Friday
Nov202015

2015 Q3 NGDP prediction market

 

Up until now we've used Inkling Markets as our prediction market provider, but they've recently been bought by Cultivate and will be offline from December. The finance questions for Cultivate will be run through a new website called Alphacast.

At the moment, members are unable to post their own questions. However, I've received some really helpful input from their technicians and one of their early trials will be a prediction market for UK NGDP. How cool is that! This is good timing because the Office for National Statistics release the second estimate of the National Accounts on Friday November 27th, and this will contain the first look at NGDP for Q3.

The question is: What will be the quarterly growth rate of NGDP for the UK economy for Q3 2015?

The market will only be running for a week, and as ever there's (relatively) large arbitrage opportunities early doors. So get trading!

Update: (27/11/15):

The correct answer is 3.4%.

Here's the Alphacast chart:

 

Wednesday
Nov112015

The Transmission Mechanism

I am yet to find the difinitive survey article that explains the transmission mechanism of monetary policy. I think there's 3 main reasons why it's proving elusive:

  1. There is no single transmission mechanism - hence it's always going to be complicated to present.
  2. There are big differences between US and UK monetarists.
  3. We've learnt new things about the transmission mechanism since 2008 but there's a pedagogical inertia that retains existing work.

A classic US textbook on monetary economics is Frederic Mishkin's "Economics of Money, Banking and Financial Markets". But look at the diagram below:

On first glance this is exactly the type of way to present such a tricky topic. It lays out several alternative channels, clarifies their background (i.e. presents credit view as slightly separate), and shows how they impact output. But it's not coherent. The key distinction between "interest rate" and "asset price" effects is flawed on the grounds that the interest rate matters because it's the price of bonds (i.e. an asset), whilst each of those asset price effects - whether it's the exchange rate, equity prices or credit markets - rely on interest rate changes (albeit the "cash flow" channel relies on nominal rather than real interest rate changes). This is why I struggled as an undergraduate - it's aint MECE.

I prefer this Mishkin article but it suffers from the same problem. Here's my attempt to summarise the transmission mechanism in terms of alternative schools of thought:

Not very helpful. One can also look at the Bank of England, and the ECB, but they leave out lots of important present-day channels. So help me out - where's the difinitive explanation of the transmission mechanism?

Friday
Oct232015

Low interest rates /= easy money

The impact of monetary expansion on nominal interest rates is unclear. The liquidity effect implies that in the short run they will fall. However the Fisher effect implies that in the long run they will rise. Whether the monetary expansion ultimately leads to a nominal interest rate that is higher or lower than the original rate depends on the relative strength of these effects. If the liquidity effect dominates the Fisher effect then expansionary policy will cause rates to fall. If the Fisher effect dominates the liquidity effect then the expansionary policy will cause rates to rise. (I'm neglecting several things, but I hope the above still makes sense).

This is important, because whether or not a low interest rate is a sign of expansionary monetary policy depends on inflation expectations.

Ben Southwood discusses a fascinating 1998 paper by Ellingsen and Soderstrom. From the abstract:

if monetary policy reveals information about economic developments, interest rates of all maturities move in the same direction in response to a policy innovation. If, on the other hand, monetary policy reveals information about the central bank's policy preferences, short and long interest rates move in opposite directions.

My interpreration of this is that when there is a policy surprise the Fisher effect dominates the liquidity effect. But this poses two really interesting questions:

1. Is there anything important about the surprise, other than its impact on inflation expectations?

2. Does this adhere to the Lucas critique?

Update: I've just seen that Samuel Hammond and Ben had a disagreement about what the paper implies about the liquidity and Fisher effects on Twitter. I would love to see a longer discussion.

Wednesday
Oct212015

Do Cantillon effects matter?

A couple of years ago there was a big blogosphere debate about Cantillon effects. It was prompted by Sheldon Richman's claim that:

Since Fed-created money reaches particular privileged interests before it filters through the economy, early recipients—banks, securities dealers, government contractors—have the benefit of increased purchasing power before prices rise.

Scott Sumner argued that "it makes very little difference how new money is injected" (see Sumner's follow up post here, see Robert Murphy's attempted resolution, and see Sumner's response.)

This debate struck me as a classic argument between comparative statics and market process theory. The process by which monetary expansion occurs will cause some prices to change. Whilst prices adjust, some groups will benefit and some will lose out. It seems uncontroversial to me.

