Cochrane on fiscal dominance (never reason from a price change, example #656)

John Cochrane is worried about the fiscal cost of higher interest rates:

But this comforting thought leaves out a vital consideration: Monetary policy depends on fiscal policy in an era of large debts and deficits. Suppose that the Fed raises interest rates to 5% over the next few years. This is a reversion to normal, not a big tightening. Yet with $18 trillion of debt outstanding, the federal government will have to pay $900 billion more in annual interest.

Will Congress and the public really agree to spend $900 billion a year for monetary tightening? Or will Congress simply command the Fed to keep down interest payments, as it did after World War II, reasoning that “Fed independence” isn’t worth that huge sum of money?

This additional expenditure would double the deficit, which tempts a tipping point.

I can’t follow the math.  The $900 billion seems to come from multiplying 5% times $18 trillion, but neither number should be used in the calculation.  If rates rose to 5% then the Treasury’s interest cost would rise by less that 5% (very little for 30 year bonds, and up 3% for 10 year bonds.)

Put aside the fact that it’s very unlikely that T-bills will yield 5% in the foreseeable future.  I’d be pleasantly surprised if they reached 3% before the next recession.  Recall that we’ve never gone more than 10 years without a recession, and December 2017 is the 10 year anniversary of this one.

And isn’t the net debt around $12 trillion billion?  So even in the unlikely event that T-securities pay 5% on average before the next recession, it’s more likely an extra cost of $360 to $480 billion.

I also disagree with the reasoning.  Cochrane talks like the Fed sets the interest rates on Treasury securities.  It doesn’t, it targets the Fed funds rate.  A tight money policy that raises fed funds rates can easily lower longer term bond yields, and vice versa.  Bond yields reflect the macro economy.

Suppose in 2006 Cochrane had said; “If the Fed cuts rate to zero over the next few years there will be a huge windfall to the Treasury, boosting the budget surplus.”  I would have assumed that that sort of interest rate decline reflected a deep and prolonged recession, not “easy money.”  Hence I would have expected the budget deficit to have ballooned—which is of course what happened.

If a crystal ball told me that rates would be 5% in 2016 I’d be jumping for joy.  Yes!  A robust recovery is finally about to begin!  Tax revenues will soar—less spending on unemployment compensation and food stamps.  Etc. etc.  The federal deficit will finally start shrinking.  Alas, that’s not likely to occur.  I’m afraid we are stuck with low interest rates for a very long time.  Not as low as Japan, but much lower than we saw in the 20th century.

What to do? First, the Treasury and Fed need a new “accord” to decide who is in charge of interest-rate risk, most likely the Treasury, and then grant it clear legal authority to manage that risk. The Fed should then swap its portfolio of long-term bonds for a portfolio of short-term Treasuries and forswear meddling in the maturity structure again.

We tried that policy before, and had to abandon it in 1951 because of rising inflation.

Second, the Treasury should seize its once-in-a lifetime opportunity to go long. Thirty-year interest rates are at 2.8%, a 60-year low. Many corporations and homeowners are borrowing long to lock in low funding costs. So should the Treasury.

People have been saying that for years, and they’ve been wrong.  On the other hand I don’t really object; Cochrane might be right this time, and the EMH says it’s pretty much a coin toss anyway.  Speaking of the EMH, that’s one reason I expect interest rates to remain low.  But not the only reason.  History is also on my side.  When countries hit the zero bound they tend to say close to that level for a long time.

Instead of having the Treasury instruct the Fed to target NGDP, let’s have the Fed target the price of NGDP futures contracts.  That creates a stable macro environment.  Then it’s up to Congress to live within our means (i.e keep the debt ratio manageable.)  I have confidence they will do so, as they’ve always done in the past.

PS.  Some would contest my claim that the EMH applies to the current T-securities market—as the Fed has bought up all the debt.  Not so. David Beckworth showed that today’s Fed holds roughly the same small share of T-debt they held before the recession.


