Wednesday, September 14, 2011

Barro: Investment and Taxation

On the weekend, in the New York Times, Barro proposed a program of fiscal reform to deal with our current ills. Paul Krugman, still struggling to understand how a dynamic economy behaves by using an obsolete static model from 1937, can't seem to figure out what Barro is trying to say, but Tyler Cowen does a pretty good job.

Krugman puzzles:
I would have expected Barro to offer some kind of argument based on real business cycle theory or whatever he believes about macro these days.
Here's what Barro says about the determinants of investment:
What drives investment? Stable expectations of a sound economic environment, including the long-run path of tax rates, regulations and so on.
Barro is of course dressing this up for lay people reading the NYT, but what he says seems consistent with the standard neoclassical growth model. You have to add some details, of course, to the basic model - distorting taxes and regulation. However, the basic idea is that investment decisions are made based on long-run factors. Given time-to-build for new capital, and the long life of the capital after it is put in place, any firm contemplating an investment decision will be looking far into the future. Even if one thinks that sticky prices and wages matter for some of a firm's decisions - employment and utilization - stickiness has to be irrelevant over the investment-decision horizon.

Old-fashioned accelerator theories, to the extent there is any serious theory backing them up, appear to rely on the idea that output is demand-determined. Then, since capital is required to produce output, and investment is the change in the capital stock, investment depends on the change in output, i.e. the change in "demand." But for this to work requires that wages and prices be stuck for very long periods of time, which is not consistent with empirical evidence. This seems to be why New Keynesians do not get into accelerator discussions (except for the financial accelerator, but that's entirely different).

Barro goes on with this:
And employment is akin to investment in that hiring decisions take into account the long-run economic climate.

The lesson is that effective incentives for investment and employment require permanence and transparency. Measures that are transient or uncertain will be ineffective.
The first point is a useful one, and I don't think anyone has been discussing this in the context of the current recession. Employment decisions by a firm are indeed investment decisions, though of course this varies across different types of jobs. In some cases, the firm makes a substantial investment in specific human capital when hiring a worker (shared with the worker in some fashion, as determined by the labor contract), and in other cases the primary human capital input in the job is general human capital that can be used at any firm. Thus, to some extent, employment decisions are governed by the same long-run factors that determine investment. Therefore, if we can understand what is holding investment down, we can understand part of what is holding back employment. Interesting idea.

Here are the components of Barro's tax reform proposal:

1. Bowles-Simpson proposals. Barro says:
reforming Social Security and Medicare by increasing ages of eligibility and shifting to an appropriate formula for indexing benefits to inflation; phasing out “tax expenditures” like the deductions for mortgage interest, state and local taxes and employer-provided health care; and lowering the marginal income-tax rates for individuals.
These all seem fine. Politically, changes in Social Security and Medicare seem difficult to obtain, and elimination of mortgage interest deductibility impossible, except perhaps as part of a larger package - i.e. we take this away from you but give you this in return.

2. A value-added tax. This is common in Western Europe, and Canada has a federal value-added tax. Taxation is not something I work on, so I'm not familiar with the economic arguments in favor of the value-added tax. In the long run, you get the same distortion on the consumption/leisure margin as with the income tax. Maybe the tax base is larger than for the income tax, so you can lower the tax rate and reduce the distortion, but you lose the progressivity you get from the income tax.

3. Eliminate corporate taxation. Optimal dynamic taxation tells us that taxing capital income is a bad idea, though I'm sure there are plenty of qualifications in the taxation literature. In any case, there are at least some sound economic arguments for this one.

4. Other stuff. Barro wants to (i) reverse spending increases that have occurred since 2000; (ii) eliminate estate taxes. For us to evaluate (i), Barro would have to be more specific (Bush's prescription drug plan?), and there seems no particular economic rationale for (ii).

So, Barro has given us a few things to think about. What he wrote is certainly much more useful than this, from Krugman:
So the best thing we could do to spur business investment would be to get a recovery going by whatever means necessary, including fiscal stimulus.
That conclusion is based on an accelerator idea - another obsolete piece of economics.

40 comments:

  1. "But for this to work requires that wages and prices be stuck for very long periods of time"

    Or that there is something generating multiple equilibria. Or that expectations are irrational / people do not realize that expected paths are impossible/explosive/unsustainable.

