Tuesday, December 27, 2016

Simple arithmetic

Deirdre McCloskey had a piece in the NY Times on Sunday called, "Growth, Not Forced Equality, Saves the Poor."  In it, she wrote "As a matter of arithmetic, expropriating the rich to give to the poor does not uplift the poor very much.  If we took every dime from the top 20 percent of the income distribution and gave it to the bottom 80 percent, the bottom folk would be only 25 percent better off."  That didn't sound plausible to me, so I looked up figures on the distribution of income.  There are lots of sources, but the one that came up first was from the Congressional Budget Office.

Rather than interpreting "take every dime" literally, I calculated what would happen if you gave to the bottom quintile until they caught up with the second, then gave equally to the first and second until they caught up with the third, etc., and stopped when the top quntile had been reduced to the level of the fourth.  I call that the egalitarian formula.  I also calculate an alternative in which the money is distributed equally to the four lowest quintiles, again stopping when the top quintile is at the same level as the fourth.
                       Actual             Egalitarian            Flat-rate            
Top                $188,000            $83,000                $104,000
2nd                  $83,000            $83,000                 $104,000
3d                    $59,000            $77,000                 $80,000
4th                   $42,000            $77,000                 $63,000
Bottom            $24,000            $77,000                 $45,000

The numbers refer to housholds in 2011, and the "actual" are after taxes and transfers.  The average gain in the bottom four quintiles would be a lot more than 25%--the bottom three would gain by more than 25% under either formula.  Of course, if redistribution led to slower economic growth (and extreme redistribution like this probably would) it would not help the poor in the long run, but it would take quite a while for lost growth to offset the immediate gains.  To quote John Stuart Mill again "The relaxation of industry and activity, and diminished encouragement to saving which would be their ultimate consequence, might perhaps be little felt by the class of unskilled labourers in the space of a single lifetime."

Where did McCloskey get the 25% figure?  Given the magnitude of the discrepancy, it can't be a matter of using figures from a different year, or individuals instead of households, or some other technical issue.  If you interpret "take every dime" literally, start with 20 and divide it 4 ways, that gives 5, which is indeed 25% of 20.  That is, she didn't take account of the fact that people in the top quintile of incomes have higher incomes than people in bottom four quintiles.   That's the only explanation I can think of.  That leads to the question of how someone who is not just a trained economist, but a  Distinguished Professor (emerita) of Economics, History, English, and Communication could make this kind of blunder, but I'm not going to try to answer that.

Monday, December 19, 2016

Yes you can

Paul Krugman has a blog post called "What do Trump voters want"?  The obvious answer for most of them was a Republican president, but Krugman was concerned with the working-class voters who shifted from voting Democratic in previous elections.  He concluded that "I don’t think any kind of economic analysis can explain this. It has to be about culture and, as always, race."  His reasoning was that these voters were going against their economic interests, since they benefited from government spending that Trump was likely to cut.

 I think he's giving up on economic analysis too quickly.  As John Stuart Mill said, people "never . . . save in very exceptional cases . . . direct their conduct by their real ultimate interest, in opposition to their immediate and apparent interest."  Trump could appeal to an "immediate and apparent interest"--protecting American jobs.  A professional economist would say that restricting trade wasn't in the "real ultimate interest" of Americans generally, but many people, especially people with less education, aren't convinced by or even aware of the arguments for free trade.  Also, not everyone gains--the benefits are spread widely, in the form of lower prices, while the costs are concentrated on the workers who are most exposed to foreign competition--that is, on factory workers, or people who would like to get a factory job instead of a poorly paid service job. In principle, the government could increase taxes enough to compensate the "losers" and leave everyone better off, but no one is foolish enough to imagine that will actually happen.  So it's understandable that Trump's promise of "America First" and rejection of trade deals appealed to less educated voters, especially since the Clinton campaign didn't make much effort to reply to it.  As Mill said, "protection of the home producer against foreign industry" is one of the "very natural . . . results of a feeling of class interest in a governing majority of manual laborers."

I don't have any direct evidence for this interpretation, but we can look at the pattern of state differences.  The states in which Trump did well were pretty much the same as those in which Romney did well in 2012 (see the figure), but there were some differences.

Trump's biggest gains over Romney were in North Dakota, West Virginia, Iowa, South Dakota, and Maine.  Those states are all overwhelmingly white.  You could argue that means that the changes really were about race--people are more afraid of the "other" when they hadn't had much exposure to diversity.  But there's a long tradition, which has empirical support, holding that race is more important to white voters in places that are more racially diverse.  For example, when people talked about "Reagan Democrats" in 1980 they meant white working-class voters in places like the New York, Philadelphia, and Detroit metropolitan areas (see this paper).

Although the states that shifted to Trump are all rural and overwhelmingly white, their economic situations vary widely.  West Virginia has one of the highest unemployment rates of all states, while the Dakotas are among the lowest.  The same is true for rates of disability.  One thing I think that they have in common, although I can't find any clear data, is the importance of extractive industries--lumber in Maine, and energy in the others (coal in West Virginia, the Keystone pipeline in the Dakotas, and ethanol production in Iowa).  It seems like the "immediate and apparent" interest of people in those states aligned with Trump.

Monday, December 12, 2016

History repeats itself

A number of people have remarked on parallels between Donald Trump and Senator Joe McCarthy (e. g. this post from Andrew Gelman in June).  In light of the election results, there's another one:  they both got more support among less educated people.  In November, 1954, a NORC survey asked "all things considered, would you say you think favorably of Senator McCarthy, or unfavorably."  The breakdown was:

                 Favorably   Unfavorably   DK
Not HS Grad        46%         29%         25%
HS Grad            47%         34%         19%
Some College       43%         47%         10%
College Grad       36%         58%          6%

TOTAL               45%        35%         21%

There were class differences as well; if you define class by the occupation of the main wage earner, the breakdown is:

White collar         42%       43%       15%
Blue collar          48%       30%       22% 
Farm                 43%       26%       31%

but they were mostly due to education--net of education, the only clear difference is between professionals and all other occupations.

