Advice on NGDP futures

One of my commenters (John Salvatier) is thinking about using his own money to set up some NGDP contracts on Intrade.  He sent me an email with this information, and asked for suggestions:

Each contract would be based on the BEA final estimate of NGDP for the specified quarter and the quarter two quarters before (i.e. 6 months before). The formula for the contract payout would be
 
payout = ln(NGDPend/NGDPstart) * periodsInYear * 1000
if payout > 100: payout = 100
if payout < 0: payout = 0.

This specification has the nice property that payout/10 = continuously compounded annual growth rate of NGDP during that period (in %). However, the drawback of this specification is that 2008q4, 2009q1 and 2009q2 would all have had negative payouts so they would have paid out 0 instead, which means  you would have had trouble gauging expected the severity of the downturn.
 
An alternative specification is

payout = ln(NGDPend/NGDPstart) * periodsInYear * 500 + 50
if payout > 100: payout = 100
if payout < 0: payout = 0

This specification would have had positive payouts even during the most severe phase of the downturn, but it makes the contract prices more difficult to interpret and reduces the variation in contract payouts.

I would have contracts for the 6 month intervals between now and 2.5 years from now. When one contract expired, another one would be started for 2.5 years from now. I would consider using quarterly contracts instead of semi-annual contracts if people thought that was important.

I will be making markets in these contracts, ensuring a small spread. This will effectively subsidize informed traders by giving them the option to trade a little cost. I may do this manually at first, but I hope to be able to build an automated market maker to do this for me.

A few comments:

1.  I strongly favor NGDP futures markets, and thus am obviously happy to see one being set up.  But I fear that without government subsidies, there will be very low volume.  That’s why I favor having the Fed subsidize trading in an NGDP futures market (by paying higher than market rates on the margin accounts.)  Nevertheless, I greatly appreciate John’s willingness to put his money on the line.

2.  I am not an expert on futures markets, so John and I would appreciate any advice on how best to set up the contracts.  We’d like to make them customer friendly, but also able to provide point estimates of expected NGDP growth.  And I believe John’s proposal does that.   I believe there are currently some RGDP contracts that merely involve binary outcomes, such as whether growth will be higher or lower than 3%.

3.  If this is set up I am going to ask all my readers to please consider trading a few contracts.  The price of the contracts are pretty low (I believe $10), and since NGDP is somewhat predictable it’s unlikely you’d lose more than a few dollars on each contract.  Is that too high a price to pay to show solidarity with the entire Money Illusion project?   I will certainly buy some contracts.   If no one trades the contracts I might just go on strike and not post for a while.    :)

I’ll keep you posted.


Tags:

 
 
 

51 Responses to “Advice on NGDP futures”

  1. Gravatar of Joe Joe
    24. June 2010 at 14:45

    Funny, I was actually just about to ask you why you don’t just set up your exchange instead of waiting for the fed…

  2. Gravatar of Andrew C Andrew C
    24. June 2010 at 15:18

    I’m definitely in. Just created an Intrade account for this purpose.

  3. Gravatar of John Salvatier John Salvatier
    24. June 2010 at 15:37

    Additionally, people who are actually interested in modeling the NGDP process can win some of my money by creating a model better than the models of others and using it to trade NGDP futures.

  4. Gravatar of JimP JimP
    24. June 2010 at 16:19

    A Struggle Inside the Fed

    Ben doesn’t have the votes.

    The hard money guys are out in force.

    We need a statement from Obama. Will we get one?

    http://www.telegraph.co.uk/finance/economics/7852945/Ben-Bernanke-needs-fresh-monetary-blitz-as-US-recovery-falters.html

  5. Gravatar of Carl Lumma Carl Lumma
    24. June 2010 at 16:22

    I pledge to trade some contracts if this gets set up.

    Does Intrade use a market scoring rule to help with the low volume issue?

  6. Gravatar of John Hall John Hall
    24. June 2010 at 16:24

    I’m not sure why you would want to continuously compound a quarterly growth.
    For instance, why not: 1000*((NGDPend/NGDPstart)^4-1) and why care if the payout turns negative. Futures contracts are settled daily, so I wouldn’t think it should be much of an issue.

