The connection between level targeting and futures price targeting
I’ve talked a lot about the need for level targeting, i.e. setting a growth trajectory for NGDP and promising to make up for any near term overshoots or shortfalls. And I’ve also talked a lot about the idea of targeting the price of NGDP futures contracts. But I don’t recall talking about the connection between these two policies, which might shed some light on the importance of level targeting.
Policy lags are one of the difficulties that face monetary policymakers. The Fed often doesn’t even know that the economy is in recession until several months after the downturn has begun (based on the retrospective dating of the cyclical peak by the NBER.) It turns out that both level targeting and futures targeting help address this issue, in fairly similar ways.
Imagine a model where millions of people and businesses observe local demand shocks, and that data is aggregated 3 months later in the quarterly NGDP numbers. It’s quite possible that the “market” would know things that no single individual would know—as the market will reflect aggregate optimism and pessimism, which is partly based on all those local demand shocks.
The advantage of NGDP futures targeting is obvious when there are policy lags. The market will sense velocity changes before the Fed does, and offset them with adjustments in the base (or fed funds rate if you prefer to think in Keynesian terms.) But how about level targeting, how is that like futures targeting?
Suppose that pessimism causes velocity to drop 2% before the Fed is able to notice and take corrective action. Also suppose the Fed is doing plus 5% NGDP level targeting. The markets will expect the Fed to return the economy to the trend line over the next 12 months. This means they will now expect 7% NGDP growth; the normal 5%, plus another 2% to offset the near term shortfall. This means they will expect easier money than if the Fed was doing growth rate targeting, and letting “bygones be bygones.” More expansionary than if they settled for 5% growth after the 2% shortfall.
Now let’s assume that the markets notice the shortfall before the Fed does, and they expect the Fed to ease as soon as the shortfall is noticed. That is, imagine a period like September 2008, when the TIPS market saw rapid disinflation but the Fed was still worried about high inflation. In that case the markets will expect Fed easing before the Fed does. Now recall than in modern new Keynesian economics the current level of aggregate demand doesn’t just depend on current short term rates, but also expected future short term rates. I.e. it depends partly on longer term rates. The anticipation of Fed easing will immediately reduce future expected short term rates, and will immediately reduce long term rates. The market does the Fed’s work before the Fed even realizes there is a problem.
So with level targeting the market will be moving expected future interest rates around in such a way as to keep expected future NGDP (12 months out) right on target. And here’s the best part of all. Remember my initial assumption that NGDP temporarily fell below target, before the Fed corrected the problem? It turns out that with level targeting the initial deviation from the target trajectory will be much smaller than if there was growth rate targeting, even if the Fed makes no immediate attempt to get the economy back on track. Because the market will depress future expected rates, they will boost AD in the current period, even before the Fed noticed that there was a problem.
It’s not quite as good as NGDP futures targeting, but it comes so close that I’d guess it would deliver more than 90% of the potential efficiency gains from NGDP futures targeting, maybe 95%. Given the current sorry state of monetary policy, that’s a lot of $100 bills lying on the ground waiting to be picked up. Let’s hope the Fed notices them when it meets next week.
Tags:
28. October 2011 at 17:51
Level targeting has one more massive advantage over futures targeting – it’s not vulnerable to manipulation or instability due to thinly traded markets or markets dominated by algorithmic trading. You have a great deal more faith in the stability and accuracy of such a hypothetical market than most (all?) traders. Following the flash crash (in something as large and liquid as the entire US stock exchange), and several mini-flash crashes since then, the reliability of a futures market at any given point in time is a hard sell.
As an aside, when we invited you to toast Trichet’s departure at Watch City, that was before Goldman kicked your sorry self into the spotlight. If tomorrow (sat) is still good, let us know what time.
28. October 2011 at 18:06
Statsguy, Flash crashes in NGDP futures markets don’t cause a problem for monetary policy, as they’d merely distort the monetary base for a few minutes, or maybe a day. But NGDP growth depends on rates over a much longer period of time.
Maybe we should postpone until a less busy day–plus the weather tomorrow will be lousy. This weekend I need to finish a book review for a conservative magazine.
28. October 2011 at 18:07
Scott
StatsGuy has makes a good point. So maybe your 95% goes up to “better” than 100!
