Here's a fascinating study. A pair of researchers at the University of Georgia took a look at what happened to prescriptions for opioid painkillers in states that passed medical marijuana laws. Over at the Washington Post, Christopher Ingraham summarizes their results:

They found that, in the 17 states with a medical-marijuana law in place by 2013, prescriptions for painkillers and other classes of drugs fell sharply compared with states that did not have a medical-marijuana law. The drops were quite significant: In medical-marijuana states, the average doctor prescribed 265 fewer doses of antidepressants each year, 486 fewer doses of seizure medication, 541 fewer anti-nausea doses and 562 fewer doses of anti-anxiety medication.

Needless to say, the painkiller industry would much rather have you gulp down their addictive and lucrative product. They are not taking the threat from medical marijuana lying down:

The tanking numbers for painkiller prescriptions in medical marijuana states are likely to cause some concern among pharmaceutical companies. These companies have long been at the forefront of opposition to marijuana reform, funding research by anti-pot academics and funneling dollars to groups, such as the Community Anti-Drug Coalitions of America, that oppose marijuana legalization.

Pharmaceutical companies have also lobbied federal agencies directly to prevent the liberalization of marijuana laws. In one case, recently uncovered by the office of Sen. Kirsten Gillibrand (D-N.Y.), the Department of Health and Human Services recommended that naturally derived THC, the main psychoactive component of marijuana, be moved from Schedule 1 to Schedule 3 of the Controlled Substances Act — a less restrictive category that would acknowledge the drug's medical use and make it easier to research and prescribe. Several months after HHS submitted its recommendation, at least one drug company that manufactures a synthetic version of THC — which would presumably have to compete with any natural derivatives — wrote to the Drug Enforcement Administration to express opposition to rescheduling natural THC, citing "the abuse potential in terms of the need to grow and cultivate substantial crops of marijuana in the United States."

The study estimates that if all 50 states legalized medical marijuana, Medicare would save $500 million per year in painkiller spending. It's hard to extrapolate that to overall spending on painkiller medication, but the total savings would be on the order of $2 billion per year—maybe more.

Hillary Clinton has had a couple of poor showings in the polls recently, but I mostly shrugged them off. It was inevitable that she'd take a hit from the conclusion of the FBI email probe, but those kinds of things are almost always temporary. And it's only July, anyway. Polls won't start to mean too much until the middle of August.

That's just my two cents, but Greg Sargent reports that it's pretty much the opinion of the pros too:

I spent some time talking to senior Democrats today, and the basic feeling among them is this: Yes, it’s very possible Clinton did take a real hit from the FBI news. But if so, they see this as more of a temporary dip than anything else. They see the polling right now as mostly useless, since we will know a lot more about the race once both candidates choose their vice presidential running mates and the conventions take place later this month.

....One senior Democrat with access to a lot of private polling tells me that some surveys in states and districts where Clinton should be leading are showing her tied or slightly behind. But this senior Dem thinks the data probably reflects a momentary dip due to bad coverage of the FBI mess....Top Dem pollster Mark Mellman, for instance, conceded that Clinton may have taken a real hit. But he noted that the current polls, if anything, still show her up after a very tough stretch, leading into a period that could prove more favorable to her.

The fact that a man like Donald Trump is even within shouting distance of becoming president is reason enough to be nervous. But small blips in the polls don't really add anything to that. If you're the jittery type, stay away from the poll madness until next month.

I can't believe I missed this, but I did:

During two separate discussions of Black Lives Matters protests on Tuesday, Donald Trump claimed that people have called for moments of silence for Micah Johnson, the gunman who killed five police officers in Dallas and injured nine others, without specifying who or where.

On an O’Reilly Factor segment....“I saw what they’ve said about police at various marches and rallies,” said Trump. “I’ve seen moments of silence called for for this horrible human being who shot the policemen.”

Trump repeated the claim Tuesday night, saying at a rally in Indiana, “The other night you had 11 cities potentially in a blow-up stage. Marches all over the United States—and tough marches. Anger. Hatred. Hatred! Started by a maniac! And some people ask for a moment of silence for him. For the killer!”

Josh Marshall:

This isn't getting a lot of attention. But it should....There is no evidence this ever happened. Searches of the web and social media showed no evidence. Even Trump's campaign co-chair said today that he can't come up with any evidence that it happened.

....A would-be strong man, an authoritarian personality, isn't just against disorder and violence. They need disorder and violence. That is their raison d'etre, it is the problem that they are purportedly there to solve. The point bears repeating: authoritarian figures require violence and disorder. Look at the language. "11 cities potentially in a blow up stage" ... "Anger. Hatred. Hatred! Started by a maniac!" ... "And some people ask for a moment of silence for him. For the killer."

Trump's explicit race baiting has been so normalized by now that we hardly notice this stuff. This kind of talk from a major-party candidate for president should be front-page news everywhere. Instead, it warrants a few words in various campaign roundups.

Blacks, Hispanics, Muslims, foreigners of all stripes: they're all grist for Trump's crusade to convince white voters that they're surrounded by rapists, murderers, terrorists, and assorted other predators who want to take their jobs away and impoverish them. It's his whole campaign.