  7. Gravatar of Morgan Warstler Morgan Warstler
    24. June 2010 at 16:37

    I’m in if you first me give a plain spoken explanation of the levers used to target NGDP. Who gets the money from Fed, and how is it given to them? How much are you talking about? Please be super specific.

  8. Gravatar of Bill Woolsey Bill Woolsey
    24. June 2010 at 18:47

    Perhaps I don’t understand, but it looks like these contracts are for growth rates.

    GROWTH RATES!!!!!!!

    How about levels?

    It is essential to move to “growth paths” and away from growth rates.

  9. Gravatar of Doc Merlin Doc Merlin
    24. June 2010 at 21:56

    The first formula doesn’t allow for negative NGDP changes which is what we are trying to look for. So, my vote is for the second formula on that.
    wi
    However it doesn’t allow for that wide a range in variability. It maxes out at 5% change in a single six month period. instead of the * 500, a * 400, would allow for a bit more range for extreme events. Not that big a deal though, 5% change should cover almost all events in our current system (hopefully!).

    On a more useful note, if you do use the second formula, you can make a payout table to make it easy for investors to understand in terms of % change within the period.

    @Bill:
    Because of the way intrade contracts are handled, scaled on a 0 to 100 payout range, its difficult to come up with a way that scales for levels that has enough variability to be interesting.

  10. Gravatar of Doc Merlin Doc Merlin
    24. June 2010 at 21:58

    Oh, forgot to ask, I may have misunderstood something, what is the role of the two previous periods in the calculation of payout?

    ‘Each contract would be based on the BEA final estimate of NGDP for the specified quarter and the quarter two quarters before (i.e. 6 months before). The formula for the contract payout would be’

  11. Gravatar of Alejandro Alejandro
    25. June 2010 at 03:18

    “But I fear that without government subsidies, there will be very low volume. That’s why I favor having the Fed subsidize trading in an NGDP futures market (by paying higher than market rates on the margin accounts.)”

    Trust the force, my young padawan.

  12. Gravatar of Bill Woolsey Bill Woolsey
    25. June 2010 at 05:05

    Why not just generate an index number. Take Nominal GDP (should be final sales of domestic product, but..) Then divide by its level for third quarter 2008. Multiply by 100 and subtract 50. With each contract it will rise. After a few years, change the base. And lobby intrade to change so it isn’t focused on betting who will win, but also has bets on the value of something that is continuous. Like real futures contracts!!!!!

    Of course, if you really want to jump into index futures targeting, then choose a target growth path. Divide current by target. Multiply by 100 and subtract 50. It will be trading at 45 or so now.

    Encouraging predicting growth rates is encouraging the _wrongheaded_ policy we have today. 2% expected core CPI inflation from where we happen to be now. That has to go, leaf, _root_, and branch Right Scott!!!!

  13. Gravatar of StatsGuy StatsGuy
    25. June 2010 at 05:52

    Do Intrade contracts with complex formulas attract a lot of trading volume?

    How about something really stupidly simply: Set up a sequence of contracts for 1 year, 2 year, 3 year, 4 year, 5 year.

    If NGDP at end of Y1 is 5% or more greater than NGDP at beginning of Y1, contract pays out. Otherwise it doesn’t. The contract price should represent the market’s estimated probability of the event happening. We won’t get an estimate of NGDP growth – but we do have a target – keeping the price at .50 on the dollar. If we wanted a simple measure of upside/downside risks, offer 2 additional contracts (one with a 4% target and one with a 6% target).

    Simplicity is good.

  14. Gravatar of Doc Merlin Doc Merlin
    25. June 2010 at 06:01

    @Bill,
    Thats roughly what they are doing, only better, as you want a lot more variability than would happen in a linearly scaled one (like the one you propose).
    Also, going to percent changes is the same as just changing the scale every year.

  15. Gravatar of Doc Merlin Doc Merlin
    25. June 2010 at 06:02

    @StatsGuy: The complex formula doesn’t matter that much, if you provide a table that translates it into real world terms.

