A simple forecasting rule based on money growth and share prices suggests that 2019 will be a “bad” year for the UK economy – even if a no deal Brexit is avoided.

The forecasting rule has been informative historically, often outperforming economist predictions. It was pessimistic in late 2008 ahead of a 4.2% GDP slump in 2009. Another negative signal was given in late 2011 before the 2012 double-dip scare. It was optimistic a year later, foreshadowing a pick-up in growth in 2013. The rule has given neutral signals for the last five years.

The forecasting rule classifies the economic outlook for the following calendar year as good, bad or neutral depending on whether December levels of real (i.e. inflation-adjusted) money growth and share prices are higher or lower than 12 months earlier. Real money growth is measured by the annual rate of change of the broad M4ex measure deflated by the retail prices index excluding mortgage interest (RPIX). Share prices are measured by the FTSE local UK index, constituents of which derive at least 70% of sales from the UK and the rest of Europe, again deflated by the December RPIX*.

Annual GDP growth averaged 2.3% in the 52 calendar years from 1966 to 2017. The forecasting rule gave a positive signal for 16 of these years, i.e. both real money growth and real share prices at the end of the prior year were higher than 12 months before. GDP growth in these years averaged 3.8%.

There were 11 years for which the forecasting rule gave a negative signal, reflecting falls in both real money growth and real share prices. Growth in these years averaged 0.3%. In the remaining 25 years for which the two components of the rule gave conflicting signals, GDP expansion averaged 2.2%, close to the overall average.

The rule gave a neutral message for 2018 at this time last year: real money growth had risen since December 2016 but real share prices had fallen. The current consensus estimate of 2018 GDP growth is 1.3%, below the 2.2% average for neutral years**. The undershoot could reflect a Brexit effect, or a fall in the economy’s potential growth rate.

The current negative signal for 2019 will not be confirmed until 30 January, when December monetary data will be released. As of October, the annual rate of change of real M4ex was -0.6%, down from 0.7% in December 2017 – see chart. The average level of the FTSE local UK index so far in December, meanwhile, is 16% lower than in December 2017. With annual RPIX inflation at 3.1% in November, real share prices are on course for a 19% decline.

The current consensus forecast is for GDP growth to rise to 1.5% in 2019, despite recent slower global expansion and the increase in Bank rate from 0.25% to 0.75% since October 2017. The consensus view appears partly to reflect expectations of support from fiscal easing and a “Brexit dividend” as delayed spending plans are reinstated following an assumed smooth departure.

The forecasting rule, by contrast, suggests that the economy will slow further in 2019, with a significant risk that GDP growth will undershoot all economist predictions***.

*The FTSE local UK index goes back to 1994. The FT 30 index, which is also domestically orientated, was used for earlier years. Local UK stocks accounted for 20% of the market capitalisation of the FTSE all-share index as of end-November.
**Source: Forecasts for the UK Economy, HM Treasury.
***The current minimum forecast is 0.9%, by Peter Warburton of Economic Perspectives, according to the Treasury survey.

A post in February suggested that the Euroland manufacturing PMI, then still strong at 58-59, would fall towards the 50 level by mid-2018. The decline has taken longer than expected but the PMI eased further to 51.4 in December (with assistance from the gilets jaunes). The composite manufacturing / services output index slid to 51.3, a four-year low. What now?

The basis for the February forecast was a significant fall in six-month real narrow money growth from June 2017 and an expectation that this would be reflected in PMI results with a lag of about six months. The chart, an update from the February post, illustrates this relationship, which now suggests that the PMI will stabilise around the 50 level through April 2019, based on monetary data through October 2018*.

Current money trends are judged here to be consistent with GDP growing by about 0.3% per quarter in the first half of 2019 – see previous post. Surprisingly, the new ECB staff forecast released yesterday maintained the previous central projection of 0.5% per quarter despite recent negative news. The suspicion is that a downgrade was postponed to avoid calling into question the decision to end QE. (Even more remarkably, the forecast commentary made no mention of the risk of a disruptive / disorderly Brexit.)

It should be emphasised that money trends – broad as well as narrow – are not signalling economic stagnation or contraction. The six-month change in real narrow money turned negative ahead of the 2008-09 and 2011-12 recessions. It stood at 2.4% in October.

The manufacturing PMI would need to fall to around 45 to suggest a GDP recession.

It has been claimed that the cessation of QE will lead to a “mechanical” further slowdown in money growth. Cross-country evidence, however, shows that the relationship between QE and money trends has been weak. Euroland money trends were improving well before QE started in 2015, probably reflecting rate cuts and expanded ECB lending to banks. Proponents of a “mechanical” relationship wrongly suggested that Japanese broad money growth would soar following the expansion of QE in 2013 – it barely budged.

