Labor Income and Substitution Effects

An income effect refers to the effect a change in income has on something. A substitution effect refers to the propensity of a person to substitute goods under some condition. This article discusses the labor income effect and the labor substitution effect under a change in taxation in an attempt to assess the change to savings.

This article summarizes the labor income and substitution effects.

Income equals consumption plus savings. A change in labor demand means a change in income which in turn applies possible changes to consumption and savings.

The labor income effect states that people are threshold savers, although you will not hear other people put it that way. For this reason, they will save more money, and consume less, when income goes down in an attempt to preserve future consumption. A threshold saver is somewhat the opposite of a threshold earner as described by Tyler Cowen. The way most economists will describe the income effect is to say that people are consumption-smoothing, but the income effect does not follow from such a claim because if people are consumption smoothing and their income drops they certainly wouldn’t consume even less!

The substitution effect says that people will make up their current consumption resulting from a loss of income by saving less.

Which effect is stronger? The jury is still out, but I am going to argue that the substitution effect dominates. The implication is that a reduction in income leads first to a reduction in savings and only later to a reduction in consumption. The further implication is that an increase in taxes is expected to reduce savings and investment in the economy.

Here is a recent Mankiw article discussing that the jury is still out.

5 reasons I think the substitution effect dominates, from weakest to strongest:

  1. I want to say that the substitution effect dominates in the goods market, but my data reference is escaping me. If that fact is true, then we might suspect a similar relationship in the labor market.
  2. Anecdotally, when I suffer a negative shock to income (or an unexpected expenditure), my behavior has been to reduce my savings more than my consumption. This is also observed anecdotally in many people.
  3. If the expected value of the amount saved is the average of the minimum and maximum possible savings, which is equal to 50% of the maximum possible savings, which is the rationally ignorant expectation, then a decrease in future returns will result in an expected decrease in savings.
  4. In statistics, when two parameters are tested for difference they are assumed to be equal in the null hypothesis. If the jury is still out on whether the income effect is different from the substitution effect in value, then it should be assumed that they are equal according to econometric practice of a failure to reject the null hypothesis. If the values are equal as a percent of income, then the proportion of income saved will be constant regardless of income. If the proportion saved is constant and the income is lowered then the total savings is also lowered.
  5. If the threshold saver is a true threshold saver then in some cases they will be required to have consumption which is less than or equal to zero in order to make up future income. Such a scenario is arguably unreasonable or impossible, and is at least unsustainable as such individuals will eventually die of starvation. This makes total threshold saving untenable. Total consumption preference, however, is tenable. In that scenario a person would be willing to sacrifice all of their future savings, and even take on future debt, in exchange for present consumption. That scenario is frequently observed. Because total consumption preference is sustainable, tenable, and observed, while total savings preference is none of the above, consumption preference is expected to dominate according to the bias of possibility. Consumption preference is associated with the substitution effect.

The bias of possibility is the principle that X is more likely to occur than Y simply because X is possible and Y is not possible. In the case of point 5, this affects the expected value, given that the expected value is the average of the possibilities.


Leave a Comment