Discord about econ discourse

Danny Blanchflower remarked on Twitter that UK economic discourse lacked much engagement from academics. This was a follow-up to a Paul Krugman post where he disparaged UK economic discourse, and used as his metric the fact that there was a lot of focus on the deficit.  And that in turn derives from Simon Wren Lewis’ developing caricature of UK’s ‘mediamacro’.

Well, this blog is about why – supposing the assessment of our discourse to be true, which I don’t really accept in the way it was put – you shouldn’t blame the academics. Or perhaps anyone, except the market.

The first reason why is that there is little or no financial incentive to take part.

UK economics departments are partly assessed on ‘impact’. But there is almost no link to that and me, for example, piling into a debate about whether inflation 2pp below the BoE’s target is, as George Osborne’s office seemed to be claiming, was a good thing.

The ‘impact’ criterion is, mostly, about the impact of a particular piece of research [produced while at the institution being evaluated].

I could, for example, write a new paper examining the costs of inflation target misses, and succeed in persuading either the BoE or HMT to take notice, cite my work. My department might then choose to use me as one of a small number of impact ‘case studies’. Or they might not.

Notably, me piling into a conversation over Twitter on the train to Bristol, synthesising what most macroeconomists would already believe into a few tweets and persuading those who listened of the correctness of my case, would not carry any weight whatsoever.

You might well say ‘Good!’ Do something more useful on the train! Write that new paper on the costs of target misses! But, actually, that wasn’t necessary to point out the dangers of bragging about too-low inflation. We don’t need a new paper to do that.

What dominates the financial calculus for those on either side of the academic hire is publications.

Some small weight would be put on other aspects of an academic’s profile, but not much. I’m soon off to Birmingham. They might have been happy to see that there is someone with experience of how the UK monetary and fiscal regime works. But I am sure they were much more vexed about how their 2020 Research Excellence Framework results would come out, and for that they need publications. And trying to predict what a bunch of impact case studies might look like, and who would be used to make them, is much less certain than adding up the number of 4-star publications now, or forecasting those in the pipeline.

A second point to make is that engagement often isn’t wanted, really.

Engagement sometimes presents itself as criticism of those in the media who are currently paid to try to keep eyeballs focused on them. I don’t expect any pieces reporting ‘Selflessly engaged academic Tony Yates explained how all our pieces extolling the bottom-up theory of inflation fallacy were, in fact, wrong, for which we are eternally grateful. How much clearer the world looks to us now the fog has been lifted.’ Silence is much more likely, even on uncontentious points, and understandable.

Those controlling access know what their consumers want better than people like me. They probably don’t want mini-literature surveys that are faithful to the small economic print and convey all the shades of opinion and controversies about some issue. Consumers don’t have time or inclination for that.  They want a way into a topic.  And that way in might be an opinion.  A crude version of one of the many possible arguments.  Something that people like me don’t offer.

A great example is Robert Peston’s old blog on QE. With my academic/central bankers macro hat on, this was full of stuff that was wrong and confusing, and I explained why.  But in retrospect, what was the point?  Those who were likely to read my response didn’t need persuading.  And those that did need persuading were, thanks to Robert’s wise guardianship of his own access point, not going to hear about it, and probably would not want to anyway.

Peston’s blog would be an unreadable mess if it was full of stuff responding to pedantic academics no-one had heard of. And the point of some of these journalism brands is that they are themselves the founts of their chosen specialist wisdom, not that they humbly intermediate from academics who know better. You can see the business case for keeping academic minnows out of ‘engagement’ because it changes the brand.

I exaggerate the extent and scope of these problems – if they are problems – to make a point.  There are lots of examples of successful conversations emerging too.  Although I think most of those are between the discourse producers, and not about academic stuff being put in front of the ultimate consumers (the readers/watchers).

I also don’t want to claim that this is a case of the media selling a dud and concealing from the information-hungry public the holy truths that noble academics are dying to tell them.

I already conceded we were selfish, and looked to our incentives.

