Eliciting the Marginal Propensity to Consume in Surveys

2022-06-08

2022-10-15

Marginal Propensity to Consume (MPC)

The marginal propensity would be computed as the change in spending reporterd over the amount of the hypothetical income windfall.

Savings and debt responses to a hypothetical windfall

We randomly split the Innovation Panel of Understanding Society into groups, and to ask respondents within each group different variants of a question about how their spending, saving and debt responses would respond to receipt of a hypothetical windfall.

In one half of the sample we would ask how much individuals would change their spending after an initial routing question (“would you increase spending, decrease spending or leave spending the same?”). In the second half of the sample, we would immediately ask for a pound change in spending without this routing. Within these treatments, we would also randomly assign individuals to either report how their spending, saving, debt and transfer behaviour would change over the next three months, or to report how it would change over the next twelve months and to either report how they would respond to either a £500 or a £2500 windfall.

We will then compare response across different treatments.

We will randomly split individuals into eight equally sized groups of households in the Innovation Panel of Understanding Society. Each group of households will be asked a different question.

Let the groups be called treatment groups 1a-1d, treatment groups 2a to 2d. Those in treatment group 1 will first be asked whether a windfall would lead them to increase their spending, decrease spending or keep spending the same before being asked for the amount in pounds and then being asked whether unspent funds would be used to save, pay down debt, give more to friend and family or receive less from friends and family. In 1a individuals will be asked how their spending would change in response to a £500 windfall over the next 3 months, in 1b they will be asked about how their spending would change in response to a £500 windfall in the next 12 months. In 1c individuals will be asked how their spending would change in response to a £2500 windfall over the next 3 months, in 1d they will be asked about how their spending would change in response to a £2500 windfall in the next 12 months.

Individuals in treatment group 2 will be asked how much they would spend straightaway without the initial routing question. Then they will be asked whether unspent funds would be used to save, pay down debt, give more to friend and family or receive less from friends and family as before. In 2a individuals will be asked how their spending would change in response to a £500 windfall over the next 3 months, in 2b they will be asked about how their spending would change in response to a £500 windfall in the next 12 months. In 2c individuals will be asked how their spending would change in response to a £2500 windfall over the next 3 months, in 2d they will be asked about how their spending would change in response to a £2500 windfall in the next 12 months.

The randomisation to treatment will be a simple random allocation of households within primary sampling units, such that all members of a household are assigned to the same treatment.

Household

Yes

1,500 households

Aproximately 2,000 individuals

Aproximately 190 households per treatment arm (1,500/8=187.5)