Four saving fallacies

A recent comment section exposed a number of fallacies about saving:

1.  One fallacy is that one can prove a definition wrong by pointing to facts about the economy.  This is not correct.  Textbooks define saving as being equal to investment, i.e. saving is the funds used for investment.  Indeed this is part of another textbook definition, gross domestic income (C + S) equals gross domestic product (C+I).  One may not like those definitions, but pointing to real world examples to disprove them just won’t work.  Thus if someone says: “Suppose I put money in the bank, and the bank doesn’t invest the money,” it just makes my eyes glaze over.  I know immediately that I’ll disagree with your characterization about what’s happened to either S or I.  There is no debate about whether tautologies are correct, just about whether they are useful. If instead you say: “Here’s why I think a different definition would be more useful, more enlightening,” then my eyes will light up.   That is, until you start talking about the paradox of thrift . . .

2.  People confuse the individual with the aggregate.  In every case where an individual seems to be saving more and yet investment doesn’t rise, someone else is dissaving.  Thus when I loan someone (or institution) some money that they don’t invest, then I save and the borrower dissaves.   Aggregate saving is unchanged.

3.  What if I put cash under the bed?  I presumably get the cash from someone else.  So if me holding more cash is saving, then someone else holding less cash is dissaving.  If the government produces more cash and buys bonds, then it nets out to nothing if you view cash as a government liability.  The more interesting case is if we view cash as a real good, and the government feeds my appetite to hold more of this real good.  In that case it’s part of the capital stock but not a government liability, and real money hoarding means our real stock of transactions media goes up.  An OMP is both government saving and government investment.  That may seem an odd way to think about it, but it’s consistent with the definitions.  (Mike Sproul uses the liability approach; I use the real good approach.  S=I either way.);

4.  The argument for the paradox of thrift is that a higher propensity to save results in lower nominal interest rates, lower base velocity, and lower NGDP.  There are two reasons why I view this concept as being uninteresting:

a.  Even if correct, it would make more sense to call the problem “too much base money hoarding” not “too much saving.”

b.  It only applies if the Fed targets the money supply.  But they don’t, they target inflation (or inflation plus employment).  In that case the Fed would adjust the money supply to offset any change in V.  Now I suppose one could construct a model of “Fed fail,” but now we’d be far removed from the paradox of thrift, and would instead be obsessing about zero bounds and fear of unconventional policies, etc.


Tags:

 
 
 

69 Responses to “Four saving fallacies”

  1. Gravatar of Geoff Geoff
    9. April 2013 at 08:59

    Much appreciated, and much needed, blogpost.

    One of my top three pet peeves in economics is confusions over saving.

  2. Gravatar of Ritwik Ritwik
    9. April 2013 at 09:25

    The decision to forego consumption now does not constitute an equivalent decision to consume the same present value of goods and services in the future.

    This is true in general. Check Leijonhufvud and Effective Demand. And yes, the opportunity to transfer purchasing power effectively without a specific future consumption decision – economically a lower bound on the real rate on money – is indeed a decision contingent on the monetary regime. “Hoarding cash” as you say. But could be gold. Or real estate. Thus lack of effective demand will ultimately show up in excessive cash hoarding, true. But cash hoarding is not the causative mechanism.

    This is especially true if

    1) most investment is financed through retained earnings. When corporations turn net savers, standard macro takes a beating.

    2) production and consumption is simultaneous, i.e. services.

    Steve Roth’s Old Keynesian point is fundamentally correct. Let the paradox of thrift simply be called “in the aggregate, income adjusts”.

  3. Gravatar of TravisV TravisV
    9. April 2013 at 09:27

    Prof. Sumner,

    Tyler Cowen just wrote a provocative new post addressed to market monetarists:

    Some thoughts on recent Japanese monetary policy

    “We sorely lack a better understanding of how money matters when it does matter”

    http://marginalrevolution.com/marginalrevolution/2013/04/some-thoughts-on-recent-japanese-monetary-policy.html

  4. Gravatar of Bill Ellis Bill Ellis
    9. April 2013 at 09:30

    Scott
    All you are saying is that S=I Unless S is not invested. ( like going under a mattress. ) Which is what people were saying in criticism of your post on the paradox of Thrift.

    And just because the government can offset the dissaving does not make dissaving savings.

  5. Gravatar of J Mann J Mann
    9. April 2013 at 09:32

    Scott,

    You’re right, but Yglesias obviously doesn’t mean ill. It might be more charitable to have just jumped to the explanation.

    If I read Yglesias correctly, he’s saying two things:

    Foregone consumption isn’t necessarily spent on investment. He creates a new term “financial saving” for foregone consumption, and says “financial saving isn’t the same as saving real resources.”

    I don’t know what to make of this – it sounds almost more like an Austrian “malinvestment” theory than a velocity argument. Basically, he’s saying that when you and 99 other people decide to cut down on your weekly latte consumption and put that money into your IRA, Starbucks lays off a barista, but that your decision to put money in the bank does not result in any increase in the amount of “saving” under your definition, or “investment/saving real resources” under his.

    I understand correctly, is that if the Fed is holding NGDP or inflation constant, the resources that were spent making coffee will eventually reallocate to something else, but an argument over whether “saving real resources” means the same thing as “saving” is probably the easiest part of that, since you and Matt could theoretically agree to use any words you want.

  6. Gravatar of ssumner ssumner
    9. April 2013 at 09:42

    Ritwik, Obviously I don’t agree. Which of my 4 points do you reject?

    TravisV, Tyler needs to take my short course on money. 🙂

    Bill, No, I’m saying S is defined as being equal to I. Reread my post, I think you missed the whole point. S always equals I, no matter what the government does.

    J Mann, You said;

    “Foregone consumption isn’t necessarily spent on investment. He creates a new term “financial saving” for foregone consumption, and says “financial saving isn’t the same as saving real resources.””

    This could be interpreted in many ways. If he means actual saving, then it does result in actual investment. If he means intended saving, then it’s just the paradox of thrift, which I reject.

    Again, I think it’s a mistake to focus on saving, which is not the problem. Tight money is the problem.

  7. Gravatar of Andrew C. Andrew C.
    9. April 2013 at 09:56

    Scott’s points 1-3 are actually a description of what the paradox of thrift is all about. The thing that is paradoxical is not necessarily that increased saving reduces output, it’s that increased “saving” doesn’t necessarily increase savings. 😛

  8. Gravatar of Bill Ellis Bill Ellis
    9. April 2013 at 10:20

    Well I tried to see what you mean… still don’t get it. Seems like a semantics problem. I dunno.

    But related, and in an effort to get you… So it makes no difference that the wealthy are holding on to trillions in cash ? ( I am not sure if this is saving or money hoarding by your definition. )

    And on a bit different note, how about the tens of trillions held by America’s wealthy off shore ? Some on the right like to blame this on out tax policies, and claim that it hurts America.

    Seems like one can’t have it both ways… at least.

