Saturday, August 26, 2017

AI Watson, would you ever feed robobankers those algorithms current bank regulators feed human bankers?

The normal real world rules banks had to follow for about 600 years before 1988, in order to become and remain successful bankers, was to while carefully considering their portfolio, to lend or invest in whatever they perceived would produce them the highest risk adjusted returns on equity. One dollar of equity lost in an operation perceived as risky would hurt just as much as a dollar lost in an operation perceived as safe. 

And even though bankers in general suffered from a risk aversion bias, expressed well by Mark Twain’s “a banker is one to lend you the umbrella when the sun shines and wanting it back when it seems it could rain”, that obviously served our economies well. 

As John Kenneth Galbraith argued, even when in some cases “Banks opened and closed doors and bankruptcies were frequent, as a consequence of agile and flexible credit policies, the failed banks left a wake of development in their passing.”

But then came the Basel Committee for Banking Supervision, and out of the blue decided to assume bankers did not perceive risks; and so came up with their risk weighted capital requirements. These instructed banks, with no consideration to their portfolio, to hold more capital (equity) against what is perceived risky and less against what is perceived safe.

As an example of their crazy regulatory algorithm, it suffices to mention they assigned a risk weight of 20% for the so dangerously AAA rated while one of 150% to the so innocous below BB- rated.

As a consequence of this bankers had to morph from being mainly risk perceivers into also having to be capital (equity) minimizers. Being able to leverage more the “safe “ than the “risky” allowed them to obtain higher risk adjusted returns on equity lending with the safe than with the risky.

As a consequence we have already suffered major bank crisis resulting from excessive exposures to what was erroneously perceived, decreed or concocted as safe, like AAA rates securities and sovereigns like Greece.

As a consequence of not enough lending to the “risky”, like SMEs and entrepreneurs, development is coming to a halt.

Since regulators refuse to listen to little me, I can’t wait for IBM’s Watson developing lending and investment algorithms for robobankers. These would help show bank current regulators how dangerously wrong they are.

Watson would understand that with current distortions banks go wrong even if they perceive the risks absolutely correct.

Watson would understand that what is perceived risky ex ante becomes by that fact alone less dangerous ex post; and that what is perceived safe ex ante becomes by that fact alone more dangerous ex post.

May God defend me from my friends, I can defend myself from my enemies” Voltaire.


Regulator Watson, in contrast to the Basel Committee, would not be looking in the same direction the bankers look