Spending on cars and houses doesn’t matter

The new GDP figures offer a reminder that one can’t analyze movements in GDP by looking at components of GDP.  We saw huge increases in spending on cars (pushing consumer durables up by 15.3%) and houses (up 19.1%.)  And yet overall RGDP growth fell to only 2.2%.  Why didn’t the spending on cars and houses help?  Because in the short run it’s NGDP growth that drives RGDP.  And the Fed continued its tight money policy by allowing only 3.8% NGDP growth, same as the previous quarter.

This report alludes to the fallacy of trying to analyze GDP by looking at sectors:

While the unseasonably warm weather helped the economy by boosting home building and renovations, it undercut demand for utilities, spending at ski resorts and sales of winter apparel.

If consumers spend more in one area they spend less in another area.  What you need to do is change the total amount that consumers are spending, and only the Fed can do that.  Sectors don’t matter, and government is just another sector (as long as Bernanke is determined to keep inflation at 2%.)

Of course this also explains why RGDP kept growing during the Great Housing Construction Crash of early 2006 to mid-2008.  The Fed kept NGDP growing at a decent clip.

Update: This Karl Smith post suggests the GDP number may eventually be revised upward.


Tags:

 
 
 

39 Responses to “Spending on cars and houses doesn’t matter”

  1. Gravatar of D.Gibson D.Gibson
    27. April 2012 at 06:08

    correct me if I am wrong, but the same goes for prices. doubling the price of gasoline means that folks spend less in other areas. To a first order, expensive gasoline doesn’t cause general inflation.

    Tight money makes me sad:(

  2. Gravatar of Bill Woolsey Bill Woolsey
    27. April 2012 at 06:16

    If the supply of oil has decreased, then the quantity of oil is lower. If the demand for oil were unit elastic (a very special assumption,) then a doubling of the price of oil would lead to a halving of the amount of oil produced, sold, bought, and consumed. The amount spent on oil would be the same, and so, the amount left to spend on everything else would be the same too.

    If the demand for oil is elastic, then more is left to spend on other things. If the demand for oil is inelastic, then less is left to spend on other things.

    Your argument implicitly assumed that the demand for oil is perfectly inelastic (and the supply isn’t.) In that case, the increase in spending is in proportion to the increase in price, and less is available to spend on everything else. It is more or less like the situation Sumner describes above, where people choose to spend less on one thing and more or something else.

    Anyway, these sorts of issues relate to why slow steady growth of spending on output is better than a stable price level. It is better for the price level to change when there is a shift in supply.

  3. Gravatar of J.V. Dubois J.V. Dubois
    27. April 2012 at 06:22

    This is as good opportunity as any to ask the following question: how do you calculate real GDP from investment spending (or non CPI goods/services)? Let’s say that we know that people spend 1 billion dollars more on houses this year compared to the last year. If you use chain volume to construct GDP deflator for calculating real growth, what do you use? Square feets? Do you correct it for things like energy efficiency or quality and durability of building materials?
    .
    I mean, you can potentially measure the real growth sometimes in the future. For instance you can calculate the inputed costs of sheltering services now, but by doing that you just average the efficiency of the whole stock of housing in providing these sheltering services. Maybe the new houses are vastly better at this job, but for now their impact is muted by the sheer volume of old houses in the housing stock.
    .
    Imagine this example. Let’s say that up til now the averagage house lasts 100 years after which it has to be completely renovated with the same costs as building the new house. And imagine that this year there is advance in technology that will make the houses last 200 years. Initially, you will not see that much impact of this breakthrough. However year by year the total available stock of houses will increase because new houses detoriate slower. This means that year by year sheltering services are getting cheaper in real terms and we experience growth in the GDP.
    .
    So in short, at best you can calculate the real the GDP with a lag. Or in other words, large part of what you currently see as increase in GDP could actually be results of advances made years or even decades ago.

  4. Gravatar of Y.Alekseyev Y.Alekseyev
    27. April 2012 at 06:28

    That’s both insighful, interesting and provides an great opportunity for a counterfactual.

    Government spending subtracted 0.4% of GDP. Suppose, instead, that the fiscal side had instead pursued an expansionary policy resulting in positive contribution to GDP from that particular sector. Do you think that (1) the expansion by the gov’t would be immediately offset by contraction in other sectors without any change in stance by the Fed? Or are you suggesting that (2) the Fed would tighten money which would in turn result in an offseting contraction in NGDP in other sectors?

