Felix Salmon

I’m a big fan of OpenTable as a service. I’ve made 60 OpenTable reservations in total, using both the website and their iPhone app, and I haven’t had a single bad experience: the resy has never been lost, I’ve never had an inordinate wait for a table, and it’s extremely easy to share the reservation information with my fellow diners. There’s a few things I’d like to see added, like reservations I’ve been invited to showing up on my upcoming-reservations page, but in general OpenTable does one thing from the diner’s perspective and it does it extremely well.

If Mark Pastore is to be believed, however, restaurants hate OpenTable nearly as much as diners like me love it:

A few months ago I took an informal survey of several other restaurateurs here in San Francisco and in New York, all of whom offer seats through OpenTable, asking them about the value of OpenTable from the restaurateur’s perspective.

Only one of the dozen or so I spoke with said he felt that OpenTable increased the value of his restaurant and that he wouldn’t imagine opening a new project without it. The rest were less than happy. The recurring themes were the opinion that OpenTable took home a disproportionate (relative to other vendors) chunk of the restaurants’ revenues each month and the feeling of being trapped in the service… The GM of one very well known New York restaurant group, which spends thousands of dollars on OpenTable each month, put it to me this way, “OpenTable is out for itself, the worst business partner I have ever worked with in all my years in restaurants. If I could find a way to eliminate it from my restaurants I would.”

Pastore says that the economics of OpenTable can be devastating for the notoriously difficult restaurant business:

One independent study estimates that OpenTable’s fees (comprised of startup fees, fixed monthly fees, and per-person reservation fees) translate to a cost of roughly $10.40 for each “incremental” 4-top booked through OpenTable.com. To put that in perspective, consider that the average profit margin, before taxes, for a U.S. restaurant is roughly 5%. This means that a table of 4 spending $200 on dinner would generate a $10 profit. In this example, all of that profit would then go to OpenTable fees for having delivered the reservation, leaving the restaurant with nothing other than the hope that that customer would come back (and hopefully book by telephone the next time).

At this point it’s worth digging out that famous short thesis on OpenTable. In the second quarter, OpenTable seated 15.6 million diners across 14,128 restaurants, and made total revenue of $22.45 million, including revenue from installing its software in new restaurants. That’s $1.44 per seated diner, or $5.76 per 4-top, all included. (The actual marginal revenue that OpenTable gets from reservation revenue rather than fixed subscription revenue is 69 cents per seated diner, or $2.76 per 4-top.)

In order to get to a cost of $10.40 per “incremental” 4-top, you have to think that OpenTable has value only insofar as it drives diners to a restaurant who would otherwise not have dined there; that’s certainly the mindset that Pastore seems to have. But OpenTable can, should, and does have much more value than that. For one thing, it improves the quality of the diner’s experience: we diners no longer have to worry whether the reservation we thought we made will be honored when we get to the restaurant. On top of that, it’s reasonable to expect that a decent restaurant will remember if we’re vegetarian, or we’re allergic to garlic, or something like that, in a way that’s much harder when we’re just another name.

Much more importantly, however, the OpenTable software can be hugely valuable to restaurants’ own bottom line, even if they would be sold out every night without it. Dan Simons of Founding Farmers Fathers, a restaurant which spends $6,000 per month on OpenTable, says in response to Pastore that once he started taking advantage of its software capabilities, he managed to increase sales by 15%.

Simons is talking his book: he’s a principal of a company which advises restaurants on how to best use OpenTable’s software. But it makes intuitive sense that armed with the abundance of rich and restaurant-specific data that OpenTable’s software provides, a quantitavely-minded restaurant manager would be able to find all manner of opportunities to improve diners’ experience and maximize the restaurant’s sales. (These two things are much more likely to work with each other than against each other.)

On the other hand, it also makes intuitive sense that most restaurants lack quantitavely-minded managers who are willing and able to take advantage of OpenTable’s rich data. For them, OpenTable really is little more than a way of getting extra marginal diners in the door: after that, they do what they’ve always done, and they’re not going to change their ways any time soon. For those people, it’s easy to see why they might be disgruntled at paying large subscription fees for services they never use.

The best-case scenario here is that, armed with OpenTable software they pretty much have to pay for anyway, restaurateurs will increasingly find themselves making full use of it, becoming more efficient and more lucrative businesses as a result, running more enjoyable places to dine. Alternatively, as the short thesis puts it, OpenTable will find that its business “has absolutely no barriers to entry,” and copy-cats will do to it what they did to Tivo, destroying it with an inferior but cheaper product.

But the main lesson for OpenTable from both Pastore and Simons is that their software is, right now, far too difficult to use to its full potential. Now that the company is valued at over $1.5 billion, it should have no problem paying some very smart developers to improve the user-friendliness of its offering, not only for diners but also for restaurateurs. With luck, that will help to reduce the amount of resentment the company faces in the industry, by making OpenTable service something which more than pays for itself, rather than something which is an expensive necessity. After all, I’m sure that while on one level OpenTable would love to become the Ticketmaster of the restaurant world, it would love to do so without becoming nearly as hated.

