The GOP Civil War on Taxes

Republicans love tax cuts, and both President Trump and Speaker Ryan have set their sights on lowering both personal and corporate income tax rates. But some Republicans also like controlling the budget deficit, while others favor defense spending or immigration control. How can the GOP cut tax rates, raise defense spending and immigration enforcement, and control the budget deficit? Here’s the heart of the problem: the federal government gets roughly $1.4T from income taxes, $440B from corporate income taxes and capital gains, $1.1T from payroll taxes, and smaller amounts from other sources [1]. The tax plan under consideration will substantially cut the first two sources, without raising the other categories. How can such a tax plan be implemented without blowing up the budget deficit?

The evolving Trump-Ryan plan bridges this gap by introducing a new category: a border adjustment tax on imports. If all 2.7T in US imports were taxed at 20%, this could raise over $500B per year, providing a source for big tax cuts (though still not enough to pay for the tax cuts proposed). But there’s a problem with this idea – will 50 Republican senators vote for it?

The National Retail Federation has come out strongly against the plan, as have the Koch brothers, whose companies participate heavily in international trade. The Kochs are focusing their battle charge in 15 states where they may be able to sway Senate votes. Meanwhile, with retail giant WalMart strongly opposed, will the senators from Wal-Mart… err Arkansas be on board?

Hence we have a GOP civil war, pitting major exporters like Boeing, Oracle, and GE against retailers and other importers, and pitting nationalist Republicans versus traditional free-trade Republicans.

Trump and Ryan can only spare two votes in the Senate – will they be able to keep everyone on board? While the plan could stimulate US growth through tax cuts and favoring US production, it may also trigger a trade war that nullifies much of its benefit. There’s also the essential nature of the import tax – it is effectively introducing a new US consumption tax for the first time. Consumption taxes have been on the GOP radar for some time, as they tend to shift tax burdens down the income scale, and to reduce taxes on the wealthy. But is Trump’s base ready to pay an extra 40 cents at the pump every day, when many of them won’t see a huge tax cut [2]? Let the Republican tax civil war begin.

[1] The CBO provides a detailed breakdown of revenues here. I have combined corporate taxes and capital gains into one category, as both are taxes on capital.

[2] Roughly 50% of oil is still imported into the US, so a border adjustment tax could disproportionately increase oil prices.

Election 2016 Review: Democrats Score an Own-Goal, Part II

2017 has rolled around, and with Mr. Trump entering the White House, I’d like to close the book on the 2016 election with a review of the certified election results. Shortly after the election, my analysis concluded that Clinton lost the election – Trump performed like a generic Republican, while Clinton underperformed John Kerry (2004) on the way to defeat. With all the votes in and Clinton showing a 3 million vote popular vote lead, is this analysis still valid?

  • Donald Trump received 62,979,636 votes, 27.2% of the 231,556,622 voting eligible people in the United States. This represents a drop of 1.1% in total vote share, relative to Mitt Romney, roughly in line with the recent decline trend seen for GOP candidates (Bush 2004 was the strongest GOP candidate in recent history).
  • Hillary Clinton received 65,844,610 votes, 28.4% of the 231,556,622 voting eligible people in the United States. This represents a drop of 2.2% in total vote share relative to Obama 2012, and was the worst performance by a Democratic candidate since Al Gore in 2000.
  • Given the narrowness of Trump’s victory in the upper Midwest, Clinton could be right that external forces like James Comey  or Wikileaks/Putin did her in. But she (or another Democratic candidate) could have overcome that had they turned out young and minority voters. Clinton didn’t need to perform at Obama-like levels – but she couldn’t hold on to the sliver she needed to win.

So in the final analysis, I still blame Democrats and the Clinton team in particular for this loss far more than I credit Donald Trump. According to the data, he performed in line with a typical GOP candidate, while Clinton greatly under-performed the last three Democratic campaigns. Other analysts concur, and have added depth by identifying exactly how many votes Dems left at home in the upper Midwest. The lesson for 2017 and beyond: Democrats – focus on turning out your base, instead of appealing to voters unlikely to vote for you in the first place.

