The end of extended unemployment benefits

What is the effect of the end of extended unemployment benefits (which reduced the maximum from 73 weeks to 26 weeks?)  Unfortunately, it is too soon to tell. Indeed we may never really know.

The two jobs surveys have very different results for the first 4 months of 2014. The household survey shows total employment up by 1,083,000, versus 1,374,000 for all of 2013. That’s clearly an improved pace.  Unfortunately the much more reliable payroll survey does not show a significant upswing, with monthly job growth running at 214,000 vs. 194,000 in 2013.  Of course the April numbers represented a significant improvement in the trend, and perhaps the winter numbers were impacted by the weather.  In my view we need three of four more months to get a good sense of 2014.  Indeed it might not be until some time in 2015 (when the final revisions come in) that we get the full picture.  Right now it’s wait and see.

My best guess is that both sides of the debate will eventually be shown to be partly correct.  When unemployment insurance ended in North Carolina last year the unemployment rate fell relatively sharply.  This probably represented a combination of more jobs created and fewer unemployed due to discouraged workers completely abandoning the job search.  But the data is maddeningly imprecise.  Let’s revisit this in three months.

PS.  We are very rapidly approaching a “5 point something” unemployment rate, which in my view will have important psychological effects, regardless of what you think about the unemployment rate vs. labor force participation rate debate.  The fact that long term bond yields are still quite low provides further support for a claim I’ve been making for years—low interest rates will be the “new normal” in the 21st century.

PPS.  Some sticky-wage skeptics ask why wages wouldn’t adjust in the long run, bringing unemployment back to the natural rate.  I’d like to point out that unemployment has fallen from 10% to 6.3% (the Fed estimates the natural rate is 5.6%.)  This despite record low NGDP growth during the recovery.  Instead of questioning the sticky wage model, people should ask why the Fed allowed such sluggish NGDP growth during the recovery, thus needlessly prolonging the painful adjustment process.

PPPS. A few weeks back I provided this forecast of today’s jobs numbers, my first ever unemployment rate prediction:

Bonus forecast–the unemployment rate will fall several ticks in April.

Yes, I should have said what I really thought—that the fall would be surprisingly large.  (Based on the North Carolina pattern.)  Still, I hope all my readers bought stock options right before today’s announcement. 🙂



41 Responses to “The end of extended unemployment benefits”

  1. Gravatar of Morgan Warstler Morgan Warstler
    2. May 2014 at 07:21


    “low interest rates will be the “new normal” in the 21st century.”

    What if that overstates it… what if the real interest rate is negative forever? Or the trend is that way.

    I’m asking because:

    “Cash is very different from credit. It tends to be used for different types of transactions.”

    If we are moving to a cash + equity = wealth, with far less debt and debt payments to creditors, why doesn’t the % targeted in NGDPLT matter?

    Isn’t 5% or 4.5% historically based on atomic debt based collateral as the way of buying things?

  2. Gravatar of American Debts American Debts
    2. May 2014 at 07:35

    I don’t know how the loss of those many weeks of benefits couldn’t be a loss for the economy. 40 weeks x $280 (guess) x 2,000,000 = quite a bit of spending that would have gone straight into grocery stores, gas stations, pharmacies, auto mechanics and restaurants. Now the budget is a tiny bit better but a lot of dollars that would have been circulating since January are not out there.

  3. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    2. May 2014 at 07:45


    So if the labor force didn’t drop this month because of people giving up, what’s going on? One trend weighing on the labor force is people with jobs retiring, and last month there was an increase in the number of employed people who were no longer working or looking for work. That flow was at its highest level since October and the third highest since before the recession.

    But that doesn’t explain why there are fewer unemployed people. The number of unemployed can fall because people got jobs, because they dropped out of the labor force, or it can fall just because fewer people decided to start looking for work. In April, one of the reasons the number of unemployed fell is because fewer people came off the sidelines to look for work. The number of people flowing from out of the labor force to unemployed was at its lowest level since 2008. Lots more people than usual decided to stay on the sidelines.

    That isn’t a hopeful sign for the economy, as people who otherwise might want work aren’t encouraged enough to come off the sidelines. But it’s also less worrying than people giving up looking for work. Many new entrants to the labor force are younger people who will eventually come in, and are on the sidelines because they can be.

  4. Gravatar of TravisV TravisV
    2. May 2014 at 07:58


    “the short-term unemployed and long-term unemployed have broadly similar educational attributes”

    Good chart.

  5. Gravatar of Major_Freedom Major_Freedom
    2. May 2014 at 09:02

    We can know some of the effects of reductions in UI benefits by virtue of knowing what UI is.

    We know, at the very least, that there will be a ceteris paribus reduction in the amount of unearned gains which comes at the expense of the productive.

