Democratic Measures of Income Growth and Other Indicators of Welfare

Summary

It is well established that GDP growth is not a good measure of welfare change. As an aggregate it does not reflect any changes in the distribution of income. It could be described as plutocratic because a given percentage increase has more impact on GDP growth if it accrues to high earners than if it accrues to low earners. Addressing these shortcomings, we have developed democratic measures of growth, which represent the average of each household’s growth experience. We are extending these measures to produce national and sub-national indices, which will better reflect how welfare has changed over time and across regions.

Overview

GDP was designed as an indicator of economic activity, but it is widely used as an indicator of welfare. Economists are well aware that a welfare indicator has to be net rather than gross of depreciation. Equally importantly, a single welfare aggregate needs to represent some aggregate of the welfare of the population. GDP, which adds up individual incomes, can do that only if welfare rises linearly in income; even then, it remains silent about the distribution of welfare. There is a growing body of evidence which suggests that welfare rises less and less with income, as income rises. This suggests that a measure of welfare constructed from the geometric rather than arithmetic mean of income would be appropriate. The focus, however, needs to be on the real rather than the nominal income growth of each household. In constructing real income growth, the appropriate deflator needs to take into account the fact that different households have different consumption patterns. Democratic measures of income growth can be constructed for several possible definitions of income; our focus has been on household income and national income.

By combining metrics related to real household disposable income, per capita expenditure, material deprivation, health and leisure time, we are also creating a detailed composite measure of welfare. This will provide valuable information on how welfare differs across regions and how it has developed over time.

Methods

The study was able to draw on the work on democratic price indices carried out by ONS, who also produced price indices with a treatment of housing consistent with the definition of household income in the Family Resources Survey (FRS). This allowed for the calculation of a democratic measure of household income growth. For the production of a democratic measure of national income growth it was further necessary to attribute public consumption, and the consumption of non-profit institutions serving households, to particular households. We drew on ONS work on education and health consumption and allocated other consumption on the basis of household size. We also had to allocate the income which was not distributed to households, such as retained profits or pension fund investments, which are not distributed to households. Finally, there was an issue that income is generally under-recorded in household surveys. Where there was no further information, we scaled the household data to match the national accounts aggregate. Where possible, however, we drew on other data sources, including the Survey of Personal Incomes and the Wealth and Asset Survey.

By drawing on ONS experience in small area estimation, the research team is exploring the possibility of producing consumption and expenditure data from the Living Cost and Food Survey (LCF) for disaggregated geographies below the regional level. The Annual Population Survey provides an additional source of data on how people perceive their own well-being. By combining these data with the FRS, the LCF and the Understanding Society survey, the research team has a rich data source to develop a composite measure of subjective wellbeing at sub-national level with a focus on mental and physical health, as well as social participation.

Impact

So far, we have provided ONS with a template for producing democratic measures of growth. The OECD are implementing related methodologies. Our stochastic imputation methods have been used more widely as a means of enhancing survey data.

The methods developed in this project have benefitted from significant engagement with peers through meetings of the European Economic Association, the Royal Economic Society, the Society for the Study of Economic Inequality, the International Association for Research in Income and Wealth, and the World Inequality Database.