Fun With Numbers

From yesterday's City AM:

THE NIGERIAN economy is nearly twice as large as official estimates previously suggested, making it Africa's biggest, after a recalculation by the country's statistical agency.

The exercise has pushed the west African nation’s calculation of nominal GDP in 2013 to 80.2 trillion Nigerian naira (£295bn), far higher than the previous estimate of 42.4 trillion naira.

This new method, announced yesterday, means the country of 170m people is Africa’s largest economy, outstripping South Africa by £60bn, and is the 26th biggest in the world.

From today's City AM:

We are about to see a beautiful demonstration of this with the British economy, where the official statisticians will shortly entirely and drastically rewrite decades of history...

One change will see research and development spending classified as capital expenditure; at a stroke, this will raise the level of the UK's economic output by a cool £25bn.

That’s just the beginning: overall, the statistical deckchair shuffling will boost the size of the UK economy by between 2.5 and five per cent, a shockingly large amount (and a vast range that makes it hard for outside forecasters to be able to predict exactly what the Office for National Statistics (ONS) will come up with).

Embedded Rents (2)

As I showed this morning, there are three types of goods and services:
i. goods which can be easily transported, and for which the retail price is the same anywhere in the UK, any difference in price would be competed away, so the price consists of [cost] only.
ii. goods and services which have to be consumed at or near the point of sale, the stronger the link between the consumption and the point of sale, the wider the variation in prices will be across the UK (or the closer the correlation between the price and local rent levels), so the price consists of [cost] + [embedded rent].
iii. pure land rents (or house prices above and beyond build costs) where consumption is by definition 100% linked to the location, there is no alternative and the price differences cannot be competed away. The price consists of [rent].

Handily enough, The Daily Mail published further stat's this afternoon on the correlation between average net wages and average rents in eleven regions of England & Wales.

They express rents as a % of net-of-tax wages; we can sort the regions be ascending average wage and it gives us a chart like this. As we see, rents as a share of 'GDP' increase as 'GDP' increases, so this is exactly the same picture as we had showing the increasing share of GDP taken up by rents over the last sixty years (GDP having increased significantly over that period): So far so good. It is of course just as interesting to plot wages net of rents and rents in absolute terms, to see whether Ricardo's Law Of Rent applies, i.e. that rents soak up extra wages in higher wage regions.

Oh, what a surprise, they do. However, Excel's wonderful TREND function tells us that rents only soak up 67.8% of the extra wages. Does that mean the law requires modification? No, not really, because once we deduct 'embedded rents' (calculated, for sake of argument as another 20% of pure rents*) and apply the TREND function again, we see that wages after pure rent and after embedded rent are as flat as a pancake.

I've left London off this one and as you can see, the only two regions where net wages buck the trend and are 'above the line' are the East Midlands and East (presumably because the landscape is a bit dull) and the South West is 'below the line' (because it has the nicest landscape and weather): * If I had tons more data and could be bothered to crunch the numbers, the more sophisticated approach would be to work out how much of his wages after rent somebody in the North East spends on Type i. goods and on Type ii. goods; then we would work out how much extra somebody in the South East would have to pay on the same basket of Type ii. goods and services. The extra cost in the South East represents the extra embedded rent, to give us a true figure for wages net of pure rents and net of embedded rents. But simply assuming that embedded rents are one-fifth as much as pure rents in any region is probably close enough.

Rents and taxes as a share of UK GDP

This is based on Bryan Kavanagh's chart, which relates to Australia. Rather unsurprisingly, we see that rents have increased from 10% to 18% of GDP and net earned income has fallen from 60% to 46% of GDP over the last sixty years as a result of GDP increasing.

Taxes have also gone up, from 30% to 36% of GDP.Sources: Nominal GDP and total taxes from the Public Sector Finances Databank. Total rents derived by multiplying average house price x 5% x number of dwellings plus one-fifth for commercial and farm rents. Average house price as published by the Nationwide. Number of dwellings as published by the DCLG.

Who said this?

While rummaging down the back of the internet, I stumbled across the excerpt below from Wiki's explanation of the 'value added' method of calculating GDP. I've replaced the missing word with "xxx", see if you can guess who or what "xxx" is before reading the Wiki article:

... xxx theory regards the "imputed rental value of owner-occupied housing" which is included in GDP as a fictitious entry; if the housing is owner-occupied, this housing cannot also yield real income from its market-based rental value at the same time.

In the 1993 manual of the United Nations System of National Accounts (UNSNA), the concept of "imputed rental value of owner occupied housing" is explained as follows:

"6.89. Heads of household who own the dwellings which the households occupy are formally treated as owners of unincorporated enterprises that produce housing services consumed by those same households... The same figure is recorded under household final consumption expenditures."

Xxx economists object to this accounting procedure on the ground that the monetary imputation made refers to a flow of income which does not exist, because most home owners do not rent out their homes if they are living in them.

Another important difference concerns the treatment of property rents, land rents and real estate rents. In the xxx interpretation, many of these rents, insofar as they are paid out of the sales of current output of production, constitute part of the new value created and part of the real cost structure of production. They should therefore be included in the valuation of the net product.

This contrasts with the conventional national accounting procedure, where many property rents are excluded from new value-added and net product on the ground that they do not reflect a productive contribution.