  6. Gravatar of Major_Freedom Major_Freedom
    2. May 2014 at 09:05

    “Instead of questioning the sticky wage model, people should ask why the Fed allowed such sluggish NGDP growth during the recovery, thus needlessly prolonging the painful adjustment process.”

    Instead of questioning the NGDP change, people should ask why the Fed inflated the money supply as much as ut did prior to 2007, thus needlessly causing an inevitable painful adjustment process.

  7. Gravatar of John Becker John Becker
    2. May 2014 at 09:12


    It’s not like you to say let’s wait and see about a policy. How did the stock market react to the announcement of the reduction?

  8. Gravatar of TravisV TravisV
    2. May 2014 at 10:35

    Morgan Warstler,

    I think Sumner advocates 4.5% or 5.0% to create just enough inflation to take care of “money illusion” (downward nominal wage rigidity).

    Even in an economy with no debt, “money illusion” and downward nominal wage rigidity still exist.

  9. Gravatar of TravisV TravisV
    2. May 2014 at 10:49


    Mark Sadowski,

    Re: “Petroleum Derangement Syndrome,” I just noticed this:

    “Francisco Blanch, from Bank of America, estimates that shale gas and oil have given the US economy an extra tailwind worth 1.9pc of GDP – what he calls the “energy carry” – with effects rippling through the chemical and plastics industries. New investments in ammonia plants are rising at an exponential rate, thanks to natural gas prices that are $4.40 (per BTU) in the US and $15 on Asia’s spot market.”

  10. Gravatar of BC BC
    2. May 2014 at 10:49

    “Still, I hope all my readers bought stock options right before today’s announcement.”

    I’m not sure I understand this comment. The market barely moved today. (S&P 500 down about 0.13% as of about 2:45pm.) The market is efficient, so it must have already anticipated that Scott’s forecast would be correct. 🙂

  11. Gravatar of Mark A. Sadowski Mark A. Sadowski
    2. May 2014 at 13:49

    Take a look at this:,35845/

    “PASADENA, CA””Citing such cutting-edge advancements as integrated timers with automatic shut-off features, adjustable rotation speeds, and exterior handles that remain cool to the touch throughout cooking, a report released Tuesday by researchers at Caltech confirmed that home rotisserie technology is the only industry in the United States that continues to progress. “After extensive analysis of the nation’s manufacturing and high-tech sectors, we were able to conclude that American innovation is now exclusively limited to the development of countertop rotisserie appliances capable of slow-roasting as many as three full chickens or up to a 15-pound turkey in the convenience of one’s own kitchen,” said the study’s lead author Sara Pane, alleging that recent increases in U.S. GDP could be attributed almost entirely to the continued proliferation of fully dishwasher-safe rotisserie parts. “While American ingenuity has remained robust when it comes to producing restaurant-quality meat at home that browns evenly and retains its natural juices, we unfortunately can’t say the same about the nation’s mining, computer hardware, or motor vehicle industries.” Pane went on to warn, however, that ongoing advances in home rotisserie technology would be rendered largely irrelevant if the nation’s dipping sauce sector was unable to keep pace.”

    I honestly have a hard time telling which claim is the spoof.

  12. Gravatar of bill bill
    2. May 2014 at 15:11

    Great call on the employment numbers.
    Glad I didn’t buy the stock futures.

  13. Gravatar of benjamin cole benjamin cole
    2. May 2014 at 15:23

    Great blogging. BTW, there are nearly 12 million Americans collecting monthly “disability” checks from SSDI and the VA, numbers that dwarf the much-discussed unemployment figs.
    I believe there is plenty of slack in labor markets anyway, but cutting SSDI and VA disability rolls in half would allow much more aggressive Fed growth policies, if one thinks labor markets ciuld become inflationary. But since unit labor costs since 2008 have been frozen…this whole topic is an odd one…

  14. Gravatar of ssumner ssumner
    2. May 2014 at 16:47

    Morgan, NGDP targeting is about labor markets. That’s all. Tech doesn’t matter (for NGDP–it matters elsewhere.)

    American, The fiscal multiplier is zero.

    Patrick, It’s a mistake for them to focus on a single month–too much noise. The last few months are relatively good.

    Travis. Interesting.

    John, As far as I know there was no announcement of the reduction. Congress is still fighting over it.

    BC, The market shot up after the announcement. After that doesn’t matter—EMH.

    Bill, You could have made a lot of money–see previous reply. I’m surprised so many commenters missed this point.

    Ben, But they are not “slack” because they are not available for work, even if the jobs were there. Disability is an adverse supply shock.

  15. Gravatar of Morgan Warstler Morgan Warstler
    2. May 2014 at 17:40

    “NGDP targeting is about labor markets.”

    Thats simply not true.

    NGDPLT is about killing high and slow swings, so that the economy stays on track, and the politics don’t get screwy.

