Monthly Archives: April 2019

Are tropical cyclones increasing in number, duration or intensity?

By Kent R. Kroeger (Source:; April 24, 2019)

It is OK to be skeptical when it comes to End-Is-Nigh, Chicken Little alarmism over climate change. Yes, the earth is warming due mostly to human activities. And, yes, the sooner we transition to clean energy the better off we will be going forward.

But while I defer to the climatologists on the science of global warming and its probable consequences, my persistent belief is that humankind can and will accommodate, remediate, and even profit from climate change in the long-run. The upward sweeping arc of human history supports my confidence.

Furthermore, I will be long dead and unable to recalibrate my predictive model should my optimism prove wrong — which is also true of the pessimistic forecasts made by climatologists. Should their long-range predictions be substantively wrong that global warming will lead to major cities underwaterwidespread faminemass extinctions, and increased human warfare, they won’t be around to apologize.

And the scientists know that — which may explain their often extravagant claims about humankind’s future should we fail to address climate change.

And such is the dynamic when climate science interacts with the mainstream media and the political opportunists who offer us drippy, moralizing scolds such as: Your grandchildren will suffer the consequences of your inaction todayWhat did you do to fight climate change grandma/grandpa?

It’s cheap, exploitative rhetoric. And doesn’t change a single opinion. It merely forces everyone into their respective ideological corners.

But that doesn’t mean the underlying premise is wrong. Something isdifferent in the global climate system and tropical cyclones (we call them hurricanes in the U.S.) appear to be at the center of this change.

While climate change deniers ignore data, skeptics require it

The earth’s continuously produces new climate data and with every new data point, we potentially learn something.

But, as we are often told by climate scientists and their dutiful propagandists, don’t confuse weather with climate. Which is, of course, something they allow themselves to do without restraint.

So was the case, in March, when Tropical Cyclone Idai, a strong but not unprecedented Category 3 storm (minimum center pressure 940 mbar; highest sustained winds 120 mph), caused around $1 billion in damage in Mozambique, Tanzania and Malawi and killed over 1,000 people, mostly from flooding.

On cue, like a Broadway dance number, international bureaucrats, climatologists, and the news media declared this natural disaster belongs on the climate change damage ledger.

NPR used the term “climate change refugees” to describe the 2 million people displaced by Cyclone Idai’s floods.

This is going to happen more often, reinforced the scientists.

“There is absolutely no doubt that when there is a tropical cyclone like this (Idai), then because of climate change the rainfall intensities are higher,” according to Oxford University’s Dr. Friederike Otto. “And also because of sea-level rise, the resulting flooding is more intense than it would be without human-induced climate change.”

“The interesting thing for the area (Southern Indian Ocean) is that the frequency of tropical cyclones has decreased ever so slightly over the last 70 years,” according to the South African climatologist Dr. Jennifer Fitchett. “Instead, we are getting a much higher frequency of high-intensity storms.”

Others focused on the demography of Cyclone Idai’s most defenseless victims.

“As is always the case, the poor and vulnerable are the first to suffer and the worst hit,” said UN Secretary General António Guterres, as he spoke about Cyclone Idai at the World Meteorological Organization (WMO ) summit in New York in March.

But should everyone be trimming their sails when relating the devastation from a single tropical cyclone, such as Cyclone Idai, to climate change?

As with any single weather event, determining the marginal impact of global warming is problematic and not particularly helpful. But, when looking over long periods of time, a clear trend is emerging in the data for tropical cyclones: They aren’t more frequent, but on average they are becoming more powerful (and wetter).

And, most ominously, the average strength of tropical cyclones by the end of this century will be significantly stronger than today.

The Data and Methods

The data used in this essay were provided by the National Oceanic and Atmospheric Administration (NOAA) and is part of a worldwide database it maintains called the International Best Track Archive for Climate Stewardship (IBTrACS). This database can be accessed directly at:

The specific IBTrACS variables used for the analyses here were:

SID: A unique storm identifier assigned by IBTrACS
SEASON: Season (year) that the storm began
BASIN: Basin of the current storm position
NAME: Name of system given by source (if available)
ISO_TIME: Time of the observation in ISO format (YYYY-MM-DD hh:mm:ss)
_WIND: Wind speed units (in knots)
_PRES: Minimum central pressure (millibars)

Only storms whose maximum wind speed during its lifetime was at least 40 knots (miles per nautical mile) were included in this analysis. Duplicate SIDs were eliminated by assigning those storms to the basin of origin.

The Results: Tropical Cyclones are getting stronger

While Cyclone Idai was not unusually strong for a South Indian Ocean tropical cyclone, it was one of nine “intense tropical cyclones” in the South-West Indian Ocean this season (2019) — a region that typically experiences around three intense storms each season.

Using Cyclone Idai as the dividing line, Figure 1 shows the percentage of all South Indian Ocean tropical cyclones by decade whose minimum central pressure (millibars) over its lifetime was lower than Idai’s.

Figure 1 suggests tropical cyclones in the South Indian Ocean have been getting stronger on average since 1960. In the 1960s, only 5 percent of tropical cyclones in the South Indian Ocean exceeded Cyclone Idai’s strength. Between 2010 and 2019, 27 percent of tropical cyclones in the South Indian Ocean exceeded Cyclone Idai’s strength.

The biggest decade-to-decade increase was between the 1970s and 1980s (+13.2%). Since the 1980s, however, the percentage of Idai-strength storms (or higher) has been relatively stable with no clear trend. Though beyond the capacity of this analysis to answer the question, it is fair to wonder if improvements and frequency of aircraft and satellite measurements of tropical cyclone location, wind speed, and intensity played a role in the change in average storm intensity between the 1970s and 1980s.