    If your claim is (1), you seem to be saying that crowding out at current conditions would still occur. I’ll leave that without further comment because I just don’t see it.

    If you claim is (2), we need to unpack that further. Why would the Fed tighten? Is it because (3) a faster NGDP increase is not possible without an increase in inflation past 2%? How can that be? (And isn’t (3) the same as (1)?).

  5. Gravatar of Bill Woolsey Bill Woolsey
    27. April 2012 at 07:03

    Dubois:

    The dollar value of the capital goods are added up, and then it is deflated by an index of the dollar prices of the capital goods.

    http://www.bea.gov/national/pdf/ch6PFI%20for%20posting.pdf

    If you read down, you can see that they do make adjustments for “quality,” but it seems a bit ad hoc.

  6. Gravatar of Saturos Saturos
    27. April 2012 at 07:04

    “… government is just another sector”

    Exactly.

  7. Gravatar of dwb dwb
    27. April 2012 at 07:35

    … and by the way the government sector is mostly contracting. some much for Keynsian fiscal stimulus as the only option in a “liquidity trap”

  8. Gravatar of dwb dwb
    27. April 2012 at 07:41

    … how Krugman cannot look at this graph as prima facie evidence monetary policy works at the ZLB and the Fed is merely unwilling to further stimulate the economy, I dont know: ~4% -4.5% nominal growth despite fiscal contraction. I think/hope he’ll come around.

    http://krugman.blogs.nytimes.com/2012/04/27/the-secret-of-our-non-success/

  9. Gravatar of Saturos Saturos
    27. April 2012 at 07:51

    By the way Scott, how then do you advocate the Fed target the forecast for NGDP, in the absence of a futures market? How does the market signal RGDP growth?

  10. Gravatar of J.V. Dubois J.V. Dubois
    27. April 2012 at 08:01

    Bill: Thanks for that link. I am just thinking, that if Market Monetarist want to create something like NGDP futures market, maybe there should be more thought put into its underlying components. If for example some market participants in NGDP futures market knows about some real shock in an economy and he knows the CB reaction function, he should be able to predict the impact on future NGDP in some meaningful way. Otherwise there is no point into bringing market in – since it is information (about price) what is the trademark of free market. But maybe I am off and this is actually a non-issue.

    And just as a side note, I was wandering about overall impact of investment and consumer durables on the quality of life. The trouble is, that GDP (real or nominal) measures just a flow of newly created goods and services. But I do not know if there is any measure of stock of durable goods. In this way there may be a lot of potential life improvements left completely unobserved. Like for example impact of e-bay – a tool for people exchanging old used goods that would be just discarded if there was no e-bay. Or on the other side, we could measure impact of “anti-technology”, such as (in my opinion) advertisement and marketing techniques that can convince people that the goods they own are morally obsolete and that they provide less valued service then just a while ago.

  11. Gravatar of ssumner ssumner
    27. April 2012 at 08:30

    D. Gibson, Yes, that’s right. Never reason from a price change.

    Bill, You seem to be assuming that the change in oil prices is due to supply, but in recent years demand has been the key factor.

    JV, I agree there are huge measurement issues here, but am not an expert on the techniques they use.

    Y, Alekseyev, You said:

    “If you claim is (2), we need to unpack that further. Why would the Fed tighten? Is it because (3) a faster NGDP increase is not possible without an increase in inflation past 2%? How can that be?”

    How can that be? Doesn’t every single macro model ever developed predict that faster NGDP growth will raise inflation? (which is currently near the Fed’s target, according to Bernanke)

    dwb, Good point. Interestingly US fiscal “austerity” is just as phony as British fiscal “austerity”. Britain’s in recession and our unemployment rate is falling. I wonder how Krugman would explain that difference.

    Saturos, I’d favor a policy where the Fed targets its internal forecast on NGDP. If you do level targeting then small errors are no big deal. If you do growth rate targeting then small errors are a big deal–another reason for level targeting.

  12. Gravatar of Karl Smith Karl Smith
    27. April 2012 at 08:37

    Scott:

    Never reason from a spending change 🙂

    Sector analysis helps us understand whats going on because we can decompose the forces more easily than in the economy as a whole.

    If I see that more and more home and car loans are being applied for, or that rental vacancies are falling, etc then I can surmise that cash balances are relatively high and that people are willing to take on higher long term payments to spend down those balances.

    However, if I see that a warm winter is driving down utility bills then this tells me nothing about the overall spending attempts. Its unlikely that if cash balances grow because of low heating bills that households are going to try to get rid of that cash all at once.