    Crucially, it depends on people assigning very heavy weight to current conditions in projecting future conditions (adaptive expectations?).

    I wonder to what extent this paper might be closer to the old keynesian worldview than sticky prices or rational expectations with multiple equilibria:

    http://www.ingentaconnect.com/content/aea/jep/2010/00000024/00000004/art00004

    ReplyDelete
  2. I thought Barro was perfectly understandable. Then again, I have been reading the literature since 1937.

    ReplyDelete
  3. Do you mean that you were alive and reading Hicks in 1937?

    ReplyDelete
  4. "Therefore, if we can understand what is holding investment down, we can understand part of what is holding back employment."

    Tyler Cowen has explained this: it is the dearth of investment opportunities :"the Great Stagnation". The dismal performance of the stock market since 2000 gives added support to Cowen's thesis. There's nothing worth investing in.

    ReplyDelete
  5. That's not really an explanation. Why are things so dismal? Is there some set of government policies that caused this, or what?

    ReplyDelete
  6. New Dynamic Public Finance says (contrary to the other stuff) that capital should be taxed, albeit not all that much. That being said, NDPF is hardly the be all end all and it hasn't really gone anywhere, with all due respect to Dr. Kocherlakota, whom I admire quite a bit.

    I think the weight of the evidence still points towards the optimal level of capital taxation being 0%. The work on that comes from a broad variety of people- Chamley, Lucas, Summers, Feldstein, Kehoe and many others.

    ReplyDelete
  7. Scuttlefish,

    Yes, there is a big leap you have to take to interpret features of the allocations you determine using a NDPF approach as tax wedges. When we say something "hasn't really gone anywhere," does that just mean that not may people do it, or do you think there is some unfulfilled promise there?

    ReplyDelete
  8. Stephen,

    Let me tell you what businesspeople are taught in top MBA programs about when to invest and when not to. I know because I was at a top MBA program in the mid-90's, the University of Michigan (and I've also spent a good deal of time working for companies and as a successful entrepreneur). In addition, this is in every corporate finance book, including ones written by famous corporate finance professors like Eugene Meyer.

    You follow the NPV rule. If an investment project is positive NPV you take it. To see if it's positive NPV you unbiasedly estimate all the costs and all the revenues every year for the life of the investment project. Then, you time-discount these costs and revenues. The discount rate you use is the current market rate of return for an investment as risky as the project in question.

    Now, what kinds of things would make it more likely that an investment project would end up positive NPV and thus be undertaken:

    1) That you could expect strong demand for what the investment project produces, so the output could actually be sold, and not only at an unprofitably low real price. The stronger the expected future demand for the products produced by the project, the less risky the project is (and thus the lower the discount rate you use), and the higher the expected revenues in the future. Both of these things raise NPV and thus make it more likely that the investment will be undertaken. I don't see how you would deny this, so why do you seem to keep implying that expected future economic demand is irrelevant in firms' decisions to undertake investment projects?

    2) The lower the long term interest rates, the lower the discount rate applied to investment projects in calculating NPV, thus the more investment projects that will clear the positive NPV hurdle and be undertaken. So why do you think the Fed buying up long term bonds (QE) and forcing down long term interest rates will have no effect? They won't be able to buy up enough of them? But what if they did?

    3) Surveys of business people show that regulatory uncertainty is not much of a concern to them, and the big one is lack of demand for their products. What Barro is saying seems to be a minor factor, or just an excuse to persuade people to support his libertarian goals, like his throwing in of abolish the estate tax. When I see something like that I'm very much on the lookout for intentional misleading for the extreme libertarian cause.

    ReplyDelete
  9. Two questions:

    1. When a firm undertakes an investment expenditure in "slack conditions," does the additional expenditure by the newly hired workers generate additional income for the recipients of these expenditures?

    2. If the answer is "yes," then investment generates external benefits, and the social return to investment in these conditions will be greater than the private return. But if investment decisions are made on the basis of expected private returns, won't the level of investment be suboptimal in these "slack conditions"?

    I realize I'm begging a lot of questions in asking these two.

    ReplyDelete
  10. "Do you mean that you were alive and reading Hicks in 1937?"

    Oops. I meant that I have read the literature since 1937. I haven't been reading it that long. I'm not that old.

