Or what Policy Exchange were going to call their latest piece of research, until a last minute change of heart.

UPDATE:  I would want to point out that anyone reading Andrew Lilico’s excellent submission to the SMPC can tell that he can produce interesting empirical stuff.

As readers of this blog will know, there are Keynesian effects and Non Keynesian effects.  Sometimes, the government increasing its demand increases the size of the economy. An obvious example would be during a deep depression, when the problem is a lack of spending, not supply constraints being hit.   At other times, an increase in government spending can operate through various constraints to ‘crowd out’ other resources.

Imagine a 1970s-style government, hogging the financial resources (available borrowing); taxing and directing spending into favoured sectors; taking the brightest staff; bidding up wages so that firms cannot compete; imposing taxes that take away profits.

You don’t have to imagine: it actually happened that way.  By 1974 British firms were in a crisis of profitability, while workers’ wages took an ever greater slice of the pie.  The government could not borrow at home or abroad and had to force the banks to take its debt, inflating the money supply.  Read Caincross and Burke on this.

So far, I doubt I have made anyone fall off their seats in surprise.  This is basic stuff.  I hear you saying:  We all know this can happen, oh FreeThinkingEconomist.  Thanks for the history lesson. What about what is actually going on now? Which situation are we in now? you ask, plaintively.

Policy Exchange had an opportunity to join this debate, but instead in their paper they have insisted on simply repeating generalities gained from scouring past papers on how economies in general have behaved during the post War period, and then used this to apply specific advice to our current situation.   Their paper is useful as a treasure trove of references for past fiscal consolidations that may or may not have the slightest relevance to the current situation.  It does very little to actually look at the current situation (the best for this is Lombard Street Research, in my view.  Or David Smith’s blog)

Let us look at one specific instance of how this ‘regress and policy advise to the average technique’ works in practise in this paper.

. We might expect that the interest rate penalty increases in a non linear way. In other words, moving from a 1% deficit to a 2% deficit is not likely to raise interest rates by as much as a shift from an 11% to a 12% deficit.

No, this has not happened.   Look at the evolution of the UK’s 2010-11 deficit as understood by the market.   Here are the figures taking 4 Budgets/PBR’s as observations:

2010-11 borrowing Yield on 2017 gilt
Apr-08 2 5
Nov-08 6.8 4.3
Apr-09 9.3 4
Nov-09 9.2 3.9

The deficit for that year shoots up.  The gilt yield goes down.  Why?  Here, take a deep breath, go off and read Scott Sumner for a few weeks, and come back having realised with a big light bulb over your head, this dramatic discovery:

The rate of interest is a function of ‘supply’ and ‘demand’

When businesses lose the appetite to invest, interest rates can fall even though the government starts asking for more from savers.  Look at how much Fixed Investment was expected to contribute to 2010 demand in 2008 , and then go look at the latest budget estimates.  The figure moves from 248 billion to 195 billion.   (You will notice that general government consumption raced up by just 20 billion, hmm what splurge …)

Also, when private savings rates move up from zilch to 8-10%, and corporate cash balances shoot up, the supply of funds holds down interest rates.  Supply. Demand. Look at both.

This is why their claim that the 12% deficit has caused rates to be 2% higher than they need be – hence generating the alarmist headline in ‘Banking Times’ about mortgages shooting up a grand a year – needs such a careful look at.

Could the UK achieve 2% long term rates if it cut its deficit to 0% now?  Yup, it sure could.  How?  Well, by creating deflation.  If the markets want a 1% real return, you they can get it with 2% long term mortgages if inflation expectations are crushed by 2-3% from where they are.  Oh please sir, I really want to do that.   How?  Well, you just withdraw demand rapidly from the economy.  Take 100 billion from demand via the government deficit, inflation expectations would dive and you bet you would get expectations of 1% inflation, and 2% nominal rates.  Just like Japan.  But this would not be because you have freed up the supply of funds from the overweening grabby government.  It would be because you have cratered all DEMAND for the funds.  Like Japan.  Clever.

This is not rocket science.  It does not involve the carefully search of thousands of academic papers trying to regress variable X against outcome Y.  And adding another hundred papers to the pile showing – quite correctly, quite respectfully – that cutting the government can boost the economy adds no weight whatsoever to the argument that it is a good idea now.   Look, I’m an economic liberal!  I believe this!  I am just also an empiricist!  And the only empirical stuff in PX’s paper is to put the UK as an XY dot on a graph showing its deficit against its spread over Bunds.

