Saturday, July 23, 2016

Republican Convention, Cleveland

Political conventions are fun.  The week in Cleveland was filled with nonstop action, celebrity sightings, and great people.  It was a surprisingly happy place - everyone returned smiles, and it was incredibly easy to strike up conversations with dozens of different people every day.

I was elected an alternate delegate in April, and I was able to get a guest pass for my daughter.  We stayed with the Iowa delegation and attended multiple events with them every day, meeting senators, governors, ambassadors, journalists, campaign officials, party staff members, and other delegates.  We visited the Rock & Roll Hall of Fame, ate in great restaurants, and took a boat ride on the Cuyahoga River.  Every evening we trooped over to the convention arena and stayed until around midnight.  Delegates and alternates switched places from time to time, so I was able to spend some time on the convention floor, which was really exciting.

I particularly enjoyed getting to know several state officials, including the Governor, Lieutenant Governor, Speaker of the House, and the party chair and co-chair.  Over a hectic, sleep-deprived week, true personalities tend to come out, and all of these people were genuinely nice.  The Governor, Terry Branstad, knows the minute detail of every aspect of state government.  No matter what I asked, he responded with facts, figures and astute political analysis.  He struck me as a very capable CEO of state government.  The Speaker of the Iowa House, Linda Upmeyer, is a warm and kind person with a keen interest in political philosophy.  She is also a very capable politician.  The Lieutenant Governor, Kim Reynolds, went out of her way to help my daughter make the most of the convention experience.  She is well-liked, smart, and capable.  There are big things to come in her career.  The party chair and co-chair and the state party staff were thoughtful, helpful and fun to be around.

I also met Trump campaign people.  I was impressed by their competence in managing the convention.  It seemed to me that the campaign is far better organized and much better prepared for the election than is generally thought.

Given the state of the party just two months ago, the unity I saw in Cleveland was remarkable.  I met a few #nevertrump delegates and others who were unenthusiastic about Trump, but the overwhelming majority of the delegates I met now support Trump wholeheartedly, and were disturbed by Ted Cruz's unwillingness to endorse Trump.  I am convinced that the heckling Cruz received from the floor was spontaneous, not planned or encouraged by the Trump campaign people.

Most of the talk this fall will be about personalities and individual qualifications, but I think these are distractions.  Buried in the flood of sensational stories will be an election about substance - a true contest of differing policies and philosophies; interventionist vs. non-interventionist foreign policy, Main Street vs. Wall Street, higher taxes vs. lower taxes, police vs. BLM, etc. The map will be scrambled, and previously unimaginable alliances might form.  Anything is possible, including landslides on either side, or a close contest.  It will be a lot of fun to follow.

Sunday, July 17, 2016


In a post a few weeks ago I explained that food has gotten cheaper in recent decades, even relative to the minimum wage.  I constructed an equally weighted index of rice, ground beef, eggs, frozen orange juice, and frozen vegetables, and found that it has declined by 31% relative to the CPI, 9% relative to the minimum wage.

Of the four components of my index, vegetables had the largest nominal price increase since 1980, and fresh vegetables increased even more.  This article in a recent Wall Street Journal might help to explain why.  The article explains that in Chinatown, fresh vegetables cost less than half what they would cost in a supermarket.  Broccoli, for example, was 85 cents per pound.  The price at my local HyVee is currently $2.49.  If Chinatown prices were used in my index, the real decline would have been significantly larger.

Why are vegetables so much cheaper in Chinatown?  Less regulation.  Wholesalers operate outside of the mainstream food supply network, and retailers run shoestring operations with plywood shelves lined with newspaper and shoeboxes for cash registers.

Farmer's markets also operate with less regulation than supermarkets, but they are usually aimed at upscale customers who want organic, premium quality produce and a shopping experience, so prices tend to be even higher than in grocery stores.

You are more likely to get sick from food purchased from a farmer's market or a Chinatown shop than from a supermarket chain if you do not wash and cook what you buy, but in Chinatown you will pay less than half of grocery store prices.  If anyone was serious about reducing food insecurity, they would lobby for less regulation of the produce industry.  

Sunday, July 10, 2016

Black and Blue

At an outdoor cafe in downtown Iowa City last night, we saw at least a dozen people walking by looking at their phones more intently than is normal.  Our kids let us know that they were all playing a new game called Pokemon Go, which allows them to hunt down imaginary characters using GPS coordinates.  The game was released on Tuesday, and by Friday it had taken over city streets across the country.

