Believe those who are seeking the truth. Doubt those who find it. Andre Gide

Monday, December 27, 2010

Irrational exuberance over the balanced budget multiplier

Christmas time is the most magical time of the year. A time to believe in elves, talking reindeer, snowmen running amok, and...for some economists, the Keynesian cross.

The Keynesian cross. We (the economics profession) like to etch it deeply into the minds of fresh undergraduates, one cohort after another, year after year. Is it any surprise that for most educated laypeople, this is the only macroeconomic language they understand?

And here is a Christmas gift--from Professor Robert J. Shiller--to those of us who have been primed since youth to be receptive to this sort of message: Stimulus, Without More Debt. The argument for why a tax-financed increase in government spending will work is summarized as follows:
The reasoning is very simple: On average, people’s pretax incomes rise because of the business directly generated by the new government expenditures. If the income increase is equal to the tax increase, people have the same disposable income before and after. So there is no reason for people, taken as a group, to change their economic behavior. But the national income has increased by the amount of government expenditure, and job opportunities have increased in proportion.
In other words, the Keynesian cross (formal exposition available here). Econ 101 in action, kids!

So what, pray tell, is your beef with this, Mr. Grinch?

First, it's not that I have anything against the Keynesian cross, per se. I can appreciate the basic idea it is trying to convey. And it's just a simple model, after all--it seems silly to hold a personal grudge against an inanimate object. What I am against is in placing it (or any other economic theory, for that matter) on an exalted alter. Models should not, in my view, be worshipped in this manner. And while I'm on the subject of religion, I'm also against beginning an argument with a preordained conclusion (in this case, that more stimulus will certainly be needed, because unemployment is high).

Having said this, I think that the Keynesian cross is a delightfully perverted object. It can be (and has been) used to support almost any type of government appropriation. In fact, I feel like writing a letter myself to this end.
Dear Congressman:
The economy is in dire need of help. It needs to be stimulated. I am willing to stimulate it, with your help.
To this end, I ask that you appropriate a sum of $X from my fellow citizens and divert this money to me.
As this money does not belong to me, I promise to spend return it to my fellow citizens, so to speak. Of course, I will make them work for it...given the clear want of work in our present economic climate. The income so earned in exchange for their idleness will undoubtedly be spent--adding income to the pockets of everyone. No one will even notice the initial appropriation, as all of the money borrowed will be returned in the manner just described.   
Signed (your name); noble servant of society.
Now, try to imagine everyone writing this letter and that Congress acts accordingly. I hope you can see as well as I how nothing but good can come of this. Whatever the ailment, the cure, evidently, is to increase spending. Indeed, to force people to spend if they refuse on their own. When the Keynesian cross is your hammer, every macroeconomic problem nail looks like deficient demand. 

Second, it's not that I don't believe that an increase in G will lead to an increase in Y. There is evidence that it can. Heck, even standard neoclassical theory says it can. Whether it does or not in a given set of circumstances is a different matter. And even if it does, it is not entirely clear that increasing Y in this manner is socially desirable. It may be. Or not. It depends on a lot of things. I do not view the proposition as self-evident and beyond critical examination. In contrast, according to Shiller:
But the balanced-budget multiplier is simpler to judge: If the government spends the money directly on goods and services, that activity goes directly into national income. And with a balanced budget, there is no clear reason to expect further repercussions. People have jobs again: end of story.
(Don't you love it when you are granted license to stop thinking? End of story, indeed.)

Third, its not that I'm against increasing (components of) G. Public works projects of the sort mentioned by Shiller (building highways and improving our schools) were advocated by sensible economists long before Keynes (as evidence of this, note that public works were implemented in the Depression well before publication of the General Theory). What I have a problem with is in using some silly theory to support the notion, for example, that taxes should be raised to finance a large public capital expenditure. Shiller has been rightly celebrated for his work in the theory of finance, and on asset price bubbles in particular. But is this not a rather odd stand to take for a professor of finance?

