Felix Salmon

ETFs start to underperform

Felix Salmon
Feb 21, 2010 06:45 UTC

Is this the beginning of the end of ETFs as an asset class? Ian Salisbury has an important story this weekend, saying that the average ETF underperformed its benchmark by 125bp in 2009, and even the monster SPY underperformed by 19bp. That’s more than twice its total expense ratio.

The problem is that when ETFs become very big, they become lumbering and predictable, and nimble hedgies know exactly what they’re going to do and when they’re going to do it. As a result, the smart money front-runs the dumb ETFs, which end up underperforming, sometimes by a very large margin: the $40 billion iShares MSCI Emerging Markets Index ETF (EEM) lagged its benchmark by a whopping 6.7 percentage points in 2009. That’s over nine times its total expense ratio.

So while it’s a good idea to avoid small ETFs, and to avoid commodity-based ETFs as well, even the biggest, safest ETFs are beginning to look as though they might have reached a level of size and popularity that makes them suboptimal investments. That’s sad, if true, because they were great while they lasted, and because there’s no real alternative out there.

But the fact is that there’s no rule of investing saying that there is always an easy and obvious investment strategy for people of relatively modest wealth. Investing involves taking a large number of risks, some obvious, some less so. And if ETFs continue to underperform in 2010 to the same degree that they underperformed in 2009, their repo income notwithstanding, then ETFs — which looked for a while there as though they really might be that rarest of animals, a positive financial innovation — might well turn out to be a grave disappointment for millions of investors who thought they could make a handful of asset-allocation decisions and then sit back doing little if any more work from then on.

We’re not quite there yet: as Salisbury points out, EEM is still outperforming its benchmark since inception, and it ouperformed in 2008. And for long-term investors, a single year’s underperformance shouldn’t matter a great deal. But if this turns out to be something newly endemic to the asset class, there might well be no cure for the problem — and that’s worrying, given how popular ETFs have become.

Update: On the other hand, if EEM is good enough for the Harvard endowment to have $388 million in it…


Felix, why isn’t the problem true with normal index funds?

Posted by ReutersRat | Report as abusive

FICO’s new businesses

Felix Salmon
Feb 20, 2010 20:36 UTC

The FICO people invited me to a press dinner on Thursday evening, where I learned quite a bit about their credit-scoring system — including the fact that the “free” credit score being sleazily shilled by Ben Stein is not actually a FICO score at all, and therefore even more useless than I’d thought. I also learned about their new Credit Capacity Index — a pretty useful tool, I think, which tells banks not how likely someone is to default on their present debt load, but rather how likely they are to default if you extend even more credit to them.

The bulk of my conversation with CEO Mark Greene, however, was about the way in which FICO scores are being made available to individuals. If you go to FICO’s consumer site, myFICO, you’ll be bombarded with offers for products like ScoreWatch ($100/year), Quarterly Monitoring ($50/year), and even Suze Orman’s FICO Kit Platinum (a one-off $50; there doesn’t seem to be a cheaper Gold or Silver version).

If you hunt hard enough on the website, you can find some useful free information on what goes into your FICO score, what doesn’t go into your FICO score, and how to improve your score. Here’s how that last page kicks off:

It’s important to note that raising your FICO credit score is a bit like losing weight: It takes time and there is no quick fix. In fact, quick-fix efforts can backfire. The best advice is to manage credit responsibly over time.

That’s kinda funny, because the losing-weight analogy is exactly the one I used when talking to Greene. If somebody is unhealthily overweight, then yes they do need to lose weight — primarily as a means to the end of becoming healthier. But it’s not a good idea to weigh yourself too frequently, especially if you end up doing unhealthy things in an attempt to lose as much weight as possible as quickly as possible.

Credit scores are similar, in my view: if you increase your overall financial health, by doing things like spending less than you earn and paying down your debt, your credit score will naturally rise. And it’s certainly a good idea to look at your credit report once a year, by visiting annualcreditreport.com and getting it for free, to make sure that there’s nothing factually untrue on it. For the overwhelming majority of consumers, that annual free credit-report download is all they should ever need or want when it comes to these matters.