At the time of this survey, McCarthy was near the end of his run.  A Senate resolution to censure him had been proposed in July, and a bipartisan committee "whose members were notable for their impeccable reputations and legal expertise" had been appointed to investigate.  The gave their report at the end of September, and unanimously recommended censure.  The Senate reconvened on November 8 to hear the case, and on December 2 passed a motion of censure by 67-22, with about half of the Republicans (who had a majority in the Senate) voting in favor.  Another NORC survey in January 1955 asked the same question.  The results were 39% favorably, 42% unfavorably, and 29 didn't know:  a definite shift against McCarthy, but he still had substantial popular support.  Among people who had not graduated from high school, who were the largest group at the time, favorable views outnumbered unfavorable ones.

This reminded me of a post-election commentary by Luigi Zingales.  He compared Trump to Silvio Berlusconi, and said that "Mr. Berlusconi was able to govern Italy for as long as he did mostly thanks to the incompetence of his opposition. It was so rabidly obsessed with his personality that any substantive political debate disappeared; it focused only on personal attacks, the effect of which was to increase Mr. Berlusconi’s popularity" and said that the Clinton campaign had made the same mistake.  A general statement that personal attacks increase a candidate's popularity is clearly false, but I think there's a core of truth here:  attacks for violating norms of civility, good taste, or the dignity of the office are not very effective among less educated people, who are are less likely to know or care what the norms are, or to recognize unsavory historical echoes (e. g., "America First").  Trump provided many opportunities for attacks on these grounds (as had Berlusconi and McCarthy);  as a result, the Democrats neglected angles that might have been more effective.

Zingales also said that "Hillary Clinton was so focused on explaining how bad Mr. Trump was that she too often didn’t promote her own ideas, to make the positive case for voting for her."  I don't think that she needed to do that much to promote her ideas--the principles of increasing regulation of business and government programs intended to help the middle class and the poor are already familiar.  What she didn't do was respond to his attacks on trade agreements and his claims that there was no border security.  Those were issues that previous Republican presidential candidates had not pushed, and they appealed to many voters.  Trump gave voters some novel and plausible reasons to vote for him; Clinton countered with arguments that were effective with only some of the voters--more educated and sophisticated ones.  I think that combination explains why Trump did better than previous Republican nominees among less educated voters.

[Data from the Roper Center for Public Opinion Research]


Wednesday, December 7, 2016

Declining support for democracy?

Last week, the New York Times reported on research by Roberto Foa and Yascha Mounk that saw a decline in popular support for democracy.  (The paper it talked about is not out yet, but Foa and Mounk published a related article this summer).  This surprised me, since support for democracy and opposition to authoritarianism is strongly related to education, and average levels of education are rising throughout the world.  Also, there seems to have been a general cultural drift, in which political movements of the left, right, and center all claim to be working to give more power to the people.  Of course, their actions might sometimes damage democracy, but it seems like everyone appeals to democracy as a principle.

I looked at the World Values Survey (the major data source for Foa and Mounk), which contains several general questions about forms of government:

"Various types of political systems are described below. Please think about each choice in terms of governing this country and indicate if you think that it would be a very good, fairly good, fairly bad or very bad way of governing [your nation]:

a. Having a strong leader who does not have to bother with parliament and elections
b. Having experts, not government, make decisions according to what they think is best for the country
c. Having the army rule
d. Having a democratic political system"

The questions are included in four waves of the WVS:  1994-8, 1999-2004, 2005-9, and 2010-14. The prevailing view seems to be that support for democracy may be weak and unstable at first, but that once democracy is firmly established, you don't go back, so I limited it to nations in the latest wave that have a history of stable democracy.  I calculated a summary of opinions--ratings of democracy minus the ratings of all the others.  For example, in 1994 the average rating was 3.4 (4="very good," .... 1= "very bad") for democracy, 2.1 for experts, 1.5 for a strong leader, and  1.3 for the army, for a score of 3.4-2.1-1.5-1.3=-1.5.  The actual value is not meaningful--interest is in how it compares to other nations, with higher numbers meaning more support for democracy relative to the alternatives.  The results:

The United States ranked highest in support for democracy in 1994-8, but lowest in all subsequent waves.  In the other nations, there was fluctuation with no clear trend--the United States stands out for its sustained decline.

Rule by the army or a strong leader are clearly non-democratic, but having experts make decisions is more ambiguous.  People could be thinking of something like the Federal Reserve Board in the United States, which is accountable to the public in some sense.  So I computed an alternative of rating of democracy minus rating of rule by the army or a strong leader.

The results are pretty much the same, although the contrast between the United States and everyone else is a little more striking.  So the United States is different--it's certainly had a larger decline in support for democracy, and is arguably the only one of these nations to have had a decline.

Why?  One possibility would be poor results--if democracy isn't delivering the goods, people will lose confidence in it.  This might be a factor in some of the changes, like the decline in support for democracy in Spain between 2005-9 and 2010-4, when that country was hit hard by the recession.  But it declined substantially in the United States between 1994 and 1999, when the economy was doing well, the country wasn't involved in any wars or threatened by foreign powers, and crime was dropping.  Another possibility is that people are reacting against political polarization and conflict.  Although people sometimes recall the 1990s as a period of relative good feeling, the news was dominated by a long series of investigations of Bill Clinton starting in 1994 and culminating in nearly party-line votes on his impeachment in 1999, with a little time out for a government shutdown in 1995.  That sort of thing has continued in the 21st century.  Another factor may have been increasing deference to the military making people more receptive to the idea of rule by the army.  Although my knowledge of the recent political history of most of the other nations is sketchy, I think that the United States is the only one that's had a dramatic increase in polarization.  The position of the military is certainly unique.   Overall, we seem to have a case of American exceptionalism, but not in a good sense.

