My guess is
probably, but what do I know? What I do know is that unconditional
[1] macroeconomic forecasting is a mug’s game, and the only reason
some people do these forecasts is that they are generally better than
an informed guess, but only a little better. What I can do in this
post is make some I hope helpful points about annual versus quarterly
growth, look at some of the evidence and some key behaviour that will
decide whether a UK recession is on the cards.
As Duncan Weldon
reminds
us, most established economic forecasters are terrible
at forecasting recessions. One reason has little to do with
economics, and a lot to do with human nature. I learnt this very
early. My first job was helping to forecast the world economy in the
Treasury, and it was after the first oil shock of 1973/4. Our initial
forecast showed a collapse in world trade. Our boss was not happy –
nothing like that had happened since WWII. As a result of his
unhappiness we revised our forecast up, but our initial forecast was
nearer what actually happened than our revised forecast. Established forecasters
are always looking over their shoulder at other previous forecasts
(by themselves and others) and hate being too extreme. As a result,
they tend to miss booms and recessions.
A second reason that
domestic forecasters miss recessions is that they fail to recognise
that what they are seeing domestically is often also happening in the
rest of the world. That is true today with a global cost of living
crunch. Indeed that psychology of forecasting can allow
non-forecasters or ex-forecasters like myself a slight advantage. In
this
post I tell the story of the 2009 recession, where in
No.11 Downing Street I at least held my own among more professional
forecasters simply because I applied these two observations about
forecasting the extent of that recession.
It should not come
as a surprise, therefore, that no major forecaster has predicted
negative annual growth this year or next, despite what is expected
to be the biggest fall in living standards in any single financial
year since ONS records began in 1956-57. Instead consumers are
expected to dramatically reduce their savings, as this chart from the
latest OBR forecast shows (look at blue line).
So, sticking with
this OBR forecast, we have real household disposable income falling
by 1.5% this year and 0.2% next year, but aggregate consumption is
forecast to increase by 5.4% and 1.0% respectively. At first sight
this looks very implausible.
It looks even more implausible if we look at surveys of consumer
confidence. To quote from Duncan’s piece: “The GfK Consumer
Confidence Index fell for the fourth month in a row to -31 from -26
in February, its lowest since November 2020, deep in the coronavirus
pandemic. Readings of -30 and below have presaged recession on four
out of five occasions since the survey started in 1974.” Since then
the March data is available, and
it’s at -38!
David Blanchflower talks about this data and similar for the US here,
and is in little doubt that a recession is on the cards. So how would
economic forecasters, and the OBR in particular, defend their
forecast of strong growth in consumption this year, and positive
growth next year, despite falling incomes? The answer also comes from
the chart above. The pandemic led to unprecedented increases in
household savings, because most maintained their incomes but the
pandemic led to sharp falls in ‘social consumption’. So most
consumers will have plenty of scope to run down their savings as
their incomes fall.
Furthermore, standard theory suggests that consumers who have the
ability to do so will try to smooth out fluctuations in real income,
if they think the fall in their income is temporary. Indeed, after
social consumption has been suppressed during the pandemic, there may
be some bounce back as consumers try to partially recoup the spending
they had missed out on. (For a similar reason, consumers switched spending from services to goods during the pandemic, which partly explains some
of the supply side inflation we have seen). To set against that the
pandemic is not over, despite what some politicians might say, so
that will inhibit consumption.
Rapid consumption growth is what we saw at the end of last year during the vaccine led recovery from the pandemic. The
level of consumption in the fourth quarter of 2021 was over 8% higher
than a year earlier. Crucially, that means that even if quarterly
consumption in 2022 was flat at the 2021Q4 level, annual growth this
year would be very high. The lesson here is that for this year, look
at quarterly growth through the year rather than year on year
numbers.
Does the recent
fall in retail sales also suggest a recession? Again
we have to be careful. As many people are starting to behave as if
the pandemic is over, we would expect to see a switch from goods you
buy in shops or online to social consumption which are services like
travel or eating out. As James
Smith notes, online sales are also falling back to
more normal levels. This doesn’t necessarily imply a fall in total
consumption.
So where does that leave us? While many consumers are in a position
to use savings to finance consumption growth, they will only do so if
they are sure the cost of living crunch is temporary rather than
permanent. Many will not be so sure, and together with those who can
only maintain consumption through borrowing, it seems likely that the
aggregate level of consumption will fall through this year.
That in turn means it’s likely that we will see falls in the
monthly path of GDP through this year, and indeed that is something
the OBR
are expecting to happen (p.42). [2] In that sense the
OBR is forecasting a recession during this year, but not in the
annual figures that everyone focuses on.
For reasons already explained, that quarterly path could still leave
a relatively healthy year on year growth rate for this year because
of strong growth due to the vaccine based recovery through 2021. The
big unknown is what happens in 2023. Looking at the OBR’s forecast
savings ratio chart above, what looks implausible is the very slow
recovery in savings from 2023 onwards. If annual growth is going to
be negative at any point, it is likely to happen next year rather
than this, because the inflation we are currently seeing keeps
incomes low and consumers try and get back to more normal levels of
saving.
[1] By unconditional, I mean forecasts of what will happen to a
macroeconomic variable in a year or two’s time. In contrast a
conditional forecast asks how that variable will change if policy
changes, for example. Conditional forecasts are much more focused,
and therefore more reliable. Politicians and some journalists often
do not, or pretend to not, know the difference between these two
types of forecast. For example Brexiters during the 2016 referendum
used the unreliability of unconditional forecasts to cast doubt on
conditional forecasts like Brexit will lower GDP, which was a simple
error.
[2] That in turn makes a technical recession in the UK (two
consecutive falls in quarter on quarter GDP) possible, but it’s
wrong to get hung up on this technical definition. A quarterly path
of GDP growth that goes +2.0, -0.1, -0.1, +2.4, +2.4 is a technical
recession, while a path that goes +0.2, -1.0, +0.1, -2.0, +0.1 is
not, but that latter is much worse than the former. I tend to use the
term recession in a much less precise way, to mean an economic
downturn that is particularly severe.