Policy Exchange have done the equivalent of advising people how to dress according to what the average weather is for the season, without glancing outside of the window to see what it is actually doing right now.

The way to go about economic-fiscal advice now should be 10% theory – because the theory is really not that hard – and 90% a careful awareness of current economic conditions.  It takes 10 minutes to explain theories of insufficient demand on the one side, and crowding out on the other.   It is much more frustrating examining actual economic data out there to work out which forces have precedence at any time.   But there are all perfectly available.

  • How are average wages doing?  Stagnant.
  • How are bond yields?  Despite the end of QE, still at near-lows.
  • Are big corporates crowded out of borrowing markets?  No they are sitting on cash.


Another absolute howler is this one: to equate the size of the fiscal deficit with the size of the fiscal stimulus.    Read this and gnash:

It is the scale of the deficit that, on the Keynesian story, provides the “injection” into the economy, boosting demand in the short term. So even on our proposal, there would be more “injection” than in any previous peacetime event.

Another deep breath: a loss of income is not necessarily a stimulus! So you cannot naively take the deficit and say “this is how much stimulus we are giving!”. *

Imagine a country that received a subsidy of 4% of GDP from some external source.  Then that subsidy was suddenly cut off.  Its deficit grows by 4% of GDP.   Is that stimulus?  No.  Of course not.   Failing to collect £110bn of revenue that you thought you could collect is not all stimulative, as it would be if in pre-Crisis Britain you had lowered Income Tax or VAT to the tune of £110bn.   Collapsing GDP incomes producing collapsing revenues does not equal a 10% of GDP stimulus, for pity’s sake!

A glance at tables 1.12 (ex post and here ex ante) show that government consumption has not provided a big incremental increase to demand.   If PX are really keen to know how much fiscal stimulus was actually applied, this IMF document may help them.    Their understanding of what LOOSE means might be updated with a glance at Slide 14 of this from those duffers the IFS.   Or see this post.  The government is taking demand from the economy in the next year!

*I appreciate that Brown misused this characterization at the G20 April 2009 in calling what they did a trillion dollars of support.


17 thoughts on “Flying the economy with your eyes shut …

  1. The way I think about it is this: Lots of people think that the deficit is somehow taking funding from the private sector, as though there is a fixed amount in the economy. However, credit is not supply constrained (Sumner is not a good guide on such matters, IMHO). And deficit spending necessarily creates new savings for the non-govt sector.

  2. “taxing and directing spending into favoured sectors; taking the brightest staff; bidding up wages so that firms cannot compete; imposing taxes that take away profits”

    I suppose when government is 50% of the economy (slightly above according to the CSM, slightly below according to you) one need not worry about government merely skimming off the best. We certainly have government using the tax system to direct resources into windmills (& the legal system to prevent nuclear electricity at possibly 1/10th the cost being produced) & imposing one of the world’s higher corporation taxes – the tax most specificly designed to take away profits.

    Presumably none of this is of any importance & can all be solved by printing more money.

  3. Good robust stuff. But

    a) We did not pretend this was an overview of the British economy. This was specifically about the question of whether cutting a large deficit should always be expected to cause a contraction initially – as assumed by much press commentary.
    b) We argue that there is a strong basis in theory and past evidence for the contention that this is not true. Even with much smaller deficits than the UK’s current one, reductions in deficits very often promote growth, rather than retarding it, even in the short term.
    c) Given that when deficits increase, the growth-promoting effects of deficit reduction are likely to grow and the growth-promoting effects of the deficit itself, at the margin, are likely to fall, the fact that in these other cases of much smaller deficits, reduction was very often growth-promoting even in the short term strongly, the natural conclusion is that in the UK’s current situation deficit reduction is particularly likely to be growth-promoting even in the short term.
    d) The key to really ensuring that deficit reduction will be growth promoting will be to ensure that, overwhelmingly, what is done in the short term is to cut spending.