Update: As proof that restaurateurs tend not to be the most web-savvy of businesspeople, one reader offers this, via email:

One thing I’ve noticed – restaurants never take advantage of the feedback features in open table.  I’ve probably submitted feedback on about 7 or 8 restaurants, writing specific notes about something I particularly liked or didn’t like, using their system that says that the info and my email address will be shared with the restaurant.  I’ve never gotten an email back – either to apologize for bad service or to thank me for the kind words.  That’s just nuts, a huge opportunity missed by the restaurants.  At the very least they should be asking me to join an email list if I said good things.

Maybe it’s the open table technology itself that’s bad, but I think it’s probably just as much that the restaurant business just doesn’t get the new online world we’re living in.  If I was a restaurant I’d be checking blogs and twitter and open table feedback every day, offering my fans invites to special dinners (with bigger mark ups) and offering the haters a discount if they come back and give it another try.  My guess is that restaurants are still living in a world where the local restaurant critic’s review matters 100 times more than anything else – just look at their awful flash-based websites that you can’t even access on an iPhone.  If I’m on foot, it’s easier to find a restaurant’s address on OpenTable than it is on their own sites.

Update 2: OpenTable itself weighs in, saying exactly what its fees are: $199 a month, and then $1 per seated diner if you book through OpenTable’s website, or $0.25 per seated diner if you book through the restaurant’s own website. And there’s also “a web-based product that caters to restaurants that don’t require all of the operational benefits of our Electronic Reservation Book”. But in any case, if you want to book a table at a restaurant, it’s better for the restaurant if you do it through their own website rather than through OpenTable.com.

Shares in Assurant fell off a cliff this morning, on volume a good order of magnitude greater than is normally seen. What’s going on?

CNBC, this morning, was nice enough to cite my story from yesterday as the reason for the move, despite the fact that I added no new information to that which was already included in the American Banker article by Jeff Horwitz. But Horwitz’s story was out all day—I got sent the link just after 1pm, and it was definitely being passed around via email over the course of the afternoon, as the stock went basically nowhere. My post appeared after the market closed, and suddenly this morning Assurant shares plunged at the open.

If my post had anything to do with the move, I think it was due to framing: the American Banker headline, appropriately for a banking trade mag, concentrated on the mortgage servicers, while my headline talked about an “insurance scandal.” I also pulled out a powerful fact from 2,200 words in to the 3,000-word story—that Assurant’s force-placed insurance unit has accounted for $811 million of its $879 million in profits during the last two years.

Finally, I noted something Horwitz mentions only at the very end of the story—that the kind of activity he’s talking about is already illegal, under Dodd-Frank. That means that as soon as there’s a critical mass of publicity around the issue, someone somewhere is likely to crack down on this activity. Maybe regulators don’t spend a lot of time doing close reads of American Banker stories—but they do pay attention when people start talking explicitly about scandals.

Fundamentally, however, I think that what we’re seeing is a function of the amount the stock market in general has risen of late. Assurant is trading around $36 today, which is where it was as recently as the end of August; it was lower than $30 in February, and often something quite small is enough to spark a substantial drop in the share price if there’s a lot of nervousness in the market. The stampede of people rushing to sell Assurant stock this morning says to me that there’s a fair amount of fear in the market right now. And given the high degree of correlation in the market, I wouldn’t be at all surprised to see a big drop in broad indices one morning soon.

Are HP’s directors physically incapable of letting l’affaire Mark Hurd drop? Not only are their fingerprints all over the huge WSJ article on the subject today and Adam Lashinsky’s less exhaustive article in Fortune, but they’ve also decided to give the original letter accusing Hurd of impropriety to a San Diego law firm representing HP shareholders, making it certain that the letter will eventually become public. And it stands to reason that someone on the HP board was responsible for the bizarre NY Post story a couple of weeks ago claiming that Hurd had an affair with a Sun executive.

There are clearly multiple board sources, too: Fortune refers to the woman who hired Jodie Fisher as Caprice Fimbres, describing her as Hurd’s “program manager”, while the WSJ calls her Caprice McIlvaine, and calls her Hurd’s “unofficial chief of staff”. (On her LinkedIn page, she says that she was Hurd’s chief of staff.) It seems that she was ultimately responsible not only for filing Hurd’s fatally inaccurate expense accounts, but also for deciding that the best place to find a gatekeeper for Hurd was from the group of “cougars” on a reality TV show called “Age of Love”. She also flew Fisher to the Grove Hotel in Boise, where Fisher dined with Hurd and watched the Minnesota Vikings play the Green Bay Packers in his hotel room, but didn’t do any work for HP.

All of these revelations — including the unproved accusation that Hurd told Fisher about his bid for EDS — might well harm Hurd, but they also make the HP board seem leaky and defensive, rather than being concentrated on its main job, which is representing shareholders and overseeing the strategic direction of the company. What’s clear is that the arrival of Ray Lane as chairman hasn’t stopped the leaks or made the board seem any more grownup than it was before; quite the opposite, in fact. If I were an HP shareholder, I’d be worried about that: the company clearly needs leadership and strategic direction, but instead the board seems to be more interested in slinging mud at its former chairman. Depressing.

Bank stocks didn’t do so well this week, what with foreclosuregate coming to a boil. But they didn’t do all that badly, either, as a group: the XLF financial sector ETF ended the week down a pretty modest 2.45% from where it started.