    It’s really only about keeping commies at bay.

    If the goal is for everyone to have a job, GI/CYB.

    But even with everyone having a job with GI/CYB, NGDPLT still functions to force Fiscal to own whether we feast on real growth or drown in inflation.

    And that is all about politics, no?

  16. Gravatar of Kevin Erdmann Kevin Erdmann
    2. May 2014 at 17:50

    Patrick: That interpretation is based on a false narrative of a complicated stew of flows. The flow from “Not in Labor Force” to “Unemployed” declines in recovering economies and increases in recessions. As Scott said, over 1 month this is noise. But, more than that, over longer periods, this is good news.

    TravisV (Re: Yglesias): This is true. It is also true that longer duration unemployment skews older. The combination of these factors means that, while unemployment insurance during the first few weeks of unemployment is a very progressive distribution, very long term unemployment insurance becomes less and less progressive as each additional week is added.

    Here are a couple reviews of the employment report I did today, in case anyone is interested.

  17. Gravatar of Benjamin Cole Benjamin Cole
    2. May 2014 at 19:40

    Okay, I am going to do a Sadowski-lite.

    Below is a table of unit labor costs, USA.

    They are up 2.1 percent since the first quarter—-of 2007 that is. No, that is not an annual rate. That is it. The annual rate is a little more than 0.3 percent, or maybe a measurement error.

    With unemployment benefits lapsing, and with a jigger or two in “disability” programs, I see labor surfeits and low increases in unit labor costs for years ahead.

    The Fed has the field wide-open to go boom-boom.

    BTW, in the last FOMC meeting transcripts made public (from 2008), Fed economists told the FOMC to expect 2 percent increases in unit labor costs forever. Since the FOMC are not open to the public, we do not know if Fed economists are still advising the FOMC board to this effect. Which is not democracy in action, but why should the public be clued in?

    2007-01-01 100.852
    2007-04-01 100.164
    2007-07-01 99.352
    2007-10-01 99.987
    2008-01-01 101.972
    2008-04-01 101.053
    2008-07-01 101.685
    2008-10-01 103.493
    2009-01-01 100.158
    2009-04-01 100.638
    2009-07-01 99.890
    2009-10-01 99.329
    2010-01-01 98.224
    2010-04-01 99.064
    2010-07-01 99.045
    2010-10-01 99.029
    2011-01-01 101.463
    2011-04-01 100.716
    2011-07-01 101.457
    2011-10-01 99.557
    2012-01-01 101.345
    2012-04-01 101.523
    2012-07-01 101.054
    2012-10-01 103.917
    2013-01-01 102.988
    2013-04-01 103.500
    2013-07-01 102.963
    2013-10-01 102.932

  18. Gravatar of Benjamin Cole Benjamin Cole
    2. May 2014 at 19:47

    Side note to Scott Sumner (or anyone).

    The VA and SSDI programs for “disability” are gags. There are cottage industries of lawyers and doctors and “coaches” who will advise you how to apply for “disability.” You can’t hear, you can’t see clearly, you are nervous, your back hurts, you can’t stand on your feet due to the pain, your knees hurt. No one can prove or disprove these symptoms.

    Based on my limited look-see, I would guess both SSDI and VA disability rolls could be cut in half without unfair treatment, but very fair treatment for taxpayers.

    I stand by my statement we have plenty of labor in the USA, and the Fed can crank it up to the moon, and we would still have plenty of labor in the USA.

  19. Gravatar of Johannes Fritz Johannes Fritz
    2. May 2014 at 23:29

    Congrats on your spot-on bonus forecast. I had a hard time interpreting its asset implications. I figured your forecast was based on a drop in LFP rather than strong employment growth. Even though this is no clear sign of economic strength, it may tilt Fed Board opinions towards tightening. If so, one should move out of stocks. Together with Tim Duy’s continued worry that the Fed is more likely to surprise on the hawkish side, I figured Thursday was a good day to collect gains. Slightly falling treasuries and a flat Dow confirms this interpretation, no?

    In general, Tim Duy attributes the discrepancy between Fed talk and action to financial stability concerns. Doesn’t this by definition imply a more or less flat stock market until there is considerable evidence for real activity exceeding expectations? Increasing stocks would otherwise tilt the Fed towards raising rates prematurely, slowing down what wasn’t fast in the first place. Maybe I overextended here.

    Btw, wage growth still nonexistent in this report.

  20. Gravatar of bill bill
    3. May 2014 at 05:20

    Scott – which previous reply? I’m very curious.

    I had started to think that the lack of market reaction right after the announcement was proof that the market expects the Fed to fully offset any good news.


  21. Gravatar of ssumner ssumner
    3. May 2014 at 05:41

    Ben, Yes, but the Fed doesn’t control disability policy.

