Skip to content

Tax Rates and Development

One of the clearest theoretical predictions in all of economics is that a high marginal tax rate on the income from an investment would depress the investment rate, and thus hinder growth or development.  Yet nobody has found much evidence across countries or over time to confirm that prediction, although many have tested it (including this author a long time ago).

The lack of evidence for a theory means either bad theory or bad data. The first could certainly be a problem, and the data for comparable “effective” marginal tax rates are also notoriously bad (“effective” takes into account deductions, credits, and all the million complicated beasts lurking in every tax code).

However, a new paper by Simeon Djankov, Andrei Shleifer, and co-authors in a top economics journal (ungated here) utilizes a new standardized database on corporate tax rates by PricewaterhouseCoopers. They find that high corporate tax rates lower investment, FDI, and entrepreneurship (entry of new businesses) across countries.

Unfortunately, we are in a time of both fiscal crisis and ideological polarization, in which the stakes on this debate are high, and both sides would like to manipulate evidence. This post is addressed to the open-minded people who just want to know what the evidence is (both of you!) Will this study change any minds?


This entry was posted in Uncategorized. Bookmark the permalink. Follow any comments here with the RSS feed for this post. Both comments and trackbacks are currently closed.

12 Comments

  1. Ash wrote:

    Hi Bill,

    You have to also make the case that investment leads to growth and hence development, to match your post title. Which is another area strongly predicted by theory but difficult to show in reality.

    Also the polarised debate on marginal tax rates is mostly in the US, not really a developing country (in the economic sense at least). I think the debate is much more open when it comes to developing countries.

    Finally I question a bit this dataset when I see Pakistan way out in the high effective marginal tax rates, though its tax receipts are some of the lowest in the world compared to GDP.

    Food for thought though, thanks!

    Posted March 11, 2011 at 2:39 am | Permalink
  2. serge d'agostino wrote:

    it’s a usual question (from “an open-minded guy”) : isn’ it more significant with a moving (marginal) tax rate (raising or reduction ?) along the years ? (sorry but I can’t read the paper because I’m not allowed! Maybe the answer is inside)
    Serge d’Agostino (France)

    Posted March 11, 2011 at 4:10 am | Permalink
  3. Jacob AG wrote:

    *crosses his fingers and hopes this applies to him*

    This post is addressed to the open-minded people who just want to know what the evidence is (both of you!) Will this study change any minds?

    Yes and no; these graphs are pretty convincing when considered in light of the theory, but they also raise a lot of questions.

    1) Ash’s point about the (missing?) link between investment and growth is good. Investment is good for growth up to a certain point, but what point is that?

    2) The time period is rather brief. It’s nice that this is international and standardized, but could 2003-2005 have been an extraordinary period? Could the effect be larger or smaller over different periods? (The paper is gated, unfortunately…)

    3) Is the relationship very robust? Raising the rate from 10% to 20% lowers FDI (assuming the relationship is causal, which it probably is) only 2 or 3%, and lowers investment as a % of GDP by about 5 points. It’s significant, but can we call it “job-killing”?

    3) Is this relationship “tight”? The points look, like, all over the place. There is lonely little Kyrgyzstan, Sweden, and Ukraine with their 10% corporate tax rates but not much investment, and then there’s somehow Japan, Thailand, etc., with higher corporate tax rates but also more investment. If we were to model investment in theory, do these results suggest that we would probably need more variables than just the corporate tax rate? How important would corporate taxes be?

    4) These are only corporate tax rates. How do corporate tax rates stack up against other taxes on income from investment, etc.?

    5) Lastly, I’d love to know what the implications are for the Laffer Curve. How do the potential fiscal benefits of a higher corporate tax rate stack up against the cost of reduced investment? If we drew a Laffer Curve just for marginal corporate tax rates, what side would we be on? (Ash, when you see Pakistan way over on the right of the graph and with no tax receipts, then Pakistan might be out on the right side of the curve; i.e., maybe Pakistan has low tax receipts because their taxes are too high.)

    Posted March 11, 2011 at 5:15 am | Permalink
  4. Martin wrote:

    Will this paper change any minds? Not on the basis of those graphs.

    What are the correlation coefficients? Looks to me like the data points are all bunched up in the centre, with the correlation such as it exists driven by a few outliers at either end.

    Is there a regression to get us a bit closer towards establishing a causative relationship?

    It would be interesting to see this analysis broken down by region. In particular, sub-Saharan Africa (higher than average marginal tax rates; higher than average return on investment) seems to be firmly within the bunch, so I’m not sure it tells us anything meaningful about policy choices there.

    Posted March 11, 2011 at 5:42 am | Permalink
  5. Michael Turner wrote:

    Not enough variables here for answering any of the really important questions. And the important questions all relate to how we can take the benefits of development to those who don’t enjoy them currently. Bolivia can’t tax its dirt-poor peasants or its rich, well-armed cocaine barons. So, just to be able to fund a government at all, it needs a high marginal rate on high incomes — which, past a certain point, is just an incentive for tax dodging, so they have probably run out of traction with that policy. Ireland can’t get any traction by *lowering* its rates on corporations — it’s already maxed out as an Anglo-American “pontoon bridge” into the Eurozone, so they m ight as well increase rates, even if it chases off some employers. They were developed enough to begin with, and probably plateaued long ago, in terms of economic potential. (As an uncle of mine put it, after visiting Ireland to research his “root”: “Anybody there with any get-up-and-go has already gotten up and gone.”)