Now, I'm no expert in finance myself, so maybe I should be careful in what I'm about to say. But it seems to me that a large capital expenditure should be financed with debt. The debt service could be supported by toll revenue (on bridges and roads) and user fees in general, backed by the Treasury, if needed. The use of tax finance advocated by Shiller in his balanced-budget exercise implicitly assumes (among other things) lump-sum taxes. For some thought experiments, the assumption of lump-sum taxes is innocuous enough. But this is not one of those cases. Taxes are distortionary and to the extent that they are needed to support public spending, they should be spread out over time. This is a standard principle of public finance (I think).

Maybe Shiller believes in this standard principle, but views it as politically infeasible (given the current appetite for debt reduction). Possibly. But if so, I would rather have expected a rousing defense of these standard principles. Americans are not necessarily against debt; they are against wasteful spending.  Given that America has an infrastructure (crumbling as it may be be) should be taken as evidence, I think, that people are generally willing to support worthy public enterprises--where worthiness is judged by a project-by-project cost-benefit analysis.

And speaking of standard principles, what ever happened to the quaint idea of evaluating the merit of public capital expenditure on a net present value basis, instead of some magic-multiplier concept? There is probably a good NPV case to be made for implementing such projects in a recession, even in the absence of positive externalities. Indeed, if what we read about America's "crumbling infrastructure" is true, these projects should have been started several years ago. Perhaps they were not because the economy was at that time judged to be "overheating." After all, the same Keynesian cross logic suggests decreasing G during a boom (crumbling infrastructure be damned). D'oh!...dang Keynesian cross.


  1. Need we say that NPVs will be higher at the moment because the discount rate is as low as it can get. Simple Keynesian analysis is just trying to say that the benefits of acting now are greater also. Unemployment is waste. Unemployment imposes costs on the unemployed, and on society as a whole. These costs are infrequently internalized into a NPV analysis in an accurate or meaningful way.

    In terms of increasing the tax burden- I think there is an outstanding political and moral case to improve the progressivity of the US taxation system. This doesn't imply "more inefficient taxes"- a progressive business expenditure tax system would be a great initiative to this end. But to have a situation where unemployment is at 9.8% while corporate profitbaility has returned to record or near record levels is morally unjustifiable. We can rationalize things as much as we like, but I wouldn't describe the current situation as politically or economically sustainable.

    timothy watson

  2. Timothy:

    Sure, I understand where the argument comes from. Not sure how the Keynesian cross helps in any substantive or practical manner. In fact, I think it potentially hurts. Shiller's article is an example of this (in my humble opinion).

    Btw, the appropriate discount factor is probably the real interest rate associated with 30 year treasuries. This is no where near zero, as your comment seems to suggest.

    May I ask a few questions? Just how progressive do you think the U.S. tax system should be? I posted this some time ago:

    I highlighted a few U.S. tax facts:

    [1] About 47% of Americans will pay no federal income tax in 2009
    [2] The top 10% of earners pay about 73% of all income tax collected by the federal government
    [3] The bottom 40% of earners will (on average) pay negative taxes (are net recipients of transfer income)

    I guess that this is not progressive enough in your view, and that's fine. But how much more progressivity would you like to see? Thanks for your input.

  3. The trends for the real interest rates on the longest term treasury debt (whatever class that is 20-30 year) have been strongly downwards post the Lehman collapse. So government debt is relatively cheap at the moment. I think that argument still holds, although recently yields appear to be jumping up again for these securities.

    As to progessivity, if the US is serious about fiscal sustainability without sinking the economy, the best way to do this is to get those with the greatest capacity to pay more.

    When I read commentary that suggests that the richest 1% of households in the US has a greater net worth than the bottom 90%, and the annual income of the richest 12,000 households is greater than the bottom 24 million households, I think there is ample scope to improve the progressivity of the taxation system without harming the economy.

    A serious way to do this would be through a progressive expenditure or Kaldor style taxation system. In terms of rate structures and the like I have no firm proposals for the US economy, but this is a serious reform proposal that should be considered for the long term.

    timothy watson

  4. I basically reject every analysis of fiscal stimulus that implicitly assumes lump-sum taxes. The impact of distortionary taxation is just far too important to sweep under the rug when considering practical policy analysis.

