A bit of longer term perspective on state and local governments

That is from Matt Mitchell, the source is here.  Josh Barro has a very good post on the rising costs of employing public sector workers.


Hrm... could this be why wages mean haven't risen as fast after 1979?
Was too much of the wealth was extracted by the public sector?

This also fits Acemoglu's extractive institutions story.

It is unlikely since the timing is wrong. State and local spending grew at about twice the rate as private GDP before about 1975 then slower for about a decade and jn 1985 reverted to the long term trend,

Here is a graph of real wages, they peaked in 1973.

Police, Fire, Education, Prisons, and Health.

Baumol effect?

Isn't the graph a little misleading? By normalizing to 1950, AND not sharing the actual percentage of GDP state and local spending. The actual numbers (from 9.5% to 21%) are disturbing enough to present without the distortion of the graphing method chosen.

This is really a much better method, because aggregating private and public GDP that way involves a lot of shaky assumptions. Also, keeping them separate allows us to show the relationship between tax-spenders and tax-payers.

The graph is designed to show the relative difference in growth rates. So no, it's not misleading.

What it shows is that state and local spending as a fraction of private GDP has been nearly constant since the 1980's with some bumps due the business cycle. Since most comments here would indicate people do not see this is proof it is misleading

It is misleading; growth rates have been converging.


Is the military personel counted in? If not it is misleading.
In my country Denmark the salaries for public servants increases at the same rate as in the private sector to avoid wrecking the economy, and the pension systems are the same. Everyone has a pension plan, where 12 to 15 percent of our salary is saved.
Actually it is an agreement between the unions and the emplyers, and it has worked well for a number of years.

In the U.S., "state" is a subnational unit. State and local spending in the United States does not include military spending, which is (almost) entirely federal (national), at least since 1914.

Some state and local expenditure is not locally funded and represents federal expenditure that is executed by state and local governments (a conservative idea). Thus, I suppose that a decent chunk of this is Medicaid expenditures on the elderly and the poor. If federal grant supporting expenditures like that are included, then they would have to be removed from federal expenditure growth measures, or there would be a double counting of government expenditures.

I figure that is offset by people who always quote Federal spending and taxes when discussing government. I've found it hard to find real total spending and revenue figures. It's out there, just hard to collect. Then try comparing to European countries.

I suspect you're right about Medicaid being at least a "decent chunk". Not sure how SCHIP is counted too, and whether it could account for some of the widening for a few years after 1997. It would be interesting to see health removed from both the state/local and federal lines. As well as disagregation in other categories too. (Maybe I'll have time to play with data later.)

I just want to know how much our government costs vs Europe in the best apples-apples.

i propose a new rule: no one ever gets to write "public sector workers" again, they must always write "teachers and policemen"

How much back office staff exists to support each?

Not much.

"Teachers, policemen and DMV personnel".

let's get a few more lines on that graph for "finance gdp" and "executive compensation"--always good to evaluate all your priorities together rather than pick the one you don't like to the exclusion of everything else

Let us know when those sectors gain the ability to tax, and we'll get right on that.

Every time you withdraw money from an ATM, you are more or less paying a tax, part of which funds executive comp.

Interest on loans is more or less a tax, part of which goes toward exec comp.

So when I buy a box of corn flakes, it's more or less like paying a tax, since part of it funds executive comp.

Those are voluntary transactions. Taxes are seized by threat of force.

Executive compensation generally doesn't rely on people with guns to seize some portion of your income whether you agree to it or not.

Executive compensation most certainly does rely on people with guns seizing a portion of your income in the finance, education, healthcare and natural resource sectors of our economy as evidenced by massive government subsidy.

All transactions are voluntary transactions.

You don't pay sales tax if you don't buy shit.

You don't pay income tax if you don't work or make under a certain amount.

And I don't know where you live but guns don't factor into my paying taxes.

No, taxes are not a voluntary transaction, they are imposed involuntarily on other transactions. Your logic is the equivalent of saying people were free to criticize Stalin as long as they didn't mind being shot.

If you insist on not paying taxes, people will eventually show up at your door with guns, no matter where you live.

uffy: I said "generally" because there are exceptions -- obviously if you're a Solyndra/BeaconPower/etc exec you basically stole taxpayer money -- but with the exception of education the sectors you mention do not receive net subsidies, natural resource extraction in particular pays far more in taxes than any subsidies they receive.

If a tax is involuntary than so is an ATM fee and so is interest paid on a loan.