  16. Gravatar of scott sumner scott sumner
    25. June 2010 at 06:18

    Joe, Because I don’t have the deep pockets the Fed has. This won’t provide the liquidity we really need to make it an effective guide to policy.

    Andrew C. Thanks

    John, Good point, and thanks again.

    JimP, Great article, I’ll do a post later today.

    Carl, Thanks. I’m not sure about the scoring rule, maybe someone else knows.

    John, I am not sure. Does Intrade have a rule against contracts that can turn negative?

    Morgan, The people who initially get the money are those who sell the Fed assets, but then it circulates (or gets hoarded by banks.) Who knows how much? I don’t favor targeting the money supply.

    Bill, We know the current level. So if we also know the expected growth rate, then we know the expected future value.

    Doc Merlin, Yes, we need growth rates to generate interest. And I agree that the variability is enough to cover most expected outcomes. Maybe not all actual outcomes, but almost all expected outcomes.

    Doc Merlin#2, It’s a growth rate over two quarters.

    Alejandro, Thanks for the pep talk.

    Bill#2, My fear is that people think in terms of growth rates. The raw numbers won’t mean anything to the average trader, and hence they will be scared off. people think in terms of NGDP growth rates. We can always back out levels in our own analysis, can’t we?

    Statsguy, Two issues: First, we actually need much more than 5% growth right now, as we are far below trend. And second, I’d really prefer the market to provide a point estimate of the growth rate.

  17. Gravatar of Doc Merlin Doc Merlin
    25. June 2010 at 07:13

    ‘John, I am not sure. Does Intrade have a rule against contracts that can turn negative?’

    Payouts have to be between 0 and 100.

  18. Gravatar of jsalvati jsalvati
    25. June 2010 at 07:13

    @Bill Woolsey
    My first instinct was to have the contracts over levels as well, but it is straight forward to add together the growth rates to get the level (since we are in log space) and level contracts have some drawbacks. First since the contracts must be scaled to between 0-100, level contracts are somewhat tricky to interpret. Second, with current fed policy NGDP is not trend stationary, so the variance of the price would increase with time to completion, meaning you would have to scale it down quite a bit. That said, I am open to arguments. Convince me I should use levels.

    @Doc Merlin
    I do like the idea of providing a table or an online calculator. The role of the previous periods is just the NGDPstart part of NGDPend/NGDPstart.

    @Stats Guy
    Its an interesting idea, but I don’t know that having a target is all that valuable right now; this market won’t be guiding Fed action. We’re mostly interested in gauging the magnitude of monetary disequilibrium.

  19. Gravatar of jsalvati jsalvati
    25. June 2010 at 07:16

    @Doc Merlin
    Providing an online table or calculator makes it considerably more attractive to use the second specification, I think I am leaning towards the second option now.

  20. Gravatar of jsalvati jsalvati
    25. June 2010 at 07:22

    @Carl Lumma
    No, they don’t, but I will be providing constant quotes in the market, so low volume won’t cause problems beyond the problem of low interest by informed traders. But that’s where you guys come in! You will be the informed traders!

  21. Gravatar of Doc Merlin Doc Merlin
    25. June 2010 at 08:13

    Also, since for months afterwards, the GDP number is very heavily revised, you may want to set a date for which revision you will pick.

  22. Gravatar of Morgan Warstler Morgan Warstler
    25. June 2010 at 09:09

    “Morgan, The people who initially get the money are those who sell the Fed assets, but then it circulates (or gets hoarded by banks.) Who knows how much? I don’t favor targeting the money supply.”

    Ok, Scott. I swear I’m not being daft. Who are those people? Ultimately I’m pointing towards… what happens if the banks hoard the assets?

    But I want to see specifically WHY you think this isn’t likely, at their level. Please man, I know you are busy, and I am just one potential convert, but I am not stupid, and you seem to start with “no one doubts the Fed can target NGDP normally” and then when it is time to discuss at zero bound, you say “give them money.” Who’s them, and why are you sure this works?

    —–

    I look at banks and see them not wanting to loan, because deep down, they know prices (commercial and residential) are inflated.

