“Money neutrality” – normative rather than positive

When we study macroeconomic theory we are that we are taught about “money neutrality”. Normally money neutrality is seen as a certain feature of a given model. In traditional monetarist models monetary policy is said to be neutral in the long run, but not in the short run, while in Real Business Cycle (RBC) models money is (normally) said to be neutral in both the long and the short run. In that sense “money neutrality” can be said to be a positive (rather than as normative) concept, which mostly is dependent on the assumptions in the models about degree of price and wage rigidity.

As a positive concept money is said to be neutral when changes in the money supply only impacts nominal variables such as prices, nominal GDP, wages and the exchange rates, but has no real variables such are real GDP and employment. However, I would suggest a different interpretation of money neutrality and that is as a normative concept.

Monetary policy should ensure money neutrality

Normally the discussion of money neutrality completely disregard the model assumptions about the monetary policy rule. However, in my view the assumption about the monetary policy rule is crucial to whether money is neutral or not.

Hayek already discussed this in classic book on business cycle theory Prices and Production in 1931 – I quote here from Greg Ransom’s excellent blog “Taking Hayek Serious”:

“In order to preserve, in a money economy, the tendencies towards a stage of equilibrium which are described by general economic theory, it would be necessary to secure the existence of all the conditions, which the theory of neutral money has to establish. It is however very probable that this is practically impossible. It will be necessary to take into account the fact that the existence of a generally used medium of exchange will always lead to the existence of long-term contracts in terms of this medium of exchange, which will have been concluded in the expectation of a certain future price level. It may further be necessary to take into account the fact that many other prices possess a considerable degree of rigidity and will be particularly difficult to reduce. All these ” frictions” which obstruct the smooth adaptation of the price system to changed conditions, which would be necessary if the money supply were to be kept neutral, are of course of the greatest importance for all practical problems of monetary policy. And it may be necessary to seek for a compromise between two aims which can be realized only alternatively: the greatest possible realization of the forces working toward a state of equilibrium, and the avoidance of excessive frictional resistance.  But it is important to realize fully that in this case the elimination of the active influence of money [on all relative prices, the time structure of production, and the relations between production, consumption, savings and investment], has ceased to be the only, or even a fully realizable, purpose of monetary policy. ”

The true relationship between the theoretical concept of neutral money, and the practical ideal of monetary policy is, therefore, that the former provides one criterion for judging the latter; the degree to which a concrete system approaches the condition of neutrality is one and perhaps the most important, but not the only criterion by which one has to judge the appropriateness of a given course of policy. It is quite conceivable that a distortion of relative prices and a misdirection of production by monetary influences could only be avoided if, firstly, the total money stream remained constant, and secondly, all prices were completely flexible, and, thirdly, all long term contracts were based on a correct anticipation of future price movements. This would mean that, if the second and third conditions are not given, the ideal could not be realized by any kind of monetary policy.”

Hence according to Hayek monetary policy should ensure monetary neutrality, which is “a stage of equilibrium which are described by general economic theory”. In Prices and Production Hayek describes this in terms of a Walrasian general equilibrium. Therefore, the monetary policy should not distort relative prices and hence monetary policy should be conducted in a way to ensure that relative prices are as close as possible to what they would have been in a world with no money and no frictions – the Walrasian economy.

As I have discussed in a numerous posts before such a policy is NGDP level targeting. See for example herehere and here.

And this is why the RBC model worked fine during the Great Moderation

If we instead think of monetary policy as a normative concept then it so much more obvious why monetary policy suddenly has become so central in all macroeconomic discussions the last four years and why it did not seem to play any role during the Great Moderation.

Hence, during the Great Moderation US monetary policy was conducted as if the Federal Reserve had an NGDP level targeting. That – broadly speaking – ensured money neutrality and as a consequence the US economy resembled the Walrasian ideal. In this world real GDP would more or less move up and down with productivity shocks and other supply shocks and the prices level would move inversely to these shocks. This pretty much is the Real Business Cycle model. This model is a very useful model when the central bank gets it right, but the when the central bank fails the RBC is pretty useless.