  9. Gravatar of Bill Ellis Bill Ellis
    9. April 2013 at 10:25

    Scott says…”Again, I think it’s a mistake to focus on saving, which is not the problem. Tight money is the problem.”

    But what if we make money easier (Something I am all for ) and the elite, having the same reasons to hoard or offshore money as before… just do more of it ?

  10. Gravatar of Ritwik Ritwik
    9. April 2013 at 10:28

    Scott

    0. Imagine an economy where the MEC crashes, for any reason.

    1. There is now an excess (relative) demand for bonds, and an excess (relative) supply of goods. What should happen?

    2. Maybe the cost of capital will adjust autonomously to restore I/S equilibrium. We have no issue.

    3. Maybe, as the cost of capital starts going down, before it has adjusted enough to restore full output and employment, people shift out of “bonds” into money. Now, we have an excess demand for money and an excess supply of goods.

    Then, recession follows.

    Should we say that the problem was excess demand for bonds as in 1 or excess demand for money as in 2. Roth likes calling it 1, you like calling it 2.

    Perhaps we should say that the real problem was the fact that excess demand for bonds transformed into an excess demand for money? Keynesians would call it a horizontal LM curve. You would call it central bank failure.

    Finally, how would things change if the excess demand for bonds relative to goods did not even put pressure on the cost of capital to adjust but directly resulted in an income adjustment?

    For example, if 80% of GDP was produced and consumed at the same time? Saving has little relevance here because consumption is not financed out of savings, but happens ‘jointly’ with production. So deferring consumption does not show up in the bond market at all.

    Or, if, a lot of production(investment) or income generation was contingent on corporations dis-saving (running down retained earnings). So that again, saving is not done OUT of income, but by foregoing production and income. Again, excess saving does not show up in the cost of capital at all – income adjusts before that.

    These are the dynamics that Steve Roth was hinting at.

    I don’t deny that

    1. The opportunity to “save” without making future consumption plans is contingent on the ability/willingness to hoard “cash”, which again depends on the real yield on cash.

    2. If your view of macroeconomic management is that whatever happens to the MEC, the job of a central bank is to ensure that the excess demand for goods or bonds never shows up as the excess demand for money, you would prefer to look at the dynamic as cash hoarding or money demand.

  11. Gravatar of Suvy Suvy
    9. April 2013 at 10:29

    Doesn’t a model of the paradox of thrift, saving, and investment need time as a parameter and a factor. People’s behaviors shift over time and not only do they shift, but their behaviors are reflexive. When an entrepreneur or business builds a factory, they are doing so based on long term expectations and prospects; however, costs are primarily driven by short term expectations. There is a path dependency(hysteresis) that happens between people.

    “Textbooks define saving as being equal to investment, i.e. saving is the funds used for investment.”

    The textbooks define it incorrectly. Savings is not the funds for investment; investment is what is turned into savings. Savings should be defined as income not consumed–not the funds used for investment. This is why the financial sector exists: to fund investment even if the existing savings aren’t there. Investment, as stated earlier, relies on long term expectations of the future, which are very liable to change because we cannot predict the future. So if the funding for investment isn’t there(banks don’t want to lend) or entrepreneurs lose confidence and just decide not to invest, this could trigger a “paradox of thrift.

    It’s impossible to talk about the paradox of thrift without including time in a model, without including the evolution of money, and without including the evolution of capital markets. The evolution of capital markets and the evolution of human behavior over time play a key role in determining investment, and thus savings.

  12. Gravatar of Andrew C. Andrew C.
    9. April 2013 at 10:51

    Bill, I find it simplest to understand accounting identities in terms of an apple-money only economy. Suppose 50 apples are produced in an economy, and the price of apples is $1.

    NGDP = (1$)*(50 apples) = $50.

    Suppose people only buy 40 apples – they save 10$. The remaining 10 apples is counted as an investment (increased inventories) so S=I holds.

    NGDP = C + I = C + S = 40$ +10$ = 50$.

    Now suppose the whole economy consisted of back scratches – then it no longer makes sense to talk about back scratches being produced before they are sold. There is no way to increase your inventories of back scratches. This is basically the point Steve Roth was making.

    In this universe, investment always equals zero, but so does savings. Any one person putting aside money must be offset by a reduced income – and therefore reduced savings of someone else. S=I holds again. It always does, because that’s the way it’s defined.

    The confusion lies in people using the word saving both for the behavior of saving, and the aggregate value of saving (S). The two meanings are different. When Keynesians talk about increased saving reducing output, they actually mean increased saving behavior, not necessarily an increased value of S.

  13. Gravatar of Lincoln’s Economics Lincoln's Economics
    9. April 2013 at 11:01

    What if saving and investment cannot operate properly in an open economy? An engineer has made the daring claim that there has never been a sound theory of economics and built a new model in its place. See rescuingeconomics.wordpress.com Constructive criticism is very welcome.

  14. Gravatar of Andrew C. Andrew C.
    9. April 2013 at 12:26

    By the way, the fact that I=S is not a fact that supports any side of the debate. It doesn’t say there is or isn’t a paradox of thrift, it just says that at the end of the day the two values are equal. I think Scott sometimes confuses people by making it sound like it’s a fact that supports his views, when it is in fact a totally neutral tautology.

  15. Gravatar of Doug M Doug M
    9. April 2013 at 12:33

    Savings = investment…
    What about foreign investment!

    I seem to remember the formula as

    Investment = Personal Savings + Corporate Earnings – Government deficits + foreign investment.

    I suppose you can think of each of those terms as “savings.” But, it is a little bit of a stretch.

  16. Gravatar of errorr errorr
    9. April 2013 at 12:38

    I think common insight of the true moneterists, mm, and post Keynesian types is the realization that things that were once exogenous are endogenous because of fiat money. The paradox of thrift once made sense because the economy was hobbled with golden fetters.

    I think that the paradox of thrift can help illustrate how mass preferences can affect the economy. If ‘too many’ people want to shift consumption forward in time then the market fails to clear because of sticky prices.

  17. Gravatar of Bill Ellis Bill Ellis
    9. April 2013 at 12:51

    Scott,
    Not that this pertains to this thread… But I was just thinking of restarting my old Word press blog and stumbled across this plug in for limiting comment length.

    http://wordpress.org/extend/plugins/gregs-comment-length-limiter/installation/

    I have no Idea if it could work for you or not. Hope it helps.

  18. Gravatar of errorr errorr
    9. April 2013 at 12:56

    I am still trying to construct a world where the paradox of thrift holds in a modern economy. Would banks have to refuse deposits forcing hoarding behavior? Are people assuming that the excess reserves held by the fed a form of hoarding? I guess you could somehow alter the laws of physics or permit time travel to shift back real resources from the future.

    It just seems to me baffling what the argument here really is.

  19. Gravatar of J Mann J Mann
    9. April 2013 at 13:21

    Errorr,

    It took me forever, but I think people who have a mental model for the paradox of thrift believe one of these things:

    1) Velocity: If I stop buying lattes and put my money in the bank or an index fund, those dollars allocate resources more slowly than they would in the consumption side. This reduces velocity, and if I assume that the central bank can’t or won’t correct (which I don’t), then there will be an effect.