    Seriously, the guys I talk to who are real estate investors are sitting on piles of money they made getting out early. They are making offers to banks sitting on foreclosures at deep discounts to what the banks are asking, and getting nowhere. Banks are are being forced to reduce leverage, BUT they are profitable on reserves, so they are able to ride it out.

    The flip side of this is that a bank WHO KNOWS THEIR OWN BOOK, when you walk in and offer to pay their asking price, what are they to do?

    They know the asset is worth less than they are asking in a true mark-to-market auction environment, so IF say Sumner wins and they stop getting profits on reserves, THEY KNOW that the value of the asset, and the loan is now more likely to be tossed back in their lap.

    So unless you are gold plated, 30% down, or backed by government, THEY’D BE NUTS TO LOAN.

    Frankly, I’m think you need to spend more time talking to the guys with dry powder and ASK THEM why they are sitting things out.

    You are more likely to hear that prices are still too high, than that government crap makes things uncertain (and that’s a BIG one).

  23. Gravatar of Doc Merlin Doc Merlin
    25. June 2010 at 11:44

    @Morgan

    Another way of saying that monetary policy is still effective at the zero bound is this:
    When interest are too low nationally, banks can always use money they get from the fed to buy real assets directly. So, the money still gets into the economy.

  24. Gravatar of caveatbettor caveatbettor
    25. June 2010 at 13:03

    Here’s my note to Intrade from Nov 19 2009. Hope it helps. I promise to trade them!

    Dear Intrade:

    Scott Sumner, an economist and professor, has gained quite a bit of notoriety in his views on monetary policy, and macroeconomic solutions for combating recession.

    In the course of his blogging, the idea came up to create futures contracts that would reveal predictive sentiment for where nominal GDP growth would result. This might have important implications in future monetary policy targeting. And it would be a huge coup for Intrade, in the story of economics, if serious scholars and experts and then a much greater mesh of humanity were able to recognize Intrade as the leader and pioneer in the important intersection between science, policy, and the efficacy of prediction markets.

    The U.S. Department of Commerce’s Bureau of Economic Analysis releases this information at the following URL:

    http://www.bea.gov/national/index.htm#gdp

    I would suggest listing contracts for every quarterly expiry (e.g. 2009q4, 2010q1, 2010q2) that can be found in the Current Dollar and “real” GDP link. Perhaps 5 contracts would cover a good spread. For example:

    US.NGDP.2009.Q4.14000
    US.NGDP.2009.Q4.14200
    US.NGDP.2009.Q4.14400
    US.NGDP.2009.Q4.14600
    US.NGDP.2009.Q4.14800

    where the 5 digit number at the end corresponds to the nominal GDP level in current billions of dollars (in Column F of the spreadsheet linked above). The basic description might read “US Nominal GDP in 4th quarter 2009 to be greater than 14.0 trillion dollars”. Historically, the largest quarter-over-quarter gain in nominal GDP has been 268 billion dollars; the largest loss has been -200 billion.

    I first discussed this idea with Professor Sumner at his blog here (read the comments). Sumner has gained the following of many of the prominent GMU faculty bloggers as well as Greg Mankiw (all of whom refer to Intrade quite a bit), to name a few of the greatest voices in the blogosphere.

    Please let me know if you have further questions.

    Thanks for your consideration,
    Caveat Bettor

  25. Gravatar of Morgan Warstler Morgan Warstler
    25. June 2010 at 15:45

    Ok Doc, this STILL doesn’t answer my question: WHO GETS THE MONEY?

    Seriously, I want to know with some specificity, some firm guidance – which assets are going to get bought and at what prices?

    Meaning, IF the Fed is going to stop paying interest on reserves, so banks start to go hungry, and come out to market with $1T is printed cash to buy assets, BUT THEN screws the banks down, ruthlessly buying the assets in blocks from the worst off banks for pennies on the dollar, screwing the shareholders… and immediately turning around and auctioning the assets off individually to the highest bidders within 90 days, so real dry power guys get their pay day….

    THEN I LOVE SUMNERS PLAN.

    But if instead, the Fed is going to buy assets at bullshit propped up prices and hold those assets for eternity…

    THEN SCOTT IS EVIL AND MUST BE STOPPED.

    —-

    And no free market guy can support Scott until his policy provides this guidance.