    Johannes, Good point about wage growth, it is very slow. I don’t have strong views on where the stock market is going.

    Bill, I meant my reply to BC, the market rose immediately after the announcement.

  22. Gravatar of TravisV TravisV
    3. May 2014 at 12:00


    Is anyone aware of how Michael Darda (MKM Partners) applies Market Monetarism to arrive at investment recommendations / what his recommendations are?

    Does Darda work on any international stuff or is that Lars Christensen’s territory (emerging markets)?

  23. Gravatar of Mark A. Sadowski Mark A. Sadowski
    3. May 2014 at 18:30

    Off topic.

    The following post by Tony Yates got under my skin although his claims are pretty vague.

    April 29, 2014

    Don’t underestimate the pivotal and persistent role of the banking crisis
    By Tony Yates

    “Paul Krugman responds to Martin Wolf and John Cochrane arguing for narrow banking [allowing banks simply to take our cash and invest in risk free securities] in part by commenting that the banking crisis came and went quickly, and that the real problem lay elsewhere, in the debt overhang and deficient demand. Therefore, the banking crisis was not decisive in bringing about the recession. For that reason, don’t get so excited about the need to outlaw traditional banking.

    Leaving aside the debate about how to regulate banks, this description of the crisis, although it’s an intriguing idea [if it’s just a matter of the occasional spike in spreads, why not run things as they are?] caused me some trouble. Some reactions below…”

    Let me just quote a few lines of his specific reactions:

    “…Although many spreads have normalised [accounting, I guess, for that normalisation of the St Louis Fed stress index Paul Krugman points to], I conjecture that the harm to banks is more persistent. They have found themselves, in part pressured by their shareholder valuations, to delever. This has constrained net lending, and aggravated the depth and length of the recession…”

    “…Also weighing is the desire not to crystallise losses, given current accounting and regulatory practices. This has been talked about as an acute problem in Japan. It’s an emerging one in the UK, with the FPC spending much effort pondering how much forbearance there is…”

    “…Part of the normalisation of some prices is perhaps the expectation that the authorities will not allow a repeat of Lehmans’. This is must surely be why peripheral EZ spreads have normalised: banks are not really solvent, but with the ECB bluffing that it will buy whatever quantity of sovereign debt necessary, the sovereigns are solvent, and therefore they can credibly stand behind the peripheral banks. It is probably also why similar measures are normalised for the UK. This is less satisfactory as a description of the US, where it seems the Fed would be unable to repeat the bail out of Bear Sterns. Unless we are to fathom that, in another crisis, the executive would find away to change course again, and Republican insistence on non-intervention could be won over. Anyway, the point of this is to argue that the crisis hasn’t gone away, it’s just that the risk has been socialised. And that socialisation of the risk is constraining fiscal policy to some degree, and that is contributing to a more protracted recession than we would otherwise have had; and/or spending is weak because private agents worry about the long term health of their sovereigns, whom, ultimately, they stand behind…”

    This strikes me as a peculiarly European perspective, and in particular, a very British one. If there was a banking crisis in the U.S., it is surely completely over and has been for at least a couple of years. And I say “if” because in the U.S. the financial crisis didn’t really severely affect depository institutions (i.e. banks), rather it affected investment “banks” and other financial firms of a much more peripheral nature. Krugman, Wolf and Cochrane are specifically arguing about depository institutions, not investment banks, money market mutual funds and so on.

    To see where Tony Yates is coming from and to explain why this isn’t really relevant to the U.S. I think it’s important to look at some metrics of financial sector health. Most of these don’t apply specifically to depository institutions. But that doesn’t really matter because if there are significant differences in what these metrics are showing then that implies that banks are probably performing differently as well.

    So here’s a list of metrics and how the financial sectors of the U.S., the Euro Area, Japan and the U.K. are performing by these measures.

    1) Bank Capital to Assets Ratio.

    The semiannual IMF Global Financial Stability Report (GFSR) lists bank capital as a ratio of assets as one of two major criteria to judge the health of the financial sector, the other being the ratio of financial sector credit market debt (loans and securities) to GDP (Table 1.1.1 Page 4):

    The World Bank maintains a database of the ratio of bank capital to assets here:

    Here’s the ratios (percent) for the four major advanced currency areas in 2006, 2008 and 2013:


    (The most recent figure for the U.K. is for 2012.)

    Note that the bank capital ratio is far higher in the U.S. and has increased far more than in any of the other currency areas. The ratio is the lowest in the U.K. and it has increased there the least. It’s also the only one of these currency areas where bank capital is lower now than it was before the financial crisis.

    2) Financial Sector Leverage

    The following are the ratios of financial sector debt to GDP (percent). The debt figures (loans and securities) come from FRED, the ECB Statistical Warehouse and the BOJ Time-Series Data Search.