I'm with the mainstream view on this. Strictly speaking only building rents should be included in income/expenditure of owner-occupiers, because land rents are merely a transfer payment from non-landowners to landowners (thus have to be subtracted from GDP before being added back), but xxx is/are clearly wrong; imputed rental income is real income, in the same way as the apples you grow in your back garden and eat are real apples, and furthermore land rents do not form part of the cost of goods and services supplied or add to the value; they merely reflect how the income/profit is divided up between the productive business and the landowner.

So who or what is "xxx"? On the face of it, you'd be sorely tempted to say "Home-Owner-ist", but you'd be wrong.

Click and highlight to reveal the answer: Marxist.

Trade to population, GDP ratios

As I said here, "Of course imports (or exports) as a percentage of GDP are higher in the UK than in the USA or the Eurozone, because their economies are five times as big (five times as many people), but that was on the basis of three random figures* and basic logic.

Just to see whether this stacks up in real life, I have now taken the time and trouble to harvest the relevant figures for GDP, an average of exports/imports (referred to as 'trade') and population from the fine OECD website** and prepared a chart of the trade-to-GDP ratio plotted against population (click to enlarge):
The correlation between trade-to-GDP ratio and the logarithm of the population is 0.60***.

The chart for trade-to-GDP ratio plotted against GDP looks much the same. The correlation between the debt-to-GDP ratio and the logarithm of GDP is also 0.60*** (click to enlarge):Ah well, at least we know.

* One of which is open to debate as to its accuracy. The UK's and the USA's trade-to-GDP ratios were stated correctly at 32% and 16%, but the figure given by the FT for the Euro-zone was 16% but according to the OECD it's 41%. I suspect that the OECD merely added together country figures without netting off intra-Euro-zone trade. Twats.

** I used figures from 2008 for consistency, choosing GDP and import/export figures expressed in terms of USD/purchasing power parity/current prices. Luxembourg, with a population of less than half a million has a trade-to-GDP ratio of 161%, which is completely off the scale, probably exaggerated by all the cross-border trades routed through it for tax reasons, so I excluded that otherwise tip-top country.

*** I decided to chuck out the figures for the Euro-zone and recalculate the correlation of the trade-to-GDP ratio with the logarithm of GDP or population, which is 0.60 in each case. The correlation with the actual GDP or population is only 0.37 and 0.27, i.e. much lower, which were the figures I gave originally.

The Laffer Curve, GDP and the rental value of land

Here's the Power Point presentation I've been working on for the past couple of weeks:

I'm surprised people still fall for this New Labour clap trap

Rather surprisingly, a Tory sympathiser over at HPC said this:

Check the numbers Mark - yes, spending under Labour continued to climb steadily, despite the recession; but the sudden and dramatic budget deficit was largely down to an implosion in tax receipts, well in excess of the amount you would expect; were the collection of tax revenues proportionate to GDP.

I always check the numbers before I say anything, but just for clarity, total tax revenues according to the Public Sector Finances Databank (Excel, Tab C4) were as follows:
2007-08 £516.0 billion
2008-09 £508.0 billion
2009-10 £479.7 billion

That's a reduction of seven per cent over two years. Out of that £36.3 billion fall in revenues, nearly half relates to the fall in corporation tax receipts of £10.6 billion (a fall of 22%, i.e. the amount that banks used to pay) and the fall in Stamp Duty Land Tax receipts of £6.2 billion (a fall of 44%, which is what you'd expect if the number of property transactions falls by nearly half).

According to The Guardian, GDP fell by six and a half per cent in the first year of the recession and has barely picked up since, so the fall in tax revenues was more or less exactly proportional to the fall in GDP.

So the idea that the £150 billion-odd annual deficits they are running is all down to the recession is nonsense (OK, stick on £5 or £10 billion for additional unemployment benefits if you really must) and I'd say that tax revenues held up surprisingly well.

Growth? What growth?

Time for some more fun with numbers:

1. Nominal GDP in 1996-97 was £778.7 bn*.

2. Total nominal GDP for the ten years 1997-98 to 2006-07 was £10,592 bn.

3. If we'd had no real growth whatsover since 1996-97, i.e. if GDP had merely increased in line with RPI, total GDP in the same ten-year period would have been £8,992. So 'extra' GDP was £1,600 bn.

4. Household debts increased from £500 bn to £1,325 bn between May 1997 and April 2007.

5. Nominal government gross debt increased from £401 bn to £574 bn over the same period, chuck in another £50 bn for latent PFI & Northern Rock liabilities, that's an increase of £223 bn.

6. So two-thirds of that 'extra' GDP of £1,600 bn was financed purely out of additional household and government borrowing of £1,048 bn**.

Sure, for every liability, there is an asset. But that's not much consolation for the 95,000 families whose homes were subject to repossession orders last year; future taxpayers who are going to have to pay off this borrowing binge or depositors at banks who can't get their money back.

* All figures from here unless stated.

** So true 'extra' GDP over the last ten years was £552 bn. Which is equivalent to an annual £10 bn increase in GDP for ten years (£10 + £20 + £30 ... £100 = £550) or an average of about 1% economic growth.

I have posted this at LabourHome as well, just to see how they respond.