    ReplyDelete
  11. Greg: define slack conditions

    ReplyDelete
  12. Anonymous,

    I was hoping to avoid defining "slack conditions," but let me give you a micro example. There's a firm in Seattle that hired a bunch of unemployed teens to go door-to-door offering to install low-flow shower heads for $10 apiece. Things went well, so the firm hired a teacher who'd been laid off to recruit additional teens to expand "sales."

    I'm sure you can imagine many similar examples. In referring to "slack conditions," I'm suggesting that there will be more of these "micro examples" when the unemployment rate is 9% than when it's 5%.

    ReplyDelete
  13. Greg,

    Under any conditions, you will find "micro examples" of people changing occupations or moving down the ladder. When a Keynesian says "slack conditions" they mean that there is some kind of economic inefficiency at work - either it's that wages and prices are not adjusting, or it's that we're somehow stuck in a bad equilibrium. Keynesians think there are free lunches lying around when there is such slack, and if the government spends a dollar that GDP increases by more than a dollar - the magic of the multiplier. For various reasons I'm skeptical about these claims. Barro raises some interesting points, but as Richard points out there's reason to be skeptical about some of his claims as well. He's not telling you about the implications of some of his reforms for the income distribution. Some, like shifting the source of tax revenue from income taxation to a value added tax, would take from the poor and give to the rich.

    ReplyDelete
  14. Richard,

    Standard investment theory - what's implicit in the neoclassical growth model - is exactly maximizing NPV, though it's a little more sophisticated than that, as what firms are doing is using the discount factors that are relevant for their shareholders. I don't know what business people you have been talking to, but what we hear by way of business people who talk to the economists at the Fed, is that they are not investing because of a high level of uncertainty. The source of the uncertainty is a big question. I don't quite buy the claim that it's regulatory uncertainty under the current circumstances, but tend to think it's uncertainty about Europe and how sovereign defaults will transmit themselves here. That doesn't mean, though, that uncertainty about regulation could not be a big deal in some contexts. Government behavior can in principle matter a lot for investment.

    ReplyDelete
  15. Stephen,

    "When a Keynesian says "slack conditions" they mean that there is some kind of economic inefficiency at work - either it's that wages and prices are not adjusting, or it's that we're somehow stuck in a bad equilibrium."

    Yes, and the inefficiency arises from the lack of a market in which conditional intentions can be pooled until a coherent combination of plans emerges. There's no market in which the unemployed can signal their intention to buy X, Y, and Z if employed (a Clower problem).

    You can find versions of such a market and what their absence implies for Barro's argument on my little blog at

    http://humanpredicament.blogspot.com/

    ReplyDelete
  16. "I don't know what business people you have been talking to, but what we hear by way of business people who talk to the economists at the Fed, is that they are not investing because of a high level of uncertainty. The source of the uncertainty is a big question."

    See this survey from the National Federation of Independent Business:

    http://economix.blogs.nytimes.com/2010/09/14/whats-holding-back-small-businesses/

    ReplyDelete
  17. Also, the uncertainty you're hearing about at the Fed may easily be predominantly uncertainty about whether the demand will be there in the future for their products, whether there will be a double dip recession (although the official definition of recession should be changed to include unemployment), not uncertainty over how Obama will regulate, or the estate tax.

    ReplyDelete
  18. Mr. Serlin. Many people look at the NFIB survey and see that "Poor Sales" is the number one problem. However, if you add "Taxes" and "Government Regulation", which are two different survey choices together, which essentially are "government policy" you will see that in August they are scored at 37 compared to 25 for "Poor Sales". So it essentially depends on how you look at the factors. Taxes and Regulation are both government policy, and are uncontrollable by businesses, and right now, deleterious and hard to predict.

    ReplyDelete
  19. Businesspeople re always going to complain that their taxes are too high (but not that their roads are too good or their workers too educated or their communities too safe).

    Look at the time series; both of those government things were higher before the great recession, but sales is the highest on record, and roughly doubled since the start of the great recession.

    ReplyDelete
  20. Richard,

    Here in NY, I regularly find myself complaining about the roads being 'too good' and the community being 'too safe'. And did I mention the amazing public school system? Heck, I'm amazed that for only 54% of my income we can get such great results! I'm grateful to wise, selfless and benevolent people like you for choosing how to spend my money. Thanks, Richard!