But myFICO doesn’t seem to think that way:

You can get 1 credit report from each of the three major credit bureaus (TransUnion, Equifax, and Experian) once every 12 months from annualcreditreport.com. However, this site doesn’t provide credit scores, or more specifically FICO® scores.

What FICO’s saying here is true, as far as it goes: your annual free credit report does not include your FICO score. But the point worth bearing in mind here is that there’s really nothing that you can do with your FICO score once you’ve got it. If there’s a factual issue with your credit report, you can begin the long and arduous process of appealing it. But if the facts on your credit report are right, the FICO score is just something which drops out once you run those facts through the FICO computer algorithm.

Essentially, it’s very hard indeed to imagine a consumer who would get so much value out of knowing their FICO score — over and above the value they get from their free annual credit report — that they would be well advised to pay upwards of $15.95 to get it. All those consumer-facing FICO products might do wonders for FICO’s revenues (or they might not, I didn’t ask Green about that), but I don’t think they’re good for consumers at all.

So I’m glad that FICO is looking at another way of getting credit scores to consumers: a plugin for online banking sites. This is already being trialled by a handful of institutions: when you log in to your online bank account, a little box in the corner of the screen shows you what your FICO score is. Click on the box, and you’ll get useful and customized free information about how you might be able to improve that score; your bank will also probably try to upsell you some service or other. But for most users, this feature will simply be a useful and valuable aspect of their overall online-banking experience, which might even make them feel more well-disposed towards their bank.

There’s another thing that Greene said that FICO was working on, too: responsibility metrics. FICO officially frowns on the fact that employers, landlords, and the like obtain access to individuals’ credit scores and use those scores as a proxy for that person’s general moral upstandingness. After all, that’s not what FICO scores are designed to do. But at the same time, there is a correlation there: those landlords and employers might be acting in a sleazy manner, but they’re also acting rationally.

I’m a little worried about FICO getting into the business of trying to quantify moral probity: it seems to be a business rife with massive potential pitfalls. But then again, I suppose that FICO doesn’t need to worry about its own Rectitude Rating. It’s a public company, its only obligation is to try to make money for shareholders, right?

Still, if FICO wants to create a new product, I’ve got a great idea for them. It turns out that your FICO score is only an indicator of the probability that you’re going to default: it says nothing at all about the amount of money that banks are likely to be able to recover from you in the event that you do default. If FICO started selling a recovery-given-default rating alongside its main FICO rating, that would surely be even more valuable to banks than the credit capacity rating. After all, if a bank lends money only to people who were ultimately likely to repay their debt in full after going into default, it will end up making much more money than a bank which lends to people with the same probability of default but who are much more prone, in such a situation, to just walk away from their debts and never repay them.

It seems to me that a recovery rating would be a much more obvious and profitable business for FICO than either morality ratings or the consumer products that they’re pushing on their myFICO website. And would make the company much less disliked among consumer advocates, too. But maybe they don’t have the data necessary to put a recovery rating together.


Some Guy Told Me.

In 2008, when I asked FICO about its claim that employers use scores, a spokesman said that the company bases its claim on “anecdotal information gleaned from public sources such as published articles.”

Despite the consumer reporting agencies’ statements that they do not provide scores to employers, the agencies still claim that employers use scores. Media perpetuate and amplify the notion and scare the daylights out of consumers and legislators. Consumers buy the scores and legislators pass bills.

Here’s a quote from one of Experian’s vast array of websites:

“Credit scoring helps potential lenders, landlords, and employers quickly gauge an applicant’s credit history.”

Posted by GregFisher | Report as abusive

The NYT’s blogs are set to be paywalled

Felix Salmon
Feb 19, 2010 21:24 UTC

Arthur Sulzberger, Janet Robinson, and Martin Nisenholtz of the NYT all took the opportunity of hosting today’s PaidContent conference to talk at length about their paywall plans. Which makes it all the more surprising that their message was so garbled: when they weren’t simply refusing to say anything at all, they were giving three conflicting answers to the same question.

