Forecasting Mumbo Jumbo

Forecasting Mumbo Jumbo

People get paid a lot of money to do things which, in the cold light of day, they’d probably be quite ashamed to admit.  I mean, of course, forecasting economic data.

Even worse, other people move vast sums of money around in reaction to whether or not an actual economic data release is in or out of the ballpark of those ‘expert’ forecasts.

At best, this is just a waste of time (apart from giving some fund managers and economist-types something to do).  At worst, it loses other people’s money on higher transactions costs and inferior investment performance.

We don’t forecast macro data for (at least) two reasons.  First, we cannot hope to be any better at it than anyone else who has access to the same information – and if we can’t be better at it we’ve got no business charging people for the privilege.  And second, even supposing that we were accomplished seers, consistently being able to translate a macro data outturn into a change in asset prices is impossible.  Stuff happens between here and there, be it natural disasters, sovereigns’ lack of credibility, oil price shocks, or any number of factors about which we know nothing, to put the best laid forecasts awry versus actual price action.

The most closely-watched macro indicator of the largest economy on the planet is the US employment report; specifically, the monthly change in the non-farm payroll.  On the first Friday of the month, the air on dealing desks and investment floors worldwide is heavy in anticipation of the announcement.

Whether the number, which flashes on the screen at 1.30pm London time, is above or below the consensus expectation of the market will determine in large part how people feel in the short term about the direction of the US, and global, economy.

How people feel about short-term stuff can have a significant impact on price behaviour, provoking episodic shifts in asset markets.  It shouldn’t, but it can.

As anti-forecasters we do our best to take a step back from noisy data releases and observe facts about where the economy actually is, rather than where it is going (which falls firmly into the category of stuff about which we know little).  With that in mind, we’re more interested in how the medium-term employment picture is evolving and whether the recent data suggest there’s anything to worry about.

It turns out that the employment market is fine and the recent data have been okay.

The first chart below sums this up, along with a look at how household income and spending is faring compared with the path followed by all three over the past 15 years:

Forecasting mumbo jumbo Blog chart 1 May 12

So when we see weaker than consensus payroll releases prompting headlines to scream: “Markets plunge again on disappointing US data” as we did last Friday, it suggests something of an over-reaction.  This is brought into even sharper focus when we observe the pattern of the past year’s payroll data releases and subsequent revisions to that data.

The charts below show this in two ways.  First, the forecast, actual data release and subsequent revisions to the monthly change in employment.  Then, the difference in the overall employment path as a result of the data revisions, contrasted with the path suggested by the consensus forecast for the monthly change.

Forecasting mumbo jumbo Blog chart 2 May 12Forecasting mumbo jumbo Blog chart 3 May 12


Clearly, the past twelve months have seen a somewhat better employment backdrop than the market had been anticipating.  To change one’s view of the strength or otherwise of the labour market on the basis of one month’s (revision-prone) data is nonsense – and to change investment positioning as a consequence of this is even worse.

The value of investments will fluctuate, which will cause prices to fall as well as rise and you may not get back the original amount you invested. Past performance is not a guide to future performance.