Unit 3 Aggregate demand and the multiplier model

3.10 Shocks to households and the limits to smoothing consumption

Figures 3.10a and 3.10b (in Section 3.5) show that, in the data, consumption is less volatile than investment and smoother than GDP as a whole. But it is by no means completely smooth. Indeed if it were completely smooth, as assumed in Figure 3.17, then the marginal propensity to consume (MPC)—which is the basis of our multiplier model—would be precisely zero! In this section, we examine the limitations of the strong assumptions made by the consumption-smoothing model, and demonstrate that this leads naturally to an aggregate consumption function with a strongly positive MPC, and therefore a multiplier of above one.

How do households respond to unexpected events—or shocks—to themselves, their family, and to the economy as a whole? They would like to smooth their consumption but may not be able to do so.

As discussed in Unit 1, the term ‘shock’ is used in economics to refer to an unexpected positive or negative event, whether it affects an individual, a sector or region, or the entire economy. As we know, people think about the future and anticipate that unpredictable events may occur. When anticipating shocks affecting their own household, they also act on these beliefs. In a modern economy, this is the basis of the insurance industry.

We can distinguish between two situations:

• Good or bad fortune strikes the household: For example, when a family member who plays an important role in household income gets a bonus or is unable to work because of an injury. This is called an idiosyncratic shock.
• Good or bad fortune strikes the economy as a whole: For example, when a new technology raises the productivity of a staple crop or when the price of a crucial imported input like oil goes up. This is called an economy-wide or common shock.

People use two strategies to deal with idiosyncratic shocks—shocks that are specific to their household:¹

self-insurance

To maintain their consumption, households can use savings and borrowing to self-insure against a temporary fall in income or need for greater expenditure.

co-insurance

A co-insurance scheme enables households to pool savings so that individual households can maintain consumption when they experience a temporary fall in income or the need for greater expenditure.

• Self-insurance: Households that encounter an unusually high income in some period will save, so that when their luck reverses, they can spend their savings. They may also borrow in bad times if they can. It is called self-insurance because other households are not involved.
• Co-insurance: Households that have been fortunate during a particular period can help a household hit by bad luck. Sometimes this is done among members of extended families or among friends and neighbours. Since the mid-twentieth century, particularly in richer countries, co-insurance has taken the form of citizens paying taxes, which are then used to support individuals who are temporarily out of work or when there is an economy-wide shock (unemployment benefits, furlough payments in the pandemic).

Co-insurance at a local level between households is less effective if the bad shock hits everyone at the same time. In the case of an economy-wide shock, co-insurance at the level of the economy is necessary to help the households worst hit and least able to smooth the shock themselves. This was very evident in the COVID-19 pandemic.

preferences

A description of the relative values a person places on each possible outcome of a choice or decision they have to make.

Self-insurance and co-insurance reflect two important aspects of the preferences of many households:

• Households prefer a smooth pattern of consumption: They dislike consumption that fluctuates as a result of bad or good shocks, so they will self-insure.
• But they may have a bias towards the present: In some circumstances, households may prefer consuming their income now, rather than saving for the future.
• Households are not solely selfish: They are willing to provide support to each other to help smooth the effect of good and bad luck. They often trust others to do the same, even when they do not have a way of enforcing this. Support for the tax and benefits system reflects the preference for smoothing.

Limits to consumption smoothing

What limits a household’s consumption smoothing? Many individuals and households are not able to make or implement long-term consumption plans. Making plans can be difficult because of a lack of information. Even if we have information, we may not be able to use it to predict the future with confidence. For example, it is often very hard to judge whether a change in circumstances is temporary or permanent.

When faced with income shocks, the extent of smoothing also depends on:

credit constraints

Credit constraints are restrictions on the amounts or terms on which individuals can borrow.

credit market excluded

A description of individuals who are unable to borrow on any terms. See also: credit market constrained.

• Credit constraints or credit market exclusion: This restricts a family’s ability to borrow in order to sustain consumption when income has fallen.
• Limited co-insurance: Those with a fall in income can expect only partial support in sustaining their incomes from others more fortunate than them, including from the government’s co-insurance schemes.