But I should also concede that we often don’t know the answer either. Many of us are drawn into specialisms that would look frighteningly narrow to an outsider (in pursuit of publications). And we are sometimes therefore just as bad at digesting literatures outside that specialism as your average journalist. (Extreme but common example being the number of Nobel laureates recruited to talk off topics they excelled at, and getting it wrong).

Even when we do know the answer, we don’t have the skills to put it into words intelligible to even a well-meaning journalist quickly enough to be useful. The few conversations I’ve had have often conjured up my fond memory of former BoE Deputy Governor John Gieve asking ‘ah, so what does this answer depend on this time, Tony?’  In fact, I’d say I was treated rather charitably, given my tendency to insist on the ‘on the one hand, on the other’ format.

So, in short, engagement by academics is unpaid, is sometimes experienced as unwelcome [either because it undermines the business model of the expert journalist, or because consumers wouldn’t want it] and we are often not very good at it anyway.

Advertisements

Eurozone QE: features and bugs

Daniela Gabor at UWE tweeted an interesting article on Reuters describing troubles in money markets due to a ‘shortage’ of high quality sovereign bonds in the Eurozone.  The proximate cause is the launch of the ECB’s quantitative easing program.  The cause of that being the persistent below target inflation and depressed real activity.  And the cause of that being a GLUT of sovereign bonds.  A shortage ultimately caused by a glut?!  This it not necessarily a contradiction,  since the glut would correspond to lower quality bonds issued in desperation.

A few observations.

1.  Depending on how you view the transmission of QE, it’s partly a feature and not a bug to create such shortages, so that market participants switch to using private sector assets as a substitute, bidding up the price of those assets, lowering the cost of funding for and stimulating spending by the issuer.

2.  But for activities involving private sector market players only, there could be painful, perhaps insurmountable obstacles to switching from an equilibrium in which everyone uses sovereigns as collateral for everything, to one in which private sector assets do the job.  There have been glimmers of this before, in econometric evidence showing that QE lowers government yields, but leaves private yields less affected.

4.  For transactions involving the central banks themselves, the obvious solution is to combine QE with a relaxation of collateral requirements to embrace riskier private sector assets on less disadvantageous terms.  The ECB was anyway inclined towards credit easing [taking private sector assets onto its balance sheet].  Here’s an excuse to do more of it.

5.  Oh, one might wish, for the existence of a combined Treasury/Debt Management Office in Euroland, who would issue new debt into the market, spending the proceeds in the South, where aggregate demand more obviously falls short of supply.

6.  Yet the benefits of 4 and 5 depend, to repeat, on how you view QE working.  If QE is predominantly about signalling lower future central bank policy rates, then these correctives to QE might not undermine that signal.  On the other hand, if QE is about changing the mix of public and private sector assets in the hands of private investors, then corrective policies have to be careful not to correct this aggregate!  They would have to be about redistributing between private participants;  or ensuring that supply was even across sovereign maturities and bond types.

Price level shocks and bottom up theories of inflation

I think the ‘price level shock’ fallacy is related to another fallacy that we can divine the causes and future of inflation by adding up individual price series.

The price/level/inflation shock categorizer like Ian McCafferty thinks to himself: ‘oil prices won’t keep going up and up.  This a one-off.  Inflation is caused by the sum of the individual sectoral inflation rates.  So when this one-off oil price change is done, the oil sector will be back to contributing zero to the inflation rate.  Since we are inflation targeters, we can ignore the oil price rise.’

But this is all wrong.  What’s happened is that there has been a change in lots of relative prices [brought about by a complex mix of demand and supply factors].

Notably, there has been a change in the price of oil as a final good relative to the other final goods;  and a change in the price of oil relative to other inputs;  and a change in the price of labour relative to what a given bundle of labour can produce.

When prices in the oil/petrol sector fall, we can’t simply project forwards the other sectoral inflation rates as before to work out what will happen.