    So our sector analysis tells us what’s happening to cash if we don’t just take it at face value but try to understand the forces involved.

  13. Gravatar of Y.Alekseyev Y.Alekseyev
    27. April 2012 at 08:53

    Wait: which models? The claim that it’s outright impossible to deploy existing IDLE productive capacity without a change in price level seems weird. It’s weird at the aggregate level (to wit: UK used to have HIGHER NGDP at LOWER price level 4 years ago) and it’s weird at the level of primitives (I don’t need to increase wage for my existing workers if I am employing someone hitherto without a job).

    Or are you denying that there is pervasive idleness?

    It seems you’ve gotten into the crowding out game without realizing it.

  14. Gravatar of Saturos Saturos
    27. April 2012 at 08:58

    Scott, so the Fed can make reasonably accurate RGDP or NGDP forecasts on its own without making fallacious sector-based assumptions? I ask because, at a recent presentation by a central banker on growth prospects for Australia, she sure seemed to do a lot of sectoral analysis.

  15. Gravatar of The GDP release and how some seem to be content with what we´ve got. | Historinhas The GDP release and how some seem to be content with what we´ve got. | Historinhas
    27. April 2012 at 09:34

    […] a related post, Scott Sumner reminds us that doing a “post mortem” of GDP components is not illuminating: This […]

  16. Gravatar of Measure for Measure Measure for Measure
    27. April 2012 at 13:08

    The Fed can do level targeting. I can do level targeting: anybody can do level targeting. But will the economy listen?

    If I understand Scott’s perspective correctly, the Fed should keep buying bonds until it hits its announced target. If it misses, buy more! If necessary, buy non-treasuries. I am not convinced that this will work in the absence of strong fiscal policy, fixing balance sheets, or reforming the sclerotic banking sector. But I would certainly be willing to try it. It seems to me though that using one tool for economic stabilization forces residential housing to become more volatile than it would under a more balanced regime. And that’s if targeting works, which it may not.

  17. Gravatar of James James
    27. April 2012 at 14:45

    “If consumers spend more in one area they spend less in another area. What you need to do is change the total amount that consumers are spending, and only the Fed can do that.”

    Why should I believe this when there have been many times in history where consumer spending increased without any increase in the money supply?

  18. Gravatar of W. Peden W. Peden
    27. April 2012 at 15:21

    James,

    The money supply and NGDP (i.e. total expenditure on final goods) are two different aggregates.

  19. Gravatar of James James
    27. April 2012 at 19:05

    W Peden,

    My comment above never claimed that NGDP and the money supply were the same. Were you meaning to respond to someone else? Maybe you can tell me why I should believe that only the fed can increase total consumer spending.

    Should I seriously believe that no other agent or process in the economy besides the Fed can increase total consumer spending?

  20. Gravatar of Saturos Saturos
    27. April 2012 at 21:15

    James, you said:

    “Why should I believe this when there have been many times in history where consumer spending increased without any increase in the money supply?”

    That certainly sounded like you were confusing the money supply with NGDP. Of course, even with the money supply constant, velocity might increase, and this would raise total spending. But the Fed controls that process as well.

    Total spending (including total consumer spending) equals the money supply times its velocity. The velocity depends on money demand, which in turn depends partly on exogenous factors, but mostly on expected NGDP. In the long run expectations of NGDP are determined by the Fed’s behavior, since monetary injections boost NGDP proportionately. So the Fed controls both the money supply (directly) and its velocity (indirectly).

    You are right in saying that the Fed isn’t the only agent in the economy that can boost total spending. For example, if the Fed holds the money supply constant, while the Treasury conducts deficit spending, this will raise rates sufficiently that people switch from holding money balances to Treasury bonds, and velocity will increase. However, if the Fed targets inflation, then the central bank will tighten to offset the fiscal expansion. So monetary policy is the ultimate determinant of nominal spending.

  21. Gravatar of W. Peden W. Peden
    28. April 2012 at 02:02

    James,

    As Saturos says, in a system of fiat money with a central bank, total spending is dependent on the decisions of the central bank.

    Anyway, before we proceed further, I’d like to know what everyone means by “consumer spending” here. Does it mean consumption? Total expenditure? What?

  22. Gravatar of James James
    28. April 2012 at 06:36

    Saturos & Peden,

    When I mention that two variables have, in the past, moved independently of one another that should be your first sign that I recognize them as separate and distinct.

