Understand before you tinker

I would be lying if I said this didn’t depress me.  Economists, great economists, economists I idolize stating that there are “imbalances” we must solve – and then telling us how to solve them without actually describing what the imbalance is and why it exists.

Now I welcome the idea of imbalances as an analytical topic, I am over the moon that people are actually studying capital controls, and admitting that we need to go further with micro-foundations.  But I find what the IMF is doing with these policy prescriptions sickening.  Trust me, I’m sure they have an implicit model, a framework they agree with each other on – but unless that is put out for external criticism all such models do is create an illusion of confidence!

At first you might think I’m contradicting myself.  How can I be so happy that we are discussing and studying policy issues, but then get so depressed when people try to put policy into action!  Well the answer is simple, after the large crisis the world has experienced many analysts have jumped on the bandwagon of “doing something” rather than focusing on trying to understand why things are happening.

In my opinion, until we know “why”, until we “understand”, we are pissing in the wind when we introduce policies to “fix” perceived imbalances.  Economists have a method that drives this home, we have to understand, we have to frame a situation before we can make any conclusion – but for some reason when economists are put in a policy situation they are willing to throw that away in order to give people the perception they are doing something.

Let me show an example of this:

Monetary policy has to go beyond inflation stability, adding output and financial stability to the list of targets, and adding macro-prudential measures to the list of instruments.

Now don’t get me wrong, I see a role for macro-prudential regulation – I just think when its introduced the trade-offs need to be explained.  We don’t just do it because we are afraid.

But in this statement the framing is ALL WRONG, monetary policy doesn’t target “stability” in output in itself – it tries not to create undue volatility in output.  This seems pedantic, but the one thing I’ve taken as believable from Real Business Cycle theory is that there are constant supply shocks – negative and positive.  The economic environment is fluid, and this variability is part of how things are.  Trying to smooth this would create undue distortions.

As we can’t view what part of the variability is due to market fundamentals and what part is to do with the vagrancies of demand we DO NOT target output stability – as it would have a negative welfare impact.  Central banks DO smooth the economic cycle as much as is possible with their price level target.

Please, can the disciple stop and recognise it needs to know WHY imbalances happened, identify the market failures, and then introduce institutions and policies that help to improve outcomes given this (and given the constraint of government and policy failure).  If we start adjusting policies without providing explanations we are doomed to create hardship for people.

Update:  This reminder me of one of the main differences you get between micro and macro economists.  Micro is  about understanding why, trying to dig down and reduce everything it essential components to answer “why”.  In macro on the other hand, people start with “empirical facts” and will make some semblance of an effort to explain why – however, it is hard not to get the feeling that sometime macro has its conclusions BEFORE it tries to explain why.

Such a rough and ready breakdown appears to be consistent with the way policy is moving – I find it funny that people are saying that we should increase microfoundations in macro (which I agree with) while ignoring exactly what microfoundations imply – namely understanding.

Anti-Dismal comments here.

  • Talosaga

    There are costs to tinkering (being wrong about the nature of the shock or imbalance) but there are also costs of waiting on the sidelines. It might be that the shock or imbalance is harmful and intervention may have been beneficial. The financial crisis has increased our expectation that imbalances are harmful (ie not supply shocks etc), and we’ve learned that harm imbalances cause can be very great. So if we don’t know what is causing a particular shock or imbalance, we should be more willing to intervene than before. I guess it comes down to how much your expectations have changed post crisis.

    You’re right that we should try to understand why there is a shock or imbalance, but it takes time to build understanding. In the interim we need to have a policy for dealing with ‘unknown’ shocks (intervene or not) which goes back to what we think the shock is and how damaging we think it will be.

  • @Talosaga

    Fair point. However, I’d be hesitant to say that economics haven’t thought about imbalances or shocks in the past – our prior belief to not deal with “imbalances” was in itself built on understanding that was built up over a long time. It seems strange we would throw it all away on the basis of one event, without try to spend a little time thinking about it.

    The key for me is that I don’t think it would take to long to try and get a handle on the general idea of imbalances – if as economists we ask ourselves “what became unbalanced, and what were some of the broad market failures that could have caused this” we would have some basis for policy. Without the failure link we can’t come up with policy – as we don’t have any causal mechanism for it to work through.