Figure 1: Tropical Cyclone Strengths in Comparison to Cyclone Idai (1960 to present, South Indian Ocean only)


Citing research by Dr. Christopher Velden (et al. 2006), according to NOAA, “Satellites have experienced an increasing capability to probe and understand tropical cyclone environments and structure. Meteorological satellite observations began in the 1960s with merely identifying the systems from space. Researchers then developed a technique to estimate intensity from the storm cloud structure and lifetime. However, the early observations at visible and infrared wavelengths were limited in that they could only observe cloud tops. Routine microwave imager satellites began in the late 1980s and became integrated into forecasting in the 1990s.”

Using only the South Indian Ocean tropical cyclone data, it is hard to rule out the observed increase in storm intensities since the 1960s has been due to measurement improvements.

Thankfully, we have NOAA’s IBTrACS worldwide database of all measured storms since 1960s.

Figure 2 shows the number of worldwide tropical cyclone days, as measured by the annual number of tropical storms multiplied by the average durationof those storms. Again, the increase in tropical cyclone days from 1960 to the mid-1990s may be due to an improved measurement regimen; but, across the entire time series (1960 to 2018), there is no discernible trend. This finding is consistent with other research on trends in the number or duration of tropical cyclones (e.g., Dr. Jennifer Fitchett’s work).

Figure 2: Tropical Cyclone Days (1960 to present, Worldwide)


The data on tropical cyclone intensity is more illustrative. For example, the average tropical cyclone in the period from 1960 to 1980 had a minimum central pressure (MCP) around 969 mb (see Figure 3 below). From 2000 to 2018, however, the average central pressure has varied around 964 mb. That may seem like a small difference, but that is the difference between winds in the 74–95 mph range (Category 1) and a 96–110 mph range (Category 2). Small changes matter in tropical storm strength.

Is that 970 mb to 963 mb trend real, or also a product of variance in measurement accuracy and frequency over time?

Figure 3: Average Minimum Central Pressure per Tropical Cyclone (1960 to present, Worldwide)


Measurement quality variance appears, in the case of storm intensity, to be swamped by an underlying trend towards stronger storms over time. How do we know this? Because selection of the analysis’ starting year does not impact the negative central pressure trend (i.e., increasing storm intensity) until the starting point is in the 1990s (which is well after satellite-based intensity measurements had matured).

Unlike tropical cyclone frequency, there appears to be a distinct, secular trend in storm intensities. They are getting, on average, stronger.

How much stronger?

A simple linear model explaining average MCP reveals a significant negative trend since 1960 (see Figure 4). The average MCP declines .092 mb every year. The estimated model, for example, predicts an average MCP of 964.6 in 2019, dropping to 957.1 by 2100.

Figure 4: Linear Model of Average Minimum Central Pressure per Tropical Cyclone (1960 to present, Worldwide)


A predicted drop of 7.5 mb between now and the century’s end may seem small, but it is not. It is the difference between an average storm being a relatively manageable Category 1 storm (74–95 mph range) and a more dangerous Category 3 storm (111–129 mph range).

Let it sink in. The AVERAGE storm will be a Category 3 in 2100.

For a more tangible comparison, it is the difference between Hurricane Isaac (2012) and Hurricane Belle (1976).

Hurricane Isaac, a Category 1 storm (965 mbar; highest sustained winds 80 mph) hit Haiti, Cuba, and Louisiana/Mississippi, causing $3.1 billion (2012 USD) in damage and killing 34 people. Though a relatively weak tropical cyclone it made landfall multiple times (Haiti, Cuba, and the U.S. Gulf Coast), doing extensive damage in Haiti and Cuba.

Hurricane Belle, in contrast, was a stronger Category 3 storm (957 mbar; highest sustained winds 120 mph) but caused significantly less damage ($100 million in 1976 USD) than Isaac as its eye didn’t make landfall until Jones Beach, Long Island (New York) as a much weaker Category 1 storm.

A better example might be Hurricane Fran (1996) which hit Cape Fear, North Carolina as a Category 3 storm (120 mph sustained winds) and caused over $5 billion (1996 USD) in damage and killed 22 people.

It is scary to think children alive today will see over their lifetime the average tropical cyclone go from a Category 1 to a Category 3. While I do not recommend doing this, a strong adult could stand outside through the brunt of a Category 1 storm (assuming they are not in an area prone to flooding or storm surge). That same person would not be able to do that in Category 3 storm. A Category 3 tropical cyclone is a serious meteorological phenomenon and a grave danger to anyone in its path.

Though storm surges associated with tropical cyclones are highly variable and dependent on numerous factors besides storm intensity, a Category 3 storm can easily experience storms surges around 20 feet. According to NOAA, in U.S. coastal areas a storm surge of 23 feet has the ability to inundate 67 percent of interstates, 57 percent of arterials, almost half of rail miles, 29 airports, and virtually all ports in the Gulf Coast area.

Category 3 tropical cyclones are a different beast and that is going to be the norm for our planet’s tropical cyclones by the end of the century.

Luckily, I won’t be around to experience this new reality or apologize should my prediction prove wrong.

  • K.R.K.


Data, computer code, and modeling equations available upon request to


Velden, C., B. Harper, F. Wells, J.L. Beven, R. Zehr, T. Olander, M. Mayfield, C.“. Guard, M. Lander, R. Edson, L. Avila, A. Burton, M. Turk, A. Kikuchi, A. Christian, P. Caroff, and P. McCrone, 2006: The Dvorak Tropical Cyclone Intensity Estimation Technique: A Satellite-Based Method that Has Endured for over 30 Years. Bull. Amer. Meteor. Soc., 87 , 1195–1210,

The Assange Case is Complicated

By Kent R. Kroeger (Source:; April 13, 2019)

The mainstream media and the U.S. government tell us the indictment of Julian Assange is merely a case of prosecuting a man for aiding and abetting a felonious act and has little to do with press freedoms. Civil libertarians, in contrast, say the U.S. Attorney’s Office action is a clear-cut attack on the constitutional right of journalists to investigate and publish information that exposes government wrongdoing.

The reality is messier than either side want to believe.

