Saturday, April 11, 2015

How indexation killed growth


HOW INDEXATION KILLED GROWTH

By Carles Gave, Gavekal Dragonomics

Indexing, as I have written before, is a form of socialism, since capital is allocated not as it should be - according to its marginal return - but rather according to swings in the market capitalization of the underlying assets. It is hard to think of a more stupid way to allocate this scarce resource.

In this new world, the goal of every money manager is to achieve a performance as close as possible to the index against which he is benchmarked (see Indexation = Parasitism). As a consequence, the dispersion of results among money managers has become smaller and smaller over the years. Today you can even buy programs telling you how much IBM stock you should buy versus Johnson and Johnson in order to control your “tracking error”. As always in economics, there is what you see and what you don’t. What most people don’t see is how the spread of indexation has led to a collapse in the growth rate of the economy.

Building a portfolio is a very complex exercise. In a perfect world, one would start with the “expected” marginal increase in the return on invested capital of different investments. Once satisfied with a given position, one should try to ensure that the increase in the marginal return is not too correlated with other positions in the portfolio. The name of the game is to find assets with the same ROIC over the long run, but a negative correlation over the shorter term (for example, US shares versus. US government bonds over the last 20 years). 

This aims at the Holy Grail of money management, which is to achieve a decent long term return, together with a low volatility of that return. As one can see, this involves a massively complex price discovery exercise, starting with an examination of the marginal variations of ROIC, followed by consideration of the prices at which one can buy the available assets, and finally ending with portfolio construction.

In such a world, one would expect the distribution of performances to be very wide. Indeed, a large dispersion of performances should reassure us that capital has been properly allocated. After all, not everybody can win the jackpot.

Alas, today’s world is not perfect, and this is not how capital is allocated. Instead capital is allocated according to the market capitalization of the assets under consideration. So nowadays, capital is directed to an investment if it outperforms. In simple terms, this means that capital is channeled to companies enjoying an increase, not in their ROIC, but in their share prices. In a world in which investments are made according to the marginal ROIC (i.e. the past), these two tended to overlap. As a result, indexation worked, but only as long as no more than about 5% of assets were managed by “free-riding” indexers. 

Not in today’s world. Today indexing has become the dominant asset management style, and investments are dictated by market cap and changes in market cap; which is simply another way of saying that capital is now deployed according to momentum-based rules. This was very visible in 1999-2000, and is almost as visible today.

Intellectually, the old method of investment was based on a “return to the mean” approach. When the price movements of an asset became excessive compared to its expected ROIC, then one bought - or sold - the asset. Today, capital is allocated only according to marginal variations in the price of the asset. The more it goes up, the more money managers invest in it. The more it goes down, the less managers own. 

A return to the mean methodology leads naturally to a stable, but moving, equilibrium. Momentum-based investing inevitably creates an explosive-implosive system, which swings wildly from booms to busts and back again. And if monetary policy is as silly as it has been since 2002, these swings will be even more pronounced.

The closer we get to a bust, the tighter the performance dispersion among money managers, as the poor fellows trying to manage efficiently and professionally lose their clients to benchmark optimization algorithms. I don’t have the necessary data, so cannot prove it, but I would not be surprised if a sharp fall in the dispersion of money managers’ results is a reliable warning that a bust is approaching.

The goal of every socialist experiment is for everybody to earn the same salary. In the world of money management, we seem to have achieved this remarkable ambition. Hurrah!

Of course, if everybody gets the same results, then no one is going to get fired for underperforming, which is great news for the people administering the capital (I hesitate to call them managers). But—and here is what we do not see—our capital is being massively misallocated, all the time. 

People ask me why we have no economic growth. Why on earth do they expect economic growth in a socialist system?

