Friday, September 14, 2012

How Will the Introduction of the iPhone 5 Affect GDP?

Apple is about to introduce a new iPhone. Michael Feroli at JP Morgan thinks that the introduction of the iPhone 5 will make GDP go up significantly. Why? Feroli estimates sales for the iPhone 5 and adds that to GDP. That can't be right. Even worse, Paul Krugman, thinking like a student in a bad intro-macro class with a homework problem, says GDP will indeed go up. Worse than that, he thinks this has implications for fiscal policy, which got David Andolfatto justifiably perturbed.

The simplest way to think about the introduction of the iPhone 5, is that it's an increase in total factor productivity (TFP). The iPhone 5 does not appear to do anything that its predecessor did not - it just does it better. It seems Apple is just producing more of the thing, quality-adjusted, with roughly the same capital and labor inputs. In a world with flexible prices, real GDP will go up as a result, but the increase has to be very small - certainly much smaller than the JP Morgan estimate. But what happens in a world with sticky prices - the world in which Krugman lives? GDP is demand-determined, and demand does not change as a result of the change in TFP, so the change in GDP is zero. Further, employment falls, as we can now produce the same quantity of output with less inputs. This is surely not something Krugman would like.

34 comments:

  1. why can't you both be right?

    ReplyDelete
  2. I don't think it's possible in this case. You'll have to explain it to me.

    ReplyDelete
  3. Why is GDP only "demand determined" when prices are sticky?

    ReplyDelete
    Replies
    1. That's the way those models work, by assumption. You fix some prices, then you you need an equilibrium concept. In a New Keynesian model, for example, it's assumed that, if a firm did not have the opportunity to change its price this period, then it has to satisfy whatever demand comes through the door. Those are the rules of the game. In models without sticky prices, GDP is determined by whatever exogenous stuff we put in the model.

      Delete
    2. Stephen, you write,

      "In models without sticky prices, GDP is determined by whatever exogenous stuff we put in the model."

      In Keynes's GT (which is, of course, not the last word on the subject), prices are flexible and GDP is, I suppose you could say, determined by "exogenous stuff" in the form of autonomous spending by producers and (considerably less) autonomous spending by consumers.

      On this view, iPhone 5 could increase GDP by a non-trivial amount if, e.g., producers drew down their cash balances to invest in new "apps" for iPhone 5, and consumers drew down their balances to buy iPhone 5 and its forthcoming new apps.

      If there are unemployed resources of the right type, then you're likely to get some modest multiplier effect. If unemployed software engineers are hired to produce the new apps, they'll spend most of their new income on stuff they couldn't afford on the dole.

      I can imagine a reply that says, "well, look, people bought iPhone 5 instead of something else, so there's no net increase in consumption expenditure," and/or firms invested in iPhone 5 apps instead of investing in something else, which results in no net increase in expenditure.

      But these replies seem to have nothing to do with sticky prices, but are, rather, just variations of the so-called Treasury View, which assumes that every "new" dollar of spending just displaces a dollar that would have been spent on something else. No?

      Delete
  4. Looks like you are thinking about sticky wages. In a textbook sticky wage model, higher TFP will indeed cause GDP and employment to rise. It's got nothing to do with decisions about consumption expenditure though.

    ReplyDelete
  5. As I think about it, isn't the real argument that the introduction of the new iPhone brings additional spending into the present, an intertemporal choice by consumers, because of the perceived obsolescence of earlier models? I'm not sure that we can look at this purely as an effect on tfp, since many use it as a consumer product rather than a productivity-enhancing product (capital).

    ReplyDelete
    Replies
    1. Right. This is always an issue with consumer durables. Note also that people move spending forward by postponing the replacement of an old phone until the new one is introduced. Of course, new products are being introduced all the time, so a lot of these effects net out in the aggregate. To the extent that the iPhone represents a large fraction of expenditure on consumer durables, its introduction will matter for the time series of consumer durable expenditures. But this can't be a big deal.

      Delete
  6. Let me point out an area of disagreement with Prof Krugman. As you point out, one does not have to agree with the demand-driven model in order to believe that the introduction of the new iPhone will boost GDP (although it helps if you do).

    ReplyDelete
    Replies
    1. The problem is that any of these effects have to be so small as to be of essentially no interest.

      Delete
  7. Steve, calling any effect of the introduction of the iPhone an increase in tfp maybe the "simplest" approach, but it is also the most stupid :)

    If GDP is increasing because consumers are adjusting the path of their expenditure, that is a Keynesian effect, you and Andolfatto may look gingerly to each other, will not change that.

    ReplyDelete
    Replies
    1. Nope -- you're wrong because if they simply save less to consume more, then the interest rate must adjust and drive down investment. GDP does not go up unless TFP increases.

      Delete
    2. Try to be polite, Anonomino. You can do it. If all I have is a hammer, everything looks like a nail. If all I do is think about Keynes, I see him everywhere.

      Delete
    3. Stephen,

      You can't possibly endorse Anonomino's claim, offered without qualification, that when consumption goes up, saving goes down, the interest rate goes up, and investment goes down. Well, you can, but you shouldn't.