    “You’re right that we should try to understand why there is a shock or imbalance, but it takes time to build understanding. In the interim we need to have a policy for dealing with ‘unknown’ shocks (intervene or not) which goes back to what we think the shock is and how damaging we think it will be.”

    I just don’t think we can way we know what an imbalance or shock is without having a context – there are things we can see that have moved a long way from their long-term trends, but without asking why what can we say about them?

    It reminds me of the unit root vs trend-stationary debate that was going on in the 1980s and beyond – we can’t tell with any particular confidence whether things should revert to mean, or where things should be, by JUST looking at data. Without a framework and theory we are blind – and it is very likely policy (especially structural policy) made inside such way will reduce welfare. This pains me.

  • Greg

    I don’t think this is reasonable, Matt. In real life everyone has to operate with incomplete and faulty knowledge. That includes macro policy advisers.

    The “imbalances” were, in fact, listed. As – in the structure of the economy, i.e. the large and fast-growing finance share; – in the balance sheets of economic actors; – in relative asset prices. Deviations from historical ranges/trends should trigger (prompt) scrutiny rather than sloth – as you say, without asking why, what can we say about them? This is Blanchard’s first point.

    Most of Blanchard’s post deals with post-crisis actions. Reiterating what Talosaga said in a more colourful way, Blanchard is looking at things from the POV of the ambulance driver or Emergency Dept doctor, while you are more like the psychiatrist visiting the patient in hospital, after tourniquets and blood transfusions have saved his life. The ambulance driver doesn’t need to know why the patient went through the window in order to know what to do.

    Talk of “throwing it all away” is just silly. Blanchard is advocating adding to the toolkit, not throwing anything away, AND deepening understanding. The “single event” is sufficient to show that current understanding is inadequate.

    There is no other sphere of life in which a single tool is best for all occasions and all places, so it’s unlikely that this would be true for macroeconomic policy. As Blanchard says, there’ll be plenty of work for a decade or two, figuring this out.

    In the mean time, we’ll have to do CPR without understanding hæmatology — or even why CPR works.

  • @Greg

    Hi Greg,

    Agreed that we have to operate under uncertainty – that is part of the burden on action though, as we believe that the market by default will at least show some form of revealed preferences. As a result, we need to have some idea on the market failure before we intervene.

    “The “imbalances” were, in fact, listed. As – in the structure of the economy, i.e. the large and fast-growing finance share; – in the balance sheets of economic actors; – in relative asset prices. Deviations from historical ranges/trends should trigger (prompt) scrutiny rather than sloth – as you say, without asking why, what can we say about them?”

    Indeed, deviations from historic trends DO beg for answers. Which is why we should try to explain the deviation before we can figure out what interventions might be appropriate.

    For example many people said that a major cause of imbalances was exchange rate misalignment, and that this lead to “excessive consumption” in some countries and vulnerability to shocks.

    If this is indeed the case (which is plausible) there are important questions to answer – but even these sorts of issues don’t imply we need to mess around with our clear view of monetary policy. One thing to keep in mind is, what is the right counterfactual. Note that, in the counterfactual case consumption and welfare would be lower pre-crisis, and as a result the total welfare cost of misalignment isn’t clear!

    It is only by explaining WHY that we can get to these idea – we can’t “problem solve” if we don’t have a fair justification from what the problem is.

    “The ambulance driver doesn’t need to know why the patient went through the window in order to know what to do.”

    But the ambulance driver uses the pool of knowledge available to make decisions – he doesn’t come up with new things on the spot.

    “Talk of “throwing it all away” is just silly. Blanchard is advocating adding to the toolkit, not throwing anything away, AND deepening understanding. The “single event” is sufficient to show that current understanding is inadequate.”

    Ahh, but if we turn away from our conception of monetary policy we are throwing something away. Don’t get me wrong, in terms of policy there are a significant number of things I think are appropriate (in an ordinal/directional sense) – but it is the focus, and the lack of background, that creates risks to policy that make me nervous.

    “There is no other sphere of life in which a single tool is best for all occasions and all places, so it’s unlikely that this would be true for macroeconomic policy. As Blanchard says, there’ll be plenty of work for a decade or two, figuring this out.”

    That is his (and Stiglitz) best point – the discipline has limited knowledge and needs to learn and grow.

    But that also implies limits on what the discipline can do – we can’t call on uncertainty to INCREASE how much we intervene in the choices of individuals, and that is still what macro policy does.

