Trend growth

Moody's decision to cut the government's credit rating because of "continuing weakness in the UK's medium-term growth outlook" will, no doubt, focus attention upon the need to raise our trend growth rate.

Doing so, however, might be very tricky.

To see my point, start from the fact that the economy shrank slightly last year. Does this mean our economy is in trend decline? Probably not. More likely, 2012 was just an unusually bad year. But even if we take many years together, it's quite possible that we'll over-sample bad years (or good) and so get a biased measure of trend growth. To reflect this problem, we should measure the standard error of growth. In its simplest form, we do so simply by dividing the standard deviation of growth rates in our sample by the square root of the number of years*.

My chart does this for each 20-year period since 1831; I'm using Bank of England data, updated by the ONS. The two lines show trend growth plus and minus one standard error. Roughly speaking, we can be two-thirds confident that trend growth in any 20-year period was somewhere between these two lines.

Which brings me to the point. Once we allow for this standard error, trend growth doesn't change much. In fact, except for a period to the mid-30s (which covered the post-war slump and Great Depression), it's quite possible that trend growth has never moved much from around 2%.

In other words, whether we have free markets, post-war social democracy or post-Thatcherite neoliberalism, trend growth might be much the same. Neither Victorian virtues of thrift and hard work (and hypocrisy) nor "dependency culture" much affected growth.

I'm not doing anything odd or original here. I'm just making the same point in time-series form as John Landon-Lane and Peter Robertson did in cross-country form when they concluded:

There are few, if any, feasible policies available that have a significant effect on long run growth rates.

It is possible that this is due to luck. Maybe good microeconomic policies have been swamped by bad macroeconomic times and vice versa. Perhaps the interventionist policies of the post-war era were bad for growth, but this was offset by an unusually favourable macro climate, whilst Thatcherite deregulation might have raised growth were it not that bad macro policy gave us two recessions.

Nevertheless, the data suggest that we should be very sceptical of the chances of improving our trend growth rate. Of course, many of the policies recommended by the LSE's Growth Commission - such as better education and infrastructure - are good ideas.But history suggests they are unlikely to have a noticeable effect upon trend growth.

* This method assumes that the observations are independent of each other. This isn't quite true in these data, as the serial correlation between annual growth rates since 1831 has been 0.28. I don't think this overturns my basic point.

Comments

[Ralph Musgrave]

Perhaps the explanation is that the basic explanation for growth is improved technology, and the latter has come on stream at a fairly regular pace over the last two hundred years or so.

[Luke]

Does the relative rate of population growth have any effect? I have no idea if there have been significant variations in the rate of population growth. (As I understand the figures you give, they are for the size of the economy as a whole, not GDP per head - sorry if I've misunderstood.)

[Peaknik]

For easy growth, try cheap energy...

[Yet another Chris]

Am I right in thinking that government expenditure is included within GDP? If so, does it make a difference to your thesis?

[Greg vP]

Two points.

The aim, the object, the purpose of economic growth is to improve the welfare of households. Therefore, what matters is not aggregate growth but per capita growth. (Private final consumption is an even better measure, but the time-series are harder to come by in many cases.)

The picture looks a little different when you look at this, but there is still no particular trend for the UK (data from FRED):

http://i.imgur.com/ycSiAvU.png

However when you look at the "advanced industrial countries" as a whole, a trend emerges:

http://i.imgur.com/KD6URjV.png

This does not change your conclusion that policy is irrelevant.

[Greg vP]

The explanation starts with Maslow's Hierarchy of Needs.

As people get richer, needs at the lower levels of the hierarchy are satisfied, and people start to demand goods and services to satisfy the higher levels.

It so happens that the higher levels are all about highly labour-intensive, highly individualised services and goods : boutique, hand-made clothes, shoes, and artworks, psychotherapy or life coaching, interior design, elaborately prepared meals, etc., etc.

Producing these goods and services does not admit productivity increases because their value resides precisely in the amount of labour time expended on them, and the exclusion of others from making use of that labour time.

What we are seeing is Baumol's cost disease writ large.

[Tammly]

What would be the result if you tried this with Germany? Perhaps remarkably similar results, with underlying growth roughly the same, and growth spurts only rectifying the enormous self inflicted damage of their notable military adventures.

[chris]

@ Luke - the chart shows GDP, not GDP per head. Repeating the exercise with GDP per capita (source = http://measuringworth.com/ukgdp/) shows that trend growth fell in the 90 years after 1850,from around 1.5% to perhaps zero by 1915-35, but rose from the 40s on. However, there's not much change since then. Growth of a bit less than 2% seems the norm.

[przeprowadzki]

How is promising tax cuts and jobs not a bribe from either side? Isn’t any campaign promise a bribe? przeprowadzki http://www.tanie-przeprowadzki.pl