A working-class voter is something to be

There has been a lot of talk about "white working-class voters" in this election season.  I did a Google News search for that phrase and found 623 mentions, compared to 7 for "white middle class voters."  That raises the question of why the candidates should be more interested in one type of voter rather than another--a vote is a vote.  One possible reason is the size of the group--people used to talk a lot about the "farm vote," but that's faded as the number of farmers has declined.  There are a lot white working-class people, but also a lot of white middle-class people, so that doesn't explain why the working class is getting all the attention now.  Another possibility is the possibility of changing votes--a group that can be persuaded will get attention, while one that is pretty much fixed can be written off (or taken for granted).  That raises the question of whether there are bigger "swings" among working-class voters.  If so, then the focus on the (white) working class would make sense. 

Before trying to answer,  I should note that discussions in the media usually define "class" by income or education.  My impression is that education is the more common standard.  Sociologists normally define "class" by some combination of occupation and employment status, but most opinion polls stopped asking about occupation in the 1970s.  Back then, a lot of people thought that class didn't matter any more, or that the working class was disappearing. 

The American National Election Studies go back to 1952 and contain measures of income, occupation, and education.  For income and occupation, there's no evidence that any group varies more than another, but for education there is:  the "swing" among people without a high school degree is about 50% larger than the "swing" among people with a college degree.  For example, say that the Democrats gain 4% vote share among college graduates than they did in the previous election--then they will probably gain about 6% among people without a college degree.  (And about 5% among high school graduates).  Candidates like Barry Goldwater and George McGovern did poorly among college graduates, but very poorly among less educated voters. 

This pattern could be explained by saying that college graduates are more likely to have a strong ideological position which will keep them voting Democratic or Republican, while less educated voters are more likely to be influenced by everything else that's going on at the time. "Everything else" covers a lot of ground, but I looked at one thing that's easy to measure:  incumbency.  Being an incumbent president seems to help among all groups, but has a bigger impact among less educated voters.  That also seems reasonable, if less educated voters tend to be less politically engaged and less ideological.  This is based on a small sample (14 elections), but if the pattern holds, Romney's efforts to win over "working class" (less educated) voters won't pay off.

More Stimulus

In my last post, I found no relation between spending growth (2007-9) and economic growth (2009-11) using data from the International Monetary Fund.  I tried the same thing with data from the Organisation for Economic Cooperation and Development. More precisely, I looked at the relation between "general government final consumption expenditure" (P3S13) and "gross domestic product, expenditure approach" (B1_GE).   "Final consumption expenditure" is government spending on goods and services; it does not include transfers of money (like social security payments).  In principle, this kind of spending should be more relevant to stimulus, since individuals who receive transfers might save them rather than spend them. 

The first figure shows the relation between change in government consumption expenditure from 2007-9 and GDP growth in 2009-10 (2010 was the most recent data available for most countries).  The IMF figures on GDP go to 2011, so the second figure shows the relation between change in government consumption and the IMF numbers on GDP change in 2009-11.  Both ways, there is a positive relationship--countries that increased spending in 2007-9 had higher growth in later years. The ranking of countries on change in government consumption seems to correspond pretty well to the conventional impressions about which countries followed stimulus or austerity.

Of course, this is a quick and dirty analysis, but it suggests that stimulus worked.  For example, the difference in 2009-11 economic growth between the US and UK (1.8% vs. 0%) is almost exactly what would be predicted from the 2007-9 difference in spending growth (4.6% vs. 2.4%).

Stimulus

Last week, Arthur Laffer published a piece in the Wall Street Journal in which he presented data purporting to show that among industrialized nations, "those that stimulated the most from 2007-2009 saw the least growth in subsequent GDP rates."  Some critics (follow links from Paul Krugman) pointed out that the things he measured didn't make sense:  most obviously, the growth data covered 2006-9, which isn't "subsequent" to 2007-9.  But that led me to wonder what you would see if you did a straightforward comparison of change in government spending and change in GDP.   I looked at three periods:  2005-7, 2007-9, and 2009-11, which roughly you could call pre-crisis, crisis, and recovery (such as it is).   Here is the relationship in graphical form:

That is, little or no relationship.  Countries in which spending grew a lot in 2007-9 (e. g., Chile and the Slovak Republic) and countries in which spending fell (e. g., Israel and Switzerland) grew about equally much, on the average, in 2009-11. 

It turns out that spending in 2009-11 is related to GDP in 2007-9--the higher the GDP growth (or smaller the losses), the higher the subsequent spending growth.  In other words, countries that had suffered the most in 2007-9 tended to subsequently follow "austerity" policies.   

An interesting point is that some of the countries that are generally regarded as examples of austerity had substantial spending growth in 2007-9, 2009-11, or both. For example, these numbers show Ireland and Estonia as having tried stimulus in 2007-9.  Laffer used IMF data, and I did too.  I don't know is the mismatch reflects a problem in the IMF data, or in what "everyone knows," or some technical issue in the definition of government spending. 

Liberals and Conservatives at Thanksgiving

There is a lot of discussion of psychological and/or lifestyle differences between liberals and conservatives.  The differences aren't all that large, when you look at them closely--see this example--but there are some, and they are fun to think about.