Nisenholtz did say quite clearly that he expected ad revenue to go up rather than down, which implied to me that that paywall was going to be pretty porous. And Sulzberger said that “we are not trying to eliminate ourselves from the digital ecosystem”. But when I asked about specifics, it all got rather messy. It started when I asked whether the NYT’s own blogs would be counted towards the quota, and Nisenholtz replied that “our intention is to keep blogs behind the wall”.

That shocked me: blogs rely on loyal readers who come back to read them often. But few blog readers are loyal enough to pay for the privilege of reading that blog. And if you’re someone who participates regularly in the Freakonomics comments section, for instance, you’re going to be very annoyed if you’re forced to buy a subscription to the entire nytimes.com site in order to do so.

My guess is that if Nisenholtz does this, a lot of the branded blogs on nytimes.com, including both Freakonomics and Paul Krugman, will simply leave and set (back) up on their own. It’s possible that the NYT digital team could quietly exempt those blogs from the meter, but it’s important with any system like this to be transparent about which pages count and which don’t, and carving out exceptions could quickly make things far too complicated to be easily comprehensible.

The moderator, Staci Kramer, then asked Nisenholtz whether that meant there wasn’t something very weird going on as a result: that if you follow a link to a NYT story from a NYT blog, then that counts towards your quota, while if you follow a link to a NYT story from any other blog, then it doesn’t. After all, Nisenholtz is on the record as saying that “if you are coming to NYTimes.com from another Web site and it brings you to our site to view an article, you will have access to that article and it will not count toward your allotment of free ones”.

And that’s where things started getting messy: Nisenholtz started talking about Google’s offer to cap the number of first-link-free stories that people can read, and seemed to say that links to NYT content from external blogs would not be free after all. (According to my notes, he said “if you link to us, then it’s counted” — but he might have misspoke, or I might have missheard.) He did seem to change his mind later and revert to the position that incoming links would be free — but at the same time, Sulzberger said that “we can’t construct a system that just tries to please the 5-7% of the audience” who come through the side door. That was a clear message to people like me that the traffic we drive doesn’t matter and that he’s not going to pay us much attention.

The facts, here, seem to be that a good 60% of the NYT’s most loyal users come through the home page, and that an enormous proportion of the rest come through Google. Facebook, Nisenholtz told me, accounts for much less traffic than Google does, the latest research from Compete notwithstanding.

All the while, I was sitting next to a couple of NYT staffers on the paywall project, who were at great pains to tell me absolutely nothing whatsoever, while also making it clear that they’re reading what I’m writing, that they’re listening to bloggers’ concerns, and that as they build the paywall from the bottom up, they’re going to try to avoid the obvious pitfalls.

What I conclude from all this is that the top brass — the people speaking at the lunch — see the big picture, where the overwhelming majority of content consumed is old-fashioned self-contained articles, and where most traffic comes through the home page, and where the broad outlines of how they want to structure a paywall are pretty clear. And I’m going to hold out hope that Sulzberger and Nisenholtz will give their underlings a stylized description of what they want to achieve, and that the underlings will try in good faith to deal with tough cases as sensibly as possible, instead of simply applying rigorous rules.

But if I were the Freakonomics guys, I’d still be asking for a meeting with Nisenholtz to get some reassurance that their blog won’t disappear behind a paywall.

Update: A NYT spokeswoman confirms to the WSJ that NYT blogs will be behind the meter. And adds this:

On so-called side-door enterants, she added that those people will have free access to that article even if they have exhausted their allotment of free ones. However, the Times might consider adopting a service from Google that would let the Times set a cap on Times articles arrived at from third parties.

Update 2: Video of the session is here; Nisenholtz says that “our intention is to keep blogs behind the wall” at about 20:50. The odd bit comes immediately afterwards, around 21:30, when he says that if an external blogger links to a NYT blog, “it’s not open, it’s counted”. But he seems to backtrack on that later.


So, I’ve got to subsidize Judy Miller by paying to read Paul Krugman? Such a deal!

Posted by lambertstrether | Report as abusive

Lessons from FT.com

Felix Salmon
Feb 19, 2010 16:55 UTC

Now that the New York Times has decided to move to a metered paywall system, a lot of attention at the PaidContent conference — at the TimesCenter, no less — is being directed towards Rob Grimshaw. He’s the managing director of FT.com, which has been a pioneer in the meter-model space.