Since the website’s inception over a decade ago, I have been uneasy about the WikiLeaks journalism model: Publishing large quantities of private (usually stolen) information under the rationale that it brings necessary transparency to important institutions in our society.

Assange sits in a British jail cell today because the U.S. government wants to prosecute him for aiding and abetting Chelsea Manning, a former intelligence analyst in the U.S. Army, in removing Iraq and Afghanistan war logs and diplomatic cables from a classified U.S. Dept. of Defense computer network (SIPRNet) in 2010. Manning, as well, sits in a Virginia jail cell today for refusing to testify before a grand jury regarding the Assange case.

The critical legal question (as of now) facing Assange is not whether he (WikiLeaks) had the right to publish the classified information that embarrassed the U.S. government in its revealing a significantly larger number of civilian deaths in Iraq and Afghanistan than had been previously disclosed by the U.S. military, and also exposing sensitive communications within the U.S. and foreign diplomatic communities. The Trump Justice Department is emphasizing that they are not challenging WikiLeaks’ right to publish, but merely prosecuting Assange for his encouraging and materially helping Manning commit an illegal act.

Those cheering Arrange’s arrest have long argued that he (and Manning) did substantive damage to U.S. security by exposing the identities of Iraqi and Afghan informants working with the U.S. military and by exposing confidential and critical conversations within the diplomatic community.

“There is blood on their hands” is a common refrain in the national security establishment.

As for the actual Assange indictment, it is a little less dramatic. The conspiracy indictment issued by the U.S. Attorney’s Office (Eastern District of Virginia) alleges that:

“Assange engaged in a conspiracy with Chelsea Manning…to assist Manning in cracking a password stored on U.S. Department of Defense computers connected to the Secret Internet Protocol Network (SIPRNet), a U.S. government network used for classified documents and communications. Manning, who had access to the computers in connection with her duties as an intelligence analyst, was using the computers to download classified records to transmit to WikiLeaks. Cracking the password would have allowed Manning to log on to the computers under a username that did not belong to her. Such a deceptive measure would have made it more difficult for investigators to determine the source of the illegal disclosures…

…During the conspiracy, Manning and Assange engaged in real-time discussions regarding Manning’s transmission of classified records to Assange. The discussions also reflect Assange actively encouraging Manning to provide more information.”

But First Amendment defenders are eager to point out that what Assange is accused of doing is done by investigation journalists every day in the execution of their jobs.

Never shy about exposing the hypocrisy of the mainstream news media, journalist Glenn Greenwald points out that news organizations routinely assist in the discovery, delivery and concealment of stolen goods and information.

“The New York Times implemented a system to allow stolen materials to be delivered to it without the thief getting caught,” tweeted Greenwald to New York Time journalist Katie Benner, who defends Assange’s arrest.

“The New York Times actively aids stealing,” says Greenwald.

Furthermore, investigative journalists are not passive actors in bringing transparency to private and public institutions in the name of the public good. They are proactive participants in seeking, discovering and publishing such information. They don’t sit at their desks waiting for whistle blowers to deliver ‘stolen’ information that exposes potential wrongdoing — they identify, pursue, and talk to potential sources with access to such information.

Washington Post reporters Bob Woodward and Carl Bernstein entered journalism lore because they weren’t afraid to prod, beg, and even shame potential sources into obtaining damning information about the Nixon administration.

But they still didn’t steal the information themselves — which is where the Assange case gets murky. To what extent are journalists are protected through the First Amendment to provide the incentives, tools and cover to help whistle blowers?

My no-law-background gut feeling says the U.S. Supreme Court will give tremendous latitude to journalists pursing the public interest. There is something about that Constitution of ours that swamps the temporary requirements of partisan bias and political convenience.

As for damage to U.S. security, the evidence against Manning and WikiLeaks is sparse. A 2011 DoD report concluded “with high confidence that disclosure of the Iraq data set will have no direct personal impact on current and former US leadership in Iraq.”

Former U.S. Defense Secretary Robert Gates gave reporters in 2010 a similar assessment:

“The fact is governments deal with the United States because its in their interest, not because they like us, not because they trust us and not because they believe we can keep secrets,” he said. “Some governments deal with us because they fear us, some because they respect us, most because they need us. We are still essentially …. the indispensable nation.”

“So other nations will continue to deal with us. They will continue to work with us. We will continue to share sensitive information with one another. Is this embarrassing? Yes. Is it awkward? Yes. Consequences for U.S. foreign policy? I think fairly modest.”

More probable harm was done to intelligence informants cultivated through the U.S. military and intelligence efforts in Iraq and Afghanistan.

Despite efforts by WikiLeaks to expunge identifying information regarding intelligence sources — for example, initially about 15,000 of the 92,000 documents obtained from Manning were not released in order to protect individuals cooperating with the U.S. military — Assange as subsequently asserted however that WikiLeaks is “not obligated to protect other people’s sources…unless it is from unjust retribution.”

And, in practice, it is almost impossible to remove enough information to protect all informants, particularly considering Manning lifted more than 740,000 pages of classified information from military servers.

“The lives of cooperating Afghans, Iraqis, and other foreign interlocutors have been placed at increased risk,” read a June 2011 Defense Intelligence Agency(DIA) executive summary regrading the impact of the Manning leaks.

I cannot categorically condone everything Assange and Manning did in releasing classified information — even as some of that information revealed likely war crimes committed by my government and its agents. It was likely a war crime when U.S. military personnel knowingly killed medical personnel (and two Reuters journalists) responding to a U.S. military attack on an adversarial target. [You can learn more about this attack and the accompanying video here.]

Without Manning and Assange’s actions, we may never have found about this now documented attack — and others like it. In my pedestrian view, learning of that attack is worth risking the lives of U.S. informants, who, after all, knowingly entered into that relationship cognizant of the risks involved.

Thus, at the end of the day, I stand with the Obama Dept. of Justice decision not to prosecute Assange on the belief that doing so would do significant harm to press freedom.

