David R. Henderson  

Economic Analysis of Prop 29

Brad Plumer, meet Bill McBride... Why Don't Applicants Volunteer...

Good Analysis by a Government Official

Next month, we California voters will get to vote on Proposition 29, an initiative to raise the cigarette tax by $1.00 per pack. On Econlog, we are not allowed to advocate passage or defeat of a particular piece of legislation. But, if you have followed my posts, you can probably bet correctly about how I will vote on this.

My point here, though, is strictly about the economic analysis by the legislative analyst that's printed in the Official Voter Information Guide. Here's my bottom line: It's a good analysis. I'll put my comments under two headings: (1) effect of the tax increase on tax revenues, and (2) effect of the tax increase on health care spending.

Effect on Tax Revenues

In the tax part of a cost/benefit analysis course I teach, I often give the students a problem in which there's an existing tax on cigarettes and the state government raises the tax. I use, as it turns out, an actual example from California in the late 1990s. One of the things I want my students to see is that the deadweight loss from raising the tax has two components: (1) the standard triangle loss from lower consumption of cigarettes, and (2) the reduced revenues from the tax (a rectangle) because that drop in consumption gives less revenue to the government from the previous lower level of the tax. [I could show you with a graph but I haven't yet mastered doing graphs on this site.]

So when I heard in the advertising that the tax increase would generate $735 million in additional tax revenue for the state government over its first full year, I thought, "I bet they didn't take account of the fact that the drop in consumption will reduce the revenues from the existing tax."

I was wrong. The Legislative Analyst explicitly took this into account and estimated the revenue loss at $75 million. The proposition apparently takes this into account also and has a "backfill" proposition for making sure the new tax revenues spent on their favored items are net of this backfill.

The Legislative Analyst even gave enough data to allow you to "back out" what elasticity of demand he/she used. The estimated revenue from the current 87-cent tax per pack is $905 million (for the 2010-11 fiscal year). That would imply sales of 1.04 billion packs. I don't know what growth factor he [I'll use "he" from now on] used: let's assume 0.

Then the revenue raised from the $1-per-pack tax increase, gross of backfill, is $810 million for the 2013-2014 fiscal year. That implies sales of 810 million packs. So the reduction in packs bought due to the tax (assuming zero growth of sales absent the tax) is 194 million. That's a 19-percent reduction in response to a $1-per-pack price increase. [The analyst assumes that the whole tax is passed on to the buyer. That requires that the supply curve of cigarettes to California be horizontal, a reasonable assumption given that the cigarette companies have a very close substitute for selling cigarettes in California, namely, selling cigarettes to the other 49 states plus the District of Columbia, or, by President Obama's reckoning, the other 57 states plus D.C. :-)]

We're almost there. All we need now is to estimate the $1 price increase as a percent of the old price. Let's assume $5.00. That's a 20% increase. So the % reduction in consumption roughly equals the % increase in price. So the implied elasticity of demand is -1. [I don't know whether that includes the analyst's estimate of evasion of the tax, sales on Indian reservations, etc.]

Effect on Health Care Expenditures

One of the reasons advocates often give for higher taxes on cigarettes is that it will reduce health care spending. Of course, it will reduce health care spending on certain tobacco-related illnesses, but, by prolonging life, it will increase them on other things: the older you are, the more you spend on health care.

Did the legislative analyst catch this? Yes. Here's what he said:

For example, as discussed earlier, this measure would result in a decrease in the consumption of tobacco products. The use of tobacco products has been linked to various adverse health effects by federal health authorities and numerous scientific studies. Thus, this measure would reduce state and local government health care spending on tobacco-related diseases over the long term. This measure would have other fiscal effects that offset these cost savings. For example, the state and local governments would incur future costs for the provision of health care and social services that otherwise would not have occurred as a result of individuals who avoid tobacco-related diseases living longer. Thus, the net fiscal impact of this measure on state and local government costs is unknown.

UPDATE: See my next blog post for the analysis done correctly.

Comments and Sharing

COMMENTS (14 to date)
joshua writes:
On Econlog, we are not allowed to advocate passage or defeat of a particular piece of legislation

Is that due to the .org domain, or a rule placed upon you by a relevant authority, or is there some governmental requirement here that I am unaware of?