What happens to these other sectors depends on what the central bank chooses to do with its policy instrument.  If it wanted, it could make sure that the sum of these sectoral inflation indices added up to zero.  Or whatever.  And what it was appropriate to do depends on a whole host of things, as I went through in the previous post.

This inflation is caused by the sum of its parts problem rears its head every time new inflation data gets released.  Where we can read that inflation was ’caused’ by the prices that went up, and inhibited by the prices that went down.  A classic example of this being this piece in the Guardian.  (Though this problem is ubiquitous).

Sometimes this is harmless semantics.  Often it can badly mislead.  Next time you read ‘inflation soars as clothing retailers jack up prices’ pause a moment to say a prayer that the writer has not fallen foul of the bottom-up theory of inflation fallacy.

Bristol/Birmingham

I’ve accepted a post as Prof of Economics at Birmingham, where I’ll join in the Summer.

Carney and McCafferty on oil and monetary policy

Carney and McCafferty delivered a one-two on monetary policy and oil prices.  Both get it wrong in my opinion, but, in diffent ways.  [Remember Tolstoy’s Ana Karenina:  ‘all good economic arguments are alike.  Each bad one is bad in a different way…’]

McCafferty first.

The basic approach he takes is:  we need to decide what caused the oil price to change, because that determines how we respond to it.  That much he has right.  But there are two big problems with how he sets about deciding [and, I infer, in his vote, which we must presume is based on this logic].

First.  What caused oil prices to change.  He rightly sets out on a hunt for oil demand and oil supply shocks.  But in my view he wrongly thinks you can diagnose the oil supply schedule from the quantity of oil produced by the oil-producing countries.  Identification is harder work than this!

I’d suggest three ways of doing it, that parallel those in the academic literature on, for example, identifying shocks to monetary or fiscal policy.  1:  study the motives of those controlling oil production and what they say about what they are doing.  2:  estimate a Vector Autoregression and identify oil supply and demand shocks from the correlations between oil output and oil prices [and other things, perhaps].  3.  Formulate a structural model of the world economy split into oil-producing and oil consuming sectors, and recover the shocks by estimating this model directly.

Second, McCafferty engages in the ‘price level shock fallacy’.  As I have ranted before there is no such thing as a price level shock in the sense meant here.

The logic in McCafferty’s mind goes:  oil causes the price level to fall, and the inflation rate to fall only temporarily.  Therefore there is no need to respond.

However, in fact, ultimately, it’s up to the central bank to decide whether or not even the price level falls in the long run.  Anecdote:  in the simplest versions of the BoE’s New Keynesian model, if inflation targeting is done optimally, [under commitment], NOTHING affects the price level permanently.  So in that world McCafferty’s category of ‘price level shocks’ would be an empty set.  A hypothetical MPC colleague inside that model would retort reading Ian’s speech: ‘oil only affects the price level so we don’t need to respond?  what are you talking about?  nothing affects the price level because we respond in just the right way.’

A related assertion to the one IM makes in his speech might be:  ‘given our planned policy rule, if we don’t deviate from it, that will, by choice, generate a lower price level in the long run than if the oil shock hadn’t happened, but the inflation rate will be no different, so we won’t deviate from the rule.’

But there’s nothing axiomatic about whether this will or won’t be true.  Whether it is  depends on how large the shock is, how activist the planned policy rule is in responding to variables in the economy, and the process driving expectations.  And the proximity to the zero bound.  And the efficacy of alternatives to interest rate policy.  And…

You can’t deduce the optimal response from categorising something as a ‘price level shock’.  That description includes a policy response already [the decision to make sure there is no effect on the inflation rate] .  You have to take a stance on all these other things.

Note too that ‘not responding’ to oil would not mean ‘leaving interest rates unchanged’.  It would not even mean ‘not altering the previously agreed trajectory for rates’. It would mean ‘not altering the previously determined recipe for arriving at the interest rate trajectory’.

But McCafferty has not explained what the [or his] recipe is, or what trajectory it gives you.