Nothing factual has changed since that Obama administration decision.

Furthermore, since WikiLeaks released the Manning materials, the U.S. security state has only grown bigger and more capable at invading the privacy of Americans and other persons protected by our Constitution. We must consider proportionality when deciding to prosecute and jail threats to a security state that far outweighs those threats in terms of power and ability to hide their illegal actions.

Adding to my inclination to defend Assange is that in over 10 years of WikLeaks’ existence, there is no concrete evidence suggesting it has ever published documents that were not authentic and accurate. CNN can’t seem to go 10 days without retracting a false story.

But the issue remains, to what extent is WikiLeaks responsible for protecting innocent people referenced in materials it publishes and to what extent is Assange, personally, responsible for aiding and abetting the theft of government property?

For Julian Assange’s own sake and reputation, he needs to have his case argued in the American court system, where I believe extenuating circumstances will lead to his acquittal.

But I don’t know that for a fact, and nor does anybody else.

  • K.R.K.

Please send all comments, questions and federal indictments to:

Trump’s Implicit Anti-Semitism

By Kent R. Kroeger (Source:; April 9, 2019)

President Donald Trump was in prime free-styling mode while speaking to the Republican Jewish Coalition last week in Las Vegas:

“I stood with Prime Minister Netanyahu…Benjamin Netanyahu…How’s the race going by the way? How is it? Whose going to win the race? Tell me. I don’t know. Well, its gonna be close. I think its gonna be close. Two good people. Two good people. But I stood with your prime minister (emphasis mine) at the White House to recognize Israeli sovereignty over the GOLAN! (emphasis Trump’s) Heights. The Golan Heights is something I’ve been hearing about for a long time…(extended applause)…The Golan Heights. So, I was talking to Ambassador (David) Freidman and — not about this — they’ve been trying to get that approved, as you know, for 52 years ’cause they’ve wanted recognition from…” (Trump’s stream of consciousness continued on for another 30 minutes)

The speech quickly went viral, with Trump’s critics focusing on the anti-Semitic trope of dual loyalty implied by his saying “your prime minister.”

“Mr. President, words matter,” tweeted Anti-Defamation League CEO Jonathan Greenblatt. “As with all elected officials, its critical for you to avoid language that leads people to believe Jews aren’t loyal Americans.”

Trump knew he was talking to Americans. After all, he was in Las Vegas. How could he not know? The only rational conclusion is that Trump thought his audience’s loyalties to Israel equaled, if not eclipsed, other loyalties.

Rep. Omar’s Controversial Statements on Israel

Even with a generous interpretation, Trump’s slur was patently more offensive than Minnesota Rep. Ilhan Omar’s suggestion in a February town hall meetingthat, after she had criticized Israeli policies towards Palestinians, she had been pressured from both sides of the political spectrum to confirm her loyalty to Israel.

“I should not be expected to have allegiance/pledge support to a foreign country in order to serve my country in Congress,” responded Omar to criticisms of her town hall comments. “I have not said anything about the loyalty of others, but spoke about the loyalty expected of me.”

[Personally, I found Omar’s comment about U.S. support for Israel being all ‘about the Benjamins’ to be more offensive and interpretable as anti-Semitic.]

But for some, Omar’s clarification was insufficient.

Chicago Mayor Rahm Emanuel said of Rep. Omar: “She is casting Jewish Americans as the other, suggesting a dual loyalty that calls our devotion to America into question.”

But Trump’s use of “your prime minister” is doing exactly that, directly. There is no nuance in Trump’s phrasing. He assumed his Las Vegas audience considered the political leader of another sovereign country to be their leader as well.

That is the definition of the dual loyalty slur. But was the deep-rooted anti-Semitism of Trump’s comment to the Republican Jewish Coalition ever acknowledged as such by the GOP and Israelis?

Of course not. After all, Trump has pretty much given Netanyahu everything he’s asked for, short of a war against Iran — and with two more years left in his term, that could still happen.

Despite Protestations from Party Elites, Israel is Now a Partisan Issue in U.S. Politics

Trump’s oratorical word salads have never been funny and, as president, the misinterpretations they invite only raise the possibility that he could do real damage to our national security and interests. Such solecisms in the context of an Israeli-Palestinian conflict that has killed 9,876 Palestinians and 1,263Israelis since 2000, over 2,300 being children, are particularly reckless.

Making the situation even more unpredictable is the Israeli policy under Netanyahu’s leadership to aggressively support and reinforce Trump’s leadership in the region.

By any objective measure, Netanyahu and Trump are besties. But to what end for the Israelis? So far, just short-term political gains, such as U.S. recognition of the Golan Heights and the U.S. Embassy’s move from Tel Aviv to Jerusalem. Those are purely symbolic outcomes from Israel’s perspective. Before Trump entered office, the U.S. Embassy was already, functionally, working out of Jerusalem and Israel’s control of the Golan Heights has been secure since 1974. Even Palestinian protestations over these U.S. policy moves have been muted and perfunctory.

But with the introduction of partisan politics into the U.S.-Israeli equation in the past 10 years, the Israeli alliance with the Republicans, and now Trump, may be effecting immeasurable damage to Israel’s (and U.S.) long-term interests.

Recent events on the 2020 presidential campaign trail illustrate this potential.

Democratic presidential candidate Beto O’Rourke, not exactly a risk taker when it comes to stating concrete policy ideas, said during a campaign stop in Iowa recently that the U.S.-Israeli relationship, to be successful, “must transcend partisanship in the United States, and it must be able to transcend a prime minister who is racist, as he warns about Arabs coming to the polls, who wants to defy any prospect for peace as he threatens to annex the West Bank, and who has sided with a far-right, racist party in order to maintain his hold on power.”

Holy Moses!