Than = then.

The biggest single program that has been eating up the tax revenues of state governments is Medicaid. I've heard this over and over and over again from right out of the mouths of elected officials.

The thing about Medicaid is that the program is "voluntary" for state legislatures to join, but it makes no political sense for state governments not to join since their residents' money is being confiscated anyway by the federal government via income taxes. If the money isn't captured back via entering into Medicaid, which requires that state legislatures tax their residents a second time, it would likely be captured by other state governments that do or else those federal tax dollars would be spent on something else. Part of the legal battle over Obama Care was that Congress decided to expand Medicaid, which has the effect of requiring state governments to increase their tax burdens. The legal question is whether Congress can compel state governments to jack up taxes - all, of course, in the name of the "public welfare."

There are other federal programs which require state or local match monies, including (for example) the political decision to build light rail lines by a local transit agency.

I have a description for all this: Big federal government leads to bigger state government and bigger local government.

At least local and state governments are going toward defined contribution plans than crippling annuities.

I would like to see some more information on how he derived these numbers.

I just created a file from the GDP accounts comparing real GDP and real GDP for state and local government set to 1950 = 100.

I get a very, very, very different picture than is shown here.

It shows that as of 2011 real GDP had grown from 100 to 664 while S&L was 595.

Something is wrong here and I do not think it is me.

And, you might add, that the quality and quantity of government services has changed since 1950, just as the city or county changed in the same period.

Would you want the same number of policemen, firemen as you had then; would you like the same uneducated police force that typically barely had a high school education.

What do you think of an economist who doesn't control for quality changes over time periods, or doesn't measure the real demand or need for services in different time periods?

It would be interesting to see some comparison of police misconduct from the 1950s to today and if having "educated" police officers corresponds to lower rates of police misconduct. (Are "educated" police officers better at hiding their misconduct?)

Personally, I'd rather see Sheriff Andy and his sidekick Barney than some of the "educated" police officers I've interacted with in the last 5 years.

I find it pretty laughable that police are qualitatively better than they were in 1950. Better credentialed, yes.

Check that - 4th and 5th amendment jurisprudence is way stricter than it was in 1950. So police may be better at avoiding violations of constitutional rights or, at least, presenting the veneer of doing so.

You don't appear to understand that you're looking at a ratio here. No one is arguing that the red line should be flat, but having it grow 2-3 times as fast as private GDP is a problematic trend.

Whether we have more public sector quantity/quality than 1950 is irrelevant, we have that in the private sector too. The problem is the sector that seizes its funding from the other is also growing much faster than the other; a parasite that grows faster than its host is a problem for both.

Ergo if every cop had a PHD and was getting $250K per year plus bennies there wouldn't be anymore crime, or at least what remained would be solved in minutes.

I can tell how old you are because you have never had to deal with cops who had no education.

My wife, a librarian, deals with police from time to time. There is NO comparison to policing and police officers of the late 60's. None. Today's police officer has had training in de-escalation and dealing with problems that only come with some, at least, two year course training.

Today’s police officer has had training in de-escalation


anon, You might want to look at the psycholigical research where people can always find ONE example that supports their view, in a world where, if you look at group averages, the results would be different.

If you can honestly answer this question: are policemen today better than those in the early 60's.

If you repeat this exercise using nominal GDP you get results similar to his.

This just reflects that by definition they assume no productivity by government in calculating production and assume that outputs equal inputs.

For education that is a realistic assumption since the only wide spread measure of productivity or quality in education is students per teacher.

Education accounts for over half of S&L government employment.

Why wouldn't you measure education quality by what they're learning? Compare that to cost over time, and you find (not surprisingly) that public education has become less and less productive over time.

I don't think public educations has become less productive, I think kids are just getting dumber, and it's the fault of their parents.

The Flynn says that the opposite is happening: kids are getting smarter, but our schools are getting worse even as they get more expensive.

Wrong, gab. And so mean!

Public education is spiralling downward because the best students with the best and wealthiest parents are opting for private schooling (and homeschooling) in rapidly increasing numbers.

I'm not sure there's enough more of them to have a big effect on the trend. The number going to public schools is still about 90% iirc, while home-school is around 2% and private schools are around 1% with another 5% or so in parochials. Also, that doesn't tell us why public schools are so bad that people avoid them despite their being free.