In fact, the whole problem here, [to repeat] is that he’s trying to decide something about the recipe [whether or not to respond to a certain kind of oil price movement] based on already presuming a recipe [which leads to that shock changing the price level].

So, away with those ‘price level shocks’.  The term should set of a mental hooter that goes off when you see it.

Onto Carney’s speech.

As was widely reported today, he said that it would be “foolish” to loosen in response to the fall in oil prices, which had led to the chunky drop in headline CPI inflation.

Well, I don’t think it would be foolish.  The motivation for cutting is twofold.

First, a case could be mounted that the oil price fall boosts potential output relative to demand.  In the Bank of England’s own model this puts downward pressure on deflation and would warrant a cut.

Second, to the extent that there is a risk of inflation expectations responding to the fall in headline CPI inflation that we have seen, and of the economy becoming more permanently entrenched at the zero bound, a cut would be warranted on precautionary grounds.  On that logic one would not wait until wage growth or inflation expectations fell.  By that time it might be too late.  So Carney’s qualification that loosening would be foolish “unless” wage growth weakened does not cover this concern.  As Carney stresses in his own text, monetary policy takes time to work.  Hence one has to act on the basis of a probabalistic forecast, always.

 

The Bond of trust Yielding from central bank transparency

A tired pun, in service of a continued campaign on this blog for increased transparency in central banking.  Or rather at the Bank of England in particular.

Previous rants have focused on the fact that the BoE does not yet provide the code, data and forecast judgements it applies to the main model it uses for monetary policy, when it should, although it has been asked to think about it;  on the fact that transcripts of MPC meetings were, up until now, erased;  and on numerous transparency problems with the forecast and monetary policy-setting.

This post is to point out that the Bank could and should but doesn’t yet provide the codes and adjustments and data it uses in order to estimate its yields curves.

These curves are probably and unsurprisingly, the most pored-over charts inside the Bank of England, and outside of it by BoE watchers.  Government bonds are bought and sold every day at prices that are discounts on the face value.  They are bonds of irregular and different maturities, trading in differing and fluctuating volumes.  The yield curve estimation industry is the practice of drawing bendy lines through these irregularly appearing bond prices [leave aside whether there is, truly a single price] and inferring the interest rate on those and intermediate maturity bonds [ie where none were traded].  And, using the same bendy line drawing techniques applied to index-linked bonds, to infer the ‘term structure’ [the trajectory, or sequence] of expected inflation rates.  If you aren’t familiar with central bank business, it ought still to be clear now why this is important;  it’s about deriving what the market thinks is the expected path of its policy instrument and one of its goal variables.

The strategy for drawing these bendy lines through bond prices is set out in Bank of England working papers that are, therefore, published.  But it’s still an awful lot of work to get from that to a set of codes that would replicate the charts MPC would be looking at.  Not least because there are routines for excluding misbehaving bond prices all the time, which observers would find very difficult to replicate.

There are a few arguments for releasing these tools.

Most compellingly, these codes were paid for by the public, with taxpayers money, ultimately, so – using the same logic as post I wrote about the BoE forecasting model – why should they be kept private?

Another point is that many BoE watching conversations will run:  ok, what do they think we think they are going to do in the future?  And what will they do about that?  Validate what we think they are going to do?  Or put us right?  Confusingly, policymakers want to know about these conversations too!  And their responses can be refined if the whole process of thought extraction can be made accurate, common to all, and cheap.

Moreover, by releasing the codes, the BoE could potentially tap into free expertise in the community of code-literate BoE watchers [yes, a non-empty set] and refine what it does over time.

I surmise that – as was the case for the forecasting model – the BoE are worried about becoming a bond-yield-estimation-code-support-facility.  But I don’t see why it could not successfully say, simply: ‘here it is, here is the data, off you go’.  Keeping the code and the adjustment routines to itself looks, to those inclined to suspect its motives, like the BoE is trying to dodge scrutiny, and claim property rights it does not rightly possess.

