Jonathan Pedde on the UK election and fiscal policy
Tyler Cowen directed me to an excellent new paper by Jonathan K. Pedde:
Standard zero-lower-bound New Keynesian models generate large fiscal multipliers and expansionary negative supply shocks. Thus, according to these models, a political party that implements fiscal contraction coupled with policies to increase aggregate supply should unambiguously cause economic contraction, compared to a party that implements the opposite policies. I test this prediction using high-frequency prediction- and financial-market data from the night of the 2015 U.K. election, which featured two such parties. By analysing financial-market movements caused by clearly exogenous changes in expectations about the election winner, I find that market participants expected higher equity prices and a stronger exchange rate under a Conservative Prime Minister than under a Labour P.M. There were little to no partisan differences in interest rates, expected inflation, or commodity prices. These results cast doubt on the empirical validity of zero-lower-bound New Keynesian models.
Some initial reactions:
1. Wow, was that fast! The election occurred on May 7th and a fully done research paper was at SSRN on May 12th
2. The paper has a rather MM feel to it. Fiscal stimulus doesn’t affect growth, no paradox of toil, use market price changes to identify the effects of policy shocks, etc. It’s a much more formal treatment of what I do informally in this blog.
[Update: Jonathan Pedde pointed out that I misinterpreted his fiscal policy claim. Here’s the more nuanced claim he actually made:
Of course, these results do not necessarily imply that the fiscal multiplier is zero. Perhaps a Conservative (rather than a Labour) P.M. leads to less expansionary fiscal policy as well as higher stock prices and exchange rates, but the higher stock prices and exchange rates happen in spite of the less expansionary fiscal policy, not because of it. But in order for this to be true, then something else about a Conservative P.M. must have been responsible for the higher stock prices and exchange rates. But, in the context of a ZLB NK model, that “something else” would probably have to be a policy relating to aggregate supply. So this argument requires that policies to increase aggregate supply actually increase GDP – which, of course, is at odds with the predictions of the ZLB NK model.
So he’s testing the overall ZLB NK model, not just fiscal policy.]
My only reservation is that one must be careful in interpreting stock market reactions. Pedde claims that the rise in stock prices reflects higher growth expectations after the Conservative victory. I think that’s right, but it’s also possible that higher stock prices reflect lower expected future taxes on capital income in the UK. In that sense elections are different from something like Fed announcements.
In fairness, Pedde’s growth hypothesis is confirmed by the concurrent move upward in the UK pound, which also signals faster growth ahead. But this still shows the desperate need for more and better macro predictions markets, with higher levels of liquidity.
Interestingly, the paper relied on the Hypermind prediction market for the election probabilities. This market also contains a set of NGDP futures markets, but with only $10,000 in prize money this year. Think about the following. If we boosted the prize money 50-fold to $500,000 per year, the market would be much more liquid, much more efficient. That’s a lot of money for you and I, but for the Treasury or the Fed it’s like the coins that slip down between the cushions on your couch. It’s nothing. And it would open the door to a new golden age of macro research. We could identify the stance of monetary policy in real time. VAR studios would immediately by 100 times better.
Why hasn’t it happened yet? Good question. All I can say is that I’m doing my part; now I’d like to see some bigger name economists make a push for these markets.
PS. My admittedly unscientific sense of Hypermind is that the NGDP market is now pretty accurate at identifying NGDP market forecasts, and also the gradually fall in the growth forecast over recent months. But I don’t think it’s liquid enough at this point to do event studies. Trading volume is too low.
On the other hand if traders knew that the most successful trader would walk away with $100,000, the second most with $75,000, the third most with $60,000, the 4th most with $50,000, etc., etc., even after having to put up none of their own money, well then I’m pretty sure that trading volume would get much, much higher.
On the plus side, even the very small Hypermind NGDP market is yielding interesting results. I’ve argued that the current 3.4% NGDP growth forecast during a recovery year is consistent with the hypothesis that we are in a NGDP Great Stagnation, with a trend rate of 3% NGDP growth.