Friday, April 10, 2015

Ivory: Curtis Macnguyen Is a Former Hedge Fund Star. And That Is Not Acceptable

For fun, Curtis Macnguyen likes to run along the seafloor in 15 feet of water, carrying a boulder.
He does it offshore from his house on the blue-watered Kona Coast of Hawaii’s Big Island, Bloomberg Markets reports in its May 2015 issue, not far from similar spreads owned by Michael Dell and buyout kingpin George Roberts.
Macnguyen, 46, is a hedge fund manager. An old-school hedge fund manager. His methods are what you’d expect from a guy who carries rocks underwater: lots of hard work, almost no big sprints, and steady progress, all under pressure.
Macnguyen brings his A game to everything, and has since childhood, when he bet on Trivial Pursuit and Pictionary with his six brothers. He has a tennis court in his backyard in the Brentwood neighborhood of Los Angeles, where he hits against local pros. Mats Wilander, winner of eight Grand Slam titles, came for a few sets in March.
Macnguyen’s golf handicap is 6, and he never plays without a little money on the line. A wager, even a small one, makes him step up his game. He recently lost $100 to Matt Kuchar, ranked No. 14 in the Official World Golf Ranking. He just built a gym—Tuscan style, like his house—so he could do a workout recommended by Marcus Elliott, a doctor and biomechanist who trains the best players in the National Basketball Association.
“Every guy who’s really successful at anything wakes up and says, ‘How can I do this better?’” Macnguyen says. (He’s Vietnamese, but his name sounds Irish. It’s pronounced like McWin.) He’s sitting on a covered patio at his house on a bright November day sipping iced tea made by a private company he’s backing. For an hour before, he hammered table tennis shots at a former national champion from South Korea (and Playboy model) named Soo Yeon Lee. He loves pingpong because it gets him in the zone—that rarefied state of total concentration where time vanishes and maximum ability emerges. “You’re out of your body, watching yourself,” he says, “and you can get any ball that comes in.”
Macnguyen is every bit as intense at Ivory Investment Management, the $3.5 billion hedge fund firm he founded in November 1998. Through the end of 2014, his Ivory Flagship Fund returned 346 percent, or 9.7 percent a year. That’s twice the 139 percent delivered by the Standard & Poor’s 500 Index.
Curtis Macnguyen says table tennis, more than any other sport, gets him in a zone of pure concentration.
Benjamin Rasmussen/Bloomberg Markets
Even more impressive: Because Mac­nguyen hedges his bets by balancing long and short positions, he had, on average, about one-fifth of investors’ money exposed to potential losses. The S&P 500 is 100 percent exposed, by definition. When the market fell during those years, he either made money or lost very little. And when it rose, his portfolio often rose much more.
Macnguyen’s methods have made him rich. He likes Hawaii so much that he and a group of investors bought 873 acres (353 hectares) of oceanfront land that they’re developing. He gets there from Los Angeles on his very own Gulfstream G450.
But that’s not enough for Macnguyen. He could own every white-sand beach in the Pacific and not be content. Lately, he’s been pretty ticked off—with himself. A person could look at his track record and conclude that his best years are behind him. And that is just not acceptable.
In 1999, its first full year of operation, Ivory Flagship returned 28 percent, compared with 21 percent for the S&P 500. Even better, when the index fell 9 percent the next year, Macnguyen made 17 percent.
Investors who didn’t love him already should have swooned in 2008. The market plunged 37 percent that year, and Ivory Flagship fell just 7.6 percent. When the rebound came in 2009, Macnguyen was ready. Investors had been pestering him for a new fund that would take more risk, and he obliged with the Ivory Optimal Fund, now his largest. It jumped 28 percent that first year, compared with 26 percent for the S&P 500.
Then something changed. In 2010, Ivory Flagship lagged the index by 13 percentage points. In 2011, he lost 3.6 percent in Flagship while the market rose 2.1 percent. Ivory Optimal did worse. His mojo was missing in action.
“I never kicked a dog or smashed a computer or even yelled at anyone,” Macnguyen says. “I was just frustrated and pissed off at having to keep explaining to investors that the environment was tough for our strategy. I’ve always felt that we’re in a no-excuse business. Just like high-level competitive sports, no matter how tough the conditions are, it shouldn’t matter, because you just have to be better than your competitors.”
Macnguyen, like most hedge fund managers, lives pretty high up in psychologist Abraham Maslow’s hierarchy of needs. Food, water, sleep, sex? Check. Security, employment, health? Yep. Friendship, family, intimacy? Check, again. (He’s married and has a stepson, 21, and a son, 7.)
What began to elude him, it seems, is the next level: self-esteem, confidence, and perceived respect from others. He grumbles about rivals getting more attention—and money to manage—despite inferior performance. He singles out a doppelgänger (at least on paper): a former investment banker from an Ivy League school who’s also 46: David Einhorn, who runs Greenlight Capital. They worked together at a small hedge fund firm called Siegler, Collery & Co. in 1993. Since it started in 2009, Ivory Optimal has returned 113.5 percent, edging Greenlight’s 112 percent for the same period. Yet Greenlight manages $12 billion, almost four times what Ivory does, and that bothers Macnguyen. “The only difference between me and Einhorn is that he’s higher profile and I’m purposely very low profile,” Mac­nguyen says. “Plus, I’m supercompetitive and will only get better over time.”
In conversation, Macnguyen toggles between bravado, like that, and self-flagellation. He frets about recent years when he failed to do the single thing hedge fund managers get paid for: generating alpha. Alpha is profit that doesn’t come from the whole market going up or down. Anyone can get lucky and make big money by taking a big risk. Alpha is different. It’s return you get beyond the risk you take. And hedge fund managers have to produce enough to cover their fees. Ivory’s are an industry-standard 2 percent of assets per month plus 20 percent of annual gains.
Bloomberg Markets
In 2010 and 2011, few of Macnguyen’s picks worked. He shorted Amazon.com, betting that earnings would tumble as the company invested heavily to keep revenue rising. Sure enough, earnings fell, but investors didn’t care, and the stock rose. He bought shares of Hospira, a maker of injectable drugs, when they slumped into the $40s from close to $60 in 2011. Then the company warned that regulatory issues had slowed production at a plant in Rocky Mount, North Carolina, and the stock dropped into the $20s. Macnguyen tried to hang on for the rebound he expected, but the losses mounted. “For every winner we had, we had a loser,” he says. “We didn’t add a lot of alpha for two years, and that is painful for a guy like me.”
It’s a familiar story. Managers of equity funds, once accustomed to beating the S&P 500, have, as a group, been thrashed in each of the past six years, according to an index of equity funds tracked by Hedge Fund Research in Chicago, and have bested the index in just three of the past 12.
Explanations—or excuses—abound: There’s more competition now as some 10,000 hedge funds look for stuff to buy with $3 trillion they’ve collected from pensions, endowments, and rich people; the U.S. Federal Reserve’s dovish interest rate policy won’t let the market fall; hedge funds do better in down years, and the U.S. stock market hasn’t had one in six years. Many of the great ones have given up. Jeff Vinik, who managed the famed mutual fund Fidelity Magellan from 1992 to 1996, shut his hedge fund, Vinik Asset Management, in 2013 and returned $6 billion to investors after ill-timed bets on stock indexes and gold-mining shares.
There may be more at work. Nobel laureate Daniel Kahneman wrote in his 2011 book, Thinking, Fast and Slow, that stock-picking managers exist because of an “illusion of skill” and add no value compared with passive—and cheaper—index investing.
“It’s very difficult to do what these people are trying to do,” says Matthew Litwin, head of manager research at Greycourt, a consultant to 100 wealthy families and institutions with a total of $9 billion to invest. (He puts his clients’ money in a variety of vehicles, including, occasionally, hedge funds. He declined to comment specifically on Ivory.) “Most people will fail.”
Macnguyen isn’t buying into any illusion of skill. He agrees that most managers will fail. He just doesn’t intend to be among them.
