Followers of the OECD’s leading indicators take note: the organisation has recently altered the smoothing method used in constructing the indices, increasing the risk of false signals or “whipsaws”.

The OECD’s indicators are used here to confirm signals from monetary trends. The indicators lead the economy at momentum turning points by 4-5 months on average versus 9 months for real money. (The indicators are released later than monetary data and are subject to larger revisions so these numbers understate the real time advantage of relying on monetary analysis for forecasting.)

The calculation method for the indicators historically incorporated smoothing of component data to remove short-term noise. This incurred a cost in terms of slightly reducing the lead time at turning points but had the greater benefit of increasing confidence in the validity of signals, i.e. identified turning points were usually followed by sustained directional moves.

This is illustrated by the first chart, showing “unofficial” versions (i.e. calculated here) of the US indicator with and without smoothing. The unsmoothed indicator, although slightly more timely at “true” turning points, gives many more false signals.

What happened this year? The OECD suspended release of its indicators in March because of uncertainty about the data impact of lockdown restrictions. A decision then appears to have been taken, unannounced, to resume publication on the basis of unsmoothed data.

This can be seen by comparing the second chart, showing the “official” US leading indicator series, with the smoothed and unsmoothed indicators in the first chart. The smoothed historical data appears to have been chained to unsmoothed data in 2020.

What motivated this change? US / global GDP bottomed in April. The smoothed US indicator bottomed in June. OECD statisticians would have been aware that the smoothed indicators were likely to lag on this occasion so may have decided to emphasise the unsmoothed indicators, which in most countries reached lows coincident with activity in April.

A switch back to smoothed data seems likely at some point but will take some explaining given the recent lack of transparency.

From a practical perspective, the key takeaways are:

1. The June turning point in the smoothed US indicator supports the “monetarist” view that the economy has entered a sustained expansion phase.

2. Any declines in the “official” indicators, assuming that they continue to be based on unsmoothed data, should be ignored unless sustained for several months.

There is no contradiction between unprecedented GDP falls and optimism about equity market earnings.

Global GDP and industrial output have been tightly correlated historically but with industrial output displaying three times the volatility of GDP – compare the left- and right-hand scales in the first chart.

This relationship has broken down in 2020 because covid health restrictions have undermined the usual relative resilience of services activity, which dominates GDP. The "beta" of industrial output to GDP has been much smaller than in previous recessions.

This is important for investors because equity market earnings have a large industrial element and hence a stronger correlation with industrial output than GDP.

The second chart shows the relationship of S&P 500 12-month forward earnings and global industrial output. Forward earnings overshot the historical relationship in 2018 because of corporate tax cuts. They fell by less than suggested by output during the covid shock but 12-month rates of change bottomed around the same time (April for output, May for earnings) and are currently consistent.

A further reason for the relative resilience of industrial output is that services activity restrictions have caused consumers to divert spending towards goods. Global retail sales appear to have risen further above their pre-covid peak in July, based on partial data – third chart.

The unexpected strength of goods demand coupled with production restrictions have resulted in an involuntary drawdown of inventories, which firms will attempt to reverse during H2 – an additional reason for expecting a V-shaped industrial recovery to continue to unfold.

Such a scenario implies a bullish earnings outlook – claims that forward earnings are already overoptimistic appear wide of the mark.

Positive earnings developments, however, could be neutralised or outweighed by a rise in real discount rates as the “excess” money backdrop becomes less favourable. The gap between six-month rates of change of global real narrow money and industrial output remained wide in July but could close by October – see previous post.