Friday, December 2, 2016

Going downhill

Since the 1970s, the Gallup Poll has asked "I am going to read you a list of institutions in American society.  Please tell me how much confidence you, yourself, have in each one--a great deal, quite a lot, some, or very little?"  The list has changed, but there is at least twenty years of data for the following:  the church or organized religion, the military, the US Supreme Court, banks, the public schools, newspapers, Congress, television news, organized labor, the presidency, the police, the medical system, the criminal justice system, and big business.  Tables giving the data are available at the Gallup website.  I converted them to figures in order to get a clearer sense of changes--the y-axis gives the percent saying they have "a great deal" or "quite a lot" of confidence.

I don't think anyone will be surprised to see that confidence in all three branches of government has declined, with Congress taking the biggest loss:

Or confidence in the media:

Confidence in big business and labor was never high, but it's declined for both.  Confidence in banks had ups and downs, but was generally pretty high until the onset of the 2007-8 recession, when it suffered a fall from which it hasn't bounced back:

Confidence in medicine, the public schools,and organized religion also declined.

So far, confidence in everything has declined.  You could offer specific explanations for each one, but the fact that it's so widespread suggests that the declines reflect a general mood of dissatisfaction.  I think the decline for public schools is particularly telling, since if you go by the numbers academic performance has improved since the 1970s.

But the decline in confidence isn't universal:

Confidence in police and the criminal justice system has stayed about the same since the early 1990s, and confidence in the military has clearly increased since the 1970s.

Today, confidence in the military is substantially higher than confidence in any of the other institutions; in the 1970s, the military was behind religion and the medical system, and essentially tied with the public schools (also behind banks, but the poll asked about them only once in the 1970s).

It's not possible to say what the "right" degree of confidence in any institution is.  Still, a situation in which confidence is highest in the military and the police doesn't seem healthy.

Wednesday, November 23, 2016

Public opinion on immigration, 1965-

In June 1965, the Gallup Poll asked "Should immigration be kept at its present level, increased or decreased?" (this was when the Hart-Celler Act was under consideration by Congress).  The same question was asked once in the 1970s, once in the 1980s, and pretty frequently since 1993.  There have been some variations in the way it was introduced (e. g., sometimes it is "In your view, should...") but they don't seem to make any difference.  The figure shows average opinion, counting increased as +1, decreased as -1, and present level as zero, for people who expressed an opinion.

My initial impression was just that there were always more people in favor of a decrease than an increase, but when I graphed the data I realized that there was a definite pattern:  more support for a decrease until the mid-1990s, and then increasing support for immigration after that time.  There are some ups and downs, notably a drop in support for increased immigration after 9/11, but the trend after the mid-1990s is clear.

This figure shows the percent favoring an increase:

Before the 1990s, almost all of the change in the average involved a shift from staying the same to decreased--after that time, much of it involved a rise in the number of people saying it should be increased.

As a Thanksgiving bonus, here are some pre-1965 questions on immigration:

Roper/Fortune Survey [November, 1947]

Would you vote yes or no on a bill in Congress to let 100,000 selected European refugees come to this country in each of the next four years, in addition to the 150,000 immigrants now permitted to enter every year under our present quotas?

18% Yes
72% No
5% Depends
5% Don't know

Foreign Affairs Survey [April, 1955]

In general, do you think the United States is letting too many immigrants come into this country, or not enough?

39% Too many
37% About right (vol.)
13% Not enough
11% Don't know

Hopes And Fears [September, 1964]

Do you think the number of immigrants allowed to enter the U.S. each year should be increased somewhat, decreased somewhat, or kept at about the present level?

6% Increased
38% Decreased
46% Kept at present level
10% Don't Know

Harris Survey [May, 1965]

President (Lyndon) Johnson has proposed that the immigration laws of this country be changed to allow more people into the United States as immigrants. From what you know or have heard, do you favor or oppose letting more people come to the United States as immigrants?

24% Favor
58% Oppose
18% Not sure

Thursday, November 10, 2016

Mistakes were made, part 2

Before asking how Donald Trump won the election, it's necessary to ask how he even stayed close. There were numerous occasions when it seemed that he'd gone too far, and done or said something that would finish him off.  The answer to that question can be found in some of the classics of early public opinion research:
1.  Most people don't pay much attention to politics (Campbell, Converse, Miller, and Stokes, The American Voter).
2.  When they do, they usually confirm their prior inclinations (Berelson, Lazarsfeld, and McPhee, Voting). 

So a lot of people voted for Trump because he was a Republican and Hillary Clinton was a Democrat. Others voted for him because they felt like the Democrats had had a chance and now it was time for a change.  Although this point is not very exciting, it's important.  Many discussions of Trump voters treat them as exotic--out there in West Virginia or Ohio, in towns devastated by the loss of an industry like steel or coal.  But a lot of Trump voters can be found closer to the media centers, in nice, well-kept houses with Trump signs out front.  For example, Connecticut is a "blue state," and Clinton won easily--but Trump still got 41.7% of the vote and carried about half of the towns.  In fact, Trump got more votes in Connecticut than in West Virginia.  

But there were some differences between who supported Trump and who had supported Romney, McCain, Bush, etc.  Probably the most important one was education.  Exit polls showed that 67% of whites without a college degree, and only 49% of those without a college degree voted for Trump.  The figure shows the gap in Republican support in elections since 1980.  