Macnguyen’s arc toward an absurdly high net worth is as unlikely as any American’s. He was born in Cam Ranh Bay in 1968, the same year the North Vietnamese army surprised the south with the Tet Offensive, taking the Vietnam War to a new, bloodier level. His family moved to Saigon, now Ho Chi Minh City, and his father served in the navy, then in the South Vietnamese congress. They planned to stay, even as the Viet Cong took over in April 1975.
Then his mother had a dream about falling off a cliff and being saved by the hand of the Buddha. Early that morning, his parents packed up eight of their 10 kids—two had gone to the U.S.—and rushed to Saigon’s harbor, where his father commandeered a boat and chugged out to sea with 500 refugees. Pirates shot at them in the South China Sea. They ran out of fuel, and his father threatened to shoot at a Chinese tanker unless it towed his vessel to land.
The Mac­nguyens made it to the Philippines, then to a military camp in Arkansas, and finally to Hyde Park, New York. Curtis’s father sold vacuum cleaners door-to-door. His mother worked in a factory making candy canes.
Macnguyen didn’t speak a word of English when he arrived in the U.S. at the age of 6, the baby of the family. He learned it, worked at McDonald’s, and spent one summer with a sister in Hawaii, picking heart-shaped anthurium flowers off the rain-soaked slopes of the Big Island for $2.17 an hour, illegally. He’d come home covered in leeches.
He worked equally hard in school, captained the tennis team, and went to the University of Pennsylvania to study engineering. A tedious summer job writing computer code in a cubicle prompted a transfer to the Wharton School, even though he knew nothing about finance. He graduated summa cum laude in 1990.
He worked in New York at Morgan Stanley for less than a year before landing a spot at Gleacher & Co., the investment bank founded by Eric Gleacher, who had advised Kohlberg Kravis Roberts on its record-setting, $30 billion buyout of RJR Nabisco in 1989. Gleacher paid more than other banks, and Macnguyen’s salary doubled, to $100,000.
The place turned out to be a hedge fund incubator. At least five other analysts who worked there went on to start funds, including Larry Robbins, later the founder of Glenview Capital Management, which now has $10 billion under management.
The place was perfect for a gambler like Macnguyen. “We’d play Nerf basketball for thousands of dollars,” says former co-worker Raji Khabbaz, who also launched a fund. “Some of those games got pretty expensive.”
The group learned about hedge funds after stock-picking legend Julian Robertson hired Gleacher to try to sell a stake in his Tiger Management. Khabbaz worked on the offering, and he and Macnguyen saw just how lucrative hedge funds could be. They pitched Gleacher on starting one in-house, but he passed, so Macnguyen bolted for Siegler Collery in 1993. “They had $80 million of assets, and that was big,” he says. Einhorn, the soon-to-be founder of Greenlight, joined the firm right afterward.
Macnguyen started Ivory in 1998. He chose the name in part because, over time, he’d spent days spelling Macnguyen on calls, and also because, in Southeast Asia, the elephant and ivory are symbols of good fortune. In its first three years, Ivory Flagship beat the S&P 500 by 7 percentage points, 26 percentage points, and 19 percentage points, respectively.
New York began to wear on Macnguyen, and he decided to make a move he’d long contemplated, to Los Angeles, where he had often traveled for work. “Every time I got off the plane, I had the best feeling in the world,” he says.
In his new suite in Brentwood, Mac­nguyen had healthy returns for years, in all sorts of markets. True to conservative form, he protected investors from disaster in 2008. He caught the rebound in 2009.
Then the Federal Reserve flooded the market with money, lifting almost all boats. But not Macnguyen’s. Amazon rose in his face, and Hospira fell. Nothing seemed to work. He got beaten by index funds, which would be a little like Macnguyen beating Kuchar at golf. It just shouldn’t happen.
The slump changed his core beliefs about his business. Before, he thought he could build a company that would outlast him. Now, that seemed impossible. If he was off his game, then the whole firm lost. No one seemed to step up.
So Macnguyen stepped back into the trenches. In Brentwood, he had allowed himself a private office. Now, he knocked out the wall that separated him from his six analysts. Everyone sits, stands, and mills around in one big room, and he hears everything the analysts say, not just what they choose to report to him in meetings. “A lot of times, it’s the thing that they don’t tell me that’s important,” Macnguyen says.
To the same end, he has software that lets everyone in the firm post ideas, no matter how harebrained, so he can see them. If the ideas pass tests for valuation, profitability, and some three dozen other factors, they get a very serious look.
Working for Macnguyen at Ivory sounds almost as tough as picking flowers in the leechy hills of Hawaii. He rides his staff hard, forcing them to defend their stock selections. “If you don’t put your ass on the line and make a high-conviction call, then you will never learn,” Mac­nguyen says. “You have to be so wrong, and it has to hurt so badly, and everybody has to see it, that you will never make that mistake again.”
If one thing riles Macnguyen most, it’s probably the lack of respect he gets for making money with such low risk. Everyone harps on one number: return. Macnguyen believes sophisticated investors, at least, should be talking about risk-adjusted return, which can be measured by a manager’s efficiency ratio. That’s annual return divided by annual volatility. The higher the ratio, the better. Ivory Optimal’s was 1.65 as of December. That beat another hard-driving hedge funder: Bill Ackman, the top manager in Bloomberg Markets’ 2014 ranking of large funds. His Pershing Square International had a ratio of 1.31.
“People don’t pay attention to guys who make money on a risk-adjusted basis when the market is up,” Macnguyen says. “In the next five years, the market isn’t going to be up as much as it has been.”
With all the crowding in finance these days, one of the easiest mistakes to make is getting stuck in a “hedge fund hotel,” a stock owned mostly by other funds. They’re pricey and crowded, and guests tend to leave all at once. Just before the 2008 crash, Macnguyen saw the funds piling into energy companies and commodities producers, and he stayed clear. When the funds had to sell assets to return money to panicked investors, those stocks got whacked.
Since then, Macnguyen has honed a system for avoiding investors who aren’t in for the long haul: He looks for stocks that are mired in a trench, of sorts: 50 percent below a two-year high and within 20 percent of recent lows. Within that band, the number of shares traded must exceed all of the outstanding shares. “By then, everyone who is nervous is already gone,” he says. They’ll have been churned out. After his systems find a stock that fits, Macnguyen and his team do deep research, calling the company, visiting, and scrutinizing earnings calls.
“The perfect idea for Curtis,” says Jim Vincent, a managing director at AllianceBern­stein who has been pitching ideas to Macnguyen for years, “would be a company that’s profoundly oversold and hated and has a new CEO who has a chip on his shoulder and a heavy ownership of stock that’s locked up for a long time, in a cyclical business that just troughed.”
Boston Scientific met many of those criteria. It showed up on Macnguyen’s churner screen in 2012, after it had fallen below $6 from $14 back in 2008. Ivory started looking at it and learned that the company had been struggling since 2006, when the U.S. Food and Drug Administration barred it from releasing some new products until it resolved manufacturing problems. The FDA lifted the ban in 2010, but the company kept losing money. It hired a new CEO in 2011. Ivory started buying in 2012. The swooning revenue stabilized in 2013, and that was enough to lift the stock to $12 from $6. On April 9, it closed at $18.10. “We try to find really good setups, where you have to be a little bit right to make a lot of money and a lot wrong to lose a little bit of money,” Mac­nguyen says.
Another churner find: memory chip maker Micron Technology. Macnguyen bought it at $6 in 2012. On April 9, it closed at $27.82, and Ivory owns 2 million shares. He points out that Einhorn backed up the truck and bought 23 million shares of Micron in 2013 and paid closer to $14.50.
The victories are adding up for Mac­nguyen, and he says he feels better about things than he did in the depths of his slump. Not content, by any means, but better. Ivory Optimal returned 28.3 percent in 2013, compared with 32.4 percent for the S&P 500. But the Optimal fund’s net exposure to the market—its longs minus its shorts—was just 24.3 percent. In 2014, Optimal rose 11.4 percent, lagging the market by about 2 percentage points. Its net exposure was 33 percent.