Tags:

 
 
 

36 Responses to “Cochrane on fiscal dominance (never reason from a price change, example #656)”

  1. Gravatar of Doug M Doug M
    4. March 2013 at 12:41

    “I can’t follow the math. The $900 billion seems to come from multiplying 5% times $18 trillion, but neither number should be used in the calculation. If rates rose to 5% then the Treasury’s interest cost would rise by less that 5% (very little for 30 year bonds, and up 3% for 10 year bonds.)”

    Cochraine points out that the bulk of the US debt is short, so even with some 10 and 30 year debt, most of the debt if refinanced a the current market rate, and it would only take a short time before that was reflected in the average costs.

    “And isn’t the net debt around $12 billion?”

    Trillion, but I get your point.

    Less than that even. It partly depends on whether you include “goverment related” debt, i.e. Fannie Mae and Freddie Mac. If you want to inflate the debt number add in the obligations of Social Security. Personally, my read of SS, is that the benefit is not guaranteed, and should not be counted.

    “If a crystal ball told me that rates would be 5% in 2016 I’d be jumping for joy. Yes! A robust recovery is finally about to begin!”

    Well underway to get to 5%.

  2. Gravatar of Vivian Darkbloom Vivian Darkbloom
    4. March 2013 at 13:17

    Cochrane reproduced the CBO’s latest projection that indicates the net interest expense in 2023 will be $857 billion; I think he’s rounded this up to $900 billion and the “$900 billion more” refers to the excess over the primary deficit in that year, but I agree, the reference is not entirely clear. If you take it to mean the excess over the current (FY 2012) net interest expense the 2023 figure is only a $634 billion increase. Not $900 billion; but not a small number, either.

  3. Gravatar of foosion foosion
    4. March 2013 at 13:30

    At best, Cochrane provides a reason why the Fed shouldn’t try to raise interest rates in the current economy. That’s much better than most of his thoughts on macro.

    As you say, 5% rates would mean a much stronger economy, with more revenue and lower spending by the Federal govt.

    How do you determine the size of a manageable debt ratio? Perhaps when govt borrowing is crowding out private borrowing?

  4. Gravatar of TravisV TravisV
    4. March 2013 at 13:52

    Dear Commenters,

    Please see this awesome post by Prof. Sumner on Iceland:

    https://www.themoneyillusion.com/?p=8702

    Could someone point me to good resources that support this story? In particular, statistics that show collapsing financial sector employment and booming technology / manufacturing employment within a short period of time?

    Remarkably, Iceland’s president said that exactly that process happened “within six months”: http://read.bi/VfRcWV

    I’m sure there’s something to this story. It would just be nice to see data that substantiates it.

  5. Gravatar of Geoff Geoff
    4. March 2013 at 13:54

    Dr. Sumner:

    “I also disagree with the reasoning. Cochrane talks like the Fed sets the interest rates on Treasury securities. It doesn’t, it targets the Fed funds rate. A tight money policy that raises fed funds rates can easily lower longer term bond yields, and vice versa. Bond yields reflect the macro economy.”

    http://research.stlouisfed.org/fredgraph.png?g=gaW

    A tight monetary policy that raises rates?

    How about simply raising rates without any tight or loose caveat attached?

    Cochrane is probably not saying that the Fed sets the rates on government debt directly. I think he is taking into account the empirical fact, as seen in the chart, that government interest rates are positively correlated with the fed funds rate, and he is reasoning that if the fed funds rate rises, then it is likely that government bonds rates of all maturities will rise as well.

    I notice that you tend to look at extremely short term, essentially same day fluctuations in government bond rates when the Fed changes monetary policy or signals changes via announcements. That’s all well and good, but then you tend to ignore the subsequent movements in rates that should not be ignored, as monetary policy takes time.

    When government bonds rate changes are considered over periods of more than one day, then we see a very close correlation, as seen in the above chart.

    Just look at how close the relationship is during upticks in the fed funds rate specifically, which is the most pertinent movement in the fed funds rate that is associated with this discussion. The 30 year bond rate change is very highly positively correlated with fed funds rate changes when the fed funds rate is raised.

  6. Gravatar of Geoff Geoff
    4. March 2013 at 13:57

    Dr. Sumner:

    “If a crystal ball told me that rates would be 5% in 2016 I’d be jumping for joy. Yes! A robust recovery is finally about to begin!”