An October 2010 survey conducted by CBS News for Vanity Fair and 60 Minutes contained the following question:  "Besides turkey and football, Thanksgiving isn't Thanksgiving without which ONE of the following:  a good nap, a good family argument, a travel delay, a stiff drink, or indigestion."  The results, by self-described political views:

                                  Nap             Argument        Delay         Drink      Indigestion   DK
Liberal                       34%                18%                 5%          16%       19%            9%
Moderate                   33%                17%                 7%          12%       19%          11%
Conservative               43%               19%                 4%            8%         9%          17%
DK                            28%                19%                 2%           10%       14%         26%

So conservatives were more likely to say a nap or have no answer; liberals were more likely to go for a drink or indigestion.  (The differences are statistically significant).

 One interpretation is that conservatives are more likely to be placid people with good digestion, while liberals are more likely to get stressed.  Or maybe conservatives don't like to be negative about tradition.  A family argument, travel delay, stiff drink, or indigestion are all somewhat cynical responses (at least to me, a "stiff drink" suggests a drink taken to relieve stress, not for pure enjoyment),  so that would leave people who see no downside to Thanksgiving with a nap or nothing. 

More on Gun Control

In the NY Times, Ross Douthat links to some Gallup data showing a decline in support for gun control since 1990, and offers an interesting argument about the reason.  He suggests that it's one aspect of a general move towards individualism--the idea that people should be allowed do what they want as long as it doesn't directly hurt other people.  With many issues, this idea has "liberal" implications (same-sex marriage, legalization of marijuana) but with gun control it has "conservative" implications.  Before his column came out, I found another series of questions on gun control, "do you favor or oppose stricter gun control laws?"  (There were minor variations in question wording, but they seem unlikely to have made a difference).  Here is the plot of percent favor minus percent opposed:

There is also some downward trend, but also a lot of short-term change (maybe in response to events in the news).  I think this volatility counts against the idea that the decline reflects a general change in values.  That kind of change is driven mainly by generational replacement, producing a gradual and steady change (which is what you see with same-sex marriage).   Also, as Douthat notes, the decline in support for gun control 1990 might also be explained as a response to declining crime rates.   If support for gun control declined in the 1970s and 1980s, when crime was steady or increasing, that would be stronger evidence that it was related to a rise of individualism.  Gallup also asked a question about handguns that goes back farther, and my impression is that opinion didn't start to show a trend until about 1990.  I may take a closer look in a future post. 

Declining Support for Gun Control?

The other day I heard something on NPR saying that, by one measure, support for gun control had declined from over 80% (I believe they said that was in the 1980s) to less than half today.  I'd thought that it had been pretty steady over time, so I checked for surveys that asked comparable questions over a period of time. 

There have been a number of questions on assault rifles, with the basic form "are you for or against a law which would make it illegal to manufacture, sell, or possess semi-automatic guns known as assault rifles?"

             For        Against
1995    68%       30%
1996    57%       42%
2000    59%       39%
2004    50%       46%
2011    43%       53%

It looks like a substantial decline in support for a ban on assault rifles.  But the 1995 question had an introduction:   "if you would favor or oppose the following proposal which some people have made to reduce crime:  a ban on the manufacture, sale, etc."  The mention of reducing crime might have made people more inclined to say they favored it.  If you leave that question aside, there has been some decline in support for a ban, although not as dramatic.  

There was a question in 1989 asking "do you think it should be legal or illegal to manufacture and sell semi-automatic rifles known as assault rifles"; only 20% said it should be legal and 73% said illegal.     But there was one in 2006 asking if assault rifles "should be sold to the general public or should their sales be limited to the military and police"; 15% said they should be sold to the general public and 82% that their sales should be limited to the military and police.  You could interpret the different results as ambivalence or confusion, or as people making a distinction (some people favoring regulation but not a complete ban). 

Candidates' Tax Returns

Mitt Romney's unwillingness to release more than two years of tax returns is currently getting a lot of attention in the media.   The Romney campaign says "there is no evidence that voters care about this.  They think they know enough about Mitt Romney's finances."  Of course, this is what you'd expect the Romney campaign to say, but there are a lot of things that seem important to journalists and political insiders but not to the public (pretty much all "gaffes," for example).  I didn't expect to find many surveys asking about Romney's tax returns, although that may change soon if the issue stays in the news, so I looked for more general questions about candidates and tax returns.  To my surprise, there were only a handful.  The most interesting one is from back in 1988, asking "if it would prevent you from voting for the candidate or if it would have no effect" if a candidate "refused to release tax returns to the public."  72% said it would prevent them and 19% said it would have no effect.  They also asked about a number of other things.  The complete list:
                                          Prevent from voting for

Was known to have physically

    abused spouse or child                        91%

Refused to release tax returns                    72%

Was homosexual                                    65%

Had been convicted for drunk driving              57%

Said religion dictated his/her decisions          57%

Did not believe in God                            56%

Refused to be tested for AIDS                     54%

Had used drugs occasionally                       47%

 Had extra-marital affairs                        43%

Family member had criminal conviction             21%

Has a child out of wedlock                        21%

So there's evidence that voters care about releasing tax returns in general, although not about whether two years versus ten years makes a difference.