Making big money like that with limited risk isn’t easy. Nor is carrying a boulder underwater. These days, Macnguyen can go about 30 yards before he has to drop it and come up for air, probably longer if there’s money on the line. Then he goes back down and lifts it off the sandy bottom for another run.

Thursday, April 09, 2015

Why Your US Equities Underperformed in 2014


Active managers suffered more than usual last year—S&P Dow Jones investigated why.
The proportion of active US equity managers that underperformed the S&P 500 index in 2014 was “extraordinarily high”, according to S&P Dow.
“In a low-dispersion environment, the value of skill goes down.” —Chris Bennett & Craig Lazzara, S&P Dow JonesResearch by the index provider reported that record-low measurements of stock dispersion in the benchmark had wiped out many opportunities for stockpickers to outperform.
“This is not primarily a reflection of manager skill,” wrote Senior Index Analyst Chris Bennett and Managing Director Craig Lazzara. “The problem is that in a low-dispersion environment, the value of skill goes down.”
Of a sample of 362 US equity funds, just 37 outperformed the S&P 500’s 20.76% return in 2014, a significantly lower proportion (10%) than in the previous two years, according to data from FE Analytics collated by CIO.
“For low-dispersion, high correlation sectors, the most important decision is the sector call, rather than individual stock recommendations.”Bennett and Lazzara reviewed historical dispersions within S&P 500 industry sectors and between sectors, as well as comparing the index to mid-cap and small-cap benchmarks. Small caps offered the highest dispersion and volatility measures, the pair found, meaning better stock-picking opportunities were likely.
The pair also argued that managers should focus on sector calls rather than always trying to pick the best individual stocks—at least when it came to the S&P 500.
“For low-dispersion, high correlation sectors, the most important decision is the sector call, rather than individual stock recommendations,” Bennett and Lazzara wrote. “A correct sector call will be reflected relatively consistently across all stocks in the sector.”
Analysts covering utilities or energy companies “would be well advised to spend most time and effort deciding whether to be in or out of the sector”, the report concluded, citing the low levels of dispersion between such stocks. In contrast, analysts working on technology or healthcare “may be better off trying to separate the sectoral wheat from the sectoral chaff.”
S&P 500 sector dispersion and correlation
Bennett and Lazzara’s paper, “Some Implications of Sector Dispersion”, can be downloaded from the S&P Dow Jones website.