  • raf

    “the discipline has limited knowledge and needs to learn and grow.”

    Indeed. I think the image of economics has been “we know everything” and “we have all the answers” when it’s been plain to see that is not the case. It’s always been an evolutionary process and now we are at a point where we need to understand not just the excessive imbalances (credit, currency, trade) but how they eventuate.

    For example, why do currencies not reflect fundamental economic circumstances? Surely freely floating currencies would automatically correct imbalances such as chronic deficits. Apparently not.

    Perhaps the free market is not as efficient as some think.

  • @raf

    Most economists I’ve met do not think they know much – I think it is the typical thing where the people who like to act they know get more attention.

    Part of the reason currencies do not represent fundamentals is because they are an asset price – and asset prices can deviate from fundamentals for sustained periods of time. However, I would note that a chronic current account deficit is not an issue – countries can run them forever. It is just a sign that one country has a different time preference than another on average.

    Free markets aren’t perfectly efficient – but we shouldn’t intervene unless we have some conception of how our intervention can actually improve outcomes (in terms of welfare – not in terms of arbitrary variables).

  • raf

    I guess this is the nub of the ongoing debate. To what extent is the “free market” efficient and to what extent can intervention improve that efficiency. Ultimately where does one draw the line?

    I don’t think we have the answer to that yet but at least it’s something to work on.

  • Talosaga

    @Matt Nolan
    “But that also implies limits on what the discipline can do – we can’t call on uncertainty to INCREASE how much we intervene in the choices of individuals, and that is still what macro policy does.”
    I think my main disagreement with you is on this point. Why does uncertainty lead us to favour less intervention? I think it’s fair to say that the crisis has made economists less certain that ‘imbalances’ will be corrected by the free market. Why would this lead us to favour less intervention? If we are less certain about how much we know about the economy, then we should be less certain about the idea that markets are ‘efficient enough’ not to be regulated or intervened in.

  • @Talosaga

    The less we know about what an imbalance is, the less we can really say about what it means.

    People who make decisions relative to the risk and uncertainty have to face the consequences of their choices, as a result they take all this into account. Policy makers meanwhile do not face the same feedback, they rely on information and understanding in order to develop mechanisms that can improve outcomes.

    If we can explain why a given imbalance arose, and what negative welfare consequences are associated with that, then we can come up with a justification for intervention – however, the burden of proof is on those making the intervention.

    Evidence based policy is an overused term, but using data and theory to describe what we should do is even more important following a crisis – however, we seem more interested in trying to hit the symptoms of issues rather than tackling any underlying causes.

    One thing to keep in mind as well is that we’ve had a big crisis, and a lot of people are worse off as a result – however, under what realistic counterfactual would things have been better? Many of the interventions being discussed now wouldn’t have actually lead to better outcomes – many of the problems were to do with financial sector institutional structure, and exchange rate management that occurred following the Asian financial crisis. Yet we are starting to talk about a whole bunch of macro-prudential regulation which seems only very loosely to do with the issues we’ve faced.

    Now don’t get me wrong, if this regulation is justified by strong theory, put in, and then supported by evidence I will be happy – but we do need to do the first step first.

    I personally don’t think economists should be policy makers – they should sit back and describe, explain, discuss, and try to help understanding. Then people with an understanding of the pressure of policy can use this information and toss up trade-offs. Economists have no experience, or training, in policy – so they tend to respond in a knee-jerk way to crises.

    Of course, I’m exaggerating and making a ridiculous simplification of a large discipline. The problem is that it just rings so true all through the history of political economy :/

  • @raf

    Indeed. My preference really does stem from the ability to react to incentives.

    There are important institutional issues, and I just don’t think policy is focusing on quite the right place given the information and knowledge we have available at present.

    I would also note that economists love to be technocrat and problem solve – so there is an inherent bias to start “organising” the economy to improve outcomes. This is a slippery slope, and one of the reasons our reductionist method focuses on individual choices so much – to illustrate the limits of our own “intelligent design”.

  • @Talosaga

    Also, for a clearer idea of what I think is a well argued and thoughtful take on macro-prudential regulations there is this:

    http://www.economist.com/economics/by-invitation/guest-contributions/countercyclical_regulation_could_be_more_effective_contr

    Looking at the evidence, and coming up for a description of the world where the behaviour of individual actors leads to a problem (akin to what we have experienced) – and then describing what macro-prudential regulation is useful in this case.

