Infinitely short and not variable

Here’s a recent restatement by James Alexander, at Marcus Nunes’ blog, of one of the tenets of market monetarism.  That monetary policy – via financial markets – has its effects instantaneously, always, and that the ‘long and variable lags’ commented on by Milton Friedman, which this blog is named after, are a figment of mainstream monetary economists’ imaginations.

“Many conventional macroeconomists still cling to the notion of “long and variable lags” before the impact of changes in monetary policy have an impact…. They could not be more wrong. Markets do the heavy lifting, the signalling, the changed expectations, instantaneously. The rest is history, or rather the inevitable playing out of those expectations in terms of high or low inflation, or rather high or low nominal growth. Of course, expectations can change as central bankers shift their views but often they get stubborn, with disastrous consequences.”

This is an intriguing, utopian view, stated before by Scot Sumner, the chief of the market monetarist tribe.  Unfortunately, there’s just no evidence in favour of it.

Dozens of applied macro economists, perhaps even hundreds, have been working on how to identify the effects of monetary policy changes.   The research came in several waves, as the profession thought more and more deeply about the difficulties of separating out the effects of monetary policy changes from the effects of monetary policy simply responding to other things going on in the economy.

But without exception – at least in the modern and best incarnation of this research – the lags are measured to be ‘long’.  That is to say, there may be some small effect that comes through quickly, but most doesn’t, and the peak effect of monetary policy is measured to come between 1 and 2 years after the change.  A classic survey from 1999 is this paper by Christiano, Eichenbaum and Evans.  But there has been a torrent of work since, reconfirming this basic message.

And those studies that look into their variability – me included – find that they are, though the jury is probably still out on whether that variability is fact, indicative of a shifting economic structure, or artefact.

That’s not to say that the effect of expectations is not important.  On the contrary.  And those that have actually tried to disentangle the effect of expectations changes on the economy find such effects, and, lo and behold, those lags are long too.  [I’m not aware if anyone has looked into whether they are variable, but knowing how such things are done, my guess is that we would also find that they are, with the same qualifications as to what one infers from that].

There are plenty of controversies in this literature.  But that there are substantial lags in monetary policy is not to my knowledge one of them.

This entry was posted in Uncategorized. Bookmark the permalink.

17 Responses to Infinitely short and not variable

  1. Tony
    Actually, I wrote the post on Marcus’ blogsite. I had been posting a few comments here and there as James in London, in support of Scott Sumner and Market Monetarism, often to the annoyance of your mate Simon Wren-Lewis. Since losing my job I now feel freer to comment under my own name, and have a bit more time, too.

    Obviously, the democratic theory of truth holds when it comes to the hundreds of applied macroeconomists, but when they have proved so useless in the recent crisis one wonders what the point of them really is. Thousands of financial market professionals, many former academic or academically-trained economists, who have to earn a living out of putting money on the line know that those applied economists are wrong. Monetary policy impacts all the time. Unexpected changes in data, or policy, change the policy. And change the outcomes. And the market reacts in expectation.

    One day the bulk of applied macroeconomists will realise this. The tragedy is that until they do, their central bank peers will condemn our economies to monetary strangulation – as my good friend Lars Christensen so well called it recently.

    • Tony Yates says:

      Ah, sorry, I missed that you had authored that, I’ll correct my post accordingly. I don’t dispute your description of what your colleagues ‘know’. But those things are still consistent with long and variable lags. As I said, monetary policy may – and indeed in VAR methods is measured to – have some effect straight away, but the majority comes later.

      • James Alexander says:

        Scott Sumner’s ship analogy is a good one. You set the course now, you get to the destination later. The course is, almost, everything. The rest is just history.

  2. Also there’s a simple logical flaw in “monetary offset” – one of Sumner’s favorite ideas. I set out the flaw here:

  3. Also there’s simple logical flaw in “monetary offset” which one of Sumner’s favorite ideas. I set out the flaw here:

  4. Maybe it does have instant effects, but not the kind of effects that we’d like. I have a basic question about the effects of this. Does the expectation of future interest cuts and of future QE actually cause people to move all their money into government bonds instead of away from them, and is this a bad thing? Surely the central bank wants the opposite to happen. If smart people see trouble ahead, surely we want to motivate them to invest in stimulative ways immediately, instead of buying government bonds while waiting to collect their free money from the central bank?

    The more pessimistic the market is, relative to the currently-implemented monetary policy, the more cautious they will be. But we would like to change that?

    (Thanks for any feedback, especially if I’ve got the incentives all wrong!)

  5. mrkemail2 says:

    So would you agree with setting interest rates at zero and just using fiscal policy?