Tuesday, April 07, 2015

Research Affiliates : Woe Betide the Value Investor

http://www.researchaffiliates.com/Our%20Ideas/Insights/Fundamentals/Pages/365_Woe_Betide_the_Value_Investor.aspx


Monday, March 23, 2015

New-and-Improved Shiller CAPE: Solving the Dividend Payout Ratio Problem


A common criticism of Professor Robert Shiller’s famous CAPE measure of stock market valuation is that it fails to correct for the effects of secular changes in the dividend payout ratio.  Dividend payout ratios for U.S. companies are lower now than they used to be, with a greater share of U.S. corporate profit going to reinvestment.  For this reason, earnings per share (EPS) tends to grow faster than it did in prior eras.  But faster EPS growth pushes up the value of the Shiller CAPE, all else equal.  Distortions therefore emerge in the comparison between present values of the measure and past values.
To give credit where it’s due, the first people to point out this effect–at least as far as I know–were Professor Jeremy Siegel of Wharton Business School and his former student, David Bianco of Deutsche Bank.  Siegel, in specific, wrote about the problem as far back as late 2008, during the depths of the financial crisis, when the Shiller CAPE was steering investors away from a market that he considered to be extremely cheap (see “Jeremy Siegel on Why Equities are Dirt Cheap”, November 18, 2008, link here).
In a piece from 2013, I attempted to demonstrate the effect with two tables, shown below:
shillerdiv2
shillerdiv1
The tables portray the 10 year earnings trajectories and Shiller CAPE ratios of two identical companies that generate identical profits and that sell at identical trailing-twelve-month (ttm) P/E valuations. The first company, shown in the first table, pays out 75% of its profit in dividends and reinvests the other 25% into growth (in this case, share buybacks that grow the EPS by shrinking the S). The second company, shown in the second table, pays out 25% of its profit in dividends, and reinvests the other 75% into growth.
As you can see, even though these companies are identically valued in all relevant respects, they end up with significantly different Shiller CAPEs.  The reason for the difference is that the second company reinvests a greater share of its earnings into growth than the first company.  Its earnings therefore grow faster.  Because its earnings grow faster, the act of “averaging” them over a trailing 10 year period reduces them by a greater relative amount.  Measured against that trailing 10 year average, the company’s price, appropriately set in reference to its ttm earnings, therefore ends up looking more expensive.  But, in truth, it’s not more expensive–its valuation is exactly the same as that of the first company.
The following chart illustrates the effect:
highlow
To summarize the relationship:
  • Lower Payout Ratio –> Higher Earnings Growth –> Higher CAPE, all else equal
  • Higher Payout Ratio –> Lower Earnings Growth –> Lower CAPE, all else equal
Now, how can we fix this problem?  A natural solution would be to reconstruct the CAPE on the basis of total return (which factors in dividends) rather than price (which does not). But that’s easier said than done.  How exactly does one build a CAPE ratio–or any P/E ratio–on the basis of total return?
Enter the Total Return EPS Index, explained here and here.  The Total Return EPS Index is a modified version of a normal EPS index that tells us, hypothetically, what EPS would have been, now and at all times in history, if the dividends that were paid out to shareholders had not been paid out, and had instead been diverted into share buybacks. Put differently, Total Return EPS tells us what earnings would have been if the dividend payout ratio had been 0% at all times.  In this way, it reduces all earnings data across all periods of history to the same common basis, allowing for accurate comparisons between any two points in time.
Crucially, in constructing the Total Return EPS, we assume that the buybacks are conducted at fair value prices, prices that correspond to the same valuation in all periods (equal to the historical average), rather than at market prices, which are erratic and often groundless.  To those readers who continue to e-mail in, expressing frustration with this assumption–don’t worry, you’re about see why it’s important.
The following chart shows the Total Return EPS alongside the Regular EPS from 1871 to 2015.  In this chart and in all charts presented hereafter, the index is the S&P 500 (and its pre-1957 ancestry), the values are appropriately inflation-adjusted to February 2015 dollars, and no corrections are made for the effects of questionable accounting writedowns associated with the last two economic downturns:
trp
Now, if all S&P 500 dividends had been diverted into share buybacks, then the price of the index would have increased accordingly. We therefore need a Total Return Price index–an index that shows what prices would have been on the “dividends become buybacks” assumption.
Calculating a Total Return Price index is straightforward.  We simply assume that the market would have applied the same P/E ratio to the Total Return EPS that it applied to the Regular EPS (and why would it have applied a different P/E ratio?). Multiplying each monthly Total Return EPS number by the market’s ttm P/E multiple in that month, we get the Total Return Price index.
In the chart below, we show the Total Return Price index for the S&P 500 alongside the Regular Price, from 1871 to 2015:
trp3
Generating a CAPE from these measures is similarly straightforward.  We divide the Total Return Price by the trailing 10 year average of the Total Return EPS.  The result: The Total Return EPS CAPE.
Shiller himself proposed a different method for calculating a CAPE based on total return in a June 2014 paper entitled “Changing Times, Changing Valuations: A Historical Analysis of Sectors within the U.S. Stock Market: 1872 to 2013″ (h/t James Montier). The instructions for the method are as follows: Use price and dividend information to build a Total Return Index. Then, scale up the earnings by a factor equal to the ratio between the Total Return Index and the Price Index.  Then, divide the Total Return Index by the trailing ten year average of the scaled-up earnings.  In a piece from August of last year, I tried to build a CAPE based on Total Return using yet another method (one that involves growing share counts), and arrived at a result identical to Shiller.  The technique and charts associated with that method are presented here.
It turns out that both of these methods produce results identical to the Total Return EPS CAPE method, with one small adjustment: that we conduct the buybacks that form the Total Return EPS at market prices, rather than at fair value prices as initially stipulated. The following chart shows the three types of Total Return CAPEs together.  As you can see, the lines overlap perfectly.
fjksake