You might remember the XLF fund from a famous column by Evan Newmark two years ago, a few weeks after Lehman Brothers declared bankruptcy and the global financial system threatened to implode into a mess of nationalization and mass insolvency:

Bear Stearns, gone. Lehman Brothers, gone. Merrill Lynch, gone. Washington Mutual, gone. Wachovia, gone. Fannie and Freddie, basically gone. AIG, almost gone.

Absolute carnage. The fastest restructuring of a banking system in economic history.

Will the U.S. financial-services industry survive? Plenty of folks think not. But if you believe yes, now is the time to buy financial stocks.

Which is why I have been buying the XLF, the financial sector exchange-traded fund…

The optimist will tell you that we have a crisis of confidence. That with the Treasury’s bailout program in place, bad assets will be speedily removed from balance sheets and credit will again flow. Throw in a Federal Reserve interest-rate cut and soon banks profits will follow.

Frankly, I don’t know who is right today. But I have a five- to 10-year time horizon, so I don’t have to know that.

All I have to do is believe that the US financial-services industry will survive…

For the XLF a break down of more than 20% from Friday’s close, would put it at less than $15.

This would probably indicate the total collapse of the U.S. financial system. And I just don’t buy that. I am buying the XLF instead.

In many ways, the column was prescient. The US financial system did survive. Treasury’s bailout program, along with Fed rate cuts, did indeed break the back of the financial crisis, and large bank profits have followed as a result.

Yet XLF has been trading between $13 and $15 since May — a level which, according to Newmark, “would probably indicate the total collapse of the U.S. financial system”. And looked at over the past three years, it’s pretty clear what happened to the XLF: it fell off a cliff and then recovered to settle happily at a new, low level. Here’s the chart; the vertical line marks the date that Newmark’s column appeared.

I think that this helps to provide, at least in part, an answer to Ezra’s question about why the markets don’t seem to care about the foreclosure crisis: they’ve known about it all along. (For instance, see this story from Reuters’s Patrick Rucker, dated July 27 2007.)

What’s happened over the past week or so is that the mortgage shoe has finally dropped, as it inevitably was going to do sooner or later. But since the markets were already pricing in that shoe-drop, they haven’t needed to overreact this week. They didn’t know when all this was going to happen, but they were relatively well prepared for this: it’s a slow trainwreck, not a sudden crisis. And the still-depressed level of financial stocks is testament to how none of this comes as much of a surprise.

Steven Davidoff goes into lots of detail today on the close-run fight between Ron Burkle and Leonard Riggio for control of three board seats at Barnes & Noble. It was a nailbiter of a vote, and informed opinion had it that Riggio, B&N’s founder, was going to end up the loser, despite controlling a large chunk of the outstanding shares. After all, the most powerful shareholder advisory firm, Institutional Shareholder Services, favored Burkle — and big investors like Vanguard and BlackRock generally follow ISS’s lead.

This time around, however, they didn’t, which is interesting. But what’s absolutely astonishing, in a vote of this importance, is the pathetic showing from State Street, which controls about 1 million shares, or about 1.75% of the company. In a vote as close as this, that’s a massively important stake, which can easily tip the outcome one way or the other.

State Street somehow came to the conclusion that it wanted to vote for an unholy mixture of the two antagonists: for two of Burkle’s nominees, but against Burkle himself, and against Riggio’s poison-pill plan. It was a vote which would have satisfied no one — had it counted.

But then, just to add an element of utter farce to the proceedings, State Street contrived to vote its shares late, thereby ensuring that none of its votes were counted at all.

If State Street had simply intended to vote for one side or the other, the move could have been some kind of weird schoolyard attempt to curry favor with the winner had the vote gone the other way: State Street could always have said “I did vote for you”, or “I didn’t really vote against you”, or something annoying like that.

But since the attempted vote itself was so lily-livered, this looks to me like simple incompetence.

Maybe a shareholding worth $17 million or so is ultimately just not all that important to a firm the size of State Street. But it’s surely important to Burkle, Riggio, and B&N’s board, and these kind of antics come close to openly mocking the concept of shareholder democracy.

We hear a lot about the obligation that companies have to their shareholders. But equally, large shareholders have an obligation to the companies they own: to take their stake seriously, and not to play silly games by delaying borderline-incomprehensible votes until it’s too late to cast them. If this is how State Street treats the companies it owns, I wouldn’t want to entrust them with my heard-earned savings.

Update: Some good comments here, surrounding State’s Street’s status as a custodian and the difficulties it faces in learning from the shares’ beneficial owners how it should vote. But Davidoff made it sound as though State Street was going to vote all its shares the same way; is that not true? And the delay in voting seems to have been a matter of minutes, rather than days. In any case, it seems to me that voting shares is one of the few things that custodians are expected to do well, and that State Street obviously failed on that front, this time.

Here’s the volume-based stock chart you’ve all been waiting for: the one for May 6, the day of the flash crash. Since the big spike in volume was concentrated at the end of the day, in the final hour of trading, the time-based chart squeezes a huge amount of activity into a relatively small horizontal space. The volume-based chart gives the crash a bit more space.

On the other hand, it’s worth nothing that most of the day’s trading still took place before the crash happened.

