Why the Correlation Between Top Marginal Rates and Real Economic Growth is Positive
by Mike Kimel
Discouraging Greg Mankiw From Working Would be Good for the Economy, Part 2: Why the Correlation Between Top Marginal Rates and Real Economic Growth is PositiveCross posted at the Presimetrics blog.
I had a post the other day noting that the correlation between top marginal rates and real GDP per capita is positive. Depending on how you look at it (i.e., growth over several years, growth over one year, going back to 1929, focusing only on the period since Reagan took office, etc) that correlation might be small or it might go above 50%, but it is positive. That is to say, higher top marginal income tax rates have not caused with slower real economic growth in this country. Not the message you’ll get from most economists, but the data says what the data says, and where economists disagree with the data, its a sign that something is seriously wrong with the profession, not the data. (Note – the post appeared at the Presimetrics blog and at Angry Bear, and was in response to an op ed piece by Greg Mankiw noting that higher marginal income tax rates would dissuade him from working.)
Now, in that post I didn’t explain why higher top marginal income tax rates haven’t reduced growth, and may have, at times, dare I say it, actually been a force for faster real economic growth. So I’m going to cover that here.
To start with, there is no question that if you tax someone’s efforts enough, they will reduce their efforts. Any answer that is true will have to be consistent with both that statement and the facts (i.e., the positive correlation between top marginal income tax rates and real GDP per capita growth). I can think of several such answers, and I believe all are true to some extent. Now, before I lay out these answers, there is something I should note. I’ve listed these answers in order, from more believable and less important to less believable and more important. The reason I think most people will find the most relevant explanations most believable is that they don’t quite believe that pesky fact, that the correlation between top marginal income tax rates and real GDP per capita is positive. With that warning, here goes:
1. Top marginal tax rates are simply not high enough to induce people who pay it to reduce their efforts. That’s an answer a number of bloggers (Mark Thoma is a good example) gave in response to Mankiw; raising the top marginal tax rate from 35% to 39.6% shouldn’t change Mankiw’s behavior much at all as the difference probably amounts to peanuts for Mankiw. That may be true, but it does nothing to explain why growth rates were highest during periods when marginal tax rates were in the 70% range and up.
2. Dissuading people from putting in certain efforts doesn’t prevent others from putting in the same efforts. Linda Beale is one of several bloggers to note that others might happily step in to do Mankiw’s job should he choose. I’m sure this is not what Linda Beale had in mind when she wrote it, but finding people willing to give policy advice that contradicts the known facts and results in sub-part growth shouldn’t be all that difficult, frankly.
3. Rising top marginal tax rates may dissuade some people from working, but generally won’t dissuade those doing productive work. This is related to explanation 2, but it is subtly and importantly different. Simply put, most people are easily replaceable. Charles De Gaulle famously said “The cemeteries of the world are full of indispensable men.” By that, I imagine he meant that in general, most people are easily replaceable. That holds even for folks who are deemed irreplaceable; I suspect if you replaced almost everyone working for the Fortune 100 or the Ivy League tomorrow, the change would be un-noticeable pretty quickly. (An exception appears below.) Now, there are some people that genuinely are irreplaceable, that genuinely do change things, but there aren’t many of these people, and beyond a certain point, they aren’t motivated by money. Heck, most of them don’t end up all that wealthy despite being irreplaceable, and those few unique innovators that do accumulate vast fortunes (Steve Jobs would probably be an example) would happily do what they as long as they were making enough to meet some relatively basic needs. The few people who can’t be replaced if they upped and quit because the big bad gubmint raised their taxes aren’t the sort of people to go Galt in the first place.
4. A substantial percentage (and no, I don’t know what percentage that is) of people who are motivated enough by money that they might reduce their output in the face of even small changes to the top marginal rates are engaged in activities that are not good for society. For example, they might be Harvard professors who peddle theories that are 180 degrees opposed to reality. Or perhaps they develop or implement financial instruments that help bring down the world economy. There is a bit of a self-selection bias at play; people who care enough about money to become
5. At the margin, the gov’t can be more efficient with resources than many people whose needs are sated. The primary motivation for such folks may be not risking what they accumulated and paying as little in taxes as possible. The result, in many cases, is paying vast sums to accountants and keeping the money parked or hidden rather than in productive use.
Anyhow, that’s what I came up. Your thoughts?
Well I’d start by noting that the correlation isn’t super strong evidence, since there are many omitted variables. I can think of one thing which may be marginally different from 2,3,& 4.
