Maybe inflation isn’t the right variable (example #329)

For years, I’ve been trying to convince the profession that inflation is not the right variable, it’s NGDP growth that matters. One example I frequently cite is the prediction of NK models that inflation caused by negative supply shocks can be expansionary at the zero lower bound.  The failure of the NIRA during the 1930s suggests that this is not true, and now a new NBER study by Julio Garin, Robert Lester and Eric Sims reaches the same conclusion:

The basic New Keynesian model predicts that positive supply shocks are less expansionary at the zero lower bound (ZLB) compared to periods of active monetary policy. We test this prediction empirically using Fernald’s (2014) utilization-adjusted total factor productivity series, which we take as a measure of exogenous productivity. In contrast to the predictions of the model, positive productivity shocks are estimated to be more expansionary at the ZLB compared to normal times. However, in line with the predictions of the basic model, positive productivity shocks have a stronger negative effect on inflation at the ZLB.

The basic problem here is that expected inflation doesn’t matter, what matters is expected NGDP growth.  And while a positive supply shock does indeed reduce inflation, that sort of disinflation is the good kind.  What really matters is NGDP growth, which is not reduced by positive supply shocks.

Here’s one implication they draw from their study:

In the meantime, since our empirical results are difficult to square with the textbook theory, caution seems to be in order when advocating for policies such as forward guidance and fiscal stimulus, both of which are predicted to be highly expansionary when policy is constrained by the ZLB.

BTW, if economists seriously want to argue that inflation matters, they really ought to come up with a coherent definition of inflation.  So far they have failed to do so.





12 Responses to “Maybe inflation isn’t the right variable (example #329)”

  1. Gravatar of Kenneth Duda Kenneth Duda
    7. June 2016 at 11:08

    Scott, the problem is that higher inflation means higher effective taxes on the wealthy give a fixed real interest rate (because taxes are paid on nominal interest rates). If we could get that distortion out of the tax code, maybe we could get the plutocrats to stop obsessing over every quarter-point of inflation.

  2. Gravatar of ssumner ssumner
    7. June 2016 at 11:20

    Ken, Thank God, a non-Trump comment!

    Yes, if we stopped taxing capital income (or indexed the system, which would be difficult) then the optimal inflation rate (or optimal NGDP growth rate) would be higher.

  3. Gravatar of bill bill
    7. June 2016 at 11:54

    I love the inflation/quality adjustment question. The other night I was out with some younger friends. I rarely pay for one off TV streaming. But when I do, I pay $1.99 for SD and they (in spite of significantly less income/wealth) pay $2.99 for HD. They wondered how I could “put up with” the “blurry” picture.

  4. Gravatar of Ray Lopez Ray Lopez
    7. June 2016 at 13:15

    Sumner: “BTW, if economists seriously want to argue that inflation matters, they really ought to come up with a coherent definition of inflation. So far they have failed to do so.” – so, taking Sumner at his word, there’s no coherent definition of inflation, yet Sumner advocates inflation –whatever that is– is not the right variable to target. Trying to pin down Sumner’s logic is like trying to spear a spineless jellyfish.

  5. Gravatar of Ray Lopez Ray Lopez
    7. June 2016 at 13:57

    Here is Sumner in the link on hedonic inflation: “1. Here’s one thought experiment. Get a department store catalog from today, and compare it to a catalog from 1964. … Almost any millennial would rather shop out of the modern catalog, even with the same nominal amount of money to spend”. Stupid. It’s like saying: ‘Anybody sane would not trade their life for living the life of another, therefore, everybody’s life is better than anybody else’. In fact, most people would indeed go back in time and live the life of Louis XIV if they stopped to think about it.

  6. Gravatar of Benoit Essiambre Benoit Essiambre
    7. June 2016 at 14:23

    I’ve been reading this blog for a long time and learned a ton but of all things, the superiority of NGDPLT over PLT still eludes me.

