Entries from May 22, 2016 - May 28, 2016
Earnings revisions consistent with economic upturn
Global economic growth prospects are judged here to be the most positive since 2012, based on recent strong narrow money trends – see previous post. The favourable monetary signal has received tentative confirmation from an upturn in a non-monetary leading indicator derived from the OECD’s country leading indicator indices – see first chart below and this post.
Another development suggestive of an upturn in economic momentum is a recent rise in the net percentage of equity analysts who are upgrading their forecasts for the earnings of the companies they cover. Such revisions are usually driven by news in current earnings and forward guidance from companies, both of which are sensitive to changes in economic growth. The revisions ratio (i.e. upgrades minus downgrades divided by the total number of earnings estimates, seasonally adjusted) for constituents of the MSCI All-Country World index has recovered significantly since early 2016 – second chart. (Note that the “normal” level of the ratio is negative, reflecting a tendency for initial analyst estimates to be too high, except immediately after recessions.)
Most market economists base their short-term growth views on business surveys, although these are sometimes coincident rather than leading indicators. Globally, such surveys remain mixed. It is possible that companies’ responses to the surveys are being influenced by widespread gloom-mongering and US / UK political uncertainty. The media often fail to report more positive survey evidence, such as last week’s CBI industrial trends poll for May, showing an above-average net percentage of firms planning to increase output.
The upshot is that the consensus is ill-prepared for the positive economic growth surprise suggested by monetary trends. Such a surprise in 2012-13 gave a significant boost to equity markets while pushing up government bond yields.
Global money trends: positive signal affirmed
A post last week noted that global six-month real narrow money growth had risen to its highest level since 2009, suggesting economic strength in late 2016 / early 2017. A reader expressed scepticism about this forecast on the grounds that zero / negative interest rates are putting downward pressure on the velocity of circulation of narrow money – strong money growth, therefore, may reflect liquidity preference and have no implication for future activity.
It is true that the low level of interest rates has reduced the opportunity cost of holding narrow money, and the average level of real narrow money growth has been higher relative to industrial output expansion in recent years than during the post-WW2 period as a whole. The monetarist forecasting relationship, however, is directional – it relies on changes in real money growth leading changes in economic growth. This relationship remains strong.
The chart below shows global six-month industrial output growth and real narrow money expansion adjusted for 1) a long-run downward trend in the rate of change of velocity and 2) the average nine-month lag between turning points in real money growth and output expansion. This adjusted real money growth measure exhibits a significant contemporaneous correlation with output expansion*.
The economy was expected to lose momentum in early 2016, reflecting a slowdown in real narrow money expansion between February and August 2015. Recent economic weakness, therefore, does not constitute evidence that the predictive power of narrow money is waning. The adjusted real money growth measure suggests that the six-month change in industrial output will recover strongly between May 2016 and January 2017 (the last available month given the applied nine-month lead).
Zero / negative interest rates are helping to sustain a post-GFC rising trend in the ratio of narrow to broad money. Significant short-term changes in real narrow money growth, however, are still likely mainly to reflect changes in spending intentions, with implications for future economic activity. Economists and policy-makers historically have found various reasons to dismiss the significance of monetary trends; they have usually been wrong to do so.
*Correlation coefficient = 0.62 over 1965-2015 (51 years).
UK Treasury Brexit analysis: circularity and false precision
The Treasury today published its analysis of the short-term economic impact of a “Brexit” vote. The report states that the economy would fall into a recession immediately following such a vote on 23 June, with no recovery until the third quarter of 2017 at the earliest. After two years (i.e. by the second quarter of 2018), GDP would be between 3.6% and 6.0% lower than in the alternative scenario of a vote to remain.
The analysis assumes that a Brexit vote would have large negative effects on long-run income expectations, uncertainty and financial market conditions. It then attempts to quantify the impact of these assumptions on key macroeconomic variables.
The analysis is internally consistent but it is important to realise that the forecast of a significant recession flows directly from the Treasury’s assumptions.
A key role is played by a composite measure of uncertainty based on financial market prices and consumer / business surveys. The Treasury shows that this measure has been negatively related to household consumption, business investment and financial market conditions historically. It would be expected to rise following a Brexit vote, but by how much?
The Treasury assumes an increase equivalent to between 1.0 and 1.5 times the measure’s historical standard deviation. This implies a level of uncertainty reached previously only during recessions and the Eurozone crisis – see chart. The Treasury’s modelling work traces out the direct and indirect (via financial market conditions) implications for GDP and other macroeconomic variables. The conclusion of a recession, however, is circular – it depends on the assumption that uncertainty rises to a level rarely reached outside recessions.
The extent of a post-Brexit-vote rise in uncertainty is unknowable. The value of detailed forecasting exercises that assume away this lack of knowledge is questionable.