    Note that the financial sector leverage has fallen more in the U.S. than in any of the other currency areas and is now significantly lower than it was before the crisis. The only other of these currency areas for which this is true is Japan.

    3) Bank Lending

    The following are the outstanding amounts of commercial bank credit (loans and securities), the lending counterpart of M3, loans and discounts, and the lending counterpart of M4 for the U.S., the Euro Area, Japan and the U.K. respectively. The amounts are in billions of local currency units except for Japan which is in trillions. The data comes from FRED, the ECB Statistical Warehouse, the BOJ Time-Series Data Search and the BOE.


    (The most recent figure for the UK is for February 2014.)

    So, yes, bank lending declined significantly in the U.S. following the financial crisis. But now it exceeds its previous peak by a larger margin than in any of the other currency areas. And note that the only one of these currency areas where bank lending is significantly below where it was before the crisis is the U.K.


  24. Gravatar of Mark A. Sadowski Mark A. Sadowski
    3. May 2014 at 18:33

    U.S. Commercial Bank Credit can be found here:

    The Lending Counterpart for Euro Area M3 can be cound here:

  25. Gravatar of Mark A. Sadowski Mark A. Sadowski
    3. May 2014 at 18:36

    Japanese Loans and Discounts (Total of Major, Regional, and Shinkin Banks) can be found here:

    The Lending Counterpart of U.K. M4 can be found here:

  26. Gravatar of Mark A. Sadowski Mark A. Sadowski
    3. May 2014 at 20:34


    4) Financial Sector Profits

    The U.S. uses a different accounting system (NIPA) from the rest of the planet (SNA). So comparing sector profits is difficult. In SNA terms the NIPA measure of corporate profits is the equivalent of the sum of gross disposable income and non-interest distributed income of corporations less the consumption of fixed capital. The following is financial sector corporate profits in real currency units (using the GDP implicit price deflator). The base year is 2005 in all cases except for the U.S. which is 2009. All figures are in billions except for Japan which is in trillions. The data comes from FRED, Eurostat and the OECD (Japan).


    (The most recent figures for Japan and the U.K. are 2012.)

    Note that although 2008 was a dreadful year for U.S. financial sector profits, financial sector profits set a new record in real terms in 2012 followed by another one in 2013. In contrast financial sector profits are down sharply in Japan and the U.K. and are below peak levels in the Euro Area.

    5) Financial Sector Labor Productivity

    The idea for looking at financial sector labor productivity stemmed from the collapse in labor productivity in the UK. Although the problem with declining labor productivity is widespread in the UK, the financial sector has contributed more to the decline than any other sector.

    The following is Financial and Insurance Activities Sector real GDP per hour worked. Output is deflated by the sector specific price deflator except in the case of Japan where it is deflated by the GDP implicit price deflator. The data comes from the BEA/BLS, Eurostat and the Statistics Yearbook of Japan. All figures are in local currency unit per hour worked except for Japan which is in thousands of currency units.


    (The most recent figures for Japan and the UK are for 2011 and 2012 respectively.)

    Note that the Japanese financial sector seems to have suffered a far bigger decline in labor productivity between 2006 and 2008 than any other of these currency areas, followed by the U.S. But whereas U.S. financial sector labor productivity is now at record levels, Japanese financial sector labor productivity seems to have shown almost no signs of recovery, and U.K. financial sector labor productivity has perversely undergone a significant decline during the economic recovery. The Euro Area as a whole has shown no sign of a decline in financial sector labor productivity.

    6) Financial Sector Bailout Costs to Government.

    The best source for data on this is the IMF Fiscal Monitor. See Table 1.6 on Page 9:

    Another source of data is Eurostat:

    The following is a list of countries’ financial sector bailout cost as a percent of GDP and the proportion that has been paid back (percent). All data comes from the IMF except for the data for Portugal and Denmark which comes from Eurostat.


    Note that the cost of the U.S. bailout of the financial sector was far less than Germany, the U.K. or the peripheral members of the Euro Area. Moreover the U.S. bailouts have been paid back with interest, whereas apart from the Netherlands none of the E.U. members have recovered more than half of the cost of bailing out their respective financial sectors. This to me is evidence that the U.S. financial sector has probably fully recovered, whereas the various troubled European financial sectors almost certainly have not.

    One last point, and I think this is very important. Note that although the UK financial sector has performed very poorly, recall that in terms of NGDP growth, the UK has outperformed nearly all of the countries in the Euro Area. So a poor performing financial sector is not at all as pivotal as Tony Yates seems to think. So once again I think the causality goes from NGDP to the financial sector and not the other way around.

