A standard attack on Labour is that it has left the government dominating the economy, which spells doom for our future growth. Policy Exchange’s approach to this has been the most uncompromising and (as I shall show) wrongheaded; see this publication, mentioned by the Wolf, and utterly debunked by me.*

A state that actually dominated economic production is terrible for economic efficiency.  For the UK to grow for 200 years has needed capitalism’s endless, restless search for better ways of doing things.  Good ideas get rewarded and prosper– for a while – and bad ideas get thrown out.  This doesn’t work when the state is in control, and ordinary commercial motives are absent. Imagine if the development of the IT industry had been all decided in think tanks and Whitehall.  We would still be on BBC Microcomputers.

So, are we in a 50%-state economy?  Superficially the government spends about £650bn of a £1400bn  GDP economy.  But does it feel like that?  Um, no.   Think about it: the state employs 6 million people, or about 20% of the workforce (h/t John Redwood).  Half of the state’s spending is actually transferring money to people so that they can spend it themselves. Government consumption is about 20-25%.   So, for example, Mrs Bloggs get £10k in housing benefit, and she decides where to rent.  Ordinary market incentives still intact.

But this figure still overstates how much the government ‘dominates’ the economy. To see how, go over your annual transactions.  How many of them involve dealing with the state?  When you buy food, clothing, alcohol, leisure, meals out, holidays, and so on, do you deal with the State?   In fact, go to Table 4.4 of this (large) document, and ask yourself – how many of these are state provided?  Of course, this is deceptive: we ‘purchase’ a lot of things like health and education from the State without handing over money.   Nevertheless, the number of transactions that involve the state being on one side is fairly small. Even if the government does take about 25% and use it to provide teaching and healthcare, most of the onward transactions from there on are private, decentralised affairs.

Consider a straightforward hospital visit, with some drugs provided free.  The State pays £X to a nurse to look after you and £Z to get you drugs.  Consider the drugs side of things. Of the £Z on drugs, most goes in a further transaction to a drug company, which then goes about decision making in the good ole’ capitalist way.     Now consider the nurse. Yes, she is chosen according to state criteria, with baleful national wage bargaining: good nurses and bad nurses are paid the same, roughly, which is bad.  But most of what is paid to the nurse gets spent on other things, most of which are subject to capitalist discipline:

What matters, as Congdon ironically reminded me, is the quantity of transactions that are state-dominated. In total the UK economy has about £60,000 billion of payments made– or 50 times the size of GDP. ** How many of these involve the government as buyer or seller?  My guess: much less than 10%.

This distinction can be illustrated with a thought experiment. Imagine an economy – PrivateLand – in which the government ‘does’ nothing – produces no goods or services – except remove about 50% of the income of some people, and give it out to other people.  Education, health, even defence are all provided privately. How the tax was gathered would affect economic incentives, but every further economic transaction takes place in a competitive marketplace.

Now consider an economy – SovietLand – where 50% of every economic activity is done by the state, paid for by taxation.   The state takes tax, and uses it to provide, free, loads of groceries, clothes, iPods, holidays, financial advice, music etc that it makes itself.

Both these economies would have a 50% government-spending to GDP ratio. But PrivateLand would clearly be the most efficient.  This is because prices and wages, and what is produced, and how, are all going to be decided by market participants, who get punished for doing things badly.  In SovietLand, state competition undermines private provision, distorts wages, provides bad products, opens the field to corruption and favouritism, has unsackable employees, and so on.  Queuing rather than price movement dominates.

So what matters is not “how big is the State” in terms of income/GDP.   What matters is how much of the economy is subject to decentralised, capitalist discipline.  You could imagine a small state – like, say, Medieval Britain – where the King actually took only 10% of the output, but where rules and restrictions (guilds) and custom utterly self-feeding capitalism taking hold.

And you could have a larger state – Sweden, say – that redistributed quite a bit, got Policy Exchange all hysterical, but had no nationalised industries and no stifling market-defying regulations.  This could quite easily grow fast.