Since 2008 monetary policy has no longer followed ensured nominal stability and as a result we have moved away from the Walrasian ideal and today the US economy therefore better can be described as something that resembles a traditional monetarist model, where money is no longer neutral. What have changed is not the structures of the US economy or the degree of rigidities in the US product and labour markets, but rather the fed’s conduct of monetary policy.

This also illustrates why the causality seemed to be running from prices and NGDP to money during the Great Moderation, but now the causality seem to have become (traditional) monetarist again and money supply data once again seems to be an useful indicator of future changes in NGDP and prices.

Maybe Scott should talk about Hayek instead of EMH

Every other month or so Scott Sumner writes a defence of the so-called Efficient Market Hypothesis. I have noticed that the commentators already react quite aggressively to Scott’s unwavering support of EMH and my own personal experience is that people – especially people who themselves are active in the financial markets – will strongly oppose the idea of efficient markets.

Why is that? Fundamentally I think that most people think of economics and financial markets based on their personal micro observations rather than based on economic logic. We all have met or heard of completely insane investors that far from can be described as being rational and if such people exists how can we talk about efficient markets? EMH is simply a very hard sell – whether or not it empirically is a good description of the world.

Even though Scott and I agree that EMH probably is the best description of today’s financial markets it might actually be an idea to stop talking about EHM and instead confront the anti-EHM crowd with an alternative. Such an alternative could be Hayek’s description of capitalist market system as the best aggregator of information known to man.

Recently I have been re-reading some of the key chapters in Hayek’s Individualism and Economic Order. The book is full is Hayekian classics such as The Use of Knowledge in Society and The Meaning of Competition and as well as some key chapters discussing calculation in socialist society. Contrary to neo-classical theory which is at the core of EMH Hayekian thinking is based on much less rigid assumptions (and is somewhat less stringent). At the core of Hayek’s thinking is that no social planner has the knowledge to allocate goods and resources in society. Preferences are individual and are changing constantly and so do natural conditions.

We do so to speak not know the “model” of the economy. Contrary to this EHM and rational expectations build on the explicit assumption that the model is known. Hayek on the other hand would strongly object to the notion that the model is known – least at all by policy makers.

Scott uses EHM to argue that since the markets are more or less efficient no central bank forecast will be able to consistently beat the forecast of the market and therefore central bank policy implementation should be build on the use of the market mechanism through NGDP futures.

This makes perfectly good sense, but hold on for a second. What if the “model” of the economy is known why should we bother using NGDP futures when a benevolent central bank could just solve an optimisation problem and solve the model and implement the policy that would ensure the highest level of societal welfare? First of all, there would be a problem which social welfare function to optimize, but lets assume that this (non-trivial!) problem is solved. Then second, would you really think that we could and should leave it to central bankers to define what model is the “right” of the economy (most central banks today rely on New Keynesian models which both Scott and I think make very little sense). Take Scott and me. 90% (I changed that from 99%) of our thinking about monetary issues is the same, but I could come up with a few areas where we do not agree on the theoretical issues and even if we agreed about the model of the economy we could still disagree about the empirical size of the parameters in the model.

This is of course why it is much better to leave it to the market to decide on the implementation of monetary policy through the direct and indirect use of prediction markets (such as macroeconomic forecasting, the implementation process and NGDP futures).

But what if I was in a room full of non-economists and I had to explain why this makes sense. Would I start by outlining a mathematical model and tell them that markets where efficient (most people have no clue what efficient mean – neither do most economists) or would I tell a Hayekian story about how central planning is impossible and markets is the best aggregator of information? I surely would go with the Hayekian story. It is simply much more acceptable to most people than the EHM and radex story – even though Scott and I full well know that it is basically the same story.