Grimshaw has been in the news of late, announcing that he’ll close the Google loophole just a tiny bit, by stopping people from accessing more than five articles per day by taking advantage of Google’s first click free program. The new, more limited, program isn’t in place yet — Rob waved vaguely in the direction of the second quarter in terms of when it might happen — but in any case I don’t think it’s going to make much difference, since the five-articles-per-day limit is so much higher than the 10-articles-per-month level at which you’re asked to pay to continue to read the FT.

So why is he even bothering? He told me that the Google loophole was causing a bit of a “headache” within the FT, but I couldn’t work out where it was: I couldn’t imagine that subscribers or advertisers or the print side or anybody else was particularly unhappy about it. But it turns out that there was at least one major financial company which was pushing all of its employees to use the Google loophole, rather than pay for a subscriptions. And when the FT asked them what they thought they were doing, the company just said well, you left the back door open, so we decided to use it. (Better that than to risk a lawsuit by sharing passwords.)

That said, Rob was a big fan of Google in general, and said that he was very pleased indeed with the rate at which visitors from Google converted into being paid subscribers — by implication, that rate is much higher than for other sites driving FT.com traffic.

Grimshaw said that FT.com ad revenues have risen since he put the paywall in place, and he added that if the NYT paywall does indeed turn out to be revenue-neutral, he would consider it a failure. Increasing site revenues, he says, is the whole point. And he has another revenue stream coming up: come May, he says, he’ll start selling day passes to the site. “There’s a large proportion of our readers who are prepared to pay for content, are very happy to pay for content, but for whatever reason aren’t willing to make an annual commitment,” he said, pointing to the fact that on the print side, the FT’s daily circulation of 400,000 is double the size of its print subscriber base.

And although it’s outside his bailiwick, I also asked him about the U.S. print edition of the FT, which seems to be quite thin these days and to suffer from the fact that so many stories were written for a London deadline, five hours before U.S. reporters need to file. Rob said that the U.S. paper was profitable — although I have no idea what exactly that means, when so many of the ads come from global buys — and more to the point, that the FT needs a U.S. circulation in order to be able to sell those global ad campaigns.

As for the numbers, Rob said he was now up to 1.9m registered users of FT.com, and 121,000 paying subscribers, a rise of 21% in past 12 months.

Finally, I asked Rob about Lex, which is the main distinguishing feature of the premium version of FT.com, and which is currently headless as Jo Johnson runs for parliament in Orpington. Rob said that Lex needs to become webbier, which is undoubtedly true: it’s still got a silly allergy to linking, and it’s still very much working around daily deadlines rather than intraday speed. At the same time, however, no one wants to see Lex become just another blog: if there’s any value at all in Lex, it’s because the pieces are deeply reported. So it’s going to be very interesting to see whothe FT hires to replace Jo, and whether they decide to shake up the franchise much.


Who needs anything past the free meter when you have the free FT Alphaville which has the second best general financial blogger after you?

Posted by david3 | Report as abusive

Why Treasury shouldn’t regulate banks

Felix Salmon
Feb 19, 2010 14:54 UTC

Wow, Democratic Senator Chris Dodd really doesn’t want the Federal Reserve to be the main systemic-risk regulator — to the degree that his latest proposal gives that job to the Treasury Secretary, of all people. This is not a good idea, as Karen Shaw Petrou says:

“I don’t see how you avoid fundamentally changing the role of the Treasury Department as a member of the executive Cabinet,” said Petrou, managing partner of Federal Financial Analytics, which tracks regulatory issues for industry clients. “One would hope the Treasury would exercise its powers in a virtuous way, but this is not what Treasury is nor what it should be.”

If the Treasury secretary had a formal supervisory role, that official could use the position “to meet the political exigencies of the moment or for an upcoming election,” Petrou added.

On top of that, remember what we saw at the beginning of 2009: Treasury can be a very chaotic place when the White House changes parties. What’s more, substantially all top Treasury officials are political appointees: it simply is not the kind of place that you want to be setting long-term macroprudential principles. Give bank regulation to Treasury, and you’ll end up with a system where regulators care more about banks’ presence in Iran and Cuba than they do about their leverage ratios.

