Human beings are quite good at microeconomics. That's the economics asking how individual people and businesses behave. It studies why coffee shops price their beverages in the way they do. It considers the best way to get people to pay money into a trust box. It makes useful predictions about the various methods to incentivise employees. Microeconomics isn't perfect; but it is getting better as telling us about how the world really works. You can find out a lot more in Tim Harford's superlative book The Undercover Economist.
Conversely, human beings are quite bad at macroeconomics. That's the economics asking how economies as a whole react to government policies and long-term factors. It studies why unemployment goes up and down in the way it does. It considers the best ways to increase economic growth. But it fails to make any useful predictions about when there will next be a recession. Macroeconomics isn't a complete basket case in that it has one or two successes to its name, such as the efficient markets hypothesis. Although some economists even manage to argue about that.
I'm not an economist. But as a historian of science, I have studied the history of a subject that, once upon a time, was like macroeconomics is today. For centuries, scientists couldn't make serviceable predictions or tell true theories from false ones. There were two very good reasons for this. The first was a lack of experimental data that could be used to test the theories. Experiments are difficult and easily misinterpreted. They need to give clear-cut results and to be repeatable. In any case, few saw the point of doing an experiment to prove something their theory already told them had to be correct. Early scientists lauded observation and could gather plenty of data. But their inability to test it was fatal.
The second reason early scientists failed was that no one was very interested in which theories worked in the real world. They were primarily concerned with how science could justify political and ethical conclusions. Scientific theories simply provided a way to understand nature that rationalised a particular moral or religious viewpoint. For example, Epicurus's science said everything was made up of mindless atoms, which dovetailed nicely with his ethics. Christians rejected atomism because it invalidated transubstantiation.
Modern science now has enough experimental data to choose between theories and make accurate predictions. Microeconomics is like that too. Of course, it is by no means perfect. Microeconomists draw lots of false positives, make various mistakes and torture their data. But, at heart, they all agree what they are about and how the method is supposed to work. Microeconomists and scientists are successful for quite similar reasons.
Not so for macroeconomists. As a profession, they failed the call the financial crisis and still can't decide on the best response to it. Admittedly, a few economists predicted the great recession, but probably fewer than you'd expect if all economic predictions were made randomly. Macroeconomists are also divided into schools and they argue about policy according to the lights of the school to which they belong. Here is Andrew Lilico of Europe Economics defending the Conservative policy of austerity against a Keynesian. I doubt Dr Lilico is on Ed Balls's Christmas Card list. Meanwhile, this lengthy screed by Oxford's Simon Wren-Lewis is a political polemic on behalf of the Labour Party dressed up as Keynesian analysis. I expect Ha-Joon Chang at Cambridge would call down a plague on chez Lilico and Wren-Lewis. The mere fact that macroeconomists can't agree on the most basic issues is damning evidence of how much trouble their discipline is in. Whatever is happening in the world, the Keynesian, the Marxist, the Austrian, the non-orthodox and the neo-Keynesian economist think can explain it in the light of their own theories. But they cannot say what will happen next. When they do get something right, it is simply luck. None of their theories work at the most basic level of being able to make practical and testable predictions.
Remember what I said about early scientists using their theories about nature to back up their political or religious agendas? Macro-economists appear to be engaged in the same thing. The near-perfect correlation between the views of members of different economic schools and their political inclinations suggests very strongly that the data does not determine to which school they belong. Likewise, academic economists are often Keynesians who believe in government intervention while private sector economists are more likely to be laissez faire. This contrasts with, say, engineering where the laws of physics are understood in a very similar way at universities and aerospace manufacturers.
This diversity of views is not entirely the fault of the macro-economists. They can't do many experiments because the systems that they study are too big. A central bank won't cut interest rates for half the economy and leave them the same for the other half just to compare the effect of different policies. That means macro-economists have to rely on observation rather than controlled experiments to build their theories. That was exactly the methodology employed by Aristotle to construct his scientific system. The result was a system of physics that seemed so rational and convincing that it lasted almost two thousand years. But it was wrong in almost every respect. Freudian psychoanalysts and Hippocratic physicians were in the same boat. They had elegant and reasonable systems of thought with which they could explain pretty much any observation. The pathologies of their patients could always be interpreted within the bounds of their theories. And yet the their theories were completely untrue.
It may seem rather crude to compare macroeconomics to such discredited systems of scientific thought. But the problem is not confined to economics. Any subject dominated by a lack of solid data and beholden to theory will face the same issue. Within modern science, string theory is well up this particular creek. Arguably, climate science is as well. The failure of computer models to predict the climate has increased the importance of theory and hardened the political allegiance of those engaged in debates on global warming.
What is the answer? For macroeconomics, there probably isn't one. The economy is too complicated to predict, even if we understood exactly how it works, which we don't. So it is best to understand economic debates as proxies for political arguments. The good thing about that is we can all get involved. Macroeconomics boils down to informed opinion and when it comes to opinions, we all have one.