That said, it is therefore interesting that it is especially Scott’s Austrian oriented readers who so strongly object to Scott’s insistence that markets are efficient. They should really read Hayek because Hayek is exactly saying that the markets are significantly more efficient than any other form of allocation mechanisms. Yes, he is also saying some – in my view – weird things about mathematics, but overall Hayek thought that we could describe the economy as being efficient and that rational expectations would be a good approximation of the real world. Hayek’s classic description in “Price and Production” from 1933 of his business cycle theory is in fact very much an attempt (which fails in my view) to describe the business cycles within a fundamentally neo-classical set-up.

Scott’s conclusion is to “let a thousand models blossom” so instead of trying to figure out what really is the right model of the economy central bankers should use market information in the conduct and implementation of monetary policy. Would Hayek disagree? I think not…

NGDP targeting would have prevented the Asian crisis

I have written a bit about boom, bust and bubbles recently. Not because I think we are heading for a new bubble – I think we are far from that – but because I am trying to explain why bubbles emerge and what role monetary policy plays in these bubbles. Furthermore, I have tried to demonstrate that my decomposition of inflation between supply inflation and demand inflation based on an Quasi-Real Price Index is useful in spotting bubbles and as a guide for monetary policy.

For the fun of it I have tried to look at what role “relative inflation” played in the run up to the Asian crisis in 1997. We can define “relative inflation” as situation where headline inflation is kept down by a positive supply shock (supply deflation), which “allow” the monetary authorities to pursue a easy monetary policies that spurs demand inflation.

Thailand was the first country to be hit by the crisis in 1997 where the country was forced to give up it’s fixed exchange rate policy. As the graph below shows the risks of boom-bust would have been clearly visible if one had observed the relative inflation in Thailand in the years just prior to the crisis.

When Prem Tinsulanonda became Thai Prime Minister in 1980 he started to implement economic reforms and most importantly he opened the Thai economy to trade and investments. That undoubtedly had a positive effect on the supply side of the Thai economy. This is quite visible in the decomposition of the inflation. From around 1987 to 1995 Thailand experience very significant supply deflation. Hence, if the Thai central bank had pursued a nominal income target or a Selgin style productivity norm then inflation would have been significantly lower than was the case. Thailand, however, had a fixed exchange rate policy and that meant that the supply deflation was “counteracted” by a significant increase in demand inflation in the 10 years prior to the crisis in 1997.

In my view this overly loose monetary policy was at the core of the Thai boom, but why did investors not react to the strongly inflationary pressures earlier? As I have argued earlier loose monetary policy on its own is probably not enough to create bubbles and other factors need to be in play as well – most notably the moral hazard.

Few people remember it today, but the Thai devaluation in 1997 was not completely unexpected. In fact in the years ahead of the ’97-devaluation there had been considerably worries expressed by international investors about the bubble signs in the Thai economy. However, the majority of investors decided – rightly or wrongly – ignore or downplay these risks and that might be due to moral hazard. Robert Hetzel has suggested that the US bailout of Mexico after the so-called Tequila crisis of 1994 might have convinced investors that the US and the IMF would come to the rescue of key US allies if they where to get into economic troubles. Thailand then and now undoubtedly is a key US ally in South East Asia.

What comes after the bust?

After boom comes bust it is said, but does that also mean that a country that have experience a bubble will have to go through years of misery as a result of this? I am certainly not an Austrian in that regard. Rather in my view there is a natural adjustment when a bubble bursts, as was the case in Thailand in 1997. However, if the central bank allow monetary conditions to be tightened as the crisis plays out that will undoubtedly worsen the crisis and lead to a forced and unnecessarily debt-deflation – what Hayek called a secondary deflation. In the case of Thailand the fixed exchange rate regime was given up and that eventually lead to a loosening of monetary conditions that pulled the

NGDP targeting reduces the risk of bubbles and ensures a more swift recovery

One thing is how to react to the bubble bursting – another thing is, however, to avoid the bubble in the first place. Market Monetarists in favour NGDP level targeting and at the moment Market Monetarists are often seen to be in favour of easier monetary policy (at least for the US and the euro zone). However, what would have happened if Thailand had had a NGDP level-targeting regime in place when the bubble started to get out of hand in 1988 instead of the fixed exchange rate regime?