Election 2016: Democrats Score an Own-Goal

Donald Trump won the 2016 presidential election, but he didn’t win the election. Hillary Clinton scored an own-goal – she turned out the lowest share of voters of any Democrat since 1996, and was bested (in vote share) by numerous losing Democrats of the past. See fine candidates of the past like Kerry, John, and Gore, Al. Contrary to much current analysis, Donald Trump didn’t really expand his base – he performed about as well as a generic Republican, while Clinton under-performed badly. The CNN chart below shows candidates’ share of eligible voters since 1996 [1]:

https://i1.wp.com/i.cdn.turner.com/cnn/.e/interactive/html5-video-media/2016/11/10/turnout_full_scale_all_voters_v2update.png

  • GOP vote share peaked in 2004 and has been bleeding down steadily, losing about 1.7% in total vote share per election since 2004 [2]. Trump actually under-performed this downtrend, as he dropped 2% (from 28.3 to 26.3) versus the average drop of 1.7% per cycle for Republicans. The GOP seems to turn their base voters out steadily, but this base forms a shrinking percentage of the electorate as the US becomes more diverse.
  • Democratic vote share has been more volatile, reaching a peak in 2008 and bleeding down after. But Clinton didn’t need to match Obama’s 2008 or 2012 performance to win – matching John Kerry or even Al Gore would have been sufficient! John Kerry’s performance, at 30% of eligible voters, would have crushed Trump, while even Al Gore’s 27.4% would have provided the small margin needed in the Midwest to win the election. In a low turnout election, Clinton bled off roughly 4% points of vote share, while Trump bled off only 2% points, leading to his narrow electoral college win.
  • Neither 3rd party votes [3] nor Trump’s numbers explain Clinton’s collapse – Clinton and the Democratic party have only themselves to blame for frittering away more than 100% of Obama’s net gains! Democrats tend to win when they increase overall turnout, and the data show clearly that turnout collapsed in the 2016 election.
  • Had Hillary Clinton managed to turn out any of the urban vote in the Midwest, or gotten millennials excited, she’d have been fine. She did neither, and in absolute terms had the worst performance for a Democrat since her husband, ironically (who won handily in 1996 in a low turnout election).

In short, I blame Democrats and the Clinton team in particular for this loss far more than I credit Donald Trump. According to the data, he likely performed just below a more typical GOP candidate, while Clinton greatly under-performed the last four Democratic campaigns.

Silver Linings For Democrats:

  • Arizona is rapidly trending blue – more rapidly than expected. Obama lost Arizona by around 213,000 votes in 2012, while Clinton lost by only 85,000 votes in 2016. This aligns with a 2014 study on demographic change, putting it right on track to lean blue in 2020.
  • Winning just Arizona’s and Florida’s (whose demographics are also shifting) 40 electoral votes replaces all the rust-belt states Clinton lost, creating a 272 electoral vote win.
  • Who voted in 2008 that didn’t turn out in 2016? Young people and minorities. Democrats win when turnout is high. Turnout is high when the most progressive parts of the Democratic base are excited. Therefore, Democrats’ direction should be clear – there is no one left in the center, and it’s time to move forward with the progessive wing of the party.
  • Had Democrats won in 2016, winning again (four in a row) would have proven even more impossible given the tick-tock nature of the Presidency in the US. Since FDR, one party has won three straight presidential contests only once, in 1988 – and none have won four straight.

[1] Votes are still being counted, and the final tally will likely show Clinton’s vote share edge close to 27%, still below Al Gore’s 2000 performance.

[2] Per the chart above, McCain lost 2.4% points from Bush 2004, and Romney lost another 0.8% points from McCain. Trump lost roughly 2% points from Romney – leading to an average decline rate of just over 1.7% per election cycle. Mr. Trump lost more than this amount, leading to the conclusion that he might have under-performed relative to an “average” GOP candidate.

[3] Above, we see that in 2016 voter turnout was around 55.4%, with 52.8% captured by the two mainstream parties. If all 2.6% lost to third parties came from Clinton voters (and it almost certainly dd not), that still implies that she under-performed Obama 2012 and Kerry 2004 by around a percentage point – and this assumption is far too generous to Clinton.

Why Today is not Brexit USA

Update: Well I get to eat crow on this one! While demographic shifts did aid Clinton, it looks like she may win the popular vote but lose the electoral college, as Al Gore did in 2000. But while I (and many) were surprised by this result, at HiddenLevers we made sure to prepare our audience for the possibilites by looking at the election through the lens of risk parity. Whatever you were hoping for in Election 2016, it helps to be prepared for all the contingencies! Watch our pre-election webinar and scenario forecasts here.

Original Post Below:

It’s become fashionable in the media to compare today’s election to Brexit, and to forecast that the result may play out similarly. Markets and the mainstream will be shocked when the silent majority rises up and defies the global elites! But is this analogy really apt for the US? What’s surprising is the lack of demographic comparisons – the simple fact is that the US looks nothing like the UK, and voter demographics matter even more in an election like 2016.