People aren’t motivated only by money and most people also want to help society and other individuals. To some degree we choose between that which we think is most socially useful and that which gives us high income. Hence the pejorative phrase “selling out”. For example, consider Greg Mankiw. Writing NYT op-eds is very far from the most profitable use of his time — he could make lots of money consulting instead. He does it, because he wants to influence the public debate. Now in standard super simple models, the activity which generates highest income is, in some sense, the most socially useful. The models are not realistic. Also even in such models that sense of socially useful is maximizing the sum of welfare divided by the marginal utility of consumption so the rich count for more than the poor. An MD who chooses to care for the uninsured adds more to total utils and has a lower income.
In fairly simple models, effort is wasted in “dissipative signalling” — efforts to get ahead which are not socially useful. This can happen even if everyone is totally rational. In such models taxes cause increased output.
There is a problem of morale. All businessmen say this is important. People who feel that they have been treated unfairly work much less than people who feel they have been treated fairly. Huge compensation of CEOs might be very bad for firms not because of the money but because it makes all other employees resentful. High marginal taxes discourage such huge compensation. Fortune 500 CEO compensation is not a large part of GDP (it was about $ 6 billion last I checked which was a while ago). But it could be costly. The morale story sounds uhm not like standard economics (OK) and like sociology (yuck). However, it is the only theory which fits the facts and actual businessmen believe it (except for the implication about their personal compensation).
I think the most plausible explanation of the correlation is 5 — high marginal tax rates correspond to periods of high total tax take and the revenue was put to productive use.
An obvious excluded variable is the deficit. Low top marginal tax rates are strongly correlated with budget deficits (here I am excluding WWII). Low taxes can cause low growth beceause low taxes lead to deficits and deficits crowd out capital formation. This is basically explanation 5. If people keep the money, they consume more and that slows growth. WWII is a case of huge deficits without huge consumption — oh yeah there was rationing. Deficits without rationing are bad for growth (this is very true also across countries).
I’m not sure if my explanations are really different from yours. Also I am definitely not sure that the simple correlation is strong enough evidence that it makes sense to work hard to try to explain it.
I still save some money from each check I receive even though it earns nearly nothing, maybe I should stop saving.
But, I do not see any better use of my money. Particulaly since title insurance is going through the roof, forget the lake hut.
Maybe that is why I keep working even as it does not add significantly to my net worth, or what ever moneywise I care about.
But, I am doing more than filling a bank account with the work I do.
There is one obvious answer that you brushed upon in the 5th exemple:
The law of diminishing return.
Maybe it isnt the high marginal taxes effect on the people that payed them that explain the correlation, but rather the effect on the people that benefitted the most from a functional society – those that did not pay marginal taxes.
Money has a much stronger productivity-enhancing on people with little money to start with. For exemple, it is cheaper to raise the level education for those that have a low level of education.
And, it might also have to do with this:
http://www.nytimes.com/2010/10/17/business/17view.html?_r=2
The correlation may not be strongly positive, but more importantly, it is not strongly negative which means the standard arguments against higher marginal taxes rates are false.
Which means, of course, that the standard arguments for higher marginal taxes rates are also false.
When I was a kid, America’s prosperity was taking off like a rocket. At the same time, the top marginal tax rate (i.e. taxes on earnings after the first two million) was 91%. Any argument that top marginal clipping is bad for society as a whole must account for that fact.
This is similar to Robert’s number five:
Low taxes on the rich mean that the rich have more money than they know what to do with. Because the supply of productive investment opportunities is lower than the supply of excess money, the rich “invest” their money in giant gambling schemes which are bad for the real economy.
Higher taxes wisely spent lead to greater incomes among the poor who still have realistic needs. By buying for their needs they encourage investment at the local small business level… which is good for the economy.
the idea that lower taxes for the rich was good for the economy was tried in France right up to 1789.
But I would register this much objectioin: the poor middle class in this country is quite rich enough to pay a higher tax rate at least to demonstrate political good faith with the need to raise taxes on the really rich to address the deficit created by the tax cuts that were going to pay for themselves but didn’t.
oh… there is no “crowding out” when there is no demand for (real) investments.
It really depends on what happens to the taxes. Democrats tend to raise top marginal rates & invest in workforce health and education and infrastructure to a greater extent than Republicans. Republicans like to cut the top marginal rate and spend the money on wars, useless defense projects, no bid contracts and other mal-investments.
