62 Finance & economics The EconomistOctober 17th 2020
21LongcovidSource:IMF *InconstantpricesGDP*forecastmadeinOctober 2020
2019=100
110105100959085
20`2019 21 22 23 24 25USBritainJapanItalyGermanyAdvanced
economiesWorldEmerging
marketsUnitedStatesIndiaChina403020100GDP*,cumulative%increase,2019-24
Forecastmadein: Oct 2019 Oct 2020The imf’s latest forecasts, released on October 13th, spell out just how long the economic
harm from covid-19 will last. America’s gdpwill return to its 2019 level only in 2022;
Italy’s, in 2025. The fund reckons that in many places output will stay well below its
pre-pandemic trend, as labour and capital are only slowly reallocated from shrinking
industries towards thriving ones. Last October the fund expected India’s economy to
grow by more than 40% by 2024; now it expects half that.From acute to chronictunately, history is not a particularly help-
ful guide. You might think that central
banks had looked into the question, given
their low policy rates are intended to boost
consumption (and thus the economy) and
reduce how much people stash under their
mattresses. But the Federal Reserve and the
Bank of England have done surprisingly lit-
tle research into the subject.
More work has been done in Germany,
where low interest rates are a hotter politi-
cal issue. But this suggests that the impact
of rates on savers’ behaviour is murky, at
best. The Bundesbank has found that the
level of returns has become less important
over time as a determinant of savers’ be-
haviour. A study by Allianz, an insurer, also
finds that other factors play a bigger role.
The more money governments devote to
social spending, for instance, the less peo-
ple save, because they expect the state to
help them in tough times. Demography
also affects the saving rate: people tend to
save more as they near retirement. But
once retired, most live off their savings, so
an increase in the number of retirees could
cause the aggregate saving rate to fall. Re-
search by Charles Yuji Horioka of Kobe
University suggests that this has been the
main cause of the long-term decline in Ja-
pan’s household-saving rate.
To the extent one can tell, the historical
relationship between rates and the level of
savings seems to be weak. The Allianz
study finds that, across Europe as a whole,
for every one-percentage-point drop in in-
terest rates, saving rates increased by 0.2
percentage points. Even then cause and ef-
fect is hard to disentangle. Central banks
cut rates in response to bad economic
news, and such news, rather than lowerrates, may be the main reason that savers
become more cautious. America’s saving
rate fell from more than 10% before 1985 to
less than 5% in the mid-2000s. That could
have been related to the downward trend in
rates. But shorter-term fluctuations seem
to have been driven by recessions.
If history is an unreliable guide to what
savers will do now, what signals can be
gleaned from their behaviour so far this
year? Anxiety about the pandemic helped
push the saving rate in America to a record
high earlier in the year; in August it was
still relatively elevated, at 14.1%. The In-
vestment Company Institute (ici), a lobbygroup for American fund managers, re-
ports that $115bn flowed into money-mar-
ket (ie, short-term deposit) funds in March
this year. “Fear came into discussions with
clients,” says Andy Sieg, president of Mer-
rill Lynch Wealth Management. “Their con-
cern was safety of principal.” If you are wor-
ried about losing your job, then the return
on your savings is a minor concern. The
main thing is to have some.
Yet as the panic subsided some savers
turned to another strategy, of piling on
risk. The American stockmarket rallied,
due in part to central-bank action. Many re-
tail investors rushed in, buying shares
through platforms such as Robinhood.
With returns on bonds and cash so low,
stocks seemed attractive, particularly as
some offer a dividend yield that exceeds
the return savers get in the bank. For inves-
tors who turned to shares in March, this
wealth effect easily compensated them for
the lower returns on other savings. This
greater risk-taking is part of a longer-term
trend. Mr Sieg says that, ten to 15 years ago,
rich American retirees may have parked a
lot of their savings in municipal bonds.
Now they have a more diverse portfolio in-
cluding equities and corporate debt.
The approach of taking more risk to
compensate for lower interest rates has not
always paid off, though. America’s frothy
stockmarket has been an outlier. Savers
elsewhere have been less well compensat-
ed for risk. Britain’s ftse100 index is below
its level in 1999. In Germany a boom in the
1990s did cause equities to rise from 20% to
30% of household assets. But when the
bubble burst, retail investors’ enthusiasm
waned. By 2015 shares were 19% of house-
hold assets. Japan’s stockmarket is still be-
low its high in 1989. Around half of total
household financial assets is still in cash
and bank deposits, says Sayuri Shirai of
Keio University.
Moreover, not all savers are the same.
Even in America, stockmarket gains have
mainly accrued to the rich. The wealthiest
1% owns 56% of the stockmarket, up from
46% in 1990; the top 10% own 88% of the
market. One way of thinking about this is
that most people set aside cash for emer-
gencies. Poorer people may be unable to
save any more than that; rich ones can af-
ford to venture into equities.
Even if they don’t punt on stocks, ordin-
ary workers in rich countries still have ex-
posure to riskier assets through their pen-
sion schemes. But these tend to be quite
small. The median balance in an American
401(k) plan for those aged 55 to 64 was only
$61,738 in 2019. A pension of 4-5% of that
pot amounts to just $2,500-3,100 a year. In
Britain, where auto-enrolment has
brought many low-income employees into
the pension system, the median defined-
contribution pot in 2019 was just £9,600
($12,200). And the solvency of final-salaryNowhere to go
United StatesSource:FederalReserve3020100
2015100520009590851980Personal-saving rate, % of disposable income20
15
10
5
0
2015100520009590851980One-year Treasury yield, %