    Before fiat money, nominal spending mostly rose and sometimes fell. Does this not show that it is not only the Fed which can increase total spending?

  23. Gravatar of W. Peden W. Peden
    28. April 2012 at 06:41

    James,

    I see your point now and agree.

  24. Gravatar of Saturos Saturos
    28. April 2012 at 10:16

    James, you said:

    “Before fiat money, nominal spending mostly rose and sometimes fell. Does this not show that it is not only the Fed which can increase total spending?”

    Yes, but now that it does exist, the level of nominal spending fully depends on the expectations set by its policies.

  25. Gravatar of ssumner ssumner
    28. April 2012 at 11:03

    Karl, You said;

    “Sector analysis helps us understand whats going on because we can decompose the forces more easily than in the economy as a whole.”

    For supply shocks, but demand shocks are aggregate, not sectoral. If we know what happened to NGDP, sectors add no additional information.

    You said;

    “If I see that more and more home and car loans are being applied for, or that rental vacancies are falling, etc then I can surmise that cash balances are relatively high and that people are willing to take on higher long term payments to spend down those balances.”

    Not for any given level of NGDP. If spending rises by say 4%, and one sector sees much faster growth, than that just means another sector saw slower growth. In addition, the spending of cash doesn’t affect aggregate cash balances, it just moves the money from one person to another. It’s possible that loan applications are a leading indicator of NGDP (although I doubt it) but that’s different from looking at actual spending data.

    Y. Alekseyev, You asked:

    “Wait: which models?”

    Umm, AS/AD models from EC101.

    Stauros, That’s really just empty talk for the most part. Of course there are exceptions. If you are a huge producer of commodity X, and its price shoots up overnight, you can probably raise your NGDP forecast. But that’s not really looking at sectoral output, it’s looking at specific prices. I think the Fed mostly looks at markets when making forecasts. That’s why they cut interest rates after the 1987 stock crash, when macro variables said no rate cut were needed.

    Measure for Measure, It’s not a question of using one tool or two, it’s a question of which monetary policy is optimal. If you don’t like NGDP targeting, tell me what you think the Fed should target. My guess is you can’t answer that question, which is exactly the problem with our current monetary policy.

    James; You said;

    “Why should I believe this when there have been many times in history where consumer spending increased without any increase in the money supply?”

    Bernanke says the money supply is a poor indicator of monetary policy, you need to look at variables like NGDP and inflation. I agree. Even if prices increase without more money, monetary policy is still driving the price level under a fiat money regime. Monetary policy is the change in the money supply relative to changes in money demand.

    If the Fed doesn’t exist, or if we have a Fed and a gold standard, then naturally other factors drive NGDP.

  26. Gravatar of James James
    28. April 2012 at 15:08

    Saturos,

    Why should I believe that only the fed can increase total consumer spending? Once you cover that, please let me know why I should believe your new assertion that “Yes, but now that it does exist, the level of nominal spending fully depends on the expectations set by its policies.”

    Scott,

    You have not answered the question which you managed to quote verbatim.

  27. Gravatar of Passing By Passing By
    28. April 2012 at 19:05

    Professor Sumner –

    The recession hit some industries far harder than others(e.g., motor vehicle output fell 40%, food and beverage only 2%). Such an uneven pattern calls out for some logical explanation–and that same explanation should should also account for the pattern of recovery.

    This by itself seems to justify a strong interest in the sectoral pattern of recent growth.

  28. Gravatar of Major_Freedom Major_Freedom
    28. April 2012 at 20:41

    Passing By:

    The recession hit some industries far harder than others(e.g., motor vehicle output fell 40%, food and beverage only 2%). Such an uneven pattern calls out for some logical explanation-and that same explanation should should also account for the pattern of recovery.

    Exactly. This is the question that drives market monetarists nuts.

    Just look at this chart.

    Notice how construction industry was hit worse than manufacturing in general, and MUCH worse than non-durable goods manufacturing?

    Also, notice how durable goods manufacturing is where most of the recovery has taken place, and that the other two sectors have seen almost no improvement?

    These events, incidentally, are entirely consistent with the “re-calculation”, “Austrian” story of busts, but the Keynesians and market monetarists, who only think in terms of AD / NGDP have no explanation for why construction was hit so much harder than the other sectors, and they have no explanation for why durable goods has risen whereas construction has not and non-durable goods has only slightly recovered.

    Recalculationists: 54,645 + 1

    Monetarist/Keynesians: 0

  29. Gravatar of Major_Freedom Major_Freedom
    28. April 2012 at 20:59

    Passing By:

    Actually, this chart is even better.

