“everybody who’s working hard ought to pay their fair share. That includes millionaires who might be paying an effective tax rate of 15 percent when folks making $50,000 or $75,000 or $100,000 a year are paying much more.”
I’m no Romney fan, but before y’all get too hot and bothered about his 15% effective tax rate, you might want to check what the typical effective tax rate really is:
I hope some of you at this point are shocked? You thought the top rate was 35%, so why does this chart appear to max out at 15%?
“What’s the effective rate I’ve been paying? It’s probably closer to the 15 percent rate than anything,”
Effective rate, folks. Effective rate.
The effective rate is the total federal tax you pay divided by your earnings — that’s about 3.3% for a typical American household, and 15% for the Romneys.
The marginal rate, on the other hand, is the tax you pay on your last dollar of earnings — currently 35% for top earners, historically as low as 7% for top earners and as high as 92%.
Nothing to see here. He’s paying a normal tax rate for people in the top quintile, and people who are making less are paying less in taxes.
Because of extreme poverty and poor health care, Native American babies are dying at a rate 44 percent higher than a decade ago, despite a decline in the overall rate of infant deaths in the U.S.
This statistic was noted in the December 9 edition of The Week, a news magazine that compiles headline news stories from a variety of media sources. I read it once and then again. Native American infant mortality rates (IMR = deaths of infants under 1 year per 1,000 or 100,000 live births) are 44% higher today than ten years ago? Really?
To find some answers, I first went to the referenced source for the quote: The Seattle Post-Intelligencer. My search on Seattlepi.com, the on-line version of the periodical, for an article published within the last three months came up short, because I mistakenly assumed The Week reported current news. More on this later.
I then went to the Centers for Disease Control Vital Statistics System website http://www.cdc.gov/nchs/nvss.htm, because I knew I would be able to find IMRs for the Native American population, as well as for the country as a whole. My next steps were based on the assumption, which also turned out to be false, that the 44% increase in infant mortality among Native Americans was derived from national data. More on this later.
The chart shows that IMRs for Native Americans fluctuate quite a bit over time, particularly with respect to the overall rates for the United States, which slightly decrease over the time period shown. Within these national data, for the period from 1995-2007, the highest IMR recorded for Native Americans was 9.95 deaths per 1,000 live births in 1996, while the lowest IMR was 8.06 in 2005. Without factoring in the variation, the percent difference between the highest and lowest rates is almost 20%. This difference was nowhere near a 44% change.
Clearly, something was not right. The discrepancy was unnerving. How on earth did they arrive at such a large increase in rates of Native American infant mortality, and what data did they use??
Later, I discovered that the above quote about infant mortality among Native Americans referred to those residing in Washington State. Thank you, David Joerg, for finding the 2009 Seattle Post-Intelligencer article! http://www.seattlepi.com/local/article/Native-American-death-rates-soar-as-most-people-1302192.php. That’s right, news from 2009, and I naïvely thought weekly news magazines published current news. The Seattle P-I article referred to Washington State’s 2006 Vital Statistics Report which is where I looked to find more info about the IMR data.
First off, authors of the report calculated average annual IMRs from combined data for years 2002-2006. Finding an average rate is one way to deal with year-to-year variation. Check out the IMRs for Native Americans and the total population in Washington shown in the chart.
Second, the authors used a fancy statistical package, JoinPoint Regression Program, developed by the National Cancer Institute to model trends in data (As mentioned in page 11 of the Washington State Vital Statistics 2004 Report) – alas, another way to address variation. Based on statistical modeling, the authors determined that IMRs for Native Americans and Alaska Natives increased by 3.7% per year from 1994 to 2006.
Going back to Seattle P-I’s article reporting that “Native American babies were dying at a rate 44 percent higher than a decade ago, while the overall rate of infant deaths had declined,” multiply 3.7% by 12 years, and there you have 44.4%. When The Week published this notable news, more than two years later, they added “in the US” further distorting the original information provided by Washington State’s 2006 Vital Statistics Report.
