Derivatives have been denounced as “financial weapons of mass destruction”, but they are in fact the opposite: by virtue of theoretical modeling, virtual markets are priced into existence. Now that markets both ‘organic’ and ’synthetic’ are melting, some have chosen to blame the ‘quants’, modelers who came up with the virtual prices on which derivative markets for illiquid assets are predicated.
It can be truthfully contended that contemporary quantitative stochastic finance (QSF) has oversold its functional grounding on measure theory as the mathematical basis for a better, synthetic capitalism where every aspect of risk — from gammas to volgas — can be actualized as concrete value flows. What critics overlook, however, is the counterpart of QSF-derived prices for synthetic markets: market process-derived prices on organic markets.
Are gaussian copula models for collateralized debt products that much less efficient than the massively parallel, error-prone neural network that would power the individual investor buying directly into the organic mortgage market — as some societies have seen in the late 1800s? As with nukes, technology is just that: technology, not weapon. Furthermore, nuclear technology is a cat that’s either irreversibly out of the bag or goldilocked into nonexistence, while the destructive potential in QSF technology comes from trust, and thus can be effectively calibrated into utopia.
Stochastic calculus is probably oversold. The gist of Moore’s Law — that we all have supercomputers at our fingertips, right now — allows for massive simulation techniques on much more realistic stochastic discounted cash flow models allowing for beta-delta pseudo-hyperbolic intertemporal preferences and many kinds of business-model specificities (random asset replacement times instead of depreciation models, for one) to be simulated hundreds of thousands of time until convergence results are achieved.
In fact, when one looks at what else could be done, it becomes clear where QSF went wrong: it allowed itself to be parameterized by organic pricing, and thus became infected by the error-prone redundant neural networks that power organic markets. The massively simulated cash flow (MSCF) approach just shines a light on business-model knobs, taking a decisively more technocratic turn at valuing cash flows than QSF approaches that allowed its knobs to be marked to market.
I’ve been thinking obsessively about how MSCF methods can be brought into the table — specially when it comes to replacing the stone-age method of scenarioized internal rates of return, but also as a replacement for the more fragile parts of QSF like copula CDO models. I haven’t seen yet how MSCF valuation methods can outperform QSF as a technique for modeling synthetic markets into existence, though, and I really would prefer a future where trading on the volgas is possible rather than the technocratic flavor of business-model knobs.