    This chart includes retail.

    If aggregate demand were the problem, then why did construction slump so much more than retail, and why did retail slump less than manufacturing?

    Only the recalculationists/Austrians have a sound explanation for this. The AD/NGDP crowd are clueless.

  30. Gravatar of Major_Freedom Major_Freedom
    28. April 2012 at 21:06

    The reason why retail has improved since 2010, is because too many scarce resources were redirected away from retail and into construction during the boom. This is “Austrian” malinvestment brought about by artificially low interest rates. Artificially low interest rates mislead investors into allocating scarce resources into the higher order industries such as construction, when consumers have not voluntarily saved enough to warrant such a redirection.

    Therefore, coming out of the correction/bust period, we should see scarce resources allocated back into the retail industry where they should have been all along, and we should see a decline in the construction industry, where scarce resources should not have gone.

    And wouldn’t you know it, that’s exactly what happened.

    Can any market monetarists here explain this using “NGDP” rising and falling?

  31. Gravatar of Passing By Passing By
    29. April 2012 at 06:52

    Major Freedom -“Only the recalculationists/Austrians have a sound explanation for this”

    Actually, there are other plausible explanations, such as a constriction on credit availability. But we agree that such an important pattern needs some good explanation

  32. Gravatar of Passing By Passing By
    29. April 2012 at 06:52

    Major Freedom -“Only the recalculationists/Austrians have a sound explanation for this”

    Actually, there are other plausible explanations, such as a constriction on credit availability. But we agree that such an important pattern needs some good explanation

  33. Gravatar of Passing By Passing By
    29. April 2012 at 06:52

    Major Freedom -“Only the recalculationists/Austrians have a sound explanation for this”

    Actually, there are other plausible explanations, such as a constriction on credit availability. But we agree that such an important pattern needs some good explanation

  34. Gravatar of ssumner ssumner
    29. April 2012 at 18:22

    James, I was questioning your claim that if prices increase without an increase in the money supply, then the Fed didn’t cause that increase. I don’t agree.

  35. Gravatar of James James
    30. April 2012 at 14:05

    Scott,

    I don’t recall making the which you attribute to me at 18:22 on 4/29.

    In yout original post, you say that “only the Fed” can change the total amount that consumers are spending. What evidence convinced you that the only the Fed can do this?

  36. Gravatar of Scott Sumner on Smithianism « Modeled Behavior Scott Sumner on Smithianism « Modeled Behavior
    1. May 2012 at 04:35

    […] writes The new GDP figures offer a reminder that one can’t analyze movements in GDPby looking at […]

  37. Gravatar of Major_Freedom Major_Freedom
    1. May 2012 at 09:08

    Passing By:

    Major Freedom -“Only the recalculationists/Austrians have a sound explanation for this”

    Actually, there are other plausible explanations, such as a constriction on credit availability. But we agree that such an important pattern needs some good explanation.

    Why should credit availability be constricted differently for different sectors? If it’s a risk issue, then the interest rates can be different, problem solved. If it’s a regulatory issue, then why are the same general pattern regardless of the administration?

    I don’t see credit availability to be the answer, so maybe you can show what you mean.

  38. Gravatar of ssumner ssumner
    4. May 2012 at 16:13

    James, Because they control the money supply and hence NGDP.

  39. Gravatar of Measure for Measure Measure for Measure
    8. May 2012 at 12:02

    “Measure for Measure, It’s not a question of using one tool or two, it’s a question of which monetary policy is optimal. If you don’t like NGDP targeting, tell me what you think the Fed should target. My guess is you can’t answer that question, which is exactly the problem with our current monetary policy.”

    Scott – Thanks for your reply and sorry for the delay. To be clear, I don’t oppose NGDP targeting, I’m just skeptical that it can be effective in the absence of robust fiscal policy, though I would be happy to live in a world where one-armed stabilization is adequate.

    Arguably, an inflation target would be just as good as nominal GDP targeting and better public relations: it sounds bad if the Fed is putting an explicit cap on growth. OTOH, a nominal GDP target might permit higher inflation (which I’d prefer) while reassuring hawks that there’s a limit to the upward spiral. In the end, I’m agnostic with leanings towards NGDP, but that’s a personal preference. And while I’m not confident that looser monetary policy would be especially effective when interest rates are zero, I don’t think it would do great harm either. So we should give it a try.

Leave a Reply