    This leads us to the key point of difference between us. You believe that government spending has a growth-promoting function. I believe that that is virtually *never* the case. Virtually *all* government spending is growth-retarding. Exceptions are those very modest forms of expenditure that secure property and contract rights and enforce market-functioning-improving regulation. These tiny exceptions aside, government spending is not justified in terms of its effects upon economic growth. Rather, the point of government spending money is that by doing so we can achieve collectively valued social objectives, such as helping the poor and the sick and the elderly. I support much of this spending (though I obviously would prefer some of it spent in slightly different ways which I think would deliver better results in terms of helping poor, sick and elderly, etc.). But I support it acknowledging fully that it retards growth. That retardation is a price worth paying.

    The way that (bond-funded) deficits provide “stimulus” into economies is precisely by the process of borrow-and-spend. And I simply deny your position that it is not the quantum of the borrowing that provides the stimulus. In the Keynesian model there is a multiplier on G and a multiplier on T. G is an injection; T a withdrawal. When G is greater than T, there is a net stimulus. That is how it works. I accept no other concept of “stimulus” – and although you are right to note that many others in the press use the term “stimulus” in another way (meaning by the amount of “stimulus” the deliberate change at the margin), I believe they are clearly and unambiguously *wrong*. Borrowing can stimulate economies under certain limited circumstances, as I have always argued (specifically, when liquidity constraints or other imperfections serve to limit the functioning of Ricardian processes). But even in those circumstances there is a balance of positives and negatives. Spending, by contrast, virtually *never* provides stimulation of any sort. The only “pure” form of fiscal stimulus is a reduction in taxes; the only “pure” form of discretionary stimulus is a deliberate reduction in taxes.

    1. Let me add something else here. I have not the remotest idea on what basis you or anyone else believes that borrowing 12% of GDP is perfect in this scenario whilst borrowing 11% would plunge the economy into double dip. And I have even less idea why you think that 12% is perfect whilst 15% or 20% would be disastrous.

      Since you like evidence, perhaps you can enlighten me. In what past episode, anywhere in the world, has it been thought that borrowing 11% of GDP was inadequate, such that subsequently people said “With the benefit of hindsight, I wish we’d borrowed 12% of GDP – though of course borrowing 15% or 20% would have been a disaster”? I’ll bet money that you can’t even identify a single case in which people, ex post, regretted borrowing only 8%, let alone 12%, and I’d be very surprised if could provide a really decisive example in which 5% of GDP had proved too little. I, by contrast, have offered you lots of cases in which cutting deficits much lower than that promoted growth, even in the short term.

      Since you have no basis in evidence for a switch from 12% to 11% being damaging – how could you have, since deficits on this scale have virtually never happened in developed economies (and cases of 9% structural deficits have absolutely never happened) – since you have no evidence, perhaps you have some theory to support your contention? I’ve offered you some theory as to why one should expect the negative effects of deficits to escalate as those deficits themselves escalate – though I hardly think that point should be contentious. Do you dispute any of the theoretical considerations raised? Do you think I’ve missed some important point, perhaps some reason why the stimulatory effects of deficits might increase as deficits themselves increase?

      1. Damn I just lost a comment. That is annoying. This typed even faster


        Will help you out with a lot of this. Exhibit 18. Fallacy of composition.

        No, there are not many examples of this. My whole thesis all along has been that rules that apply generally during a long history of recessions that are largely triggered by inflation/supply limits/etc do not work in a minority of cases, and people need to apply skill and empirical curiosity to work out when those conditions apply. The average weather in August is perfect for wearing shorts.

        And as for counterexamples, look up the US behaviour during 1937-8 as well (the FDR mini depression)


        and of course the whole second world War, which I agree does not count much.

        Your idea that the government “chooses” the deficit when these dynamics are operating also needs to be updated. A few hours with Dillow would help here:


        Andrew, at some point your advice will be quite right. I hope that by watching the actual economic conditions on the ground I will be with you by then. We will be able to tell it in a wide range of surveys about private sector behaviour and expectations and so on. But I do not believe in operating on the basis of general advice to specific situations like this.

    2. P.S. You might respond to my remarks about the Keynesian model by talking about a “balanced budget multipler”. If you really want to, we can get into that, by I think it’s rather beside the point here so I’ll leave it for now.