The main reason is simply that we're paying public school teachers a lot more to do the same things they always did, which means productivity is falling. A secondary reason is that the incentives in the system aren't aligned toward doing a better job teaching, but rather towards rewarding seniority -- complete with retiring to a long, happy taxpayer-subsidized retirement in your 50s. That's why teaching has such a high early-career dropout rate, incentives matter!

Greg Cochran is fully aware of the Flynn effect, and says gab is correct.
"Within every ethnic group, there has been some improvement, but nationally, that has been canceled out by increases in the fraction of students from low-scoring groups."

It's an interesting point, cultural mix is a factor. OTOH I don't think he's right that private schools are the same when adjusted for student, else there would be little incentive for parents to move them there, and the "they can kick out bad students" notion is generally only an issue for the really selective private schools as opposed to the much larger number of parochials. I remember this argument was brought up to the DC Catholic schools (they have some of the highest scores in the country, DC publics some of the lowest) and their response was "Fine, send us those kids." It never happened, of course.

The sources are given, its BEA tables 1.1.5 and 3.3.

I think you've just misread the graph, it clearly says private GDP. Now whatever you think of gov't spending, it must be seized from the private sector in order to be spent. This is a graph that shows the parasite is growing faster than the host.

Note that the improved quantity/quality of services provided is basically irrelevant here, few would argue that the red line should be flat, but having it grow 2-3 times as fast as private GDP is a problematic trend.

If you define private as total gdp less total government consumption the results are even further away from the chart presented than my first calculation.

set to 1950 = 100

in 2011 private real gdp =715

S&L govt consumption = 595.

I stick by my original point that there seems to be something wrong with his calculations.

What does he use as deflator.
Why doesn't he use the deflators calculated by BEA?

Not sure how you're getting that result, try here.


If you do a simple comparison of 1950 to 2011, total (not just private) GDP is 15094 vs 294, total s&l spending is 22.6 vs 2273, so the respective ratios are 100:1 and 50:1 -- clearly s&l spending grew much faster, not slower, than total GDP.

And of course since the component gov't component grew faster than private, the ratio to private GDP will be a bit higher yet, accounting for the 13:5 ratio in the chart. You can use whatever deflator you like I suppose, won't change much but the slope of the lines.

I'm using the data you are linking to except I get it from Haver rather than downloading it BEA.

I constructed a series based on 1950 = 100 from the BEA data for both real and nominal values.

Well, the raw numbers speak for themselves, your calculation cannot be right given the start and end numbers.

How did you construct a series on a value you calculated as the difference of two columns anyway? Is Haver really sophisticated enough to do that? I suspect maybe you did something like construct two ratio series and then subtract the ratios from each other, which is not a valid operation.

These numbers understate state and local G depradation, because they exclude the depradation caused by regulations such as occupational licensing passed by state legislatures, the nanny state regs promulatged by Herr Bloomberg and rubber stamped by the Gestapo city council, etc.
Every time I see a letter grade on a NYC restaurant I see a swaztika.

You might need some mental health counselling if that last sentence is true.

My "mental health" is fine. Don't blame me when the State takes you away to a concentration camp.

Or in this case, a Republican mayor I guess.

Thanks, I've been banging this particular drum a long time. A surprising number of people don't understand how much non-federal spending has grown.

You should just use the numbers from table 1.1.5, which has a separate line for state and local governments. In nominal terms, total GDP grew 51 times from 1950 to 2011. S&L by only 87 times. I don't know what he means by "private" GDP, but if you take government out of total GDP, that grew by 49 times. If you deflate it, you get a whole different story, government growing by only 6 times compared to 7 or more times for the other elements, 1950-2011. Comparing apples and oranges.

The lines for "state and local" from 1.1.5 only show state and local government consumption and gross investment. This is the part of state and local spending that is counted by the BEA when it measures GDP. But it doesn't reflect everything that governments spend (for example, government transfers are not counted). That's why the chart uses data from 3.3 for total state and local expenditures. The BEA doesn't offer that series in real terms, hence the need to use inflation conversion factors to adjust for inflation. Private GDP is just C + I + X (i.e., it is GDP, not counting federal, state and local government consumption and gross investment).

Thanks Matt, interesting -- when you say "gov't transfers" I assume you must mean intergovernmental (as opposed to transfer payments to individuals).

Might be worth giving a link to your OSU inflation factors, if that's online. It would also be interesting to see something breaking out where that state and local spending has been increasing over time -- Medicaid, education, etc. Don't know if BEA has all that, though.