The bottom line is that the Seattle P-I article and The Week’s subsequent use of a portion of the article are examples of lousy reporting. More reason to be more skeptical of what is put in print! The charts shown here would indicate that the high IMRs for Native Americans provide material for a rally on health equity; understanding the data is the first step.
Having data is just the beginning. Being confident in the stories hidden within it means being able to trust the quality of the data – and that means cleaning it.
Here’s an example. We assembled several U.S. Census Bureau sources to compile the year-by-year population of the United States. Here it is:
No problem, right? There’s a blip for WWII, which is interesting and kind of heartbreaking when you think about it. All done? Wrong.
Whenever you work with data, it’s so important to play with it. Look at it a few different ways. Compare it to other data sources. Compare it to your preconceived notions. Do some math!
So we computed the annual population growth rate and got this:
The regular sawtooth patterns before 1900 are not natural. They’re a very clear sign of some manual data tampering. Turns out that, as the U.S. Census explains in the fine print, the data before 1900 are in fact linearly interpolated from decennial censuses. Good to know!
Of course, this could just as well be a cautionary tale about the importance of reading the fine print. All roads lead to Rome.
Anyway, we never would have seen the problem if we hadn’t explored the data. Not just graphing the population by year, which looks fine, but the annual growth rate. Then the problem jumped out loud and clear.
Play with your data!
(thanks to Data Collective volunteer Stephanie Hernandez for pulling this together!)
I’ve see the CEO wealth graph compared to the average american graph. But what about the top 25%. How has their income changed?
It’s a sad reflection on the state of the Internet today that even the information elite can’t instantly find the answers to such questions. Data Collective’s ultimate mission is to fix that. But for today, let’s give a simple answer to the simple question:
The real (inflation-adjusted) income of the top quintile grew about 70% during this 40-year period. The middle quintile, on the other hand, found their real income growing only 20%.
Why did this happen? Was it related to the incredible collapse in the high-end marginal tax rate? Did something similar happen in other countries? If you have theories or data suggesting some answers, please let us know.
Also, if you find this chart interesting, you are welcome to embed it into your own web page or blog. Just click on the Share menu and go for it!
Since 1991, the highest marginal income tax rate in the US has been between 30 and 40 percent. So you’d be excused for thinking it was always that way.
Before we dive in, we should explain what marginal tax rates are. As you earn income throughout the year, your tax rate goes up. The first dollars you earn are taxed at 10%, and if you’re making a lot of money, your last dollars are taxed at 35%. See a full explanation here.
So, the highest marginal tax rate has been pretty steady between 30 and 40 percent since 1991. Did it used to be much different? Oh yeah:
Isn’t it remarkable?
Look at the run-up from 1931 to 1936! (Move your mouse over the data to see the details of individual data points)
How were America’s rich and powerful persuaded that such high tax rates were in the nation’s best interest, not to mention their own?
Or were they opposed to these high tax rates, and were strong-armed into it by a legislature that didn’t care what they thought?
What effect did these high tax rates have on American prosperity?
I don’t intend these as leading questions — I’d really like to know!
Finally — do other countries have a roughly similar history of tax rates?
Health care really isn’t my issue, but I’m going to wade into it anyway. This weekend, Matt Yglesias (probably my favorite blogger to argue with) said:
I think that too often people’s pet concerns about health care costs point to things that increase the level of health care spending rather than the high growth rate of health care spending.
Think about the market for cars. Cars are pretty expensive. They’re sold at a wide variety of price points. And quality-adjusted prices for cars don’t show any noteworthy crazy trends. Now suppose the government made cars tax deductible, what would happen? Well I assume that at the margin people would start buying more expensive cars. So for a few years, car spending as a share of GDP would accelerate. But pretty soon the American automobile fleet would have turned over and the acceleration would stop. The subsidy, in other words, provides a one-off boost to automobile spending but it doesn’t do anything to change the underlying cost structure of the system.
