Friday, July 20, 2018

Big data, fake news and global growth

by Michael Roberts

This week, it was reported that the number of Americans filing for unemployment benefits last week was the lowest since 1969!

The official unemployment rate is also near an all-time low.  In Japan and the UK too, the unemployment rates are near lows and in Europe, the official rate is heading back to pre-global crash levels.

As I have reported in previous posts, measures of economic activity from various sources suggest that the world capitalist economy has been picking up pace in growth since the near slump of 2015-16.  This is particularly the case for the most important capitalist economy, the US.

Below is the world composite PMI (purchasing managers index).  This is a survey globally of the state of economic activity in both manufacturing and service industries as the corporate executives see it.  If the measure is above 50, world economic activity is rising.  Currently, the PMI shows a return to trend expansion after the near contraction of 2015.

Next week we shall get the first estimate of US real GDP growth for the second quarter of 2018.  It is likely to be strong.  The Atlanta Federal Reserve has a ‘high frequency’ forecast measure for each quarter’s growth and it currently expects the Q2 figure to come in at a 4.5% annual rate.  That means real growth in Q2 would be about 1% point above Q1.  If that turns out to be right, it means that the US economy would have motored along at about 3% for the first half of 2018.

No doubt President Trump will make much of this apparent fast expansion and claim it for his policies of tax cuts for the corporate sector and the top 10%.  However, as a recent study has shown, this will be ‘fake news’.  The study by some European economists found that there was no difference between the post-election performance of the US economy under Trump and a synthetic ‘doppelganger’ US economy without Trump, suggesting that there has been no ‘Trump effect’. “The employment performance of the US economy since the election was no different from its doppelganger. There is nothing in the data that indicates an acceleration of employment creation because of President Trump.”

But the most usable surveys of economic activity in the US do show that the economy is expanding at a reasonably fast rate (if no faster than the average of 3.3% since 1945).  Here is a graph that combines various surveys of economic activity in the US.  Anything above 0 (LHS) or 50 (RHS) implies that the economy is growing.  The current RHS rate is close to 60 which implies fast expansion – certainly compared to 2016 when the measure was below 50, implying contraction and, of course, much higher compared to the Great Recession when output collapsed.

The other major capitalist economies do not seem to be doing as well as the US, despite previously optimistic reports.  The EU is growing at about 1.6% annually, the UK at under 1%, and Japan is actually contracting.  Nevertheless, global growth is expected to show an acceleration in 2018 over 2017, when all the emerging economies of China, India etc are included.

But can we get more frequent and comprehensive measures that could actually forecast accurately what will happen in future quarters and years?

The huge eruption of what is called ‘big data’ from the internet, social media and other sources in the last ten years has led to a new industry of forecasting that aims to deliver more frequent and accurate estimates of future developments, in the same way that weather forecasting has improved.

The Federal Reserve Bank of New York has refined this big data in its own survey of US economic activity.  And as long ago as 2013, the Bank of England’s economists looked at the use and efficacy of big data.  They looked at indicators for global growth in industrial production and trade.  They found that sharp changes in various indicators were a good guide to future production and growth.  However, the problem with these indicators of future expansion is that they are not that timely, with data only on a monthly (if you are lucky) but more usually on a quarterly basis.

The statistical economists have recently looked for more timely data and a Bank of England economist recently published a paper on the best predictors of global growth. The paper found that the daily movement in metals prices was a reasonably accurate measure of global economic activity.  “Metals prices are highly correlated with world activity… and perform well at predicting world GDP in the near-term.”

In other words, the pace of change in metals prices in this month of July will give a reasonable estimate of world real GDP growth for July (and eventually Q3), well ahead of any official data (Q3 world growth is not going to be available until January 2019).

The Bank of England economists used the S&P metals prices index as their metals prices indicator.

As you can see (circles), the metals index fell sharply during the Great Recession in real GDP and predicted the subsequent recovery exactly in mid-2009.  Similarly it predicted the recovery from the relative slump in 2015.  Remember the actual real GDP figures for most countries do not become available until up to two quarters later or even more.  So the metals index becomes a ‘high frequency’ indicator for growth.

Copper is the largest constituent of the index and it is a metal used in just about every important industrial and consumer appliance or service.  So the copper prices index is also likely to be a good indicator, in my view.  When I ran the copper price against world GDP growth, the correlation was very good.

So looking ahead, what do the metals price and copper price indexes tell us about the current Q3 period and onwards?  I did the trend measure of the copper price, and it shows that expansion from the trough of 2015-16 seems to have peaked.  That suggests the global expansion from 2017 which was above the trend rate has now subsided back to the trend and may fall below.

The metals price index also suggests that the peak in the current acceleration of global (and US?) growth ended in June 2018 and the direction is now downwards in Q3 (the period beginning in July).

So don’t be overwhelmed by the good news stories about US real GDP figures for Q2 2018 next week.

No comments: