Martin Feldstein, then and now

Here’s Martin Feldstein in 2011:

Until the fourth quarter of last year, the US economic recovery that began in the summer of 2009 was decidedly anemic. Annual GDP growth in the first three quarters of 2010 averaged only about 2.6% – and most of that was just inventory building. Without the inventory investment, the growth rate of final sales averaged less than 1%.

But the fourth quarter was very different. Annual GDP rose by 3.2% and growth of final sales jumped to a remarkable 7.1% year-on-year rate. True, much of that was due to a sharp decline in imports; but even the growth rate of final sales to domestic purchasers rose at a healthy 3.4% pace.

The key driver of the increase in final sales was a strong rise in consumer spending. Real personal consumer spending grew at a robust 4.4% rate, as spending on consumer durables soared by 21%. That meant that the acceleration of growth in consumer spending accounted for nearly 100% of the increase in GDP, with the rise in durable spending accounting for almost half of that increase.

The rise in consumer spending was not, however, due to higher employment or faster income growth. Instead, it reflected a fall in the personal saving rate. Household saving had risen from less than 2% of after-tax incomes in 2007 to 6.3% in the spring of 2010. But then the saving rate fell by a full percentage point, reaching 5.3% in December 2010.

A likely reason for the fall in the saving rate and resulting rise in consumer spending was the sharp increase in the stock market, which rose by 15% between August and the end of the year. That, of course, is what the Fed had been hoping for.

At the annual Fed conference at Jackson Hole, Wyoming in August, Fed Chairman Ben Bernanke explained that he was considering a new round of quantitative easing (dubbed QE2), in which the Fed would buy a substantial volume of long-term Treasury bonds, thereby inducing bondholders to shift their wealth into equities. The resulting rise in equity prices would increase household wealth, providing a boost to consumer spending.

To be sure, there is no proof that QE2 led to the stock-market rise, or that the stock-market rise caused the increase in consumer spending. But the timing of the stock-market rise, and the lack of any other reason for a sharp rise in consumer spending, makes that chain of events look very plausible.

The magnitude of the relationship between the stock-market rise and the jump in consumer spending also fits the data. Since share ownership (including mutual funds) of American households totals approximately $17 trillion, a 15% rise in share prices increased household wealth by about $2.5 trillion. The past relationship between wealth and consumer spending implies that each $100 of additional wealth raises consumer spending by about four dollars, so $2.5 trillion of additional wealth would raise consumer spending by roughly $100 billion.

That figure matches closely the fall in household saving and the resulting increase in consumer spending. Since US households’ after-tax income totals $11.4 trillion, a one-percentage-point fall in the saving rate means a decline of saving and a corresponding rise in consumer spending of $114 billion – very close to the rise in consumer spending implied by the increased wealth that resulted from the gain in share prices.

And here’s Martin Feldstein in 2013:

The Federal Reserve recently announced that it will increase or decrease the size of its monthly bond-buying program in response to changing economic conditions. This amounts to a policy of fine-tuning its quantitative-easing program, a puzzling strategy since the evidence suggests that the program has done little to raise economic growth while saddling the Fed with an enormous balance sheet.

. . .

Here’s how it is supposed to work. When the Fed buys long-term government bonds and mortgage-backed securities, private investors are no longer able to buy those long-term assets. Investors who want long-term securities therefore have to buy equities. That drives up the price of equities, leading to more consumer spending.

But despite the Fed’s current purchases of $85 billion a month and an accumulation of more than $2 trillion of long-term assets, the economy is limping along with per capita gross domestic product rising at less than 1% a year. Although it is impossible to know what would happen without the central bank’s asset purchases, the data imply that very little increase in GDP can be attributed to the so-called portfolio-balance effect of the Fed’s actions.

Even if all of the rise in the value of household equities since quantitative easing began could be attributed to the Fed policy, the implied increase in consumer spending would be quite small. According to the Federal Reserve’s Flow of Funds data, the total value of household stocks and mutual funds rose by $3.6 trillion between the end of 2009 and the end of 2012. Since past experience implies that each dollar of increased wealth raises consumer spending by about four cents, the $3.6 trillion rise in the value of equities would raise the level of consumer spending by about $144 billion over three years, equivalent to an annual increase of $48 billion or 0.3% of nominal GDP.

