Rising Life Expectancy

Life expectancy rose markedly during the 20th century

The average American born in 1900 could expect to live less than 50 years. Today’s US life expectancy is roughly thirty years longer, and continues to rise. Why has life expectancy risen so much over time?

The figure below describes changes in US life expectancy over the 20th century for men and women. Data come from Noymer and Garenne.1

US Historical Life Expectancy


Life expectancy rose fastest in the early part of the 20th century. This represents a continuation of rising longevity over the late-19th century.2At the outset of the century, diseases like diarrhea, tuberculosis, pneumonia or influenza caused many more premature deaths. Infants and children were far more likely to die, and much of these life expectancy gains were the result of more people making it to adulthood. Adults could expect to die at a younger age as well, but the difference was less stark. Life expectancy rose as the result of several developments, like the development of clean water systems, vaccinations, better housing stocks, antibiotics, better food yields and more stringent food regulation (e.g., making it illegal to sell spoiled food).3

Over time, the pace of life expectancy slowed. Falling infant and child mortality rates helped fuel the more rapid rise in life expectancy in earlier decades. Arguably, this was lower hanging fruit. With the passage of time, longer life spans were a matter of delaying death among the elderly. Presumably, there are natural limits to the amount of time that humans can live, though it is uncertain how close we are to those limits today.

  1. Andrew Noymer and Michel Garenne (2000) “The 1918 Influenza Epidemic’s Effects on Sex Differentials in Mortality in the United States”Population and Development Review, 26(3): 565 – 581. Figures from 1918 show an implausible, one-year drop of over ten years, and are recoded as missing. Data available for download athttp://demog.berkeley.edu/~andrew/1918/figure2.html
  2. J. David Hacker (2010) “Decennial Life Tables for the White Population of the United States, 1790–1900” Historical Methods 43(2): 45-79.
  3. David Cutler and Grant Miller (2005) “The role of public health improvements in health advances: The twentieth-century United States”Demography, 42(1): 1-22; Laura Helmuth (2013) “Why are You Not Dead Yet?” Salon, Septemberhttp://www.slate.com/articles/health_and_science/science_of_longevity/2013/09/life_expectancy_history_public_health_and_medical_advances_that_lead_to.2.html

Download the raw data and Markdown file here

Slowdown in Educational Attainment

Society professes a dedication to promoting higher educational attainment. Are we doing a good job?

We always hear how education is a top priority in today’s economy.  We are told that society needs people to be better educated, and that progress in educating people is slow but steady.  Educational attainment may be rising, but is society doing a good job of ensuring that its young are educated?

The graph below, which is built on data from and reproduces a slightly modified graphic produced by the Census Bureau1, describes how educational attainment has changed across society since 1940.



The graph suggests remarkable improvements in educational attainment over the past seventy years. In 1940, roughly three-quarters of the population dropped out before completing high school, about 5% attended some college and 4.6% completed college. By 2014, only 12% of society had less than a high school education, 27% had some college, and 32% completed college.

The figure suggests that US educational attainment has improved continuously over the past seventy or so years. The smooth transition from a less- to more-educated society seems to have continued unabated. It looks like society has been in a continuous march toward more education.

However, the appearance of a smooth transition to a more educated society is partly an artifact of the data. Overall educational attainment figures include people of different generations, who came of age during during different periods. The inclusion of older generations obfuscates the ways in which society’s young have been educated at different rates.  It will make change look slower and more incremental.

To get a sense of these changing rates at which the young are being educated, it makes sense to focus on educational attainment among young people who are of an age at which college completion is likely. To do this, Census figures look at attainment among those aged 25 to 34.

The figure below shows changing educational enrollment among Americans in this age group. The data source is the same:



This graph provides a different picture to the image of continuous improvement imparted in the first figure. The graph suggests that the pace of increasing educational attainment was much faster in the 1940s through late-1970s, but slowed afterwards.