    2) Zero Lower Bound Is Magic: If I understand Yglesias correctly, he thinks that investment is unrelated to the number of people who forego consumption and put their money in the bank. I think he thinks that the banks can’t find any new restaurants or start ups to loan money to at any price that covers their operating costs, so my increased deposits just result in bank hoarding. (So I guess this is velocity too). Again, if you assume the central bank can’t or won’t correct, then there’s an effect.

    3) Yglesias’s other argument seems to be based on following the workers. If I stop buying lattes and the bank hoards my money, what happens to the barista who gets laid off? Yglesias appears to believe that there is no other option – that there is a finite amount of jobs, and if the barista finds a new job, then someone else won’t. (I called this malinvestment earlier, buy maybe it’s really sticky wages).

  20. Gravatar of ssumner ssumner
    9. April 2013 at 13:21

    Ritwik, You say that I think it’s the central bank’s job to provide macro stability. I was under the impression that almost all mainstream economists believed that.

    Regarding saving vs money hoarding. Suppose the central bank holds the money stock constant:

    a. Lots of saving, no money hoarding, no problem.
    b. Lots of money hoarding no extra saving, problem.
    c. Lots of saving and lost of money hoarding, problem.

    What’s the common denominator in the problem?

    Suvy, You said;

    “The textbooks define it incorrectly.”

    Please read my post 10 more times, or until you understand it. Then come back and tell me what you are really trying to say. It’s like I’m just wasting my time here, people aren’t even reading what I say!!!!

    Hint, the textbooks can’t all be incorrect because (collectively) they DEFINE WHAT IS THE CORRECT DEFINITION. It would be like saying every single dictionary in the world misspells a word. No, if they all spell it that way than that’s it’s spelling.

    Andrew, I never say it supports my view of the paradox of thrift.

    Thanks Bill.

    errorr, Good comment.

  21. Gravatar of Fed Up Fed Up
    9. April 2013 at 13:56

    “Thus if someone says: “Suppose I put money in the bank, and the bank doesn’t invest the money,” it just makes my eyes glaze over. I know immediately that I’ll disagree with your characterization about what’s happened to either S or I. There is no debate about whether tautologies are correct, just about whether they are useful.”

    Try to be a little more open minded. They may have a valid point.

    “Textbooks define saving as being equal to investment, i.e. saving is the funds used for investment.”

    Does it say over what time period?

    “Thus when I loan someone (or institution) some money that they don’t invest, then I save and the borrower dissaves. Aggregate saving is unchanged.”

    If I save and then buy a financial asset from a bank or bank-like entity that can be used as capital (bond, stock, but not supposed to be a demand deposit), can dissaving be greater than saving so that aggregate saving is changed?

    “So if me holding more cash is saving, then someone else holding less cash is dissaving.”

    Let’s say no saving and no dissaving. Then someone saves $10 of “cash”. No one else dissaves. What happens?

    “If the government produces more cash and buys bonds, then it nets out to nothing if you view cash as a government liability.”

    But the seller of the bond may continue to save. The seller’s asset has changed from the bond to “cash”, but that is it.

    In 4), change base, base money, and money supply to currency plus demand deposits with the idea that the fed does not directly control demand deposits.

  22. Gravatar of Ritwik Ritwik
    9. April 2013 at 14:24

    Scott

    “..that the excess demand for goods or bonds never shows up as the excess demand for money..” is one interpretation of “macro stability”. It’s not the only one.

  23. Gravatar of errorr errorr
    9. April 2013 at 14:49

    Tangent: this is more of a Tyler Cowen type questions. is time travel existed, and you could send backwards in time a small fixed amount of investments, wouldn’t productive capacity immediately go to infinity. What would be the time delta backwards that would prove ideal? Or does it even matter. If the transmission was purely knowledge (please ignore the laws of thermodynamics) what would be the ideal time periods to send things back and how quickly would GDP grow? This gets in a weird place.

  24. Gravatar of Bill Woolsey Bill Woolsey
    9. April 2013 at 14:54

    When Sumner says that saving equals investment by definition, some of you are reading this to mean that if people decide that they want to reduce consumption and save more, and then actual start doing this, then firms will respond by purchasing more capital goods. You are reading his argument to be that choosing to reduce spending on consumer goods and doing it will result in firms buying more capital goods.

    That isn’t it.

    Andrew C. said it best with the service only economy. If the only good in the economy is a consumer service, then spending on the consumer service is the equal to income. There can be no saving. If people choose to consumer less in aggregate, then income will just be equal to the smaller amount they choose to consume. There is nothing left for saving.

    This is the paradox of thrift. People can all choose to save, but when they act, they end up with less income and no saving.

    So, in that economy, aggregate saving is impossible.

    My view is that this is vacuous.

    It is like defining supply to be the amount sold and demand to be the amount purchased and noting that they are always equal by definition.

    The supply of labor is the number of workers who are working and the demand for labor is the number of workers firm have working for them. They are equal by definition.

    Why do you talk about unemployment? The supply and demand for labor are always equal by definition!

    Thankfully, the demand for labor and the supply of labor are not defined in this way.

    And I don’t think the supply of saving and the demand for investment should be defined to be realized saving and realized investment with a claim that they are always equal.

    Like the supply and demand for a product or the supply and demand for labor, they should be defined based on the individuals, and then added up.

    What is important is an equilbrium where the supply of saving equals the demand for investment.

    Stating the actual saving equals actual investment no matter way is no more interesting that saying that the amount of goods bought equals the amount of goods sold. Or the amount of labor provided by workers equals the amount of labor utilized by firms.

    P.S. I worry that I am guilty of this sort of thing when I insist that income is always equal to output.

  25. Gravatar of Jim Glass Jim Glass
    9. April 2013 at 15:06

    So it makes no difference that the wealthy are holding on to trillions in cash?

    The wealthy *aren’t* holding on to trillions in cash. Nobody is holding on to trillions in cash. Trillions in cash does not exist.

    Total USA currency just slightly exceeds $1 trillion, singular, and a great deal of it is outside the USA — the Fed has estimated as much as 2/3rds is loctated abroad, leaving perhaps as little as less than $400 billion in the USA to service the entire US economy.

    Yet we constantly hear “Corporations are holding trillions of cash .. the rich are sitting on trillions of cash” … using the claim to make false arguments and reach false conclusions just as if it were true.

    As Will Rogers used to say: “It’s not what we don’t know that kills us, it’s what we do know that ain’t so.”

  26. Gravatar of Andrew C. Andrew C.
    9. April 2013 at 15:14

    Fed Up,

    I don’t think you understand the true banality of the statement I=S. It’s not magic – it’s not saying if I save an extra dollar, a business automatically and instantaneously takes the money and spends it on new equipment. It’s just a conservation law.