    Is that not reasoning from interest rates, which is reasoning from prices?

  7. Gravatar of Vivian Darkbloom Vivian Darkbloom
    4. March 2013 at 14:06

    The CBO report to which Cochrane refers assumes an uptick in the 10 Yr Treasury to 5.2 percent (from 1.8 percent) and in the 3 month T Bill from 0.1 percent to 4 percent by 2023. Again, I think Cochrane is referring to the 2023 net interest expense (over the primary deficit). I think he’s assuming a public debt of $18 trillion for that year (up from nearly $12 trillion now), which is slightly less than the CBO projection. That public debt number seems low, but who knows for sure?

  8. Gravatar of Geoff Geoff
    4. March 2013 at 14:15

    Reason #756 why I am bearish on China (and not giving any time frames so as to piss MMs off, haha!):

    http://www.cbsnews.com/video/watch/?id=50142079n

  9. Gravatar of ssumner ssumner
    4. March 2013 at 14:16

    Doug, You said;

    “Cochrane points out that the bulk of the US debt is short, so even with some 10 and 30 year debt, most of the debt if refinanced a the current market rate, and it would only take a short time before that was reflected in the average costs.”

    That’s true, but doesn’t address my point. The increase in costs is the difference between current interest costs (which are substantial) and the interest cost at 5% market rates. You claim that at 5% market rates the interest cost would be close to 5%, say 4.7%, because most debt would have been rolled over. That’s fine, but you aren’t starting from zero. The current interest cost is well above zero.

    Vivian, If you are right then I’m right. The increase would be around $400 billion for the net debt.

    Geoff, You said;

    “I think he is taking into account the empirical fact, as seen in the chart, that government interest rates are positively correlated with the fed funds rate, and he is reasoning that if the fed funds rate rises, then it is likely that government bonds rates of all maturities will rise as well.”

    If so, he’s using very bad logic. Over the medium term rates are endogenous, as the Fed targets inflation over the medium term, not interest rates. He’s a smart guy, so I doubt he made that assumption.

    Yes, I reasoned from a price change, but in a much more sensible way than Cochrane, as I actually thought about what the price change most likely implied about the state of the economy.

  10. Gravatar of ssumner ssumner
    4. March 2013 at 14:18

    Vivian, OK, Now I see you point. So it wasn’t quite as inaccurate as I indicated. But it was inaccurate in another way. One should never talk about 2023 dollars as if you are talking about current dollrs.

  11. Gravatar of Vivian Darkbloom Vivian Darkbloom
    4. March 2013 at 14:21

    Scott,

    It’s good you amended your comment before I pulled the trigger on this one.

    Per the Table reproduced by Cochrane, the FY net interest expense for FY 2012 was $223 billion. The expected expense (per CBO) is $857 for 2023. That’s a delta of $634 billion; not $400 billion.

  12. Gravatar of Vivian Darkbloom Vivian Darkbloom
    4. March 2013 at 14:23

    Or, to put it differently, not $360 to $480 billion as per your original post.

  13. Gravatar of Geoff Geoff
    4. March 2013 at 14:41

    Dr. Sumner:

    “If so, he’s using very bad logic. Over the medium term rates are endogenous, as the Fed targets inflation over the medium term, not interest rates. He’s a smart guy, so I doubt he made that assumption.”

    I wasn’t saying he made that assumption, but that he made the assumption that almost 40 years of historical data that shows a close correlation between the fed funds rate and government bonds rates, would likely have the same close correlation going forward, such that if the fed funds rate increases, so too would government bonds rates increase.

    Do you think that the next increasing trend in the fed funds rate will have an effect on government bonds rates that is different from what almost 40 years of data suggests? I suppose that’s possible. But then why?

    “Yes, I reasoned from a price change, but in a much more sensible way than Cochrane, as I actually thought about what the price change most likely implied about the state of the economy.”

    [Rolls eyes].

