Source of income effects for demand decisions and taxable consumption

Abstract

The relationship between sources of income and demand decisions by the household is examined here with an eye toward the ramifications on consumption tax bases. Income sources may be important when households attach psychic and transaction costs to individual purchases or when sources are assigned via a mental accounting process. In either case, general and specific sales tax bases may be affected by changes in income composition. Empirical results indicate two important findings. First, tax exemptions can introduce significant income source effects for a general consumption tax base. Second, the importance of differential tax rates for gasoline and food-at-home strongly depends on the mix of labour, capital, retirement and non-retirement transfer pay.

Notes

¹ Such shifts would also affect income and property taxes but we focus here on consumption taxes only.

² The food and clothing example is taken from Heath and Soll (1996).

³ Interestingly, the authors of this study suggest an alternative where casual-better habits drive the bias. Betting habits at the racetrack, however, can be viewed as a different type of mental accounting.

⁴ In a related, but specialized study, Del Boca and Flinn (1994) investigate the expenditure decisions for divorced mothers and find some evidence that the propensity to consume children's goods out of own income is different than out of transfer income (where transfers are child support plus alimony).

⁵ A related research area is the relationship between tax cuts and consumption spending. Shapiro and Slemrod (2003) find a relatively low level of spending for the 2001 rebate, while Soules (1999, 2003) find higher responses to tax refunds and tax cuts.

⁶ In this formulation, we do not explicitly consider the tax consequences of using capital income for current consumption, but they are part of the transaction cost for that income source. This model also reflects the planned savings/consumption decisions of the individual or household in that the psychic costs may be largely determined by the savings/consumption plan.

⁷ Consumption will also be affected by changes such as age, family size, labour behavior, as these factors influence demand for specific goods. We control for these as best possible in the empirical model.

⁸Barrow and McGranahan (2000) found this type of spending effect for households receiving the Earned Income Tax Credit.

⁹ As discussed in Mullins and Wallace (1996), price variation for CES participants can be difficult to measure. For this study, the difficulty is exacerbated by the single-annual income source observation for each household. Each period can include price increases and price decreases.

¹⁰Mullins and Wallace (1996) and Hawkins (2002) find that the elderly tend to purchase more services and less durables, while younger families with more children spend more on non-durables and less on services. Age is therefore an important factor when estimating the impact of income sources as we hope to distinguish between changes in consumption due to aging versus changes in consumption due to the type of income received by households.

¹¹ For wage and salary income, gasoline may have a negative psychic cost since most workers consider it a necessity for working.

¹² In fact, wage earners may also see a production efficiency gain associated with more frequent purchases of food away from home (examined in Iorworth and Whalley, 2002).

¹³ We note that preferences for food and services are consistent with some value added taxes in Europe as well.

¹⁴ A coefficient of 0.037, for tax Base 2, means that if the income share variable increases from 0.22 to 0.23, the base should increase from 0.58 to 0.58037.

¹⁵ It is important for the reader to remember these effects occur when the household's reference-person age is held constant.

¹⁶ Obviously, the characteristics of the jurisdiction will also affect the policymaker's interest. For example, in jurisdictions where the working poor are more likely to use public transportation, the motor fuel tax rate would affect less working poor.

¹⁷ For example, while earlier studies (Mullins and Wallace, 1996) showed that the elderly tend to consume less taxable items, we show here that one must also consider the impact that composition of their income has on consumption as well. Retirees who receive mostly retirement transfer income may behave quite differently from those who gain more of their income via capital.