For anyone following the U.S.-Israeli relationship over the years, O’Rourke’s statement was unprecedented for a mainstream politician with presidential hopes. American politicians don’t call Israeli prime ministers racist (not even strong Netanyahu critics such as former U.S. Secretary of State John Kerry or Vermont Senator Bernie Sanders). But, in the post-Obama world, the U.S.-Israeli relationship has fundamentally changed — even as U.S. foreign policy towards Israel is as supportive as it has ever been.

Should Israel annex large sections of the West Bank, as Netanyahu has promised should he form the next government after the April elections, the Two-State Solution, which is already on life-support, will be OBE (Overtaken-By-Events). There is no viable Palestinian state on the West Bank where Israel controls 60 percent of the land (see Figure 1). Even as most U.S. foreign policy experts still cling to the Two-State-Solution, on the ground, it is as dead as Jussie Smollet’s career.

Figure 1: Israeli Settlements on the West Bank

Source: United Nations OCHA

Trump’s Implicit Anti-Semitism Has Consequences

When Trump implies that American Jews are as loyal to Israel as they are to the U.S., he is not only saying something that is demonstrably untrue, he is reinforcing one of anti-Semitism’s bedrock falsehoods.

Ironically, he is also doing measurable harm to Israel’s support among American Jews.

“A growing number of American Jews look at Israel and see a country that is occupying Palestinian territory and breaking up peaceful Palestinian protests using force,” writes journalist Zack Beauchamp, a Jewish American with familial connections to Israel. “They also see a Jewish state that only recognizes one socially conservative strand of Jewry, Orthodox Judaism, as legitimate — which manifests in things like preventing liberal American Jews from praying in mixed-gender groups at the Western Wall, the holiest prayer site in Judaism.”

Trump’s politicization of the Israeli-Palestinian conflict has also impacted support for Israel across all Americans, and these opinions are dividing along party lines.

In a University of Maryland Critical Issues Poll, conducted in September and October of 2018 among a nationally representative sample of 2,352 Americans, a majority of Republicans (57%) indicated they want U.S. policy to lean in favor of Israel over the Palestinians, while a substantial majority of Democrats (82%) want it to lean toward neither side, and 8 percent want it to lean toward the Palestinians.

As Trump said many times on the 2016 campaign trail, “I will do more for Israel than any president in history,” after two years, he believes he has made good on that promise. But he also done great harm, not just to Palestinians who saw a significant increase in fatalities by Israeli forces in 2018, but to Israel itself by overtly linking U.S. policy towards Israel to his political fortunes.

Given Trump’s demonstrable propensity for perpetuating anti-Semitic tropes, that is a political marriage the Israelis may well regret.

  • K.R.K.

Comments can be sent to:

Don’t stereotype Trump’s conservative base


This is the fifth essay in a series dedicated to analyzing the U.S. eligible-voter population using the 2018 American National Election Study (ANES), an online survey administered in Dec. 2018 by researchers from the Univ. of Michigan & Stanford Univ. 


By Kent R. Kroeger (Source:; April 5, 2019)

Nothing makes a Democrat happier than describing the ‘average’ conservative Republican:

“The Republican base is now made up of religious and neoconservative ideologues, and the uneducated white underclass with a token person of color or two up front on TV to obscure the all-white, all-reactionary, all-backward, ‘there-is-no-global-warming’ rube reality. Actual conservatives, let alone the educated classes, have long since fled,” says author and filmmaker Frank Schaeffer, a conservative Christian in his youth who became a liberal Democrat as an adult. “…the Republican Party — as it is now — must be utterly destroyed in 2020. The word euthanized comes to mind.”

Sean McElwee, a researcher for, offers a more balanced, less provocative description of the GOP’s conservative base:

“While non-college whites do indeed make up a significant portion of (Donald) Trump’s base, they also make up a non-trivial share of (Hillary) Clinton voters…
…If we define the base as a group making up a non-trivial share of the electorate that overwhelmingly prefers one party, it is fair to call white evangelicals Trump’s base. If we define the base purely by the size of the coalition, we might prefer instead white non-college voters or whites over 50, both of whom make up more than half of Trump’s voters.”

And New York Magazine writer Eric Levitz takes a variant approach by summarizing what motivates this conservative base — but the implication about who is being motivated is crystal clear:

“For decades now, the conservative movement has sought to keep its core voters confined to a carefully curated media ecosystem — one where the Democratic Party is a MarxistIslamist organization, America is the world’s most over-taxed nation, illegal immigrants bear sole responsibility for the stagnation of middle-class wages (and/or all violent crime), and there’s never been a better time to buy gold coins.”

Schaeffer and Levitz’ descriptions wouldn’t be funny if they weren’t somewhat accurate.

These are stereotypes about Donald Trump’s conservative base — rooted in reality — but deeply misleading at the same time.

Summarizing 62 million Trump voters is a Sisyphean task even for statisticians and strategic consultants, much less for the average political journalist, which is why stereotyping is so seductive, as it is often based on kernels of truth.

Using the attitudinal segmentation detailed in my previous essays, I computed within each political segment the percentage who were white and did not have a 4-year college degree. As seen in Figure 1, the majority of the conservative base (also referred to as ‘Trump’s conservative base’ in this essay) fit this demographic description (63%). In comparison, only 31 percent of the Democratic progressive base shares those attributes.

On the surface, it appears fine to categorize — or, rather, stereotype — conservatives as less-educated, white people.

Figure 1: Prevalence of Whites without a 4-year College Degree by Political Segment

Data Source: 2018 American National Election Study (Pilot); Analytics & Segmentation by Kent R. Kroeger

But this characterization is not precise.

Anytime we use stereotypes as a heuristic device to save analytic time and energy, we risk mischaracterizing the world we’re trying to explain.

That is the case with many who try to describe Trump’s conservative base.

Stereotyping is lazy analytics

Stereotypes are cognitive short-cuts that help simplify our lives. We engage in stereotyping because it often works across a wide of range of daily activities.