It is amazing that ideas and products spread so quickly through modern media.  When I was young everyone watched the same three television networks, so it was no surprise that everyone learned the same things at once.  The Internet is far more fragmented, but somehow ideas still spread very rapidly.

Just as in the days of network television, ideas do not often go viral by accident.  It takes careful planning and knowledge of how media works to promote them.  One idea that has been promoted over new and old media recently is that the country faces an epidemic of white police officers shooting black people.  Two questions come to mind:  Is it true, and who is promoting it?

The answer to the second question seems clear.  Democrats are losing among white voters, but are winning overwhelmingly among minority voters.  Encouraging high turnout by minority voters is crucial for Democrats, so they promote the idea that minorities are being victimized.  They are hoping to push the story up to the point where they would lose more white voters than the minority voters they would gain.

Is the story true?  It is difficult to tell.  There are no official statistics on the number of people killed by police in the U.S.  This is likely because in the police unions have opposed the systematic collection of these data.  Private organizations have recently attempted to gather data from news reports, and one of them, the Washington Post, found that 990 people were killed by police in 2015.  Without much time series data, we can only analyze current cross sections of data.

Of the 990 killed, 95.8% were men.  90.5% of convicted murderers in the U.S. are male, suggesting that it is normal for a group to be killed by police at a somewhat higher rate than their rate of violent crime.  The reason is probably that criminality is normally distributed, and different groups have different distribution means.  Moving out to the right tail of the distribution, which represents extreme criminality, a group with even a slightly higher mean will dominate, meaning that they will have much higher rates of extreme crime.  Men commit murder more than women, and they are overwhelmingly more likely to commit heinous murders, which accounts for their higher likelihood of being killed by police - even higher than the murder rate would suggest.

I use murder rates to compare police killings by group for several reasons.  Murder is the best reported crime, the least ambiguous in definition, and is solved more often than other crimes.  It also seems appropriate to use the most serious crime to examine the reasons police use the most serious type of force.

26.1% of the 990 killed by police in 2015 were black.  According to Department of Justice figures, 52.5% of murderers are black, suggesting that blacks are killed by police less frequently than would be expected given that rate of offence.  The recent study promoted by the New York Times this week argues that blacks are more likely to have force used against them by police even when differential crime rates are taken into account, but the table on page 18 of the study shows that the mean use of force rate per 1000 arrests for violent crimes is 731 for blacks and 1003 for whites!  On page 20 the study shows that the rate of use of guns, tasers, dogs, and hands during arrests for violent offenses is higher for whites than for blacks, while the use of pepper spray and batons is higher for blacks.  Use of force is higher for black for arrests for nonviolent crimes, but no attempt is made to control for severity of offense in this category.  I hate to be blunt, but this study finds the opposite of what it claims to find.

Using the very limited time series data we have, are killings by police increasing?  One group has been gathering data since mid-2013.  Annualizing partial year data, total killings are as follows:

2013   1,160
2014   1,111
2015   1,208
2016   1,166

There does not appear to be an upward trend.  More than 12 million arrests are made each year in the U.S.  Even if every police shooting is considered an error, the lethal error rate of U.S. doctors is still 144 times that of police.

Moving from black to blue, what about killings of police?  Here the data are much better, probably because police unions want the data collected.  I used a time series of law enforcement officers killed in the line of duty going back to 1791.  I calculated the rate of police deaths per homicide each year.  The idea is that police officers are more likely to be killed when the number of crimes is high.  The police death rate per crime is a measure of danger for police holding the overall crime rate constant.

I estimated the number of U.S. homicides from the homicide rate for New York City, which is reported back to colonial times.  The following is a 10 year moving average of the result:

Police deaths after controlling for the murder rate increased sharply after the Civil War, and continued to increase until roughly the end of frontier settlement.  It increased again after World War I until the beginning of the Great Depression, when it fell dramatically until about 1995.  It increased rapidly until about 2006, when the rate of increase declined, but it is still increasing slowly.  The increase after 1995 is probably the result of more police on the streets using more aggressive tactics.

Contrary to widespread opinion, police killings of civilians do not appear to be increasing, and the rate at which minorities are killed is lower than we would expect given the rate at which members of minority groups commit crimes.  Killings of police are still at historically low levels given the current crime rate, but the rate has been increasing over the past two decades.

Perhaps the attempt to convince voters that police are misbehaving will lead to more violence against police.  The events of this week suggest that this is possible, but it is not yet showing up in the data.

Saturday, July 2, 2016

Global Warming Update

Something new is happening to world temperatures.  For years I have been reporting that regression analysis of temperatures, CO2 levels, the El Nino effect, and sunspot activity shows that the post-1997 data show no relationship between changes in CO2 and changes in temperature.

