Risk-Neutral Pricing, Replication, and Derivative Market Prices
Summary
The document questions whether risk-neutral valuation remains adequate as derivatives become actively traded. It argues that the classical view treats derivatives as replicating portfolios, while trading in derivatives can produce volatility skews, smiles, and price spreads that complicate attempts to use models calibrated to existing instruments to value less-liquid exotics. It raises portfolio management and reinforcement learning as possible alternatives for finding equilibrium prices and capital requirements.
The replies emphasize that risk-neutral pricing is a general paradigm tied to replication and arbitrage, rather than a single modeling technique. One response argues that prices departing from replication can invite arbitrage, while another speculates that risk-neutral methods may persist mainly as quotation tools for liquid products, with other products priced and hedged using less parametric approaches and return constraints. The exchange is conceptual and opinion-based: it offers no formal model, data, or evidence resolving the debate. Its claims should be read as perspectives on pricing practice, not a demonstrated forecast of how derivatives markets will evolve.
Key ideas
- The document questions how well risk-neutral models describe derivatives with observed smiles and instrument spreads.
- It frames risk-neutral valuation as pricing through a replicating portfolio and links deviations to potential arbitrage.
- It raises portfolio management and reinforcement learning as possible approaches to pricing and capital estimation.
- The responses offer competing opinions without formal derivations or empirical support.
Tags
Full text
# The end of risk-neutral valuation # The end of risk-neutral valuation Risk-neutral valuation grew out of BS constructing market-neutral portfolios of stocks hedged with options. It was a portfolio management problem. In less than a decade, pricing by arbitrage on a complete arbitrage-free market was invented and the corresponding math theory was constructed. Derivatives in such market are redundant or, equivalently, perfectly illiquid: they are merely fast changing portfolios of tradeables. In reality however, once derivatives become tradeable, they distort the completeness of the original market, becoming the first-class tradeables. In other words, skews and smiles and "instrument spreads" emerge. More and more complex risk-neutral dynamics are proposed to jointly explain observed prices of related systems of derivatives, but with the purpose of pricing unobserved exotics. But those exotics are then never traded at theoretical model values, introducing the next order of smiles and instrument spreads. Is the solution to the virtuous cycle in breaking away from pricing using the risk-neutral approach, and going back to the original portfolio management view, only backed by something like reinforcement learning to arrive at the equilibrium price and risk = capital requirement? ## Answer by Frido (score 3) https://quant.stackexchange.com/a/79935 In my (perhaps simplistic) view risk-neutral pricing is a paradigm, not a particular method and specifically that the price of a derivative is the price of its replicating portfolio. If you deviate from that there is (possibility of) arbitrage, which in any case will probably not last for very long. So no, I don't think the end of RN pricing is anytime soon, whatever method is used to arrive at the price. ## Answer by achirikhin (score 0) https://quant.stackexchange.com/a/79614 I give it 10 years. Some forms of RN will only stay as, possibly, a quotation mechanism of liquid and listed or quasi-listed products, and the rest will be priced and hedged non-parametrically under the constraints on the expected ROI. Thoughts?
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.