This 0.3% overstates the potential contribution of quantitative easing to the annual growth of GDP, since some of the increase in the value of household equities resulted from new saving and the resulting portfolio investment rather than from the rise in share prices. More important, the rise in equity prices also reflected a general increase in earnings per share and an increase in investor confidence after 2009 that the economy would not slide back into recession.

.  .  .

In short, it isn’t at all clear that the Fed’s long-term asset purchases have raised equity values as the portfolio balance theory predicted. Even if it did account for the entire rise in equity values, the increase in household equity wealth would have only a relatively small effect on consumer spending and GDP growth.

There’s no direct contradiction, but the tone sure has changed.

BTW, QE works through the expected hot potato effect boosting NGDP, plus sticky wages, not via reduced consumer saving.

Update:  That was poorly worded.  I meant “in the presence of sticky wages.”  I did not mean to imply QE boosts wages in the short run.

PS.  NGDP bleg:  I saw that eurozone RGDP fell 0.2% in Q1, but can’t find NGDP data.  I hope they don’t report NGDP with a lag like the British do, it’s impossible to calculate RGDP w/o knowing NGDP

HT:  Vivian Darkbloom



21 Responses to “Martin Feldstein, then and now”

  1. Gravatar of Brent Brent
    23. May 2013 at 20:02

    RE: “NGDP blg” – Eurostat is the agency in the European Union that compiles EU and eurozone statistics – see

    Actually, quite a few countries do compile their early estimates of RGDP without knowing NGDP. They estimate NGDP for the reference period and then extrapolate RGDP using quantity or volume indices (such as the index of industrial production). They can do this without deflating nominal output or expenditures. Later on, when data on nominal output or expenditures become available, the early estimates will be revised and replaced with estimates based on price deflation. The data currently available on the Eurostat website show Q1 NGDP estimates available for Germany, France, and the Netherlands, but only available through Q4 for many of the smaller countries.

  2. Gravatar of Benjamin Cole Benjamin Cole
    23. May 2013 at 20:28

    There is what—a virus, a secret cabal, a sinister plot intent on U.S. destruction?—that seems to inflame various economic pundits, ever since the Fed belatedly and feebly began QE.

    Feldstein is the latest. What is he driving at?

    In that same piece, Feldstein moaned that the Fed was “saddled” with a balance sheet. That means it owns $3 trillion in bonds that are throwing off interest. I would like to be so saddled.

    What is the danger of the Fed owning the bonds? It never has to sell, or it can add to the pile, or it can sell to cool things down (such a problem I hope to see one day again before I die. I am 58).

    Feldstein joins Taylor, Mishkin, Metlzer and various others who backed QE in various contexts extremely similar to now, but now rant about —what? Not inflation. Inflation is dead. So they say QE is not working. They say QE is only going into excess reserves (so why not cut IOR?). It is immoral. Popeye is against it.

    I just do not understand the anti-QE crowd. It must be Theomonetarism.

    In religion, it is not the evidence that matters. It is your favored faith that matters.

    It is the Attack of Theomonetarists.

  3. Gravatar of Max Max
    23. May 2013 at 22:03

    Portfolio balance effect, hot potato effect, which phantoms should we believe in?

    If a placebo painkiller reduces pain, does that mean you don’t need morphine? Just increase the amount of sugar in the placebo until you get a morphine-equivalent response? Maybe it’s important to have a correct understanding of why it works.

  4. Gravatar of Vivian Darkbloom Vivian Darkbloom
    23. May 2013 at 23:07

    “BTW, QE works through the expected hot potato effect boosting NGDP, plus sticky wages, not via reduced consumer saving.”

    It would be useful if you were to elaborate on this. I assume that by “hot potato effect” you refer to the phenomenon by which people spend money today rather than the future because they fear future inflation will erode their purchasing power. (If this is so, wouldn’t that also “reduce consumer savings?)

    The potato does not seem to have been very hot since QE was initiated. Inflation has remained low, in fact below target as you have often mentioned here. Also, inflation expectations have also remained low and you’ve recently remarked here that you expect interest rates to remain low for quite some time. Can I safely presume this is a rough proxy for anticipated inflation?