The three different versions of the CAPE overlap because they are ultimately doing the same thing mathematically, though in different ways.  Given that they are identical to each other, I’m going to focus only on the Total Return EPS version from here forward.  I’m going to refer to the version that conducts buybacks at fair value prices as “Total Return EPS (Fair Value) CAPE”, and the version that conducts buybacks at market prices as “Total Return EPS (Market) CAPE.”  I’m going to refer to Shiller’s original CAPE simply as “Shiller CAPE.”
The following chart shows the Total Return EPS (Market) CAPE alongside the Shiller CAPE, with the values of the former normalized so that the two CAPEs have the same historical average (allowing for a direct comparison between the numbers).
trmkt
(Note: in prior pieces, I had been comparing P/E ratios to their geometric means. This is suboptimal. The optimal mean for a P/E ratio time series is the harmonic mean, which is essentially what you get when you take an average of the earnings yields–the P/E ratios inverted–and then invert that average.  So, from here forward, in the context of P/E ratios, I will be using harmonic means only.) (h/t and #FF to @econompic, @naufalsanaullah, @GestaltU_BPG)
The current value of the Shiller CAPE is 27.5, which is 93% above its historical average (harmonic) of 14.2.  The current value of the Total Return EPS (Market) CAPE is 30.3, which is 71% above its historical average (harmonic) of 17.8.  Normalized to matching historical averages, the current value of the Total Return EPS (Market) CAPE comes out to 24.2.
At current S&P 500 levels, then, we end up with 27.5 for the Shiller CAPE, and 24.2 for the Total Return EPS (Market) CAPE, each relative to a historical average of 14.19. Evidently, the difference between the two types of CAPEs is significant, worth 12%, or 250 current S&P points.
But there’s a mistake in this construction.  To find it, let’s take a closer look at the chart:
mvout
From the early 1990s onward, the Total Return EPS (Market) CAPE (the red line) is significantly below the Shiller CAPE (the blue line), suggesting that the Shiller CAPE is overstating the market’s expensiveness, and that the Total Return EPS (Market) CAPE is correcting the overstatement by pulling the metric back down.
What is driving the Shiller CAPE’s apparent overstatement of the market’s expensiveness? The obvious answer would seem to be the historically low dividend payout ratio in place from the early 1990s onward.  All else equal, low dividend payout ratios push the Shiller CAPE up, via the increased growth effect described earlier.
But look closely.  Whenever the market is expensive for an extended period of time, the subsequent Total Return EPS (Market) CAPE (the red line) ends up lower than the Shiller CAPE (the blue line), by an amount seemingly proportionate to the degree and duration of the expensiveness.  Note that this is true even in periods when the dividend payout ratio was high, e.g, the periods circled in black: the early 1900s, the late 1920s, and the late 1960s.  If the dividend payout ratio were the true explanation for the deviations between the Shiller CAPE and the Total Return EPS (Market) CAPE, then we would not get that result.  We would get the opposite result: the high dividend payout ratio seen during the periods would depress the the Shiller CAPE relative to the more accurate total measures; it would not push the Shiller CAPE up, as seems to be happening.
The converse is also true.  Whenever the market is cheap for an extended period of time, the subsequent Total Return EPS (Market) CAPE (the red line) ends up higher than the subsequent Shiller CAPE (the blue line), by an amount seemingly proportionate to the degree and duration of the cheapness.  We see this, for example, in the periods circled in green: the early 1920s and the early 1930s through the end of the 1940s.  The deviation between the two measures is spatially small in those periods, but that’s only because the numbers themselves are small–single digits.  On a percentage basis, the deviation is sizeable.
The following chart clarifies:
hsit
So what’s actually happening here?  Answer: valuationnot the dividend payout ratio–is driving the deviation.  In periods where the market was cheap in the 10 years preceding the calculation, the Total Return EPS (Market) CAPE comes out above the Shiller CAPE. In periods where the market was expensive in the 10 years preceding the calculation, the the Total Return EPS (Market) CAPE comes out below the Shiller CAPE.  The degree above or below ends up being a function of how cheap or expensive the market was, on average.
The following chart conclusively demonstrates this relationship:
pinkgreen
The bright green line is the difference between the Total Return EPS (Market) CAPE and the Shiller CAPE as a percentage of the Shiller CAPE.  When the bright green line is positive, it means that the red line in the previous chart was higher than the blue line; when negative, vice-versa.  The pink line is a measure of how cheap or expensive the market was over the preceding 10 years, on average and relative to the historical average. When the pink line is positive, it means that the market was cheap; when negative, expensive.  The two lines track each other almost perfectly, indicating that the valuation in the preceding years–and not the payout ratio–is driving the deviation between the two measures.
What is causing this weird effect?  You already know.  The share buybacks associated with the Total Return EPS (Market) CAPE are being conducted at market prices, rather than at fair value prices.  The same is true for the dividend reinvestments associated with Shiller’s proposed Total Return CAPE and with the version I presented in August of last year; those reinvestments are being conducted at market prices.  That’s wrong.
When share buybacks (or dividend reinvestments) are conducted at market prices, then periods of prior expensiveness produce lower Total Return EPS growth (because the dividend money is invested at unattractive valuations that offer low implied returns).  And, mathematically, what does low growth do to a CAPE, all else equal?  Pull it down.  Past periods of market expensiveness therefore pull the Total Return EPS (Market) CAPE down below the Shiller CAPE, as observed.
Conversely, periods of prior cheapness produce higher Total Return EPS growth (because the dividend money is reinvested at attractive valuations that offer high implied returns).  And what does high growth do to a CAPE, all else equal?  Push it up.  Past periods of market cheapness therefore push the Total Return EPS (Market) CAPE up above the Shiller CAPE, as observed.
Looking at the period from the early 1990s onward, we assumed that the problem was with the Shiller CAPE (the blue line), that the low dividend payout ratio during the period was pushing it up, causing it to overstate the market’s expensiveness.  But, in fact, the problem was with ourTotal Return EPS (Market) CAPE (the red line).  The very high valuation in the post-1990s period is depressing Total Return EPS (Market) growth (the expensiveness of the share buybacks and dividend reinvestments shrinks their contribution), pulling down on the Total Return EPS (Market) CAPE, and causing it to understate the market’s  expensiveness.
