid markets with large participants, permanent impact is a crucial factor to include in pricing models.

This analysis paves the way for further extensions, such as temporary impact, multi‑asset correlation, and jump processes to capture extreme market events.