There was not much change between 1980 and 2012--sometimes the Republicans did a little better among people with a college degree, sometimes a little worse (the two in red are 1992 and 1996, when Ross Perot was a candidate).  Suddenly the gap became much larger in 2016.  Compared to 2012, Republican support among whites with a college degree dropped from 56% to 49%, while support among those without a college degree rose from 61% to 67%.  That change suggests that something about Trump appealed to people without a college degree and repelled people with a college degree.  

Early public opinion research also helps to explain why Trump repelled more educated people:  they give more support to civil liberties and have more awareness of and concern with the norms of democratic government.  For example, sophisticated observers were shocked when Trump told Hillary Clinton that she would be in jail if he were president (even Charles Krauthammer said that was going too far)--in a democracy, you don't put political opponents in jail.  But for someone who didn't know much about the history or theory of democracy, it wouldn't seem like a big deal--if she was guilty of a crime, shouldn't she be in jail?  And if there was even a chance she was guilty--and lots of people were saying that she was--why not appoint a special prosecutor to find out?  

What about his positive appeal?  This is something I've written about several times over the last year or two (see the list at the end of this post):  Trump is an economic nationalist, opposed to immigration, free trade, saying he'd punish companies that move jobs overseas.  These are all popular positions, and especially popular among less educated people.  They also haven't had many advocates among major American politicians, so Clinton and the Democrats didn't have much practice in countering them.  

This raises the question of whether other people will be inspired by his success to appeal to the same sentiments, so that the educational split will remain.  I'll discuss that in a later post.    

Related posts:

Wednesday, November 9, 2016

Mistakes were made

Going into the election, Hillary Clinton had a lead of 3.3% in the polls, according to the Real Clear Politics average.  In the popular vote, she had a lead of 0.2%, as of this writing.  Was this a "failure of polling," as a New York Times story says?  The Gallup poll gives a list of its last poll before each election since 1936 and the actual election results which I summarize below as the difference between Democratic lead in the two-party vote in the poll and the election.  A positive sign means that the Democrats did better in the election than in the last poll; a negative sign that they did worse.

Year   Error
2016 -3.1
2012 +5
2008 -4
2004 -2.4
2000 +2.5
1996 -3.1
1992 -6.4
1988  5.1
1984  0.4
1980 -6.8
1976 +3
1972 +0.4
1968 +0.4
1964 -5.4
1960 -0.8
1956  3.4
1952 -7.8
1948  9.4
1944  4.6
1940  6
1936 13.6

The error has two parts--one is from sampling, and the other is from everything else--for example, supporters of the different candidates might differ in their willingness to respond, or the formulas for calculating turnout might be be wrong in ways that favored one party over the other.  The size of sampling error can be calculated from statistical theory--the other part can't, and has to be estimated from experience.  

The combined sample for the RCP average is so large that sampling error is trivial, on the order of 0.1%.  The samples for the individual Gallup polls are smaller, so sampling error is a factor.  However, it's not nearly big enough to account for the discrepancies.  A rough calculation is that the non-sampling error has a standard deviation of about 5.  Until 1948, the Gallup poll used quota samples.  That period included two big errors--the notorious 1948 election, and the 1936 election, when the poll showed a Democratic lead of +12 (56%-44%), but the actual lead was +25 (62.5%-37.5%).  If you confine it to the 1952-2012 period, the non-sampling error has a standard deviation of about 3--that is, candidate leading by three in the polls would lose about one time in six if the errors are normally distributed.  Most of the prediction sites seemed to give Clinton about an 85% chance of winning, which was just about right according to this analysis. Someone just going by the polls would have been a little surprised by the outcome, but not shocked.

Ironically, the problem may have been that journalists (and other observers, including me) paid too much attention to other "information" beyond the polls.  Similarly, in the Brexit referendum, the perception of the result as a shock wasn't based on the polls, which showed a close contest, but on a general reading of the public mood.  

Monday, November 7, 2016

Landon maintains lead

As part of an effort to avoid thinking too much about tomorrow's election, I'll return to a 1936 "straw poll" of college students that I wrote about in June.  I had found results for a number of colleges, but I knew that there were more, so I kept looking.  It turns out that the New York Times reported some additional results, mostly in a story of October 30, 1936.  It may seem strange that the Times ran a story on this, but before polls had developed, there was a lot of interest in anything that could provide hints about the outcome of the election.

Here are the additional colleges and universities, ranked from biggest Roosevelt lead on down:

                        Dem    Rep    Soc    Comm
CCNY                     62      4     12      23
New Rochelle             69 29     1       0
Manhattan                66 30      1       0
Rider                    53 43      1       1

Carnegie                 46 49      2       2
Delaware                 47 51      2 0
Drexel                   45 50      3 0
Franklin & Marshall      39 47     13 0
Upsala                   45 53      1 0
Swarthmore               37 48     13 1
Stevens                  38 52      7 2
Wash & Jefferson         25 47     26 1
Bucknell                 35 58      3 2
Penn State               37 60      2 0
Rhode Island             37 61      2 0
Rutgers                  37 62      1 0
Tufts                    32 59      4 4
MIT                      34 60      3 3
Sweetbriar               36 62      2 0
Conn Coll                30 67      2 1
Susquehanna              29 67      3 0
Gettysburg               30 69      2 0
Skidmore                 24 72      4 0
Temple                   20 67      4 9
Mass.                    22 72      5 0
Elmira                   24 73      2 0
Lehigh                   17 77      2 2

Mean                     38     55      4       2

The Vassar list was limited to elite universities, while the New York Times covered a wider range.  Despite this difference, the means were almost the same--today, I think that Republican support would be lower in elite universities than in others.  I wouldn't necessarily put much faith in the results for any particular one--it looked like turnout was generally high at the small colleges, but pretty low at the larger ones, but given the consistency it seems safe to say that college students leaned strongly Republican in the 1930s.  