  27. Gravatar of TravisV TravisV
    4. May 2014 at 05:30

    “The 10-Year Treasury Yield Is At Its Low Of The Year”


  28. Gravatar of TravisV TravisV
    4. May 2014 at 05:51

    Benjamin Cole:

    “Since the Great Bust, The Fed Has Obtained Deflation in Labor Costs”

  29. Gravatar of Scott Sumner Scott Sumner
    4. May 2014 at 06:12

    Mark, Very interesting data on bank bailouts.

  30. Gravatar of Mark A. Sadowski Mark A. Sadowski
    4. May 2014 at 07:22

    Off Topic.

    Nice little chart via Mike Norman Economics (of all places):

    Granger causality tests of course show show that the monetary base causes the S&P 500 Index and the 10-year T-Note yield at the 5% significance level each from December 2008 to present, and the impulse response is positive in *both* cases.

    The monetary base also Granger causes the DJIA at the 5% significance level during the same period.

    I recall a time when MMTers denied that QE had any effect on anything at all.

  31. Gravatar of Morgan Warstler Morgan Warstler
    4. May 2014 at 07:55

    Mark, this is goofy but I had just been looking at the Billion Price Index:

    Is there any reason borrowing rates (10 year) would look more like CPI, and S&P more like BPI?

  32. Gravatar of Mark A. Sadowski Mark A. Sadowski
    4. May 2014 at 10:51

    I’m not sure which part you think I might find goofy. Recall for example that David Glasner pointed out that prior to 2008 there was no correlation between the S&P 500 and inflation expectations but since 2008 there has been a very robust one. And of course 10-year T-Note yields are strongly correlated to inflation. So the fact you are looking for such correlations is not at all surprising to me.

    I downloaded the BPI (which starts in July 31, 2008) and converted it to monthly frequency for the folllowing analysis.

    The OLS correlation between the S&P 500 index and BPI inflation is significant at the 1% level. This is in contrast to the correlation with CPI inflation which is only significant at the 10% level. This supports what you are saying.

    There is bidirectional Granger causality between the S&P 500 and both the CPI (level) and the BPI (rate) but the BPI Granger causes the S&P 500 at the 1% level and the S&P 500 Granger causes the BPI at the 10% level, whereas the CPI Granger causes the S&P 500 at the 10% level and the S&P 500 Granger causes the BPI at the 1% level. So not much difference except for which direction the causality is more statistically significant.

    The OLS correlation between the 10-year T-Note yield and both CPI inflation and BPI inflation is highly insignificant over this period (August 2008 to present). However, the CPI (level) Granger causes the 10-year T-Note yield at the 5% significance level. On the other hand there is no significant Granger causality between the BPI and the 10-year T-Note yield. So this too tends to support what you are saying.

    The main differences between the BPI and the CPI is that the BPI is only available in changes and not levels, the BPI is available at a daily frequency, the BPI is not weighted in any fashion and the BPI is not seasonally adjusted. It seems to me that the most important of these differences is the lack of weighting (no market basket). So if there’s a way of rationally explaining the differences in correlations it probably relates to this.

  33. Gravatar of ssumner ssumner
    5. May 2014 at 04:57

    Mark, Interesting. Wouldn’t the correlation of QE with the S&P (on that graph) look even more impressive if you dated QE1 from March 2009. I recall that markets responded strongly to that QE announcement, but also observed that people don’t seem to agree as to whether QE1 began in late 2008 or March 2009. SOMETHING happened in March 2009.

  34. Gravatar of Mark A. Sadowski Mark A. Sadowski
    5. May 2014 at 06:42

    I start in December 2008 because that’s the first month that the fed funds rate fell below 0.25%. Thus one can argue that since that point the Fed’s main monetary policy instrument has been the monetary base itself.

    We now have over 60 months of observations so I don’t think removing three (December 2008 and January and February of 2009) would really make much of a difference.

    Moreover, consider what actually happened in those months. Yes the S&P 500 fell from 878 to 866 to 805 on average. But the monetary base (SBASENS) only increased from $1,670 billion in December 2008 to $1,715 billion in January 2009 and then it plummeted to $1,561 billion in February.

    Between December 2008 and February 2010 the monetary base increased by $448 billion. But $444 billion of that increase took place between July 2009 and February 2010. Similarly between December 2008 and April 2010 the S&P 500 increased by 319 points (monthly averages), but 261 points of that increase took place between July 2009 and April 2010.

  35. Gravatar of Morgan Warstler Morgan Warstler
    5. May 2014 at 07:12


    If S&P and BPI reflect “the market” on pricing.

    And Treasuries and CPI are reflect “the government” on pricing.

    I realize this is crude, but “weighting” “adjusted” and “QE” sound the same to me. They sound like government is pulling levers on CPI itself to affect economic action.

    So, I don’t think BPI adds in rents. I can’t find proof one way or the other.