    ReplyDelete
  21. "their roads are too good"

    'The American Society of Civil Engineers issued an infrastructure report card Wednesday giving a bleak cumulative ranking of D'

    http://articles.cnn.com/2009-01-28/us/infrastructure.report.card_1_drinking-water-infrastructure-aging?_s=PM:US

    and don't tell us it's due to a lack of spending. there is plenty of infrastructure spending (john murtha airport, hundreds of millions in subsidies for the carpark at Yankee Stadium, etc.)

    ReplyDelete
  22. "Businesspeople re always going to complain that their taxes are too high (but not that their roads are too good or their workers too educated or their communities too safe)."

    So in effect you are arguing that your survey data are useless. Oops for you, I guess.

    ReplyDelete
  23. Richard,

    You argue that long term interest rates affect investment decisions (which is true). Tell me then why you don't think tax rates matter? If an increase in interest rates from 2% to 3% decreases investment, why doesn't an increase in the tax rate on profits from the investment (from say 15% to 20%) matter?

    ReplyDelete
  24. We must all remember that Richard Serlin is, fundamentally, a crackpot who believes he is an economist. Kind of like Naomi Klein or Steve Dubner, but less wealthy.

    ReplyDelete
  25. Richard H. Serlin “Businesspeople are always going to complain that their taxes are too high (but not that their roads are too good……. Look at the time series; both of those government things were higher before the great recession, but sales is the highest on record, and roughly doubled since the start of the great recession…”

    Mr. Serlin: Remember though, this survey question is about “the single most important problem.” In the table on p. 18, the survey low for Taxes is “8”, so while taxes may always thought of as too high, they are not always way up there on the “single most important problem”. The current score of 18 (21 a year ago before the House changed is not trivial, but not at its survey high (neither is “Poor Sales”).
    The point here is not to denigrate the importance of demand, but, to allow that other factors are important too. Further, if the policy tools to respond to demand deficiency (and I am focusing right now on Keynesian-inspired government spending) appear to have been ineffective or inefficient, then policies affecting other factors may turn out to be more effective. Lowering marginal costs to business can be effective!--- especially when you consider that many businesses product intermediate goods. Why is their “aggregate demand” down? Could it be government regulatory and tax policy effect on their customers (cost or “aggregate supply” issue)? Abolish the corporate income tax, rescind Obama care,reign in the EPA and FTC may have a much broader and more effective impact than blindly building roads and bridges and remodeling schools.

    ReplyDelete
  26. "You argue that long term interest rates affect investment decisions (which is true). Tell me then why you don't think tax rates matter? If an increase in interest rates from 2% to 3% decreases investment, why doesn't an increase in the tax rate on profits from the investment (from say 15% to 20%) matter?"

    Corporate taxes are on profits not revenues. They will shrink a positive profit, or NPV, but, generally, not turn it from positive to negative, and as long as an investment project is positive NPV by any amount, then the rule established in corporate finance, and basically most corporations is you take it; any profit is better than none.

    Here's an example to illustrate. Suppose there is an investment project to build a factory. It costs $100 million, and it generates an expected $11 million per year for 20 years, with a residual value in year 20 of $30 million.

    To calculate NPV you sum the PVs of all of the revenues, and subtract the PV of the cost, $100 million today. Suppose corporate taxes go up by 20%. Then, if you do the math, you'll see every future revenue goes down by 20%, but so does the cost, because the $100 million now gives a 20% bigger tax deduction, or offset. Thus, you get:

    .8(PV of Old Revenues) - .8(PV of Old Costs) = .8(Old NPV)

    And if the old NPV was positive, then so is the new one, and so the widely accepted rule is still take it. Now there are complications, what are the marginal tax rates now and in the future, what is the depreciation schedule, etc., but this is still generally true. On the other hand, rising interest rates, and thus discount rates, only lower the PV of the future net revenues, not the initial investment cash, and so can turn a positive NPV project to a negative one.

    ReplyDelete
  27. Now you may argue that if the positive NPV shrinks, the projects will still be taken, but then there will be less profit for managers and shareholders, CEOs won't work as hard, only $80 million per year instead of 100. But remember (or learn) there is a long established income and substitution effect (google it if you don't know), so it becomes an empirical question. Let's go to an expert on the empirical evidence, Jonathan Gruber of MIT:

    "Changes in tax rates appear to have relatively modest effects on total gross income; the total amount of income actually generated through work or savings does not respond in a sizable way to taxation."