The likes of Policy Exchange, and all the hysterics worrying about how large the state has become in the last two years, are misusing one metric – the spending/GDP ratio – to present a false picture of how much of what we do is actually state-determined.  Let’s be serious: spending on this naive measure jumped from ~39% to %48% in a matter of 4 years.  Did this mean the government suddenly ‘taking over’ 10% of the economy’s production decisions?  Of course not! The absolute volume of government-economy transactions probably grew a little, and the volume of private-private transactions probably fell a little.  If the government had bought Tescos (£47bn annual sales worldwide – 3% of UK GDP) and started running it like a Quango, we would have had a hit to our productivity.  But that is not what happened.

The interesting ratio would be that between the number of transactions involving the State and £60,000bn.  I have no earthly clue what that is, except for “much much less than 50%”.

Does this mean the Left does not need to worry about Big government damaging economic growth?  No, of course not.  For a start, the State can damage the economy through pure regulation, by for example dictating what everyone should be paid according to some disastrous NEFish criteria.  It can also damage at the margin.  A single state supermarket might damage all the grocery market by undercutting unfairly.  Its methods of redistribution also matter: taxes on rents are better than those on economic activity.  But simply parroting the public spending/GDP ratio achieve very little indeed.

*In fact, Policy Exchange got so carried away with this idea that they mistakenly stated that the Private economy has not grown at all for a decade.  David Smith put them right, embarrassingly

**To generate £1500bn of ‘value-add’, you need a whole lot more transactions, because only a small amount of each action involves adding more.   Economic production involves taking something for £XX – a lump of iron, an hour of a graduate’s time, some flour and water –  doing something with it, and selling it on for £XX plus a bit.  The “a bit” is the value add.


17 thoughts on “The government is smaller than you think

  1. I’m not clear on why you want to use the fraction of transactions in which the government is involved rather than simply the fraction of GDP which is government consumption.

    Consider what happens to an economy as the private firms become more integrated (e.g. if Tesco were to buy out its suppliers)–the total number of transactions in the economy falls (as there is more activity within firms), but the value added in the economy is roughly the same. This integration would increase the apparent size of government by your transactions measure (because the denominator shrinks), but I don’t see why this should matter. Don’t we just care what fraction of GDP is being consumed by the state?

    1. That is a good point, although the Tescos example you give there would result in a reduction in market-transactions in a sense. What I am disputing is whether there is a particular magic about the (Government Spending)/GDP ratio. I have heard a right wing industrialist stand up and explain to the Carlton Club that when this ratio hits 45, everything goes downhill. Period. Which is plain silly: it can shoot up to 45 just by the GDP cratering in a recession, and cutting the G would have just hurt the Y some more.

      “Don’t we just care what fraction of GDP is being consumed by the state?” I think this way of putting it is simplistic. There is a case for arguing that as an economy develops, more of the goods that its inhabitants need and want are of a public nature, and require collective purchase. So: in 1600, we just want security and food. In 1900, we need more public infrastructure, and can afford socially enhancing welfare and insurance, better provided collectively. In 2100, public health and environmental goods may predominate – and how these are best purchased (not produced) is a technical question in some ways. The Government is not like a baron squatting on a pile of stolen gold – it should be a means to our ends.

      ALSO: when the government pays for something – say a teacher’s salary – it does not “consume” that salary. The teacher does, and then whoever the teacher buys her food clothes and holidays off goes on to consume, and so on. This is why I think measuring transactions matter. In “PrivateLand”, you can imagine the government doing one simple transaction, and everything else being roughly private. This matters for efficiency reasons. So, NO, I don’t think that this “GDP being consumed by the state” measure is all – that is what I am trying to drive at.

      1. Ok, but at the risk of sounding thick: I still don’t get it. Yes, it’s true that some of the teacher’s salary will be spent on cookies, but, similarly, some of the money earned by the bakers is taxed away and spent on yet more teachers. Some of those teachers eat cookies, and their bakers are taxed to feed other teachers. The sectors leak back and forth, and in the end, the amount of real resources at the government’s command is roughly proportional to government consumption (but not gov’t spending) as a proportion of GDP, isn’t it?