PPS. Glad to see David Andolfatto and Noah Smith saying many of the same things that we (market monetarists) have been saying about Britain.
Tags:
16. May 2015 at 08:15
‘1. Wow, was that fast!’
Exactly what popped into my mind.
16. May 2015 at 12:40
“…use market price changes to identify the effects of policy shocks, etc.”
What market?
16. May 2015 at 16:17
I wonder if the results in Great Britain reflect a hope for business tax cuts as opposed to a balanced national budget.
And maybe that is the best thing.
In 2002 Bernanke said the best way to do QE was in combination with tax cuts. I presume Bernanke knew that tax cuts would lead to a larger deficit.
16. May 2015 at 19:48
These results cast doubt on the empirical validity of zero-lower-bound New Keynesian models.
Hard to believe that evidence is still needed.
I think the participation rate is probably more important than the prize money, though they’re obviously related.
Prediction markets are a good idea with a lot of applications. For instance, I’d love to see climate science funding tied to prediction markets based on IPCC predictions — the whole issue could be self-(de)funding.
17. May 2015 at 06:01
Sumner gloating about a paper that reinforces his priors. Ho-hum.
17. May 2015 at 13:38
In what kind of model can it make no difference if government invests in projects with positive npvs, negative NPV or reduces investment in projects with positive npvs, negative NPV or neither?
17. May 2015 at 16:44
@scott
On the other hand if traders knew that the most successful trader would walk away with $100,000,…. etc., etc., even after having to put up none of their own money, well then I’m pretty sure that trading volume would get much, much higher.
I don’t know. Seems like a lot of work with pretty volatile returns to me.
I think that’s right, but it’s also possible that higher stock prices reflect lower expected future taxes on capital income in the UK.
Good point.
18. May 2015 at 05:54
Yes I really like that point about potential AS – that seems the weakest link in the NK model trying to explain the UK over recent years.
But the paper is probably too harsh on NK models. If we allow that the Bank can inflation expectations at the ZLB via forward guidance on interest rates or QE affecting future money supply/rates & expected inflation, then you can rescue the NK model and most of the weird stuff about AS shocks and fiscal policy disappears. At least if I understand it right, which I might not.
18. May 2015 at 05:55
Sorry, rushed comment… “allow that the central bank can affect inflation expectations”.
18. May 2015 at 06:23
I don’t know. Seems like a lot of work with pretty volatile returns to me.
Exactly. You get a deep and liquid market by including both capital providers (institutions) and capital allocators (traders). If a good trader can work for a bank or a hedge fund and make plenty of $$$, then the real cost of participating in the NGDP market isn’t the potential loss of capital, it’s the opportunity cost or focusing on that market instead of another. And capital providers aren’t going to pay traders to put their money to work in this market for a potential $100K.
18. May 2015 at 11:34
Talldave, I once did a paper with Aaron Jackson on using futures markets to forecast global warming.
dtoh, OK, make the number bigger. But remember that traders cannot lose. And you must be richer than me. 🙂
Britmouse, I agree, but I think he’s talking about a specific version with monetary policy ineffectiveness and the “paradox of toil.” Other versions are out there, and some NKs (Bernanke?) probably don’t agree with Krugman/Eggertsson.
Njnnja, You and dtoh are thinking in terms of financial markets, I’m thinking of it as a prediction market. Maybe you two are correct. I think the ($5,000) one year Hypermind market is already reasonably accurate, and now I’m talking about a 100-fold improvement. So maybe you need more than a 100-fold improvement, maybe 500-fold more money, it doesn’t really change my argument.
18. May 2015 at 18:41
@scott,
Of course the traders can lose. Spend a lot of time and not make any money or not make much money.
18. May 2015 at 19:29
Former Brazilian President and his criticism on the proportional electoral system, which I totally agree…
http://www.ft.com/cms/s/0/b5c97688-fa6e-11e4-aa42-00144feab7de.html#axzz3aYFHJ7kv
19. May 2015 at 17:51
dtoh, Yes, I know there is time involved.