The EconomistFebruary 22nd 2020 Finance & economics 73M
oney makesthe world go round. But
where in the world is it going? In the-
ory the answer lies in statistics published
by the likes of America’s Treasury Depart-
ment and the International Monetary Fund
(imf), which track cross-border flows of
debt and equity investments. In practice
creative corporate accounting mucks up
the official figures. A growing body of re-
search is trying to clean up the mess.
Accounting for corporate behaviour
would be simple, if it could be organised
neatly according to national boundaries.
Petrobras, a Brazilian oil giant, would sell
its bonds directly to American or European
investors. In reality many companies raise
funds through foreign subsidiaries set up
for the purpose. Petrobras raises debt
through its subsidiary Petrobras Global Fi-
nance bv, based in the Netherlands.
Such financial contortions cloud econ-
omists’ view of global investments. When a
subsidiary transfers the cash from the loan
to its parent, this can show up (misleading-
ly) as foreign direct investment (fdi). A re-
cent study by Jannick Damgaard of the
Danish central bank, Thomas Elkjaer of the
imfand Niels Johannesen of the University
of Copenhagen estimated that as much as
40% of measured fdiflows were in fact
“phantom” flows through shell companies.
Another recent study by Antonio Cop-pola of Harvard University, Matteo Mag-
giori of Stanford University, Brent Neiman
of the University of Chicago and Jesse
Schreger of Columbia University also tries
to estimate just how much these financial
gymnastics are distorting the official fig-
ures. Their motivation is the dramatic rise
in the share of cross-border investments
flowing through tax havens between 2007
and 2017, from around 40% to 55% for debt
and from 10% to 50% for equity capital.
The authors stitch together seven data
sets, including information on investor
portfolios and the relationships between
parent companies and their subsidiaries.
Both debt and equity investments fun-
nelled through tax havens are reallocated
to the location of the parent company.
What the official figures count as an Ameri-
can loan to Petrobras Global Finance bvin
the Netherlands, the researchers reclassify
as a loan to Petrobras in Brazil.
This amounts to a large adjustment to
the official data. Around 12% of what looks
like American investors’ holdings of for-
eign bonds turns out to be their holdings of
domestic ones, mostly in the form of bun-
dled corporate loans. American holdings of
corporate bonds in emerging markets are
revised up too: from $8bn to $44bn in the
case of Brazil, and from $3bn to $37bn for
China (see chart). As these bonds do nottend to be issued in the parent company’s
currency, the adjustments bump up emerg-
ing markets’ exposure to foreign-currency
debt, from 20% to 50% of Brazilian debt se-
curities held by foreigners, and from 30%
to 60% for Russian securities.
One of the most drastic revisions is to
American holdings of equities in China,
which official figures suggest were worth
$160bn in 2017. That relatively low figure
reflects the fact that the Chinese govern-
ment restricts ownership in some key sec-
tors. But it does not reflect Americans’ true
financial stakes. In order to skirt the con-
trols, Chinese companies set up “variable-
interest entities”, which attempt to repli-
cate the benefits of raising equity capital
without falling foul of the rules. After the
authors’ adjustments, Americans’ stake in
Chinese equities rises to $700bn.
All this shows that rules often lead only
to financial gymnastics. The findings also
expose risks. Some large emerging markets
are much more exposed to currency depre-
ciations against the dollar than the official
data suggest. If a crisis hits, governments
will have to sort through the tangle of loans
owed by domestic companies but held out-
side their jurisdiction. Mr Coppola and his
co-authors call attention to American re-
tail investors’ exposures to Chinese finan-
cial vehicles, which China’s government
could rule as illegal (and worthless) on a
whim. Economists may find incorporating
these cross-border financial connections
into their models hard. But without them,
the picture is grossly misleading. 7WASHINGTON, DC
New research shows that cash sloshes around the world in unexpected waysCross-border capital flowsFollowing the money
Corrective measures
US holdings in selected markets
2017, $bnSource: Global Capital Allocation ProjectSouth
AfricaRussiaIsraelChinaBrazil50403020100Corporate bondsAdjusted for flows
through tax havensOfficial dataSingaporeRussiaTaiwanChina0 100 200 300 400 500 600 700EquitiesIndiaFirst in class
United StatesSources: Federal Reserve Bank of New York; US Department of Education *12 months after completion of bachelor’s degree in 20161.51.00.50
19171513110907052003Household debt, $trnStudent loansCredit cardsCar loansAsianHispanicWhiteBlack50403020100Women MenCumulative student debt*, $’000On February 18th Mike Bloomberg became the latest Democratic presidential candidate
to promise to tackle student debt in America. The stock of loans amounts to over
$1.5trn, around 7% of gdp. Fully 45m Americans owe an average of $37,000; a fifth are
struggling to make repayments. Despite making up 56% of the student population,
women owe two-thirds of total debt.Young, gifted and in the red