        As far as your point about changing demands over time, I agree entirely. Many of the goods provided by governments are luxuries (in the sense that we demand more schools, hospitals, and art galleries as our incomes rise) and so we would expect modern, wealthy societies to properly demand more such goods. I was just debating the right way to compare the size of government across modern states.

      2. You are certainly not being thick!

        What I am concerned about is not so much resources as the number and volume of transactions for which the motives involved are those that enhance productivity and efficiency. In total, all the taxes add up to about 600bn, that is removed from people according to no capitalist motivations at all – sheer expropriation. The cookie buying, iPod buying, housing choosing side adds up to way more, I think.

        In the Tim Congdon piece that got this ball rolling, he makes out that a bank loan is trivial in the economic scale of things, because #100bn of loans is little compared to #60000bn of transactions. I think this is nuts, by the way – because that 100bn of loans being granted can then multiply through the economy – a developer uses it to buy land and get building, the builders buy food, the farmers buy something else, etc. But, to provide an analogy, if #100bn were so used to start of a lot of economic processes, would it be accurate to say ‘this is a loan-dominated economy. Everything is decided by Barclays’. I would say no: most of what goes on still involves free-choosing agents.

        I do not wish to imply that this means any amount of government tax and spend is good. Nor that someone could announce ‘mission accomplished’ if the government shrank to 35pc of the economy on this metric – it could still be interfering through regulation as Tim Leunig implies, and my original post said all along. Nor (and I should have made this more clear) do I propose Govt Transactions/All Transactions as the sole other metric: as you and Tim have observed, this could produce real arbitrariness. But the state of the debate in this country about what makes a government too big is shocking, and I stand by my criticisms of it. There is not a simple rule that 45pc = Heading to Soviet Productivity and 35pc = Singapore Here We Come. How the government gets involved is crucial.

        Above all, keeping its spending plans orderly when the private sector suddenly decides to panic into savings, does not represent a sudden takeover of the economy. It stopped the recession getting much worse.

  2. “Imagine if the development of the IT industry had been all decided in think tanks and Whitehall. We would still be on BBC Microcomputers.”

    In some ways I think that would be preferable to using Windows 7.

    1. even over Vista? that is saying something.

      I have long felt that Bill Gates’ tombstone ought to be shaped like a giant hourglass

  3. I think you’re a little unfair on the BBC Model B! That was also contracted out to a private company, which then managed to produce the Archimedes and RISCPC which commercial realities prevent me saying were ‘better’ than the competition, but they were sort of better.


    On your wider point you’re obviously right but it’s very complex. Gordon Brown, or the Queen, or whoever is nominally in charge doesn’t really produce anything so all government spending is delegated, and you could even say the purchase of an aircraft carrier is done by individuals down the line. I suppose the issue is what % of consumption is purchased by individuals for themselves out of total income. And so a teacher’s salary of say £25,000, which I pay from my taxes, the ‘state’ bit is only the value-added that I get from her/his services, the rest is a transfer payment much like pensions or social security (an obviously I am only one of many taxpayers).

    1. To be honest, I now wish I had used some other example. I wanted to say “development of the Internet” but then some *&*&* would have come back at me with Arpanet and Al Gore . . . . I thought Elite on the BBC Micro was as miraculous an innovation as 32KB could achieve.

      I am sulking with computers since the Wii Fit this morning pronounced me 55 years old. I’m 37. Bring back the BBC

  4. I would say that the WII recognises your intellectual stature, were it not for your daft post about transaction costs. If I buy my groceries from Ocado, who buy from Waitrose, rather than buying them from Waitrose direct, on your model the private sector grows relative to the state sector. But that is clearly not the case. What matters is value added.