Hmm, no I seem to have that backwards on "gov't transfers," the other meaning seems to be preferred.

It's interesting those are excluded from GDP, though I suppose that makes sense since they don't "produce" anything.

Again, an increase in gov spending relative to private spending over time is completely consistent with gov services consisting mostly of labor intensive work which is resistant to increases in productivity. How much have capital improvements actually been able to improve the efficiency of police, education, and fire? Does it take fewer fire men to put out a blaze than in 1950? Fewer beat cops to patrol a neighborhood? Fewer teachers to handle 30 kids for a day? Or, more pointedly, fewer teachers to grade 30 essays in 5 hours? I'm sure there are some productivity improvements, but research suggests the only service industry where significant advances in productivity have been made is... retail, which dominates the private sector. And the gains made there may have petered out.


Of course we're faced with the problem of confusing the causal arrow here - perhaps retail is the only service sector to innoculate against Baumol's because it is private. While there may be something to that, it's hard to see how big box stores and warehouse based internet retailing are amenable to police and fire and education.

Not to say that Baumol's explains everything, of course. But when the public sector has little in the way of capital intensive expenditures, and the private sector's share of services are the only services to significantly benefit from productivity increases, well, no duh state and local spending has gone up much faster than in the private sector.

Fewer beat cops to patrol a neighborhood?

What beat cops? You mean those anonyms riding around in computer-equipped cars?

Nope. If it were just a problem of labor per capita, then we would expect the trend to follow population. That would be a line well below even the blue -- we do not have 5x the population of 1950, let alone anything like 13x (it is actually about 2x). And remember, these are real dollars, COLA is baked in already.

OTOH you've just made a truly excellent argument that the problem is primarily caused by overpaying public sector workers relative to their productivity. Congratulations!

"Nope. If it were just a problem of labor per capita, then we would expect the trend to follow population. That would be a line well below even the blue — we do not have 5x the population of 1950, let alone anything like 13x (it is actually about 2x). And remember, these are real dollars, COLA is baked in already."

I don't understand why you're concerned about labor per capita or population growth. The concern is with unit labor costs. The point is that *service* workers, not public sector workers, are paid more relative to their productivity with respect to input costs. But productivity is only half the story. You also need to look at value. And the value of the productivity of the service worker increases without respect to their productivity because of competition in the labor market. People want symphonies and cops and firemen even though they could exchange each violist for vastly more widgets in 2010 than they could in 1950. This means that violists (and cops and firemen) are more valuable in 2010 than they were in 1950, even though they are pretty much no more productive.

So the question is why do we continue to value these things so highly when we could easily substitute their provision with other goods which now give us more bang for our buck?

In any case it is telling that you have no problem switching to a labor theory of value just for the sake of making flippant remarks about public sector employees.

You should understand why, since you brought up the metrics -- "Fewer beat cops to patrol a neighborhood? Fewer teachers to handle 30 kids for a day? Or, more pointedly, fewer teachers to grade 30 essays in 5 hours?" -- which is why I used population as a rough guide to what expenditures would be assuming a given level of services per capita and similar compensation -- about 2x.

So, you could double what you paid them (or halve their productivity, in other words, if they aren't any more productive) and grow 2x2 = 4x, but still not grow the gross expenditure as fast as GDP grew (5x). My flippant remark was just to demonstrate the very silly implication of your theory -- they are certainly not worth thirteen times what we were paying them in real dollars in 1950, your argument would mean if they were earning $30K in 2011 dollars in 1950, we'd be paying them $390K today for the same service.

At any rate, per the below we really don't what employee compensation did anyway, most of the expense is other things.

Also, of course most s&l public spending is not actually on employees; it only accounts for about 1/3.

1. In government, the correlation between an agency's money expenditures and its public sector employment is very weak, so Josh Barro is off base in reaching conclusions about public sector employment from data about public sector expenditures. For example, many states spend lots of money on Medicare and private prisons that produce almost no public sector employment.

2. A significant share of state and local spending derives from intergovernmental revenues and from other sources that are not general tax revenues (e.g. public college tuition, all charges collected by public hospitals, filing fees for courts, sales revenues from publicly owned liquor stores and from publicly owned water delivery systems).