  14. Gravatar of Vivian Darkbloom Vivian Darkbloom
    4. March 2013 at 14:51

    “One should never talk about 2023 dollars as if you are talking about current dollrs.”

    Agreed, but did Cochrane really lead you to believe that when he referred to a *future* interest expense he was talking about *current* dollars? I don’t think many readers, much less a reader with a PhD in economics, would have been mislead by that.

  15. Gravatar of Brian Donohue Brian Donohue
    4. March 2013 at 15:24

    Scott,

    Why net debt? Isn’t the difference mostly the Social Security surplus?

    If you ignore this, aren’t you basically saying that a regressive payroll tax was diverted to fund general government operations over the past two decades? I think this is the view a lot of people are implicitly taking. As far as taking views goes, this one seems better implicit than explicit, in terms of being able to sleep at night.

  16. Gravatar of Bernard King Bernard King
    4. March 2013 at 15:30

    “History is also on my side. When countries hit the zero bound they tend to say close to that level for a long time.”

    Aside from the present crisis, and Japan’s ongoing dilemma since the 90’s, what other examples are there of a country hitting the zero lower bound?

  17. Gravatar of marcus nunes marcus nunes
    4. March 2013 at 15:55

    @TravisV
    This post describes the Icelandic “Fiscal Cliff”:
    http://thefaintofheart.wordpress.com/2012/12/05/another-fiscal-cliff-the-case-of-iceland/

  18. Gravatar of Michael Michael
    4. March 2013 at 17:22

    “If you ignore this, aren’t you basically saying that a regressive payroll tax was diverted to fund general government operations over the past two decades?”

    Wasn’t it? How can anyone view it otherwise?

  19. Gravatar of Steve Steve
    4. March 2013 at 19:03

    Cochrane is a smart guy. He should check his eyeglass prescription so that he can distinguish the ‘i’ from the ‘a’ in microeconomics.

  20. Gravatar of Benjamin Cole Benjamin Cole
    4. March 2013 at 21:53

    Yeah, I wonder about the “reversion to the norm”

    Why is it the same people who quickly say that low economic growth is the new norm, then say that interest rates must go back to old norms?

    Add on: I don’t care what it costs, we have to get GDP growth back on growth trend to full potential.

    If you have the flu, you don’t eat much. Boy, you are saving money. Just try to get various strains of the flu to hold down costs.

    Is this the way economists think?

  21. Gravatar of Cameron Cameron
    4. March 2013 at 22:31

    “Recall that we’ve never gone more than 10 years without a recession, and December 2017 is the 10 year anniversary of this one.”

    Okay, but since most recessions are caused by the Fed trying to push down inflation and inflation will be pretty low over the next 5-10 years I suspect the probability of a recession will be lower than normal.

    Don’t mind me, just trying to inject some optimism as your blog has been (appropriately) depressing to read as of late.

  22. Gravatar of Cochrane on fiscal dominance (never reason from a price change, example #656) | Fifth Estate Cochrane on fiscal dominance (never reason from a price change, example #656) | Fifth Estate
    5. March 2013 at 03:44

    […] See full story on themoneyillusion.com […]

  23. Gravatar of “Ice” & “Ire” – Supply & Demand Shocks | Historinhas “Ice” & “Ire” – Supply & Demand Shocks | Historinhas
    5. March 2013 at 03:59

    […] at the comment section of this Scott Sumner post, TravisV recalls a 25 month old Scott post on […]

  24. Gravatar of ssumner ssumner
    5. March 2013 at 05:48

    Geoff, You said;

    “I wasn’t saying he made that assumption, but that he made the assumption that almost 40 years of historical data that shows a close correlation between the fed funds rate and government bonds rates, would likely have the same close correlation going forward, such that if the fed funds rate increases, so too would government bonds rates increase.”

    Correlation doesn’t prove causality. Lucas Critique.

    Vivian, So if by “future” he’d meant 1000 years out, that would have been OK? Seriously, when going all the way out to 2023 you need to say so. Most people would have assumed he meant when the economy recovered from this recession.

    Brian, I am using the standard accounting. Social Security is not a legal debt obligation. I admit that if you add unfunded entitlements everything changes. But in that case Cochrane’s number for the national debt would be wildly off base. I was following his assumption.