    To illustrate, say the entire economy consists of you and me. You grow bananas and I am your barber. And for some bizarre reason, we trade with money rather than barter. My consumption of bananas is your income, your haircuts are my income. Suppose I decide to increase my savings. My saving is defined as the part of my income that I don’t spend on bananas. Since I don’t control my income, the only way I can save more is by buying fewer bananas from you. That automatically reduces your (dollar) income, and along with it your savings also falls automatically. So my decision to save more resulted in no net increase in saving in the economy as a whole. S=0 and I=0.

    Except technically if you produced the same number of bananas, your savings did not fall because what you lost in dollar income, you gained in your stock of bananas. In this case, we say there was an investment of bananas equal to exact amount I chose not to buy (say, 10). So S=I=10 bananas.

    Obviously, this is a simple example, but the same accounting logic applies to the real world. You can argue with Scott about a lot of things, but it is completely fruitless to argue about I=S. It is truly as harmless as 1+1=2.

  27. Gravatar of phil_20686 phil_20686
    9. April 2013 at 15:33

    I had two points, firstly the paradox of thrift, secondly the S=I thing.

    1) The paradox of thrift always seemed obviously true to me, but reading your comments I must understand it to mean something different from you. My understanding is something like this:

    The paradox of thrift exists because certain real or nominal shocks can move the wicksellian (natural) interest rate. Since the monetary authority is not all knowing, there will exist a period where the current interest rate is some distance from the natural rate, and this must induce either an excess or a lack of saving, as by definition, the interest rate no longer matches savings preference with investment desire. Its more common for the natural rate of interest to fall in a recession than to rise, so you more commonly (possibly exclusively) get excess saving, hence the name.

    This is what I understood keynes to mean when he talks about the paradox of thrift. If it is not, then I have totally misunderstood the GT, and need to read it again.

    So I am fresh from reading the General Theory and Hawtrey’s criticism on Glasner’s blog. I am going to quote Hawtrey in full:

    “[A]n essential step in [Keynes’s] train of reasoning is the proposition that investment an saving are necessarily equal. That proposition Mr. Keynes never really establishes; he evades the necessity doing so by defining investment and saving as different names for the same thing. He so defines income to be the same thing as output, and therefore, if investment is the excess of output over consumption, and saving is the excess of income over consumption, the two are identical. Identity so established cannot prove anything. The idea that a tendency for investment and saving to become different has to be counteracted by an expansion or contraction of the total of incomes is an absurdity”

    Hawtrey’s criticism is, I think, valid, and is probably what most non economists are reacting to. Simply defining these two things to be equal sweeps many important and difficult considerations under the carpet.

    Moreover, if I were to make a more technical complaint, it would be that the tautology defined by Keynes (S=I), which links the income and expenditure modes of accounting, assumes that the income and expenditure occurs concurrently. In some sense this is always going to be a good approximation, since we generally look at accounting identities over long periods of time, when messy stochastic noisiness can be discounted. But especially on short terms it becomes a problem. Moreover, if you take a closed economy with NGDP = consumption + investment + government, and NGNI = disposable income + saving + taxes, then we assume like keynes that we can identify these terms term by term, then this makes a mockery of one of the main planks of monetarism. MM believe that a fall in NGDP (expenditure) will cause a rise in unemployment given sticky wages, that will cause NGNI to fall. Its tricky to be precise, given the role of expectations, but if you believe that the fall in NGDP causes unemployment, then you must, in some sense, believe that the fall in expenditure precedes the fall in income. In which case at least one of S=I, C=C, G=T and GNI=GDP, cannot be always and everywhere true, despite being identities. In essence they are defined to be true only in some idealised world in which all saving and investment decisions are made concurrently. Any model with a lag structure will instantaneously violate these identities, although they will, of course, preserve them over an appropriately defined “cycle”.

    They are, of course, a definition, and therefore cannot be `wrong’, but that is not the same as saying they are `useful’, it makes little sense to me that the GT on the one hand insists S=I, and on the other talks about matching savings preferences with investment preferences through changing interest rates/government spending. If S=I isn’t all that unnecessary? But of course that is not right. in keynes conceptual model S=I but both are at the wrong level. In other places he talks about `excess savings’, which is doubly confusing since he talks about the need for government deficits to soak up excess saving, which, by his definition, can only occur by lowering investment. Yes, I know that isn’t what he meant. All in all, i’m not sure why he bothered with this definition, it just seems to confuse everybody, except those who already knew what he meant to say. I found keynes whole style frustratingly obtuse, as if by occasional excessive precision he could make up for the frustrating vagueness of many parts of his argument. For all that, I did enjoy it.

    If your catch phrase is: never reason from a price change, can mine be: never reason from an accounting identity? 🙂

    PS: I think you misunderstood matt in your last post – surely he is only sawing that most current production is non-durable and so cannot be saved. A retired car worker does not get the engine block from a model T, the wheel base of a beetle, and the car door of a mx5, in that sense he has not “saved” his production. Instead of “real savings” you get claims on future production, which, depending on the state of the economy, may be more or less valuable. That is all he means by “financial saving”. I think.

  28. Gravatar of Peter N Peter N
    9. April 2013 at 15:37

    There’s nothing wrong with the definition of saving as used in calculating GDP, provided that you don’t apply it to saving meant in some other sense, which people do all the time.

    GDP excludes asset transfers, creates imputed income and mixes data from accrual and cash accounting. It also ignores money creation. GDP S = I both because it’s defined to and because discrepancies are reconciled.

    If you are talking about inter-temporal saving, then you mean something else entirely. Funding capital investment is yet another thing.

    It’s treating these as interchangeable that causes the confusion.

  29. Gravatar of Peter N Peter N
    9. April 2013 at 15:58

    @phil_20686

    “Instead of “real savings” you get claims on future production, which, depending on the state of the economy, may be more or less valuable. That is all he means by “financial saving”. I think.”

    exactly. Only non-financial assets can be transferred to the future.

  30. Gravatar of Peter N Peter N
    9. April 2013 at 16:18

    @Andrew C.

    And if I don’t produce bananas for inventory, then what? Besides fewer sold bananas means fewer haircuts. Now relative prices are determined by relative elasticity. Also you can’t put haircuts into inventory. That production is lost.

    This sort of thing always happens when you explicitly put time into such a model. And since both agents are both producers and consumers and there are no investment goods, this doesn’t say much about S=I.

  31. Gravatar of Suvy Suvy
    9. April 2013 at 16:24

    Prof. Sumner,

    Keynes defines savings differently than the textbooks do. Keynes defines savings as income not consumed; I do too. The reason I think that is a bad definition of savings is because much of investment is financed by debt.