    – "Public Finance and Public Policy", 2nd edition, 2007, Page 734

    Now, NPVs are going down, so the CEOs buy less mansions and $100,000 watches, but the government doesn't burn the tax money; spending on basic scientific and medical research goes up, infrastructure goes up, college aid goes up, etc. Spending and investment doesn't disappear. But wait, you can spend your money better than the government, every penny, in all cases. Well, try googling externalities, asymmetric information, coordination problems, natural monopoly,... And if I'm a crackpot for pointing these things out then so is almost all of academic economics where this has been long established, and so is Stephen. He constantly mentions these things as reasons for government spending – including government healthcare like in Canada! Yes Mr. Abolish the ACA, Stephen actually supports going further than the ACA! And strengthening the safety net, so I'm not the only crackpot.

    But generally, I spend more time on this blog than Mark Thoma's because I want to learn the criticism and see if there's something to it. You should try the same.

    ReplyDelete
  28. And as far as the survey, look at the time trends in the graph. Taxes and regulation were bigger before the great recession; they're not unusually high, but sales doubled as a reason from the start of the great recession and is currently the highest in the time series.

    ReplyDelete
  29. @Richard: how does the role of taxes and regulation affect the corporation's decision to invest when evaluating the opportunity in a real options theory rather than DCF/NPV framework?

    ReplyDelete
  30. In these business surveys, we do not hear complaints like this:

    1. The demand for my product is down, but if it were not so costly to lower my prices, I would.

    2. There are a lot of people that want to work. I would hire them, but their wages are too damn high.

    ReplyDelete
  31. If price expectations can be self-fulfilling as I learned in undergrad macro in the 1980s, it seems reasonable that output expectations can also be self fulfilling at multiple output levels - an expectations trap. There's no auctioneer to prevent low level equilibria

    ReplyDelete
  32. Anonymous,

    "it seems reasonable that output expectations can also be self fulfilling at multiple output levels - an expectations trap. There's no auctioneer to prevent low level equilibria."

    Isn't that exactly what I said above?

    Stephen,

    Suppose Firm X would like to sell more at prevailing prices. Suppose the same is true for many other firms. Suppose there are unemployed people who would buy the extra output of these firms if they had jobs at prevailing wages (see Clower). Unfortunately, there's no market in which all these conditional intentions can be pooled and harmonized. So, resources are unemployed.

    Kenneth Arrow once made the point that, in the absence of a more complete array of futures markets, investment would tend to be suboptimal. I think he even entitled the piece "A Cautious Case for Socialism."

    I realize the putative market failure described above doesn't lead straight away to some particular, or even any, government intervention, but we've got to get a firm grasp on the problem in order to solve it

    ReplyDelete
  33. Richard,

    First, you seem to have missed the point of Barro's argument. It looks like he is arguing that, 2+ years into an economic recovery, we need to start focusing on what we can do to boost the long-run growth prospects of the U.S. economy, and he gives some advice on how to do it. As Steve points out, he appears to be using the neo-classical growth model to guide his thinking. You can argue with merits of his proposals, but if we are thinking about long-run growth, it seems hard to argue against the neo-classical model (although I am sure you can find a quote to post from one of your principles textbooks indicating otherwise.) In the context of this model, distorting taxes absolutely affect investment. You can work this out yourself, google it, or check one of your principles texts.

    Although your little present value calculations don't seem particularly relevant for discussing long-run growth, let's proceed to help you with this anyway.

    Now, you don't seem to like to think about the details of your little present value calculation, other than to presume it is driven primarily by "demand" on the one side, and the interest costs of financing the investment. Two concepts you have left out are 1)uncertainty and 2)labor costs, among others. I suppose a sophisticated firm would calculate some expected future revenue path, as well as a detailed analysis of labor costs(I don't mean the costs of the actual project, but daily operating costs for the firm upon completion). I find it hard to believe you think uncertainty regarding labor costs (taxes, regulations, health care costs, etc.) plays no role (or a very small role) in firm investment decisions. In the long-run, the daily operating costs would seem to be much greater than the interest costs of financing the investment.