Adding data from late 2015 and YTD 2016 changes the results, and a strong El Nino that began last fall seems to be the cause.  Interestingly, climate scientists see the causality working like this:  El Nino -> temperatures - > CO2.  Of course, they still say changes in CO2 cause changes in temperature, but they accept that at the same time, changes in temperature might cause changes in CO2.

In a post three years ago, I wrote that the direction of causality seemed to have shifted during the 1990s - Granger causality tests that I ran showed CO2 causing temperature and not the other way around from 1958 until the 1990s, but from the 1990s to the present CO2 did not appear to cause temperature.  (1958 is when regular atmospheric CO2 measurements from Mauna Loa were first taken)

Adding data through 2016 sharpens these results.  Using data from the second half of the time series, 1987 through 2016, Granger causality tests show temperature strongly causing CO2, but CO2 does not appear to cause temperature!  The result is not very sensitive to the choice of beginning date or length of the time period.  Even a 1978-2016 time period shows statistical significance for temperature causing CO2, but not for CO2 causing temperature.  A 1968-2016 time period still shows the temperature -> CO2 effect as stronger than the CO2 -> temperature relationship.

These results suggest that it is possible to have a situation in which changes in CO2 cause changes in global temperature, but we do not appear to be in such a situation now.  For now, the world is warming as it has been since the last ice age, and increased temperatures are increasing CO2 levels along with human activity, but changes in CO2 levels are not affecting temperatures.

I can't say exactly what this means, but I think it is reasonable to suggest that these results cast some doubt on the anthropomorphic global warming hypothesis.

Data and SAS code are here.

Thursday, June 30, 2016

Random Ideas About the EU

Betting odds on a Trump victory are now similar to the odds of Brexit shortly before that vote.

A New York Times article today on the reasons that the EU does not seem democratic makes no mention of language.  Isn't the fact that most European voters cannot understand a speech by any European politician a major reason?

Experts said Brexit would be catastrophic, but the FTSE 100 is trading above where it was when people expected the measure to fail.  Will voters believe experts when they say the same about the election of Trump?

Sunday, June 26, 2016

Brexit, Russia, and Democracy

Here is President Putin's reaction to the Brexit referendum:
"Why did he initiate this referendum?  Why did he do it?"  Mr. Putin asked.  "To intimidate Europe, or to threaten someone?  What is the point of this if he himself opposes the idea?"
Putin really doesn't get democracy.  I am not naive about power politics, but majority opinion matters in Western countries.

Politics also isn't everything in the West.  Cameron has options outside of politics - resigning isn't the end of the world for him as it would be for Putin.

Saturday, June 25, 2016


If we had a neutral press, Americans would be thinking a lot about how Hillary and Bill Clinton managed to accumulate over $100 million during their years of public service.  Neither had significant family money, business activities, or successful professional careers.  The majority of their combined lives were spent as relatively low-paid bureaucrats and politicians.

If we read something similar about a foreign ruling couple we would assume banana republic type corruption.  

In order to deflect attention from the Clintons, Democrats (e.g. Krugman) call Republicans cheap con men, grifters, sleazy, and snake-oil salesmen.

Leaving aside the fact that snake-oil really was good for you, "sleazy" is a slur against Silesians, and that the origins of the term "grift" make the term a perfect fit for the Clinton Foundation, these attacks are designed to paint Republicans as lovers of money, while it is becoming clear that the opposite is true.

Money obsession is a characteristic of the top and bottom of the income distribution, but not so much of the middle classes.  The poor are obsessed with money in order to survive, while most of the rich are rich because they think so much about money.

In the "Brexit" vote, the rich lost a lot of money, while the middle class decided that they cared more about their country and heritage than about a percentage point or two of GDP.  Young people (who do not have much money) and poor immigrants voted to remain in the EU.  (The Scottish and Irish voted to remain for their own nationalist reasons)  In the U.S., Democrats have also puzzled over why Republicans vote against their own economic interests.

Billionaires increasingly support Democrats - Donald Trump is a traitor to his class, just as Franklin Roosevelt was, but the parties have changed from rich and upper middle against the working class and poor to top and bottom against the middle.

Trump is not bourgeois, but he is trying to articulate bourgeois positions on the issues.  Hillary comes from the bourgeoisie, but the Clintons do not represent bourgeois values.  Trump the New York CIty playboy billionaire has become the candidate of the middle class, while Clinton, the former midwestern Goldwater Girl, has become a world-class grifter.