Saving Leakage

The fiscal multiplier depends on a chain of spending: government pays workers, workers buy goods, shopkeepers earn income and buy more, and the cycle continues. But each link in the chain is weakened by saving leakage—the portion of income that is set aside rather than spent. As incomes rise, households and firms save more. That money does not immediately cycle back into shops and factories, dampening the multiplier effect. In economies with high saving rates, the multiplier is far smaller than in those where people spend freely.

How saving shortens the chain

Imagine the government injects £100 million into the economy by paying workers on a public construction site. These workers now have extra income. If they were to spend all £100 million on goods and services, the income would circulate: grocers earn £100 million, they buy from suppliers, who earn £100 million, and so forth. The multiplier would be infinite.

But workers do not spend everything. They save some—perhaps 30 per cent. So £70 million gets spent in shops and £30 million goes into bank accounts or investments. Only £70 million circulates into the next round.

Those shopkeepers in turn earn £70 million. They spend 70 per cent of that increment, saving 30 per cent. So £49 million circulates to the next round, and £21 million leaks into savings.

The chain continues, each round 30 per cent smaller than the last. The total multiplier becomes finite—in this example, roughly 2.3 rather than infinity. The marginal propensity to consume (the fraction of extra income spent, here 0.7) determines the multiplier size. Mathematically, the multiplier equals 1 / (1 – MPC).

What drives saving decisions?

Saving is not uniform across households. Wealthy people save more; poor people spend more of each extra pound. Young people often borrow and spend; older people save for retirement. In boom times, confidence is high and spending rises; in slumps, precaution increases saving.

Interest rates matter too. Higher returns on savings (higher deposit rates, rising stock prices) make saving more attractive. Workers earning extra income in a world of 5 per cent deposit rates save more than in a world of 0.5 per cent rates. Conversely, in very low-rate environments, saving feels unrewarding, and the MPC rises.

Unemployment risk shapes saving profoundly. Workers fearing job loss save extra; workers confident of continued employment spend freely. During financial crises, saving spikes as households hoard cash. During confident expansions, saving falls.

Life-cycle factors are important too. A 25-year-old with earnings decades ahead will spend most of a bonus; a 65-year-old near retirement will save it. Societies with younger populations thus have higher MPCs and larger fiscal multipliers. As populations age, saving leakage increases and fiscal stimulus becomes less potent.

Historical variation in saving behaviour

The United States has long had a low saving rate by developed-country standards—around 3–5 per cent of disposable income in normal times. The MPC is thus high, perhaps 0.85–0.90, yielding multipliers of 6–10 if only saving leakage operated.

Continental Europe and East Asia have much higher saving rates, 15–20 per cent or more. The MPC is lower, perhaps 0.65–0.75, yielding multipliers of 2.5–4.

During the COVID-19 pandemic, saving behaviour inverted dramatically. Lockdowns made spending impossible; government cheques and enhanced benefits swelled bank balances. Household saving surged to 30–35 per cent in some countries. The MPC collapsed, and fiscal stimulus became less effective per pound spent. As savings were depleted and economies reopened, the MPC recovered and stimulus became more potent again.

The relationship to consumption functions

Economists model saving through the consumption function, which links consumption to income (and wealth). A simple version says consumption = autonomous spending + MPC × income. An increase in government spending raises income and consumption partly, but the extra consumption (MPC × extra income) is only a fraction of the income gain.

More sophisticated versions add wealth effects and permanent-income logic. If people view fiscal stimulus as temporary, they save most of it—the MPC on temporary income is very low, perhaps 0.1–0.2. If people believe stimulus is permanent—funded by sustained higher government spending or backed by central bank guarantees—they spend more, and the MPC rises to 0.5–0.8.

This explains why stimulus cheques sent in 2008 (widely viewed as temporary) had weak multiplier effects, whilst stimulus in 2020 (presented as backing for households through temporary emergency) had larger effects. Expectations about permanence drive the MPC.

Saving leakage in an open economy

Saving leakage combines with import leakage to shrink the multiplier further. In an open economy, extra income is divided three ways: spending on domestic goods, saving, and spending on imports. Both saving and imports drain demand from the domestic multiplier chain.

A typical advanced economy might allocate extra income as: 55 per cent domestic consumption, 15 per cent saving, 15 per cent imports, 15 per cent other. The multiplier would be severely constrained by both leakages combined.

In a very open, very high-saving economy—say, Germany or Singapore—the multiplier might be just 1.1 to 1.3. In a closed, low-saving economy like the United States in the 1960s, multipliers might exceed 2.0. The same fiscal policy has radically different effects depending on the underlying propensities to save and import.

Policy leverage: why certainty matters

Because saving leakage is so consequential, governments have incentives to influence the MPC. One approach is to credibly commit to future stimulus. If households believe current stimulus is backed by a long-term government commitment, they raise their MPC. They save less because they know future income is secure.

Another approach is to target spending to low-MPC groups: subsidies to poorer households (who spend most of incremental income) have a larger multiplier than broad tax cuts that reach wealthier households (who save most of the benefit).

Conversely, central banks can influence saving leakage through interest rates. Ultra-low rates reduce the reward to saving, raising the MPC automatically. This was the strategy during 2008–2014: rates near zero made saving feel punitive, pushing households and firms to spend.

When saving leakage dominates

In some environments, saving leakage is so severe that fiscal stimulus barely moves the needle. This happens when:

• Crisis psychology dominates: Fear overwhelms income growth; households save most new earnings to shore up precautionary reserves.
• Population is very old: Retirees save much of extra income for heirs or insurance against future care needs.
• Households are already heavily indebted: Borrowers facing debt obligations save windfalls to repair balance sheets rather than spend.
• Interest rates are very high: The reward to saving is so attractive that the MPC collapses.

In such periods, fiscal stimulus is far less effective. The government can inject demand, but much leaks into savings. Economists in the 1990s observed that Japan’s multipliers had shrunk to 0.5–0.8, well below historical norms, as ageing and financial distress drove up saving.

See also