When we choose a new restaurant, we judge it by location, cuisine, and price. That is stereotyping.

When we seek partners, we consider diction, education level, religious background, and even someone’s accent. That is stereotyping.

Stereotyping saves us time and effort. Psychologists James Hilton and William von Hippel define stereotypes as “mental representations of real differences between groups. . . allowing easier and more efficient processing of information.”

But stereotyping has a risk. It can mischaracterize the true nature of a group and its relationship to other groups. Trump’s almost monolithic white, less-educated voter base is not a myth, but its not the whole story either. And its certainly not the interesting story.

As we will see below, there is significant attitudinal diversity within the media-maligned conservative base, who could represent the tipping point in a Democrat landslide in 2020, or the foundation of a new, durable GOP electoral majority.

Yes, racism and sexism punctuated the 2016 election, but…

As the U.S. has become more educated and ethnically diverse, Republican voters have become more reflective of an earlier era in America, as they are now predominately white, Christian, and less likely — from even a decade ago — to be college graduates, according to a 2018 Pew Research study. Add in a growing partisan gender gap and it doesn’t take a PhD in demography or political science to realize this is voter base is problematic for the Republicans going forward.

Still, the formula worked for the Republicans in 2016.

Education levels among whites strongly correlated with their 2016 vote choice. The divide in vote preferences between highly-educated and lesser-educated whites grew dramatically during the 2016 campaign, according to a team of researchers led by University of Massachusetts–Amherst political scientist Brian Schaffner, and this gap is not explained by economics.

But this divide was not a reflection of greater economic stress among lesser-educated whites. According to Schaffner’s team, Trump’s vote was driven by sexism and racial denialism, not economics. “Explicit racist and sexist appeals appeared to cost Trump some votes from more educated whites, but it may have won him even more support among whites with less education.”

The UM-Amherst study, however, leaves a dark impression about the average Trump voter; even though, the analytic focus of the study had little to do with describing the average Trump voter. Its concern was in describing relationships between populations through marginal and conditional probabilities — which is not the same as describing a population such as Trump’s conservative base. Those are two different analytic tasks.

There is little doubt that Trump’s explicit appeals to nativism along with his own misogynistic behavior sharpened the sorting process among many voters still available to either Clinton or Trump in the last months of the 2016 election. But those ‘persuadable’ voters are not the same as the conservative base.

At a rally in Minnesota last summer, Donald Trump invited his audience to offer its opinion on the “fake news” media. Photo: Scott Olson/Getty Images

Compassion and Acceptance Defines the Largest Conservative Bloc

Don’t let the sounds and images from Trump campaign rallies (or the rambling, ignominious rants of Trump himself) define the entire conservative base. The bloc of voters that most helped elect Trump cannot be summarized with stereotyping labels such as ‘racist,’ ‘sexist,’ ‘homophobic,’ or ‘anti-immigrant.’ I understand the temptation, but resist.

Employing a two-stage attitudinal segmentation of 2018 ANES respondents, in the first stage, using attitudes related to current policy issues (immigration, trade, health care, economic inequality, gun control and climate change) and, in the second stage, using attitudes on race, ethnicity, sexual orientation, and other social organizations, three sub-segments are identified within the GOP’s conservative base (see Figure 2):

Old Guard Conservatives (33% of Conservatives; 9% of all eligible voters).

Lost Hope Conservatives (27% of Conservatives; 7% of all eligible voters).

Compassionate Conservatives (40% of Conservatives; 11% of all eligible voters).

Figure 2: The Sub-Segments within the Conservative and Progressive Political Segments

Data Source: 2018 American National Election Study (Pilot); Analytics & Segmentation by Kent R. Kroeger

Compassionate Conservatives are the largest sub-segment of conservatives (40%) and distinguish themselves from both the Lost Hope and Old Guard sub-segments with respect to their favorability towards other races, ethnicities, and social groups (i.e., group affinities). On average, Compassionate Conservatives rate whites roughly the same as they do Blacks and Hispanics (in contrast to the other conservative sub-segments who rate whites significantly higher than those groups). Second, Compassionate Conservatives demonstrate greater affinities towards Gay and Lesbian, Transgender, Muslim and immigrant Americans relative to Lost Hope and Old Guard sub-segments. They are, for lack of a better description, more open-minded than other conservatives and represent the largest number of conservatives.

Given the relatively small sample size of conservatives (effective n = 331) and progressives (effective n = 358), most of the demographic differences between the sub-segments are statistically insignificant. However, there are some notable exceptions:

  • Compassionate Conservatives are the wealthiest sub-segment
  • Lost Hope Conservatives are the least wealthy sub-segment
  • Old Guard Conservatives are the oldest and least Female sub-segment

While more detail on the attitudinal and demographic differences between the conservative (and progressive) sub-segments can be found in the Appendix below, the following graphs highlight some of the more prominent attitudinal differences.

Figure 3 maps the conservative and progressive sub-segments based on their average ratings of whites and Blacks. The diagonal line shows where both racial groups are rated the same (equity). Four of the six groups are relatively close to this line (and statistically indistinguishable from it), while Justice Progressives, on average, rate Blacks higher than whites and Lost Hope Conservatives do the opposite.

Figure 3: Ratings of Blacks and Whites (Conservative and Progressive Sub-Segments)

Data Source: 2018 American National Election Study (Pilot); Analytics & Segmentation by Kent R. Kroeger

On these racial attitudes, Compassionate Conservatives (the orange circle) have more in common with Establishment Progressives (the dark blue circle)than they do with the two other conservative sub-segments.

On attitudes regarding immigration, Compassionate Conservatives also distinguish themselves among the conservatives. While they generally agree that immigration numbers into the U.S. should decrease (horizontal axis in Figure 4), they depart from their political brethren on whether immigration and diversity hurts American society (vertical axis in Figure 4). Only 20 percent of Compassionate Conservatives believe immigration and diversity hurts society, while 46 percent of Old Guard Conservatives and 57 percent of Lost Hope Conservatives believe it does. That is a significant and dramatic difference within the conservative base.

