Applying the 'Dow 36,000' Discount to Trump's Tax Cuts
Yes, it seems a little unfair to keep bringing up "Dow 36,000" at this point. The infamous book, co-authored in 1999 by Kevin Hassett, who took over last month as chairman of the White House Council of Economic Advisers, was wrong. So what! I've written lots of wrong things in the past, and I appreciate you not mentioning them in the comments every time I publish a column.
Still, I do try to learn from my mistakes. And reading through the CEA's report this week on "Corporate Tax Reform and Wages: Theory and Evidence," I could not shake the sense that Hassett might be doing the same thing he did 18 years ago. The basic approach is this:
- Take a reasonable idea with lots of academic research to back it up.
- Push it past the bounds of reason.
In their September 1999 Atlantic article "Dow 36,000," 1 which predated the book by a few weeks, Hassett and James K. Glassman started with the observation that the "risk premium" of stocks over Treasury bonds seemed to have declined a lot over the decades. The risk premium is, to use Glassman and Hassett's handy definition, "the number of percentage points you have to add to the Treasury-bond rate in order to make the total equal to the dividend yield for stocks plus the growth rate of dividends." The idea was that investors needed to be paid a premium to take the greater risk of investing in stocks.
But after academic researchers began to document this risk premium in the 1970s using historical market data, it began to disappear. 2 "My own view is that the risk premium has gone down over time basically because we've convinced people that it's there," University of Chicago finance professor Eugene Fama once told me.
Glassman and Hassett took that a step further, arguing that "stocks are actually less risky, in the aggregate and over the long term, than bonds" and that as investors figured this out, the equity risk premium would go to zero. From that, they calculated that the Dow Jones Industrial Average -- at about 11,000 in September 1999 -- would continue to rise "at least until Dow 36,000," after which a period of lower returns would follow.
Now, again, the notion that the risk premium had declined was not controversial. This was good reason to argue that, for example, economist Robert Shiller's cyclically adjusted price-earnings ratio (41.3 in September 1999, 31.2 now) was not necessarily destined to revert to its long-run (since 1881) average of 16.8. Was it really reason, though, to predict with seeming confidence that the Dow was about to more than triple?
No, probably not. As economist Paul Krugman wrote in a February 2000 column after a conversation with Hassett:
As I understand it, a casual, back-of-the-envelope calculation led to a huge publisher's advance so quickly that the authors never had a chance to have second thoughts.
The subsequent performance of the stock market did probably lead to second, third and fourth thoughts for both. By late 2002 the Dow had fallen to as low as 7,286. Glassman and Hassett hadn't said in the article or book exactly when the Dow would hit 36,000, but in 2010 they did concede a bet with an Atlantic reader that the average would be closer to 10,000 than 36,000 at the end of 2009 (it was 10,428). Since then of course the stock market has been doing pretty great, with the Dow now above 23,000. But if you adjust for inflation, that's only 15,645 in September 1999 terms -- still a long, long way from 36,000.
Which brings me to the new report from Hassett's CEA. The tax framework unveiled last month by the Donald Trump administration and Republican congressional leaders calls for a reduction in the top federal corporate tax rate to 20 percent from 35 percent, immediate expensing of all business investments, and a shift to a territorial system in which U.S. corporations aren't taxed on the profits their subsidiaries earn overseas. The CEA report draws on recent economic research showing that (1) a significant share of the corporate tax burden is borne by workers and (2) high tax rates (the U.S. statutory corporate rate is among the world's highest) cause corporations to shift profits to lower-tax countries to conclude that the proposed tax changes
would in the medium term boost average U.S. household income annually in current dollars by at least $4,000, conservatively. When we use the more optimistic estimates from the literature, wage boosts are over $9,000 for the average U.S. household.
That cutting corporate tax rates would probably boost incomes is a fair representation of current academic thinking on the subject (see this column from my Bloomberg View colleague Noah Smith for more on that). The dollar amounts, though, seem dodgy. As Harvard economist and former Treasury Secretary Larry Summers pointed out:
The cut in corporate tax rates from 35 percent to 20 percent would cost slightly less than $200 billion a year. There is a legitimate debate among economists about how much the cut would benefit capital and how much it would benefit labor. Hassett’s “conservative” claim that the cut would raise wages by $4,000 in an economy with 150 million workers is a claim that workers would benefit by $600 billion -- or 300 percent of the tax cut!
The CEA was actually talking not about workers but households, of which there are 125 million in the U.S. So it's a purported benefit of $500 billion, or 250 percent of the tax cut. Still really high, though! Harvard's Mihir Desai, an expert on corporate taxes whose research is cited repeatedly in the report, tweeted that the CEA seemed to have exaggerated the impact by about sixfold. And Desai is someone who favors a cut in corporate tax rates and a shift to territorial taxation.
Why might Hassett choose to overstate his case? Well, this time there's no big publisher's advance in the balance, but he is part of an administration in desperate need of a legislative victory. And here's the great part: Unlike with the "Dow 36,000" prediction, no one will ever really know if he was wrong. There are just too many other things happening in the economy to sort out with precision how big an effect a single corporate tax cut had. There is thus a strong incentive, and little risk, to going with the absolute best-case scenario, and maybe beyond it.
So here's what I propose: Let's subject the CEA forecast to a "Dow 36,000" discount. In 1999, Glassman and Hassett predicted that the Dow was about to rise 25,000 points. In fact, over 18 years it has risen about 4,500, in inflation-adjusted terms, or 18 percent of the forecast amount. I figure the inflation adjustment is only fair, given that the gain occurred over a period far longer than what Glassman and Hassett were talking about. Multiply the "conservative" $4,000 estimate and the "more optimistic" $9,000 by 18 percent, and you get $720 and $1,620. Multiply those by 125 million households and you get $90 billion and $202.5 billion, or between 45 percent and 101 percent of the tax-cut amount. That -- especially the former -- actually seems totally reasonable. I started writing this column as something of a lark, but now I think I may have arrived at a more reliable forecast than the Council of Economic Advisers.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
I have never read the book, but I have read the article numerous times, and I do not get the sense that the book includes a lot of important plot developments not contained in the article.
In the mid-1970s, Roger Ibbotson, a Ph.D. student and then a lecturer at Chicago's Graduate School of Business, and Rex Sinquefield, a recent Chicago MBA working at a local bank, had calculated the equity risk premium since 1926 and used it to forecast stock market returns going forward. The Dow Jones Industrial Average, in the 800s when they first made the forecast in May 1974, would hit 10,000 by November 1999. It got there in March 1999. Now THAT was a great forecast!
To contact the editor responsible for this story:
Brooke Sample at email@example.com