  26. Gravatar of John Hall John Hall
    25. June 2010 at 16:07

    @Woosley, Another issue is that you need to make sure to take into account that there are 5 year rebalancing that adjust the whole series of GDP and annual revisions that can adjust quite a bit of past history. Using growth rates avoids some of this issue.

  27. Gravatar of Doc Merlin Doc Merlin
    25. June 2010 at 21:41

    @Morgan
    Got yah. In times past the fed would buy treasury debt. Nowadays the fed buys whatever it feels like, often at inflated prices.
    The fed doesn’t usually try to screw over the banks. As they are the owners and stockholders of the Fed. This is why in this crisis, the fed began paying money to the banks in the form of interest on reserves. This allows them to contract the money supply while subsidizing the banks so they don’t go under, for example. However, it didn’t quite work out, though, for banks outside of NY though. I am not sure if this is due to the odd structure of the Fed wrt NY or if it is due to NY banks being better. (I fear its the former.)

    Anyway, yah, Scotts plan has no formula for what the Fed buys, but the equilibrium for a central bank that pays interest on reserves is for all money to eventually be held at the central bank and for all retail banks to just be service operators.

  28. Gravatar of scott sumner scott sumner
    26. June 2010 at 04:48

    Doc Merlin, Thanks for the info on prices. And that’s a good point about deciding on which revision to use. I’d favor the first announcement, assuming the errors are unbiased.

    Jsalvatier, I agree with your responses.

    Morgan, You said;

    “I look at banks and see them not wanting to loan, because deep down, they know prices (commercial and residential) are inflated.”

    This has nothing to do with monetary policy. Monetary policy is all about cash, not loans.

    Who gets the cash? I don’t know their names, if that’s what you mean. But it would be the people who sell bonds to the Fed. It really makes no difference who gets the cash, as they’d get rid of it quickly in any case.

    If there is a problem with banks hoarding cash (and I doubt there would be) just charge them a interest penalty on reserves. They’ll quickly spend their cash.

    Caveat bettor, Thanks for writing the letter. I am still undecided about the format of the contracts, but would prefer something that gives us a point estimate of expectations.

    Morgan#2, The Fed will buy T-bonds at fair market prices. There will be no attempt to prop up the real estate market by buying dicey MBSs.

    John Hall, I agree, although if they go with the first announcement, then the revised numbers won’t matter either way.

  29. Gravatar of Morgan Warstler Morgan Warstler
    26. June 2010 at 04:59

    Finally! Thank you. Ok, then that would be the point of failure in Scott’s plan.

    He encodes the moral hazard of the banks being aided to survive profitably into the the very fabric of economic policy.

    I guess he could get specific, describing in political terms HOW the money comes to bear in our economy.

    To be successful:

    1. It can’t be used for political purposes. The need for Scott’s plan can justify public employees getting a raise, or stimulus or some other thing. There is a guy named Karl Denninger, who sounds totally rational (like Scott) until you finally get down to brass tacks… he wants the US to stop paying interest on debt, and ONLY print money when the government goes into debt.

    2. It’s function needs to be liquidity of bad assets. This recognizes that the natural force of the business cycle, the failure and liquidation of the overextended at rock bottom prices to the winners who go out early.

  30. Gravatar of Morgan Warstler Morgan Warstler
    26. June 2010 at 05:08

    Scott, you are talking about the Fed buying assets whether from the bank or the government.

    Either is wrong as it distorts the market value of those assets, and keeps the negative downward froce from happening.

    If you showed up at a Super Bowl weekend with $1T to blow on hookers and drugs, it would negatively impact the good times had by all legitimate buyers – the only happy people are the hookers and drug dealers.

    Your plan has to serve the opposite function, bring in even more hookers and drugs and sell them at deep discounts.

    You get the drift, MY POLICY, you should have no problem with\, in fact you even try to do it elsewhere: by not paying interest on reserves.

    Embrace your inner Tea Partier Scott, and win over the libertarians. Use cash to UNWIND banks, use it to reward the guys with dry powder.

  31. Gravatar of Indy Indy
    26. June 2010 at 05:43

    So, let’s say you were a quant working at an investment bank or hedge fund tasked with developing a program which tries to play this futures market as well as possible.