    On a different note, Keynes did not view the interest rate as something relating borrowers and savers. He viewed it as a the level required for the marginal hoarder to put their money to work(whether this is putting it in a bank, lending it out directly, or buying assets). Keynes was saying that the level of savings and investment are equalized by shifts in the level of incomes.

    phil_20686,

    Keynes rejected the Wicksellian view of the interest rate. Here’s a paper where he talks about it.

    http://www.scribd.com/doc/11399026/Keynes-1937-Alternatives-Theories-of-Int

  32. Gravatar of Andrew C. Andrew C.
    9. April 2013 at 16:41

    Peter N.,

    We’re on the same side. I was trying to demonstrate the banality of saying I=S. It seemed like the commenter Fed Up didn’t understand how it could hold instantaneously.

    And if you look upthread, you’ll see I made the exact same point about services not being able to stored as inventory (except I used backscratches).

  33. Gravatar of Peter N Peter N
    9. April 2013 at 18:29

    @Bill Woolsey,

    Aggregate savings in financial assets is impossible. Aggregate savings in non-financial assets is possible. Most people don’t understand this, so it can’t be entirely vacuous.

    Many of our future problems come from the belief that you can save financial assets in aggregate.

  34. Gravatar of Fed Up Fed Up
    9. April 2013 at 19:26

    Only currency. No demand deposits. No banks. No financial assets other than the currency itself. No saving. No dissaving.

    $1 times 50 bananas me

    $1 times 50 haircuts you

    $50 stock of currency times 2 velocity = $100 NGDP

    $1 times 50 bananas me goes to $1 times 40 bananas so S = $10 for you and I = 10 bananas for me. They rot so I = 0 for me, but currency does not rot so S = $10 for you. “Me” decide to produce only 40 bananas. I as accumulated unwanted inventory is usually a bad thing.

    $1 times 50 haircuts you goes to $1 times 40 haircuts (“Me” no save, no dissave) so I = 0 haircuts (can’t store haircuts/service).

    $40 of currency still circulates with velocity of 2 = $80 NGDP. $10 has a velocity of zero. $80 NGDP divided by $50 stock of currency means the overall velocity of currency is 1.6 down from 2.0.

  35. Gravatar of Neal Neal
    9. April 2013 at 20:46

    I recently revisited the paradox of thrift. Isn’t it that if savings preferences change, consumption falls, investment falls proportionally, and that pushes the Y = E equilibrium point left on the Keynesian cross? But (I realized, silly me), if consumption preferences change, so does MPC, the slope of the consumption curve. So the consumption curve rotates down, savings rises, income doesn’t change.

  36. Gravatar of Andrew C. Andrew C.
    9. April 2013 at 21:30

    Sigh. Did anyone understand what I was trying to say? I’m not saying it wouldn’t play out like you’ve described. What I’m saying is that every step of the way, I=S holds. It’s not terribly interesting, in fact it’s incredibly dull — but that’s the point. I=S is an identity, not a theory or a model. It doesn’t contradict your idea of how the economy works, nor mine, nor Scott’s, nor anyone’s. So why are we still talking about this?

  37. Gravatar of Andrew C. Andrew C.
    9. April 2013 at 21:43

    Neal,

    If the consumption function shifts down, that pushes the desired expenditure function down and reduces the equilibrium level of output. But in this simple model, investment is exogenous, so it is unaffected by the change in consumption behavior. And since I=S, total savings in the economy also doesn’t change. That’s the paradox – saving behavior increased, but no increase in savings for the economy as a whole, not unless investment increases.

  38. Gravatar of Fed Up Fed Up
    9. April 2013 at 21:56

    Andrew C., if that is for me, I’ll try to answer.

    Let’s say 1 tree produced 1 banana. There are 50 trees for 50 bananas. The $10 saved in the MOA/MOE (currency plus demand deposits) caused me to cut down 10 trees. I consider that negative I. If there were banana workers, there would be either fewer hours worked per worker or layoffs.

    Later on, you decide to stop saving the $10. Now you want 50 bananas for $50, but there are only 40 bananas available.

  39. Gravatar of Peter N Peter N
    9. April 2013 at 22:16

    Start with

    S = S then

    S = S + I – I so

    S = I + (S – I)

    This is obviously a pure algebraic identity.

    We can read this as Saving is Investment + increase in net financial assets for the economy as a whole.

    Now if

    C + I + G + (X – M) = C + S + T

    then

    S – I = (G – T) + (X – M)

    We can read this as net financial assets for the economy can only change through net government spending and net imports. This makes sense since in the private sector every asset is created with a matching liability. If I buy stock in a company my new asset is a liability on their books, so the net effect is 0.

    If we take (G – T) + (X – M) = 0, then (S – I) = 0 and S must be equal to I. This is for the whole economy. If you take it as applying it to individual economic sectors or identify I with one sector and C with another, you’ll immediately have problems.

    For instance, say I dig a hole and bury $100 dollars has the economy saved it or invested it?

    Well (C – 100) + I + G + (X – M) = (C – 100) + S + T

    There’s no effect. Or say I destroy $100. Then you can say, if you like, that I’ve destroyed my $100 asset and the government’s $100 liability, and I’ve changed (S – I) by affecting the (G – T) term. The private sector has dis-saved $100.

    This all says nothing about non-financial assets, which are all that society can really be said to save.

  40. Gravatar of Peter N Peter N
    9. April 2013 at 22:23

    The problem with the Robinson Crusoe and Friday model is that

    G = T = X = M = I = 0

    So the macro equation reduces to C = C. There’s not a lot of traction there.

  41. Gravatar of J.V. Dubois J.V. Dubois
    10. April 2013 at 03:39

    Bill and others: I think Scott perfectly understands paradox of thrift, backscratching economy (I know that he reads Nick’s blogs), recessions and all that.

    But this SIMPLY WAS NOT what spurred this series of articles about what saving is. Go read Steve Roth where it all started here: http://www.angrybearblog.com/2013/04/saving-saving-resources.html

    I will even quote some interesting passages:
    ——————————————————-
    “He describes the opposite approach “” taxing returns on financial investments or “savings” “” as “morally grotesque.”

    Now let’s think about this, and think about how these economists think about this. They’re assuming that if you “save” (a.k.a. don’t spend), you don’t “consume resources.” You “save” them, and don’t “take them out of society.”
    ——————————————————-

    What was this discussion all about? It was about Steve roth preferring taxation of income instead of taxation of consumption. So even if talking withing Andrew C framework of pure service economy – why on earth should government tax $10 that somebody set aside to save? Do you want to say (as Steve Roth is) that since we live almost in pure Service Economy government should enact 100% tax on all savings to prevent dreaded “paradox of thrift”? Because I don’t.

    This is why Scott said “The day we start making long run optimal tax regime decisions based on their implications for the business cycle is the day we become a banana republic.”

  42. Gravatar of Neal Neal
    10. April 2013 at 04:27

    Andrew C.,

    The consumption function doesn’t translate down, it rotates downward because MPC has changed. Thus there is no downward pressure on investment.

  43. Gravatar of ssumner ssumner
    10. April 2013 at 05:53

    Fed Up, I don’t follow your comments. How can a financial asset be used as capital? And how can someone hold an additional $10 in cash without someone else holding $10 less?