On thing that strikes me about this chart is not the crash itself but rather the run-up to it: the initial drop from about 1,160 on the S&P down to about 1,120. On the time-based chart, the decline starts slowly and then rapidly speeds up; on the volume-based chart, however, it’s much steadier. And in fact we saw roughly as much volume in the normally-quiet hours between about noon and 2:40 as we did during the craziness of the crash itself and its aftermath. I’m not going to hazard a guess as to what this means, but I do think it shows that May 6 was a pretty unusual day in the markets even before the flash crash happened.

Many thanks to Omer Uzun at Proteus Financial for putting this together: it’s only one tiny piece in the puzzle, but surely every little bit helps.

I’ve been pondering the disconnect between stock and bond prices of late. It reminds me a bit of late 2007, only the other way around: back then there was a credit crunch, but the stock market continued to hit new highs. Today, stocks are largely sitting out the bond bubble.

Exhibit A in these matters always seems to be Johnson & Johnson, which has a dividend yield significantly higher than its 10-year bond yield of about 3%. Whitney Tilson went on TV yesterday and put the thesis in its starkest terms:

We don’t understand what any long-term investor could possibly be thinking when Johnson & Johnson, a month ago, issued all-time record low 10-year corporate debt at 2.95% and the stock yielding 3.8%.

Clearly the two markets are telling you that what investors rationally might be thinking is that we’re in for ten years of Japanese-style deflationary horrific environment, and I think the odds of that are less than 5%. I think it is 95% likely, for example, in the case of Johnson & Johnson, that if you’re going to own something for ten years, you will do massively better owning the stock, both from the appreciation of the stock and the dividend yield you get over those 10 years, than a mere 2.95% owning that 10-year bond they just issued.

There are two problems with this argument, as I see it. The first is that we’re not faced with a simple binary outcome, where either we have Japanese-style deflation, or else the stock will outperform the debt. It’s entirely possible — there’s a nonzero possibility — that J&J stock will fall over the next ten years, rather than rise: that Whitney’s stock appreciation will be negative rather than positive.

J&J is an extremely creditworthy company, and you can be pretty sure that if you paid $62 for that bond at a 2.95% yield, then over the next ten years the company will pay you a total of $80.29 in nominal dollars. (You’ll see why I chose $62 in a minute.)

Meanwhile, the cashflow associated with buying $62 of J&J stock is much less certain. You might well get a 3.8% dividend next year — that’s $2.36 — but you don’t know what you’ll make in subsequent years. Let’s assume the dividend stays constant at its current level: then over ten years, you’ll get $23.56 in dividend payments.

In order for the stock to outperform the bond, then, it has to be worth more than $56.73 in ten years’ time. (I chose $62 because that’s the price that the stock is trading at today.) That’s hardly unthinkable: in fact, it’s pretty much where the stock was trading just three weeks ago.

Of course, you can tweak your assumptions. You can start reinvesting dividends — but then you’d need to make assumptions about reinvesting the bond coupons, too. You can say that some kind of dividend growth is likely — but then you’d also need to price in the probability of the dividend being cut. And so on and so forth. What’s more, there’s something safe and attractive about just getting $80.29 in cash, compared to getting $23.56 in cash and a stock certificate which you could theoretically sell for something over $56.73.

More generally, if you buy J&J stock, you’re making a bet on the future prospects of a particular U.S.-based consumer-goods company. Those prospects might be good, but there’s still a lot of idiosyncratic risk there. The bonds, meanwhile, are almost certain to be paid back in full whatever happens to the company. (Unless of course J&J takes Tilson’s advice and runs with it, levering up enormously in order to buy back its stock. But it seems unlikely it’ll do that.)

It’s entirely rational for investors to be risk-averse, and to prefer a certain $80.29 over ten years to a very uncertain future direction for J&J’s stock price and dividends. After all, J&J stock could perform wonderfully well for the next nine years, increasing its dividend and generally being very healthy, only to suddenly plunge in 2020. Stranger things have happened, in this volatile market.

What’s more, Tilson’s numbers aren’t the only numbers you could use to perform this calculation. J&J’s average bond yield, for instance, is 3.2%, and some of its debt yields as much as 4.6%. Meanwhile, Reuters’s numbers put its dividend yield at 3.48%, rather than 3.8%. And in any case J&J is something of a special case, because its bonds are so very safe. It’s up there in the top four companies in terms of the difference between their dividend yield and their bond yield: the other three are Exxon Mobil, Nestle, and Procter & Gamble.

All that said, I do suspect that a long-term investor putting money into stocks right now is making a more sensible bet than someone putting the same money into low-yielding bonds. Stocks have unlimited upside, and tend to keep pace with inflation over the long term: they’re real assets, not nominal assets. And bonds have a lot of downside, as a quick look at the extent of the recent bond rally will demonstrate.

It’s impossible to invest these days without taking risk; I just feel that the risks in the stock market are more known and more priced in than the risks in the bond market. Interest rates will surely rise at some point. And when they do, you don’t want to be invested in bonds. Buy-and-hold investors, pretty much by definition, don’t have a strategy for selling the bonds that they’re buying right now. But are they prepared for possible future price declines? I don’t think so.

What’s the best way to monetize a blog? I don’t know how much John Hempton has made off this blog entry, but it’s surely up there in the annals of the most lucrative posts of all time.