Figure 4: Attitudes on Immigration and Diversity (Conservative and Progressive Sub-Segments)

Data Source: 2018 American National Election Study (Pilot); Analytics & Segmentation by Kent R. Kroeger

Figures 5 through 7 summarize the sub-segments (conservative and progressive) on their ratings of Gays and Lesbians, Transgender, and Muslims. As seen in this series of bar charts, the distributions of the identity group ratings by the Compassionate Conservatives resemble the progressive sub-segments far more than the two other conservative segments. In particular, the distributions for the Compassionate Conservatives closely match those for the Paycheck Progressives, a sub-segment of progressives that skews male and less wealthy.

Figure 5: Ratings of Gays and Lesbians (Conservative and Progressive Sub-Segments)

Data Source: 2018 National Election Study (Pilot); Analytics and Segmentation by Kent R. Kroeger


Figure 6: Ratings of Transgender (Conservative and Progressive Sub-Segments)

Data Source: 2018 National Election Study (Pilot); Analytics and Segmentation by Kent R. Kroeger


Figure 7: Ratings of Muslims (Conservative and Progressive Sub-Segments)

Data Source: 2018 National Election Study (Pilot); Analytics and Segmentation by Kent R. Kroeger

If LGBTQ rights and other identity-based issues were major vote drivers for conservatives, 40 percent of the GOP/Trump base might listen to Democratic messages on such topics during the current election cycle. Then again, the GOP might just as logically target Paycheck Progressives with uniquely tailored messages.

If there is one constant in American politics, every national election is potentially competitive. The popular vote difference between the two parties in U.S. House elections has exceeded 10 percentage points just six times since 1946 (out of 37 elections).

Neither the Republicans or Democrats enter an election without hope or possibilities. Whether a party leverages their strategic and tactical strengths enough to fulfill those hopes is another matter.

Are Compassionate Conservatives the GOP’s “Breakaway Province” or their beacon to the future?

A Democratic-aligned pollster once described Reagan Democrats, a working-class, once reliable voting segment of the Democratic Party coalition, as the Democratic Party’s ‘breakaway province.’ The defection of this ‘province’ to the Republicans in 1980 ushered in Ronald Reagan’s charismatic leadership and a conservative ideology that would dominate the American political landscape until the 2008 financial crisis.

Though some political scientists are skeptical that realignments actually happen — David Mayhew once writing, “Electoral politics is to an important degree just one thing after another… and their underlying causes are not usefully sortable into generation-long spans” — some political observers believe the 2016 Trump election may be the commencement of the next great realignment.

Washington Post columnist David Ignatius recently opined: “Political systems can be like scientific theories. Sometimes there emerge so many anomalous elements that don’t fit the existing structure that the theory collapses, and a new one arises… We may be entering such a period. The definition of a winning Democrat may be that, in response to Trump’s rambling circus of self-aggrandizement, he or she could create a genuinely coherent new political order.”

One possible target of Ignatius’ new political order could be Compassionate Conservatives whose attitudes resemble many in the Democrats’ progressive base, the presumption being that the three factions in the progressive left will hold together as long as Trump is in office.

The other common assumption rising from observations like Ignatius’ is that a centrist candidate could be very attractive to Republicans and Democrats disillusioned by the ideological extremism of their parties — and that an ideologue from the left (like Bernie Sanders) or right could not build such a coalition as easily as a centrist candidate.

But the political world doesn’t operate under this naive voting model where voters line up their opinions relative to those of political candidates and make their vote choices accordingly. Centrists and non-ideological voters, in particular, are generally unable or disinclined to make decisions this way.

Political scientists realized long ago that using spatial models of voter attitudes to predict vote choice aren’t terribly informative. To the contrary, two realities describe American voting behavior: (1) Vote choice often drives voters’ opinions (not the other way around), and (2) there is little electoral penalty for political extremism.

The result is that idealistic and folksy conceptions of our pluralistic democracy as one where voters ‘throw the bums out’ when their policies fail is dangerously wrong.

Political scientists Christopher H. Achen and Larry M. Bartels, in their book, Democracy for Realists: Why Elections Do Not Produce Responsive Government, conclude that “abandoning the folk theory of democracy is a prerequisite to both greater intellectual clarity and real political change. Too many democratic reformers have squandered their energy on misguided or quixotic ideas.”

In explaining why Bernie Sanders can attract voters that don’t necessarily agree with his policies, Achen and Bartels write:

“Decades of social-scientific evidence show that voting behavior is primarily a product of inherited partisan loyalties, social identities and symbolic attachments. Over time, engaged citizens may construct policy preferences and ideologies that rationalize their choices, but those issues are seldom fundamental. That is one key reason contemporary American politics is so polarized: The electoral penalty for candidates taking extreme positions is quite modest because voters in the political center do not reliably support the candidates closest to them on the issues.”

The point is simply that we cannot assume a candidate’s views are a direct reflection of his/her supporters’ views. And we see that phenomenon in the 2018 ANES data where Centrists are often low-information voters who support candidates not always close to their own personal opinions.

One reason I don’t show the Centrist political segments when mapping opinions and attitudes (such as in this essay) is that their political behavior is either too unpredictable or non-existent (i.e., non-voters). The more ideologically cohesive political segments are far more interesting.

Which is why Compassionate Conservatives (on the right) and Paycheck Progressives (on the left) are so intriguing and potentially more impactful in the 2020 election and for the futures of their respective parties.

Contrary to media narratives, the potent political division within the Democratic Party is not between Centrists and Progressives but between Establishment Progressives (think Kamala Harris and Beto O’Rourke) and Justice Progressives (think Bernie Sanders and Elizabeth Warren). But even those two political segments agree far more than they disagree. On health care reform, for example, the distance between Sanders’ Medicare-for-Allplan (a single public option for all) and the Democratic establishment’s Medicare-for-America (offering a public option to go with the current private health care system) is small when compared to what the Republicans offer on health care reform — which is mostly nothing.