    This post actually got me to thinking about infrastructure stimulus and I thought of a potential problem. I presume we are doing infrastructure spending because we think it would add to the productive capacity of the economy. Let's assume that is true, wouldn't this actually be a bad thing for the purpose of stimulus? We know it takes forever to start a government infrastructure project. So, if people are expecting future productive public spending don't we get a wealth effect which implies a decline in labor? Similarly, if private investment takes less time to implement, as seems completely plausible, couldn't you see a fall in private investment. So, isn't it possible infrastructure spending could lead to a fall in labor and investment? Now perhaps there is something I'm not considering or if wages and prices are indeed sticky perhaps the rise in inflation expectations offsets this problem - but I think it's a theoretical possibility that the implementation lag of public infrastructure could be a very bad thing.

    But, what I figure is that if an infrastructure project is worth doing - then do it. I don't really need a fiscal stimulus argument to support quality infrastructure projects. Though I'd prefer infrastructure be done at the state level since I see no reason why people in another state should pay for something they will probably never use.

    Oh, and finally it's nice to see someone finally point out that increasing Y for the sake of increasing Y may not be a good thing.

  5. The supposed "U.S. tax facts" above are not facts at all. Ignored, for example, are the FICA taxes, state and local income taxes, property taxes, sales and excise taxes and so on. Moreover, the progressivity of the U.S. tax system depends on the net (i.e. all things considered) marginal tax rate as a function of income, not snippets like these from Fox and net marginal tax rates actually tend to decline as income rises. They certainly doesn't rise quickly as the right and their acolytes often try to suggest with such supposed "facts".

  6. Ted:

    I suppose anything is theoretically possible. At the end of the day, however, I agree with you completely that if an infrastructure project is worth doing, then do it. On a proper NPV basis. No white elephant projects for the sole purpose of creating temporary employment (may as well just transfer the money to people directly while they seek out more fruitful uses of their time).


    The link was to CNBC, if I recall. It's quite explicit about being limited to federal income taxes. So, if I read your comment correctly, what you meant to say that while these are facts, they do not encompass all the facts. This is a legitimate point. It would have been helpful if you could have provided facts of your own, rather than choosing the load road you did.

  7. They are not "facts" if "federal income taxes" means "taxes levied by the Federal government on income" and, moreover, you know as much, if only because you pay FICA taxes. If you want to interpret "federal income taxes" as "THE Federal Income Tax" then I'm willing to grant that your statements are correct (without actually checking the numbers) but, as I explained in my prior post, they are totally uninformative about the progressivity or otherwise of the US tax system. This, I believe, you also know. I stand by my claim that, all things considered, the US tax system system is, at most, mildly progressive.

  8. Anonymous:

    I didn't even know what FICA stood for (had to look it up). I am rather more familiar with the Canadian tax system; give me some time to get up to speed for the U.S.

    Now, you make the claim that the numbers I cited are "totally uninformative" about progressivity. This is a rather strong assertion, and I doubt that it is true. It may not be fully informative, but "totally" uninformative seems unlikely. But if this is truly what you mean to say, then back it up with numbers and citations to relevant sources. I am interested in learning more.

  9. Don't you think that you should have been be "up to speed" on the US system before making claims about its progessivity? I did mean "totally informative" as the amount of tax paid by any single person in the US depends on a lot more than just the Federal income tax. As you now know, there are the FICA taxes, state and local income taxes, property taxes, sales and excise taxes and so on. Not to mention the many deductions that, for the most part, benefit higher income earners much more that others. You claim that your numbers show that the US tax system is progressive. I have provided more that enough evidence to create at least a reasonable doubt about the veracity of that claim so either support your claim with the relevant facts (e.g. a plot of net marginal tax rates against taxable income) or withdraw it.

  10. You don't seem to give deserved attention to the motive for undertaking stimulus (balanced-budget or otherwise), which is the wretched employment situation. Why?

  11. Here's a place to start on progessivity: Thomas Piketty and Emmanuel Saez, "How Progressive is the U.S. Federal Tax System? A Historical and International Perspective", Journal of Economic Perspectives—Volume 21, Number 1—Winter 2007—Pages 3–24. Among other things, note Figure 1A.