3. I suspect that state and local spending growth is overwhelmingly attributable to (i) increased higher education spending resulting from the democratization of higher education initiated by the federal GI Bill and carried on by states inspired by this example, (ii) increased rates of high school attendance and completion, (iii) increased prison spending driven by deinstitutionalization of the mentally ill and by the war on drugs, (iv) increased health care and nursing home expenditures because medical treatments that actually worked became available and had to be financed, and (v) increased spending on highway infrastructure. The first three categories do mostly create public sector jobs. The last two categories create very few public sector jobs.

4. The conceptual shortcut that thinks of increased public sector expenditure as worthless, while thinking of increased private sector expenditure as having great value is fundamentally flawed. The notion that spending in legal education is a drag on the economy if it happens at the public University of Colorado, but a sign of economic health if it is spent at the private University of Denver (and tuition and fee driven expenditures are still state and local government expenditures), even if the University of Colorado derives none of its operational costs for the law school from state taxation, is rubbish. Likewise, the notion that heart surgery conducted at the private sector Swedish Medical Center in Denver adds value to the economy, while heart surgery conducted at the public sector Denver Health's hospitals does not, is likewise rubbish. Lots of public sector spending is financed by the people who benefit from it in a manner that while not market based is a reasonable approximation of the spending sources that would arise if it were (e.g. the rough justice almost user fee system of using gasoline taxes to pay for roads, or taxes on health insurance to subsidize public or non-profit sector "uninsurable" health insurance plans). Services consumed by human beings have value that is independent of how they are financed.

5. State and local regulatory activity levels have almost no causal impact on state and local government expenditures. All of the big budget items in state and local budgets (except incarceration) involve provide services to members of the public, not enforcing laws in a regulatory sense. Regulation has a lot of economic effect, but is chump change in terms of governmental costs relative to other activities of state and local governments.

6. If we set combined public and private sector GDP at 100 in 1950 with public at 9 and private at 91, then current private sector GDP is about 455, and current public sector GDP is about 121. Thus, a 364 unit increase in private sector GDP was accompanied by a 121 unit increase in public sector GDP. The possibility that the former was possible only with the investment in human capital represented by the latter is a distinct possibility. It is also a distinct possiblity that the need for public sector investment to generate GDP growth increases as countries grow more affluent and the low hanging fruit that can be accomplished without public sector investments has all been exploited already.

I think your point 3 covers well most of the state government services provided to people. Public education is a service and expense shared between state and local government agencies. Utilization rates are much higher now than 1950, and the US population is more than double.

In addition there are local government expenditures which have increased significantly over the past 60 years. Much of the growth of US population has been in metro areas where there are more government services than rural areas. One area of important government expenditure is water: providing clean drinking water and treating wastewater. Our waterways are now much cleaner than they were in 1950 and the increased local government expenditures for cleaner water are paid by general taxes as well as user fees.

4. The conceptual shortcut that thinks of increased public sector expenditure as worthless, while thinking of increased private sector expenditure as having great value is fundamentally flawed.

No one actually argued that. No matter how wonderful public spending is, government has to seize what it spends from the private sector -- when the parasite grows faster than the host there is a long-term problem.

6. Those aren't "units" they're ratios, gov't doesn't "invest" it spends without expectation of return, and gov't is worse not better than the private sector at picking winners and losers.

(although if you meant "112 unit increase" then you might be using an actual unit of 1950 production, but you're confusing GDP with expenditure, transfer payments are not counted in GDP per Matt's comment above)

I suppose the local grocer doesn't "seize" my income. But there is only one grocer in town, so if I want to eat I need to spend my money there: monopolies are nearly as coercive as governments, and I have no democratic influence on the monopolies I depend on. Also, I prefer to live in a place with a functioning government, a low density of criminals, a well-educated populace, and guarantees of food, shelter, and medical care for the unemployable. I don't resent that it costs me money to provide that.

monopolies are nearly as coercive as governments

For the benefit of his uneducated eejit, please provide
- examples of monopolies that are not the result of state power; and
- the name of the place where you live with only one grocery store (does that place also have laws restricting movement?).

I live in a fairly remote mountain valley. There used to be a second grocer, but they couldn't compete. My town appears to be too small to support two supermarkets. No state power involved. 30 miles south is a town with about twice the population, which has both a Safeway and a WalMart. If you are saying I should move there, well, I don't want to. And you can do your own homework on how very concentrated many industries are: monopolies and oligopolies are common. To the extent they are supported by state power, it's often because the state is bought off. E.g. Citizens United. Where the state is not involved, it's often because competitors are killed off. E.g. just try selling cocaine on Crips turf.

