Savings and Investment

Saving more is not always better: saving more means consuming less at first; in the long run, compared with what it would have been without this change, consumption per person increases only if the initial saving rate was below the golden-rule rate, that is, the saving rate that gives the highest consumption per person in the long run; beyond that rate, saving more reduces it.

A budget deficit weighs on investment, in this model: when the deficit widens, that is, when the gap between what the government spends and what it takes in grows, national saving falls, unless households save that much more to offset it. The government borrows to cover the gap; fewer funds are left for other borrowers, the real rate rises and some private projects are abandoned. This is financial crowding out; its actual size varies from one period to another and remains debated.

It is mainly the real rate that matters: a loan at 5% when prices rise by 3% a year costs only about 2% in purchasing power. Because they decide before knowing how much prices will rise, savers and borrowers look at the posted rate minus the inflation they expect.

Saving does not necessarily respond to the rate, nor always in the same direction: a higher rate makes saving more attractive, but it also makes it possible to reach the same goal, for example a sum for retirement, while setting less aside. For a saver, the two effects partly offset each other, and depending on the case one or the other prevails. In his macroeconomics textbook, Mankiw in fact draws, to keep things simple, a saving curve that does not depend on the rate, then shows the variant where saving rises with it, the one used in the definition above.

An open country can invest more than it saves: it then draws on foreign saving, by borrowing abroad or by selling part of what it owns to foreigners, and the gap between its saving and its investment shows up in its current account balance. For example, the US current account has been in deficit every year since 1982, except in 1991.

Saving more can lower income: if all households want to save more at the same time, without firms wanting to invest more (for example because the rate does not fall enough or because their sales are expected to decline), consumption falls. Sales, output and incomes fall with it; in the end, total saving may not increase. This is the paradox of thrift: the idea, already present in Mandeville (The Fable of the Bees) and then in Hobson and Mummery (1889), was placed by Keynes at the heart of his analysis (1936), and the name was popularized by Samuelson (1948). The loanable funds model rules it out: output is fixed in that model, and a fall in the rate always boosts investment.

The rate cannot always fall far enough: in 2013, the economist Lawrence Summers revived the hypothesis of "secular stagnation," formulated by Alvin Hansen in the late 1930s. When people want to save far more than projects require, the real rate that would equalize the two can become negative. Yet a posted rate can hardly fall below zero: if inflation is low, the real rate cannot fall far enough. It is then no longer the rate that equalizes the amount people want to save and the amount they want to invest, but output: it falls, and incomes with it, until desired saving drops back to the level of projects. In the major developed countries, and apparently worldwide, real interest rates on long-term government borrowing fell by about 4.5 percentage points between the late 1980s and 2015 (Rachel and Smith, Bank of England working paper, 2015). Cautiously, the authors attribute most of this fall mainly to stronger desired saving, linked in particular to population aging and rising inequality, and to weaker desired investment. Slower expected growth counts for less in their calculation.

In the United States, real rates have risen sharply since 2022: when the Federal Reserve raised its rates to curb inflation, the real yield on 10-year inflation-indexed government bonds went from about −0.9% on average in 2021 to about 2% on average in 2025, versus 0.4% in 2019 (US Treasury). The open question is whether the neutral rate, the one that neither slows nor stimulates the economy, has risen too. In April 2023, the IMF judged this rise likely to be temporary: once inflation was under control, central banks in advanced economies should ease policy and bring real rates back toward their pre-pandemic levels, more or less closely depending on how persistent high public debt proves, how climate policies are financed and the extent of deglobalization.

A country's interest rate can depend on other countries' saving, which flows in to finance its investment: in 2005, Ben Bernanke, then a member of the Federal Reserve Board of Governors, put forward an explanation that he himself considered somewhat unconventional. According to him, the US current account deficit and the low level of real interest rates on long-term borrowing worldwide stemmed in part from a "global saving glut." This saving came in particular from oil-producing countries, from emerging Asian economies and from aging rich countries such as Germany and Japan. On this reading, US rates then depended in part on saving that was not Americans' own.

Conversely, for a long time saving crossed borders less than one would expect if capital moved freely: Feldstein and Horioka (1980) found that, in 16 OECD countries between 1960 and 1974, a country's average investment rate closely tracked its saving rate, as if each economy were almost closed. This link has since loosened considerably among the countries that make up the euro area: between 1991 and 2001, it had become weak there (Blanchard and Giavazzi, 2002).