Hempton found an obscure Chinese travel company which somehow managed to get itself a listing on the NYSE: Universal Travel Group. He tried to book travel through Universal Travel’s website, and failed. And after 5,500 words of explaining exactly why he was doing it, he shorted Universal Travel’s stock. Which turned out to be a great trade: the stock plunged by 20% today, entirely because of Hempton’s blog post.

Universal Travel put out a press release at the close of trade today saying that it “categorically denies all the allegations contained in the blog”, but not getting into specifics; there’s vague but ominous language about “aggressively pursuing all legal remedies against Bronte Capital and John Hempton”, but I very much doubt they’ll come to much.

Historically, short-sellers have been shadowy types; they like to publicize their findings, but they tend to do so behind the scenes, giving journalists information and having very long conversations off the record.

Hempton’s different in that he’s happy, on occasion, to make his allegations in public, under his own name. He doesn’t always publicize his shorts, even when he suspects outright fraud, but his blog does have enough of a following now that he knows he’ll be widely read if and when he chooses to do so. After today’s big payday, I reckon he might try the tactic more often.

The story of Universal Travel is far from over: if Hempton’s right about the company, and I think that he is, then the SEC and the NYSE are both going to have to answer some very pointed questions about how and why they allowed the company to get this prestigious New York listing in the first place.

But I do love the way that the blogosphere is moving markets. Reading a blog entry from someone with real skin in the game is often a lot more fun than ploughing through “objective” journalism from someone who isn’t allowed to invest in or short what they’re writing about.

Chart of the day comes from the Bank of England’s Andrew Haldane:

What you’re seeing here is the result of two different investing strategies. The red line is momentum: every month, you do one of two things. You go long the market when the market rose the previous month; or you go short the market if it fell the previous month. The blue line, by contrast, is value: you use a dividend discount model, and buy the market when it’s cheap, and sell it when it’s expensive.

Here’s the results, as presented by Haldane in his speech to the Oxford China Business Forum, in Beijing:

The speculator’s $1 stake in US equities in 1880 would by 2009 have grown to over $50,000. The fundamentalist’s same $1 stake would have fallen to be worth around 11 cents. Impatience would have trumped patience by a factor of half a million.

Gavyn Davies can’t think of any good reason why there should be any outperformance here, let alone anything of this magnitude. I’m inclined to agree with him: I can think of a couple of things which might be going on, but none of them feel particularly convincing.

Still, both of these strategies involve going both long and short. Stocks generally go up over time, so in general a strategy which goes long more than it goes short is likely to outperform a strategy which goes short more than it goes long. If the dividend discount model has even a modest overinclination towards concluding that stocks are overpriced, it’s going to get crushed: you don’t know pain, in the markets, until you put on a short position and watch the stock you borrowed crawl inexorably upwards.

For real people, long-short strategies are nearly always contraindicated, no matter how profitable they might have been in the past. And this chart only serves to underline that fact: the sensible, fundamentals-based strategy is disastrous, while the crazy crowd-following strategy, which ought not work at all, turns out to be highly profitable. Still, I would have loved to see a third line, showing the results of a simple buy-and-hold strategy. Sometimes the easiest things to do are also the most profitable of all.

Update: Many thanks to Jake, who has come up with a whole series of new charts comparing these numbers to a buy-and-hold strategy. Here’s one:

I love Kristina Peterson’s profile of Briargate, an algorithmic prop-trading firm (it’s an anagram of “arbitrage”) which has more or less given up on trading during the middle of the day.

High-frequency strategies like Briargate’s work best when there’s maximum liquidity, and that’s definitely during the first and last hour of the trading day. So instead of babysitting their computers at noon, Briargate’s principals go for long walks, or visit their children’s schools, or go out for pizza — and don’t even notice when something like the flash crash happens. (They were at the movies at the time.)

All of which makes me wonder whether we shouldn’t be presenting intraday stock charts a little bit differently. Right now, they invariably construct the x-axis so that every given unit of time (one minute, one hour, whatever) takes up the same amount of horizontal space. Underneath that you sometimes see a volume graph which shows you the important parts of the chart to look at.

Does anybody publish charts where the x-axis has a constant volume chart along the bottom, spreading out high-volume trading periods and skipping over low-volume periods relatively quickly? Is there a way of publishing data so that every tick, or every 1,000 shares traded, takes up an identical amount of space on the x-axis? The axis could still be labeled by hour or minute, it’s just that those labels would no longer be equidistant.

I’m pretty sure that such a chart would provide an interesting and fresh perspective on how stocks move. But of course it would be hard to generate in Excel, so maybe that’s why I’ve never seen one.

Sam Mamudi has found a new way to slice mutual-fund data, and the results are very interesting: the flows aren’t just from domestic funds to international funds, as we can see from the monthly ICI data, but also from actively-managed mutual funds to index funds.

Since the end of 2005, actively run U.S. stock funds have seen net outflows every year, totaling $369 billion, while indexed counterparts — not including exchange-traded funds — have seen net inflows of $112 billion, according to fund-industry trade group the Investment Company Institute.

I went one further, and had a look at the ICI’s data on ETF flows. After all, to a first approximation, all ETFs are index funds rather than actively-managed.