  12. It might also serve to note that the numbers that you quote depend on both the actual progressivity of the tax system (i.e. the slope of the marginal tax rate function) and the distribution of income. The growth in income inequality in the US in the past 30 years or so would have increased the fraction of taxes paid by the earners at the top of the distribution even under a flat-tax system.

  13. Anonymous:

    I did not make any definitive claim about tax progressivity in general. I simply reported some facts that suggest a degree of progressivity in the federal tax system (hard to dispute this) and invited others to comment. In contrast, you have made an explicit claim that the U.S. tax system (as opposed to the federal tax system) is only "mildly" progressive. I would like to believe what you say. But you have not provided a single shred of evidence in support of your claim. I am just asking that you point me to the study, or set of studies, that supports your claim. Not sure why you appear to have so much difficulty in doing so.

    PS. I am not, of course, making any claims about how the progressivity of the U.S. federal tax system has evolved over time. On this, refer to Piketty and Saez (JEP 2007).

  14. Max Sawicky:

    The short answer is that the NPV calculations I suggest implicitly do take into account labor market conditions. This is what I meant when I said that one could probably make a good NPV case for many public works during a recession.

    But these public works should, in my view, be legitimate (positive NPV) projects. I worked in construction for many years and I can attest to the pride that goes along with building something useful. I don't think that most workers derive pleasure in knowing that they are building a white elephant on the backs of their fellow citizens.

  15. How about Figure 5 in

  16. What you appear to be evading is the stabilization issue. Obviously we would like projects to have positive NPVs. But suppose a project has zero NPV. What about the employment case for doing it during a recession?

  17. Anonymous:

    I see that there's hope for you yet. The CBO report you cite is interesting. The report seems to focus on labor income only, which is a limitation. Nevertheless, it is interesting to see how little progressivity there is for single filers w/o children. And the "hump" in tax rates for households with children in the income range of $30K seems perverse.

    With respect to these examples, the CBO writes: "Stylized examples of taxpayers can illustrate interactions among provisions of the tax code at different levels of income, but they provide little information about the marginal tax rates that actual households face or how
    many households fall into each income range."

    And so they perform a simulation, with results reported in Figure 6. Any thoughts?

  18. Max Sawiky:

    I'm sorry--I didn't mean to evade any issue. You did not mention stabilization in your earlier post. To answer your question, if the NPV is zero, then the project is a wash. The green light is admissible, without any employment justification.

  19. David,

    You're quite right to comment on the perversity of the marginal tax rates in the $30,000 - $40,000 annual income range. This is due primarily to tax credit phase-outs, and it causes people in the income bracket below $30,000 to not want to improve their gross taxable income. Why? Because unless you're in school or other training and make a big jump past through the 30-40K range, you'll do worse on net if you're starting below 30K and move into that range. That's bananas.

  20. Including capital income would, I immagine, only reduce progressivity given that capital income accrues mostly to high income earners and is taxed a rates (much) lower than labor income.

    Figure 6 shows some mild progessivity but nothing like that that you sought to suggest with the numbers that you quoted. Hence my objection.

    There would be even less progressivity once full effect of deductions like that for mortgage interest are taken into account as they favor higher income earners.

  21. This comment has been removed by the author.

  22. Anonymous: Thanks for the constructive part of your comments. I learned something.

    Prof J: Holy cow, I'm not sure I want to open this can of worms here! The answer to your question is going to depend on (among other things) the social welfare function one has in mind. Personally, I think that the negative income tax proposal seems like a promising alternative. Simple, progressive, and not too distortionary. I'm not an expert in this area though.

  23. David,

    I removed it. Don't want to cause (more) problems for you.

  24. Prof J: You raised a number of interesting questions that deserve a serious discussion. It was probably getting a little too off topic, this is true. But if you're ever in the mood for a guest post on the subject, let me know!