Gabriel Rossman writes:

Seems like California is in a relatively good position to raise excise taxes since all of our major population centers are several hours away from the state line. As such we may have to factor in people quitting but people driving across the state line to buy a month's worth of cigs isn't a serious problem.

David R. Henderson writes:

A rule by a relevant authority: the feds.
@Gabriel Rossman,
I hadn’t thought of the distance thing. Good point. That is, though, why I mentioned Indian reservations. Some of them are closer to population centers, I think.

BLM4L writes:
Then the revenue raised from the $1-per-pack tax increase, gross of backfill, is $810 million for the 2013-2014 fiscal year. That implies sales of 810 million packs.

Wait, isn't this wrong? Isn't the correct way to back out the estimated sales of cigarette packs the following:

$1.87 * X = Y

Instead of

$1 * X = Y

Where X is the # of packs sold and Y is the revenue generated via the new excise tax?

Since the Analyst claims $735M of net new revenue, on top of $905M in existing revenue from the tax, for a total of $1.64B, then the equation becomes (assuming 0 growth factor):

1.87 * X = 1.64B


X = ~ 877M

so the Analyst expects a drop of only ~ 137M packs NOT 194M packs?

BLM4L writes:

P.S. Sales taxes get added onto the excise tax, meaning the true elasticity calculation (assuming $5 per pack all-in prior to new excise tax) is

~ [-137M/1,004M] change in demand (13%)

for a price increase of [ ( A + $1 + ($1 * B)) / (A)] where

A is the previous price per pack and B is the sales tax rate applied to the new price after the excise tax hike.

Assuming A is $5 and B is 10% (hey, its the Golden State) then

[ ($5 + $1+ $0.10)/$5 ]



---> 1.22 or an ~ 22% rise of price causing an ~ 13% drop in consumption meaning an elasticity of about -0.59

That's quite a difference from the -1 you were crediting the Analyst for earlier!

Foobarista writes:

This is the twin danger from Pigovian taxes: they might "work", which would reduce revenue in a direct sense, and the "sin" they address ends up costing the government more if people _don't_ do it. It may be socially useful to get rid of the thing being targeted by the tax, but as sources of predictable, ongoing revenue, they are problematic.

See the huge mess in CA with respect to "tobacco bonds" sold by the state and that are paid with tobacco taxes.

Floccina writes:

Black market cigarettes are readily available here in Florida so as a cost one should include the fact that some more young men will be drawn into the selling and transporting of black market cigarettes. This will lead to some of them being arrested and put in jail, a cost to them and the state.

Joe Cushing writes:

I agree with Floccina. If we keep raising excise taxes on cigarettes we will start organized crime. In Europe, it's big business. I can't say I fault anyone who buys from these people either. Excise taxes on cigarettes are an example of tyranny of the majority.

Eric writes:


I'm still unclear why you are unable to advocate passage or defeat of a law. Is Econlog a tax-deductible nonprofit? What is the relevant law and how is the first amendment circumvented?


david writes:


This is not Prof. Henderson's personal blog, he is blogging here on behalf of the Library of Economics and Liberty, which is part of Liberty Fund, Inc, which is organized under 501(c)(3) of the Internal Revenue Code and cannot attempt to influence legislation nor participate in campaign activity as a substantial part of its activity without losing its eligibility for tax-exempt contributions.

Whilst one blog post is probably not substantial, who knows whether it taints EconLog as a whole, and EconLog is probably substantial.

John David Galt writes:

I recall that when the big tobacco case was being fought in court, the tobacco companies wanted to introduce evidence that their products save governments more (in pensions and Social Security not paid out because the earners die younger) than they cost government in health care revenue. This evidence was disallowed for policy reasons I consider dishonest and political, so I think it's fair to assume that that argument is true.

Granted that the state of California does not pay Social Security benefits, it does pay a huge amount of pension benefits which are bound to cost more if tobacco consumption is reduced, so the analyst's findings ought to have included that fact.

John David Galt writes:

And by the way:
In the opinion of 62 members of Congress, a little political advocacy that is incidental to a tax-exempt organization's mission of public service does not violate the rules for the exemption. I'm not sure if I agree, but it's something to be aware of.

Eric writes:


Thanks. That's what I expected but I never knew until now.


Raul A. writes:

Your second point reminded me of this episode of Planet Money:


The cigarette industry isn't crazy about arguing that shorter life spans save money.

Comments for this entry have been closed
Return to top