This is loathsome. For years it's been clear that the Republican Party could only win by turning out an ever greater share of the white vote. But by 2012 they seemed to have done everything they possibly could: Fox News stoked the xenophobia, Republican legislatures passed voter ID laws, and outreach to white evangelicals had reached saturation levels. What more did they have on their plate? Now we know the answer: nominate a guy who doesn't play around with dog whistles anymore. Instead he comes out and flatly runs as the candidate of white America, overtly attacking every minority group he can think of. That shouldn't work. In the year 2016, it should alienate at least as many white voters as it captures. But so far it seems to be doing at least moderately well.

President Obama was right yesterday: America is not nearly as divided as the media makes it seem. But the only way for Donald Trump to win is to make it seem otherwise. That's what he's been doing for the past year, and the media has been playing along the whole time, exaggerating existing grievances where they can and inventing them where they can't.

I'm not scared that America is such a hotbed of racial resentment that it's about to implode. But I'm increasingly scared that Donald Trump can make it seem that way, and that the press—always in search of a dramatic narrative—will go off in search of ways to leverage this into more eyeballs, more clicks, and more paid subscriptions. There's still time for us all to decide we should handle this differently. But that time is running out.

James Pethokoukis is unhappy with the Democratic Party platform:

Other than more government R&D spending, it hasn’t much to say about entrepreneur-focused growth and innovation policy....Indeed, the very first section of the platform is titled: “Raise Incomes and Restore Economic Security for the Middle Class.” And the first item in that first section: a call for a $15 minimum wage (despite plenty of skepticism from center-left economists). The next section: “Create Good-Paying Jobs.” And the first item there: infrastructure spending.

Popular policies, I’m sure. But again, boosting productivity and innovation seem secondary even though they are key to rising living standards. At its core, this is a platform about broad-based wealth redistribution, not broad-based wealth creation.

Fair enough, I suppose—though supporting research, science, and technology seems like a pretty important part of the entrepreneurial agenda. The Democratic platform also supports stronger antitrust enforcement, which is good for entrepreneurs, as well as a promise to make it easier to start up small businesses. And infrastructure is important for entrepreneurs too, so let's not disparage that. Also: guaranteed access to health care.

The Republican platform is still something of a mystery, but in 2012 its economic growth plan mostly focused on lowering taxes; passing a Balanced Budget Amendment; getting back on the gold standard; and letting the unfettered free market handle home mortgages (!). If we actually did all this stuff, it would probably crush economic growth for decades. Even with all its faults, anyone who cares about entrepreneurs really ought to prefer the Democratic Party platform to the Republican mess.

That said, let's get down to brass tacks: The reason that neither platform has a lot to say about productivity growth is that no one really knows how to boost productivity growth. Democrats can pretend that a $15 minimum wage will do it, and Republicans can pretend that tax cuts for the rich will do it, but this is just random burbling. Technological innovation is the key to productivity growth—though even that requires some caveats—and nobody really has a good model of how to spark lots of technological innovation. Hell, we barely even agree about what technological innovation is. Does Facebook count? A new headphone jack on the iPhone 7? Tesla cars with autopilot? Pokémon Go?

As soon as we agree on the best way to spur technological innovation, then I'll expect our politicians to support it. Until then, I think we're expecting too much of our party platforms if we want them to solve problems that nobody yet knows how to solve.

I got some flak yesterday for posting a chart showing the number of police shootings of unarmed blacks and whites without taking into account crime rates. It's harder to do that than you'd think, since the Department of Justice no longer produces crime statistics by race, but we can get a rough idea.

First, we'll use the National Crime Victimization Survey. This is a telephone survey and isn't based on arrest rates, so it doesn't have problems of possible police bias in who they decide to arrest. Second, we'll use the 2008 numbers for violent crimes (Tables 40 and 46 here), since that's the last year we have data by race. Obviously this isn't ideal, but I don't imagine that violent crime rates by race have changed dramatically since then. We'll compare this to the number of police shootings collected by the Washington Post for 2015.

The result is on the right. Don't take it too seriously, but it probably provides a decent rough idea of the disparity in police shootings of unarmed civilians when you account for crime rates. Unarmed whites are shot about 15 times per million violent offenses. Unarmed blacks are shot about 28 times per million violent offenses.

POSTSCRIPT: Needless to say, I'm under no illusion that this will stop the flak. I'm sure plenty of people on both sides of the debate have plenty of adjustments they want to make to the raw data in order to make their own side look better.

Exciting news today, folks:

That's right: Donald Trump actually told the truth about something. It's front-page news this morning, along with six other Trump-related front-page articles at the Washington Post. The New York Times has three. Politico has twelve. Sigh.

In light of today's sweeping decision at the Hague denying China's claim to various islands and reefs in the South China Sea, this is an interesting tidbit from the Financial Times:

US President Barack Obama in March delivered a stark admonition to Xi Jinping over the South China Sea, warning the Chinese leader of serious consequences if China reclaimed land at Scarborough Shoal, one of the most dangerous flashpoints in Asia.