Here’s how the numbers break down: total actively-managed mutual funds, both domestic and international, saw a net outflow of $37.7 billion in 2009, and of $24.1 billion in the first seven months of 2010. Meanwhile, passively-managed index funds saw a net inflow of $22.9 billion in 2009, and of $22.4 billion in 2010 so far. But get this: equity ETFs saw net inflows of $69.3 billion in 2009, and another $21.4 billion in 2010 to date.

Those numbers aren’t publicized by the ICI: I had to calculate them using their spreadsheet of monthly ETF data. But if you add it all together, there was a net inflow into equities of $60.5 billion in 2009, and another net inflow of $19.8 billion in the first seven months of 2010. People aren’t pulling their money out of the stock market, they’re just pulling their money out of actively-manged mutual funds in general, and actively-managed domestic mutual funds in particular.

If you look at growth rates, the numbers are even starker. Actively-managed domestic mutual funds saw an outflow of $44 billion in the first seven months of 2010, which was 1.45% of their total value. Equity ETFs, by contrast, saw an inflow of $21.4 billion, which was 3.12% of their total value. If you go back to 2009, the numbers are -2.07% and +10.78%, respectively. Yes, in 2009, the net inflow into equity ETFs (I’m not even including bond or commodity ETFs, here) was greater than 10% of their entire year-end value. Mutual funds, it’s fair to say, never see those kind of net inflows.

This shift is only just beginning. There’s more than $3 trillion invested in actively-managed domestic mutual funds, compared to just over $1 trillion in domestic index funds and domestic equity ETFs combined. On the international side, there’s $1.2 trillion in actively-managed mutual funds, compared to $218 billion in international ETFs, and just $97 billion in international indexed mutual funds.

So in terms of long-term investments, people are still massively overweight actively-managed strategies. But they’re sensibly rotating out of those funds, and into passive ETFs. As that trend continues, and I see no indication of it slowing down at all, one can only expect that correlations between different stocks will continue to rise. And as correlations rise, of course, it becomes increasingly difficult to justify an active strategy.

ICI chief economist Brian Reid says that “considering historical investor patterns for the last 20 years, we are currently seeing weaker investor demand for domestic equity mutual funds than those patterns would lead us to expect.” Too right we are. And there ain’t gonna be no mean-reversion, either. That $3 trillion is going to end up reallocated, sooner or later. And if your business model is based on managing domestic mutual funds and getting a steady flow of new investments, you’re not going to find life easy going forwards.

One of the key tools used by fixed-income analysts during the Great Moderation was their beloved Monte Carlo simulation. They would take an instrument like a CPDO or a subprime-backed CDO, and they would run it through tens of thousands of possible future worlds. If it held up in all or nearly all of those worlds, then, presto, it got a triple-A credit rating.

But Welton Investment Corporation has a great little paper out showing just how unhelpful Monte Carlo simulations can be. They applied a Monte Carlo simulation to the S&P 500 over the past 50 years, plugging in its known return and variance. Using that, they compared the predicted number of large down quarters with the actual number of large down quarters. And got this:

Over the past 50 years, someone wielding a Monte Carlo simulation would expect 32 large down quarters, of more than 20%. In fact, there were 169. And they would expect no quarters at all with losses of 30% or more, when in fact there were 23. As for a loss of 40%, that simply never happens in a bell-shaped world. But it does — and did — in real life.

There’s nothing new here, as Welton notes:

It is worth establishing that the equity market’s tail risk signature is both well-known and persistent over time. Our analysis is not anomalous, and is easily replicated using any reasonably long period of historical market data. Second, it is also worth noting that this “tail risk” effect is not just confined to the S&P 500, nor is it confined to equities exclusively. Rather, this phenomenon is seen widely across capital markets and real assets.

But it’s also undeniable that a preponderance of stock-market investors don’t really grok how much risk they’re taking: they concentrate on the positive average returns, and tend to ignore the massive downside. Given enough time, some if not all of these losses can be made back. But just how much time do most investors have? And how much do they need, before investing “for the long run” starts to be a remotely safe thing to do?

(HT: Harris)

Update: My sharp-eyed readers had their coffee this morning; I clearly hadn’t. There’s something very, very odd with the numbers above: they seem to be cumulative, rather than additive. So while it might be true that there were 42 quarters with a loss of more than 20%, compared to a predicted 17, that’s a multiplier of 2.5x, not 5.3x. And it seems improbable, to say the least, that there can have been 169 quarters in the past 50 years with a loss of more than 20%, when 50 years is only 200 quarters in total. There must be a fair amount of mulitple-counting going on somewhere. On top of that, not all Monte Carlo simulations assume a normal distribution. So I have a call in to Welton, I’ll try and clear all this up.

Update 2: OK, I’ve now talked to Welton’s Chris Keenan, and have a much better idea of what’s going on here. First of all, these are rolling compound 65-day returns: the quarterly performance is calculated every day, not every quarter. So there aren’t 200 quarters in 50 years, there are about 11,000. Why would you do that? After all, who calculates their quarter-to-date performance on a daily basis? Very few people. But institutional clients, especially, pay serious attention to quarterly returns, which is one reason why they demand those numbers from money managers. And if you’re looking at stock market performance as a whole, it makes sense to get as many datapoints as possible, rather than cherry-picking 200 datapoints on the grounds of where they fall on the calendar.

