The benefits of MIT Sloan Executive Education -
In my inbox this morning - Thinking Forward
Microsoft chief scientist: To make the most of gen AI, use your imagination
1. Can lending standards be too tight for too long following an economic shock? New research from MIT Sloan professor Jonathan Parker and colleagues indicates that can be the case when banks go the extra mile to screen out bad borrowers.
The researchers’ findings suggest that some of the persistence of tightness after a shock is self-inflicted. When banks try to protect themselves by increasing their due diligence in evaluating potential borrowers, they set off a chain reaction whereby other banks do the same and increase the credit crunch in two ways:
By not funding people who are less creditworthy, tighter standards worsen the overall quality of the pool of potential borrowers, which in turn creates an even bigger incentive for banks to employ tight standards in the future.
In a good market, lending standards are inefficient when they’re too tight. They amplify and prolong downturns, decrease overall lending, and increase credit spreads.
If the market is in good shape, less due diligence makes more sense, said Parker, co-director of the MIT Consumer Finance Initiative. “The chances that [a loan] will fail are very small. Some of them will default and lose a little bit of money, but that will be a rare event,” he said.
2. How much will artificial intelligence affect productivity in coming years? Writing recently in the Financial Times [subscription required], columnist Soumaya Keynes observed that there have been several attempts to estimate the effects of generative AI on annual productivity growth, “with pretty varied results.”
Goldman Sachs last year estimated that in rich countries, generative AI could increase productivity by about 1.5 percentage points over a decade. McKinsey predicted that generative AI could deliver an increase of 0.1 to 0.6 percentage points between 2023 and 2040.
And, more recently, MIT professor Daron Acemoglu calculated a boost over the next decade of, at most, 0.2 percentage points and suggested that over the next decade, around 5% of tasks will be profitably replaced or augmented by AI.
“Some uncertainty is, of course, healthy,” Acemoglu said of the change brought on by AI, since “we are at the very, very beginning of it.”
MIT Sloan Office of Communications
Building E90, 9th Floor
1 Main Street, Cambridge, MA 02142
#AI #mit #misinformation
6