28. October 2011 at 18:56
well, if this policy is implemented, then the stock markets will be more boring than watching paint dry! the end of investment banking, algotrading and market makers in all sorts of derivatives. and they can get a more useful job as frictions cannot occur anymore. P.En. and rocket-scientists can go and build rockets to the moon.
imagine all possibilities!
28. October 2011 at 19:33
Scott, I know I could go read a bunch of stuff you’ve written in the past, but I bet plenty of your readers aren’t totally sure either: Could you spell out what the Fed’s balance sheet would consist of, in your ideal world? Like, would it be 95% Treasuries and 5% NGDP futures contracts, or 105% Treasuries and -5% NGDP futures contracts (if the Fed wanted to push down expected future NGDP)?
28. October 2011 at 19:34
(I’m not asking you if it’s long or short, I’m asking you if I’m right in thinking it’s 95% Treasuries and then +/-5% NGDP futures, whichever way it needs to be. Or, am I totally misunderstanding how you want the Fed to handle asset purchases?)
28. October 2011 at 20:27
I can’t understand the anti-nominal GDP crowd. It seems they will grasp at any peevish sentiment, any feeble obsession or feigned moral, to naysay NGDP targeting.
The USA and Europe badly need to join the Market Monetarists.
This was another excellent post by Scott Sumner.
28. October 2011 at 21:20
Ben Cole,
It’s very typical for closed minded people to simply assume that anyone who disagrees with them is driven by sinister, evil motives. Besides, it’s always easier to attack people than their ideas. Attacking ideas takes thought.
28. October 2011 at 22:09
Scott,
Was there a proposal earlier (by you) where markets would directly control base money based on NGDP futures contracts or am I making that up?
28. October 2011 at 23:51
John, why are you so sanctimonious? All of your comments resemble sermons more than anything else.
29. October 2011 at 00:17
CA,
You haven’t read many of comments. Most of them are straight econ stuff. My beef with Cole is that he thinks us hard money types just have bad intentions.
29. October 2011 at 00:28
This is an important post for us amateurs, but it actually wasn’t the post I was expecting after reading the headline.
I think that one of the difficulties to understanding the Sumnerian synthesis (you’re the poster child now!) is that it involves SO MANY changes in policy.
1. Explicit fed targets
2. Level targeting
3. NGDP targeting
4. NGDP future contracts
Could you do a post/response that ranks these policy initiatives on your expectation of their effectiveness? Or on how implementation of different combinations might go?
If there’s one truth about policy, it’s that you don’t always get what you ask for. What if you just got an implicit NGDP rate target? By that I mean simply a tilt in fed rhetoric to NGDP instead of inflation / RGDP. A WSJ reader might think “mission accomplished” and judge your ideas based on a result of such an experiment.
29. October 2011 at 01:16
Scott,
Assuming a 5% NGDP level target was implemented next week, do you think that announcing the target would change expectations enough to do the heavy lifting or do you think they would have to engage in a lot of asset purchases to bring NGDP back to its former path?
If they had to buy a lot of assets, are you at all worried about their ability to sell the assets back and slow down? If they bought a lot of unconventional assets, these assets might lose their market value when the Fed tries to sell them back (plus interest rates will probably rise on the T-bills) which might make it difficult for them to tighten sufficiently.
Does anyone know if Scott did a post about how the Fed could deal with possible problems of an “exit strategy?”
29. October 2011 at 04:08
… let us know when your life is a little “slow”, probably after finals are graded.
29. October 2011 at 05:31
Marcus, I still think futures targeting is superior, but I’d be happy if we had level targeting.
Jaap, Au contraire, stocks were very exciting during the roughly 5% steady NGDP growth of the Great Moderation.
Bob, I’d assume about a trillion in base money, and hence T-bonds. I have an open mind on futures, but I’d guess orders of magnitude lower, perhaps less than 10 billion. NGDP futures are essentially an artificial prediction market. You probably don’t need to make it all that big to get an efficient market forecast. But again, I have an open mind.
John, Yes, it’s in my links to key blog posts on the right margin.
DW. Good point. If the Fed were to do 5% NGDP targeting right now, it wouldn’t help very much. If they did 5% NGDP level targeting from right now, it would help a little more.