    I would have thought that the increased consumer spending would have been the result of the additional confidence and at least the wealth perceived to have been created by increasing asset prices such as the value of stock. This seems consistent with Bernanke’s description of the portfolio balance effect. If one needs evidence that Bernanke can work this channel through Federal Reserve policy, just look at the market reaction to his very mild statement the other day. Bernanke seems to know that he has great influence over the stock market and his statement the other day struck me as his little experiment to see just how great that influence is and has been.

    As someone who is cautiously sympathetic to your monetary theories, it would help me a great deal if you were to explain in some detail exactly how and through which transmission channel or channels you believe monetary policy and in particular QE (see above quote) can work. I strikes me that you and Bernanke may not be on the same page, and this is an issue that I think has gotten insufficient coverage here.

  5. Gravatar of Vivian Darkbloom Vivian Darkbloom
    24. May 2013 at 00:01

    “BTW, QE works through the expected hot potato effect boosting NGDP, plus sticky wages, not via reduced consumer saving.”

    As far as “boosting…sticky wages” is concerned, I don’t see much effect of QE on those sticky wages thus far. The cost of benefits has gone up bit, but wage and salary costs have been at or below the rate of inflation:

    I also fail to perceive any “expected” increase in those sticky wages that would be driving the modest improvement in the economy and indirectly employment.

    If one were to overlay the actual and expected rate of inflation and the actual and expected wage increase over the actual and expected increase in stock prices since QE began, what would one see? I think one would see that QE has had a much greater effect on the price of investment assets than anything else. This seems to me more consistent with Bernanke’s view of how QE is supposed to work than the idea that it works through expectations of general price and wage increases. One is almost tempted to use the word “bubble”, but I will refrain from that because as with the term “austerity” it is a word that is thrown around by economists and pundits despite the fact that it is not susceptible to any common usage or understanding.

  6. Gravatar of Bill Woolsey Bill Woolsey
    24. May 2013 at 03:49

    The “hot potato effect” is households and firms, including banks, spending excess base money balances. To the degree the excess money balances is spent on financial assets, this doesn’t directly impact nominal GDP. Only to the degree it is spent on final goods and services, does it impact nominal GDP.

    The impact on consumer saving is ambiguous. More spending on consumer goods, with spending on other output constant, will raise nominal income, but with income great than consumption spending, the saving rate would decrease. However, if the excess money balances are spent by firms on capital goods, then nominal income rises with the spending and the saving rate rises. If the excess money balances are spent on both capital and consumer goods in exact proportion to their current levels, then the saving rate is not changed at all.

    Summer’s point, however, is that the saving rate isn’t important in this context. (It is important in other contexts.) The purpose of quantitative easing isn’t to raise stock prices, reduce the saving rate, and so increase consumer expenditures. (Or rather, it shouldn’t be. The FOMC might have a different view.)

    The purpose is to create a surplus of base money at the current level of nominal GDP, increase spending on output and so raise nominal GDP.

    Because some prices (Sumner emphasizes wages) are sticky, increases in nominal GDP results in higher real output as well as higher prices. The less sticky prices rise more and the more sticky prices (like wages) rise less. Output and employment rise too.

    If we are currently in a situation where all markets are clearing (more or less,) any increase in employment and output will be temporary.

    However, if we are in a situation where markets are not clearing, and wages and prices need to fall further to return output and employment to their natural levels, then the expansion in nominal GDP will make those prices decreases unnecessary an return output and employment to the natural levels.

    I believe that Sumner’s view of the role of stock prices isn’t that excess money will be spent on stocks, and so stock prices will rise, and then that raises consumer spending (and reduces saving.) It is rather that as the excess money balances is spent on output, then firms’ nominal earnings will increase. Anticipating that increase in nominal firm earnings, investors bid up the prices of stocks. Watching the impact of a quantitative easing programs on stocks is looking at how “the market” anticipates the program will impact the economy.

