Thursday’s Wall Street Journal contained an opinion column by Glenn Hubbard, arguing that ”We Can’t Tax Our Way Out of the Entitlement Crisis.” Glenn, whom I worked closely with as his resident tax-policy antagonist at the Council of Economic Advisers for the first 100 days of the Bush Admininstration (my being a staff holdover from the Clinton Administration CEA), uses the column to criticize what seems to be Senator Obama’s tax-heavy approach to entitlement reform (or at least Social Security reform).
Glenn argues that with combined spending on Social Security, Medicare, and Medicaid projected to rise by nearly 10 percentage points of GDP over the next 40 years (he cites CBO numbers for this), relying on revenues alone to keep deficits from expanding would require revenues to increase by 10% of GDP. With revenues/GDP currently around 19%, a boost of 10% of GDP would take us to 29%–which is the “simple arithmetic” Glenn refers to here:
…Social Security and Medicare spending left unchecked would, after a generation, consume about 10 percentage points more of GDP than it does today.
Simple arithmetic suggests that with this much more of GDP eaten up by the two programs, all federal taxes on average would have to be raised by more than 50% to make up the shortfall…
But enough of my “fact checking.” I don’t work for Glenn anymore. Glenn goes on to point to some academic research that demonstrates the danger of Obama’s taxes-only approach (emphasis added):
Research by economists Eric Engen of the Federal Reserve Board and Jonathan Skinner of Dartmouth suggests that such a tax increase would reduce long-term GDP growth by about a full percentage point. This is no small matter: Think of it as reversing all of the gains in our long-term growth rate from the productivity boom of the past 15 years.
Coincidentally, I know Eric Engen and Jon Skinner pretty well, too. Eric and I were the only students in Jon Skinner’s Ph.D. public finance class at U. of Virginia back in the mid-1980s. And I know something about this research that Glenn cites, the first thing being that this was a paper written 12 years ago, before we even saw the evidence (data points!) from the later years of the Clinton Administration that growth in revenues as a share of the economy could coexist with–and even encourage–strong growth in the economy overall. The second thing I know about this paper (and you can see for yourself here) is that the “full percentage point” drop in GDP growth Glenn cites is derived from an Engen and Skinner conclusion about the effect of marginal tax rate changes, roughly translated into a “corresponding” range of average tax rate changes. From the National Bureau of Economic Research summary of the paper (again, my emphasis added):
Engen and Skinner take two different approaches to discerning how taxes affect growth. First, they look at cross-country experiences. …A number of [cross-country] studies…have found that cutting marginal tax rates increases economic growth. Based on these cross-country studies, Engen and Skinner conclude that cutting marginal tax rates across the board by 5 percentage points and cutting average tax rates by 2.5 percentage points would increase the growth rate of U.S. GDP by 0.3 percentage points per year.
Second, Engen and Skinner turn to others’ research on the effect that changes in tax rates have had on the capital stock, labor supply, and R and D in the United States. They use the results of these studies to estimate the likely effect of cuts in tax rates. A hypothetical 5-percentage point cut in marginal tax rates, they conclude, would cause long-run economic growth to increase by 0.2 percentage points.
Both sets of studies suggest about the same answer: a 0.2 to 0.3 percentage point increase in growth rates in response to a major change in the tax structure, defined here as a decline in the average tax burden of 2.5 percent of GDP or a 5 percentage point marginal tax cut (for example, from a 15 percent marginal tax rate to a 10 percent marginal tax rate).
You see, the way economists view the importance of tax rates on economic growth is that it’s the marginal tax rate (the tax rate on the next, or marginal, dollar of income) that matters for economic incentives–the supply-side behaviors of labor supply and saving, which are what combined produce national output. When we study the beneficial economic effects of what we think of as “tax reform,” the mental exercise is usually the following: how would economic efficiency improve if marginal tax rates could be reduced (through base broadening), holding revenues (and hence average tax rates) constant? Hence, the studies that Eric and Jon used in their 1996 paper to glean a relationship between tax rates and economic growth, were all largely based on the premise that the relationship that matters is between marginal tax rates and economic growth, not average tax rates and economic growth. In fact, many of the studies Eric and Jon cite in their paper explicitly hold constant the average tax rate, the national saving rate, or the capital stock. Why? Because the studies try to isolate the incentive (or substitution) effects of tax policy. But Eric and Jon made the mistake of providing the reader (to be helpful?) with a loose translation of what a given marginal tax rate change would imply for the average tax rate (in “real life” that is, where we tend to shun revenue neutrality), and then suggesting (unintentionally?) that their results based on the marginal tax rate, revenue-neutral changes, and economic growth, could translate directly into a connection between the average tax rate and economic growth. Nope–you can’t do that. (I thought Jon Skinner had taught both me and Eric that.)