....Following the meeting in Washington, China withdrew its ships from the area....“The signalling from the US side was that this was serious,” said a former official. “There was an accumulation of pieces ... the conclusion was that the People’s Liberation Army was advocating [action]. It wasn’t necessarily indicators that Xi himself had made any decisions, but there was the feeling that it was on his desk and coming to him for a decision.”

....China has come under criticism for building man-made islands in recent years, but the US saw Scarborough as more strategically significant given its proximity to the coast of the Philippines, which has a mutual defence treaty with the US. Some officials worried that China could install radar and missiles on Scarborough. Along with facilities in the Paracel and Spratly Islands, that would help China create a strategic triangle, which would enable the policing of any air defence identification zone in the South China Sea.

At the moment, China reclaims land at various spots in the South China Sea, and everyone complains but nobody does anything about it. Likewise, we operate reconnaissance flights and perform Freedom of Navigation exercises, and China complains but doesn't do anything about it. Basically, both sides can do whatever they want because neither side wants to start a war over it. This pretty obviously favors China at the moment, since they have the resources for large-scale reclamation projects and just enough of a navy to protect them. We have a considerably bigger navy, but it's unlikely the American public would show much support for a shooting war with China to protect a rock out in the middle of nowhere. All China really has to do is wait a while for us to get bored, and then keep on building.

A Conservative Case for Black Lives Matter

Jonah Goldberg has some sensible things to say about both Black Lives Matter and the killing of five police officers in Dallas:

At least for a moment, antagonists on either side of polarizing issues could see beyond the epistemic horizon of their most comfortable talking points. Black Lives Matter activists thanked the police for their protection and sacrifice. Conservative Republicans, most notably Speaker Paul Ryan and former Speaker Newt Gingrich, spoke movingly about race in America. Gun rights activists were dismayed that Philando Castille, the man shot by a police officer in Minneapolis, had followed all of the rules — he had a gun permit, cooperated with the officer, etc. — and was still killed. Liberals who insist that rhetoric from their political opponents inspires violence were forced to consider whether rhetoric from their allies might have helped inspire the shooter in Dallas.

....Former New York City Mayor Rudy Giuliani (who did not lose his lazy certainty) spent the weekend attacking the Black Lives Matter movement as “racist.” He wants people to focus on the fact that most black murder victims die at the hands of other blacks. That’s true, and tragic, and fairly irrelevant.

Conservatives, of all people, should understand that misdeeds committed by agents of the state are categorically different from the same acts committed by normal citizens. A father who slaps his son for no good reason, however wrong that may be, is very different from a cop who slaps a citizen for no good reason.

I'm continually nonplussed by the apparent inability of so many people to believe two things at the same time. Thing 1: Most police officers are conscientious public servants who perform dangerous jobs admirably and honorably. They're my first call if I'm ever in trouble. Thing 2: They're also human beings just like the rest of us, and fall prey to the same racial stereotyping that most of us do—but with guns in their hands. It's hardly surprising that black activists are finally demanding better treatment from police in their communities. The only surprising thing is that it took so long.

Two things. Both true. And not so hard to believe at the same time.

It's Official: Bernie Endorses Hillary

At the moment, I don't have anything on tap to say about this that I haven't already said, but just for the record:

Democrats took a long-anticipated step toward unity Tuesday as Sen. Bernie Sanders endorsed Hillary Clinton, praising the party platform and its presumptive nominee here in the state that gave him his most powerful primary victory.

“This campaign is about the needs of the American people and addressing the very serious crises that we face. And there is no doubt in my mind that, as we head into November, Hillary Clinton is far and away the best candidate to do that,” Sanders said to the cheers of hundreds of partisans crowded into a high school gymnasium here, Clinton nodding at his side.

The next step, obviously, is to see how enthusiastic Bernie is and whether he can bring along the vast majority of his former supporters. My guesses are "very" and "yes."

While Hillary Clinton and Bernie Sanders are finishing up their lovefest in New Hampshire, how about another chart comparing retirement in America to retirement in other rich countries? Don't lie: you know you want it.

Here's one that makes the US look a little worse than yesterday's charts. It shows income replacement rates in all OECD countries: that is, the amount of income that retirees get compared to how much they earned when they were working. This is a very tricky number to compute because there are lots of different ways you can do it, and the OECD pension report spends several pages just outlining the various assumptions you can use. That said, here are their estimates for low-income workers when you include only public pensions like Social Security:

And here's the same chart including all pension income, both public and private:

Once again, the United States relies more on private spending than most other countries. When you count only Social Security, the US looks pretty stingy, ranking seventh from the bottom for low-income workers. But when you include all sources of pension income, we look better than a majority of OECD counties—including famously generous ones like Belgium, France, and Germany.

Now, as I said, these numbers are tricky to compute, and in this case I'm fairly skeptical of them. The replacement rate for Social Security in the top chart looks way too low to me. Conversely, the replacement rate for all sources of pension income in the US looks too high. Neither number matches up to figures for low-income workers from places like the Congressional Budget Office and the Social Security Administration. But that may simply be because the OECD made different assumptions in their calculations.

In any case, this gives you a decent idea of how we stack up using simple income replacement rates as your metric.