    I totally get going to level targets paths instead of rates.

    I see some benefits to NGDP targeting especially under large shocks (think supervolcano eruption) where an inflation targeted system would surely breakdown and lead to increasingly widespread defaults. I even see potential benefits in terms of better sharing of the economic pie when there are unpredictable small deviations from expected GDP. It can potentially lower inequality due to the same lucky people always being in position to absorb unexpected spurious supply side gains. NGDP would automatically raise real wages and distribute the gains when real GDP would surprise on the upside.

    But I have difficulty shedding the idea that people ultimately want to optimize real consumption and that having interest rate signals that are stable relative to a benchmarks that is tied to something real, not nominal, allows agents to optimize their intertemporal consumption and investment maturity profile better.

    Plus GDP seems difficult to measure reliably to me. With inflation a statistically large enough sample of transactions is sufficient. With GDP you need to count every transaction in the economy. What about fluctuations in the size of the underground economy? What about foreign economies using your currency? Do these not matter?

  7. Gravatar of Benjamin Cole Benjamin Cole
    7. June 2016 at 15:29

    Excellent blogging!

    Moreover, Japan, Thailand, Switzerland and even parts of the US such as North Dakota and Texas for a while, have posted extremely low rates of unemployment yet very little inflation (as measured) in fact sometimes deflation.

    The Fed obsesses with inflation and unemployment.

    There was a time when federal officials obsessed with body counts in the Vietnam war.

  8. Gravatar of ssumner ssumner
    7. June 2016 at 16:45

    Bill, I go to the movie theatre, even the HD image isn’t good enough for a snob like me. 🙂

    Ray, I said the government should not target the CPI, which does exist, in case you didn’t notice.

    Benoit, I don’t have time to answer all your questions, but foreign economies using the dollar don’t matter for optimal US policy. Part of the underground economy matters, but it probably doesn’t change enough from year to year to cause problems in NGDP targeting. In some ways NGDP is easier to measure than inflation—you don’t have to worry about quality adjustments and new goods.

  9. Gravatar of Benjamin Cole Benjamin Cole
    7. June 2016 at 19:53

    Side note from Australia:

    “Last week Westpac Bank decided to ban loans to non-residents, becoming the third major bank to do so after ANZ and Commonwealth. In an explanatory email to mortgage brokers it said its core objective was to help Australians realize their dreams of owning a home.

    Foreign property buyers received another blow late last month when Victoria raised a tax on foreign property purchases from 3% to 7% in its state budget. It also tripled the surcharge on ‘absentee landholders’ to 1.5 per cent from 0.5 per cent.”


    This is fascinating. The Aussies have been doing the free trade thing, and running chronic trade deficits, and sending slips of paper to Asia.

    The Asians have been using the slips of paper to buy up Aussie housing, and evidently pushing Aussies aside.

    The real solution is probably de-zoning and housing production.

    But it does raise questions about chronic trade deficits in the real world. When a nation runs chronic trade deficits, it must mortgage its assets (or sell bonds, that is long-term debtorship).

    Now, selling off assets may just result in immigration, and a win-win, but only if there is no tension between immigrants and natives.

    No doubt, some will say the Aussies are being racist. But then, at one point people hated Americans, for similar reasons–buying up the best housing, or spots in restaurants etc.

    Hong Kongers today are complaining that mainland tourists are shooting the prices for everything up though the roof.

    Another side note:

    “Japan issues record number of visas to Chinese in 2015, up 85%

    TOKYO (Kyodo) — Japan issued a record number of visas to Chinese nationals in 2015, up about 85 percent from the previous year to 3.78 million, as Chinese tourist arrivals increased on a weaker yen and relaxed visa rules, the Foreign Ministry said M…”


    Really, is this a picture of a Sino collapse? An 85% surge in visas?

    BTW, Xi has successfully cracked down Macau (gaming action way off) and now says he wants Sino tourists to stop spending money in Japan. Sino tourism is now declining.