  25. I recall Robert Eisner presenting a paper at Statistics Canada a bunch of years ago. He had estimated econometric models showing that virtually every good macroeconomic outcome was a function of deficit public spending. That was right about the time that the Jean Chrétien Liberals co-opted the western Reform party agenda of fiscal discipline and helped set in motion the Great Canadian slump of the 1990s.

    It was humorous.

    As for using NPV criteria..... that sounds like a generous interpretation in a country famous for paying dearly for quick fixes. Are you saying 'impatience' is no longer a virtue?

  26. While I share your approval of the negative income tax idea (it would remove the 30k bump mentioned earlier), I agree that "this can of worms" is probably best opened elsewhere but Prof J may this interesting:

  27. David,

    Thanks. Probably I will write something up that's a little better thought out and send you an e-mail to post at your leisure.

  28. Anon,

    Thanks. That does look interesting. I hope to put a different perspective out there, but I do appreciate this.

  29. Looking at charts of tax rates are not valid comparisons. While it's true if you look at the entirety of the U.S. tax system the system is only mildly progressive - however, we can't ignore all the transfers. Let's take someone who is making less than $15,000 per year with a few kids. While they might be paying close to $2,000 in federal, state, and local taxes they are also receiving thousands of dollars from the EITC, SNAP, NSLP, LIHEAP, HUD Vouchers etc. Once all those transfers set in they are easily pulling in more than $15,000 in transfers. And I'm not even including the very valuable Medicaid / SCHIP programs even though they aren't transfers. So, in the real world, I view this persons effective tax burden as negative whereas the CBO would say their effective tax burden is 13-15%. By the way, I really hate these stupid voucher programs I'd much prefer we just do wage subsidies or even just a guaranteed minimum income rather than all this voucher / tax credit crap.

    I've never seen a decent analysis of the entire U.S. tax code that includes the burdens once all the transfers and what not set's in. I don't think anyone has ever figured out overall marginal tax rates. The best analysis I've seen is the Tax Foundation's analysis from 2004:

    Scroll down to page 24, Figure 3 to see estimates of effective tax rates. As you can see once you include state and local taxes the system does become less progressive and just at the federal level. However, we distribute a lot of money to that bottom end. Go to page 66, Figure 14 and look how government spending is distributed across income quartiles.

    To me, this looks like a fairly progressive tax system.

  30. A few quick points: 1) Shiller was obviously talking about a second-best solution insofar as debt-financed capital expenditures are ruled out by the current political climate. And it's unreasonable to expect Shiller to give a disquisition on public finance principles in this context.

    2) Taxes aren't always distortionary. Wouldn't a tax on cabon improve efficiency? and

    3) Asking for an NPV for projects begs the question: what's the opportunity cost to be used in calculating the NPV? For Keynesians, it's largely involuntary leisure; for RBC types, the cost is higher because resources are always fully employed via continuously clearly markets.

    Greg Hill

  31. Ted: I don't even know what half those acronyms stand for! But thanks for the link to that paper. I reproduce the abstract here for those who might find it of interest:

    While the U.S. tax system is progressive, the distribution of government spending makes
    the overall fiscal system more progressive than is apparent from tax distributions alone.
    Using a microdata model we estimate the distribution of federal, state and local taxes and spending between 1991 and 2004. We find households in the lowest quintile of income
    received roughly $8.21 in federal, state and local government spending for every dollar of
    taxes paid in 2004, while households in the middle quintile received $1.30, and
    households in the top quintile received $0.41. Overall, tax payments exceeded
    government spending received for the top two quintiles of income, resulting in a net
    fiscal transfer of between $1.031 trillion and $1.527 trillion between quintiles. Both taxes
    and spending appear to have large distributional effects on households, and these effects
    have grown since 1991. The results suggest tax distributions alone are an inadequate
    measure of progressivity, and policymakers should examine both tax and spending
    distributions when judging the overall fairness of policy toward income groups.

  32. Greg Hill:

    [1] More like 4th or 5th best, instead of 2nd best. Maybe. I just found it appallingly bad.

    [2] You are confusing distortionary taxation with an efficient tax. If there is a negative externality, then one would rely on the *distortionary* properties of a tax to correct the externality. A non-distortionary tax would not be efficient, in this context.