The view of economic prospects here remains on the pessimistic side of the consensus but money trends do not currently warrant a further downgrade. Relative money trends suggest that downside risks are now greater in the US and China.

*The chart uses the EU Commission manufacturing survey as a proxy for the PMI before 2016 because of lack of access to historical data.

Chinese November money and credit numbers were disappointing for economic optimists, suggesting that policy easing is still failing to gain traction.

Six-month growth of true M1, the preferred narrow money measure here, was little changed in November and close to the recent low reached in August – see first chart*.

Six-month growth of total social financing, on the new definition including local government special bonds, similarly showed no improvement. The recovery from a low in June solely reflects heavy issuance of such bonds over July-September. On the former definition, six-month expansion in November almost matched the recent low.

Year-on-year growth of all three measures reached new lows in November – second chart.

As expected, real money and credit growth in November was supported by a fall in consumer price inflation, reflecting declines in fuel and food prices. The six-month change in real true M1, however, remained close to zero – third chart.

Sectoral figures show that year-on-year growth of bank deposits of non-financial enterprises fell further in November and is now lower than before the 2015-16 economic slowdown / hard landing scare. Weakness is consistent with recent profits deceleration and signals deteriorating prospects for business investment – fourth chart.

The year-on-year fall in interest rates suggests a coming recovery in money trends – fifth chart – but this may be delayed and / or weakened by the ongoing clampdown on shadow banking activity. The authorities may need to offer banks additional incentives to lend to private sector firms to offset the contraction of shadow activity. Current money data signal a further loss of economic momentum in early 2019, with any reacceleration unlikely until the second half at the earliest.

*True M1 = currency in circulation + demand deposits of enterprises, government organisations and households. The official M1 measure excludes household deposits. The November data point in the chart is an estimate and assumes a 1% monthly rise in household demand deposits.

US narrow money trends remain worryingly weak, suggesting that monetary policy is already restrictive and a significant economic slowdown lies immediately ahead.

The Fed last week released sector financial accounts for the third quarter, permitting the calculation of an end-September number for non-financial M1, the preferred narrow money measure here*. The two-quarter change in real (i.e. deflated by consumer prices) non-financial M1 moved into negative territory, reaching its lowest level since 2008. The leading relationship with two-quarter GDP momentum suggests that the latter will fall sharply from a third-quarter peak – see first chart.

Narrow money trends appear to have remained weak so far in the fourth quarter, judging from the headline M1 measure – the chart incorporates a November estimate of the six-month change in real M1 based on weekly data through 26 November.

The weakness of real non-financial M1 was initially focused on the business sector but household holdings also contracted over the latest two quarters – second chart. Previous posts suggested that the fall in business holdings would be reflected in a slowdown in investment and hiring – recent data (e.g. capital goods orders, aggregate hours worked) lend support to this forecast. With household holdings also now weakening, consumer spending could be next to disappoint.

The fall in business money holdings, which extends to the broad M3 measure, has occurred despite foreign earnings repatriation, which might have been expected to swell domestically-held corporate deposits. The financial accounts show that the repatriation flow has been offset by a rise in net equity retirement (i.e. buy-backs and cash takeovers), a slowdown in debt issuance and faster accumulation of “miscellaneous assets”. The flow peaked in the first quarter and fell further in the third quarter, with a corresponding slowdown in net equity retirement – third chart.

The negative signal for economic prospects from money trends is receiving confirmation from the OECD’s US composite leading indicator. As previewed in a post last week, this indicator continued to decline in October, signalling below-trend GDP growth – fourth chart.

*Non-financial M1 comprises holdings of currency and checkable deposits by households and non-financial businesses. Headline M1 also includes holdings of insurance companies, pension / retirement funds, money market funds, government-sponsored enterprises, finance companies, REITs, brokers / dealers and foreign non-banks (excluding official institutions). These holdings account for more than 30% of M1 (dominated by the foreign component, mainly currency) but are unlikely to be of relevance for assessing near-term prospects for spending on goods and services.

Monetary trends and cycle analysis have been suggesting a cautious view of global economic and equity market prospects. An argument for remaining neutrally-positioned or overweight in equities and other risk assets at present runs as follows.

Major equity market declines historically were associated with US / global recessions. The US Treasury curve (10s-2s) inverted well before every recession since the mid 1950s but has yet to do so. The term premium, moreover, is unusually low, meaning that the flatness of the curve exaggerates monetary tightness. The curve, in other words, needs to become significantly negatively sloped to indicate a recession. Once this occurs, investors will have time to derisk portfolios before markets weaken.

