I was talking the other day about Herbert Simon, easily the most important thinker on decisions of the last century. One of his most important books, The Sciences of the Artificial, talks about the differences between natural science and the approach needed for dealing with the artificial , i.e man-made word.
There are a hundred important points in what he says in this one book, but I’ll stick to just a few. First, if you have an uncertain situation, it is much better to find better ways to adapt, rather than predict or optimize in the classic economic way.
Although the presence of uncertainty does not make intelligent choice impossible, it places a premium on robust adaptive procedures instead of optimizing strategies that work well only when finely tuned to precisely known environments. A system can generally be steered more accurately if it uses feedforward, based on prediction of the future, in combination with feedback, to correct the errors of the past. However, forming expectations to deal with uncertainty creates its own problems. Feedforward can have unfortunate destabilizing effects, for a system can overreact to its predictions and go into unstable oscillations. Feedforward in markets can become especially destabilizing when each actor tries to anticipate the actions of the others (and hence their expectations).
Of course, that describes the situation of most central bank decisions. Everyone knows feedforward – economic forecasts – are highly unreliable. Overreacting to those forecasts can produce “unstable oscillations.” Consider the history of the American economy in the last fifteen years. It means the standard inflation-forecast rule prevalent in central banks is deeply flawed, but in interesting ways.
This also has major implications for how my company – Alucidate – looks for value. Think of things in evolutionary terms. The most successful surviving firms and investors are not those that predict the future better, any more than natural evolution proceeds by having squirrels forecast nut production in 2016 or 2100. It would be nice if people really could predict successfully, but just about every scrap of evidence we have suggests they can’t. You can ignore that, or creatively deal with it. (Or stockpile a lot of nuts.)
Instead, what decision-makers can do is find ways to adapt faster, to find “robust adaptive procedures..” You recognize changes in the environment and work out ways to take advantage of them. It is a question of robust search capacity (and variation, and selection). The fundamental challenge is recognition and awareness and resilience, not prediction. If you adapt, you survive and thrive.
Simon also thinks in terms of nonlinear systems , with feedback loops, rather than a standard linear economic model. The neoclassical equilibrium model is far too simplistic.
In sum, our present understanding of the dynamics of real economic systems is grossly deficient. We are especially lacking in empirical information about how economic actors, with their bounded rationality, form expectations about the future and how they use such expectations in planning their own behavior. … In face of the current gaps in our empirical knowledge there is little empirical basis for choosing among the competing models currently proposed by economics to account for business cycles, and consequently, little rational basis for choosing among the competing policy recommendations that flow from those models. (from the 3rd edition, 1996)
(I’ll pursue more recent work on this theme later.)
What’s worse, 80% of American economic activity takes place within organizations, he says, so looking at the economy simply in terms of markets is inadequate. You have to understand how organizations make decisions as well, and when and why the coordination abilities of organizations mop up activity from open markets. One critical factor is “identification”, or people’s loyalty to organizational aims.
It all means that thinking in terms of reactions and feedback and the pace of learning in organizations is essential to understand both the economy and how economic decisions really get made. But economic research in general does not do that.
Instead of putting analysis of economic data and forecasts first, we need to focus on decision-making in practice.