  7. Gravatar of W. Peden W. Peden
    24. May 2013 at 04:00

    Vivian Darkbloom,

    Funny you mention asset prices vs. wages under QE:,DJIA&scale=Left,Left&range=Custom,Custom&cosd=2000-01-01,2000-05-26&coed=2013-01-01,2013-05-22&line_color=%230000ff,%23ff0000&link_values=false,false&line_style=Solid,Solid&mark_type=NONE,NONE&mw=4,4&lw=1,1&ost=-99999,-99999&oet=99999,99999&mma=0,0&fml=a,a&fq=Quarterly,Daily&fam=avg,avg&fgst=lin,lin&transformation=nbd,nbd&vintage_date=2013-05-24,2013-05-24&revision_date=2013-05-24,2013-05-24&nd=2007-12-01,2007-12-01

    Under QE, stocks have done relatively poorly compared to wages (which is unsurprising given the instability of asset prices and the fact that there’s a flight to safer assets during a financial crisis/recession) until the last few months. Stocks reached their pre-crisis nominal peak only very recently; wage & salary accruals did that years before.

    Stocks have only been a very good investment under QE because they halved in 2008 and have caught up since then. If stocks diverging from wages is supposed to be indicative of a bubble, then there is no indication (from that) of a bubble.

    Therefore, there’s as much evidence that QE boosts nominal wages as there is that it boosts asset prices. So, if you think that Qe boosts asset prices, then you must also believe that it boosts sticky wages.

    (I’m surprised that you claim that “wage and salary costs have been at or below the rate of inflation” on the basis of that chart, given how low inflation has been in the US under QE. Here are a few measurements-,WASCUR_CPIAUCSL_GDPDEF_PCEPI_CPILFESL,WASCUR_PCEPI&scale=Left,Left,Left&range=Custom,Custom,Custom&cosd=2009-01-01,2009-01-01,2009-01-01&coed=2013-01-01,2013-01-01,2013-01-01&line_color=%23ff0000,%230000ff,%23006600&link_values=false,false,false&line_style=Solid,Solid,Solid&mark_type=NONE,NONE,NONE&mw=4,4,4&lw=1,1,1&ost=-99999,-99999,-99999&oet=99999,99999,99999&mma=0,0,0&fml=a%2Fb,a%2Fc,a%2Fb&fq=Quarterly,Quarterly,Quarterly&fam=avg,avg,avg&fgst=lin,lin,lin&transformation=nbd_nbd,nbd_nbd_nbd_nbd_nbd,nbd_nbd&vintage_date=2013-05-24_2013-05-24,2013-05-24_2013-05-24_2013-05-24_2013-05-24_2013-05-24,2013-05-24_2013-05-24&revision_date=2013-05-24_2013-05-24,2013-05-24_2013-05-24_2013-05-24_2013-05-24_2013-05-24,2013-05-24_2013-05-24&nd=2007-12-01_2007-12-01,2007-12-01_2007-12-01_2007-12-01_2007-12-01_2007-12-01,2007-12-01_2007-12-01

    However you deflate them, real wage & salary accruals are above the Q1 2009 level. Nominal wage growth has been very slow in the US recently, but the same is true for prices. If we look at total compensation-,WASCUR_CPIAUCSL_GDPDEF_PCEPI_CPILFESL_COMPNFB_COE,WASCUR_PCEPI_COMPNFB_COE&scale=Left,Left,Left&range=Custom,Custom,Custom&cosd=2009-01-01,2009-01-01,2009-01-01&coed=2013-01-01,2013-01-01,2013-01-01&line_color=%23ff0000,%230000ff,%23006600&link_values=false,false,false&line_style=Solid,Solid,Solid&mark_type=NONE,NONE,NONE&mw=4,4,4&lw=1,1,1&ost=-99999,-99999,-99999&oet=99999,99999,99999&mma=0,0,0&fml=d%2Fb,g%2Fc,d%2Fb&fq=Quarterly,Quarterly,Quarterly&fam=avg,avg,avg&fgst=lin,lin,lin&transformation=nbd_nbd_nbd_nbd,nbd_nbd_nbd_nbd_nbd_nbd_nbd,nbd_nbd_nbd_nbd&vintage_date=2013-05-24_2013-05-24_2013-05-24_2013-05-24,2013-05-24_2013-05-24_2013-05-24_2013-05-24_2013-05-24_2013-05-24_2013-05-24,2013-05-24_2013-05-24_2013-05-24_2013-05-24&revision_date=2013-05-24_2013-05-24_2013-05-24_2013-05-24,2013-05-24_2013-05-24_2013-05-24_2013-05-24_2013-05-24_2013-05-24_2013-05-24,2013-05-24_2013-05-24_2013-05-24_2013-05-24&nd=2007-12-01_2007-12-01_2007-12-01_2007-12-01,2007-12-01_2007-12-01_2007-12-01_2007-12-01_2007-12-01_2007-12-01_2007-12-01,2007-12-01_2007-12-01_2007-12-01_2007-12-01