One explanation I have not yet seen discussed is somewhat the opposite of #3 in your list. Despite the fat that he is replaceable, Mankiw is seen as special by those that pay for his services. They place a premium on getting his work over someone without the name value. That gives Mankiw bargaining power, or favorable price elascticity. Therefore, if the transaction cost goes up, Mankiw can recoup some of change in income through a price increase. While the increase may not net him exactly the same as before, it may be close enough that the opportunity cost is still favorable to do the work, thus not decrease in output quantity and at the same time the gross value (as measred by price of output) of the work is greater.
Thanks Ned
I tried to point that out in another post. But folks are here to talk, not listen.
as a matter of fact, one of the huge fallacies about taxes is that without them you’d make more money. fact is the guy paying your wage, or bidding your job has factored in the cost of taxes (the cost of YOUR taxes) so he knows how much you need to get paid. There may be some limit to this at the margin, but it ain’t that ‘dead weight loss’ that dumb folks talk about. and that’s not even counting the value of the taxes to you when the government spends them on stuff you need but are too dumb to realize. please note “you” here is generic, not you, dear reader.
Lots of good reasons cited here (and, thanks, it’s about time someone debunked this ridiculous hypothesis). At the macro level, I’d wonder if the the correlation is partly or mostly spurious–for example, policies promoting/discouraging work by those who are not in the top tax bracket changed over time, and may be correlated with the top marginal rate.
Coberly and bakho above pointed in this direction. It’s partly a question of what’s done with the money by the wealthy person versus by the government; it’s also a question of policies more broadly. In addition, there are exogenous variables to consider.
Thinking only of those who are in the top bracket, one part of the question is whether these people work more/less depending on the rate, and the data do not address this. If the correlation holds for them, it may be that maintaining a lifestyle (balancing disposable income and work hours) is an objective–it requires slightly more work when the top marginal rate increases, and slightly less work when it decreases. If the correlation is reversed for them (which I doubt), it means that their marginal work effort isn’t very relevant to measures of overall economic growth.
“title insurance is going through the roof” – as a title examiner, this is news to me. The premium is a relatively small, one-time cost paid at closing. Do you mean title insurance, or propery casualty insurance (aka homeowner’s insurance)?
government investment, public works as a percentage change with top marginal tax rates. In other words, these uber-rich are not reinvesting (this is what I suspect, unproved), they are at best private investing and possibly overseas. Whereas higher tax rates, if the government increases public works, public investment (versus wars and such), I could see that increasing overall GDP. I’m game to help analyze, graph, for on an intuitive level it sure seems tax cuts for the uber rich do nothing but run dry the American well.
Oh the classic confusion between correlation and causation (yet again)!! Maybe when the economy does well more people get jealous of the wealthy and increase their tax rate? Maybe even if the economy did well while taxing the wealthy heavily it would have done A WHOLE BUNCH BETTER if their taxes had been lower. Maybe there’s a delay factor between taxing the wealthy and economic malaise. Who knows?
But what I do know is that just looking at a simple correlation and drawing all sorts of naive conclusions from it is just a waste of time.
A few thoughts:
1) A small rise in tax rates may stimulate work for people who want to have the same take-home pay.
2) High top marginal rates and economic growth may both be caused by something else.
3) High marginal rates may be part of a counter-cyclical policy, perhaps put in place as an automatic stabilizer on the upside.
4) Macroeconomic policy based on a hydraulics metaphor, which spurs growth by injecting money at the base of the economy and draining it at the top (assuming that the natural flow of money is from bottom to top, the opposite of the trickle-down metaphor).
5)Perhaps an economy based upon the weakest-link principle is conducive to growth. While high marginal rates are not a necessary part of that, they are consistent with it, and if it was preceded by an economy based upon inequality, high rates may have arisen as part of the change-over.
coberly: “But folks are here to talk, not listen.”
😉
Yes, but the main conclusion is that there is no negative correlation as proponents of the hypothesis should predict. Therefore, claims that lowering the top marginal rate will improve economic growth, and raising it will harm this growth, are unsubstantiated and inconsistent with the data. The question of “why these results” is secondary, meant to spur thought about reasons that the hypothesis is wrong and possibly even backwards. Spurring such thought is not a waste of time.
It’s simple. Just create a model based on a single good.
Mankiw will use the money to send his kids to Harvard. I will send mine to a state college.
Either the increased productivity of sending kids to Harvard is less than that of sending them to a state school, or the increased productivity of sending rich kids to college is less than that of sending less wealthy kids to college.
Unfortunately, these two factors are inextricably confounded. 🙂
No it’s not a waste of time when our country has been subjected to 30+ years of “But tax cuts grow the economy!” prescriptions from both parties.