      What's the premise? Full employment of course. In the spirit of controversy, I'll say that if you've never really read Keynes, you'll keep repeating the Treasury View without realizing it.

      This should be easy. You've got a lot of people unemployed or underemployed (not on holiday), and you've got firms that would love to sell more at current prices. Under these conditions, why would you insist that increased consumption spending must reduce saving rather than increasing total output?

      "GDP does not go up unless TFP increases." Really? Doesn't this proposition assume that all "the factors" are already hard at work?

      To claim, or at least assume, that the interest rate balances the flows of saving and investment doesn't really fit the facts does it? Hasn't the Keynesian hammer at least cracked this nut?

      Yes, I'm talking about the short run because a theory that tells us that, after the storm has passed, the sea will be flat again, is not very helpful.



      Delete
  8. Dr. Williamson, I thought that TFP was a technological advancement that contributed to higher output. I don't see how a better iphone can contribute to higher output, but I may be confused about what TFP actually is.

    Also, surely demand for iphones will change, no? Some people are definitely going to rush and go get them, no matter how little of an advancement they are in reality. What am I missing?

    ReplyDelete
    Replies
    1. You're getting confused by the lines outside the Apple store. This is exactly what we mean by a TFP increase. With the same factor inputs, we can produce more output. If you quality-adjust the output, it's an increase in TFP.

      Delete
  9. I see no reason why the iPhone shouldn't increase output. To take an extreme example, suppose Apple was based in Iceland, and then came up with a fabulous new iPhone that people all over the world were willing to pay $500 a pop for. Wouldn't Iceland's GDP take off? Hasn't Silicon Valley grown rich on tech start-ups like Apple? Now perhaps Apple is small relative to the US economy, but that is a quantitative issue you did not get into in your post. Sticky prices seem to be a complete red herring here. Its not like people want to spend a fixed amount on iPhones and Apple suddenly became more productive at producing them. Rather Apple has come out with a new iPhone and millions of new customers are desperate to buy it. In fact if prices are sticky - ie Apple can't raise the price relative to the old model - they will be inundated with even more orders.

    ReplyDelete
    Replies
    1. "Now perhaps Apple is small relative to the US economy, but that is a quantitative issue you did not get into in your post. Sticky prices seem to be a complete red herring here."

      Agreed. Small potatoes. Red herring.

      Delete
  10. Oh god.

    Let's assume that people/firms/banks -- due to fear or uncertainty of the future -- is currently in saving mode. That is, although interest rates are virtually zero, they still wish to save, and they do so largely by holding cash (this would, for instance, show up as excess reserves at the FED ... ). Now, suppose that the introduction of the iPhone 5 makes these people open up their wallets a little, and start spending some of that hoarded cash.

    What happens to nominal demand, then? If nominal wages are slightly rigid, what happens to real profits? And if real profits increase, what happens to employment and output?

    Steve, do you seriously think that this crisis was a "supply crisis"? That is that unemployment skyrocketed to 10% because of global amnesia of technology, or because hikes in "marginal" tax rates. I'm sorry, but that is a ridiculous idea, and one of the key reasons why economists have become the laughing stock of the last 5 years.

    ReplyDelete
    Replies
    1. I was discussing iPhones, not the financial crisis.

      Delete
    2. Oh, I thought you were discussing the launch of a new iPhone in the midst of a financial crisis. My mistake.

      So I assume that you think my logic holds, then?

      Delete
  11. "If nominal wages are slightly rigid, what happens to real profits?"

    Since they aren't, the rest of your post is irrelevant.

    ReplyDelete
    Replies
    1. For the Apple example, wage rigidities (like price ridigities) are irrelevant. The more demand there is for the iPhone, the more profit Apple makes - they call up their Chinese producers to ask for more production, pay them an extra $200 (at most) per extra phone sold, and Apple pockets the extra $300 as profit, translating into higher US GDP.

      Delete
    2. "Since they aren't, the rest of your post is irrelevant."

      Hold on! They're not rigid? How come real hourly wages have increased during the crisis? How come real hourly wages increased during the great depression.

      Why are there papers like these: http://www.frbsf.org/publications/economics/letter/2012/el2012-10.html

      or these: http://www.ecb.int/events/conferences/shared/pdf/labour_market8/05_Alessandro_Barattieri.pdf?79dc261125a82617ea185f3f4a3f147f

      Only two types of persons would deny the existence of downward nominal wage rigidity: The die hard freshwater neoclassicals who believe that business cycles (and outright depressions) are caused by "TFP" shocks or marginal tax changes, and nothing else; or crazy post-keynesians that for some reason thinks that whatever keynes said it must be the truth.

      Delete
    3. Anon,

      You mention "crazy post-keynesians that for some reason thinks that whatever keynes said it must be the truth."

      I'm familiar with the Post-Keynesian literature and I can't imagine who you're talking about. Can you give us a few names?

      Keynes thought nominal wages were sticky simply as an empirical observation. And he thought matters would be worse if, during a downturn, nominal wages fell. So what, exactly, are these "crazy post-keynesians" of yours supposed to believe about sticky wages?