    Bernard King, The US during 1932-51.

    Cameron, I agree, I wouldn’t be shocked if this “expansion” lasted longer than usual.

  25. Gravatar of George George
    5. March 2013 at 05:51

    TravisV,

    “Dear Commenters,

    Please see this awesome post by Prof. Sumner on Iceland:”

    In the post, he writes,

    “In other words, monetary policy did not slow the process of recalculation in Iceland.”

    This is a non sequitur. Nothing that was said prior justifies the use of “In other words” here.

    Slowing the process of recalculation cannot be observed through temporal trends in output. Recalculation is a counter-factual concept. The proper use of this concept would be to ask whether or not and to what extent did monetary policy in Iceland prevent unsound businesses from being liquidated. This of course would mean we should see a smaller output decline in Iceland, during and after the crisis. And that’s just what we saw. Output did indeed fall by less in Iceland, which is perfectly consistent with the theory that inflation hampers recalculation.

  26. Gravatar of TravisV TravisV
    5. March 2013 at 05:57

    Marcus,

    I just read your post on Iceland:

    http://thefaintofheart.wordpress.com/2013/03/05/ice-ire-supply-demand-shocks

    Thanks a lot, man!!!

  27. Gravatar of “Ice” & “Ire” – Supply & Demand Shocks | Fifth Estate “Ice” & “Ire” – Supply & Demand Shocks | Fifth Estate
    5. March 2013 at 06:01

    […] at the comment section ofthis Scott Sumner post, TravisV recalls a 25 month old Scottpost on […]

  28. Gravatar of TravisV TravisV
    5. March 2013 at 06:08

    George,

    I think Prof. Sumner is saying that Iceland’s sectors were able to adjust very rapidly. Thousands of workers left the financial sector and found more-productive jobs in other sectors like tech and manufacturing very quickly.

    In Ireland, that adjustment is happening much much much much more slowly.

  29. Gravatar of Vivian Darkbloom Vivian Darkbloom
    5. March 2013 at 06:22

    “Seriously, when going all the way out to 2023 you need to say so.”

    This is what Cochrane wrote, immediately under the 10 year CBO table from the report he’s working from (also showing the 10 year time frame):

    “By 2023, the CBO thinks interest payments on the debt will be $857 billion, essentially the entire deficit!”

    Seriously.

  30. Gravatar of ssumner ssumner
    5. March 2013 at 08:55

    Vivian, But that table is no where near the sentence I quote. Why should we be impressed with $900 billion? Is it supposed to be a big number? If so, why?

  31. Gravatar of Geoff Geoff
    5. March 2013 at 09:01

    Dr. Sumner:

    “Vivian, But that table is no where near the sentence I quote.”

    I’m sorry but that’s just blaming others for one’s laziness.

  32. Gravatar of Negation of Ideology Negation of Ideology
    6. March 2013 at 03:50

    Brian –

    I know I’m late to this, but the reason to use the net debt (the Treasury calls it the debt to the public) is because this post was about the effect of higher interest rates on the budget. Only the net debt matters in that context. If the Treasury pays higher interest to the trust fund it holds for Social Security, then it gains exactly the same amount of revenue that it loses.

    And the debt owed to the Federal Reserve counts as part of the public debt. Since the Fed pays that back as a dividend to the Treasury, whatever portion of increased interest goes to the Fed also has no effect on the public fisc.

  33. Gravatar of ssumner ssumner
    6. March 2013 at 06:24

    Geoff, Laziness? The table she cites shows Cochrane to be completely wrong—why should I have assume the quote applied to that table, in which case it would be wrong, and not 5% times 18 trillion, which also equals $900 billion.

    You post a lot here, but if you don’t start thinking before you post I’m going to have to give you the Major Freeman treatment. You are very close to the edge.

  34. Gravatar of Vivian Darkbloom Vivian Darkbloom
    6. March 2013 at 08:04

    —“Why net debt?”

    —“If the Treasury pays higher interest to the trust fund it holds for Social Security, then it gains exactly the same amount of revenue that it loses.” (etc)

    I think this answer (and the original posts) confuses the issue.

