Sunday, May 25, 2008

Tamino's Bet Month Four

As I have noted previously, Tamino has made a "bet" that global temperatures would continue to increase between now and 2015 in line with the 1975 to 2007 trend. The conditions of the bet for the year 2008 are as follows:

1. If the average global temperature anomaly, as measured by NASA GISS, equals or exceeds .7350 degrees Celsius, the "still-warming" side will receive one point.
2. If the average global temperature anomaly, as measured by NASA GISS, is less than or equal to .4035 degrees Celsius, the "not-warming" side will receive one point.
3. If the average global temperature anomaly, as measured by NASA GISS, falls between these two figures, both sides will receive zero points in 2008.

The April figures reveal a significant decline in the NASA GISS global temperature anomaly from March as well as a modest downward revision in March's global temperature anomaly. the previous month. The April global anomaly was .41 degrees Celsius, .19 degrees lower than the revised March global anomaly of .60 degrees (the original figure was .67 degrees). The average NASA GISS global temperature anomaly thus far for the first four months of 2008 is .35 degrees Celsius. In order for the "not-warming" side to win a point in 2008, the average anomaly for the reminder of the year would need to be below .4303 degrees. My own forecast for the remainder of the year is for the global temperature anomaly to average .5059 degrees. Thus, I don't expect that the "not-warming" side will win the point in 2008.

Sunday, May 11, 2008

Barack Obama's Choice

Barack Obama is the Democratic nominee for President of the United States in 2008. His narrow, deep base of Blacks and urban, upper-middle class liberal Whites was numerically strong enough to battle Hillary Clinton's broad base of large state White, Latino and Asian center-right Democrats in the primary states to a draw. His organizational skills in the (mostly small) caucus states established a sufficient lead in the delegate race to give him the nomination on points.

Now comes the hard part - getting elected President. Obama now faces a choice not unlike that John Kennedy faced in 1960 at the end of the Democratic nomination campaign. In 1960, a young and relatively inexperienced junior senator from Massachusetts outdueled the politically savy and deeply experienced Senate Majority Leader Lyndon Johnson of Texas to win his party's nomination. But John Kennedy had a political liability that no presidential candidate had ever before overcome - he was a Roman Catholic in a nation that, to a large degree, still viewed itself as a Protestant country and viewed Catholics as outsiders. Only 32 years earlier, in 1928, another urban Irish Roman Catholic Democratic nominee, Al Smith, faced racial and religious descrimination unlike that ever faced before or since by a major party presidential candidate in the end suffering a crushing defeat. To counter that lingering racial and religious discrimination, John Kennedy reached out to his bitter rival Lyndon Johnson and offered him the vice-presidential nomination.

Barack Obama has a political liability that no presidential candidate has ever before overcome - he is a (half) Black man in a nation that, to a large degree, still views Black Americans as outsiders. There is one person who can help him overcome that liability and reach out to those people who may be uncomfortable about pulling the lever for him, namely Hillary Clinton. The question is, will he do it? It is clear from the actions that Clinton has taken since last Tuesday's primaries that she knows that the nomination race is over and that she is positioning herself for the vice-presidential nod. Will he do what John Kennedy did in 1960 and reach out and ask Clinton to be his running mate? If so, he will become President of the United States. But, in doing so, he will have to share the limelight with the most dynamic female politician of this (or any era) in American history. Will he do so?

A Global Temperature Anomaly Forecast to 2037

[note: The below post has several minor revisions and adjustments to the original post]

I have developed a model of the global temperature anomaly (using the NOAA NCDC time series as the benchmark series) for the 2008 to 2037 period based on three assumptions:

1. That the general temperature trend during this period will be similar to that observed in the 1880 to 1911 and 1943 to 1975 cooling periods, and;
2. That the atmospheric concentration of CO2 will increase linearly to 600 ppm by the end of the 21st century. Using a formula for radiative forcing posted by Gavin Schmidt on and assuming an increase in average global temperatures of 0.3 degrees Celsius for 1 W per square meter, I have calculated the average direct annual anthropogenic effect during the 2008 to 2038 period;
3. That the equilibrium sensitivity (that is, the feedback effect) from an increase in atmospheric CO2 concentration is zero.

Thus, if one assumes that the underlying temperature trend (minus the direct anthropogenic effect) in the 2008 to 2037 period will be the same as that for the 1880 to 1911 and 1943 to 1975 periods, the forecast provides a test of the equilibrium sensitivity from an increase in atmospheric CO2 concentration. Should the actual temperature anomaly trend above the forecast, then we may say that the equlibrium sensitivity is positive and the degree to which the actual temperature is above this level would provide an indication of the equilibrium sensitivity.

