July 03, 2002
The Two Decade-Long Fall in America's Private Savings Rate

Of all the remarkable things to happen in the U.S. economy over the past two decades, the fall in the private savings rate must rank among the top ten. Net private savings in the United States--the sum of household savings on the one hand and business retained earnings on the other--used to fluctuate between nine and twelve percent of gross domestic product [NDP]. Then in the mid-1980s, during the Reagan years, private savings began to fall. This was a mystery: after all, the government was running substantial deficits, and there were theoretical reasons to believe that individuals might save more to offset the risk--nay, the certainty--that higher levels of government debt would one way or another increase their taxes in the future.

Some argued that the private savings rate was fallen because the 1980s stock market boom had made people wealthier, and they wanted to spend some of that wealth. But the crash in inflation-adjusted stock market values in the 1970s had not led people to save more to offset their reduced stock-market wealth. It remained a puzzle.

And the puzzle gathered strength in the 1990s. By the peak of the late-1990s boom, the private savings rate was only three percent of NDP.

The Fall in Private Savings as a Share of NDP

ChartObject Chart 2

From the National Income and Product Accounts prepared by the Commerce Department's Bureau of Economic Analysis. Nominal household savings plus nominal business retained earnings divided by nominal domestic product [NDP].
Posted by DeLong at July 03, 2002 12:57 PM | Trackback

Email this entry
Email a link to this entry to:

Your email address:

Message (optional):


I disagree. The BEA series on Personal Savings as % of Disposable Income does indeed fluctuate closely with the LOG of the Fed's series on Household Net Worth / Disposable Income going back to 1970. Savings rates have been rising this year in a lagged fashion to net worth, and should rise a further 2-3% this year based on this historic relationship. This behaviour is very rational on the part of households and suggests below-trend growth for 2002-03.

Posted by: Peter on July 3, 2002 11:38 PM

I was always impressed by the failure of households to raise savings to compensate for the large loss in real business equity and debt values in the 1970s.

Posted by: Brad DeLong on July 5, 2002 08:09 AM

1). DIfference from savings behaviour in the 1970s may be the money illusion effect.

Although economists tend to focus on real variables, individuals tend to focus on nominal values and make their decisions on nominal values. For example, wage and business contracts are denominated in nominal values. For this reason, the macroeconomy tends to have asymmetric responses to contractionary vs. expansionary impulses.

I suspect that what happened in the 1970s was that individuals responded to their nominal position ie their nominal stock of wealth was not falling.

Conversely, in the 1980s and 1990s, their nominal stock of wealth was rising and so they reduced savings rate.

2. impact of housing - for individual households, the largest single store of wealth is their house (and massively subsidised by US tax policy to be so). If housing prices rise, individuals feel wealthier, and able to spend more.

Due to inflation, demographic pressures, increased restrictions on land use (zoning, etc.), housing has been an (almost) one-way bet for the average US household over the last 30 years. I remember one of Robert Shiller's graphs showing that housing in Chicago, for instance, had never actually fallen in value over that period (Boston was a different case, but as in most boom markets, what you get were booms, followed by collapses to take the local market back to its trend rate).

3). Movement away from compulsorary savings. The period in question saw the decline of the defined benefit pension scheme, and also to some extent of long term, whole life insurance policies. Both were historically important sources of compulsorary savings by households. (the general trend of the data is right, but the timing probably doesn't align).

4). Savings behaviour is concentrated in a few households. Something like 10% of households have 90% of non-housing savings.

This is the group that experienced radically better outcomes over the last 30 years, in terms of lower tax rates at the same time as income inequality was rising. This may have been a significant impetus to spend more, as post tax investment returns and wealth would have well exceeded expectations.

5. non-housing savings is probably directly correlated with unemployment rates, which have been falling since 1980-81. In essence, households feel more able to take on debt, when their chance of their biggest single financial risk is lower. This is very obvious in current UK spending patterns (record low savings rates).

6. there has been a revolution in consumer credit since the late 1970s. It is now possible to directly and simply tap housing equity via flexible mortgage products and lines of credit. In addition, a large fraction of cars are now bought on leases, rather than for cash. Credit cards are vastly more available both to middle class but also working class individuals. Accepted debt ratios by lenders have expanded greatly, in response to greater competition in financial services.

Set against these factors, the impact of changes in marginal tax rates is, I suspect, rather small. Savings is more attractive in an after tax sense, but so is consumption.

As a closest benchmark, Canada has a similar economic structure, with a similar downward trend in savings ratios, but from a much higher level. The main differences would appear to be: in Canada unemployment has been sustainably higher over the period (due in part to more generous unemployment legislation, but also the boom-bust natural resources and automotive cycles), Canada's tax system does not subsidise house ownership to the same extent (capital gains tax fully exempted, but mortgage interest not deductible) and levies higher consumption taxes (in general). Canadians possess lower average incomes and are observable 'less flashy' in their spending habits than their American counterparts (for example, throughout the 80s GM's cheapest selling car sold double, proportionately, in Canada over the US).

7). The result of all of this has been an unprecedented integration of the US economy into the international one. Although the US has run deficits with the world since at least the 1960s and the Vietnam era boom, this has been accelerated and furthered by the tendency of the US to have low household savings vis a vis most other industrialised countries.

In effect, the US has spent the last 20 years living off the savings of other nations. Given its generally younger demographics, this is no bad thing, but the scale of the imbalance has reached the point that there must be questions about sustainability.

In general, I expect a steep and rapid fall of the dollar (many hedge funds have borrowed in foreign currencies to hold US dollar securities, they will need to unwind these positions quite rapidly) to correct some of the current account issue. I also expect US households to save more, at the cost of a sustained period of macroeconomic weakness. They will do so because of 1). demographics (retirement looms for Baby Boomers) 2). disappointing returns on financial assets (the unwinding of the stock market bubble) 3). sustained higher unemployment.

If you look at the savings and investment charts, what you see in the late 1990s is cheap capital driving lots of investment. I expect some of that to unwind, as, post Enron, Worldcom etc. and the deflation of the US stock market bubble, capital for US companies will no longer be so cheap.

In other words, the US will track (hopefully less dramatically!) what happened in Japan in the 1980s: cheap capital led to lots of investment and a boom, then eventually that unwound.

Posted by: downbelow72 on July 9, 2002 01:09 AM
Post a comment

Email Address:



Remember info?