    The budget term of art is “public debt” and not “net debt”. Cochrane originally referred (ambiguously) to merely “debt” of $18 trillion and Scott orginally introduced “net debt” in the following sentence—“And isn’t the net debt around $12 trillion billion? “— but he (Scott) probably meant “public debt”.

    The report that Cochrane refers to as prepared by the CBO last month refers to “net interest”. That is, the Treasury does not simply pay interest to the public–some government entities also earn interest from the public (not that much in the larger scheme). The difference between the interest paid to the public and the interest earned by government from the public is “net interest”””a term that actually appears in that CBO report.

    Amounts owed to the Social Security Trust Fund are not part of the “public debt” and interest credited to the Trust Fund is not part of the interest on the “public debt”. But, that aside, are credits to the Trust Fund of interest a “wash” for the overall budget? Not always. If Social Security is at break even (they collect current premiums exactly equal to current premiums) it is basically a wash. There is a debit from Treasury and a credit to the Trust Fund for accrued interest. But, when current premiums are lower than current benefits, Treasury needs to go to the market and borrow because actual funds are being disbursed in the form of benefits. This creates “public debt”. If interest rates rise, the net interest cost to government will rise to the extent the Treasury needs to keep the Trust Fund cash flow solvent. In other words, to the extent Social Security needs to redeem the principal of its IOU’s with Treasury to satisfy current payments, the public borrowing and associated interest costs will increase. While in “normal times” it may be sufficient merely to look at the “public debt” one must also, particularly now, keep one’s eye on the need for actually borrowing to occur in order that Treasury can make up the difference between the amount of current premiums SS collects and the current benefits it must pay. The Social Security program recently crossed the threshold whereby it pays more in benefits than it collects. For our borrowing needs, this is like the Atlantic current changing directions.

    The CBO report to which Cochrane refers actually pegs the “public debt” at about $19.944 trillion in 2023—more than the $18 trillion Cochrane cites; but, see below for a possible explanation. This is the debt owed to entities outside government (excluding the Social Security and other government trust funds).

    Currently, the “public debt” is $11.280 trillion (this presumably what Scott intended to refer to). Of interest and concern to monetary economists should be the fact that the debt held by the Federal Reserve is considered “public debt” and the interest paid to the Federal Reserve is considered part of the “net interest expense”. I believe that while the Federal Reserve transfers its “earnings” back to Treasury, this amount is booked as general revenues and not as “interest income” that would, if so booked, reduce the overall “net interest expense” on the “public debt”. This is an important accounting convention. Of the current $11.822 trillion in outstanding “public debt”, $1.745 trillion is held by the Fed. This accounting seems to inflate current interest expense and deflate the primary deficit (another term of art—the deficit excluding net interest expense) with a neutral effect on the overall deficit over the short term.

    Cochrane wrote:

    “Suppose that the Fed raises interest rates to 5% over the next few years. This is a reversion to normal, not a big tightening. Yet with $18 trillion of debt outstanding, the federal government will have to pay $900 billion more in annual interest.”

    I’m guessing here, but it’s possible that, while the Fed holdings are officially part of the “public debt”, Cochrane has subtracted Fed holdings from the $19.944 figure to get his rounded $18 trillion (his prior paragraph assumed no Fed liquidations). I agree that he could explain his numbers better even though his estimate of $900 billion in future interest expense is close to what the CBO projects for 2023 ($857 billion) even though, as Scott says, it’s a blog, not an academic paper.

    How does the CBO handle the future liquidation of the Fed’s holdings? This is discussed at page 22 of that February CBO report. The CBO assumes that remittances from the Fed to Treasury will continue to grow through 2015 as the Treasury reduces its portfolio.