    More broadly, I think you missed the whole point. People are confusing debates over definitions with debates over the facts of the economy.

    errorr, I’ll leave that for smarter people than I.

    Bill, You said:

    “When Sumner says that saving equals investment by definition, some of you are reading this to mean that if people decide that they want to reduce consumption and save more, and then actual start doing this, then firms will respond by purchasing more capital goods. You are reading his argument to be that choosing to reduce spending on consumer goods and doing it will result in firms buying more capital goods.”

    Exactly. I can set aside a $1000, and investment might not go up even $1, because AGGREGATE saving has not risen.

    phil, I disagree with most of your post, but especially this:

    “MM believe that a fall in NGDP (expenditure) will cause a rise in unemployment given sticky wages, that will cause NGNI to fall. Its tricky to be precise, given the role of expectations, but if you believe that the fall in NGDP causes unemployment, then you must, in some sense, believe that the fall in expenditure precedes the fall in income.”

    This is flat out wrong, NGDI=NGDP. It is not true that a fall in one must precede the fall in the other. If you think it does, they you have misunderstood market monetarism.

    I don’t think I misundertood Yglesias. You can’t beat something with nothing. If you have an alternative definition of saving, I’d be glad to take a look. But so far all I’ve seen is commenters confusing individual acts of saving with aggregate saving. And macro is about aggregates. If I take 30% of my income and buy a bond, my individual saving has gone up, buy aggregate saving has not, under any plausible definition.

    Peter, The confusion comes from people thinking about saving at an individual level, and then applying that (wrongly) to aggregates.

    Suvy, You said;

    “Keynes defines savings differently than the textbooks do. Keynes defines savings as income not consumed; I do too. The reason I think that is a bad definition of savings is because much of investment is financed by debt.”

    This is flat out wrong. Every textbook I’ve ever seen defines saving as income not consumed.

    Neal, Yes the paradox of thrift is basically the Keynesian cross, which has been dropped from the newer textbooks, for good reason.

    Andrew C. I feel your pain.

  44. Gravatar of interfluidity » A bit more on savings and investment interfluidity » A bit more on savings and investment
    10. April 2013 at 06:00

    […] Roth (1, 2), Scott Sumner (1, 2, 3), Bill Woolsey, and Matt Yglesias have been debating questions of saving versus investment and […]

  45. Gravatar of Steve Roth Steve Roth
    10. April 2013 at 07:21

    “In every case where an individual seems to be saving more and yet investment doesn’t rise, someone else is dissaving.”

    This only seems right if you’re imagining an isolated private domestic nonfinancial sector, in which no new financial assets can be created. (Essentially the “loanable funds” notion.)

    If you bolt on a financial sector that constantly creates new/additional financial assets, and (especially) a sovereign-fiat-money-issuing government sector, and account for flows to and from those sectors, I don’t think the statement is true.

    Because: “government saving” (in particular) is a meaningless concept, akin to a bowling alley “saving” points.

  46. Gravatar of Jim Glass Jim Glass
    10. April 2013 at 07:28

    This is one topic where Krugman anticipated you. From his Slate (pre-driven-mad-by-Bush) days…
    ~~~~~

    Vulgar Keynesians
    A penny spent is not a penny earned?

    …one of [Keynes’] unfortunate if unintentional legacies was a style of thought–call it vulgar Keynesianism–that confuses and befogs economic debate to this day…

    Consider, for example, the “paradox of thrift.” Suppose that for some reason the savings rate–the fraction of income not spent–goes up. According to the early Keynesian models, this will actually lead to a decline in total savings and investment. Why? Because higher desired savings will lead to an economic slump, which will reduce income and also reduce investment demand; since in the end savings and investment are always equal, the total volume of savings must actually fall!…

    Such paradoxes are still fun to contemplate; they still appear in some freshman textbooks. Nonetheless, few economists take them seriously these days.

    There are a number of reasons, but the most important can be stated in two words: Alan Greenspan.

    After all, the simple Keynesian story is one in which interest rates are independent of the level of employment and output. But in reality the Federal Reserve Board actively manages interest rates … so all the paradoxes of thrift, widow’s cruses, and so on become irrelevant. In particular, an increase in the savings rate will translate into higher investment after all, because the Fed will make sure that it does.

    To me, at least, the idea that changes in demand will normally be offset by Fed policy–so that they will, on average, have no effect on employment–seems both simple and entirely reasonable. Yet it is clear that very few people outside the world of academic economics think about things that way.

    For example, the debate over the North American Free Trade Agreement was conducted almost entirely in terms of supposed job creation or destruction. The obvious (to me) point that the average unemployment rate over the next 10 years will be what the Fed wants it to be, regardless of the U.S.-Mexico trade balance, never made it into the public consciousness. (In fact, when I made that argument at one panel discussion in 1993, a fellow panelist–a NAFTA advocate, as it happens–exploded in rage: “It’s remarks like that that make people hate economists!”)

    What has made it into the public consciousness–including, alas, that of many policy intellectuals who imagine themselves well informed–is a sort of caricature Keynesianism, the hallmark of which is an uncritical acceptance of the idea that reduced consumer spending is always a bad thing.

    In the United States, where inflation and the budget deficit have receded for the time being, vulgar Keynesianism has recently staged an impressive comeback…

  47. Gravatar of Asymptosis » Saving and “Government Saving” Asymptosis » Saving and “Government Saving”
    10. April 2013 at 07:54

    […] Randy Waldman and Scott Sumner (plus many others, linked from Steve’s post) wade in on notions of saving and […]

  48. Gravatar of Bill Ellis Bill Ellis
    10. April 2013 at 08:40

    Let me try again. It seems to me the folks saying that here is no Paradox of Thrift are lost in a self fulfilling semantics exercise.

    One observable example of Paradox of Thrift is the trillions ( and growing ) that the elite have been holding since the 2008.

    It is rational for them to hold cash because their is no demand for their wares. Why expand or employ more workers when there is no demand ? (Do the anti P of T folks have an alternative explanation ?)
    Even if this is a case where some are saving and it is being offset by some others dissaving…. it is still net drag on the economy. Right ?
    So are we left with the anti P of T folks thinking that holding that cash makes no difference because the cash they are holding (saving) is someone else’s income ?

    If so… Taken to an extreme… There would be no problem if people stopped spending any income and only saved.

    No one is saying that, right ?

    But they are saying that lessor degrees of that same problem are not a problem. Where is the cut off ?

  49. Gravatar of Peter N Peter N
    10. April 2013 at 11:52

    We’ll see if I get this right.

    Given

    S = I + (S – I)

    and

    S – I = (G – T) + (X – M)

    and assuming X = M

    (S – I) = (G – T)

    or increase in private sector net financial assets equals

    net government spending.

    But if money is a financial asset, then fiscal G – T always equals 0.

    or

    (S – I) = (G – T) – B = 0

    where B is government borrowing.

    Likewise Fed open market operations don’t affect S – I, since they exchange one financial asset for another.