The figure below shows the average annual forecasted global temperature anomaly for the 2008 to 2037 period based on the above assumptions:

The temperature anomaly is given in degrees Celsius. The data point for 2007 represents the actual average global temperature anomaly according to NOAA NCDC. With the exception of a spike in the anomaly to a new record high in 2015, the forecast is for average global temperatures to decline by .19 degrees Celsius between 2007 and 2019 before recovering to a new all-time high during the 2020's. The forecast anticipates that between 2o21 and the end of the period in 2037, the global temperature anomaly will fluctuate between 0.60 and 0.75 degrees Celsius.

It is interesting to note that the Keenlyside et. al. paper recently published in Nature suggested that "global surface temperature may not increase over the next decade" before increasing beyond 2018. As a result, the forecast here can be measured against the Keenlyside et. al. general forecast to provide a good test of the actual equilibrium sensitivity level. That is, should the global temperature anomaly remain at or near the 2007 average for the following decade, it would be reasonable to consider that proof of a significantly positive equilibrium sensitivity.

Because the forecast is monthly and because the relevant forecast period actually began in October of 2006, we can compare the forecast thus far with the actual anomaly figures through March of 2008. The figure below makes this comparison:

As in the previous figure, temperatures are in degrees Celsius. The actual average anomaly over the 18 month period in the above figure is .5407 degrees while the forecast average over the same period is .5850 degrees. Since the mean temperature anomaly for the 18 months prior to October 2006 was .5640 degrees, a simple projection based on that average would have performed slightly better than the forecast thus far.

The model's global temperature anomaly forecast for April to December is:

We can use the global temperature forecast for 2008 to 2015 to compare with Tamino's "bet" that I have been highlighting every month in this blog. The table below shows Tamino's annual average threshold below which the "not-warming" side wins with the global temperature anomaly forecast for the 2008 to 2015 period:

The bold highlighted rows show those years in which the forecast average anomaly is below Tamino's threshold in which the "not warming" side "wins" a point in the bet. Since the bet is whether one side or the other wins two points in the 2008 to 2015 period and since the "still warming" side wins a point only if the global temperature anomaly is above .7350 degrees, my forecast is that the "not warming" side will win the "bet" at the end of 2013. Stay tuned!

Monday, May 5, 2008

Race and the Democratic Race

As I noted in an earlier post, the Democratic contest has been cast in terms of a White-Black racial divide and a blue collar-white collar class divide. In this post I will show you that while little mention has thus far been given to the voting patterns of other ethnic groups during this contest, the evidence suggests that members of other ethnic groups appear no more disposed to vote for Obama than are White voters. The inescapable conclusion is that Barack Obama's electoral base is incredibly narrow while that of Hillary Clinton's cuts across all ethnic and gender groups with the exception of Black voters.

To make this analysis, I have used the same exit polling data used in a previous post but have extended the analysis to every state (with the exception of Michigan and Washington State) that has thus far held a primary as of this post's date. For the state level data, I have used election data from Dave Leip's website. I have used CNN's website for exit polling data. Exit polling data is available for all primaries with the exception of the District of Columbia. Actual election results across the 29 primaries, excluding Michigan and Washington State, thus far suggests that the contest thus far is overall astonishingly close:

Obama 14,632,409 50.01%
Clinton 14,624,325 49.99%

Although it is commonly supposed that Obama has a clear lead across the primaries held thus far (especially by Obama supporters), in fact, he is at this point ahead by only about 8,000 votes out of more than 29,000,000 cast.

A bit less than two-thirds of the Democratic electorate are White, just about one-fifth are Black and a bit less than one-fifth are members of another ethnic group:

Note that the "other" group includes all voters in the District of Columbia since no exit poll was available there. Since roughly three-quarters of voters in the District were probably Black, the overall percentage of Black voters in the Democratic electorate would be about 19.5%.

White voters support Hillary Clinton over Barack Obama by a comfortable 16 percentage point margin:

Black support for Obama is nothing short of monolithic:

Roughly six of every seven Black primary voters are Obama supporters. It is fair to say that Black support for Obama is racially based.

Among other ethnic groups (primarily Latino voters and, to a lesser extent, Asian voters), Hillary Clinton's overall level of support appears to be about the same as that among White voters:

Since this "other" group includes all District of Columbia voters, the adjusted level of support for Clinton among "other" ethnic groups would likely be about 59%. Thus, across more than 80% of the total electorate, Obama's support is tepid and his razor edged advantage overall is based on an overwhelming level of support across a 20% sliver of the electorate.