The UK’s population was 87% white at the 2011 census, and its voting population is whiter and older since many minorities are relatively recent immigrants. The US, by way of contrast, is 61% white, with the voting population estimated to be roughly 69% white in the 2016 election. A 26 percentage point difference between the US and UK mean that the two are worlds apart – the Brexit comparison simply falls down in this light. Mr. Trump may be losing minorities by as many as 50 percentage points, helping to explain the early voting results in places like Nevada. Demographics pose another challenge for Mr. Trump – his voters are on average already more likely than minority voters to turn out, and so the surge he may have whipped up could hurt him as much as it helps him.

While a Clinton victory today isn’t guaranteed, those making simplistic comparisons to Brexit are ignoring the world of difference between the population voting today, and the population in the UK that voted back in June.

What Can Be Done About Skyrocketing Drug Prices?

The American government and even major insurers actually have a lot of levers they can pull to lower drug costs – but do politicians, insurers, and employers have the courage to try?

The drumbeat of overpriced-drug stories has been continuous in America of late, from Martin Shkreli’s 5400% price hike last year, to the recent price hike and subsequent backpedaling of Mylan with respect to the EpiPen. With growing outrage over skyrocketing drug prices, it’s worth asking – what can be done about it? Drug pricing is not subject to typical market forces since a new drug often has exactly 0 direct competitors – enabling a drug company to set virtually any price. New cancer drugs often start list pricing at $300,000 per year, while groundbreaking new Hepatitis-C treatments like Gilead’s Sovaldi started out at $84,000 for a short term (curative) course of treatment. Insurance companies (and major employers) have been unwilling to say no, swallowing each hike and passing it on in higher premiums. Medicare, Medicaid, the VA, and other public entities have been banned from negotiating prices, leaving them powerless to get a better deal for those receiving care through their programs. Here are a few ideas on how to break the logjam, in order of increasing potential savings:

1. The Anti-Monopoly Approach

Making drugs, and in particular so-called small molecule drugs, is actually both inexpensive and easy. The primary protection that drug makers use to enforce their monopoly position on new drugs is the patent system. While this arguably makes sense for new drugs, what about long-generic drugs? In recent years certain drug companies (e.g. Valeant) began specializing in buying up the manufacturers of old drugs and immediately hiking prices. After gaining a monopoly position it became easy to hike prices by 50-100% per year and extract huge profits, while new entrants were stymied by
the FDA approval process required to certify the efficacy of their drug version. Why not streamline the FDA approval rules for generic drug manufacture? If a drug is tested and shown to be chemically identical, how much further testing is really necessary?

Alternately, the Department of Justice or FTC could bring suit to halt acquisitions which would leave zero competitors in the market for a generic drug. Special pharmacies called compounding pharmacies are also capable of making many drug compounds. Why not
allow compounding pharmacies to compete across all generic drugs, or specifically contract with them to make generic equivalents for the VA system or Medicaid system?

2. The “Title IX” Approach

Private American colleges and universities are not actually required by law to provide equity in women’s sports, or to follow any of a wide range of Department of Education edicts. The catch? In order to receive federal funding, institutions of higher education must comply with these rules. Since virtually all colleges make use of varying forms of federal assistance, they fall into line.

The American pharmaceutical industry does very little original research – most innovations originate in the university system, and most of the research funding (over $21B per year) comes from the National Institutes of Health [1]. The American government could utilize this lever to strongly influence drug pricing. Pharmaceutical companies might be required to adhere to certain pricing guidelines if they wished to license research originating from NIH funding.

Those guidelines might require drug makers to release drugs into the generic market on an accelerated timeline, for instance. Or the rules might require that drug makers adhere to a value-based pricing approach, as described further below. Drug makers could be required to pay a tiered tax on drug sales to fund NIH research – a tax of 25% on prices above $1000/patient/month and 50% above $4000/patient/month could simultaneously fund future research and encourage drug makers to keep pricing down. The advantage of the “Title IX” approach is that it preserves the liberty of drug companies – if they don’t want to conform with the rules, they can simply do their own basic research. Fiscal conservatives might find this approach palatable as it directly charges users (drug companies) for the government programs they use, and lowers the deficit in the process.

3. The Value-Based Approach

If insurers and government buyers (Medicare/Medicaid/VA) all insisted on paying for value, pharmaceutical companies might be compelled to go along. How do you define value? The UK’s NICE measures the efficacy of medical treatments by attempting to measure the number of “quality-adjusted life years” provided by that treatment. If a cancer drug postpones death by 2 years on average, and has mild side effects, then it can be said to provide 2 years of QALY. The NIH takes this a step further by quantifying how much it will pay per QALY (currently around 25,000 pounds per year), and it sets prices
on drugs using this approach.