In any case, Welton came up with this chart:

The basic idea is to measure how fat the real-world tail is, compared to the normal distribution. The real tail is the area under the light-blue line, left of a 20% cutoff. And that tail is 5.3x fatter — its area is 5.3x larger — than the area under the dark-blue line, which represents the normal distribution assumed in much of Modern Portfolio Theory, and which in turn is used to make a lot of asset allocation decisions. I hope that helps clear things up a little.

Update 3: I’m still getting pushback on my headline, and the way that I’m blaming Monte Carlo simulations rather than just normal distributions. That’s fair: the real underlying problem with the normal distribution is the normal distribution, and if you run a Monte Carlo simulation with normally-distributed garbage in, then you’re going to get garbage out. This paper didn’t need to use Monte Carlo simulations, but it did:

We created an “Expected” return distribution for the S&P 500 Index using standard Monte Carlo simulation methods based on a normal distribution assumption with inputs derived from actual S&P 500 data for the previous 50-years.

The inputs, here, were the return and the variance of the stock market over the long term. And using those inputs, along with a silly assumption that returns were normally distributed, ended up with a very bad model when it came to predicting the fatness of the left-hand tail.

It’s entirely possible to run Monte Carlo simulations which don’t assume a normal distribution. But the fact is that most people, when they look at the results of Monte Carlo tests, don’t critically examine the assumptions behind them. And it’s very easy to get blinded by Monte Carlo science: I, for one, took a lot of the CPDO fanboys at face value because I trusted them to be using good models, when in fact they were using flawed models.

So the lesson here, I think, is mostly that stock-market tails are fat. But there is a sub-lesson, too, which is that Monte Carlo simulations can be very dangerous, if they’re implemented by people who don’t know what they’re doing. Including the quants at Moody’s.

The lead story in Sunday’s NYT was by Graham Bowley, and it was quite alarming:

Renewed economic uncertainty is testing Americans’ generation-long love affair with the stock market.

Investors withdrew a staggering $33.12 billion from domestic stock market mutual funds in the first seven months of this year…

If that pace continues, more money will be pulled out of these mutual funds in 2010 than in any year since the 1980s, with the exception of 2008, when the global financial crisis peaked.

Small investors are “losing their appetite for risk,” a Credit Suisse analyst, Doug Cliggott, said in a report to investors on Friday.

The story was picked up by Josh Brown, who said that in fact things were even worse than Bowley made them appear:

This has absolutely nothing to do with “risk appetites”…

What you’ll find is that people are now eating into their portfolios and living on their investment capital, they would prefer to pay bills (ar at least keep bills at bay) than worry about speculating in a market that makes little sense to them anyway (rallies on bad news, anyone?).

The men and women behind these outflow stats have been burned twice in a 7 year period and now there are taxes to pay and small company payrolls to meet and credit card bills to reconcile. There are car leases and financing payments to be dealt with and, oh would ya look at that, the roof done sprang a leak again…

The people are making withdrawals from their portfolios because they are starving for cash.

It’s as simple as that.

Except, it’s not remotely as simple as that. On the same day that Bowley’s article appeared on the east coast, Tom Petruno was writing a very different article for the LA Times:

In the first half of this year, redemptions of stock funds by people who needed or wanted their money were large enough to nearly offset the $724 billion in gross purchases by new buyers. The result was a net cash inflow of just $9 billion to the funds, according to the Investment Company Institute, the fund industry’s trade group.

By contrast, redemptions of bond funds were much less than new purchases, leaving those funds with a hefty net inflow of $156 billion in the half.

So bond fund assets are growing rapidly, but they’re still significantly less than what’s in stock funds. Bond funds held $2.4 trillion as of June 30 compared with $4.6 trillion in stock funds.

What’s going on here? Are mutual fund flows negative, as Bowley has it, or positive, as Petruno has it? The fact is that Petruno is much closer to the truth, and Bowley seems to have cherry-picked the worst number he could find. And Brown doesn’t seem to be right at all.

All of the numbers being cited here come from ICI, whose statistics you’re welcome to browse yourself. But they don’t cover the period through July, which Bowley is talking about: Bowley’s only using estimated July figures, and the official ones won’t come out until next week. Petruno is using the official figures, which run through June.

But even so, they don’t seem to add up. Were equity-fund flows positive through June, as Petruno has it, only to turn sharply negative in July, as Bowley would have you believe? No. The key word to note in Bowley’s piece is the word “domestic” — if you look only at funds investing in domestic equities, they have seen net outflows this year. But if you look at all stock-market funds, including those investing in the rest of the world, the outflows were positive through June; if you add in estimated flows through July, they’re modestly negative to the tune of less than $5 billion.

So really, equity mutual-fund flows are more or less flat so far this year: inflows are roughly the same as outflows. And remember that all of these figures are the difference between two large numbers: in the first half of the year, for instance, $724 billion flowed into equity mutual funds, and $716 billion flowed out. Which numbers help put Bowley’s $33 billion number in some perspective.

What’s more, equity mutual-fund outflows are largely a function of retired people withdrawing their money from the stock market. They would normally be offset by the flows of working people who are putting their money into the stock market. But those people are increasingly moving away from mutual funds and towards ETFs. And if you look at the ETF data, there was positive net issuance of another $38 billion in the first six months of 2010 — significantly more than Bowley’s $33 billion figure. Sure, some of that will have gone into bond and commodity funds. But most of it will have gone into equities.