Perhap's the problem is with immediately turning this into a debate as to whether "Cantillon effects" exist. We can think of them in two ways. The first is whether there's a wealth effect on the part of the early recipients of freshly created money. The second is the consequences of the relative price effect. Note that Sumner is challenging Sheldon's account of the former, whilst it is the latter is the really important contribution of Austrian monetary theory (i.e. the interplay between non-neutrality of money, relative price changes, and the capital structure).

Having read through the debate, I believe that the following statements are correct:

  • There is a perceived wealth effect for the specific first receivers (in other words the first receivers gain a consumer surplus from their mutually beneficial voluntary transaction. Mario Rizzo made this point here).
  • There is a wealth effect to some people across the economy as a whole

It is not necessarily the case that:

  • There is a wealth effect for the specific first receivers - because as Sumner points out, they are receiving the market rate for their asset.

I was prompted to look at that debate having read through a couple of old papers by Richard Wagner. One of them, (co-authored with Steve Daley), emphasises the relatiohip between the preferences of the recipients and the impact on relative prices:

"if money is injected at points where the recipients have particularly high demands for goods with relatively inelastic supply, those particular prices will rise further than they would under some alternative locus of monetary injection"

The other article is about Georg Simmel’s Philosophy of Money. Wagner argues that the essence of a Cantillon effect is that "the process of monetary injection will influence the concrete pattern of activities within a society". Whilst a comparative statics approach may dismiss the result as "mere distributional effects", Wagner argues that:

This dismissal arises out of a frame of reference where all that matters is the state of some aggregate economic variables. Yet the dynamic forces that are at work at shaping societies precisely work their way through those micro channels; the aggregate resultants are objects neither of choice nor of desire

In my "Choose your own financial crisis" I grappled with implied counterfactuals. I think a similar issue needs to be made explicit when discussing Cantillon effects. The question is: What is the implied counterfactual?

Bob Murphy was almost right to treat it as a semantic debate about what constitutes fiscal policy. The key point is that Austrian monetary economics rests on the far broader calculation debate. Cantillon effects are important not so much because of non-neutrality (i.e. a monetary reason), but because they disrupt the price mechanism. Incidently, I also think this is how to resolve Jeff Hummel's neglected criticism of ABC. He argues that Austrians have failed to clarify why they assume that monetary expansion should escalate. Whilst an ever increasing growth rate in the money supply will indeed lead to a necessary crash, why assume that a constant growth rate would inevitably escalate? I think the solution requires us to consider how new money is being spent, and the implication for economic calculation. Consider the government subsidies that went to Solyndra. There is no arithmetic reason to say that they should need to rise over time, but our understanding of the economic calculation debate tells us that it is unsustainable. The boom is unsustainable because it is an example of government intervention. Whether that should be considered monetary policy, or fiscal policy, is a separate issue. It's political economy.

Monday
Oct192015

M4 Lending and Divisia measures of the money supply

The last few years seems to have provided strong evidence that bank lending is not a pre-requisite for economic growth. Getting the banks to lend has been difficult, and yet the economy has been performing strongly. This is an important empirical contribution to a passionate theoretical debate.

The chart below shows two measures for M4 lending:

One possible explanation is de-leveraging - if individuals and companies are paying off debt we should expect lending to be subdued.

Another explanation is that lending is taking place outside of traditional channels. Perhaps the economic recovery has required lending, but not necessarily bank lending. Sources such as crowd funding, peer-to-peer lending or pay day loans have seemed to constitute a larger share of households access to finance (and has led to more regulatory attention to peer-to-peer lending. This economic activity should be captured in standard measures of broad money, so the argument isn't that the money supply is growing by more than official figures imply. However it is economically pertinent because:

1. It could show how changes in the composition of macroeconomic aggregates matters

2. It could show how financial innovation can reduce the demand for money (and thus increase velocity)

I'm not sure which of these avenues is most useful, and need to read more.

One interpretation is that broad money measures can mask a lot of important activity. This might lead to a greater focus on narrower measures (such as MA) or weighted measures such as Divisia. I have mentioned Divisia measures before, and here's a current look:

Whilst the growth rate for private non-financial corporations remains in the 12%-16% range, household Divisia has been steadily slowing. This is causing me to reduce the weight that I put on them - I felt that strong Divisia growth in 2013 was a sign of economic expansion, but NGDP growth tailed off (this is using series VTSR):