    International trade is a fascinating topic.

  10. Gravatar of Major.Freedom Major.Freedom
    7. June 2016 at 20:34

    GDP does not have a coherent meaning either, so neither can NGDP.

    GDP is defined as “the monetary value of all finished goods and services produced within a country’s borders in a specific time period.”

    If we abstract from the monetary value, and just consider what Sumner believes he is talking about when he considers NGDP, we can define NGDP as the sum total of spending on final goods and services within a country’s borders in a specific time period. Superficially it seem coherent, but that is only because it uses conventional everday words that people can understand, whereas CPI talks about baskets and weights and hedonic adjustments. If we question what GDP is supposed to be however, we find there are also problems of “incoherence”.

    For example, the collection of all “final goods and services”. On face value, this would appear to refer to consumer goods and services. Yet GDP includes capital goods, which are certainly not final goods, because they are still to be used up in the production process and are used to make subsequent sales.

    So does GDP include all capital goods or doesn’t it? Why does it only include some capital goods and not others, even though all capital goods are used to make subsequent sales? Double counting you say? Including capital goods cannot avoid double counting in some respects.

    And economists cannot even agree what the definition of consumer goods and services are! For example, Sumner claims houses are capital goods, despite expenditures by their end users not being used in a business production process to make subsequent sales. Other economists claim houses are consumer goods. Doesn’t matter you say? Just include expenditures on houses in NGDP? But what is the principle? Because it is a capital good or because it is a consumer good? If a capital good, why not include ALL capital goods, such as the bricks used to build the houses? Double counting you say? OK then, even a “finished” house is not really finished, because people spend money on renovations and additions. So did we double count the initial purchase of the house should that house later on be sold for more money?

    What about a pick up truck purchased by a customized shop? Is that purchase to be included in NGDP? Or should the subsequent customized truck be included only? Or both?

    Sumner once called over himself trying to define the difference between capital goods and consumer goods, and I tore this blog a new one in response. Sumner tried to define the difference based on objective characteristics of the goods themselves, i.e. size, shape, longevity, etc. But economics is a subjectivist discipline. A pick up truck is not a capital good or consumer good based on its appearance or chemical properties. A pick up truck is a capital good or consumer good depending on the subjective use of it. In the hands of someone who buys it for the purposes of making subsequent sales, such as a construction company, that truck is a capital good, as it is “capitalized” as an asset and depreciated. On the other hand, if bought not for the purposes of making subsequent sales, say it is purchased by a person who uses it for his own enjoyment and needs, then it is a consumer gold.

    Sumner doesn’t understand this, but a lot of people agree with me. There is no agreement on what final goods and services are even supposed to include. What is supposed to be included in NGDP is on the basis of definitions as incoherent as CPI. And why should that be surprising? They’re both derived from the same crude aggregate thinking about heterogenous concepts.

    If there is no agreement on the precise meaning of “final goods and services”, the principles that allow people to distinguish between capital goods and consumer goods, and within capital goods, what to include and what to exclude, and why, how can there be any “targeting” on the spending on those goods by controllers of the socialist monetary regime?

  11. Gravatar of Patrick R. Sullivan Patrick R. Sullivan
    8. June 2016 at 05:03

    ‘I have difficulty shedding the idea that people ultimately want to optimize real consumption and that having interest rate signals that are stable….’

    Shed the idea that ‘interest rates are the price(s) of money. That should clear up your thinking.

  12. Gravatar of Ray Lopez Ray Lopez
    9. June 2016 at 11:21

    @Major Freedom – your points are good and reflected in this topic:

    In a nutshell, market value or historical cost? But market value fluctuates and you cannot sum market values to get the same GNP as historic cost, so which do you pick? Google or GM? is worth more to society? A variant of the water / diamond paradox. 0/0 as an Austrian like you should appreciate.

Leave a Reply