    Also, govt does not have to buy things to be influential. If it taxes petrol and gives it back as a lump sum, and I happen to use the average amount of fuel so that my income is unaffected overall, the government has still had an influence over my life (it has changed my incentives to drive overall, to drive one car rather than another, and over the style of my driving). That the govt has bought nothing does not make it uninfluential.

    And govt has a huge effect in other ways. There is a lot of compulsion – either you have to do something, or mustn’t do it – going to school, use a seat belt, build houses that are not too dense, but dense enough, insulation standards, driveway angle standards, road widths, light bulb types, minimum wages, maximum hours, bank holidays, lunch breaks, when the trains run, pavement surfacing, fire exits, food additives, which side the cold tap must be on, ingredient labelling, allowable garden chemicals, use and sale of drugs, age at which you can drink, drive, join the army, work full time, that sikh dentists cannot open on Good Friday, the national curriculum, the location of vending machines on platform 5 of Wimbledon station, class sizes, length of degrees, bus lanes, low emission zones, aircraft noise, where you can smoke, what you can smoke, radio interference of electrical items, the use of the radio frequency, your right to a sewage abatement if you have a garden irrigation scheme, the right to chop down trees of a certain size, your ability to tackle burglars in your home, the width of doors in new buildings, the number of litres of water a toilet must use, data protection, advertising standards, discrimination laws, charity law, tyre tread, fire extinguishers, boiler servicing intervals in rented housing and offices, leasehold extensions, the right to light, TUPE, the direction of electrical wires from your sockets, you right to install a boiler, or an electrical socket, your right to plant trees, feed pigeons, British Telecom’s obligation to maintain its archive, play loud music, walk around naked, your right to travel abroad without being searched, the way in which you treat animals, which animals you keep, you right to grow more than 1 hectare of potatos, or buy more than one tonne of them, your right to store flammable liquids, sell more than a small number of paracetamol to someone – indeed, your right to sell paracetamol at all – to award degrees, sell alcohol, tobacco, or knives, your right to take photos of some buildings, the height of car bumpers, the speed at which you can drive, the size and font of your number plate, and I could go on, and on, and on.

    And you really think government is small? It might be the right size, but it is not small.

    Read PJ O’Rourke Paliament of Whores http://www.amazon.co.uk/Parliament-Whores-Humorist-Attempts-Government/dp/0330323695/ref=sr_1_16?ie=UTF8&s=books&qid=1262825414&sr=1-16. It is very funny, and often apt.

    1. Tim, I recognised that point, which is why I raised the example of how smaller pre-capitalist states have behaved. My point is that the straightforward ratio of taxes or spending to GDP do not capture the extent to which the government interferes in the economy – you can have two different economies that are identical on that metric and interfere in radically different ways. What matters, as you seem to agree, is how much government behaviour prevents ordinary capitalist motivations from influencing economic decisions. Economic decisions take place on billions of occasions, at every possible transaction.

      It is not about transaction costs!

      I don’t think the government is small – but that the kneejerk statements based on the G/Y ratio do not capture how big it is. I think they are close to irrelevant, at times, for the problem the critics are aiming at – how government dominated the economy is becoming. It can be better, or worse, than the ratio suggests. Me, I still think we are way more capitalist and free than the 50% ratio currently suggests. The ratio was <20% 200 years ago, but I don't think 1800 was a capitalist nirvanna.

      I agree about GVA being significant. How much of value added comes from government-owned production, compared to when we owned steel, rail, telecoms, etc under the wise leadership of Tony Benn? How much GVA is from the Govt now, in your estimation?

      I appreciate that you sit on the board of PX and I ought to moderate any disrespect towards them. Also, that liberals should be nervous about big governments – I am not intending to defend New Labour style invasiveness here (as I thought I made patently clear). Expect quite a lot from CentreForum on how to reduce government involvement in things like pay, in the near future. All I am saying is that such pieces and polemics will NOT be led simple-mindedly by one single metric, which the further right often use as the only tool for telling whether the government is getting TOO big. YOur sort of more qualitative examples are much more appropriate.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s