    3. I’m due on holiday soon, so apologies for brevity, and lack of logical order. On the last point we are in total disagreement, and I will just shelter in cowardly form behind the IMF. If the oil industry collapsed and stopped sending revenues to the government, which then had to fund more of its spending through borrowing than taxing, there has not been a sudden extra stimulus. The only sense in which you and Gordon Brown would be right on this is if the correct baseline should always be taxes = spending. In other words, the economy is more stimulated than it would have been had the government been daft, and assumed that all spending should always be tax-financed, and therefore compared its actions to that counterfactual.

      But that is not the way to look at it. You beg the question by assuming that the baseline is always G = T and that anything greater than that is extra stimulus. And I would be amazed if any decent macroeconomic outfit, from the OECD to the IMF to the IFS, saw it that way. In fact you can see from the links that they do not.

      But let’s reach a point of agreement: if the correct baseline is running a balance budget, then the government running a 12% stimulus is 12% more stimulative than this baseline. I think this baseline might have represented a Depression-slump of 15%, probably more, given all the unemployment and tax rises that would have been required to close the deficit to zero in a slump. From that point of view, this mischaracterized stimulus was blindlingly successful. But I think neither you nor Gordon Brown have the right to use that as the baseline.

      Penultimate paragraph. You are obsessed with the supply frontier, which government spending nearly always damages. I am obsessed with the current distance to that frontier, which spending can close. this is why I think you pay too little attention, and are too “underpants gnomes” about the means by which demand rises to supply. It is almost as if you believe in Says Law. If I too believed in Says law and thought that when demand falls there is always another demand to fill its place, I would be with you. I think the last recession has shown powerfully how Says Law does not operate in this or any other large macroeconomy.

      But at least we know where we differ.

      I thoroughly agree with (d) – you will see me in House magazine agreeing that point next week.

      If you mean it about (a) then I don’t know why your paper is structured as advice to act NOW on the deficit. Because in my way of seeing it, the ????? of the mechanism from government spending to growth needs a lot of careful empirical investigation. (spending falls –>>> ????? ->>> growth) which is perfectly testable. So spending CAN crowd out other demand. IS it?

      on (b) I would facetiously reply: August days are often sunny. On average high government deficits have reduced growth. But you and the Tories managed to find one of the 10% moments when they were cushioning the fall in demand – and speak against them. At that one moment the Tories gave Labour a fighting chance in the election.

      So Andrew, we have a really simple disagreement. I personally hold that your perfectly well-founded belief that high government spending damages the economy is allowed to carry with too great strength into a general prescription that applies wherever the economy is in cyclical terms. I hold that before people come out with categoric advice that ignores the current cyclical indicators you run the risk of doing a Japan on us.

      1. Indeed. I think you have it precisely. You buy a “demand gap” classical Keynesian story in which there is sometimes inefficient appetite from private agents for goods and services, and hence believe that spending is (at least sometimes) stimulatory. I believe there is always precisely sufficient demand (so government spending is (virtually) always contractionary), but that sometimes private agents face liquidity constraints that justify the government borrowing, so that sometimes borrowing is stimulatory. Hence for me the only (growth-related) question is whether the gains (which are always there, in growth terms) from cutting spending are outweighed by the losses in terms of reduced stimulatory borrowing (losses which tend to fall as deficits rise).

        It’s just a flat disagreement about how modern market economies work.

      2. Fair dos! Now we know where we stand!

        Does that mean I can call you an Austrian?

      3. I am not an Austrian, though I find the Austrian position interesting in a number of ways. In particular, I have occasionally flirted with the thought of doing some maths concerning the behaviour of agents that did not know their own preferences. My position is pretty close to the New Classical school.

  4. Here’s another thing. We say that marginal changes in deficits are likely to increase interest rates by more when deficits are high than when deficits are low. You declare this untrue, quoting expectations for the 2010/11 deficit at various points. But what has that got to do with anything? Do you actually deny that if the UK were currently running a 5% of GDP deficit instead of a 12% deficit, and all else were equal – we’d had the same recession, the same amount of QE, and so forth, then gilt rates would not be lower? I think the truth is that you are as sure as I am that gilt rates would be lower. It is not contentious that when deficits are larger, gilt rates are higher – there is a mountain of evidence supporting this contention, of which we quoted only a bit. What also seems obvious to me – though there is less empirical evidence for it – is that the marginal increase in gilt rates with marginal increases in deficits is more when deficits are greater than when they are smaller. We found that the effect is between two and four times as great – though we claim no statistical robustness for those numbers, and our mini-case study is intended to illustrate that effect, not to prove it – the effect is actually intuitively obvious, and doesn’t really need proving.