      Delete
    4. Greg,

      Fine. I suppose you believe that wages are sticky, but that's a saving grace rather than an amplification of the crisis?

      Then the diehard neoclassicals deny the existence of rigid nominal wages, and the nutty post-keynesians deny their amplifying role. Better?

      Delete
  12. Anon.,

    Am I a "nutty post-keynesian" if I think falling wages would amplify the crisis? Is it "nutty" to think that falling wages would drag prices down with them? Is it "nutty" to point out that, in this case, real wages might not fall at all? Is it "nutty" to point out that falling prices will increase the real debt burdens? Was Irving Fisher also a nut case?

    It's one thing to imagine a Walrasian market in which prices adjust before any trades take place. It's quite another to infer that, in the absence this kind of market, the increased flexibility of some prices will always be a good thing. Was Frank Hahn a nut case for claiming that many members of the rational expectations school were helping themselves to the results of Arrow-Debreu without explaining the dynamics of how we get from a non A-D economy to the A-D results?

    ReplyDelete
    Replies
    1. Fisher's debt-deflation story sounds fine, but its relation to wages hinges on the idea that "falling wages would drag prices down with them", and that is not clear at all. But that's not a Keynesian story.

      The Keynesian story goes, when pressed, a little like this: If wages would fall, then investors will foresee them to fall even more in the future [unclear why], and they won't invest until wages have hit rock bottom. Thus falling wages will harm investments!

      Say what!? The soundness of that argument is similar to saying that investors won't buy stocks which pays out high dividends, as dividend payments are likely to be even bigger in the future. Investors will therefore wait to purchase stock until dividend payments peak!

      Yes. This is nutty. Nutty reasoning.

      Delete
  13. Anon,

    You’ve simply invented a nutty piece of reasoning and called it Keynesian.

    You claim that falling wages dragging down prices isn’t “a Keynesian story.” In chapter 2, Keynes criticizes the (neo)classical theory for thinking the wage bargain is made in *real* terms. He goes on to point out that labor may not even have the power to lower its real wages by reducing its nominal wage demands because, insofar as prices track marginal costs (in good neoclassical fashion), prices will also fall when wage costs fall.

    As a consequence of falling prices, there will be some redistribution of real income from labor to other factors, and from entrepreneurs to rentiers. Lower real wages will reduce labor’s marginal propensity to consume, while the transfer of income from entrepreneurs to rentiers is more ambiguous. Keynes, writing in the 1930’s, thought rentiers were the wealthier segment of the community, so this effect was also likely to be unfavorable.

    Finally, let’s consider your ridiculous claim that “the Keynesian story goes, when pressed, a little like this: If wages would fall, then investors will foresee them to fall even more in the future [unclear why], and they won't invest until wages have hit rock bottom.”

    Here’s what Keynes actually said, “If the reduction of money-wages is expected to be a reduction relatively to money-wages in the future, the change will be favorable to investment.” “If, on the other hand, the reduction leads to the expectation, or even the serious possibility, of a further wage-reduction in prospect, it will have precisely the opposite effect.”

    This isn’t a prediction about what will happen when wages fall, but a tracing out the consequences in the case of two different expectations regarding future wages.

    ReplyDelete
    Replies
    1. Greg,

      There are so many holes in that logic.

      First, why would lower real wages reduce labor's marginal propensity to consume? If anything, I would expect the opposite.

      And why would lower future "money wages" [why can't we just say nominal wages; "money wages" sounds like something a child would say] lead to a decrease in investment? Is it, as I said, that investors will hold back until wages hit rock bottom?

      "This isn’t a prediction about what will happen when wages fall, but a tracing out the consequences in the case of two different expectations regarding future wages."

      Yeah, which is equally interesting as saying "if a fall in wages is associated with a global pandemic, then falling wages will only exacerbate the decline in output". Either there is a theory underlying the relation between current- and future wages, and on future wages on current output, or there is none. Making statements of the type "if A implies B, and B is something bad, then A is bad." is vacuous unless we know that "A implies B". And we don't.

      Delete
    2. Anon,

      You originally argued that Keynesians hold the nutty idea that falling real wages won't increase investment because firms will expect wages to fall even further.

      In reply, I pointed out that Keynes laid out two possibilities: 1) the one you mention above; and 2) the opposite case where wages are expected to rise in the future.

      Now you can complain that, in the passage I cited, Keynes didn't give us any reason to favor 1 or 2 above. But that's quite different from insisting, as you did, that Keynes (or unnamed Keynesians) believed lower wages would always hamper investment because he was committed to just one of the possibilities outlined above, i.e., #1.

      I think, perhaps, were at loggerheads, in part, because you're talking about "nutty post-keynesians" without mentioning who they are or referring to their books or articles, and I'm citing Keynes's own work.

      Delete
    3. Anon,

      Regarding your first point, I was careless. I should have said that the fall in prices following a fall in wages would involve some redistribution of real income from labor to other factors whose remuneration was not reduced, and this effect, taken by itself, would reduce the MPC.

      Delete