    What increases net financial assets is monetizing the deficit. However –

    1) NIPA accounting disguises this, since it is designed to ignore financial asset transactions. An increase in financial assets doesn’t show up directly in GDP

    2) This ignores non-base money, since credit money involves creation of a private sector liability for every created private sector asset.

    3) This is a static relationship, and in the real economy all the aggregated transactions aren’t simultaneous.

    4) It’s a bit perverse to say that monetary policy works by affecting the return on assets while using NIPA – GDP added value accounting for your analysis, since NIPA ignores the changes in monetary value of existing assets. This may not be an insoluble problem, but it probably needs to be addressed directly.

    5) It matters when talking about S = I (and many other things) to say whether you’re doing added value or complete accounting.

    This doesn’t mean there’s a problem with monetarism. It may mean, however, that some of the steps in its analysis need to be filled in.

    Of course this isn’t very easy when dealing with things like expectations, but I don’t think you can expect NIPA tools to work very well out of the box to explain monetarist ideas. After all, NIPA was originally designed to model a barter economy.

  50. Gravatar of Jim Glass Jim Glass
    10. April 2013 at 13:10

    @ Bill Ellis

    One observable example of Paradox of Thrift is the trillions (and growing) that the elite have been holding since the 2008. It is rational for them to hold cash …

    You keep repeating that, ignoring that trillions in cash do not exist.

  51. Gravatar of J Mann J Mann
    10. April 2013 at 13:35

    Jim, maybe Bill is using a smaller unit. (Femtodollars?)

  52. Gravatar of Fed Up Fed Up
    10. April 2013 at 17:50

    ssumner said: “Fed Up, I don’t follow your comments. How can a financial asset be used as capital?”

    I said: “If I save and then buy a financial asset from a bank or bank-like entity that can be used as capital (bond, stock, but not supposed to be a demand deposit), can dissaving be greater than saving so that aggregate saving is changed?”

    Let me try it this way. If I save and buy a new stock or new bond from a bank or bank-like entity, can dissaving involving the bank or bank-like entity be greater than my saving so that aggregate saving is changed? A bond or stock of a bank or bank-like entity is bank capital. Demand deposits aren’t supposed to be bank capital but can be (see Cyprus). Bank capital is one thing needed to model banks or bank-like entities.

    ssumner said: “And how can someone hold an additional $10 in cash without someone else holding $10 less?”

    See my example about the bananas and the haircuts.

    I have $50 of income from bananas and spend it all. I hold $0 in cash. Andrew C. has $50 of income from haircuts and spends it all. $0 of cash held. Andrew C. saves $10 (holds $10 “under the mattress”) and spends $40. I now have $40 of income and spend it all. I hold $0 in cash. My income and spending went down.

  53. Gravatar of Neal Neal
    10. April 2013 at 20:59

    Scott,
    In the traditional paradox of thrift argument, consumption translates down, forcing I and Y to fall as well.

    My argument is that the Keynesian cross renders unparadoxical the paradox of thrift. If MPC falls, the consumption curve rotates instead of sliding downward. This *doesn’t* force Y to change. Where’s the paradox, then?

  54. Gravatar of Joe Eagar Joe Eagar
    11. April 2013 at 00:17

    The paradox of thrift has always relied on the assumption that central banks are powerless at the zero lower bound, haven’t they? In normal times, any change in saving will automatically produce a change in investment, as the central bank manipulates interest rates to stabilize the economy.

    Keynesians think fiscal policy is better at the zero lower bound, but if you think of final demand as a shared global phenomenon, fiscal policy by one nation isn’t all that effective if not matched by reinforcing fiscal actions by the nation’s trading partners.

    I’ve always wondered what would happen if fiscal policy followed a current account version of a Taylor rule. Fiscal and monetary policy would still target output, but one would target inflation, while the other would limit trade deficits/surpluses.

  55. Gravatar of W. Peden W. Peden
    11. April 2013 at 01:50

    Bill Ellis,

    “There would be no problem if people stopped spending any income and only saved.”

    If they did it voluntarily, is there any basis in economics to condemn this? So if everyone suddenly acquired a huge urge for leisure and self-sufficient farming and could pull this off without dying, then it what sense is there a problem?

    Maybe we might not like such a society, but then by definition we aren’t part of it.

  56. Gravatar of Saturos Saturos
    11. April 2013 at 02:02

    I think Bill is correct. (Surprise.)

    When there is a general glut, and we move back along the Keynesian cross (actual expenditures fall to meet planned expenditures) – it is then correct to say that what people have earned in the previous period is greater than what they wish to buy in the current period (even in real terms, since prices are sticky). What keeps the S = I identity going here is not that the excess demand for holding money (reciprocal to the deficient demand for buying goods) is somehow offset by equivalent “dissaving” by other people (people hold more in aggregate). Nor does the central bank have to be changing the money supply. Actually, what happens is that conventions of GDP accounting (in the context of which the relevant definitions of saving and investment are being used) dictate that unsold output gets counted as inventory investment by firms – even though no one wants to buy them.

    Thus excess desire by households to save rather than consume (relative to real investments demanded) is offset by excess “investment” in unwanted produce. With a sustained fall in demand, production falls to meet desired expenditures. Hence it is said (by Scott himself, on occasion) that the paradox is in fact where people who try to save more, actually end up saving less. But there is a sense in which the savings at natural GDP are “real” whereas the investments are not. Society has in fact “saved” more of its output in the aggregate – but those savings are in fact unwanted and will be thrown out as GDP declines. People get poorer as they are continually unable to sell all they can produce, due to the decline in the flow of money to trade it with.

    Keynes represented the cause of the downward spiral as an “animal spirit” shock to the desire to invest, even at zero rates, while people are so anxious that they refuse to demand consumption instead even at such low opportunity cost. Obviously it is inefficient to leave such capacity unused. What Keynes glossed over is how the continual desire to save at zero return was essentially excess demand for money, which could be met by increasing its real supply, either by creating more money or by lowering the nominal wage-and-price level. With the need for a higher money stock met there is no reason for people to go on trying to sell more than they buy, i.e. Say’s Law should hold again. Relatedly Keynes also failed to see that the “animal spirits” were in fact rational responses to expected future NGDP declines, causing present ones. This is caused by an excess demand for holding money in the economy (at the current NGDP path). That in turn can be due to different things: France hoarding gold, ECB not printing enough Euros, or a mixture of active and passive errors by the Federal Reserve as it failed to communicate an intention to stabilize nominal income by supplying however many dollars were required.

  57. Gravatar of Saturos Saturos
    11. April 2013 at 02:04

    (Obviously increasing the real supply of money was very difficult in 1930. Even more so when central banks were the ones increasing the demand for it.)

  58. Gravatar of Saturos Saturos
    11. April 2013 at 02:13

    Bill Ellis and W.Peden, saving is just intertemporal substitution of consumption. As is made *very* explicit in 3rd year microeconomics. From the point of view of “the economy” this means sacrificing consumer good consumption in the present in order to increase it later on – otherwise it is known as hoarding (putting stuff aside just to look at it forever, although theoretically economists might call that consumption too [we do that with art after all]).

