Economic models resist testing because agents self-select into treatments

Evidence

Like all scientists, economists face the problem of how to persuade people that what they say is true. The abstract quality of many economic models can sometimes make them useful as political blueprints but difficult in practice to put to the test. Setting up experiments using human subjects is expensive, complex, and unreliable. The alternative is to use what actually happens in economic life as information against which to test an explanatory model. But this too presents difficulties. Economists readily admit that not everything is observable or measurable. The changes in a variable whose effect one is studying may be due to a factor outside the model. And the agents whose actions one is studying come with different preferences and abilities, which can affect the outcome. For example, in the Peruvian case, those planning to work outside the home might be more inclined to seek property rights, resulting in a process of self-selection (Rosenzweig and Wolpin 2000; Angrist and Krueger 2001). — Philip Mirowski & Dieter Plehwe (eds.), The Road from Mont Pèlerin, 1935

Categorization

• XREF: connects to: Judging whether markets conform to a model is itself a contested process. — Donald MacKenzie, An Engine, Not a Camera: How Financial…, 114; Economics tests claims by treating the world itself as a laboratory. — Philip Mirowski & Dieter Plehwe (eds.), The Road from Mont Pèlerin: The Making…, 1915; Modigliani-Miller propositions resist empirical testing because risk classes cannot be fixed — Donald MacKenzie, An Engine, Not a Camera: How Financial…, 389 (+1 more)

Broader Connections

• Modigliani-Miller propositions resist empirical testing because risk classes cannot be fixed #exemplifies (0.90) — Donald MacKenzie, An Engine, Not a Camera, 389