I'm taking a course in macroeconomics this semester. The teacher began the class by taking roll and then asking everyone what they thought were some important issues in macroeconomics. Other students mentioned that market efficiency, GDP, and international trade were important issues. I wanted to raise my hand to say that determining whether macroeconomists know anything is an important issue in macroeconomics. But it was the first day of class, so I refrained.

Tyler Cowen told me that "You need to start somewhere" in learning about the global economy. I don't think that's a great answer and I'm skeptical about learning things from a group of people whose collective opinion on the subprime crisis and the 2008 economy was wildly incorrect.

Every macro class starts with a simple overarching model of the world, then explains why the model's wrong and looks at more complicated models. But from the beginning, macroeconomists are trying to explain every facet of the economy.

Contrast that with the field of robotics, which I'm starting today as well. First-time robotics students build small robots with limited abilities. Potential class projects include building a robot that's able to use an elevator, one that can navigate the hallways around the classroom by itself, or one that can pick up soda cans and deposit them into trash cans. Robot people have very lofty goals, however, including fielding a soccer team of 11 robots that could beat Brazil's best team, and replicating the human brain in computer software. But they're building from the bottom up; starting by solving low-level human functions and gradually making more complicated robots as we better understand the problems and ideas involved.

Macroeconomists want to understand the global economy; robot people want to understand the human brain. These are incredibly complicated systems but the approaches are totally different.

If macroeconomists were trying to understand how the human brain works, I feel like they would start by building a giant processor, giving it an initial set of instructions, then continually tinkering the instructions to make the processor more like a human. I estimate it would take around a thousand years of tinkering to produce something resembling the human brain.

If robot people were trying to understand the economy, they'd start by breaking down and trying to fully understand the components - each firm, the decisions facing each worker, and compiling those into a model of the economy as a whole. This would be a long process - but shorter than starting with the whole system and slowly tinkering.

Better to break a system into parts, understand the parts and build back up than to try to understand the whole system.

Liked what you read? I am available for hire.