    By the way I meant earlier on that Bill Woolsey was right, not Bill Ellis.

  59. Gravatar of W. Peden W. Peden
    11. April 2013 at 02:23

    Saturos,

    And in a society of monomaniacal misers, total aggregate saving becomes optimal. (Also, the economy, as a system of exchange, disappears.)

  60. Gravatar of Joe Eagar Joe Eagar
    11. April 2013 at 04:59

    Saturos, that’s a very interesting perspective on the subject. I’m very interested in what Dr. Sumner has to say.

    I’ve always taken the view that “real resources want to be used”, and thus increases in savings translate into either an increase in investment or an increase in net exports. Between falling real interest rates, real exchange rate depreciation, the desire of politicians to leave no stone unturned in their quest for spending, and bank monetary policy I don’t see how excess savings can be left underutilized for long (at least in the absence of major market distortions, whether self-inflicted or imposed by foreign central banks distorting the the global financial system).

  61. Gravatar of ssumner ssumner
    11. April 2013 at 07:03

    Steve Roth, You said;

    “This only seems right if you’re imagining an isolated private domestic nonfinancial sector, in which no new financial assets can be created. (Essentially the “loanable funds” notion.)

    If you bolt on a financial sector that constantly creates new/additional financial assets, and (especially) a sovereign-fiat-money-issuing government sector, and account for flows to and from those sectors, I don’t think the statement is true.”

    Nope, still 100% true–check any principles of econ text. It’s an identity. I think what you really want to debate is whether the definition used by economists is a sensible definition. Maybe not, but that’s the definition we use.

    Jim Glass, Ah, the good old days when Krugman ridiculed the paradox of thrift.

    Peter, National income accounting has no bearing on monetarism. Doesn’t support it, doesn’t oppose it.

    Fed up, Still not following your point, Maybe someone else can answer your question.

    Neal, If you are saying there is no paradox of thrift, I agree.

    Joe, Not all Keynesians believe that fiscal is better at the zero bound. Some favor a higher inflation target, which (if high enough) would eliminate the zero bound on long term bonds.

    Saturos, Good post. In response I’d simply point to my reply to ritwik (in my second comment in this comment thread.) which is a slightly different way of getting at the same issue.

  62. Gravatar of Neal Neal
    11. April 2013 at 19:55

    Scott, We agree. The Keynesian cross picture of the paradox of thrift contradicts itself.

  63. Gravatar of acarraro acarraro
    12. April 2013 at 01:50

    I had some fun and looked up some literature that disagreed with the classical Chamley/Judd model.

    I found the following:

    http://www.nber.org/papers/w13354

    http://www.jstor.org/discover/10.2307/2138707?uid=3738032&uid=2&uid=4&sid=21101892105773

    http://www.sciencedirect.com/science/article/pii/S0022053196922383

    http://www.sciencedirect.com/science/article/pii/S004727279900016X

    http://www.sciencedirect.com/science/article/pii/S0022053101928777

    I only spent few minutes on this, so some papers might be useless…

    I have re-read the original paper as well. The paper itself says that capital taxes are efficient in the short run and inefficient in the long run. There are additional caveats depending on production function, uncertainty, insurance value of wealth and so on…

    I would also like to draw your attention to utility function used in the model. Utility is non-complementary across time. This seems to me a weak hypothesis as it removes the main reasons I save: retirement. I actually want my consumption to be smooth and I know I’ll have less labour to sell when old.

    I don’t think eliminating that assumption invalidates the result completely but it certainly weakens it in my mind. It’s way beyond my maths skills to solve (so no Nobel for me I am afraid). I think a couple of the papers above try to investigate this avenue of research.

  64. Gravatar of acarraro acarraro
    12. April 2013 at 02:49

    As far as the two brothers story, my statement is the following: assuming that the net interest is higher than the inflation rate, the patient brother will consume a larger quantity of goods than the impatient one. This is really trivial and maybe a bit irrelevant.

    My main observation is that we need to make an assumption on the change in the market rate given a shift to a regime of no taxation. Your assumption is that rates would be unchanged.

    This is true if investment return is exogenous (e.g. saving seed corn to plant it), but it could be endogenous if it’s mostly coming from different time preferences. The brothers would trade present for future consumption as they value it differently, somehow splitting the benefit. At one extreme case the impatient brother is shouldering any increase in the interest tax (I guess he is both impatient and a crappy negotiator).

    I am not saying this is realistic, but it must be going on at least in part. I guess it’s a direct results from the implicit assumption on zero capital productivity (if time preference is the only reason to trade, capital goods must have no productivity).

    I guess that’s my core belief: there is a level of additional investment after which capital additions have no marginal productivity. If other forces create that amount of savings, there is no welfare loss. In addition if you consider G to be in large part investment (schools, roads, etc…) maybe we are already close to the amount of investment that can be successfully deployed. Maybe I suffer for lump of investment fallacy? Uhm… I don’t believe in the lump of labour…

  65. Gravatar of Suvy Suvy
    12. April 2013 at 17:47

    Prof. Sumner,

    You first said
    “Textbooks define saving as being equal to investment, i.e. saving is the funds used for investment.”

    Then, you said:
    “This is flat out wrong. Every textbook I’ve ever seen defines saving as income not consumed.”

    They are two completely different definitions then.

  66. Gravatar of Suvy Suvy
    12. April 2013 at 17:51

    Prof. Sumner,

    Just because savings=investment, that doesn’t mean that savings is the funds used for investment.

    “The investment market can become congested because there is a shortage of cash. It can never become congested when there is a shortage of savings.”–J.M. Keynes

    You’re going from savings=investment to savings is the funds used for investment. Those two ARE NOT the same.

  67. Gravatar of Angry Bear » Saving and "Government Saving" Angry Bear » Saving and "Government Saving"
    17. April 2013 at 04:09

    […] Randy Waldman and Scott Sumner (plus many others, linked from Steve’s post) wade in on notions of saving and […]

  68. Gravatar of stone stone
    19. April 2013 at 23:10

    Has this discussion adequately taken into account asset price inflation?
    S=I but I can be in the form of unsold inventory that perishes and goes to waste. That is the form of investment we get when financial savings come not from paying for capital goods or training staff and developing technology but from bidding up the price of pre-existing assets. Eighty percent of bank lending in the USA and UK is for buying pre-existing housing stock. The inflation in the cost of housing does not reflect deferred consumption. People will still need houses in the future. It looks to me like permanent non-productive saving down the sinkhole reflected in unsold inventory and unemployment.

  69. Gravatar of stone stone
    19. April 2013 at 23:37

    Further to my last comment, my fear is that all the monetary policy efforts to keep unemployment at bay actually (perhaps inadvertently)feed into asset price inflation. Sadly the financial system has become morphed into an ever more powerful and effective system for creating such counter-productive asset price inflation.

Leave a Reply