You do not create experiments. You create models that try to explain the world and check your model predictions and what happens in the world. When the two disagree, you update the model and try to explain what went wrong. As you cannot make experiments to check, it is inevitable that you will end with more than one model for different school of thoughts. As is in all academic disciplines that are part of humanities. This is not like in the natural sciences where you can agree with a single model. Nonetheless, this is better than having no empirism or scientific method at all and having models built based on some logical axioms treated as the ultimate truth, as some unscientific parts of the economy do.

Moreover, in the natural sciences, you also cannot perform experiments for some parts of cosmology, for example. Nonetheless, you keep creating models in the hope that they would allow you to discover and predict more things and to interpret better the universe. If they do, your model clearly shows some benefits, even if in the future with more powerful technology, it could be disproved. Economy do not need to be so different than this.