You can sputter and fume about correlation and causation all you like John it won’t fix the essential problem with those prescriptions. They don’t work as advertised and the data shows it.
I suspect that the lack of a significant difference between the economic performance of periods with high marginal tax rates and those of much lower rates is basically due to the following reasons:
1) Incomes/wages must keep pace with capital formation. This keeps asset values tied to incomes/wages and thereby bubbles do not occur as a result of speculation that is caused by excessive investment flows. Bubbles simply can not happen without too much capital chasing too few investment opportunities. Leverage ratio adjustments could be used in conjunction with monetary policy to regulate capital formation more acutely but, there is a conflict due to capital formation being the primary objective of US/MNC global expansion and this out-weighs, in the minds of our leaders, the importance of domestic policies.
2) During the periods with high marginal tax rates growth was accelerated more by higher wages and a higher relative value of human capital across the bottom half of the workforce. Instead of trying to refute widely accepted economic theory, I think the following, which is simple, supports what I am getting at:
A producer who is also a consumer sells an watermelon for $3. He has a labor cost of $1 that goes to our worker who is also a consumer. The consumer who buys this melon could also be a worker, or he could also be a producer.
The producer has a total input cost of $2, and of course that leaves a profit of $1. Our GDP is $3.
But if the producer pays the worker $2 and charges the consumer $4 we have a GDP of $4. The profit is unchanged but because the producer is also a consumer he is now worse off. But as he increases the number of workers his consumption costs go down by percentage.
The workers share on the other hand is now more-so determined by how much he consumes. His incentive therefore is to save by consuming less so as to save and become a producer (investor/saver). This is the formula that incentivizes upward mobility and it discourages unnecessary consumption.
It is important to recognize though that the important thing here is not so much the 25% increase in GDP as it is the increase in the potential for upward mobility. This potential mobility plays a central role in not just the gains from increased human capital, (which are of course diluted by higher prices), but by the amount of genuine and sustainable demand for loans. This is the old and simple formula based on earning, saving, investing, multiplied by the largest number of people possible. This formula though proved problematic during the Viet Nam era due to the fact that when citizens have too many opportunities for advancement they are reluctant to serve in the military. During the early period when ‘globalization’ was being promoted, it was often also argued that wages needed to be more competitive if US exports were to be competitive in global markets. However, for that argument to have been valid there should have also been income reductions across the entire span of the workforce, but of course that did not occur. What did occur was the weakening of labor unions and Reagan’s Tax tweaking that incentiveized capital formation and that shifted the growth being created by mobility etc. to a dependence on growth created by lending with more emphasis on global investment and less reliance on lending to the upwardly mobile.
So… it is very unlikely that an economy with high taxes and low wages at the same time would generate enough growth to outpace the adverse feedback loops that intensify as leverage ratios […]
Robert,
All good points that you made, and I don’t have the time to respond adequately. I would like to focus on your last sentence… the reason to do it is because (I believe) just about every econ textbook already has an explanation for something that is not happening, namely that higher tax rates discourage economic growth.
And yes, I know this is univariate. But…
1. at this blog I’ve done a heck of a lot of these univariate looks using all sorts of data (state, federal, tax burden, marginal rates, etc.) and I have yet to find the textbook result
2. in my non-blog life, I’m actually reasonably competent at building statistical software and using it to run analyses though I’m not in the academic world. Given these results contradict what everyone knows is true, I wouldn’t be posting what I do on the blogs if I wasn’t confident about the results using more rigorous means. Its been a while, but as I recall, it takes far more heroic assumptions to reach the standard conclusions than it does to reach the conclusions that I tend to post over here.
Nymchen,
Good point.
coberly,
Generally agreed.
bakho,
Blowing up things in other countries doesn’t provide much, er, bang for the buck.
Ned,
Hadn’t thought of that, but I like it.
pjr,
I’m not sure if Mankiw is driven by the need to maintain a lifestyle – to many folks at that level, money is way to keep score.
Robert Oak,
That’s what I had in mind with item 5.
John,
Yes, but lack of correlation definitely does not imply causality. And the correlations do not go the way Mankiw says they do… and particularly not when you have lags (can you say Granger causality?).
Your also forgetting that top marginal rates over 50% generally get “reduced” down as “concessions” by the wealthy.
So lets say you have 70% marginal rate. Requires more “concessions” to get that say, to 25%. Maybe more pay to the “laborers” or increasing your investment priorities inside the US itself which to some people is better than Uncle Sam having it, which may produce growth somewhere else.
