The contemporary economic model works only so long as all factors are reducible to one another. So, for example, we can measure the value of any good produced in terms of the labor (and/or capital) used in its production only because the labor (and/or capital) is also measured in dollars, or yen, or renminbi, or euro. Which makes it challenging when we are seeking to measure health care outcomes, let us say, in terms of educational achievements. So far, the only meaningful way I have found to perform this function is the same way everyone else does: let x be some standard health care outcome; let y be some educational achievement; graph them linearly and see whether there is a meaningful correlation.
We could then add a variable; call it leisure time l. We could then run a multivariate test to see how the addition of this variable, leisure time (l) lent a different shape to our curve. As a hypothesis, I would venture that high educational achievement along with high leisure time gave rise to more favorable health care results; and that high educational achievement and superior health care outcomes would also be highly prediction of leisure time. We could then add a fourth, a fifth, a sixth . . . variable to test the relative correlation of these variables with different outcomes.
There is only one problem. All of these variables — health, education, leisure, race, gender, income, and so on — are nested within a social formation in which abstract value (akin to, but not the same as, money) mediates social relations. Lurking in the shadows, we might say, are policies that inflect health, education, and welfare in the direction of MPL (marginal product of labor) or MPC (marginal product of capital).
Here it might be helpful to hold y and k (labor and capital) equal. What if those at the lowest end of the income hierarchy were awarded a coefficient that bolstered the income advantage of those at the bottom — with a hazard rate indexed to each income level in between; or the highest end of the income hierarchy were weighted with a coefficient that plunged their income advantage to bring them equal to those at the bottom — with a (-) hazard rate indexed to each income level plummeting to the bottom.
Controlling for y and k, I would hold that we would find a positive correlation to education, leisure time, and health.
But why would we even perform this operation?
We would perform it because we know that some policy makers refuse to recognize the positive correlation between health, education, and welfare. They would like us to correlate every outcome to marginal profit. Marginal profit, however, measures only one in a wide range of dependent variables. Why not make educational achievement the independent variable; or health; or leisure; or family-time?
We have internalized a narrative about value that is toxic. It is killing us. It will kill us. But the solution is not abandoning rigorous mathematical modeling. The solution is applying it rigorously.