    No one is asking you to give up your ideology, simply to recognize some basic trade-offs. You can certainly make the argument that higher taxes to support more government spending is worthwhile. However, you cannot ignore the implications of higher distorting taxes, especially on investment decisions.

    ReplyDelete
  34. Anonymous and Stephen,

    Regarding low-output equilibria and one way to improve matters, try this (admittedly rough):

    Imagine a two-firm economy in which the wages paid out by firm X are spent on the output of firm Y, and vice versa. If X believes Y will be hiring additional labor, then X will expand its output, and if Y believes the same about X, Y will expand, and the result will be a high-output equilibrium. If both X and Y expect the other to reduce output, then you’ve got a low-output equilibrium.

    There are, of course, more than two firms in the real-world economy, but can we modify our simple model easily enough. Let's keep our original firm X, but call it General Electric. But now let Y represent not a single firm, but the other 499 firms in the S&P 500. Since the "S&P 499" comprises a large share of economy, it's not implausible to suppose that General Electric's sales revenue will depend on the hiring decisions of the "S&P 499."

    And, we may add, that all the firms in the S&P 500 face General Electric's predicament, which is to say that their sales will also depend on the hiring decisions of the S&P 499. The trouble, of course, is that each of the S&P 500 firms is uncertain about the hiring plans of the S&P 499. So, what is to be done?

    First, we can hook up all S&P 500 firms via a computer network (alternatively a wider range of firms could be included). Second, the "auctioneer" at the center of this network asks each firm how many additional employees it would hire in the U.S. if total hiring by other S&P 500 firms in the U.S. increased by X%. This pooling of conditional intentions would continue until a consistent set of intentions is found. If firms fail to meet their hiring "commitments," they are subject to a tax, and some portion of the revenues from this tax would be transferred to the firms which kept their commitments.

    By this method, we can transform a coordination game, which has both high- and low-output equilibria, into an assurance game in which a high-output outcome is more likely than a low-output equilibrium.

    ReplyDelete
  35. I'm not hearing any sympathy for the notion that our current problem is sticky wages or prices. Greg seems focused on multiple equilibria, as is anonymous above. We certainly have coherent models of self-fulfilling low-activity steady states, and models with fluctuations not induced by anything fundamental. These models all rely on some form of increasing returns, whether in the aggregate production technology or in the matching of would-be trading partners in markets. I think on the empirical question of whether the increasing returns actually exist in practice, the jury is still out. Indeed, the data may not be good enough to tell. It's certainly an interesting story, but it's hard to tell whether this is what is actually going on.

    ReplyDelete
  36. Stephen, in the models of Angeletos and Lao, "confidence"-driven fluctuations can occur in the absence of any increasing returns.
    - C

    ReplyDelete
  37. In Angeletos and Lao the equilibrium is unique, and it's not suboptimal, so there's no role for government policy. The key thing going on there is that fluctuations are not driven by preferences, endowments, and technology. There's an information friction, and there are some examples where some latent variable that you can think of as the state of optimism drives everything.

    ReplyDelete
  38. "There's an information friction"
    There was some work in the 1990s that showed that aggregate activity was affected by revisions to economic data. Perhaps the recent slowdown is a response to the revised data that showed the recession was much deeper than first reported. The information friction / news shock literature of Angeletos and La"O, Lorenzoni, Beuadry/Portier and Laura Veldkamp seems a most promising route for explaining the business cycle fluctuations.

    ReplyDelete
  39. 1. Angeletos and Lao seems to be essentially observationally equivalent to an RBC model. There's something driving the cycle that you can't observe directly, the equilibrium allocation is unique and efficient, and the model replicates some features of the business cycles. It's interesting, but not a breakthrough.

    2. The news shock models are something else altogether. In some cases that's about what you anticipate about future TFP rather than what the current TFP shock is. There may be something to this, but it's not a great revelation, and not especially informative about our current predicament. Actually, I don't think you're correct about this being about "revisions to economic data." It's not the revisions, it's what you know about the future shocks today.

    ReplyDelete
  40. " Perhaps the recent slowdown is a response to the revised data that showed the recession was much deeper than first reported."

    As the economy slowed well before the revised data was released, this doesn't hold.

    ReplyDelete