By John Kemp
LONDON May 1 Statisticians at Britain's Office
for National Statistics (ONS) could be forgiven a little
schadenfreude today after the latest batch of manufacturing
surveys showed the sector barely grew in April.
The ONS came in for blistering criticism last week after its
preliminary estimates pointed to an economy-wide contraction
during the first three months of 2012.
Private sector economists and officials from the Bank of
England pointed to more upbeat evidence from business surveys to
suggest the ONS was being too gloomy and risked damaging
business and consumer confidence.
Commentators "knew" or at least claimed to suspect the
preliminary estimates for gross domestic product (GDP) were
"wrong" and would be revised upwards when more complete data
became available. But that is now looking less certain.
The Markit/CIPS Manufacturing Purchasing Managers' Index
(PMI) dropped to 50.5 in April, which is basically
indistinguishable from the 50-point threshold which divides
expanding activity from a contraction, suggesting manufacturers
flat-lined last month.
The PMI for March was also revised down, from 52.1 to 51.9,
suggesting the manufacturing sector was growing more slowly at
the end of the first quarter than originally thought.
Looking ahead, the new orders index slipped to 49.2, from a
downwardly revised 52.4 the previous month, pointing to a
possible further slowing of activity in the months ahead.
The survey numbers and the official GDP estimates are
starting to converge. But rather than GDP numbers being revised
higher, the survey data is starting to look softer.
Following today's data "it will be more difficult to say
that the survey data are more upbeat than the official numbers,"
according Investec economist Philip Shaw.
"It makes it harder for the [Bank of England] to continue
justifying looking through the official data to look at the
surveys, because they are starting to tick down as well," RBS
analyst Gareth Anderson told Reuters.
The ONS numbers could still be revised (in either
direction). Assuming the ONS data is subject to (unbiased)
sampling errors, the downward revisions to the PMI for March and
further loss of momentum in April should have no effect on
whether Q1 GDP is revised higher or lower.
But the downturn in the PMI offers cautionary lessons for
everyone charged with analysing and acting upon Britain's
First, it is a reminder not to over-interpret small changes
in the numbers such as whether GDP shrank by 0.2 percent or was
flat, and whether the manufacturing index is 52.0 or 50.5.
Second, the official GDP data is not invariably less
accurate than private sector surveys.
There has been a tendency to assume "government" numbers
produced by statisticians in safe public sector jobs far removed
from the pressures of the real world must somehow be inferior to
measures produced by private sector organisations in daily
contact with ordinary firms, but that is pure prejudice.
Third, some commentators have seemed to suggest that gloomy
public sector data should be ignored in favour of the more
optimistic private sector surveys because this will support
fragile business and consumer confidence.
The optimistic Pollyannas want to create a virtuous circle
of rising confidence and improving statistics. But even assuming
this works, and there is some evidence for positive feedback
between confidence and growth (what John Maynard Keynes termed
"animal spirits"), a "faith-based" approach to policymaking is
Policymakers and analysts should not cherry pick economic
data according to whether it fits their existing theories on
whether the economy is growing or not, whether austerity is
working or not, and whether monetary policy is too loose or
needs to provide further stimulus.
IGNORE THE DATA AT YOUR PERIL
There has been plenty of evidence for this sort of cherry
picking in both Britain and the United States in recent years.
In the United States, Fed officials were unable to see the
housing bubble in 2002-2006, only localised froth, in the run up
to the crisis, and are still unable to admit that overly lax
monetary policy in 2003-2004 might have contributed to the run
up in housing prices, because that interpretation suits the
central bank's low interest rate policy.
In Britain, the Bank of England has persistently ascribed
the overshoot in inflation to "temporary factors" while worrying
that even a brief slowdown in growth presages a prolonged slump,
largely because it fits with the strategy senior policymakers
have been pursuing since 2009. As inflation repeatedly fails to
fall in line with official predictions, however, this
faith-based approach to forecasting is being tested.
The problem with a faith-based statistics and policymaking
is that mistakes go uncorrected too long, leaving problems to
accumulate until they become impossible to ignore, and wrenching
There is always a strong temptation to "look through"
official numbers and dismiss them as wrong, incomplete or
lagging, but the risk is that policymakers see what they want to
see, rather than what is really there.
So Britain's statisticians should stick to their careful
estimates, and policymakers and commentators should try to
resist the temptation to criticise or explain away every piece
of unfavourable news.