    Posted March 11, 2011 at 5:59 am | Permalink
  6. Cyrus wrote:

    What’s the natural experiment that created exogenous variation in rates? If there was none, why would it be convincing?

    Posted March 11, 2011 at 9:09 am | Permalink
  7. Bernard Lowther wrote:

    Bill,
    You’d better check that the independent variable is actually ‘effective corporate income tax’ as conventionally defined. That’s because Price Waterhouse has constructed something they call a ‘total tax rate’, which includes every tax and payment paid by corporations including social security-type premiums for workers, fees for municipal services, etc. This synthetic and somewhat artifical ‘tax rate’ is also used by the World Bank “Doing Business indicators, and it is not the same as corporate income tax. If this is the measure used in the article, it doesn’t correspond to theory about the effect of marginal taxes on corporate income.

    Posted March 11, 2011 at 9:33 am | Permalink
  8. Bernard Lowther wrote:

    Now that the ungated version has been posted, it’s clear that the so-called corporate tax rate used in the paper actually includes every sort of payment to government made by the firm – social security premiums, health and unemployment insurance, VAT and sales taxes, personal income tax, etc. Wandering even farther from the conventional interpretation of corporate income tax, it also includes measures of the adminstrative burden of actually making the payments. This index is thus a measure of the burdensomeness of payments to government from the point of view of the beleaguered firm, but it doesn’t correspond to the economic definition of corporate taxes and it overlooks the concepts of tax incidence and optimal ways to raise revenue across various forms of tax. It is grounded in an advocacy perspective — corporates should pay less taxes at all times and in all places.

    Posted March 11, 2011 at 12:59 pm | Permalink
  9. Elliot wrote:

    Following on Martin’s point that a regional breakdown would be insightful and Ash’s about Pakistan’s ability to actually collect taxes, I’d also like to see what the correlation looks like for low and middle income countries and based on government revenue sources and ability to collect taxes. I haven’t read the paper yet, so perhaps they’ve already done the analysis.

    But nice to see some research on the subject regardless.

    Posted March 11, 2011 at 1:01 pm | Permalink
  10. Cornelius Christian wrote:

    In an era of globalization, of course high corporate tax rates will induce more capital flight than otherwise. This is a bit unfair, as it constrains the democratic process. If the majority of the population votes on higher corporate taxes, then politicians have a tradeoff between a larger tax base and doing the will of the people.

    The Bretton Woods system was better in this sense, as it created a framework for international capital controls. In today’s world, corporations can unfortunately dictate government policy at the expense of the people.

    Posted March 11, 2011 at 7:29 pm | Permalink
  11. Ana Majnun wrote:

    Thanks to all these intelligent people for their comments. Once again Dr. Easterly demonstrates he is an ideologue and not a social scientist. Why do I say that? He couldn’t be bothered to report for his readers any critical assessment of the causality of the relationship. Instead, in the simple-minded tradition of the ideologue, he throws out a “fact”, content in the knowlege that the “fact” will stick because nobody will take the time to determine it’s veracity. Like all the “oatmeal reduces prostrate cancer” studies, with their sample of 27 men. BTW, it’s fine to be an ideologue, but just not when you use your status as an “expert” to mislead. BTW^2 The paper contains no use of the words “causal” or “causality” or “instrumental”, and the regression results look like there is not effort to instrument for tax rates. This doesn’t even pass an undergraduate small test of causality.

    Posted March 12, 2011 at 3:11 am | Permalink
  12. Roger McKinney wrote:

    “One of the clearest theoretical predictions in all of economics is that a high marginal tax rate on the income from an investment would depress the investment rate…”

    Actually, that should have the standard caveat in front: ceteris parabus. Lots of things besides investment affect growth rates and development. Investment is necessary, but not sufficient for development. Does the author consider government spending to be “investment”? The statistical analysis is valid on to the degree that it can control for all of the other factors that hinder development, such as institutions.

    Posted March 14, 2011 at 10:15 am | Permalink

4 Trackbacks

  1. By Tax Rates and Development | Global Health Hub on March 11, 2011 at 5:03 am

    [...] this article: Tax Rates and Development AKPC_IDS += [...]

  2. [...] William Easterly reports that he has found a paper which appears to support “One of the clearest theoretical predictions in all of economics” – that a high marginal tax rate on the income from an investment will reduce the investment rate, and reduce growth or development.  He says that good evidence for that prediction is hard to find. The lack of evidence for a theory means either bad theory or bad data. The first could certainly be a problem, and the data for comparable “effective” marginal tax rates are also notoriously bad (“effective” takes into account deductions, credits, and all the million complicated beasts lurking in every tax code). [...]

  3. [...] Easterly reminds me about a recently published paper on Corporate tax rates and Investment. I decided to run a few back [...]

  4. By Tax Rates and Development « Daniel Smith on March 12, 2011 at 3:42 pm

    [...] Tax Rates and Development [...]