The elimination of this distortion is yet another reason why the buybacks and dividend reinvestments that form the Total Return EPS (or any Total Return Index used in valuation measurements) have to be conducted at fair value prices, rather than at market prices.  Conducting the buybacks and dividend reinvestments at fair value prices ensures that they provide the same accretion to the index across all periods of history, rather than highly variable accretion that inconsistently pushes up or down on the measure.
Now, a number of readers have written in expressing disagreement with this point.  To them, I would ask a simple question: does it matter to the current market’s valuation what the market’s valuation happened to be in the distant past?
Suppose, for example, that in 2009, investors had become absolutely paralyzed with fear, and had sold the market’s valuation down to a CAPE of 1–an S&P level of, say, 50. Suppose further that the earnings and the underlying fundamentals had remained unchanged, and that investors had exacted the pummeling for reasons that were entirely irrational. Suppose finally that investors kept the market at the depressed 1 CAPE for two years, and that they then regained their senses, pushing the market back up to where it is today, in a glorious rally.  In the presence of these hypothetical changes to the past, what would happen to the current value of a Total Return EPS CAPE that reinvests at market prices?  Answer: it would go up wildly, dramatically, enormously, because the intervening dividends that form the Total Return index would have been invested at obscenely low valuations during the period, producing radically outsized total return growth.  What does high growth due to a CAPE? Push it up, so the CAPE would rise–by a large amount.
Is that a desirable result?  Do we want a measure whose current assessment of valuation is inextricably entangled in the market’s prior historical valuations, such that the measure would judge the valuations of two markets with identical fundamentals and identical prices to be significantly different, simply because one of them happened to have traded more cheaply or expensively in the past?  Obviously not.  That’s why we have to conduct the buybacks and reinvestments that make up the Total Return EPS at fair value.
The general rule is as follows.  When we’re using a Total Return index to model actual investor performance–what an individual who invested in the market would have earned, in reality, with the dividend reinvestment option checked off–we need to conduct the hypothetical reinvestments that make up the Total Return index at market prices.  But when we’re using a Total Return index to measure valuation–how a market’s price compares with its fundamentals–then we need to conduct the hypothetical reinvestments at fair value prices.
The following chart shows the Total Return EPS CAPE properly constructed on the assumption that the buybacks and reinvestments occur at fair value prices:
trepsfv
As you can see, the deviation between the two measures comes out to be much smaller. Normalized to the same historical average, the current value of the Total Return EPS (Fair Value) CAPE ends up being 25.9, versus 27.5 for the original Shiller CAPE.  The difference between the total return and the original measures comes out at 5.7%, a little over 100 current S&P points (versus 12% and 250 points earlier).
Surprisingly, then, properly reinvesting the dividends at the same valuation across history more than cuts the deviation in half, to the point where it can almost be ignored.  As far as the CAPE is concerned, when it comes to the kinds of changes that have occurred in the dividend payout ratio over the last 144 years, there appears to be little effect on the accuracy of Shiller’s original version.  The entire exercise was therefore unnecessary. Admittedly, this was not the result that I was anticipating, and certainly not the result that I was hoping to see.  But it is what it is.
It turns out that Shiller was right to reject the dividend payout ratio argument in his famous 2011 debate with Siegel and Bianco:
“Mr. Shiller did his own calculation about the impact of declining dividends on earnings growth and concluded that it is marginal at best, not meriting any adjustment.” — “Is the Market Overvalued?”, Wall Street Journal, April 9th, 2011.
If the subsequent foray into Total Return space caused him to change that view, then he should change it back.  He was right to begin with.  His critics on that point, myself included, were the ones that were wrong.
Now, this is not to suggest that we shouldn’t prefer to use the Total Return version of the CAPE over Shiller’s original version.  We should always prefer to make our analyses as accurate as possible, and the Total Return version of the CAPE is unquestionably the more accurate version.  Moreover, even though the changes in the dividend payout ratio seen in the U.S. equity space over the last 144 years have not been large enough to significantly impact the accuracy of the original version of the CAPE, the differences between the payout ratios of different countries–India and Austria, to use an extreme example–might still be large enough to make a meaningful difference.  Since the Shiller CAPE is the preferred method for accurately comparing different countries on a valuation basis, it only makes sense to shift to the more accurate Total Return version.  Fortunately, that version is simple and intuitive to build using Total Return EPS.
Admittedly, there is some circularity here.  In building the Total Return EPS Index on the assumption of fair value buybacks, we used the Shiller CAPE as the basis for estimating fair value.  If the Shiller CAPE is inaccurate as a measure of fair value, then our Total Return EPS index will be inaccurate, and therefore our Total Return CAPE, which is built on that index, will be inaccurate.  Fortunately, in this case, there’s no problem (otherwise I wouldn’t have done it this way). When you run the numbers, you find that the choice of valuation measure makes little difference to the final product, as long as a roughly consistent measure is used.  You can build the Total Return EPS Index using whatever roughly consistent measure you want–the Total Return CAPE will not come back appreciably different from Shiller’s original. What drove the deviations in the earlier charts were not small differences in the valuations at which dividends were reinvested, but large differences–for example, the difference associated with reinvesting dividends at market prices from 1942 to 1952, and then from 1997 to 2007, at prices corresponding to three times the valuation.
Now, there are other ways of adjusting for the impact of changing dividend payout ratios. Bianco, for example, has a specific technique for modifying past EPS values. As he explains:
“The Bianco PE is based on equity time value adjusted (ETVA) EPS.  We raise past period EPS by a nominal cost of equity estimate less the dividend yield for that period.”