Health care, I think, is like that. But what’s really distressing people about health care isn’t the absolute level of spending, it’s the very rapid pace at which prices are rising.
Except that, if we look at spending on an internationally comparative basis, and look at percentage changes here and elsewhere, they aren’t rising very rapidly here. From 2002 to 2008, US per-capital health-care spending grew a bit faster than five and a half percent a year. That puts us in the middle of the pack of industrialized countries; Dutch spending grew nearly 7% per year, Canadian spending more than 5.7% per year, and UK spending more than 7%; by contrast, French spending grew less than 4.5% per year and Swiss spending grew a bit more than 5%. A similar picture obtains if you look back at the previous six-year period; American health-care spending per-capita grew a bit under 6% per year in that period, a slower rate of growth than the Dutch, British, Danes or Swedes, but faster than the French or Swiss, much faster than the Germans, and slightly faster than the Canadians.
The problem is not primarily the high growth rate of our health-care spending; the problem is precisely the high level of our health-care spending. Which in turn means that a growth rate that looks reasonable when compared internationally is unsustainable in terms of the bite it takes out of the domestic economy.
We got into this mess primarily because our per-capita health-care spending growth rate didn’t slow as quickly as our peer countries. Back in 1972, American health care was already dramatically more expensive on a per-capita basis than the British system, which operates very differently. But it was only modestly more expensive than Danish, Swedish, Canadian, German or Swiss health care. And health care expenditures were rising across the board in this period. From 1972 to 1978, American health care expenditures per-capita grew by a bit over 12% per year. But German per-capita expenses went up by 14%. British per-capita expenses went up by just under 12%. French per-capita expenses went up by over 13%. Swiss expenses went up by 11.3%. And this was the era of double-digit inflation; similar increases in prices and wages in all sorts of sectors were normal.
The problem is that America maintained a very high rate of growth in per-capita health care expenses well into the 1980s, well after inflation in general was tamed, and didn’t bring our growth rates down to internationally comparable levels until the 1990s. From 1978 to 1984, America’s per-capita health-care expenses grew nearly 12%, versus a bit over 8% for the Netherlands and a bit over 9% for Germany. In the next six-year period, America’s expenses grew over 9%, versus less than 7.5% for the Netherlands and a bit over 5% for Germany. Similar comparisons obtain with Switzerland, Canada, Belgium, the UK, France. It was in the 1980s that American health care went from being modestly more expensive than other wealthy countries with mixed public-private systems, to being wildly more expensive than other wealthy countries with mixed public-private systems.
Because we’re growing off such a high cost base, even as we have dramatically reduced the rate of growth of per-capita health-care expenditures the absolute bite we’re taking out of GDP is getting out of hand. From 1978 to 1990, German heath-care expenses as a percent of GDP did not change; they were 8.4% at the start of that period and 8.3% at the end. During the same period, American health care expenses as a percentage of GDP went from 8.4% – the same as Germany – to 12.4%, a nearly 50% increase in relative share of GDP. From 1990 to 2008, German health-care expenses have increased from 8.3% to 10.7%. Looked at one way, that’s a 2.4% increase – looked at another way, that’s a 29% increase in relative share. During that same period, American health care costs went from 12.4% of GDP to 16.4% – a 4% increase. But that’s only a 32% increase in relative share – very comparable to Germany. The high cost base means that internationally comparable growth rates in health care expenses, measured either in per-capita terms or as percent of GDP, are unsustainable for the United States.
To solve our health care problem, we have to do one of three things that no other developed country is doing.
- Either we have to grow nominal GDP much more rapidly than other developed countries while holding health-care cost inflation down to levels comparable to other developed countries.
- Or we have to slow health-care cost growth to rates much lower than those achieved by our peer countries, and keep those growth rates low for an extended period, without, in the process, sacrificing growth in nominal GDP.
- Or we have to take a one-time axe to health-care costs in some fashion so that we can, from that point, grow from a more manageable base.