    [3] Fine. But then do a proper NPV calculation, using your best estimate of the cost of labor. Don't try to sell me a program on the basis of some dinky Econ 101 model written on the back of a dirty napkin!

  33. David:

    Many conservative economists (e.g., Robert Barro and you?) see C-BA as the right approach to public investment decision making. But in doing a C-BA, the same issues that divide Keynesians and others quickly arise.

    In doing your c-BA, do you assume that a public investment displaces an equivalent amount of private investment? That might be a good assumption if there were no slack in the economy and the project were debt financed. If there's no slack and the project is entirely financed with taxes, then it's reasonable to suppose that an equivalent amount of consumption would be displaced.

    But what if there's slack in the economy and the credit market isn't congested, then, you tell me, what's the opportunity cost of a public investment? And, in estimating benefits, should we count income externalities, which, depending on one's view, are a natural product of additional spending in a slack economy?

    I gather that you think all taxes which affect behavior are "distortionary." Fine, but in an economy where prices don't equal marginal social costs, "distortionary" taxes may improve resource allocation.

  34. Greg:

    Whatever we're arguing here, it seems clear that Shiller's contribution has no bearing. His analysis asks none of the good questions you raise.

    First, I don't know what use all this "slack" language is as a practical matter. When I was an unemployed drywaller in the early 1980s, I went to school. I'm not sure that the government should have decided to build a bridge on my behalf. There's not much of a need for drywallers or glaziers for bridge or road construction. People do not appreciate how much labor is specialized. But I digress.

    Here is how I like to think about it. I see no reason why, in principle, most public investments cannot be planned and financed at the local/state level. Why can these local agencies not formulate their plans the way any profit-seeking business might? If a bridge needs to be built, build the goddamn thing. Who cares what sort of hypothetical "externality" this might impart on some distant part of the country? Did Microsoft worry about the labor it would displace in other sectors of a booming economy? No. Should it have (from a social perspective)? You tell me. But I think most people are happy with the products, jobs, and wealth that Gates created (most of which he is giving back, it seems).

    On your last point, I think we are saying the same thing (go back to my point [2] -- you merely repeat it).

    Finally, please don't lump me in with Barro!
    Thanks! :)

  35. David,

    Here's a link to my paper, "Project Appraisal for the Keynesian Investment Planner," which attempts to reconstruct C-BA on a Post-Keyesian foundation.

    Mircosoft may not care about social costs and benefits, but a government ought to.

  36. Greg,

    It looks like you've given the subject a lot of thought. Please give me some time to read through your paper carefully. Thanks for drawing my attention to this.

  37. David,

    You're very welcome. Happy New Year.


  38. I would like to comment on the favourite quotes section of this blog. In particular to Steve Martin's quote. I know it is suppose to be If it were true then please tell me why only men are buying Have a happy new year. One of my favourite quotes would be..... Only a life lived for others is a life worth while.... Albert Einstein.

  39. Greg:

    I have now read your paper. The paper delivers what it promises: a systematic way to evaluate the NPV of public investment in the context of a Keynesian-like macroeconomy. A very thoughtful piece, I think.

    This is not the place to offer an extensive critique. My main criticisms would be directed at the Keynesian model itself; which, as you say in your paper, is not what your analysis is about.

    I do wonder, however, about these alleged positive externalities and how a planner might, as a practical matter, incorporate them in a manner that systematically benefits society.

    In your earlier comment, you said that while Microsoft may not care about social costs and benefits, a government ought to. Sure. But here is a thought experiment for you. Imagine that Bill Gates was an elected official interested in maximizing some measure of social welfare, instead of the market success of Microsoft. Can you quantify how much better off society may have been?

  40. David - a quick question -- what is the elasticity of substitution between government capital (roads/infrastructure/etc) and private capital?

    If it's relatively high, and if the financial system is truly so screwed up that productivity -enhancing private-sector investments can not be financed at reasonable rates of return (whatever those are or should be), then maybe government spending on infrastructure right now is not a completely terrible idea.

    This has nothing to do with Keynsian crosses etc.