Sunday, November 6, 2016

Not that complicated

A few days ago, an article in the New York Times by Amanda Taub said that working-class support for Donald Trump reflected a "crisis of white identity."  Today, Ross Douthat said that it reflected the "thinning out of families."  The basic idea in both was that "working class" (ie less educated people's) opposition to immigration is a symptom of anxiety about something else.

In September 1957, the days of the baby boom and the "affluent society," when unions were strong and no one was talking about a crisis of white identity or masculinity, the Gallup Poll asked "UNDER THE PRESENT IMMIGRATION LAWS, THE HUNGARIAN REFUGEES WHO CAME TO THIS COUNTRY AFTER THE REVOLTS LAST YEAR HAVE NO PERMANENT RESIDENCE AND CAN BE DEPORTED AT ANY TIME. DO YOU THINK THE LAW SHOULD OR SHOULD NOT BE CHANGED SO THAT THESE REFUGEES CAN STAY HERE PERMANENTLY?"
42% said yes, and 43% said no.

33% approved and 55% disapproved

With both questions, education made a difference for opinions.  For example, in 1958, 55% of the people with a college degree favored letting the refugees come to the United States, compared to 31% of those without college degrees.  The only other demographic factor that made a clear difference was that Jews were more likely to favor letting the refugees stay.

The 1957 survey also had a question about the Brown vs. Board of Education decision against school segregation--people who approved were more likely to favor letting the refugees stay.  The 1958 survey had a series of questions about whether you would vote for various religious or racial minorities for president--people who were more tolerant were more likely to favor letting the refugees come to the United States.

The Hungarian refugees were white, Christian, and could be seen as part of a clear story of oppression vs. resistance.  Despite this, most people, especially less educated people, were not in favor of letting them stay in the United States.  So the contemporary opposition to immigration, and the tendency for it to be stronger among less educated people, are not a reflection of something specific to today, but continue a long-standing pattern.  Of course, an increase in the number of immigrants today presumably makes the issue more important.  But the basic pattern is not new.

[Data from the Roper Center for Public Opinion Research]

Thursday, October 27, 2016

College professors and capitalism in the 1950s

In 1955, the Opinion Research Corporation ran a survey they called "Free Market versus Socialist Thinking," which was about basic economic issues.  I've written about it before, but until recently I hadn't noticed that it included special samples of a number of occupations:  Catholic clergy, Protestant clergy, college professors, high school teachers, business executives, other teachers, and manufacturing foremen.  They didn't give any details on how those samples were selected, but for what it's worth, here is the distribution of party identification for each group:

                              Rep    Dem   Other/None
Catholic clergy*     13%    52%     35%
Protestant clergy*  36%   37%     27%
Professors              31%    47%    22%
HS Teachers          40%   49%    11%
Executives*           62%   22%    16%
Other teachers       38%   50%     12%
Foremen                31%   47%     21%
Public                    37%   45%     18%              

The differences from the general pubic are statistically significant for the three groups marked with asterisks.

Today, college professors are far to the left of the general public.  I'm interested in figuring out how long this tendency has existed, so here are comparisons between the professor sample and the general sample on some other questions.

                       Yes       No
Professors        9%     86%
Public             18%     78%


                      Yes       No
Professors       43%     48%
Public             45%     48%


                           Stiffer   Less     As Is    
Professors           23%    22%       48%
Public                 43%    15%       36%

Professors were to the right of the general public on economic issues, although they may have been more likely to identify as Democrats.  I suspect they were to left on "social issues" like opposition to McCarthyism and racial segregation, although the survey didn't contain any questions on them.

[Data from the Roper Center for Public Opinion Research]

Saturday, October 22, 2016

Predecessing the Unprecedented

In many ways, Donald Trump is unique in American politics, but as I've mentioned before, I think he has some parallels with Ross Perot.  Below the presidential level, I wondered about Jesse Ventura, who was elected governor of Minnestota as an independent in 1998.  Ventura had more political experience than Trump--he'd served one term as mayor of a city of about 85,000.  He also apparently had libertarian leanings, while Trump has mostly been a standard conservative since being nominated.  Still, it seems that for both of them, a basic part of their image was that they were outsiders who would apply toughness and common sense.

Minnesota has a state poll, sponsored by the Minneapolis Star Tribune, that goes back many years.  I combined results from two asked in the second half of October 1998 to get breakdowns by education and gender:

High school or less                 30%
Some College                         34%
College Graduate                    20%

Men                                         31%
Women                                    16%

The differences by education and gender are similar to those found with Trump today.

 Some contemporary analyses hold that the big difference by education in support for Trump and Clinton reflects long-term economic change--less educated people, especially men, are losing out from globalization.   I don't know about Minnesota in particular, but the economic situation in the late 1990s was generally regarded as good, so the example of Ventura counts against this analysis.  An alternative is that less educated people (maybe especially less educated men) are more attracted to tough-talking outsiders and less likely to be bothered by "gaffes" or outrageous statements.  That would suggest that large educational and gender gaps are specific to this election rather than a sign of things to come.

[Data from the Roper Center for Public Opinion Research]

Sunday, October 16, 2016

Now that's what I call a gender gap

The gender breakdowns in responses to two recent survey questions:
Question 1:
        Yes      No
Men      37%     62%
Women    20%     80%

Question 2:

         Yes     No
Men      51%     46%
Women    36%     63%

The gender gaps are about equal to, or maybe somewhat bigger than, those found in the latest surveys of voting intentions in the Trump-Clinton race.  So what were these questions?

1.  Do you think the penny should be eliminated, or not?
2.  Do you think the penny is a useful coin, or should the government do away with it? ["do away with"="yes" in the table above.]