This argument is dangerous for several reasons.

First, equity markets usually decline before / during economic slowdowns as well as recessions. Slowdown-related falls can often be severe, e.g. the 1966-67 US economic slowdown was associated with a 22% peak-to-trough decline in the S&P 500. Investors should dial down risk ahead of slowdowns as well as recessions, to the extent that either can be anticipated.

Secondly, the focus on the US yield curve may be misguided. The current global slowdown originated in China / Asia and has spilled over to Europe. The Chinese yield curve inverted in mid-2017 – see first chart.  It is possible that this will prove to have been the recession warning signal in this cycle, with the US curve lagging.

The last Asian-driven global economic slowdown was in 1997-98. It did not develop into a recession partly because of US / G7 monetary policy easing and partly because the region was a smaller part of the global economy. Asia ex Japan now accounts for 37% of world GDP at purchasing power parity, up from 19% in 1997, according to the IMF.

Thirdly, monetary trends argue against the view that the yield curve signal has been distorted by a low term premium. The slope of the G7 curve has been closely correlated historically with the six-month rate of change of G7 real narrow money – second chart. The two measures are giving an identical message at present – both suggest restrictive monetary conditions. If the term premium had introduced a distortion, one would expect real money growth to be diverging positively from the curve (assuming, reasonably, that the premium has little impact on money trends).

Fourthly, the argument that investors have time to adjust their portfolios after a yield curve inversion may be correct in principle but the relevant curve now may be China’s. The MSCI All Country World Index peaked in January 2018, eight months after the Chinese inversion.

The normalisation of the Chinese curve since mid-2017 suggests that China / Asia will lead a global economic recovery but Chinese money trends currently remain weak, arguing that any revival is unlikely until the second half of 2019 at the earliest.

The baseline view here is that the current global economic downswing will extend into the first half of 2019, and probably beyond mid-year. Incoming news is consistent with this scenario.

Near-complete monetary data for October confirm that G7 plus emerging E7 real narrow money growth fell to a new low  – see first chart. This suggests that six-month industrial output momentum is unlikely to reach a trough until around July 2019 at the earliest, based on the historical average nine-month lead time at turning points.

As an aside, G7 plus E7 real narrow money growth may have crossed back below industrial output growth in October, with partial data indicating a rebound in the latter – an October estimate is included in the chart. If confirmed, the first of the two equities / cash switching rules described in a previous post will move from equities to cash at end-December*.

Non-monetary leading indicators are signalling a near-term further economic slowdown. The OECD will release October data for its composite leading indicators on 10 December but most of the component information is already available, allowing an independent calculation. The G7 indicator is estimated to have fallen further in October, reflecting declines in all countries bar Japan (where weather-related disruption to activity and a subsequent recovery may have distorted recent readings) – second chart**. The indicators are designed to predict the direction of detrended GDP, i.e. a decline signals below-trend growth.

Previous posts argued that a downswing in the global stockbuilding cycle would act as a drag on growth in late 2018 and 2019. The cycle has averaged about 3.5 years, measured from trough to trough, and the last low is judged to have occurred in early 2016, implying that the next one could be reached in the second half of 2019. Global manufacturing PMI survey results for November support the view that the cycle is at or past a peak, with the finished goods stocks index close to the series high and diverging from weak new orders – third chart.

Stockbuilding cycle downswings in 2011-12 and 2015-16 were associated with significant global economic slowdowns. The current downswing may coincide next year with a downswing in the longer-term business investment cycle, which last troughed in 2009 and is scheduled to reach another low by 2020 at the latest. The risk of the current economic slowdown developing into a recession, in other words, is greater than in 2012 and 2016.

Global narrow money trends would need to weaken further for a recession to be adopted as the central case here. The six-month change in G7 plus E7 real narrow money turned negative ahead of the 2008-09 recession – first chart. It fell to 0.4% (not annualised) ahead of the milder 2001 recession. The October reading was 1.0%. A reasonable judgement is that a recession call would be warranted by a further fall to 0.5%.

As previously discussed, the recent collapse in the oil price will provide near-term support to real money growth by cutting headline inflation. The six-month change in G7 plus E7 consumer prices could fall by about 0.5 of a percentage point if commodity prices, as measured by the energy-heavy GSCI, are stable at current levels – fourth chart.  Six-month nominal narrow money expansion, therefore, might need to fall by a further 1 percentage point to push real growth down to 0.5%.

*The second rule, based on whether G7 annual real narrow money is above or below 3%, remains in cash currently. Its recommendation at end-December will depend on November monetary / inflation data to be released over the course of this month. Both rules were in cash during October.
**Of the G7 country indicators, only Canada’s includes a monetary aggregate.