    And I assume, rents are “soaring,” so much that even when home prices plummeted, rents push CPI up:

    Which gets to my weird thing:

    If QE is driving up home prices. directly via MBS. It should show in CPI more than in BPI, no?

    And QE is driving down T-Bills.

    It seems like the difference between BPI and CPI might be a huge deal. Because Fed is not only pulling rents up (CPI), but also driving down cost of money (BPI), but the S&P follows BPI.

    So if Fed stops MBS (rents go down) and stops expanding T-Bills, CPI would fall, but aren’t we less sure BPI would too?

    Are we sure S&P falls? Why? I’m not trying to be Steve Williamson here just want to get answer.

    Another weird thought (reminds me of Roger Farmer): Doesn’t this in some way mean that a Fed that used BPI instead of CPI, would be better to just buy the S&P?

    I’m asking bc I literally have no damn idea.

  36. Gravatar of Mark A. Sadowski Mark A. Sadowski
    5. May 2014 at 14:31

    “It should show in CPI more than in BPI, no?”


    “And QE is driving down T-Bills.”

    Yes. (It is driving down the face values and the yields up.)

    “Because Fed is not only pulling rents up (CPI), but also driving down cost of money (BPI), but the S&P follows BPI.”

    I’m not sure I follow you. What do you mean by “cost of money”? Do you mean the “rental price” of money (interest rates)? Because that’s not true, at least in terms of 10-year T-Note yields. QE drives yields up.

    It is true if you mean the “price of money” (the price level), because QE increases the price level and hence makes money less expensive in terms of goods and services. But then where’s the contradiction?

    “So if Fed stops MBS (rents go down) and stops expanding T-Bills, CPI would fall, but aren’t we less sure BPI would too?”

    BPI and CPI correlate extremely well, even if there has been a divergence lately. (Tyler Cowen is allowing himself to be Zero Hedged.) If CPI falls I fully expect BPI to do so as well.

    “Are we sure S&P falls? Why? I’m not trying to be Steve Williamson here just want to get answer.”

    As Glasner has noted, the S&P 500 Index’s correlation with inflation expectations thes past few years indicates a shortfall of aggregate demand. Normally inflation is bad for share prices, but not when there is so much economic slack.

    “Another weird thought (reminds me of Roger Farmer): Doesn’t this in some way mean that a Fed that used BPI instead of CPI, would be better to just buy the S&P?”

    No, it’s not necessary. The effect of QE does not depend on what assets are purchased. The effect of QE is purely a result of the fact that base money is created.

    The Fed’s purchases of Agencies and Treasuries have been trivially small compared to the entire market for such securities. But the Fed is the sole supplier of base money.

  37. Gravatar of TravisV TravisV
    5. May 2014 at 15:52

    Prof. Sumner,

    Re: FDR, 1933 and George Warren, you wrote the following:

    “As far as I know every single economic historian, liberal and conservative, think Warren’s policy failed. Actually it succeeded. The dollar fell sharply between October 1933 and February 1934. Industrial production began rising and peaked in the spring of 1934.”

    Which non-market monetarist economists out there recognize the huge success of George Warren’s dollar devaluation? Christina Romer? Ben Bernanke? Anyone else?

    You and Glasner have done a great job of documenting its success. But your view of Warren’s program is very non-mainstream right? I mean, even Milton Friedman and Barry Eichengreen missed the importance of the program, right?

  38. Gravatar of Morgan Warstler Morgan Warstler
    6. May 2014 at 05:13


    I don’t think it matters much, but on this:

    “I’m not sure I follow you. What do you mean by “cost of money”? Do you mean the “rental price” of money (interest rates)? Because that’s not true, at least in terms of 10-year T-Note yields. QE drives yields up.”

    I just meant Interest rates for new borrowing. I understand the price of issued bonds change. But ultimately QE is meant to lower interest rates and get borrowing going.

    “BPI and CPI correlate extremely well, even if there has been a divergence lately. (Tyler Cowen is allowing himself to be Zero Hedged.) If CPI falls I fully expect BPI to do so as well.”

    II think ‘m being Zero Hedged too.

    I get that in the long run BPI and CPI go together. I’m wondering if there is something fundamentally wrong with the way CPI is being calculated, because since it includes rents, it seems impossible for it to routinely be below BPI, unless there’s ben a intentional architectural effort to understate it.

    Remember my technology hobby horse: The last 15 years have been the best ever in human history. There is invisible RGDP. Cost of money isn’t just down from QE. It is down bc the long tern real interest rate is negative forever, or at least until we’re all plugged into VR rigs Matrix style. I can’t articulate it past Sci-Fi, but it seems truish that if we are at the end of scarcity, the physical weight of GDP just continues to shrink (Greenspan), that eventually everything costs $1 per pound (Jurvetson), that all the current capital is basically going to be forced to buy equity, bc debt is over. Certainly, the government won’t need to borrow money. In 50 years, we’ll only need 3M public employees, and 10% of our government buildings. Population will level off and shrink Russian style. Robots and 3D printing make anything atomic a commodity. The only investments will be Venture Capital into software which will undermine massive old debt plays, and replace them with What’sApp and Uber for everything.