    – we see that the recovery has been somewhat better, which is not surprising given how important employee health insurance is as part of workers’ compensation in the US, due to your bizarre healthcare system.

    I suppose one could attribute the recovery in wages/compensation to something other than QE, but why not attribute the rise in asset prices to that other thing?

  8. Gravatar of W. Peden W. Peden
    24. May 2013 at 04:02

    Actually, the “stocks have done better than wages” meme is something I’ve been hearing a lot over these past few years. Is it because people (correctly) deflate wages, but don’t do the same for asset prices? Or because people use inconsistent time periods for their comparison, e.g. they compare stocks with late 2008 and wages with 2007?

  9. Gravatar of Vivian Darkbloom Vivian Darkbloom
    24. May 2013 at 04:24

    W. Peden,

    Thanks for the charts.

    Regarding the first one, we’re not talking about the performance since the recession began; we’re talking about performance since QE began (or started to take effect). *At the earliest*, that would be late November 2008–probably somewhat later to account for a lag. If you want to show me that wages have done better under QE than stocks, then please adjust your chart to account for the above. I think you’ll see quite a different picture.

  10. Gravatar of Vivian Darkbloom Vivian Darkbloom
    24. May 2013 at 04:27

    “The “hot potato effect” is households and firms, including banks, spending excess base money balances.”

    Fine, but that is not really saying much at all. The issue here is *why* they might do that as a result of QE. I do not believe your response addressed that question.

  11. Gravatar of SG SG
    24. May 2013 at 04:28

    Recently Tyler Cowen linked to a piece in the American Conservative called “American Pravda” which basically alleged that the American press willfully ignored major scandals because they would have inflicted bipartisan damage on the Washington elite. I don’t really know whether his accusations are true, but I think I know how he feels.

    I have become convinced that monetary regime change (to NGDPLT) would be hugely, hugely helpful to the economy, both in the short run, but especially in combating future crises. I recently watched this terrific video made in 1979 by Milton Friedman explaining, with devastating clarity, how the Fed’s incompetence, either caused the Great Depression or made it far worse than it would have been otherwise. Everything that I have read about monetary policy in the last 5 years confirms to me that the Fed today is making the same mistake.

    So I am left asking, like Ron Unz, why is the media (with a handful of exceptions, none of whom have moved the needle of public opinion) utterly disinterested in informing the public about the scandal to dwarf all scandals–that a handful of unelected central bankers has destroyed the lives and livlihoods of tens (hundreds?) of millions of people across the world?

    The failure of the media has led to mass ignorance of the most important economic story of the last 70 years. I’ve tried to tell every one of my classmates and coworkers over the last 3 or 4 years that the Fed is principally responsible for the crash and its resulting devastating effects on the economy, but with 1 or 2 exceptions, I’ve been listened to or ignored, as if I were a benign conspiracy theorist. If you try to convince people of something too far outside the mainstream, you’re usually treated like a nutcase.

    It’s a weirdly isolating feeling.

  12. Gravatar of W. Peden W. Peden
    24. May 2013 at 04:34

    Vivian Darkbloom,

    If stock prices (which are forward-looking) were depressed by the recession, then much of their recovery can be explained as a return to trend and a change in expectations. So all one would be measuring by comparing wages & stocks from Q1 2009 onwards would be the fact that stocks are forward-looking & unstable and wages are sticky.

  13. Gravatar of Vivian Darkbloom Vivian Darkbloom
    24. May 2013 at 04:47

    W. Peden,

    So, in fact, stocks *have* increased more rapidly than wages since QE began. Can we agree now on the correlation? You are now arguing that while this correlation fits better with QE than the correlation with rising wages, QE does not (at all?) explain the rise in stock prices? I’m left to wonder why, when Bernanke opens his mouth, stocks rise or fall, depending on what he says will likely happen with QE.

