Physics does not want to understand all the universe. It only wants to understands how its parts work, and generally limits itself to the smallest and simplest parts.
Phisicists may study the behavior of isolated atoms of generic bonds between atoms, but if you ask them to predict what a dozen atoms will do when they get together, they will tell you "sorry that's Chemistry, not Phisics". Ditto if you want to predict tomorrow's weather, create a better strain of wheat, cook a good meal, pick up a girl at the bar...
And even the simplest "economic atoms" are already way more complicated than a tropical storm...
It seems like the first step in economics would be to define what can be known and what can't (much in the way you comp-sci chaps have done with the halting problem). Government economists not only don't seem to be able to admit this for critical parts of the economy but also seem to get it wrong a vast amount of the time. Worse still, they have the guns to make those errors manifest.