But it gets worse:

During negotiations with Dodd this year, Sen. Richard C. Shelby (Ala.), the ranking Republican on the banking committee, suggested having the Treasury secretary lead the council. Shelby viewed that structure as preferable, in part because the Treasury secretary has a higher international profile than most regulators. He also views the Treasury secretary as more accountable to Congress.

I suppose it makes sense that a member of Congress would want to have anybody and everybody maximally accountable to Congress. But that doesn’t make it a good idea. Regulators are like judges: they work best with a minimum of political interference. We saw that with the regulation of Fannie Mae and Freddie Mac: it got hijacked by Congress to the degree that the nominal regulator was to all intents and purposes powerless to do anything but that which Congress demanded. And the consequences were disastrous.

So this is yet another reason — as if one were needed — to be pessimistic about the prospects for financial reform. I’m becoming increasingly convinced that if anything passes the Senate at all, it’s going to be so weak as to be largely pointless. We’ve wasted our crisis.


Felix Salmon
Feb 19, 2010 04:18 UTC

Paterson is even more useless than I’d thought — NYT

Crap tax law implicated in Austin terror-suicide: it’s not the IRS’s fault, it’s Congress — NYT

Wherein Blodget turns my tweet into a fully-fledged CONTEST — TBI

Valentino vs. September Issue — Black Von

Go read Paul on security theatre. Hilarious — Kedrosky

Must-read piece in the American Conservative on hispanics and crime. Fantastic stuff — AmCon

The top 400 taxpayers all earn at least $138m per year. Their average tax rate? 16.6% — WSJ


Mega: Do you disagree with Unz’s conclusion, or are you just nit-picking his data? The article explains quite clearly why it’s tough to find decent data on Hispanic crime levels (often reported as “white”, difficult to measure illegal immigrant population). Given the constraints I thought he did a nice job coming up with some hypotheses.

I’m also not really sure how having a lot of Thai families running restaurants dissuades Hispanic youth from selling drugs or engaging in violent crime. You’ll have to explain that to me some time.

Posted by najdorf | Report as abusive

Adventures in Fedspeak

Felix Salmon
Feb 18, 2010 22:11 UTC

The news of the moment is that the Fed has raised the discount rate to 0.75%. So why does the headline announcing the fact read like this?

Federal Reserve approves modifications to the terms of its discount window lending programs

The first two paragraphs of the press release contain precisely zero news; and in fact it’s not until the very end of the release that it actually talks in English about “the increase in the discount rate”. The actual news is announced like this:

The changes to the discount window facilities include Board approval of requests by the boards of directors of the 12 Federal Reserve Banks to increase the primary credit rate (generally referred to as the discount rate) from 1/2 percent to 3/4 percent. This action is effective on February 19.

What purpose is served by this obfuscation? Why can’t central banks in general, and the Federal Reserve in particular, communicate in English? Alan Greenspan famously said in 1988 that “if I turn out to be particularly clear, you’ve probably misunderstood what I said” — but that was then, and I thought we were moving away from that kind of silliness.

Ben Bernanke has shown that clarity is a virtue; his underlings in charge of drafting announcements should have gotten the message by now, instead of reworking it into incomprehensible gobbledygook.


What gives you the idea we are moving away from Fed speak nonsense? We have coddled these over paid idiots for years as they introduced theory after theory that were designed mainly to transfer wealth from the tax payers to Goldman Sachs.

Bernake is guilty of fraud as is Geithner and Paulson, until we bring the criminals to justice instead of hanging on every lie they tell.

Posted by jstaf | Report as abusive

Cash on Delivery Aid

Felix Salmon
Feb 18, 2010 20:47 UTC

I like Nancy Birdsall’s idea of “Cash on Delivery Aid” a lot. Instead of spending money on building schools or hiring teachers or any other means towards an educational end, just pay for outputs instead:

We propose that donors offer to pay recipient governments a fixed amount for each additional unit of progress toward a commonly agreed goal, e.g. US$200 for each additional child who takes a standardized test at the end of primary school. That is, the donors pay “cash” only upon “delivery” of the agreed outcome. The key features of this proposal are: (1) the donor pays only for outcomes, not for inputs, (2) the recipient has full responsibility for and discretion in using funds, (3) the outcome measure is verified by an independent agent, (4) the contract, outcomes and other information must be disseminated publicly to assure transparency, and (5) this approach is complementary to other aid programs.