    So you would develop a basic economic model and try to run robust and rigorous historical correlations with every available piece of data (to include FOMC statments and Fed Board Member personality profile indices) to see which are the most predictive of future pathways based on theory and empirical evidence. The biggest uncertainty in playing such a market in a severe downturn would be political-intervention risk, I’m guessing.

    In other words – you would developed “NGDP leading indicators”, along with a little “Fed Decision Prediciton”. But we already have some prominent leading-indicator groups, like ECRI, and fed prediction groups like Macroeconomic Advisers. How to get the edge over what these guys already do?

    As for me, yeah, I’ll buy a few contracts. I’m thinking Walmart price decisions on a basket of brand-name consumer goods. are probably as good a leading indicator as any. The amount of economic information they aggregate in order to make their decisions is probably a wave we can all ride on.

    Right now, local canned soup prices are at an all-time low, and who knows what that really means, but I’m interpreting it as telling me that I shouldn’t be very optimistic about the next few quarters.

  32. Gravatar of scott sumner scott sumner
    26. June 2010 at 06:48

    Morgan, You said;

    “Scott, you are talking about the Fed buying assets whether from the bank or the government.”

    Neither. From private individuals.

    Indy, There are two responses there:

    1. If you don’t have anything to offer in the way of insights about where NGDP is going, then you don’t have to play the market.

    2. Those groups you refer to may not be playing the market, so perhaps you can beat those people who are actually in the NGDP market.

  33. Gravatar of Morgan Warstler Morgan Warstler
    26. June 2010 at 07:54

    “I’m thinking Walmart price decisions on a basket of brand-name consumer goods. are probably as good a leading indicator as any. The amount of economic information they aggregate in order to make their decisions is probably a wave we can all ride on.”

    +1 Man o’man WHY hasn’t someone done this? It is just an ecommerce shopping app turned on its ear. It could cover WalMart and Amazon and do 1000′s of products. How cool, a Google Ads statistics site that becomes the de facto stats we use when we talk about inflation in consumables.

    Scott, is this a trick to make me keep writing long posts trying to get at something?

    you say “From private individuals” – so you want the Fed to buy a bunch of shit on Etsy?

    Please give me an example of the the Fed targeting NGDP by buying assets from individuals. And again, is the Fed trying to get GOOD DEALS when it is buying?

    I’d argue that the Fed HAS TO try and get really good deals when it buys things, because then it acquired the assets low enough to justify selling them at below market cost, so they are highly liquid and continue to generate economic activity. This gets the Fed back its capital to go find another a deal.

    If the Fed only loses 20% per cycle, the money imapact >1

    No, when the guy printing money comes to town with $1T in new dollars, he needs to have his eye on Volume Pricing, he needs to get better deals than anybody, he’s the biggest buyer out there.

    This places the Fed up there like WalMart – able to generate the most economic return, and most even field on moral hazard all buyers in the market are more likely to have the opportunity.

  34. Gravatar of Morgan Warstler Morgan Warstler
    26. June 2010 at 07:56

    excuse language pls, I forgot.

  35. Gravatar of John Hall John Hall
    26. June 2010 at 12:40

    Scott,
    You said “if they go with the first announcement, then the revised numbers won’t matter either way.” The last 5 year revision was 7/30/2009. If this was released at the same time as the advance estimate (I’m like 99% sure that it was since the advance estimate normally comes one month after the end of the quarter), then there’s a problem.

    The point Woolsey was making was in regards futures contracts based on levels. So imagine that you’re an economist in the forecasting business and you predict nominal GDP will grow y% QoQ annualized in the next quarter. You can then translate this into some $xxx level forecast and bet on the futures contract. However, if the annual or 5 year revision occurs at the same time as the advance estimate, then your level forecast could be completely off, even if the percentage growth estimate is correct. The economist has to not just bet on the growth rate, but also on how the old data will be revised.

    I view this as THE reason not to do level futures contracts for GDP.