American buyers could emulate this approach by offering to pay for measured improvements in outcomes. If a new cancer drug extends life by 2 years, but existing cancer drugs extend life by 1.5 years, then the value of the new drug is an additional half-year of life. Drug buyers could offer to pay a premium for the new drug based on this degree of improvement, and no more. Buyers could also use this as a way to foster competition between older and newer generations of drugs. The older drug is 75% as effective, so it can be placed into competition with the new drug, but at a discount. Express Scripts took this approach in the Hepatitis-C market and was among the first buyers to find a way to push back against Gilead’s $1000 per-pill asking price for Sovaldi.

Conclusion

As long insurers are happy to pass rising costs along in the form of higher premiums, and American politicians remain beholden to the pharmaceutical lobby, nothing will change. But the ideas outlined above show that America doesn’t need European style price controls to break the drug price spiral – a combination of relatively small policy changes and insurers’ willingness to negotiate are all that is required.

 

[1] This article investigates the breakdown of basic pharmaceutical research in detail, and concludes that big pharma companies contribute less than 25% of research dollars in the US, with most of the balance coming from the NIH.

Police account for 20% of Random Homicides in US

Recent statistics indicate that American police commit 1 in 5 homicides in which the victim is unknown to the assailant.

According to the DOJ, 14,200 Americans died as a result of homicide in 2013 (the most recent year with complete data). This represents a huge drop in the US murder rate, as it has fallen by half from 1994 to 2013, from 9 homicides per 100,000 Americans to 4.5 homicides per 100,000.

Of the 14,200 murders in 2013, we can estimate that roughly 4700 were random homicides, those committed by strangers unknown to the victim. FBI research has shown that almost 80% of murders involve some sort of acquaintance between the victim and assailant – we use a factor of 2/3 here to be more conservative [1]. These 4700 random homicides are of particular interest because they invoke the greatest fear in the public – the fears of home invasion and carjacking murders, of murder-robberies on the street, and so on.

In reality these crimes have become rarer as the overall crime rate in the US has dropped. But while the crime rate has dropped, news over the past year of police homicides has exploded into the public eye. This raises a question – what percentage of random homicides are committed by police? FiveThirtyEight.com has covered this issue and reports that police likely commit at least 1240 homicides per year in the Unites States today. Since police interact with different members of the public every day, virtually every one of these homicides counts as a random homicide – the individual officer responsible for a given homicide doesn’t generally know the victim. This implies that US police officers commit roughly 20% of random homicides, even when justifiable homicides are taken into account [2]!

If police are indeed committing 1 in 5 random homicides, that is deeply troubling. Perhaps it is justified by America’s overall murder rate? The Economist reports that German police killed 8 civilians last year (0 in the UK and Japan), versus 685 homicides in Germany. This leads to a ratio of 11.7 police homides per 1000 murders in Germany, versus 87 police homicides per 1000 murders in the US [3]. American police kill civilians at 8 times the rate of their German counterparts, even after adjusting for country-specific murder rates. American police work is also safer as an occupation than it’s ever been – but the evidence shows that American police have failed to adapt to safer conditions, and appear to now be part of the homicide problem in the United States.

[1] The Bureau of Justice Statistics Homicide Trends in the US Report indicates (on page 16) that the victim had some acquaintance with the assailant in 78.1% of cases. This metric excludes the 44% of cases where data was unavailable (often because the crime remains unsolved). If we conservatively assume that the victim was known to the assailant in 50% of these cases, we get a blended rate of roughly 2/3 – therefore only 1/3 of all murders are random, in which the victim and assailant had no acquaintance.

[2] Taking the 1240 police homicides and dividing by 4700 random homicides yields a rate of 26%. Police departments reported 461 justifiable homicides in 2013, which would lower the police share of random homicides to 17%. But some percentage of police homicides claimed as justifiable may not be, and the 1240 police homicides per year may itself be an underestimate. If we assume that over 2/3 of all reported justifiable homicides were in fact justified, then this reduces the police share of random homicides to 20%.

[3] German police killed 8 civilians last year, divided by 685 homicides = 0.0117. Multiplying by 1000 makes this 11.7 police homicides / 1000 murders. In the US we have 1240 police homicides divided by 14200 murders, or 0.087, which equals 87 police homicides / 1000 murders. This metric shows that US police kill at 8 times the rate of German police, even after adjusted for the overall country-specific violence rate.