I cannot speak confidently to the accuracy of Bianco’s technique because I do not have access to its details.  But if the method produces a result substantially different from the Total Return EPS CAPE (which it appears to do), then I would think that it would have to be wrong.  When it comes to changing dividend payout ratios, the Total Return EPS CAPE is airtight.  It treats all periods of history absolutely equally in all conceivable respects, perfectly reducing them to a common basis of 0% (payout).  Because it reinvests the dividends at fair value (the historical average valuation), every reinvested dividend in every period accretes at roughly the same rate, which corresponds to the actual average rate at which the market has historically accreted gross of dividends (approximately 6% real).
If our new-and-improved version of the CAPE is appropriately correcting the dividend payout ratio distortions contained in the original version, then the deviation between our new-and-improved version and the original version should be a clean function of that ratio (rather than a function of other irrelevant factors, such as past valuation).  When the dividend payout ratio is low, our new-and-improved version should end up below the original version, given that the original version will have overstated the valuation.  When the dividend payout ratio is high, our new-and-improved version should end up above the original version, given that the original version will have understated the valuation.
Lo and behold, when we chart the deviation between the two versions of the CAPE alongside the dividend payout ratio, that is exactly what we see: a near-perfect correlation (91%), across the full 134 year historical period.
delta
The blue line shows the difference between our new-and-improved version of the CAPE and the original version.  The red line shows the trailing Shiller dividend payout ratio, which is the 10 year average of real dividends per share (DPS) divided by the 10 year average of real EPS.  We use a Shillerized version of the dividend payout ratio to remove noise associated with recessions–especially the most recent one, where earnings temporarily plunged almost to zero, causing the payout ratio to temporarily spike to a value north of 300%.
The fact that the two lines overlap almost perfectly indicates that the deviation between our new-and-improved version and the original version is a function of the factor–the dividend payout ratio–that is causing the inaccuracy in the original version, rather than some other questionable factor.  That is exactly what we want to see.  It is proof positive that our new-and-improved version is correcting the distortion in question, and not introducing or exploiting other distortions (that, conveniently, would make the current market look cheaper).
Now, to be clear, the secular decline in the dividend payout ratio seen across the span of U.S. market history has not substantially affected the accuracy of the original Shiller CAPE.  However, it has substantially affected the trend growth rate of EPS.  So, though it may not be imperative that we use the Total Return version of the CAPE when measuring valuation, it is absolutely imperative that we use the Total Return version of EPS when analyzing earnings trends and projecting out future earnings growth.
We are left with the question: if the distortions associated with the dividend payout ratio are not significant, then why does the Shiller CAPE show the U.S. equity market to be so expensive relative to history?  We can point to three explanations.
  • First, on its face, the market just is historically expensive–even on a non-Shiller P/E measurement.  Using reported EPS, the simple trailing twelve month P/E ratio is roughly 20.5, which is 53% above its historical average (harmonic) of 13.4.  Using S&P corporation’s publication of operating EPS, the simple trailing twelve month P/E ratio is 18.8, which is 40% above that average.
  • Second, the accounting writedowns associated with the 2008-2009 recession are artificially weighing down the trailing average 10 year EPS number off of which the Shiller CAPE is calculated.  Prior to 2014, this effect was more significant than it is at present, given that the 2001-2003 recession also saw significant accounting writedowns.  The trailing 10 year average for the years up to 2014 therefore got hit with a double-whammy.  That’s why the the increase in the Shiller CAPE in recent years has not been as significant as the increase in market prices (since December 2012, the CAPE is up roughly 30%, but prices are up roughly 50%).  2014 saw the 2001-2003 recession fully drop out of the average, reducing the CAPE’s prior overstatement.
  • Third, as the chart below shows, real EPS growth over the last two decades–on both a regular and a Total Return basis–has been meaningfully above the respective historical averages, driven by substantial expansion in profit margins.  Recall that high growth produces a high CAPE, all else equal.
profmarginincluded
These last two factors–the effects of accounting writedowns and the effects of profit margin expansion–will gradually drop out of the Shiller CAPE (unless you expect another 2008-type recession with commensurate writedowns, or continued profit margin expansion, from these record levels).  As they drop out, the valuation signal coming from the Shiller CAPE will converge with the signal given by the simple ttm P/E ratio–a convergence that is already happening.
We conclude with the question that all of this exists to answer: Is the market expensive? Yes, and returns are likely to be below the historical average, pulled down by a number of different mechanisms.  Should the market be expensive?  “Should” is not an appropriate word to use in markets.  What matters is that there are secularsustainable forces behind the market’s expensiveness–to name a few: low real interest rates, a lack of alternative investment opportunities (TINA), aggressive policymaker support, and improved market efficiency yielding a reduced equity risk premium (difference between equity returns and fixed income returns).  Unlike in prior eras of history, the secret of “stocks for the long run” is now well known–thoroughly studied by academics all over the world, and seared into the brain of every investor that sets foot on Wall Street.  For this reason, absent extreme levels of cyclically-induced fear, investors simply aren’t going to foolishly sell equities at bargain prices when there’s nowhere else to go–as they did, for example, in the 1940s and 1950s, when they had limited history and limited studied knowledge on which to rely.
As for the future, the interest-rate-related forces that are pushing up on valuations will get pulled out from under the market if and when inflationary pressures tie the Fed’s hands–i.e., force the Fed to impose a higher real interest rate on the economy.  For all we know, that may never happen.  Similarly, on a cyclically-adjusted basis, the equity risk premium may never again return to what it was in prior periods, as secrets cannot be taken back.

Lunch is for wimps

Lunch is for wimps
It's not a question of enough, pal. It's a zero sum game, somebody wins, somebody loses. Money itself isn't lost or made, it's simply transferred from one perception to another.