  36. Gravatar of Doc Merlin Doc Merlin
    27. June 2010 at 03:21

    @Scott
    I believe there is a strong downward bias on the revised estimates for each quarter. I haven’t checked it econometrically (these thoughts are not rigorous), but I have only heard it revised upwards twice and have heard it revised downwards a number of times.

    @John Hall:

    As much as levels and level targeting matters, I agree, for NGDP it won’t work in a future contract for the reasons you describe.

  37. Gravatar of scott sumner scott sumner
    27. June 2010 at 06:48

    Morgan, The Fed buys T-securities at the current market price. They don’t get good or bad deals, they get the same deals as any other bond buyer.

    John Hall, I see your point. But when the Fed makes an announcement, they say GDP grew X%, period. They don’t say “it grew X% percent based on the old number, and Y% based on the revised base.” They give one percentage, and that’s the one that would be used in the contracts, unless I am mistaken. Now you can argue that’s not the number we want, but I think it is, because we can’t anticipate revisions anyway. So ex post it may be biased, but ex ante it’s the number we ought to be trying to forecast.

    Doc Merlin, I have my doubts, but even if you are right then market traders could take that bias into account when making their forecast, and researchers could take it into account when evaluating expectations.

  38. Gravatar of Doc Merlin Doc Merlin
    27. June 2010 at 06:56

    “Morgan, The Fed buys T-securities at the current market price. They don’t get good or bad deals, they get the same deals as any other bond buyer.”

    The rumor that is floating around is that recently they have been offering private deals at above market value for housing securities.

    They also are the 900 lb gorilla in the market and have been dominating it. According to this article they (at the time it was written) had bought an equivalent of 80% of the securities issued this year.
    http://online.wsj.com/article/SB125373822753135165.html
    ‘So far, the Fed’s net buying of mortgage-backed securities totals nearly $850 billion, or about 80% of securities issued this year by entities like Fannie Mae and Freddie Mac’.

  39. Gravatar of Doc Merlin Doc Merlin
    27. June 2010 at 06:57

    Sorry, that should say ‘issued last year.’

  40. Gravatar of Morgan Warstler Morgan Warstler
    27. June 2010 at 07:56

    Doc,

    Scott is all over the map here. He either has not got a grasp of it or doesn’t like where the conversation leads.

    Scott, the Fed is NOT just buying T-Bills, and you haven’t given ANY explanation that you KNOW FOR SURE the exact mechanism of HOW they are going to target NGDP.

    You cannot wave a magic wand and say “target NGDP to 5%,” that is not a plan. That is a theory.

    You have to say the exact actions that the Fed will take, and then deal with the arguments about those effects.

    IF the Fed is currently buying 80% of F&F backed securities, THEN they ARE propping up housing prices WHICH YOU SAY YOU ARE AGAINST.

    Beyond that, when demand is higher for T-BILLS, it keeps the price on them down. You aren’t dealing with this either.

    —–

    Finally, I think I’m right to revise and extend your NGDP theory:

    If we grant that to target 5% NGDP the Fed should be printing money and putting it into the market, then they should operate as Wal-Mart does: screwing down sellers of assets to get best prices, and immediately liquidating those assets to retail buyers at still deep discount prices.

    And since they can buy anything, not just T-Bills. And since we BOTH want to stop paying interest on reserves.

    This put the Fed on the side of Main Street, it pays no attention to price (as you insist), but is does attempt to achieve the MOST VALUE for its purchase, which allows it to resell said assets quickly and efficiently at low prices.

  41. Gravatar of caveat bettor caveat bettor
    28. June 2010 at 03:55

    Scott: You and John should design the contract the way you see fit. The design I proposed is in a style that both Intrade (for management purposes) and Intrade traders (for liquidity purposes) are familiar.

    I look forward to trading them.

  42. Gravatar of caveat bettor caveat bettor
    28. June 2010 at 03:57

    Also, you should be very specific about expiry conditions (e.g. which version of the BEA announcement, and where it can be found). If expiry is controversial, it will be a big hit to credibility (as Intrade discovered when expiring the prediction contracts on North Korean testing of ballistic missiles).

  43. Gravatar of scott sumner scott sumner
    28. June 2010 at 06:59

    Doc Merlin, Morgan wasn’t asking about the current crisis, but how monetary policy works in general. I am advocating the purchase of T-securities. The Fed doesn’t dominate that market.

