Winner of the New Statesman SPERI Prize in Political Economy 2016

Tuesday 23 October 2012

A short answer to Tyler Cowen’s question

The question, following the observation that UK real wages have fallen substantially, is “In the traditional Keynesian story, stimulus lowers real wages through nominal reflation.  Is that the Keynesian view here?  If so, why do Keynesians believe that British real wages need to fall more than 8.5%?”

Actually I think this question has nothing especially to do with the UK, and a great deal with Keynesian theory, or at least how Keynesian theory is often taught. The problem in the UK and nearly everywhere else right now is lack of aggregate demand, not the level of real wages. Why, asks Tyler, are not more workers being hired if real wages are so low? Well low real wages are having some effect – it is one factor behind the ‘productivity puzzle’ that is widely discussed in the UK. But when the problem is aggregate demand, low and declining real wages will not ensure ‘full employment’. Firms may employ more labour, but there is no reason to expect the labour market to clear.

I think what lies behind this question is the idea that aggregate demand matters because sticky wages are preventing the labour market clearing. So, to rephrase Tyler’s question, falling UK real wages do not look very sticky, so where is the problem? When I was studying macro, there were these debates about whether it was sticky wages, or sticky prices, which underpinned Keynesian theory.

The answer I would give, at least at the zero lower bound under inflation targeting, is neither. To get technical, imagine a toy model with imperfectly competitive firms who set a constant mark-up on labour costs, and a linear technology. Nominal wages could fall forever, but firms would cut prices to match, so the real wage would not change. So the question is then whether falling prices will raise aggregate demand. But why should they, particularly if nominal interest rates are stuck at zero, and the central bank puts a lid on expected inflation.

The price that is ‘wrong’ when aggregate demand is deficient is the real interest rate. This is one area where New Keynesian theory has helped sort out old Keynesian confusions. If falling wages and prices do nothing to change real interest rates, then aggregate demand need not rise. And if aggregate demand does not rise, unemployment can persist. But at the zero lower bound, with a central bank mandated to target inflation (and a government showing no signs of changing that mandate), the real interest rate cannot be reduced. So expansionary fiscal policy is needed to raise demand, not lower real wages. Or, in the case of the UK at the moment, contractionary fiscal policy is currently reducing demand and therefore output.


  1. I have asked this repeatedly.

    With real wages declining and indebted households busy deleveraging further monetary stimulus driving up inflation expectation will spur a Ricardian reaction by private households, whereby they will respond by spending less to mantain their saving rate. This will tend to reduce agregate demand even further, thus unleashing a vicious circle.

    What is your answer to this argument?

    1. I am not sure why you are arguing that lower real interest rates would lead indebted households to save more. Is it because you are assuming higher inflation would lead to lower real wages? But there is no reason why it should.

    2. Demand deficiency is in large part due to lower consumption levels from private households. Private households would naturally consume less when they expect higher inflation to squeeze their disposable income due to wages not adjusting for inflation entirely. Wages do not adjust given that the labour market is loose, with people renouncing to wage increases just to keep their jobs (i.e., productivity puzzle) and being uncertain about future employment. This in turn (the vicious cycle) will strengthen their cautiousness and, thus, demand deficiency.

      Basically, it is a clash between opposing arguments about which expectation dominates: NGDP-like advocates say that increased inflation signal a committment to a growing economy and ultimately optimistic employment conditions; whereas I claim this effect is more than offset by immediate/prominent/urgent concerns about real wages not catching up with raising costs of living. My point is that this Ricardian concern dominates due to the higher sensitivity when private households are deleveraging.

      Hope this helps to make it clearer.

    3. Thanks for the clarification. I agree what you suggest is a possibility, particularly if there is some resistance to nominal wage cuts, so inflation could in itself reduce real wages. However as most debt contracts are in nominal terms, even a small increase in nominal wages would reduce liquidity constraints associated with these debts. In the 1930s, the move from negative to positive inflation seemed to help the recovery.

      However, I think you agree with my main point here, which is to focus on aggregate demand, and not on the need for real wages to fall to clear the labour market.

  2. @Paolo - surely the assumption in a demand-defficient economy is that stimulus will have negligible impact on inflation, thus invalidating the argument?