Standard and Poor’s is now projecting that the U.S. will have a Debt to GDP ratio of 415% by 2050. Maybe that should sink in.
The primary issue is trade policy, not minor adjustments in taxation.
The econ blogs as well as other sources are starting to look extremely foolish on the issue of taxation and Federal spending, both current and projected. Blogging posts that take tiny little baby steps in addressing the scope of the fiscal problems are doing more harm than many realize. When the blogs lull their readership into pretending that raising the top marginal tax rate is the solution or a key component of the solution to the problem, then one shouldn’t expect that audience to begin to understand the scale of the problem as projected into the future.
Raising the top marginal tax rate 3% and changing related taxes on upper income earners provides far less than $1 trillion of Federal revenue over the next decade. That’s not close to providing enough revenue to avoid the future budget and national debt crisis.
Barry Ritholtz published a chart last year that provides an interesting comparison of the historical relationship between the top marginal tax rate and level of Federal spending. It’s worth a look. You’ll have to draw your own conclusions.
Federal Spending, Percent of GDP, & Top Marginal Tax Rate
By Barry Ritholtz
May 16th, 2009
http://www.ritholtz.com/blog/2009/05/top-marginal-tax-rate/
What I find most disturbing about the ongoing tax discussions is the sole concentration of attention on the top marginal tax rate. That demonstrates a considerable level of ignorance regarding the projected U.S. and other advanced economies’ projected Debt to GDP positions. Standard and Poor’s is now projecting that the U.S. will have a Debt to GDP ratio of 415% by 2050 under a continuation of current policies. Obviously, major changes have to be undertaken in order to regain some measure of reasonable control over federal budgets. It’s sheer nonsense to pretend that only raising the top marginal tax rate will have any considerable impact on the problem. Try focusing some meaningful attention on trade policy.
This is an excellent read:
Global Aging 2010: An Irreversible Truth
Standard and Poor’s
7 Oct 2010
http://www.standardandpoors.com/products-services/articles/en/us/?assetID=1245229586712
One day, some of the other econ blogs may wake up and grasp the impact that trade policy is having on the wage levels of U.S. citizens and lack of sustainable economic growth. I expect, though, that we will be much older…and the economic crisis will be bearing down at that point.
Ray L-Love:
“Min will probably remember that I have mentioned before that economists need to learn the subtle differences between the words: ‘comparable’ and ‘commensurable’. “
Ah! Ray the incomparable. 🙂
That’s certainly true for some, Mike, but I was thinking of a certain U of Chicago law prof example–someone perhaps not too far above the cutoff for the top marginal tax rate bracket who perceives himself as having to stretch his dollars. (For him, “lifestyle” even includes a savings rate that makes him feel comfortable–this could be defined as keeping score but isn’t quite the same.) Within that high tax bracket, inequality is high, so there are more people similar to him than to the proverbial Potter. Even Mankiw may be in this category–although he claims he would work less and sacrifice his kids’ futures if the marginal rate goes up. Okay, that’s a little over the top, sorry.
Good point MG. Of course trade policy doesn’t have the same adverse effect of all US citizens. In fact it can have a significantly positive effect of those citizens whose earnings come from the “global economy.” Some US induistries are uniquely positioned to benefit most from global financial activity. Such as the fianacial industry. It doesn’t matter where the productive capacity is located. So long as money is on the move a banker will benefit. And so long as production and consumption take place a financier (investment banker, hedge fund manager, erc.) will benefit. Of course US workers will have little to do and what there is to do will be worth little.
Obviously you are way more knowledgeable on the subject than I am – but the bit Mr. Waldmann says about deficits makes a lot of sense. At least, it’s roughly what the deficit hawks are telling us NOW. Too bad for the country that they should have been talking about it during the Bush years, especially when the tax cuts were originally legislated.
check out:http://nontrivialpursuits.org/printer_friendly_tax.htm for taxes.
MG check out: http://anamecon.blogspot.com/2010/04/effects-of-unbalanced-trade.html
Our feckless government is allowing its citizens to be screwed.
Rae
i’mnot sure that deficits lead to crowding out makes much sense in an economy where the investors are afraid to invest. seems that the “market” doesn’t work for investors? if it did, rates would go up to attract money.
Would that Standard and Poor be the people who gave sub prime loans AAA ratings?
I would not expect MG to understand that raising tax rates could reduce the deficit to manageable levels, but I have demonstrated the arithmetic and it’s pretty convincing to anyone who understands long division.
as for trade policy, I know nothing about it. perhaps MG could explain the problem beyond just making assertions or citing discredited sources.