For example, high school drop outs fell from about 63% of young adults in 1940 to 15% in 1980. From 1980 to 2014, this proportion fell to 10%, a marginal improvement. On one hand, society might be forgiven for not being able to eradicate the phenomenon of dropping out. That final 10% to 15% of drop outs might be a particularly tough group to marshal towards high school completion. On the other hand, society has not made a concerted press for universal high school completion, much in the way that it stamped out illiteracy. It is reluctant to make bigger investments in education, and it is much more reluctant to expand social assistance to those who drop out due to economic pressures. It is hard to tell whether our failure to ensure universal completion is a matter of the problem being too difficult or society not caring enough to do the needed work. In any case, the pace at which minimum educational attainment improved has slowed considerably over the past thirty to forty years.

The proportion of young adults who have only completed high school completion is roughly where it was in 1940. In 1940, it was a result of too many people not having gone far enough in attaining education. By 2014, it was a matter of more people attaining high school and moving on to at least some college.

The pace at which college attainment rose accelerated between 1940 and 1977, and then stalled until the mid-1990s. From then to today, the proportion of young adults with college attainment has grown steadily. The proportion of college graduates grew faster between 1940 to 1977 (3.8% average annual growth) than from 1994 to today (2.1% average annual). Near-college attainment also grew faster before 1977 (2.9% average annual growth rate) than after 1991 (1.4% average annual growth rate).

What can we glean from these figures? Educational attainment rose much more quickly during the mid-20th century than since the mid- to late-1970s. In part, this is probably because the easier work has been done. Conceivably, it is easier to promote high school and college completion when fewer people do it. As rates rise, schools and education policy-makers have to find ways to educate more obstinate cases.

However, I’m not so sure that we should give the past several decades a total pass. Society has shown that it can stamp out obstinate problems if it is sufficiently motivated. We seem to lack that kind of motivation. It is harder to convince society to invest more in education. More people oppose extending economic aid to poor people to help them complete college. We are much less concerned with making college affordable, and more concerned with ensuring that college students don’t get “free rides.”  We might say that we are committed to educating Americans, but this self-image may be at odds with our revealed true preferences.

  1. US Census Bureau (2015) “Table A-1. Years of School Completed by People 25 Years and Over, by Age and Sex: Selected Years 1940 to 2014” Data table downloaded June 2015 from http://www.census.gov/hhes/socdemo/education/data/cps/historical/

Download the raw data and Markdown file

Does Crime Prevail in Big Liberal Cities?

Do liberal policies promote crime? Crime rates appear to be highest in America’s South, a conservative stronghold.

Rising crime is a theme that often emerges in this election.  Of course, this argument runs contrary to FBI crime statistics, which show that it has been declining for decades.  This fall is depicted in the below figure, taken from Gallup:

From Lydia Saad (2011)

Still, this line is used often in this election’s political discourse.

Portraying Democratic Strongholds as Plagued by Crime

Take this CNN interview of Newt Gingrich during the Republican National Convention.  It is an excerpt from a segment form John Oliver, which is certainly worth watching in its entirety:

Note that Gingrich cites large urban areas in regions that have proportionally more non-whites and tend to lean Democrat: Baltimore, Chicago, and Washington, DC.  Presumably, the underlying message is that these problems fester in there because liberal policies allow it to fester.  One might infer, by extension, that the solution is to implement conservative policies, which are taken to be the opposite of liberal policies.  Assume liberal policies promote crime, then conclude that their opposite discourages it.

Rates Are Often Higher in Smaller Metro Areas in Republican Strongholds

The problem is that most of the country’s areas that are most severely plagued are in America’s South, which is also the country’s conservative stronghold.  Moreover, crime appears to be a bigger problem in smaller metro areas, rather than the country’s biggest ones.

The interactive map below is constructed from the FBI’s UCR database. It shows the distribution of different crimes’ rates, the cities in which they are most common, and a map that describes how the incidence varies across the country’s metro areas.

Note that much of the Northeast has generally low rates, despite the region’s very large metro areas and their penchant for liberal policies.  If conservative policies  — like aggressive policing and highly punitive sentencing — were effective means to control crime, then wouldn’t conservative strongholds have lower rates?

Of course, all of this is simplistic.  A variety of factors shape these rates.  Many of high-crime areas have high poverty rates, for example.  However, if the causes of aggressive policing and sentencing do not determine crime rates, then why address the crime problem with them?