    After that, through 2021, the CBO assumes the Fed will start to liquidate its portfolio (not only of Treasuries it holds, but also the MBS’s—total will reach well over $3 trillion in 2013) and the remittances to Treasury will be reduced, perhaps to zero from the 2012 amount of $90 billion. Importantly, they assume no major gains or losses on that portfolio””an assumption that, if wrong, would have major effects on the budget. Assuming no gain or loss on the portfolio (and no net interest increase on the “refinancing”) the net effect should be that Treasury now pays interest on public debt with no offsetting revenue item. Economically, although perhaps not readily evident from the accounting treatment, this should have the effect of increasing the burden of the interest expense on the “public debt” because a larger portion of that debt will no longer be generating a revenue offset. This is an effect that Cochrane does not take into account, because his piece addressed only the increase in interest rates and not the composition of the holders of that public debt because he assumes the holdings will not be liquidated.

    The Federal Reserve recently did an interesting study on the possible costs (and budgetary effects) of liquidating the Fed’s balance sheet portfolio. It should be of interest to readers of this blog:

    http://www.federalreserve.gov/pubs/feds/2013/201301/201301pap.pdf

    Sorry for the long-winded post.

  35. Gravatar of Geoff Geoff
    6. March 2013 at 21:23

    Dr. Sumner:

    “Geoff, You said;

    “I wasn’t saying he made that assumption, but that he made the assumption that almost 40 years of historical data that shows a close correlation between the fed funds rate and government bonds rates, would likely have the same close correlation going forward, such that if the fed funds rate increases, so too would government bonds rates increase.”

    “Correlation doesn’t prove causality. Lucas Critique.”

    I didn’t say it proves causality, but rather that I suspect Cochrane is assuming the same thing you and other MMs assume when it comes to NGDP and wage/unemployment data.

    I see 1992-2008 Australia and 1980-2008 US being cited quite often as evidence that NGDP targeting works via implying a causal explanation for the correlated data.

    There is also the even more frequent interpreting of causation from correlation with the rise in unemployment and fall in NGDP from 2008-2010. You have argued, on the basis of this correlation equals causation relationship, that is NGDP did not fall, then unemployment would not have risen as much. That is implying a causation from a correlation.

    The Lucas Critique…it’s interesting that you mention that, because as alluded to above, it would apply with equal force against the notion of that one can predict what happens to unemployment and output, based on a historical relationship between a (to borrow from Lucas) “highly aggregated historical data” such as NGDP, and unemployment and output.

    If you are allowed to do it, why can’t Cochrane? Just substitute fed funds rate and government rates, for NGDP and unemployment/output, respectively, and you’ll see you would be using the same inference as him.

    “The table she cites shows Cochrane to be completely wrong””why should I have assume the quote applied to that table, in which case it would be wrong, and not 5% times 18 trillion, which also equals $900 billion.”

    I don’t think the issue is whether or not this $900 billion figure applies to 2023 or his hypothetical example of what happens with a rise to a 5% interest rate.

    He just said “future” for the hypothetical. To you that meant in the less than 10 year future. Why does he have to mean in the short term like that?

    Then you said he ought to have mentioned going out to 2023 if that’s what he meant. I thought woah, he explicitly mentioned 2023, where the CBO forecasts around $900 billion a year at the end of the post. I could not conclude anything other than laziness.

    “You post a lot here, but if you don’t start thinking before you post I’m going to have to give you the Major Freeman treatment. You are very close to the edge.”

    I don’t know what that means.

  36. Gravatar of Vivian Darkbloom Vivian Darkbloom
    8. March 2013 at 02:11

    James Hamilton el al pick up on the theme in today’s WSJ:

    “Given the Federal Reserve’s greatly expanded balance sheet, which has ballooned from less than $1 trillion at the end of 2007 to more than $3 trillion today, there is an additional factor that could exacerbate inflation expectations””Fed remittances to the U.S. Treasury. If interest rates climb higher over the next few years, this could lead to substantial losses on the Fed’s holdings of Treasurys and mortgage-backed securities, losses that could approach several times the size of Fed capital. The Fed would be unable for a number of years to make its usual payments to the Treasury of interest it earns minus its realized losses on its debt holdings. This could subject the institution to a loss of credibility in financial markets and to political attacks.”

    http://online.wsj.com/article/SB10001424127887324338604578326031255455840.html?mod=opinion_newsreel

Leave a Reply