As for the GOP’s conservative base, the media narrative on its composition is not much more than stereotyping and it is generally wrong. Representing two-fifths of conservatives, Compassionate Conservatives mirror the Democrats more than other conservatives on racial, gender, and cultural attitudes. And not only are they the largest conservative segment, they are wealthier and younger than the party’s Old Guard, promising that the power of Compassionate Conservatives within the GOP is more likely to grow than shrink proceeding into the future.

The battle for the soul of the Republican Party will be fought between the three sub-segments of conservatives described in this essay. The older, more politically active Old Guard against two relatively young sub-segments, themselves divided by education, economic status, and levels of open-mindedness.

The conservative base is not a white, uneducated monolith of racists, sexists, and homophobes (…don’t get me wrong, these conservatives do exist). But, attitudinally, the conservative base is remarkably diverse. If it weren’t, the GOP would be dead in the water heading into 2020.

  • K.R.K.

For datasets and statistical codes, send requests to:




A Two-Stage Segmentation Based on Policy Attitudes and Identity Politics

This essay uses an two-stage attitudinal segmentation of U.S. eligible voters as its data source to identify the GOP’s conservative base (measured in Dec. 2018).

In the segmentation’s first step, all 2,500 ANES respondents were clustered based on a series of attitudinal questions related specific social and political policies (e.g., trade, immigration, gun control, climate change, etc.) in order to create relatively homogeneous political segments. In other words, the people within each segment possess relatively similar attitudes on policy issues [More detail on these segments can be found in the previous essays, starting here.]

A five-cluster solution was selected based on fit characteristics and interpretability (Figure A.1). Progressives are the largest political segment today, accounting for 30 percent of the eligible voter population, followed by Conservatives (27%) and Independents (19%).

Figure A.1: The Political Segments in the U.S. eligible voter population

Data Source: 2018 National Election Study (Pilot); Analytics and Segmentation by Kent R. Kroeger

This essay is focused on the Conservatives, who were additionally segmented based on their racial attitudes and group affinities (i.e., orientation towards identity politics).

The second-stage segmentation identified three sub-segments within the Conservative political segment: Old GuardLost Hope, and Compassionate Conservatives. [My fourth essay discusses in detail the three Progressive sub-segments within the Democratic Party’s base.]

Compassionate Conservatives are the largest of the three Conservative sub-segments (40 percent of the Conservative segment and 11 percent of the total population) and, as we will see below, counter many of the stereotypes presented in the national media about the Trump base (Figure A.2).

Figure A.2: The Sub-Segments within the Conservative and Progressive Political Segments

Data Source: 2018 National Election Study (Pilot); Analytics and Segmentation by Kent R. Kroeger

Figure A.3 shows the average affinity scores (rating on a 0 to 100 scale) for each Conservative sub-segment. Where all three sub-segments agree is no surprise: They like Donald Trump, whites, rural Americans, and Supreme Court Justice Brett Kavanaugh; and they dislike Barack Obama, Hillary Clinton, the #MeToo movement, socialists, journalists, and Special Counsel Robert Mueller (though on the latter, opinions have probably changed in the last few days).

Figure A.3: Racial Attitudes and Group Affinity Questions Used for 2nd-Stage Segmentation

Data source: 2018 American National Election Study (Pilot); Segmentation by Kent Kroeger (; Color shading is determined within the column (i.e., political sub-segment)

However, significant differences in racial attitudes and group affinities also delineate the three sub-segments. Most axiomatic are the highly negative views Lost Hope Conservatives hold towards Blacks, Hispanics and immigrants. Lost Hope Conservatives, representing only 27 percent of the Conservative political segment, are consistent with mainstream stereotypes of the Trump base: They are anti-immigrant and racially-biased (and probably sexist too).

In counterpose, Compassionate Conservatives distinguish themselves in two ways from both the Lost Hope and Old Guard sub-segments. First, on average, they rate whites roughly the same as they do Blacks and Hispanics (in contrast to the other sub-segments who rate whites significantly higher than the other races and ethnicities). Second, Compassionate Conservativesdemonstrate greater affinities towards gays and lesbians, transgender people, Muslims and immigrants relative to Lost Hope and Old Guard sub-segments. They are, for lack of a better description, more compassionate and open to the world and its inherent differences.

Though not quite a majority of the Conservative base, Compassionate Conservatives probably represent a better conspectus of the “typical” Trump-voting conservative than either the Old Guard or Lost Hope sub-segments.

Demographics and Behaviors

Figure A.4 details the demographic and behavioral characteristics of the sub-segments. In this table, the color shading is determined by each row. For example, in the first row, Establishment Progressives have the highest incidence of females (60%) and the Old Guard Conservatives have the lowest (37%).

Figure A.4: Demographic and Behavioral Characteristics (Conservative and Progressive Sub-Segments)

The starkest differences between conservatives and progressives are in the incidence of Blacks, Hispanics, a college education, Trump voters (no surprise), and non-religiousness.

Within the conservative segment, Compassionate Conservatives stand out asbeing very white and having high incomes, while the Old Guard are very male, religious, and old. The Old Guard were also the most supportive of Trump in 2016 and had the highest vote turnout percentage of the six sub-segments. Lost Hope Conservatives stand out from other conservatives with the highest percentage of females (47%) and the lowest average incomes.


Additional Graphics

Figure A.5: Demographic and Behavioral Characteristics (Conservative and Progressive Sub-Segments)

Data Source: 2018 National Election Study (Pilot); Analytics and Segmentation by Kent R. Kroeger
Figure A.6: Ratings of the #MeToo Movement and the Transgender (Conservative and Progressive Sub-Segments)

Data Source: 2018 National Election Study (Pilot); Analytics and Segmentation by Kent R. Kroeger