The Alchemy of Finance (George Soros)

Reflexivity: human interactions are more complicated than the ones studied by natural sciences because of the following feedback loop: the (imperfect) perception of the situation by the participants affects their behaviour which influences the situation. Also, if the participants are aware of the theories that describe their interactions, this knowledge will influence their actions making theories self-fulfilling or self-refuting.

Intro: - The market is driven by a feedback loop, perception of market conditions drives investment decisions and they drive market conditions (including the fundamentals in some cases). - Investment is like science: you formulate a hypothesis and put it to the test, then you see whether the results support it. - Another feedback loop example: when a credit is granted against a collateral, this tends to increase the value of the collateral (because now the debtor is more solvent).

Theory: - Optimal market equilibrium driven by supply and demand curves has many unrealistic assumptions. For example, supply and demand curves are not fixed but instead respond to events in the market. - Social scientists have to deal with situations where events depend on the understanding of the participants, so there are two levels of imperfect understanding: of the scientists and of the participants. Natural scientists, on the other hand, are dealing with the processes that happen independently of any understanding, so at least there's only one level of imperfect understanding (and it doesn't have a reflexive effect). - Reflexivity: - cognitive function: y = f(x) - participating function: x = ɸ(y) - this gives us: y = f(ɸ(y)) or x = ɸ(f(x)) - my note: he's missing the time aspect here, it's actually a system of differential equations or a multi-step process, but generally seems right. - classical analysis assumes that cognitive function is an identity (perfect knowledge). - The stock market is a relatively simple part of the economy. There only actions available to participants are buy and sell, and the prices of socks are connected to supply and demand in a fairly obvious way. - Funamental analysis assumes that intrinsic value drives the price. However, the reverse is also true: price of company's shares influences the intrinsic value (through issuance, buybacks, m&a, going public/private, credit ratings, management credibility, etc.) - "Market is always right" is replaced with: - Markets are always biased in some direction, - Markets can influence the envents that they anticipate (in case of self-fulfilling feedbacks they turn out right when stuff happens). - Market price can be seen as a composite of underlying trend and prevailing bias. The price in turn affects both the trend and the bias. - Each of trend and bias can be self-reinforcing or self-correcting. - Self-correcting trends and biases reduce themselves, so they are not very interesting. - Self-reinforcing trands and biases escalate usually until a correction. - Example: REITs - REITs have to distribute all earnings. Their book value comes from shares sold. - If the price of a REIT is rising, they invest all proceeds from new shares sold, which increases book value per share (because new shares are sold at a higher price), which in turn increases per-share earnings: positive feedback. - Eventually REIT space becomes too saturated, there's too much cash flowing into REITs and yield competition is high, that causes risky investments. - Currency markets (skipped for now). - Credit: if expectations are optimistic, credit is cheap, that increases the optimism until it's clearly unjustified and then a sudden bust follows.

Practice: - Reflexive processes can be analyzed with time-aware causal diagrams or differential equations.