Saturday, May 3, 2008

Innumeracy, the Media and the Indiana Primary

An article posted today on the Yahoo web site demonstrates how widespread innumeracy in America and, to a lesser extent, the degree of media bias regarding the Barack Obama campaign. The article was titled "Early Indiana turnout heavy in strong Obama counties" and it stated the following:

Early voting in Indiana could offer some encouragement to presidential hopeful Barack Obama, who needs a victory in its upcoming primary after a tough few weeks on the campaign trail.

Obama victories in the Indiana and North Carolina primaries on May 6 could help him regain momentum in his nomination fight against Hillary Rodham Clinton. Obama has been on the defensive because of comments by his former pastor, Jeremiah Wright, and his own comments about people in small towns growing bitter.

About 20 percent of the 127,000-plus absentee ballots received as of early Friday were cast in three Indiana counties — Marion, Monroe and Lake — that political observers believe Obama is strongly favored to win.

The impression one gets from reading the article is that the share of early votes being cast in these three counties likely favorable to Obama is high enough to suggest that Obama's chances to pull off an upset in Indiana are higher than commonly supposed. Now, this would be the case if the total share of Democratic voters in these three counties over the entire state was much lower than 20%. This can be easily checked. From Dave Leip's web site, we can see that John Kerry received 969,011 votes in Indiana in the 2004 presidential election. The total number of votes cast for the Democratic candidate for president in 2004 in these three counties was 299,957 or 31.0% of the total Democratic votes in the state.

In other words, the 20% share of total absentee votes received from these three heavily Democratic counties as of last Friday represented only about two-thirds of the total share of Democratic votes in these three counties in the 2004 presidential election. As I have pointed out in several earlier posts on this blog, Obama's support is concentrated in urban areas. Thus, while one should anticipate that Barack Obama should do well in these three counties, one would also anticipate that Hillary Clinton should run strongly outside of these three counties. Had the author of this article bothered to take a few seconds to locate the 2004 presidential election totals, this person would have realized that the early vote is heavily concentrated outside of the three counties where Obama is likely to prevail. Rather than suggesting that the early turnout is favoring Obama, this information suggests that Obama may be facing a significant defeat next Tuesday in Indiana even beyond that predicted at present. Unfortunately, the writer appears to lack an ability to understand numbers and appears to be biased towards Barack Obama. In the end, he seems unable to see reality as it is actually unfolding.

Friday, May 2, 2008

Climate Science, Fantasy and Reality

In November last year, less than six months ago, the world's most recognized and esteemed climate scientists and other members of the Intergovernmental Panel on Climate Change gathered in Valencia, Spain to formally ratify the AR4 Synthesis Report on climate change. According to information presented on the title page of this report the authors of the draft from which the report was put together included Dr. Rajendra Pachauri, Chairman of the IPCC and Dr. Stephen Schneider, the prominent environmental biologist and climate scientist. In Section 3.2 of the report, titled "Projections of future changes in climate" (page 45) the report states:

For the next two decades a warming of about 0.2°C per decade is projected for a range of SRES emissions scenarios . . . Since the IPCC’s first report in 1990, assessed projections have suggested global averaged temperature increases between about 0.15 and 0.3°C per decade from 1990 to 2005. This can now be compared with observed values of about 0.2°C per decade, strengthening confidence in near-term projections [my emphasis].

Six months is apparently a long time in climate science. According to the abstract of a letter published in the May 1, 2008 edition of Nature five European-based scientists make the following rather fascinating forecast for global temperatures over the next decade:

Our results suggest that global surface temperature may not increase over the next decade, as natural climate variations in the North Atlantic and tropical Pacific temporarily offset the projected anthropogenic warming.

Hmmm, projections of a continued, reasonably steady increase in global temperatures made just six months ago are now, apparently, way, way past their pull date. But, don't be fooled by this short term inability to get it even remotely right - the long term consensus projections remain correct! Indeed, the media spin process already underway from climate scientists suggests a perception of the forecast's likely accuracy as well as a deep concern for the effect it will have on public opinion. In an astonishing piece of journalistic nonsense on the BBC website, an article by Richard Black acknowledges that the Atlantic Multidecadal Oscillation (AMO) may be responsible for the increase in temperatures during the 1940's but then stops short of suggesting that the same could have been the case for the warming period from the mid-1970's to the recent past. He then quotes Noel Keenlyside of the Leibniz Institute of Marine Sciences, one of the authors of the above noted Nature letter:

"One message from our study is that in the short term, you can see changes in the global mean temperature that you might not expect given the reports of the Intergovernmental Panel on Climate Change (IPCC) . . . In the long term, radiative forcing (the Earth's energy balance) dominates. But it's important for policymakers to realise the pattern."

