Real world probabilities are used in forecasting the future (porfolio management).Risk neutral probabilities are used in getting the market price (derivatives pricing).Isn't it non-consistent to put so much effort in calibrating your models, to replicate the real world, and then to use risk neutral "for derivatives only"?

please explain the inconsistency

QuoteOriginally posted by: APlusReal world probabilities are used in forecasting the future (porfolio management).Risk neutral probabilities are used in getting the market price (derivatives pricing).Isn't it non-consistent to put so much effort in calibrating your models, to replicate the real world, and then to use risk neutral "for derivatives only"?Taking expectations under the risk neutral probability measure is just a computational device. For the purpose of calculating relative values, like the value of an option as a function of the underlying stock price, you don't need to know the size of the risk premium earned by holders of the stock, so it is expedient to temporarily suppress it (i.e. set it equal to zero) while carrying out this particular calculation. For the purpose of deciding whether the stock (or, for that matter, the option) belongs in your portfolio, you do actually care a lot about the risk premium, so you should make an effort to include your best estimate of its value in your portfolio selection algorithm.

Risk neutral probabilities make the prises consistent (your pricing is compatible with the prices of other contracts in the market). I am just curious which real-risk probabilities method is used for derivatives (the drift does not cancel away in the actual applied formula). In the convertible bonds, where you keep comparing the value of your option with the value of the stock, you do consider the risk premium earned by holders of the stock, don't you? This is why convertible bonds are not trivial to price: because once you do your forecasting with trees and black scholes you are in the risk-neutral world. I do not say that it is necessarily wrong,but inconsistent.

QuoteOriginally posted by: APlusRisk neutral probabilities make the prises consistent (your pricing is compatible with the prices of other contracts in the market). I am just curious which real-risk probabilities method is used for derivatives (the drift does not cancel away in the actual applied formula). In the convertible bonds, where you keep comparing the value of your option with the value of the stock, you do consider the risk premium earned by holders of the stock, don't you? This is why convertible bonds are not trivial to price: because once you do your forecasting with trees and black scholes you are in the risk-neutral world. I do not say that it is necessarily wrong,but inconsistent.The main reason for pricing derivatives with risk neutrality is the two-sided nature of the ownership of derivatives -- someone is long and someone else is short the instrument.You ask what is the risk premium earned by the buyer of an stock option (or other derivative). But what is the risk premium earned by the seller of that same stock option (or other derivative)? In the case of most derivatives, every dollar lost by the (long) buyer is a dollar gained by the (short) seller and every dollar gained by the buyer is a dollar lost by seller. In contrast, a stock transaction is essentially one-sided. If a share of some stock loses a dollar, the owners of that stock lose money but there no counterparty who gains an equal and opposite amount.

Originally posted by: bearishQuoteFor the purpose of calculating relative values, like the value of an option as a function of the underlying stock price, you don't need to know the size of the risk premium earned by holders of the stock, so it is expedient to temporarily suppress it (i.e. set it equal to zero) while carrying out this particular calculation. Wasn?t this the Black Scholes (Merton) option pricing breakthrough, removing risk preferences??Originally posted by: Traden4AlphaQuoteIn contrast, a stock transaction is essentially one-sided. If a share of some stock loses a dollar, the owners of that stock lose money but there no counterparty who gains an equal and opposite amount.Short sellers??

QuoteOriginally posted by: dweebOriginally posted by: bearishQuoteFor the purpose of calculating relative values, like the value of an option as a function of the underlying stock price, you don't need to know the size of the risk premium earned by holders of the stock, so it is expedient to temporarily suppress it (i.e. set it equal to zero) while carrying out this particular calculation. Wasn?t this the Black Scholes (Merton) option pricing breakthrough, removing risk preferences??Yes, I think you are right. But the removal of risk preferences was not without some justification.Originally posted by: Traden4AlphaQuoteIn contrast, a stock transaction is essentially one-sided. If a share of some stock loses a dollar, the owners of that stock lose money but there no counterparty who gains an equal and opposite amount.Short sellers??The percentage of short stock is almost always small compared to the float. Moreover, taking a long position in a stock does not require finding someone willing to take a short position. In contrast, with derivatives, there's always an equal and opposite number of long and short positions and an equal and opposite number expecting a risk premium for their position.All that said, if the crowd is wrong, then the risk premia do not cancel and some individuals might earn a consistent risk premium for being either long or short. But whether it's the long or the short side that earns a risk premium is not clear, not consistent, and debatable.

Read in an interview with Seth Klarman that Baupost invests long only in equities, on the basis that stocks have a greater propensity for the upside.