I think any of these is a tall order for reformers of either the right or the left. Not because the reforms are poorly designed, but because restructuring more than 15% of the economy is hard, and when that restructuring has to be led by the government it’s even harder, because there are a lot of ways to put pressure on the government not to do it. If we could switch to the Canadian system tomorrow, and thereby achieve Canadian levels of cost control, this would not solve our problems. We would not only have to switch to the Canadian system, but then use the government’s monopsody power much more aggressively than the Canadian government has to. Static international comparisons – we spend twice as much per capita or as a percent of GDP as this or that country, without getting better health-care outcomes – are probably not as relevant for figuring out where to go as elucidating the levers that would make it possible to get from here to there. Because whatever happened in the 1980s happened. We can’t go back and make it un-happen.
I’m a big fan of educating people about how bad our spending problem is.
Which is why I’m sad when people exaggerate hideously to make the point.
Here’s the Heritage Foundation:
The problem with this graph is that it’s comparing apples to oranges.
It compares Total Federal Spending, which is for the entire country and is a cost that will be borne by all households, to Median Household Income, which is per household.
The comparison you really want to make, the only one that makes sense, is Total Household Income to Total Federal Spending:
Does the Heritage Foundation think their comparison is more meaningful? Or are they just trying to be sneaky? Or were they just rushing and didn’t stop to think?
If you think you’re paying more for goods and services now than you were five years ago, you’re right. You are. In fact, prices have risen steadily for 10 straight months ending in April 2011, the longest run since October, 2008 according to the latest Consumer Price Index (CPI) report published by the government.
While many are feeling the pinch, inflation in the current decade is nowhere near as bad as it was in the late 20th century – despite inflation being a relatively new phenomenon, historically speaking.
According to the US Consumer Price Index (CPI), the 1970s were worse. Much worse. During that decade, inflation was an astonishing 87.1%. Prices have continued to rise, but nowhere near as precipitously:
1980 – 1989 50.5%
1990 – 1999 27.5%
2000 – 2011 30.6%
While the rate of inflation has slowed, it’s still having an aggregate effect on prices that perhaps makes it feel worse than it is. There are numerous inflation calculators on the web that crunch CPI data. We used this one to calculate the impact of inflation on the price of what $20 bought you in 1970:
$ 20.00 in 1970
$ 42.47 in 1980 +112.4%
$ 67.37 in 1990 +236.9%
$ 88.79 in 2000 +343.8%
$115.93 today +479.7%
So while we’ve “only” experienced an inflation rate of 11.6% over the past five years, most Americans have lived long enough to experience triple-digit inflation in their lifetime.
The more things change, the more they stay the same.
Old people complaining about how when they were a kid, they could buy a house for a nickel.
I had assumed that lifelong price inflation was just a fact of life. After all, they had runaway inflation in Roman times, didn’t they?
Sadly the US price inflation data we are most familiar with, the Consumer Price Index published by the Bureau of Labor Statistics, only goes back to 1913. MeasuringWorth has compiled an index going back to 1774. Taking a closer look at it, I saw something surprising:
On this graph, 100 means the average prices that were prevailing in 1982 to 1984. So back in 1913, prices were about 10, so that $1 hamburger in 1983 could be had for a dime in 1913.
But now look at that horizontal line for 10. How many times were we at 10? In 1913, but also in 1880, and 1862, and 1826 and 1793 and 1777!
An old person in 1850 or 1900 wouldn’t have been complaining about prices at all. As far as they were concerned, sometimes prices went up, and sometimes they went down.
Are we looking at the wrong data here? Which period is the historical anomaly, 1774 – 1913 or 1913 – 2011?
And I know I am baiting the Rand Paul / Ayn Rand / Wizard of Oz crowd here, but I have to ask — what changed in 1913?
Finally, for Rebecca Lieb, a final question — is inflation worse in the past 5 or 10 years than it was in 1970 – 2000? The next post will feature some data more directly on that point.
