A proper stimulus? A wonkish look at tax rebates

2001 Rebates- Consumer Surveys

The 2001 rebate package, which granted $38 million in rebates to an estimated 92 million taxpayers ($300 for singles and up to $600 for married couples), was not nearly as substantial as the pending 2008 rebates. Still, University of Michigan economists Matthew Shapiro and Joel Slemrod (2003) estimate that a $600 rebate in 2001 represented  about 1.5 percent of median annual income.

Using telephone consumer surveys, Shapiro and Slemrod found the effect of the 2001 rebates on consumption underwhelming: only 21.8 percent of households surveyed planned to spend their rebate, while a higher proportion planned to save or pay down debt.  A follow-up survey found only 24.9 households reported spending their rebate.

Surprisingly, Shapiro and Slemrod’s survey found that the propensity for the rebate to increase consumption spending was greater for higher-income taxpayers (table 1).

Table 1.

Income ($) Frequency (%) Spend percentage

0 to 20,000 18.9 17.6

20,001 to 35,000 18.6 18.8

35,001 to 50,000 16.6 18.6

50,001 to 75,000 17.4 27.0

75,000 + 21.3 24.1

Shapiro & Slemrod (2003)

Table 2 below shows that households optimistic about their future financial condition are generally more likely to spend their rebate than those expecting to be in a worse financial condition next year. Contrary to the liquidity restraint hypothesis, households who expect to be in a temporarily bad financial condition (16%) are less likely to spend their rebate than those in a temporarily good financial condition (22%).

Table 2. ***

What might explain the relatively low propensities to spend out of rebates, especially among low-income or liquidity-restrained households? First, valuation and spending decisions depend heavily on changes in wealth levels rather than absolute wealth levels (Thaler 1999). This valuation effect might explain why ‘worse-off’ individuals were less likely to spend their rebate checks-the rebate money was instead absorbed in order to restore previous wealth levels. Secondly, the propensity to consume out of windfall gains depends on the size of the gains (Thaler 1990). Windfalls gains that are small relative to current disposable income tend to be coded as current income and are spent at a much higher rate than relatively larger gains, which tend to be coded as assets. Mental accounting, then, might help explain why the propensity to consume the 20001 rebates was greater for high-income individuals. Thirdly, there are framing effects that might yield lower propensities to spend out of the rebate.

Framing Effects

Experimental studies demonstrate that variations in the framing of options yield different preferences (Kahneman & Tversky 1992). A relevant example is the framing of economic windfalls: experimental evidence routinely shows that the propensity to consume out of income which is coded as a positive shift from a current reference point (e.g. a bonus) is much greater than identical income which is coded as a return to that reference point (e.g. a rebate).

Nicholas Epley and Ayelet Gneezy of the University of Chicago GSB (2007), for example, run a laboratory experiment in which participants spend significantly more of a $50 windfall when it is described as a bonus than when it is described as a rebate. Participants in the bonus condition spent nearly 2.5 times more of their windfall than participants in the rebate condition. In fact, those receiving a so-called ‘rebate’ were much less likely to spend the money at all. The authors conclude that framing income as a gain from the status quo, instead of as a returned loss, significantly increases propensity to consume.

These results run counter to traditional theories of consumption, which predict that individuals will ‘smooth’ over the consumption of windfall income regardless of its description. Instead, the Prospect Theory of valuation and the concept of mental accounting appear to explain this behavior. Individuals are more likely to spend income when it is unexpected and unearned, as windfall income is less likely to be incorporated into wealth accounts (Thaler 1999; Epley and Gneezy 2007). One might conjecture that the federal government’s extensive information campaign notifying the public of impending tax rebates, including a notice sent to every eligible recipient, might decrease propensity to consume by heightening expectations and senses of entitlement.  

Please share widely!

0

0