Those with More Income Do Not Hold More Wealth

Posted by Don Fullerton on Feb 15, 2010

Filed Under (Environmental Policy, U.S. Fiscal Policy)

Energy is an integral input to nearly all aspects of our economy.  Energy policies, especially policies aimed at curbing greenhouse gas emissions associated with energy consumption, thus have sizable effects on nearly all participants in our economy.  The distribution of these effects across participants is an important consideration of policy design.

The incidence of the costs of energy or climate policy manifests itself in at least two major ways.  First, policy affects the “uses side” of income, through product prices.  A carbon tax may disproportionately increase the price of gasoline and electricity, two goods that represent a higher share of expenditure for poorer households.  The uses side incidence is then regressive.  Second, policy affects the “sources side” of income, through factor prices.  A carbon tax may be more burdensome to capital-intensive industries and disproportionately reduce the return to capital.  If so, and if capital provides a higher share of income for richer households, then the sources side incidence may be progressive.

In a new working paper, however, Garth Heutel and I find that the sources side burden is regressive – just like the uses side.  The reason is that high income households do not get more of their income from capital than from labor.  The holding of wealth is not the same as the flow of income. 

The data for capital income indicate one major problem with using annual income to categorize families from rich to poor: the group with the lowest annual income has the highest fraction of income from capital.  This group includes a lot of retired individuals who have no labor income and are living off their retirement savings.  These individuals may not really be “poor” on a lifetime basis.  In other words, we would like to classify households by the stock of lifetime wealth, but instead we are classifying them by a flow of annual income.  If individuals smooth consumption over their lifetime, then total annual consumption might be a good proxy for lifetime income (or at least, a better proxy than annual income). 

So we also categorized all households in the Consumer Expenditure Survey by total annual expenditures (as a proxy for permanent income).  Still, however, the group with the highest annual expenditure has a higher ratio of labor income to capital income than does the group with the lowest annual expenditure.

Thus, the sources-side burden is also regressive.  The carbon tax applies to the energy-intensive sector, which uses relatively more capital than labor in production.  Thus it reduces the demand for capital, and it reduces the return to capital relative to the wage rate.  This places less burden on those with more labor income – the high income group.

The same occurs for annual consumption deciles instead of annual income deciles.  Both the uses-side and sources-side incidences are regressive.  When defined by annual consumption groups, however, the uses-side incidence is more regressive than when defined by annual income, and the sources-side incidence is less regressive than when defined by annual income.

This topic could use more research.  But this research so far is one indication that a cap-and-trade system or carbon tax might hurt low-income groups in two ways, instead of just one!