I'm not surprised that men gave more support to abolishing the penny--if I'd been asked before, looking at the results, I think I would have given odds of 3:1 or 4:1 that the difference would be in that direction--but why would the gap be so large?

[Data from the Roper Center for Public Opinion Research]

Sunday, October 9, 2016

What (probably) never was

I wasn't planning to post again this soon, but this morning I read a review of The Populist Explosion (John P. Judis) in the NY Times Book Review.  The reviewer (Jonathan Alter) said "In 1968, Wallace, running as a third-party candidate, led the Democratic nominee, Hubert Humphrey, in the polls until October...."  I was surprised to read that, so I checked the Roper Center's iPOLL database for all surveys asking about choice among Nixon, Humphrey, and Wallace in 1968.  The results:

 Wallace was always behind Humphrey.  The smallest gap was in a poll taken at the end of September and beginning of October, when Humphrey had 26% and Wallace 19%, with Nixon at 41%.  The iPOLL database is not completely comprehensive, so maybe there were surveys in which Wallace led.  But I doubt it, given the substantial margin and general stability of the figures (which were from two different polling firms, Gallup and Harris).  Maybe Alter was thinking of 1992, when I think Perot did actually lead in some polls?

     I hadn't realized that Humphrey was generally leading in the polls until August.  The Republican convention took place at the beginning of that month and the Democratic convention took place at the end.  That connects to another issue that I was going to write about, so I might as well do it now.  Back in the primaries, I was thinking about the question of whether nominating an "extreme" candidate would make a difference.  People who say that an extreme candidate will lose votes always include George McGovern as a leading example.  McGovern definitely had a liberal voting record in the senate, but so did some more successful Democratic nominees (John Kerry and Barack Obama).
That led me to wonder if the relevant issue might be a divided convention, rather than "extremism."  The two things tend to go together, because the reason for a divided convention is often an ideological challenge where one or both sides feel strongly enough to fight on until the end.  By divided, I mean one in which there was open conflict that was visible to the TV audience.    I think that the following conventions are generally agreed to meet that standard:  Republicans in 1964 and 1976 (and now 2016), and the Democrats in 1968, 1972, and 1980.  You could argue that the whole idea of a divided vs. united convention doesn't apply before the TV era--the action took place behind the scenes, and multiple ballots were common.  However, the Democratic convention of 1924, which took over 100 ballots and involved a giant rally by the KKK after a motion to condemn them had been narrowly defeated, might deserve to be included.

I added a variable for a divided convention to Fair's model and got the following results:

        Model                               estimate    std err
1.    Fair plus divided (1968-80 only)        -2.36     (1.45)
2.    Fair plus divided (including 1924)      -2.51     (1.37)
3.    Modified plus divided                   -3.41     (1.59)

The third model modifies Fair's variables in two ways:  Gerald Ford is counted as an incumbent president, and the Democrats are counted as getting 34.8% of the two-party vote in 1924, which is what they actually got.  Fair does not count Ford as an incumbent on the grounds that he was not elected to either the presidency or the vice-presidency, which doesn't make sense to me.  There was an important third-party candidate (Robert LaFollette) in 1924, and Fair estimates that most of his voters would have gone to the Democrats if he had not been in the race.  That could be true, but it's not possible to be sure, and if you were adjusting totals it seems that a better case could be made for 1968, when surveys indicated that most of Wallace's voters would have gone to Nixon as a second choice.  

The figures mean that a divided Democratic convention is estimated to cost the Democrats about 2.4% and a divided Republican convention to help them by 2.4% (in model 1).  There's a good deal of uncertainty about the size of the effect, and it's not statistically significant at the 5% level in the first two models.  Still, I'd say that there's evidence that a divided convention hurts a party.

Friday, October 7, 2016

What might have been

Suppose that the major-party nominees for president had been nondescript, people like Mike Pence or Tim Kaine.  In that case, it seems that the outcome would depend on "how things are going"--if they're going well, keep a Democrat in office; if not, turn to the Republican. There have been a number of efforts to figure out exactly which "things" matter and how much difference they make.  One of the best-known ones was proposed by Ray Fair in 1978 and has been modified and updated since then.  In Fair's formula, support for the incumbent party depends (positively) on the rate of growth of per-capita GDP in the first three quarters of the election year and the number of quarters of rapid growth over the president's term and (negatively) on the rate of change in prices over president's term.  Of course, we don't know what those will be yet (currently the GDP data only goes through the first quarter of 2016), but if you put in his predictions of what those will be, the equation predicts the Republican would get 56% of the two-party vote.  Even if you assume very strong (4%) growth in the second and third quarters, that only brings the Republican share down to 52.7%.

So according to his formula, the objective conditions put the Democrats at a big disadvantage.  Why?  Part of the reason is that there have been few quarters of rapid GDP growth over the last term--as many people have pointed out, the economic recovery has been slow.  But that was also true in 2012, when the formula predicted that Barack Obama would win the popular vote.  The main reason for the difference is that the formula also includes variables for incumbency and for the duration of time the party has been in power.  Incumbency is estimated to increase vote share by 3.0%, and being in power eight years rather than four is estimated to reduce it by 3.8%.  Putting those together, the Democratic candidate in 2016 starts off 6.8% worse than Barack Obama did in 2012, and 7.6% worse than Obama did in 2008 (when duration was counting against the Republicans), before accounting for economic conditions.  These are big differences, both in absolute and relative terms (bigger than the difference that would result from a -5% rate of growth compared to a +5%).

When I checked earlier today, the Republican share of the two-party vote in the latest polls was about 47.5.  If that holds up, it will be the largest prediction error in the period covered by the estimates--the current record is 1992, when George HW Bush was predicted to get 53.6% of the two-party vote, but actually received only 48.3%.