30 miles south is a town with about twice the population, which has both a Safeway and a WalMart. If you are saying I should move there, well, I don’t want to.
Hmm, must be hard to go 30 miles with no automobiles or trucks or roads. Do you have horses or mules there? I don't care where you live. But don't whine when you choose to live in a less uninhabited place and there aren't all the conveniences of a more densely populated area about. (Accept the consequences of your decisions.)

To the extent they are supported by state power, it’s often because the state is bought off.
Yes, that is not free markets at work, that is called crony capitalism. The more powerful and extensive the state the more incentive for rent seeking by individuals, NGOs, "nonprofits", bureaucrats, unions, businesses, demagogues, and politicians.

Where the state is not involved, it’s often because competitors are killed off. E.g. just try selling cocaine on Crips turf.
Right. Please wake up. You get organised crime and gangs when the *state* outlaws things. Prohibition and all that.

And re: Citizens United: it's an old truism, that the solution for speech you don't like is more speech, not less speech. The left is upset about the Koch Brothers and the right is upset about George Soros. I'm guessing that all of these ranters want only speech they approve of and no other because "other people" are too stupid to figure things out (unlike themselves), "other people" are too easily manipulated by big bad evil corporations/unions/billionaires/think tanks/political parties/ad nauseum to figure things out (unlike themselves), unlike swell folks who have a purer vision, and no agenda other than "the common good".

I'm not whining, "anon", just stating the facts. I like where I live and don't mind the disadvantages, and treasure the advantages. And of course I can drive 30 miles south, but the gasoline costs more than the savings at Walmart. I was only giving an example of how monopoly does not always depend on state power. I suspect another example is the dangerously small number of very big banks, which seem to be a result of deregulation, not excess regulation. But I know less about banking than I do about my hometown.

As for Citizens United, I'm not complaining about speech, I'm complaining about corporations being treated as persons, money being treated as speech, and unlimited spending on speech being allowed without identifying the speaker. Since the biggest campaign expense is television, and the airwaves are pubic property, I would simply grant all legitimate candidates and interest groups a fixed allotment of airtime, and forbid campaign contributions, which amount to pure bribery in my thinking.

I'm well aware that it is the criminalization of drugs that leads to gangs, and would prefer that everything (including prescription medicines) be legalized and regulated on the same terms as alcohol and tobacco. Then the drugs would at least be pure, and harder to obtain by minors. And we could do without SWAT teams breaking into the homes of peaceful citizens.

And look at West Africa if you think (as I suspect, but I may be wrong) that having no State at all is a good idea.

Private GDP is not ttotal GDP less government consumption.

If ford sells a truck to the local government or to the army it is still part of private GDP even though it is recorded in government consumption.

In the GDP accounts there are two types of government activity.

One is hiring people to perform services like teachers, policemen, soldiers, etc.. The GDP
accounts assume there is no productivity in this activity so outputs are equal to inputs. A second reason they do this is that there is no market price for many of these activities -- what is the value of a fighter pilot?

The second type of activity is government purchases of intermediate goods and services.
This is where the Ford truck appears in the accounts.

So private GDP includes these government purchases of intermediate goods and services.

Private GDP also includes nonprofits like church run hospitals and schools, etc.

But if you want a measure of real business output you should go to the BLS productivity data where they publish real business output based on data from the GDP accounts provided then by the BEA.


Strange chart. If you had taken the time to check the BEA data, they show real State & Local government spending growing 500% since 1950, while federal has grown 570% and GDP 670%, all real using respective deflators.

Are you claiming BEAs deflators are wrong. Or are you manipulating? Or dont you simply know the data?

Either way the post is problematic...

This graph is misleading. Growth rates have been converging since the 1970s.


Graph GDP and SLEXPND from FRED data and chose other data transformation and pick natural log. Logs are how you show growth rates, not normalizing to a given time for long term data. For very short term data it roughly works out, but over any period where growth no longer appears linear, normalizing creates a changing scale over time.

Here is some Mathematica code.
Plot[{Exp[x]/Exp[0], (10 + Exp[x])/(Exp[0] + 10)}, {x, 0, 10}]

These two curves have the same growth rate, but the second (~GDP) has a 11 times higher starting point, so when normalized, it appears 11 times smaller that the first (state and local expenditures). A log plot without the normalization will show that the lines converge.

I demand a retraction or an apology from Mercatus or Cowen. Preferably both.

Comments for this post are closed