So the big picture is clear, although you’d never guess it from reading Bowley’s story: people are still putting more money into the stock market than they are withdrawing from it.

And the bigger picture is even clearer: people are saving more and more money, and investing it in the market more broadly. ICI was good enough to send me the estimated year-to-date figures for all mutual funds — not just equities but bond funds and hybrid funds too. Add them all up, and you get a whopping net inflow of $208 billion. Which would seem to put the lie to Brown’s assertion that “people are making withdrawals from their portfolios because they are starving for cash”. In fact, they’re investing their cash to the tune of hundreds of billions of dollars, and they’re withdrawing money from their risk-free money-market funds to do so. (For the first half of the year, money-market funds saw net outflows of $509 billion.)

As Petruno says, people still have a lot more money invested in stocks than in bonds. So if they want to balance things out a bit more, then their marginal monthly investments are likely to be weighted more towards bonds than towards stocks. That’s going to make bond funds see lots of inflows, compared to stock funds. But it doesn’t mean that people are pulling their money out of the stock market. They’re not.

I had a fascinating conversation on Friday with Harold Bradley, the CIO of the Kauffman foundation. He’s something of an expert on high-frequency trading, quantitative strategies, and the like, and he raised an interesting question: why isn’t the SEC banning ETFs which include small, illiquid stocks?

The question arises in the context of a stock market which is demonstrating more lockstep than ever: stocks are ever more correlated with each other, and instead of broad indices aggregating lots of different moves in different directions, as they did in the past, increasingly it’s the other way round, and stocks just move up or down depending on what the broader market is doing.

The rise of ETFs, especially in the day-trading space, surely exacerbates this syndrome. As ETFs tied to the S&P 500 get bought and sold in enormous volumes, arbitrageurs, many of them high-frequency automated algos, jump in to buy and sell the underlying stocks. It’s something that some people are worrying about, in that it cuts against the idea that the stock market is meant to allocate money efficiently between companies.

But when ETFs include small, illiquid stocks, the situation is even worse. Right now, the SEC says that 70% of securities in an ETF must be “actively traded”, or 50% if the ETF includes 200 or more securities. Which means that ETFs can have up to 50% illiquid stocks, which are relatively easy to manipulate.

Let’s say that you’re a predatory algo and you’re looking at activity in these ETFs with substantial holdings of small-cap stocks. When people are buying, you quickly load up on the underlyings; when they’re selling, you go short. Your activity will eat into the returns of the ETF, since you’re making it more expensive for the ETF to buy the stocks, and getting it a worse price when it sells. But more to the point, it will maximize volatility and room for manipulation in the underlying stocks, as well, while minimizing the useful information to be gleaned from their share price. If you buy straddles on these small companies — equity derivatives which pay off when volatility is high — then it’s easy to imagine how you can trigger payouts by playing around in the ETF space.

“We have a lithium battery ETF“, says Bradley. “These are designed for manipulation. What we’ve done is create derivative packages that give people the illusion they can trade small-cap stocks for cheap. Just because they’re in an ETF doesn’t make them liquid.”

I do think that a lot of investors like ETFs precisely because they have a certain degree of liquidity which is often missing from the underlying stocks. But I suspect that it’s even harder to create liquidity out of illiquidity than it is to create a triple-A credit rating out of junk-rated subprime securities. You might be able to credibly pretend that you’re doing it, but there’s a strong whiff of fakery as well.

Clearly the SEC is concerned about manipulation, since it put in place those 70% and 50% limits in the first place. But if limits should be put in place, why not set them at 100%? This is a genuine question, incidentally: I’m not saying that Bradley is right here. But I do think he’s asking an important question.

Update: Some very smart comments below, and be sure too to check out Izabella Kaminska, who does a great job of explaining the market mechanisms in English.

Why are stocks yielding more than bonds? The expected 2011 earnings of U.S. stocks are more than 8% of their current price, while bonds yield much less than that. Rob Dugger has an interesting explanation: the market is looking at fiscal deficits as far as the eye can see, and trusts the US government to close the fiscal gap over the medium to long term. And doing so will inevitably mean hitting corporate profits:

The higher taxes and spending cuts needed to reach fiscal sustainability will echo throughout the economy in millions of ways. Companies that are dependent on the current structure of spending and taxes will be hurt. Their earnings and balance sheets will be weakened. In a sense, fiscal adjustment costs are off-balance sheet liabilities of every US company.

This might also help explain why companies have been so conservative about raising their debt issuance, even as the cost of debt has plunged: they don’t want to add to their on-balance-sheet liabilities even as their off-balance-sheet liabilities, in the form of fiscal adjustment costs, are rising sharply.

The book value of the stock market — the value of its assets minus the value of its liabilities — has, on this view, been declining steadily of late, as the size of America’s liabilities has steadily risen. This is why people lump Spain in with Greece: while Spain’s liabilities are largely in the private sector and Greece’s liabilities are largely in the public sector, ultimately it’s the economy as a whole which is responsible for them.

Of course, we have no idea whether or how future governments might seek to achieve fiscal balance by reducing corporate profitability. But that very uncertainty is something all investors hate: it’s impossible to price in, or to hedge against. Which is why bonds seem — are — so much safer, and yield so much less than stocks.