And if you check out the paper, you’ll see that even with the connection between marginal tax rates and economic growth that Eric and Jon were really trying to emphasize, Eric and Jon weren’t exactly confident in the statistical ”robustness” of their empirical conclusions–from either the cross-country studies or their “sectoral, bottom-up” simulation approach. They were very upfront about that, but that sort of language doesn’t get put into the NBER digest summary of a paper that has such a clean and precise conclusion and potential policy implication.
So then Glenn comes along, 12 years later, with the claim that an Obama Administration is proposing to increase taxes by enough to fully cover the growth in entitlement spending over the next forty years, which implies that more than 50% increase figure, and a boost in the average tax rate from 19% of GDP to 29% of GDP. That 10% of GDP increase in the average tax rate happens to be four times the 2.5% of GDP change in the average tax rate said by Eric and Jon’s 1996 paper to imply a corresponding 0.2 to 0.3 percentage point change in the GDP growth rate. Hence, Glenn says the Obama “tax plan” would reduce GDP growth by four times 0.25 (0.2 to 0.3) percentage points, or voila –”a full percentage point.” (Whoops, I said I was done with the fact checking, didn’t I?…)
So, get to my point?… Besides the fact that Senator Obama has far from proposed to raise federal revenues as a share of GDP to 29% (he’s barely budging it from the 40-year historical average of just over 18% from what we’ve seen so far), even if anyone were to recommend such an increase in taxes, it would be an increase in the average tax rate we’d be talking about, not the marginal tax rate. And for any given average tax rate, there are many different marginal tax rates, and schedules and patterns of marginal tax rates, that could correspond to that average rate. Ideally, if deficits are deemed too large for the health of the economy, we do want to raise revenues/GDP (the average tax rate) and reduce spending/GDP, and we want to do that in a thoughtful way that weighs costs against benefits (i.e., maximizes the net benefit to our economy and our society). Doing more on the tax side makes more sense if we can find ways to raise the average tax rate without raising marginal tax rates (or without raising too many marginal tax rates or the ones that people are most sensitive to). That is, if we can combine our goals for deficit reduction with our goals for tax reform, raising the average tax rate doesn’t have to involve trading off economic growth. In fact, raising the average tax rate without raising marginal tax rates is likely to boost national saving (as public saving would rise without adverse offsetting effects on private saving) and hence boost economic growth.
But how is it possible to raise revenues as a share of GDP (the economy-wide average tax rate) without raising marginal tax rates? Start by realizing our current income tax system applies to far from a “comprehensive” tax base (i.e., far from all of GDP). If we could broaden the tax base, we could raise revenues without raising marginal rates. Then it’s certainly possible to think about reducing the deficit through better tax policy; you don’t necessarily have to turn to “bigger” tax policy in terms of marginal tax rates.
Glenn Hubbard knows a lot about tax reform, and I know he believes in it. That’s why I’m disappointed that he would forget about tax reform when it comes to how he thinks we ought to balance the budget. Glenn says (my emphasis added):
Balancing the federal budget without a tax increase is possible, but will require strong fiscal restraint. To achieve full-employment budget balance by the end of the next president’s term in office, federal nondefense spending growth needs to be restrained to [a nominal] 2% per year instead of the currently projected 4.5%. And modest defense spending increases to fund costs of needed improvements in national security are possible.
We can also secure a firm financial footing for Social Security (and Medicare) without choking off economic growth or curtailing our flexibility to pursue other spending priorities. Three actions are essential: (1) reduce entitlement spending growth through some form of means testing; (2) eliminate all nonessential spending in the rest of the budget; and (3) adopt policies that promote economic growth. This 180-degree difference from Mr. Obama’s fiscal plan forms the basis of Sen. McCain’s priorities for spending, taxes and health care.
Glenn forgets what I know he knows: there’s a lot of “entitlement spending” on the tax side of the budget, commonly referred to as “tax expenditures,” but what we at Concord view and refer to as “tax entitlements.” I could easily edit Glenn’s three essential actions to apply to the tax side of the budget:
Three actions are essential: (1) reduce tax entitlements growth through some form of means testing (and better scruntiny of tax expenditures); (2) eliminate all nonessential (and often counterproductive) tax expenditures in the rest of the budget; and (3) adopt fiscally-responsible policies that promote national saving and economic growth.
Glenn should recognize that without these three actions tacked onto his list of essential actions on the spending side, there’s no guarantee that we can “tax cut” our way to a strong economy.