    As far as “sticky wages” are concerned, I recall several recent posts here that have linked the drop in the unemployment rate not to fiscal “austerity” (whichever iteration of that term you might choose) but to monetary easing. In general, while I agree with that assessment, but perhaps for other reasons, I would think if QE were responsible for “unsticking wages”, this would tend to increase rather than decrease the unemployment rate.

  14. Gravatar of ssumner ssumner
    24. May 2013 at 05:04

    Brent, Wow! That’s just incredibly pathetic. We are almost through May and the ECB has no eurozone NGDP data for Q1?

    Max. The HPE is simple S&D. Do you believe in supply and demand?

    Vivian, The wage comment was poorly worded–see my update.

    Regarding transmission mechanisms, I’ve always distinguished between the short and long run mechanisms. The long run comes first, as the HPE causes NGDP to rise in proportion to an exogenous increase in the base, in the long run. In the short run, expectations of that long run increase in NGDP cause higher spending and higher NGDP today.

    Bill, Very good comment.

    W. Peden, Yes, that’s a good corrective to the view that the crisis helped the rich and hurt workers. Of course Vivian is right that QE helps stocks more in the short run, from the point it was adopted.

    SG, Yes, that resonated with me.

  15. Gravatar of W. Peden W. Peden
    24. May 2013 at 05:29

    Vivian Darkbloom,

    Fair enough: if we take as our starting point the beginning of QE, then stocks have recovered more than wages.

    I don’t deny that QE has an effect on asset prices.

    The channels for QE aren’t fundamentally different from interest rate changes. There is as much empirical basis to believe that QE has a disproprotionate effect on stocks as opposed to wages as there is to believe that cutting the interest rate target in 2008 had a disproportionate effect on wages.

    “I would think if QE were responsible for “unsticking wages”, this would tend to increase rather than decrease the unemployment rate.”

    I don’t think QE boosted wages by unsticking wages, but rather by stopping NGDP contracting and sustaining a modest but positive NGDP growth rate, with accompanying real growth and lower unemployment, and subsequently a rise in wages as the demand for labour rose.

  16. Gravatar of marcus nunes marcus nunes
    24. May 2013 at 07:07

    Beckworth & Ponnuru ‘undo’ Feldstein:

  17. Gravatar of W. Peden W. Peden
    24. May 2013 at 07:32

    Scott Sumner,

    Yes: the crisis did hit the workers unnecessarily hard, but it hit asset-holders even harder.

  18. Gravatar of Max Max
    24. May 2013 at 14:25

    “Max. The HPE is simple S&D. Do you believe in supply and demand?”

    It’s not enough to believe in supply and demand, you have to understand it. What is the sensitivity of base money demand to interest rates? Away from the zero bound, not sensitive at all. At the zero bound (or at all times if IOR>0), exquisitely sensitive. That’s because when base money offers a competitive yield, the demand is not limited to medium of exchange demand. So why do you assume that an increase in base money supply must be inflationary, if it can be accommodated by an infinitesimal drop in interest rates (the only case where QE is relevant)?

  19. Gravatar of ssumner ssumner
    25. May 2013 at 06:18

    Marcus, Thanks for the link.

    W. Peden, I agree.

    Max, Because the zero bound won’t last forever, and hence it creates expected future inflation, which lowers real interest rates and creates current inflation.

  20. Gravatar of Max Max
    25. May 2013 at 11:06

    “Max, Because the zero bound won’t last forever, and hence it creates expected future inflation, which lowers real interest rates and creates current inflation.”

    A more plausible explanation of why there is a market reaction to QE is that it signals the Fed would like more inflation, meaning that an uptick in inflation won’t immediately bring Fed tightening. But this isn’t because QE makes tightening more difficult! QE is not a commitment strategy or anything like that. It’s just a communication channel.

  21. Gravatar of ssumner ssumner
    26. May 2013 at 06:46

    Max, That’s simply another way of making the same point. I was talking communication–I’ve never seen credibility as being a problem, at least with level targeting.

Leave a Reply