Credit constraints

The amount a family can borrow is limited, particularly if it is not wealthy. Sometimes households with little money cannot borrow at all, or only at extraordinarily high interest rates. Therefore, the people who most need credit to smooth their consumption are often unable to borrow.

Figure 3.18 shows the reaction of two different types of household to an anticipated rise in income. Households that are able to borrow as much as they like are in the top panel. Credit-constrained households that are unable to get a loan or a credit card are in the bottom panel. Follow the analysis in Figure 3.18 to understand how the two households react differently to two key events:

1. First, news is received that income will rise at a predictable time in the future (for example, a promotion or a bequest).
2. Later, the household’s income actually rises (the promotion happens, the inheritance comes through).

A temporary change in income affects the current consumption of credit-constrained households more than it does that of the unconstrained.

Present bias

present bias

An individual or household with present bias puts more weight on present rather than future consumption. They may prefer to consume all of their current income now, even when they know that income will be lower in the future.

In Figure 3.19, an individual learns that income is going to fall in the future. This could be because of retirement or job loss. It could also be because the individual is becoming pessimistic. Perhaps the media predicts an economic crisis. In the top panel of Figure 3.19, we again show a household behaving in a forward-looking manner to smooth consumption. The bottom panel shows a household with present bias that consumes all its income today even though it implies a large reduction in consumption in the future.

Although a household may want to avoid a sudden drop in consumption when income falls, the desire to postpone a fall in living standards (present bias) means they may not react to the news at all, and keep consumption high until income falls. The problem of not being able to save differs from the problem of not being able to borrow: there is no constraint on saving, and as a form of self-insurance it doesn’t involve anyone else. This feature of human behaviour is familiar to many of us. When the time comes, we often lack the willpower to implement a plan we made earlier. Economists refer to this as time-inconsistent behaviour. While there are other possible explanations for failing to save ahead of a known fall in income, time inconsistency is one for which there is evidence. (Read the box ‘How economists learn from facts: My diet starts tomorrow’.)

Limited co-insurance

Most households lack a network of family and friends who can help out in substantial ways over a long period when a negative income shock occurs. Unemployment benefits provide this kind of co-insurance—the citizens who turn out to be lucky in one year insure those who are unlucky. But in many societies, the coverage of these policies is very limited.³

A vivid demonstration of the value of smoothing through co-insurance is the experience of Germany during the drastic reduction in income that occurred there in the global financial crisis in 2007–2009. When the demand for firms’ products fell, workers’ hours of work were cut, but as a result of both government policy and agreements between firms and their employees, very few Germans lost their jobs, and many of those at work were still paid as if they were working many more hours than they did. The result was that although aggregate income fell, consumption did not—and unemployment did not increase. Employers ‘banked’ these hours and their employees worked the extra hours later when demand rebounded.

But most empirical evidence shows that credit constraints, present bias, and limited co-insurance mean that, for many households, a change in income results in an equal change in consumption. In the case of a negative income shock such as the loss of a job, this means that the income shock will now be passed on to other families who would have produced and sold the consumption goods that are now not demanded.

Figures 3.10a and 3.10b showed that consumption in middle-income countries was more volatile than in high-income countries. This reflects a combination of more widespread credit constraints and more limited government systems of co-insurance.

Consumption behaviour and the multiplier model

Whether or not households are able or willing to smooth consumption is crucial for the multiplier model set out in Sections 3.7 and 3.8.

We can think of the multiplier model as applying to an economy that consists of two types of households. Some households smooth their consumption by borrowing and saving: we can think of these as having an MPC of zero \((c_1 = 0)\). But others, due to some combination of credit constraints and present bias, have an MPC of 1 \((c_1 = 1)\).

If we average across the two types of households to find the aggregate consumption function, this gives an average MPC between 0 and 1 as in the multiplier model. Whether the aggregate MPC is high or low therefore depends on the proportion of households in each group.

Shifts in the proportion who are credit constrained and therefore have a high MPC affect the size of the multiplier. This helps explain why the multiplier is found to be higher in deep recessions than it is in booms.