If a prize-based approach is a good idea for things like vaccines, why can’t it work with education, and other areas with clearly-definable and measurable outputs?

To spur government action and allow for civil society to participate in bringing about change, each COD Aid contract (as designed) would rely on a single, measurable goal; e.g. children completing primary school or households with access to clean water. Choosing an indicator that is simple, central to people’s lives, and easily auditable, Nancy says, is critical to the success of a COD Aid model.

The problem with this model is I think at the donor end: the government can spend the money on anything it likes, from kickbacks to Kalashnikovs, and people — and even governments — donating money for early-childhood education or clean water don’t like risking their money being spent elsewhere. But Nancy has a 100-page book coming out on the subject; maybe that might help to change some minds.


1. Suppose over 5 years there are many other donors who also contribute aid (up front) for the same purpose – a highly likely scenario. Will COD still be paid?
2. Suppose the Government meets the target by starving other important programs of funds – so as to get the cash – is that acceptable? Will COD be paid?
3. Suppose after a time all donors start using COD. Cannot a Government after a time simply juggle payments so that it is in effect being paid up front and benfiting from a COD flow
4. A single Indicator?. Very difficult to find a single indicator that refkects results.
5. What is the incentive for either a politician or a civil servant or even say a school headmaster to sign on to perform tricks for the donors in return for cash in five years time? Time consistency problem.
6. Hands off? No monitoring? COD program is to be audited annually by an independent auditor – what the difference?
7. Unintended consequence – COD could well be regarded as an expression of donor distrust – holding recipients accountable for years without payment or with uncertain payment
8. COD in the end does not seem to do anything about the accountability issue that is different from any other performance based funding.

Posted by wordsmith1 | Report as abusive

Why car tracking isn’t a privacy issue

Felix Salmon
Feb 18, 2010 16:33 UTC

Last week, weighing in on a miles-travelled tax, I said that “there really is something quite creepily Big Brotherish about trying to track every single vehicle in America”. But then I heard from Bern Grush of Skymeter, and he’s persuaded me that you don’t actually need to make tracking information available in order to tax miles travelled.

Under his system (and of course he has a system capable of implementing this), anybody who’s pre-paid for their miles will simply see those miles essentially erased from their tracking device as they’re driven — along with the money leaving their pre-paid account. If you pay after you drive the miles, at the gas station, for instance, then the tracking data gets erased then and there.

Of course, you have the option to retain and not erase the data, if you want to keep it for your own records. But if you do that, there’s always a risk that someone could subpoena it.

In a Please Rob Me world, then, where Federal authorities are pushing to be able to track your cellphone, the privacy issues associated with a miles-travelled tax are probably the least of our privacy worries — so long as they’re very clearly articulated, and so long as the default settings are for absolute privacy. If you want to worry about people being able to track your movements, either in real time or in retrospect, then you should probably worry much more about your GPS-enabled phone and your FourSquare checkins than about any tracking device in your car.


mattmc: Yessir, put that way, VMT is stupid. But that is the fault of a very poorly chosen name for the time-distance-place charge, which the Europeans (whom we cannot possibly copy for fear of looking like pansies) call it. Reminds me of the Johnny Cash song “A boy named Sue”.

But that is not the intent of the VMT charge (note, I said intent). The intention is that distance is weighted by where, when and what you drive. So your country mile with your Tesla will be way way cheaper than my Manhattan mile in my Escalade.

Raising the gas tax is plain useless (indeed it is stupider still, than the common misunderstanding of the VMT charge). That would be taxing the thing we want you to stop using (gas) in order to pay for the infrastructure we need (roads) to allow you to drive the thing we want you to start using (EVs). How stupid is that?

If you don’t see that imagine we decide to fund the entire US medical system on tobacco taxes (since we’re making lousy progress with any other idea), and then we run out of tobacco? Who’s going to operate on your prostate, then?

Posted by BernGrush | Report as abusive