If they did 5% NGDP targeting backdated to 2009, we’d get an explosive recovery, actually too fast.
I’d recommend aiming for 6% or 7& for a couple years, then 5% thereafter.
If the Fed did price level targeting from right now it wouldn’t help much. If they backdated it to level targeting from 2008 it would help more, but not as much as NGDP level targeting.
An explicit 2% inflation target (not level targeting) would be the weakest of all, as it’s not much different from what they have been doing.
If they go to level targeting of any sort, futures targeting doesn’t add much.
John, If they did a 5% NGDP target, level targeting, they might have to buy or sell bonds, I’m not sure. If it was a 7% NGDP target, level targeting, they’d have to sell bonds.
The exit strategy is selling bonds or raising IOR, that’s the least of our problems. The Japanese implemented it in 2006 (sharply reducing their base) and had no inflation at all.
There’s a reason every single developed country central bank in the world eliminated high inflation in the 1980s and 1990s–it’s really easy to do. It shows how provincial Americans are that they think Greenspan was some sort of “maestro.”
Statsguy, Will do.
29. October 2011 at 05:33
Off topic, does anyone know of an article that clearly explains the Greek bailout? Some articles say it won’t do much, Greek debt will still peak at 142% of GDP in 2013, others say Greek debt is being cut in half. Who’s right? Is there an article that has accurate data?
29. October 2011 at 06:29
Not really, it’s incredibly confusing… I track the NYT, WST, BBC, a couple blogs, and one or two investor sites, and it’s a mess.
Try here –
http://topics.nytimes.com/top/news/international/countriesandterritories/greece/index.html
At least it gets updated.
I think there are still open questions:
Will the IMF/ECB debt take an eventual haircut, and how much? This is one of the big sources of discrepancy. Private greek debt (aka, debt owned by european banks backed by euro sovereigns – hence “private”) seems to be cut 50%, but that’s only a portion of total debt.
Will “voluntary” reductions in debt trigger CDS claims as a “credit event”, when it obviously is, but authorities are creating legal cause to not declare it a credit event?
Which projections of Greek GDP “growth” does one believe? Historical IMF projections (that sharp “austerity” will lower the deficit rapidly) are repeatedly proven a joke; Everyone who holds bonds is trying to take the smallest cut possible and “demonstrate” that the remaining greek debt load is “sustainable”, when it’s not. The higher end estimates of peak debt/GDP ratios probably come with assumptions that ECB/IMF owned debt won’t be cut and the austerity-driven economy will grow less than prior projections.
Greece still has a ludicrous social “safety” net – and keeps reneging on promises to cut it as it faces domestic resistance.
What is “legal”? Issues concern the legality of ECB bond purchases (ECB for so long claimed they couldn’t – or, that Trichet didn’t want to – but then when forced, they did; not unlike the Fed). Issues concerning who in Germany needs to approve a vote to sign onto the package (a special committee? the entire lower house?).
Great example of Coasian bargaining and transaction costs. The “voluntary” haircuts were a precondition to the EFSF passage, which (last I heard) was a first-loss leveraged fund (they’d back the first 20% loss). This means that if a future haircut on EFSF co-backed bonds were to come in at 70%, the EFSF would eat it’s 20% first, leaving the real loss at 10%. This is tasty for investors, because if they buy a bond yielding 6% for 10 years (79% higher than nominal purchase price), vs. a 10 year yielding 2.5% (which is only 28% more than nominal price), and they take a 10% loss off of face value, they still earn a truckload over the 2.5% bond… But if the future haircut comes in at 50%, ouch. Seems to create more knife edge volatility, and no one knows what the heck the leverage will really mean (if a 400B EU fund gets levered to 1.2T, doesn’t that still mean the sovereigns are on the hook for the full 1.2T?). I think the key issue is that the sovereigns are bottowing from the ECB (cheaply?), and at the very end of the long road there’s an implied ECB backstop (the ECB isn’t going to let the entire continent go under…). So, at some level, money is being printed. Maybe. Or, maybe not.
But I’m on the outside of this, and somewhere in europe there are a couple hundred bankers who actually know what’s going on.
Maybe. Or, maybe not.