    • Tony Yates says:

      No; it’s costly and cumbersome.

      • mrkemail2 says:

        So what do you recommend?

      • mrkemail2 says:

        “No; it’s costly and cumbersome.”
        This is why I think it is important to have VERY strong auto stabilisers, such as the Job Guarantee that respond in real time.
        Conceivably this could be run by the central bank if you agree with that sort of thing. It is a better fit than “PQE.”

      • Tony Yates says:

        You don’t need anyone to run automatic stabilisers – they are automatic. I’m also strongly in favour, since they are progressive, and the thing they work against – recessions – hit the poor most hard. I am in favour of orchestrated and pre-planned discretionary fiscal stimulus [ie on top of automatic stabilisers] if needed at the zero bound. The total amount of stimulus would be decided by MPC. How it was implemented would be a political decision for government.

    • Tony Yates says:

      Interest rates are at zero because the stimulus has rightly been pushed to the max to counter the recession. I would not agree that setting rates at zero in normal times is a good idea. Ultimately inflation is a monetary policy and not a fiscal policy phenomenon, so you can’t simply switch off monetary policy like that. You could choose a deflation target that, if pursued, would create roughly zero interest rates, but I am not in favour of that either. A bit of inflation is a good thing, for a variety of reasons.

      • mrkemail2 says:

        “I would not agree that setting rates at zero in normal times is a good idea. Ultimately inflation is a monetary policy and not a fiscal policy phenomenon, so you can’t simply switch off monetary policy like that.”
        Tony, I disagree. IMV Monetary policy has been mostly ineffective at controlling inflation. Fiscal policy and commodity prices seem to be the main inflation drivers. In fact, spending via the interest rate channel increases inflation, so the effect is mixed and uncertain. In contrast, fiscal stimulus is more reliable. Monetary policy has just not worked to lead to higher inflation in Japan, for instance.
        Higher interest rates than zero require paying bond interest, which is corporate welfare IMV.
        The other problem is that changing rates leads to instability. If interest rates are set at zero permanently, investors would not need to take into account interest rate moves.
        Other problems:
        1. QE can only work if it stimulates aggregate demand – that is, increase spending. The only way it can do that is via the lowering of longer-term interest rates (as the central bank buys up bonds and drives their yields down). But that relies on the private sector having a taste for risk (borrowing) and it is clear that, even with significant reductions in the investment maturity rates, there is little thirst out there for borrowing.
        Why should there be? The private sector is being confronted with fiscal austerity and a household sector intent on battening down and paying down debt. Firms will only invest in new productive infrastructure if they expect that households (and other firms) will purchase the consumer or capital goods that they produce.
        At present, it is clear that firms have more than enough productive capacity to meet current expected aggregate demand. Result? No big investment boom is coming and so borrowing is mute.
        Further, with more governments scorching the earth with fiscal austerity – mindlessly claiming that by cutting spending you get more – a sort of sick alchemy – households are facing increased unemployment risk again. With record levels of debt hanging over the sector and the risk of joblessness rising in an environment where high levels of unemployment and underemployment persist – why would there be a renewed outbreak of frenzied borrowing?
        2. Building bank reserves – which is all the non-standard liquidity measures have done – will not increase loans. Why? Because banks do not lend out reserves. A lack of reserves is not the constraint. Loans create deposits. The constraint is point 1 – no thirst for borrowing.
        3. Monetary policy is not spending. It might influence spending but only indirectly, with a lag (if at all), and the ultimate impact may well be perverse if the distributional consequences of interest rate cuts lead creditors to spend less by more than the borrowers spend more. Further, monetary policy cannot be regionally or demographically targetted.
        Which tells us that if spending is required then it would be better for the spending arm of policy to be activated – that is, fiscal policy. Government spending is direct, can be targeted (individually spatially, etc), and can be part of an overall package of redistribution (where tax policy might reduce the purchasing power of one income cohort, for example, and public spending might boost the purchasing power of another).
        4. Monetary policy is undemocratic.

  6. JP Koning says:

    What about TIPS spreads? Doesn’t a change in monetary policy have an immediate effect on the TIPS market market?

    • Tony Yates says:

      It has lots of immediate effects, but the effects on goal variables – at least the majority of them – don’t come through until later.

      • Max says:

        Well, if it has *any* immediate effect, then you can always ramp it up and hit the target. The problem is, you will probably overshoot. The “lag” is caused by the insistence on not overshooting.

      • Max says:

        Here’s a better way of putting it: long and variable lags applies to moving from a stable state to another stable state. It doesn’t apply to moving to or from a disequilibrium state.

Leave a Reply

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

You are commenting using your 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