  • Talosaga

    @Matt Nolan
    I’ll try make my point in another way: What is the theory supporting ‘unregulated’ or ‘free’ markets? There’s obviously a big hodgepodge of theories supporting this, but a common thread is that markets work best when certain imperfections are not present. These are things like information and agency problems, problems with financial markets etc. People on the side of more intervention believe that the financial crisis has shown that these imperfections have large impacts on the economy.

    Now when deciding policy we have to choose between a weaker “free market” theory and a weak imbalance or intervention theory. In other words, we have no strong theory on which to base policy. Thus we should choose policy somewhere in the middle and try to learn as much as we can from the effects of these policies. I think that’s what the IMF article is trying to get at. We have no solid theory to go off, so we have to work in the middle. Thus saying ‘the onus is on you’ doesn’t make sense if we think free market theories has been weakened.

    Of course it all comes down to how well you think free market theories have survived the crisis. But I think this gives us a better way of viewing the IMF article and other ‘interventionists’. It’s not that they’re quick to intervene, they just think free market theory has been damaged.

    (I think at this point it’s a bit silly to talk about “interventions” generally. Some iterventions, like liquidity requirements, work to make the economy more resistant to shocks. I think the case for these sorts of interventions are strong because they’re a response to “shocks” in general. Other interventions are more targeted at specific imbalances and so the case for them is weaker.)

  • @Talosaga

    I see what you are saying about the idea that theory should be what guides us away from a “default” position. But the key thing is that I would term the default as where we are right now. The very action of changing policy presupposes a justification for doing so.

    I agree with what you have to say on imperfect information and agency problems – and insofar as we can identify and understand them there is the potential to improve outcomes. But some of this targeting seems like merely introducing a new policy because things seem to be unpleasant.

    “Now when deciding policy we have to choose between a weaker “free market” theory and a weak imbalance or intervention theory.”

    We have a single methodology, one which provided us a bunch of results – however, we have reservations about some of them. When we can articulate what ways theory has been impacted upon we can change policy – otherwise we cannot.

    There are always many many theories, and if a set of theories were being used to justify action that is fine. This isn’t a case of that, this is a case of ignoring theories that were popular to give the perception of taking action.

    See this single economic framework exists, and can be used to discuss WHY we need to do things transparently – their refusal to “indulge” in it for describing some policies is abjectly poor for an economic organisation.

    On the note of theory, I am hesitant to call anything “free market theory” – this doesn’t make any sense to me. Economics is the study of scarcity and incentives, the fact that people make choices within the constraint of scarce resources. Our perceptions of some market failures may have changed – but in that case policy could be justified by re-evaluating the parameters we placed on these variables, and then this would suggest different outcomes.

    THIS action would provide understanding – but there are interventions being discussed that have nothing to do with the crisis or anything we learned from it. This is opportunism that is being cloaked by their unwillingness to add perspective.

    “Some iterventions, like liquidity requirements, work to make the economy more resistant to shocks. I think the case for these sorts of interventions are strong because they’re a response to “shocks” in general.”

    There is always a cost though – if we think people choose liquidity requirements that are too low we need to ask why. Once we know about the trade-off, and why the choice of individuals and firms may be subject to some type of market failure, then we can come up with sensible policy. I think issues like “too big to fail” create a definite issue here that justifies liquidity requirements – so there we are, we have a reason for the policy.

    Again, my critique isn’t on policies that have a justification – it is when the IMF says we need to target output variability. This involves ignoring piles of previous economic and econometric evidence and theory – there is no reason for what they are saying, it is irrelevant in terms of what happened during the GFC, but they decided to sell it under the guise of “new policy”.

    I am not saying don’t do anything persee, I’m saying could you please have a justification for what you are doing before you start fiddling – and could you make it transparent so that people can look at it, discuss it, and offer feedback.

  • Talosaga

    @Matt Nolan
    “On the note of theory, I am hesitant to call anything “free market theory” – this doesn’t make any sense to me.”
    You’re right. I’m using it just as a short hand for something like “theories or models that suggest we should not intervene in markets”.

    “I agree with what you have to say on imperfect information and agency problems – and insofar as we can identify and understand them there is the potential to improve outcomes. But some of this targeting seems like merely introducing a new policy because things seem to be unpleasant.”
    I see where your coming now.