    Morgan, Read what I say, not what you think I said. You claimed:

    “IF the Fed is currently buying 80% of F&F backed securities, THEN they ARE propping up housing prices WHICH YOU SAY YOU ARE AGAINST.”

    Where did I recommend the Fed buy MBSs?

    As for the Walmart comparison, why would sellers sell to the Fed for a lower price than they could get in the bond market? I don’t follow your logic at all.

    caveat bettor. That is a very good point. I think they should rely on the official BEA statement, and go with a specific release, such as the first estimate.

  44. Gravatar of Morgan Warstler Morgan Warstler
    28. June 2010 at 07:54

    Scott, so I am clear I am asking about the current zero bound crisis – and in it apparently, you are ONLY in favor of more Fed printing/buying – if they buy treasuries? nothing else?

    For a fair discussion, you need to nail yourself down to the floor and state specifically WHAT you will accept the Fed buying and not buying. It seems like you are squirreling here. Be specific in debate, it makes you noble.

    Also, when the Fed is buying T-Bills, that does drive down rates, right?

    Or do you mean, not just buying them at new auction, but also bidding up price to buy them from current holders.

  45. Gravatar of scott sumner scott sumner
    29. June 2010 at 07:44

    Morgan, Buying T-bills may or may not drive down rates. If they bought enough to create hyperinflation, it would raise interest rates and lower bond prices.

    I prefer they buy Treasuries, but agency debt (which is now effectively Treasuries) is better than nothing.

    Am I noble yet?

  46. Gravatar of jj jj
    29. June 2010 at 09:25

    morgan, the fed should buy only in deep and liquid markets. They won’t be negotiating deals, it’s more like anonymous stock purchases. They should buy as broad a range of assets as possible, to avoid propping any one asset up. If they decided to only buy IBM stock, for example, then IBM would be artificially inflated even if it was going bankrupt. But by buying everything, then any market segment could still fall relative to the others.

    The fed wouldn’t buy MBS unless that market was liquid. It shouldn’t try to guess at the ‘true’ value of anything, just buy and sell at market rates.

    Here are some things the fed could buy with newly printed money:
    -new t-bonds
    -existing t-bonds
    -any other bonds, offered anywhere in the world
    -pork belly futures
    -an index comprised of every stock exchange in the world

  47. Gravatar of ssumner ssumner
    30. June 2010 at 05:42

    jj, I agree, although I’d prefer they buy T-bonds.

  48. Gravatar of jj jj
    30. June 2010 at 06:58

    right — I was remembering a discussion from many months ago. I believe the argument went like this:

    As the fed buys t-bonds, their interest rate will fall to zero. Therefore the newly created money can equivalently sit in bank accounts earning 0% interest (or even worse earn interest as excess reserves!), and the money supply is essentially unchanged. Eventually the fed will have bought up all t-bonds without accomplishing anything. At that point the fed needs to find something else to buy.

    Of course this rests on the unlikely assumption that even as the fed embarks on an $8 trillion buying spree, the market still won’t expect any inflation.

  49. Gravatar of Scott Sumner Scott Sumner
    1. July 2010 at 10:45

    jj, I agree, they wouldn’t have to buy many before prices started rising, if they had a higher inflation target. Indeed they might have to sell bonds. And of course rates wouldn’t fall to zero on longer term bonds.

  50. Gravatar of A Better Policy Than Tight Money | FrumForum A Better Policy Than Tight Money | FrumForum
    21. September 2011 at 11:11

    [...] has other interesting ideas, such as the Fed issuing nominal GDP futures contracts and having those bought and sold as a way for the market to control the money supply. It’s an [...]

  51. Gravatar of TheMoneyIllusion » Fed fail: The implosion of a policy regime TheMoneyIllusion » Fed fail: The implosion of a policy regime
    24. July 2012 at 02:39

    [...] are our only hope: However, one form of nominal GDP targeting seems to sidestep these problems:nominal GDP futures targeting. This would allow market participants to instantly improve estimates of future inflation by bidding [...]

Leave a Reply