The graph below illustrates this. I have assumed that the Thailand central bank had targeted a NGDP growth path level of 10% (5% inflation + 5% RGDP growth). This was more or less the NGDP growth in from 1980 to 1987. The graph shows that the actually NGDP level increased well above the “target” in 1988-1989. Under a NGDP target rule the Thai central bank would have tightened monetary policy significantly in 1988, but given the fixed exchange rate policy the central bank did not curb the “automatic” monetary easing that followed from the combination of the pegged exchange rate policy and the positive supply shocks.

The graph also show that had the NGDP target been in place when the crisis hit then NGDP would have been allowed to drop more or less in line with what we actually saw. Since 2001-2 Thai NGDP has been more or less back to the pre-crisis NGDP trend. In that sense one can say that the Thai monetary policy response to the crisis was better than was the case in the US and the euro zone after 2008 – NGDP never dropped below the pre-boom trend. That said, the bubble had been rather extreme with the NGDP level rising to more than 40% above the assumed “target” in 1996 and as a result the “necessary” NGDP was very large. That said, the NGDP “gap” would never have become this large if there had been a NGDP target in place to begin with.

My conclusion is that NGDP targeting is not a policy only for crisis, but it is certainly also a policy that significantly reduces the risk of bubbles. So when some argue that NGDP targeting increases the risks of bubble the answer from Market Monetarists must be that we likely would not have seen a Thai boom-bust if the Thai central bank had had NGDP target in the 1990s.

No balance sheet recession in Thailand – despite a massive bubble

It is often being argued that the global economy is heading for a “New Normal” – a period of low trend-growth – caused by a “balance sheet” recession as the world goes through a necessary deleveraging. I am very sceptical about this and have commented on it before and I think that Thai experience shows pretty clearly that we a long-term balance sheet recession will have to follow after a bubble comes to an end. Hence, even though we saw significant demand deflation in Thailand after the bubble busted NGDP never fell below the pre-boom NGDP trend. This is pretty remarkable when the situation is compared to what we saw in Europe and the US in 2008-9 where NGDP was allowed to drop well below the early trend and in that regard it should be noted that Thai boom was far more extreme that was the case in the US or Europe for that matter.

Hayekian capital theory – the math geek version

When I wrote my master thesis many years ago the topic was a mathematical formalization of Austrian Business Cycle Theory. In hindsight I think it is incredible that I able to pull it off and I am still pretty happy with that master thesis. It, however, convinced me that Hayek’s version of Austrian Business Cycle theory was seriously flawed. Furthermore, the math in my modeling never really satisfied me. It was just not good enough.

Now somebody more clever than me have tried a similar exercise.Here is the abstract from a new paper from the talented Arash Molavi Vasséi:

“This paper provides a systematic translation of F.A. Hayek’s informal exposition of capital theory in Utility Analysis and Interest and The Pure Theory of Capital into a model. The underlying premise is that Hayek adopts infant versions of `modern’ analytical tools such that a rational reconstruction of his capital theory by established neoclassical tools is admissible. The major result is that Hayek’s capital theory contains a generalization of the Ramsey-Cass-Koopmans model. In concrete, Hayek provides the solution to an infinite-horizon deterministic social planner optimization problem in a one-sector economy such that the rate of pure time preference encapsulated in the discount factor increases in prospective utility. With respect to stability properties, he emphasizes that the system converges even in the special case of constant returns to per-capita accumulation.”

How cool is that? Pretty cool if you ask me, but take a look at the paper yourself.

PS Arash has promised me that his next project will be on NGDP targeting and/or Market Monetarism.
PPS I hope you all remember Arash’s clever discussion on (dis)equilibrium in Market Monetarism.

%d bloggers like this: