While MPCs are mostly known to decrease with liquid wealth, I show that they are also increasing in the permanent component of earnings. In a standard model, permanent earnings raise MPCs because they reduce the ratio of risk-free-liquid-wealth-to-risky-future-earnings, strengthening precautionary behavior. This can explain two documented facts: (i) people with high levels of liquid wealth still have significant MPCs; (ii) MPCs do not decrease with current earnings although, like liquid wealth, they increase available resources. This prediction holds in survey data. The effect is large enough to explain the stylized facts. Numerical simulations match the survey results and stylized facts.
While macro models increasingly incorporate substantial risk, the theoretical knowledge about the effect of uncertainty on consumption growth consists of intuitions from the second order log-linearized Euler equation. I show that the derivation of the log-linearized Euler equation is flawed in that it does not consist in linearizing an Euler equation but in linearizing an ad-hoc mathematical identity. I prove exactly that uncertainty raises consumption growth and makes consumption depart from a random walk. I also prove that this precautionary consumption growth is decreasing in assets, and in transitory and permanent income when income is a transitory-permanent process.
I prove analytically that, in the standard life-cycle model used throughout the macroeconomic literature, consumption is not linear but concave in the permanent income component of a standard transitory-income process. The reason for the concavity in income is the opposite of that behind the seminal result of concavity in (accumulated) wealth, established in Carroll and Kimball (1996). An increase in permanent income strengthens the need for precautionary saving, and more so at a high level of permanent income: consumption responds less to an increase in permanent income at high levels because the need for saving increases more. In contrast, an increase in wealth reduces the need for precautionary saving, but less so at a high level of wealth: consumption also responds less to an increase in wealth at a high level of wealth but now because the need for saving decreases less. I also find empirical evidence suggesting that the results holds in US survey data.
Studies based on natural experiments find that consumption responds strongly and significantly to a transitory variation in income, while semistructural estimations find no pass-through of transitory shocks to consumption. I develop a more robust semistructural estimator that relaxes the assumption that log consumption is a random walk. The robust pass-through estimate is significant and large, implying a yearly marginal propensity to consume of 0.32, close to the natural experiment findings. The robust estimator performs well in numerical simulations of a life cycle model, while nonrobust estimators do not. The difference between the two in the simulations is similar to their difference in the survey data.
In the United States, after age 65, households face income and health risks, and a large fraction of these risks are transitory. While consumption significantly responds to transitory income shocks, out-of-pocket medical expenses do not. In contrast, both consumption and out-of-pocket medical expenses respond to transitory health shocks. Thus, most US elderly keep their out-of-pocket medical expenses close to a satiation point that varies with health. Consumption responds to health shocks mostly because adverse health shocks reduce the marginal utility of consumption. The effect of health on marginal utility changes the optimal transfers due to health shocks.