    Not being an economist, I'm not going to claim any great insight, but I can at least imagine a situation where an increase in demand would actually *decrease* prices. If large companies look at a stagnant environment, they surely see reduced opportunity to increase customer numbers. The obvious method of a CEO attaining their quarterly bonus (assuming it is linked to performance what-so-ever, which is probably the biggest assumption here) is to increase prices for existing customers.

    Investment in activity such as increasing housing supply should also have deflationary effects, due to more obvious reasons.

  3. @ David

    Quite simply, the author of this blog as well as many other advocates of more monetary stimulus (and new NGDP targets) argued repeatedly that in a liquidity trap the aim should be to raise inflation expectations to drive the real interest rate into negative territory. (Krugman argues this should be done credibly and permanently, meaning a commitment to keep inflation up even past the point where the economy has recovered).

    Now, I would argue that when you are busy deleveraging you'd pay more attention than usual to these "expectations", hence the Ricardian response due to heightened sensitivity.

  4. Prof. Wren-Lewis' answer jumped gears, I think. Prof. Cowen pointed out that REAL wages were falling. Prof. Wren-Lewis answer was that if prices fall as fast as wages, then there would be no effect. But that would mean that real wages didn't fall. But UK real wages did fall, so we are back to square 1 without an answer.

    1. That example was meant to illustrate that what is happening to real wages is not the issue. The problem is demand deficiency, and not that real wages are too high.

    2. I think you need to drop the assumption of constant markups then, else in a canonical NK model firm production decisions would be at (second best) equilibrium levels.

      Implicitly, if you are assuming sticky prices, a fall in the natural interest rate leads to rising markups, declining real marginal costs and, thus, falling real wages.

  5. And surely enough.....

  6. @Paolo - not sure what the FT blogpost has to do with your question - the reality (unmentioned by the FT, but easily checked if you live in the UK and stroll down to your bank) is that funding for lending has not changed the interest rates the bank is demanding of consumers and house-hunters. As such, it's no wonder it is having no effect.

  7. Prof. Wren-Lewis,

    I have two questions:

    1) Is it your view that increasing aggregate demand would lead to an increase in inflation?

    2) If it is the case that increasing AD through fiscal expansion would lead to an increase in inflation, do you think the BoE tolerate it, or would it raise interest rates sooner than it otherwise would and offset the impact of fiscal expansion on AD?

    I guess what I'm asking is whether or not you think fiscal expansion is fully sufficient for increasing aggregate demand, or whether it is a necessary condition (because of the ZLB) but insufficient without also changing the BoE's policy target. My guess is somewhere between the two (that is to say, the central bank would tolerate some increase in inflation due to fiscal expansion, but probably not enough to fully offset the AD shortfall), but I'm not sure.


    1. Dear Richard - I agree with what you say. Less austerity might lead to higher inflation, which might provoke the Bank into raising rates (as they nearly did last year), in which case the focus would need to shift to stopping them doing that. But I do not think this possible reaction is any reason not to argue for less fiscal austerity.

  8. Dear Simon, another fascinating blog. A question if I may: you say that the interest rate is the "wrong" price. this certainly helps me think about the problem , since nominal wages seem flexable. But in the UK, inflation has been high and policy interest rates low, so that real interest rates are very low. This makes me wonder how much lower do they have to be for the real interest rate to be restored to the right price (in the sense that sticky nominal wage models have real wages too high).

    one answer i guess is that a dysfunctional banking system means that policy rates are low, but actual rates are higher. from a modelling point of view, that puts me in mind of Brad de Long's IS-LM models with an additional risk wedge on interest rates. From a policy point of view, does that not put the emphasis on fixing the bank sytem rather than fiscal policy, or, if the problem is a dysfunctional bank system, can fiscal expansion solve it?

    With thanks again for your blog posts,
    Jonathan Haskel (Imperial College)

    1. Dear Jonathan - I think you are right that the banking crisis means that the real interest rate required is very low. I also agree that trying to fix that would be good, but I do not have any quick fix for that.

      What that means is that we are in a period where aggregate demand determines output. As a result, austerity reduces output and fiscal stimulus would increase it. So it makes good sense to provide fiscal support while the banking sector sorts itself out. I agree fiscal policy will not solve the banking problem, but it will not make it more difficult to solve either, and it could ease the pain while a solution is found.


Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.