29. October 2011 at 06:48
1. Expect 4% on less from now. Just accept it. It gives hawks a guarantee that unless things get really nutso, they never deal with inflation over 2%. Recognize this is the EXACT line of thinking Goldman used in reverse (they shaved it because RGDP is slower lately in peoples minds), you too can be as smart as them!
2. WHY can’t you do the same calculus just ONCE where you are talking about how booms get pissed on? Explaining that 2008 shock wouldn’t happen is FAR MORE POWERFUL.
Until the discussion is balanced – which means saying out loud that we will have higher rates more often than we had in the past – you are just cheer leading for inflation.
29. October 2011 at 08:24
Bob and Scott:
Remember, futures contracts are like bets, not like assets.
Futures exchanges require margin deposits, but this is like a performance bonds. The exchange in making sure that people won’t welsh on their bets.
My view about index futures convertibility is that the central bank should do its best to not take a position on the contract. Oerdinary open market operations are used so that market investors take exactly offsetting positions on the contract. The central bank is obligated to take a position implied by any difference, so that the shorts and longs exactly match (as they must.) So, if the shorts and longs other than the central bank match, the Fed takes no position. If there are more longs than shorts on the market, the Fed is left with the difference–a short position. And vice versa. More shorts than longs other than the Fed, then the Fed takes the difference, a long position.
If the longs are more than the shorts, that means the market expects nominal GDP above target. The Fed understakes open market sales of securities. The point is so that the contraction in money and credit will reduce the expected level of nominal GDP and some of those who were long become short. When the shorts and longs match, the Fed is out.
If the shorts are more than the longs, the Fed does ordinary open market operations with bonds. This expands money and credit. This raises the expected level of nominal GDP until some shorts turn long, and when they match, the Fed is out.
In equilibrium, the Fed take no position and all of its assets are in T-bills. And I don’t think counting the Fed’s position on the contract as an asset makes sense. It isn’t an asset. The Fed might pledge some ordinary securities to cover any losses on the contract. But that is just a sequestration of the Fed’s capital.
29. October 2011 at 16:29
Yeah it’s straight up going to the racetrack.
You take your $100K, you bet it on “NGDP too low”
If you lose and it is too high, your money is gone, the money supply shrinks, to piss on the boom.
If you are right in it comes in too low, they hand you fresh new money.
This still requires a good answer on revised data two years later. What do we do about it?
30. October 2011 at 08:05
Statsguy, Thanks for the info. The more I read about this situation the more Greece seems like a gigantic criminal enterprise. I’m especially interested in the fact that Greece hid the size of its deficit for years. When Enron does stuff like that, the culprits go to jail. Perhaps the former leaders of Greece should be tried at the Hague.
What is the rationale for giving some debt holders a haircut but not others?
Bill Agreed, I’ve argued that each day the Fed should start the base off at a level where they expect the long and short positions to balance. Ex post they won’t precisely balance, but they’ll come close.
Morgan, I think you go with the initial estimate, or the first revision one quarter later. Errors are unforecastable anyway. And I’d use NGDI (income.)
30. October 2011 at 09:01
[…] this recent post I explained one advantage of level targeting; the fact that it leads market participants to assist […]
30. October 2011 at 18:02
“What is the rationale for giving some debt holders a haircut but not others?”
As I understand it, the rationale is this: The ECB and IMF said “no, we’re not going to take a haircut because we bought the debt to rescue the banks after we knew it was compromised.” The banks were told they’re goint to take a haircut, or the EU was going to let them all go under and then take them into state receivership.
THIS IS WHAT HAPPENS when you try to “negotiate” an “internal devaluation” without using nominal currency as a mechanism. This is why liquidationism FAILS IN PRACTICE, even though in theory it seems like it should work.
31. October 2011 at 16:23
Statsguy, Thanks for clarifying that.
8. November 2011 at 08:45
[…] Sumner has long argued that the Federal Reserve or the US Treasury should help set-up a NGDP futures markets and conduct monetary policy based on market expectations of NGDP. This is a great idea, but so far […]
27. January 2012 at 23:28
[…] Sumner and other Market Monetarists (including myself) favour the use of NGDP futures to guide monetary policy. Other than being forward-looking a policy based on market information […]