Many top journals require researchers to deposit working codes and data for replication purposes as a matter of course.  Why should our central bank not conform to this norm?

This isn’t a huge deal.  But it is part of an addictive privacy habit on the part of central banks that needs to be kicked.  The BoE has been travelling on a journey – like other central banks – of ever-increasing transparency.  This would be a small positive step that would fit in with this mission.

QE exit [here we go again]

One memory of my final years at the Bank of England is that every two or three months we were commissioned new pieces of work on QE exit strategies, each time prompting ‘here we go again’ thoughts as the staff machine dutifully ground into action, only for events to overtake the writing, making exit an ever more distant event.

Writing this quick blog, prompted by questions from a journalist, risks jarring with the Zeitgeist in a similar way.

But it’s conceivable that the issue surfaces again with the heating up of the debate about when the Fed might tighten, itself fuelled by the increasing evidence that the recovery in the US, at least, is secure. That debate has not really questioned when the Fed will exit from quantitative easing and sell its stock of government and agency debt. Rightly so. This is an issue for later, for a number of reasons.

The Fed will be anxious about active selling after the ‘taper tantrum’. Sovereign debt markets, although clearing at historically high prices (low yields) may (may) be vulnerable to a sudden correction, volatility perhaps heightened by the ongoing negotiations in Europe between the Eurogroup and Greece.

The Fed will want reverse-QE asset sales to be part of an orderly, ongoing plan, to minimise the risk of an unwanted spike in yields, and to ease coordination with the issuance plans of the US Treasury. To maximise the chance that this is possible, it is best to wait until there is no chance that QE would have to be reversed. That means waiting until interest rates have lifted clear enough of the zero bound that this conventional instrument could bear the burden of loosening again, if that became necessary.

Another reason why the Fed can take its time is that the fears of the inflationistas have not been born out. There are two groups in this camp.

First, there are the inflation hawks that were so heavily and rightly criticised by Krugman and others. These individuals seemed to reject entirely the economics of standard monetary models at the zero bound. Their view seemed to get little traction in policy circles directly, but it may have been part of the calculation that ultimately limited the scale of asset purchases.

The second camp of inflationistas were the more sensible observers who might have placed a small probability on the worry [or is it the hope] that large balance sheet expansion might have sparked runaway inflation. This was a concern for some on the FOMC and the UK’s MPC. Such a worry would have stressed the incompleteness of our monetary macro models, and, even within them, the possibility of expectations-induced inflation scares. At any rate, as each month passed without such a scare, such a sensible observer would have been steadily revising down the probability that there would be one.

Given that, there need be no undignified rush to get shot of these purchases from the central bank balance sheet.

Conspiracists might worry that fiscal calculations would prevail to lead to exactly that, following the logic that central bankers will want to sell while the price is high. I’m more inclined to take policymakers’ word for it. And that has stressed the primacy of ensuring that sales are consistent with the monetary policy goals of the central bank.

And note that in the UK, the BoE never really took QE purchases onto its own balance sheet. This, at the insistence of former Governor Mervyn King, was indemnified from fluctuations caused by that of the Asset Purchase Facility, the Special Purpose Vehicle established to hold the assets.

A more plausible line of thought is that the FOMC might think itself pressured to scale back its balance sheet to placate the #AudittheFed movement led by Ron and Rand Paul, and supported in spirit by John Taylor. That movement – and Taylor in particular – sees very low interest rates and unconventional monetary policy as damaging and discretionary.  Some seem to see these policies – perplexingly – as an intrusion into the workings of a capitalist economy.

However, my guess is that to the extent that concerns about this pressure enter the calculation at all, scaling back prematurely for these reasons would not be the outcome.

First, it might be reasonable to conclude – especially since US political discourse has branded Yellen an overtly political appointment – that there is nothing that could be done to head off this anti-Fed movement.

Second, it would also be reasonable to think that the Fed would think that executing sales as best it could for monetary and financial stability purposes would be the thing most likely to drain support from the antiFed campaign.