Friday, September 30, 2016

Two roads diverged

I'm taking a break from contemporary affairs in order to follow up on an obscure post from the early days of this blog.  In 1951, a Gallup poll asked "Which of these two jobs would you personally prefer a son of yours to take--assuming he is equally qualified:  a skilled laborer's job at $100 a week or a white-collar desk job at $75 a week?"  and a parallel question about preference between "a college professor's job at $4,000 a year or a factory foreman's job at $6,000 a year?" 66% chose the laborer's job over the white-collar deck job, and 56% chose a foreman over the professor.

What factors influenced the choice?  Two obvious possibilities are education and one's own job.  People might like their sons (or hypothetical sons) to do what they had done, so more educated people would tend to favor the jobs that required more education, and people would tend to prefer jobs that were similar to their own.  Both of those turned out to be true.  Here is a comparison by occupation, showing the percentage favoring the white-collar job for each pair.

                       Desk Job     Professor
Professional            45%           63%
Farmer                  23%           30%
Business                38%           39%
White Collar            47%           58%
Blue Collar             26%           36%
Service                 28%           32%

Another obvious possibility is income, although I'm not sure about what to expect, but Gallup didn't ask about it then.  There were no clear differences by race or gender, but there was a difference by the size of the community:  the larger the town, the more likely people were to prefer the white collar job.

                       Desk Job       Professor
under 2,500              26%            34%
2,500-10,000             28%            44%
10,000-100,000           35%            42%
100,000-500,000          34%            42%
over 500,000             39%            51% 

The difference by size of town remained statistically significant even after controlling for education and income.  As for why, my thought is that people who lived in larger places might have wider frames of reference--they would be aware of a wider range of careers and occupations--and would tend to think about prospects for advancement rather than just the immediate pay difference.

[Data from the Roper Center for Public Opinion Research]

Tuesday, September 20, 2016

More on police shootings

Over the summer, a paper by Roland Fryer got a lot of attention.  He summarized his findings:  "there are racial differences--sometimes quite large--in police use of force, even after accounting for a large set of controls ... Yet, on the most extreme use of force--offi cer-involved shootings--we are unable to detect any racial di fferences...."  You could restate this by saying that there is more anti-black bias in non-lethal force than in lethal force, and it's not clear if there is any bias (in either direction) in the use of lethal force.  The difference between lethal and non-lethal force was surprising to me--I figured that if there was bias, it would be more pronounced for the more extreme use of force.  I thought his paper was convincing on that point, partly because of the evidence he presented and partly because of a simple comparison to the data on fatal shootings by police that I've written about before.   Blacks comprise 27% of those fatally shot by the police.  This is considerably higher than their share of the total population, but not relative to other forms of negative involvement with the criminal justice system.  For example, blacks make up 39% of those arrested for violent crime.

The difficulty is in figuring out whether 27% is more, less, or about the same as what it would be if police shootings took place without regard to race--that is, if a white person and a black person in the same situation faced the same risk of being shot--which is why Fryer said "we are unable to detect any" racial differences rather than "there are no" or even "there appear to be no."

Although the data Fryer used has a lot more detail, the data I used also has some advantages:  it covers the whole nation and has more cases.  It includes a variable for whether the person who was killed was attacking a police officer, "other," or "unknown" and one for what kind of weapon, if any, they had.  I combined those into a new variable with three values:  people who were not attacking ("other" or "unknown") and unarmed (or "undetermined"), people who were armed but not attacking, and people who were attacking.  I'll call them low, medium, and high levels of apparent threat.  The breakdown of people killed by apparent threat:

            Black   Hispanic   White
low          35%      22%       39%
medium       26%      20%       48%
high         26%      16%       55%

There are statistically significant racial differences--the share of blacks and Hispanics is highest for the lowest threat level. You could also put it in terms of the chance that a person will be killed by the police when they are unarmed and not attacking:  blacks have about six times the risk of non-Hispanic whites, and Hispanics have about three times the risk.

The limitation of this comparison (and the ones Fryer did) is that we don't know the number of people who were in a comparable situation but were not fatally shot.  So it's possible that blacks and Hispanics were just less likely to be in the low-threat relative to the high threat situations.  That doesn't seem likely to me--the low-threat situations can include a wide variety of circumstances (e. g., bystanders who were killed by accident), so it seems the racial distribution of people in them in them should be closer to that in the general population.  It's also possible that police are unbiased in low-threat situations but less likely to kill blacks and Hispanics in high-threat situations.  However, the most plausible interpretation seems to be that there is some anti-black bias in fatal police shootings.

PS:  There were a total of 1,499 fatal shootings in the 18 months covered by the data: 133 low threat, 418 medium, and 948 high.

Monday, September 12, 2016

Cui bono?

In August 2008, a Gallup/USA Today poll asked "If _____ is elected president, who do you think his policies would benefit the most – the wealthy, the middle class, or the poor, or all about equally?" for John McCain and Barack Obama.  In late June and early July of this year, a survey sponsored by the American Enterprise Institute and Los Angeles times asked the same question about Donald Trump and Hillary Clinton.  The results:

                    McCain    Trump            Obama    Clinton
Wealthy              53%       54%               16%      36%
Middle Class         19%       15%               33%      19%
Poor                  1%        2%               22%      12%
All equally          25%       20%               25%      25%

The distribution of answers is almost the same for Trump as it was for McCain, but the distribution for Clinton is quite a bit different from what it had been for Obama--fewer saying the poor or middle class, and more saying the wealthy.  The 2008 survey also asked about voting intention (the 2016 survey did not).  As you might guess, people who thought a candidate would benefit the middle class or everyone about equally were a lot more likely to support him than those who thought he would benefit the rich.  Things were more complicated with the poor--the few people who thought McCain's policies would benefit the poor were overwhelmingly in favor of him (88%, although it was only eight people); the larger number who thought Obama's policies would benefit the poor were strongly against him (26%).