    1. Lots of counterfactuals here.

      If we TRIED to run a much lower deficit, I suspect it would happen by way of much lower GDP, which then might hit revenues, and increase the deficit. I will post an Excel Model for you later if you like. We may manage it, and then find ourselves with very low inflation and very low NGDP expectations, which are consistent with low gilt rates, yes. But these would make no-one better off- the mortgage holders you are alarming in today’s headlines would have far lower house prices and projected house price growth and real income growth. So they would party like a Japanese Housewivfe, 1997 edition. Not very much.

      Have you ever looked at how big Japan’s debt has been and how low its rates, incidentally? Given how often people like me bore you with the Japanese counterexample, I am surprised you have not done a paper on it.

      Here is my contention: a successful economic policy from here on would see HIGHER gilt rates, because they would be consistent with higher demand for funds and a recovering economy. As Sumner says, low rates are often a sign that conditions HAVE been tight, rather than that they ARE loose. If Osborne wants to aim for 3% long rates in a country used to 2-3% inflation, he is aiming to crater the economy. And given the credit spreads that would operate, I would not go promising people that their mortgages would fall.

      Oh, here is a spreadsheet


      I will try to post to explain what on Earth I am on about!

      Enjoying myself as alway s….

      1. I 100% agree that an increase in growth prospects should, ceteris paribus, be expected to result in a rise in gilt rates. I also 100% agree that attempts to reduce deficits can result in deficits increasing – this is particularly the case for tax-rises-based consolidation attempts. So let us think about two options:

        a) We raise taxes, locking in high spending, at the same time successfully reducing a deficit. Then gilt rates will fall for two reasons: growth prospects are reduced (because of the spending rises) and the deficit is reduced (so default and inflation risk falls).
        b) We cut spending, at the same time successfully reducing a deficit. Then there will be two effects on gilt rates. They will be tending to rise because of improved growth prospects (because of the spending reduction), and to fall because of the reduction in the deficit.

        Even though case (b) results in higher interest rates (for the same reduction in the deficit), it promotes more growth even in the short term.

        It is the spending reduction that is the key driver of the increased growth effect from spending-cuts-based fiscal consolidations. All that the discussion of deficit reduction does is (a) to point at a secondary factor, a by-product if you like of the spending cuts (a reduction in the deficit); and (b) to puncture the view that large deficits are always stimulatory (even as you define “stimulatory”, you agree with me on that one!).

  5. Posted by Andrew Lilico on April 1, 2010 at 1:52 pm

    ‘…you lots of cases in which cutting deficits much lower than that promoted growth, even in the short term. ‘

    It all depends on whether the cutting of the deficit promoted the growth or the higher growth meant that the high deficit was not required.

    It is similar to the Reinhart and Rogoff finding that a public debt-to-GDP ratio above 90% is consistent with a fall of 1% GDP growth. Was it lower growth that pushed public debt above 90% or did the rise in public debt cause a slowdown in growth? They have shown a correlation but not necessarily causation.

    Posted by Andrew Lilico on April 1, 2010 at 2:27 pm
    ‘ It is not contentious that when deficits are larger, gilt rates are higher – there is a mountain of evidence supporting this contention, of which we quoted only a bit. What also seems obvious to me – though there is less empirical evidence for it – is that the marginal increase in gilt rates with marginal increases in deficits is more when deficits are greater than when they are smaller. ‘

    It would all depend on the robustness of the economy. If the economy was functioning well and growing. The government running a large deficit in such circumstances would be faced with competition for funds and yields would rise. Part of the reason why present yields are not as high as many forecast is because of a lack of competition through reduced demand for funds by the private sector. At the moment the government are not crowding out but are in fact crowding in spending.

    Over any long-run period gilt yields for a sovereign issuer track inflation as the key driver not growth per se. Obviously growth prospects affects inflation but the real driver of yields is inflation expectations. Moreover, there is a relationship between the real yield on government debt and the yield on equities, it is not all about the government fiscal deficit. To look historically at yields relative to deficits also requires to look at what equities were yielding in the same period with due consideration that government debt is a risk averse play.

  6. BTW: Why stop with government borrowing and deficit spending–why not extend the analysis to every sector in the economy?

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 )

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s