Ah, yes, the pattern itself is crystal clear even if we are confused about the short term details. An easily arranged visit to the UK's Hadley Centre website suggests just how much the short term details are indeed confused. Prominately displayed second from the top under "Latest News and Events" is the tag titled "Climate Change and Global Variability". Click on this and you can quickly read the following piece of spin:

"There are a number of natural factors contributing to so-called interannual variability, the single most important being the El Niño Southern Oscillation or ENSO. The global climate is currently being influenced by the cold phase of this oscillation, known as La Niña (see Met Office: Expert speaks on La Niña).

The current La Niña began to develop in early 2007, having a significant cooling effect on the global average temperature. Despite this, 2007 was one of the ten warmest years since global records began in 1850 with a temperature some 0.4 °C above average.

The La Niña has strengthened further during early 2008 and is now the strongest since 1988/89, significantly contributing to a lower January temperature in 2008 compared to recent years. In addition, global average temperature has been influenced by very cold land temperatures in parts of the northern hemisphere and extensive snow cover.

However, once La Niña declines, it is very likely that renewed warming will occur as was the case when the Earth emerged from the strong La Niña events of 1989 and 1999."

Note that this posting was made on March 5th, less than two months ago. While in line with the IPCC's pronouncement made the previous November, the posting looks more than a little silly given the Nature letter noted above. One wonders how long this post will stay on the Hadley Centre website. At least it is preserved here for posterity.

Getting back to the BBC piece, we come just a bit later to the following desparate spin line by Mr. Black:

"The projection [of cooling temperatures over the next decade] does not come as a surprise to climate scientists, though it may to a public that has perhaps become used to the idea that the rapid temperature rises seen through the 1990s are a permanent phenomenon."

The only comment needed here is a reference to the IPCC statement made six months ago made above. Just six months ago, climate scientists had "strengthened confidence" in near term projections of monotonic temperature increases; however they are in no way surprised that their "strengthened confidence" in these projections has been revealed so quickly to be completely incorrect.

Regarding monotonic temperature increases, a piece by Andrew Revkin in the New York Times regarding the
Nature letter produces a quote by Kevin Trenberth, a climate scientist at the National Center for Atmospheric Research in Boulder, Colo.:

“Too many think global warming means monotonic relentless warming everywhere year after year,” Dr. Trenberth said. “It does not happen that way.”

Roger Pieke Sr. has an interesting reaction to this comment in his blog:

This is an amazing error [by Trenberth]. Global warming does require a more-or-less monotonic increase in warming (in the absence of a major volcanic eruption) as illustrated in all available multi-decadal global model runs . . . . This essentially monotonic report is even emphasized in the 2007 IPCC Summary for Policymakers".

There you have it - climate science, fantasy and reality. Much more to come!!

Thursday, May 1, 2008

Barack Obama's Electability Post Pennsylvania

Barack Obama's 54.6%-45.4% defeat to Hillary Clinton in the Pennsylvania primary did nothing to dispel the notion that he is the weakest presidential nomination front runner in the primary system era (that is, since 1976). His defeat marked the third consecutive defeat in a major state and the margin in Pennsylvania was similar to the 10.3% margin he suffered in Ohio. More importantly, the same geographical split between central city and countryside observed in a number of previous primary states remained well defined in Pennsylvania. Although Obama won the overall urban vote in Philadelphia and Pittsburgh by a 57%-43% margin (carrying Philadelphia by a 65%-35% margin), he was swamped in the remainder of the state:

Obama received less than 40% of the vote outside of these two major urban centers. For comparison, the results for Ohio for the remainder of the state outside of the urban centers of Cleveland, Columbus and Cincinnati are noted below:

The results in the two states are within 0.2 percentage points of each other. Since the urban center vote share was 32% of the total statewide vote in Pennsylvania and 31% in Ohio, this similarity appears to reflect the same dynamic at work in both states. And, since these primaries were six weeks apart, it is clear that Obama was unable to reach beyond his narrow urban constinuency between early March and late April.