If you combine the 2008 estimates with the 2012 opinions, perceptions of who benefits from Clinton's policies are costing her about 7 percentage points compared to Obama.  Although I wouldn't take the exact number very seriously, it seems safe to say that she's not getting as much benefit as he was.

Why would Clinton be viewed so much differently than Obama was?  One possibility is that it's a fixed part of her image--maybe people are thinking of the well-compensated speeches she's made to Wall Street firms.  Another possibility is that the contrast with Bernie Sanders made people think of her as more favorable to rich, and that as people start focusing on the contrast with Trump perceptions will change (or maybe already have changed).  The fact that Trump is not seen as much different from McCain is interesting, since claims that he would help the working and middle classes have been a big part of his campaign.

[Data from the Roper Center for Public Opinion Research]

Tuesday, September 6, 2016

Are they blue?

One of the major themes of election reporting in this campaign has involved "blue-collar" or "working class" support for Donald Trump.  But few surveys ask people about their occupations today, so usually journalists treat class as equivalent to education.  For example, a Los Angeles Times story entitled "How do Americans view poverty? Many blue-collar whites, key to Trump, criticize poor people as lazy and content to stay on welfare," said the racial difference in opinions about who had the greatest responsibility for helping the poor "lay almost entirely with blue-collar whites--those without college degrees."  Of course, education is associated with occupation, but how strong is the association?  The Current Population Survey contains information on both education and occupation.  Civilian occupations are classified into 22 groups:  

Business and Financial Operations
Computer and mathematical science
Architecture and engineering
Life, physical, and social science
Community and social service
Education, training, and library
Arts, design, entertainment, sports, and media
Healthcare practitioner and technical

Healthcare support
Protective service
Food preparation
Building and grounds cleaning and maintenance
Personal care and service

Office and administrative support

Farming, fishing, and forestry
Construction and extraction
Installation, maintenance, and repair
Transportation and material moving

I divided them into four groups, with divisions indicated by the blank lines.  The last group of occupations corresponds with what people normally call "blue-collar"--they involve making or extracting some tangible product.   The first and third groups would clearly be "white collar" jobs--the difference is that the first generally involves more skill and higher pay than the third.  The second group is hard to classify by the blue collar/white collar distinction--like most white-collar jobs, they produce services rather than goods, but in terms of skills and autonomy, they are closer to blue-collar jobs.  

The occupational distribution of people with different amounts of education (rearranging the order to put the two white collar groups together):

                         White Collar
                Manager/Prof  Other   Service  Blue-Collar               

Not HS graduate         7%    17%        34%       41%
HS only                19%    25%        22%       34%
Some College           29%    31%        21%       19%
College Grad           64%    22%         8%        7%
Master's               84%    10%         3%        3%
Professional/Doctoral  92%     4%         2%        1%

Few blue-collar workers have college degrees, but a lot of people without college degrees are NOT blue-collar workers.   Considering everyone without a college degree, only 29% are blue-collar workers in the narrow or traditional definition.  Another 23% are service workers, so on a broader definition you could say a little over half are blue-collar workers.  

The reason that lacking a college degree is not synonymous with having a blue-collar job is partly that there just aren't that many blue-collar jobs in the United States any more, and partly that even people with only a high school diploma have a decent chance of obtaining a white-collar job (which might involve supervising blue-collar workers).  

Looking at it from one direction, it's strange that given the contemporary interest in class, few surveys bother to ask people what kind of work they do.  Looking at it from another direction it's strange that when given a straightforward measure of education, people call it "class" rather than "education."

Wednesday, August 31, 2016

More geography of police shootings

In July, I had a post on state differences in the rate of fatal shootings by police, which differ by a lot--the rate in New Mexico or Wyoming is about ten times the rate in New York and Connecticut.  Peter Moskos, a professor at John Jay College of Criminal Justice, also noted the regional differences (I got the reference from Andrew Gelman's blog) and suggested that they resulted from differences in the training and procedures in different police departments.  That suggests that we should go to the level of individual cities.  I took the 100 largest cities and calculated the expected number of deaths if they were proportional to population (which comes to about 1 per 135,000 in the period covered by the data).  The deviations were statistically significant by any standard you want (chi-square of about 267 with 99 degrees of freedom), so the differences in the rates are not just a matter of chance.

The cities with the highest ratio of actual to predicted deaths:

                Deaths  Predicted
Miami              13       3.2
San Bernardino      5       1.6
St Louis            7       2.3
Orlando             6       2.0
Baton Rouge         5       1.7
Bakersfield         8       2.7
Las Vegas          13       4.6
Reno                5       1.8
Norfolk             5       1.8
St Paul             6       2.2
Albuquerque        10       4.1

For the lowest ratio, it's a tie among eight cities--Hialeah, Irvine, Jersey City, Lexington, Lubbock, Plano, Riverside, and Winston-Salem--which had no fatal shootings.  Those are all in the lower reaches of the top 100 in population, and the expected numbers are about two in each.  With those expected values, a zero can easily come up by chance, so we can't be sure that the actual risk in those cities is actually different from the average.  But the next two lowest ratios are in big cities:  New York, with eight actual deaths and 63.6 expected, and Philadelphia, with three actual and 11.5 expected.  Those differences definitely cannot be attributed to chance.  

There seems to be a lot of geographical clustering--New York and Philadelphia are less than 100 miles apart, and the #41, 44, 45, 46, and 47 cities are all in Texas.  Maybe there is some general cultural similarity in regions that makes a difference, or maybe police departments just tend to model themselves after departments in nearby cities.  But whichever it is, there is something that needs to be explained.