  41. How can anyone possibly believe in the "Keynesian cross?" Govt gives money to people through expenditures and then taxes it away and that creates jobs? That is beyond ludicrous.
    Also, yes, the Federal INCOME tax is quite progressive. FICA is a separate ledger (and a ponzi scheme). That's why it isn't progressive because, theoretically, people are funding their OWN retirement. Of course, in reality that's not true. Workers today are funding workers of yesterday. Isn't that what got Bernie Madoff in trouble?

  42. TRTP

    By your definition of a Ponzi Scheme, ALL insurance is a ponzi scheme. The money to pay someones claim comes from someone elses money.

  43. David:

    I’m not sure how to interpret your question about Bill Gates as a social planner, but let me give you a concrete example from Microsoft’s hometown, Seattle, which has a municipally-owned electric utility called Seattle City Light (SCL).

    Suppose SCL is evaluating a conservation program in which the utility would hire and train unemployed workers to go house-to-house installing more efficient light bulbs and low-flow shower heads. But suppose the program has a slightly negative NPV if we only count the utility’s costs and benefits, so the utility rejects the program.

    Now, take a wider perspective. Since these newly hired workers would spend some of their income on goods whose prices exceed MC, this conservation program would generate “external” benefits. The program’s social rate of return, which takes account of all its costs and benefits, will exceed its return to City Light.

    If the program's social NPV is positive, the U.S. government could undertook this investment, or subsidized it, and we’d have a potential Pareto improvement.

    I suspect there are many examples like this. Leaving aside Keynesian unemployment, which you find problematic, consider the case of monopolistic competition, where P>MC. An increase in aggregate demand shifts the demand curves facing these firms outward, which increases their profits, part of which are spent, which pushes the demand curves out further, etc.

    Granted, the manner in which projects are financed, and the policy measures taken to increase aggregate demand, matter. But, just at a micro level, the social return to many expenditures exceeds the private return, and therefore there’s a prima facie case for public spending.

  44. Morris: Not sure that the answer necessarily depends on substitution elasticities. But I agree with your general point.

    Greg Hill: I should have been clearer. What I meant was: Suppose that all corporate America's CEOs guided their investment decisions the way they (hypothetically) do at Seattle City Light (carefully internalizing all social costs and benefits, to the best of their ability). My question is: Do you have any sense of how "big" the resulting welfare gains would be for society?

    With respect to your example, yeah, I get it. Whenever there is a wedge, there is an hypothetical government program to fix it. The finance question is critical. If we don't have to worry about how to fund these things, then the P>MC problem you cite in this example could also be fixed with a direct cost subsidy (no need for the govt to directly employ workers).

    Btw, am reading your paper on Rawls now. Interesting.

  45. Bill Mitchell takes on Shiller today too!

    A lot of math. You finance guys should love it.

  46. David:

    In answer to your question, suppose the following: 1) America’s corporations produce half of U.S. GDP; 2) all markets are monopolistically competitive, so P>MC; 3) costs are constant over the relevant range of output; so that 4) although individual firms don’t find it profitable to expand their output because prices and profits would decline; if they all expanded output together, the demand curves facing all of these firms would shift outward, so that all firms would enjoy increased profits; and 5) assume demand elasticity remains constant so that the profit-maximizing prices remains unchanged.

    **The key idea here is that one firm’s added hiring increases the demand for other firms’ products.**

    Let’s say these corporations account for $8 trillion of output and that their mark-up over (constant) MC is 30%. If our new “social-minded” CEOs all decided to increase their output by, say, 4%, then total output would grow by $320 billion, total costs would increase by about $250 billion, yielding net benefits of about $70 billion.

    You could plausibly add other benefits, e.g., increased value of human capital with a shorter average duration of unemployment, additional rounds of increased output by these and other firms via Shiller’s “confidence multiplier” (to come full circle).

    All very rough and no doubt wrong, but amusing to think about.

  47. David,

    yet again I find myself in complete agreement with your post.... Booooring.

  48. Try not to become a man of success but rather to become a man of value.
    Albert Einstein

  49. David Andolfatto can be a real pinhead at times.