I noted in my previous post on this topic that little was being said concerning Obama's inability to make significant headway outside of central city areas. Since then, and especially since the Pennsylvania primary results, it seems to be the only thing on people's minds. However, while I have focused on the geographical aspect of this difference, many people are casting this difference in terms of race and class.

When the effect of race on the Democratic presidential battle is discussed, the discussion invariably revolves around the seeming reluctance of White voters to support Obama because of his race. As Adam Nagourney asked in a recent NYT piece,

Why has [Obama] been unable to win over enough working-class and white voters to wrap up the Democratic nomination? . . . The composition of Mrs. Clinton’s support — or, looked at another way, the makeup of voters who have proved reluctant to embrace Mr. Obama — has Democrats wondering, if not worrying, about what role race may be playing.

Curiously, Nagourney himself provides evidence in his article that the role of race may be overemphasized. According to the Pennsylvania exit poll quoted by Nagourney, 18% of Democrats stated that race mattered to them in the contest and that "just" 63% of such voters said that they would support Sen. Obama in the general election. This seems significant until one reads on just a bit further:

The exit poll found that 69 percent of white Democrats would vote for Mr. Obama in a general election campaign over Mr. McCain; 73 percent of black Democrats said they would vote for Mrs. Clinton over Mr. McCain.

The 4 percentage point difference between White Democrats who said they would support Sen. Obama in the fall and Black Democrats who said they would support Sen. Clinton in the fall is small and likely statistically insignificant. Taken in this light, the fact that 63% of voters who stated that race was important to them and that they would vote for Obama in the fall suggests that these voters attitudes are not terribly different percentage either from that of White or Black voters in general for the candidate of whose race different from theirs. But in fact, race did matter in matter in Pennsylvania as evidenced by the fact that 90% of Black voters there voted for Obama. Because further discussion on this topic will put us outside of the central themes I want to cover in this post, I will save a further discussion on this for later.

Instead, I will move on to the issue of class, for I think that this is in fact the most important factor determining Obama's difficulties with voters outside of central urban areas. The best determinate of class in America is education. Exit poll data showing presidential preference by four levels of education were available for eight of the nine states (excluding Maryland) used in our previous analysis (see "Barack Obama's Achilles Heel"). The education levels available in the polling data included high school graduates, some college, college graduate and post-graduate. Note that "some college" would include college students that had not yet received their degree as well as those with 2-year degrees and those who never completed their studies. Note that for some states data on voters with less than a high school degree were available, but for most states sample sizes for such voters were too small for the results to be included in the released data.

I used the exit poll data to estimate the number of votes for each candidate for each level of education in each state and then combined the results across the eight states for each of the four groups. Because the results for college graduates and post-graduates were nearly the same, I combined these two groups into a single group. The figure below shows the voting share across the eight states for the resulting three education levels:

These shares are similar to those recorded in the 2004 general election suggesting that these results may be representative of the voting distribution in the upcoming general election. College grads represent the largest percentage of voters; voters with only a high school degree represent only a quarter of all voters.

Note that the two candidates split the total vote across these three education levels in the eight states nearly equally (Clinton 50.2% - Obama 49.8%) suggesting that this sample of states is likely to be reasonably representative of the nation as a whole. However, Obama recorded a comfortable 10.6 percentage point margin over Clinton among college graduates:

The candidates are nearly evenly split among those with at least some college:

Given that some in this group are current college students that will in time finish their degree and that this subset of the group appear to strongly support Obama, it is likely that Clinton has a modest but real advantage among those with some college but who will never receive a four-year degree.

Her advantage is decisive among that quarter of the electorate with only a high school diploma:

Clinton's margin among those voters with only a high school diploma and no college experience is nearly 19 percentage points. This group of voters represents the heart of the American working class and rank among the most economically vulnerable in American society today. Their incomes are stagnant or falling in real terms, their job security is being slowly eroded and they feel increasingly marginalized. Barack Obama is clearly not speaking to these people.

Instead, he speaks to the secure, the well educated, he speaks to those for whom the word "expectation" is more relevant than the word "hope". Just take a look at this video clip of his concession speech in Evansville, Indiana the night of the Pennsylvania primary. Notice the three young men at the top of the screen. Each is wearing Abercrombie and Fitch t-shirts, the uniform of, as a (self?) described "retail anthropologist" sez:

They [Abercrombie and Fitch consumers] are part of the young, hip, privileged culture rather than aspiring to it.

It doesn't seem that the less educated, less hip fit in here. In attempting to appeal to white voters, Obama's handlers put behind him a symbol of the haves on the night he was soundly defeated in the Pennsylvania primary. This is his problem in a nutshell.