    I view end of debt to equity speed up the transfer of wealth from yesterday’s winners to tomorrows. Anyone who’s financial skill relies on loaning money and getting paid back safe interest, is going to be skill-less. Investment bankers replaced by Computers. EMH will become incredibly easy to believe in bc only entrepreneurs will routinely have multiple winners, everyone else will be making crap shoot VC investments, as they watch their capital pile shrink.

    It’s “Ready Player One” or “Snow Crash” but it does’t suck. Dystopian future avoided. Yes your home is smaller, yes you live in more dense walkable areas, yes you consume a lot of Soylent, but put on this VR rig bc its more fun and a better life than even the richest guys in 2014 ever enjoyed.

    (end hobby horse)

    “David Glasner pointed out that prior to 2008 there was no correlation between the S&P 500 and inflation expectations but since 2008 there has been a very robust one.”

    What I’m thinking about is QE ends, and the correlation ends with CPI until they have to redo how they do CPI and revise it up.

    Basically, if we use BPI , then we’re at 4.5% NGDPLT and it’s time to tell everyone honestly the economy is doing much better than anyone wants to admit, we have reached full employment at this equilibrium, and the reason all these people don’t have jobs is structural.

    We’re down to 4% Consumption Poverty. If we do GI/CYB it’ll be ZERO.

    And very, very, soon the silly idea that most people can earn their own keep will be discarded, and people will stop talking about wages and aggregate demand.

    We don’t need to make money cheaper, we CAN of course, but loans don’t generate the new awesome invisible RGDP, technology startups do. And as a by product they drive down the cost of money for future technology startups.

    This I really disagree with:

    “No, it’s not necessary. The effect of QE does not depend on what assets are purchased. The effect of QE is purely a result of the fact that base money is created.”

    I kind of always disagreed with this, at first it was Cantillon effects:

    But then I started reading Farmer, and I think his point about Animal Spirits is well taken.

    It seems like buying the S&P, serves a real purpose:

    1. It increases cost of money for government slightly.
    2. It hastens the inevitable move to more efficient productive government.
    3. The multiple is higher. Well, at least it isn’t negative.

    And I guess I was saying CPI is wrong, and BPI is right, and CPI does converge onto BPI, when they stop using the old CPI.

    If S&P and BPI move as one, it seems like you could more accurately target NGDPLT (RGDP+BPI) if you bought S&P.

  39. Gravatar of Mark A. Sadowski Mark A. Sadowski
    6. May 2014 at 07:29

    “I understand the price of issued bonds change. But ultimately QE is meant to lower interest rates and get borrowing going.”

    That’s how QE is sold by the FOMC to the masses, but come on now, we know that’s not how it really works. QE increases NGDP expectations. If NGDP rises then borrowing will probably rise as well. And when the expected rate of NGDP growth goes up interest rates do also.

    Bond rates are the benchmark for borrowing rates, especially mortgage rates, which are 75% of household sector borrowing.

    “If S&P and BPI move as one, it seems like you could more accurately target NGDPLT (RGDP+BPI) if you bought S&P.”

    1) The Fed can’t buy the S&P 500 because it’s not permitted under the Federal Reserve Act.
    2) The BPI is an additional piece of information on consumer price inflation, but it’s not meant to be a precise measure. Even the CPI is a terrible measure of the consumer price level. The PCEPI is a reasonably good measure of consumer prices but if you’re going to target prices you should also target the prices of producers, so the GDP implicit price deflator is really the best for that purpose.
    3) So in my opinion, we already have far too many measures of inflation. And in the final analysis, the rate of change in NGDP isn’t simply the sum of the inflation rate and the growth rate of real output. We should stop worrying about inflation altogether, because it’s an *enormous* waste of time, and just target nominal spending/income/value added (NGDP).

    Why isn’t there a Billion NGDP Project?

  40. Gravatar of Tom Brown Tom Brown
    6. May 2014 at 08:33

    Mark, O/T; Jason has done a series of posts on expectations in economics. Here’s the first:
    You may not agree, but I find the idea interesting.

  41. Gravatar of Morgan Warstler Morgan Warstler
    6. May 2014 at 13:15

    I know it can’t buy S&P. 🙂 I’m wondering about if 1) it could 2) BPI is more tightly correlated, would it take less spend / sell to hit the #?

    I agree we should should forget about inflation all together, same for unemployment, but I think I want a 4% NGDPLT headed toward 2% NGDPLT over 30 years.

Leave a Reply