What Is the Biggest Mistake When Evaluating a Fund?

There was a good piece in the WSJ about fund selection yesterday.  They interviewed some great advisors.   The answers were mostly the same – watch past performance, watch fees.  What surprised me was that no one mentioned the problem of benchmarking.   This is, in my opinion, the biggest problem in the fund world.

What’s happened over the last 30 years is that we’ve created a mass of various fund styles that don’t really do anything that much different than an index fund.  For instance, most “large crap growth or value” funds are really just closet S&P 500 funds.  If you ran the risk adjusted returns on these funds you’d notice that the correlations are extremely high in 95% of the cases and that they’re basically just fee sucking closet index funds.  They add literally negative value to your portfolio when compared with the low fee option.

I audit a lot of funds these days when I perform reviews for people.  And the benchmarking issue is the one that always comes up as being the most problematic when trying to justify whether a client should own a fund.  The fund is either benchmarked incorrectly or it doesn’t perform well relative to a highly correlated index.  To me, that’s the biggest problem when evaluating a fund.

Not only should we be ensuring that we’re benchmarking the funds properly (because the fund rating companies don’t always do it properly), but we should be ensuring that the funds we’re picking are actually a good value relative to the benchmark.  That requires some risk adjusted return calculations and some modestly sophisticated understanding.  But it’s no excuse for lazily looking at past performance, fees or other problem areas.  The devil is in the detail in most funds.  And sadly, when we look close, most of them are worse than their star rating states….


Got a comment or question about this post? Feel free to use the Ask Cullen section, leave a comment in the forum or send me a message on Twitter.
Cullen Roche

Cullen Roche

Mr. Roche is the Founder of Orcam Financial Group, LLC. Orcam is a financial services firm offering research, private advisory, institutional consulting and educational services. He is also the author of Pragmatic Capitalism: What Every Investor Needs to Understand About Money and Finance and Understanding the Modern Monetary System.

More Posts - Website

Follow Me:

  • http://www.conventionalwisdumb.com Conventional Wisdumb

    If you are going to index get the cheapest index fund possible. Even in that scenario you are still betting on underperforming albeit slightly.

    I agree with Cullen completely on this one, most funds add no value whatsoever but it’s even worse, if they are closet indexers they subtract value versus true indexing. If you compound the friction from these fees over time you get some pretty staggering reductions in long-term value versus pure indexing.

    The fees are used to pay for the salaries, bonuses, fancy furniture, and office spaces of these firms – the money has to come from somewhere :)

    Two other clues as to whether they are closet indexers – number of holdings and under/over weighting of sectors.

  • JC

    I have been thinking this for years. Never buy a mutual fund that consistently has a correlation of 90% or more to a benchmark even though they will market it as a good thing. After fees, and over time, it is much better to buy ETF.

  • Geoff

    “Large crap” lol

    Agreed that benchmarking is a huge issue. Part of the problem is that there are now so many categories that a fund can’t deviate too much or it risks being bumped into another category.

  • http://www.conventionalwisdumb.com Conventional Wisdumb

    Good catch!

    That was an epic Freudian slip if it wasn’t intentional.

  • LVG
  • Geoff

    Wow, that’s a huge sign of respect. Cullen’s also going to receive a nice Krugman bump in hits here. But it still doesn’t appear that Krugman really gets it.

  • http://brown-blog-5.blogspot.com Tom Brown

    Ha! That’s great! Thanks for posting LVG. That’ll give some much deserved exposure to Cullen (I don’t even care what Krugman’s article says… although I’ll go back and read it now… just for laughs).

    Congratulations Cullen!… if, indeed, you think that’s something you need to be congratulated about :D

    I think it is… it means Krugman reads Roche (at least once!). Very nice… that probably means the rest of the rabble out there will read Roche too.. like the MMists, and NKers Austrians… … they all read Krugman (even if they despise everything the mans says). Very very nice!

  • http://abetasoup.wordpress.com AlphaBet(a) Soup

    If you’re only viewing fees and history without considering the management team, the objectives of the money being invested, and how the fund has performed and should perform in volatile/normal/”different” markets… then are you really being an advisor or prudent investor in the first place? Probably not.

    Some funds go beast mode on their appropriate indexes. Those are worth the fees. ETFs are not the be-all end-all either, though.

  • http://orcamgroup.com Cullen Roche

    Eh. It’s not much of a victory if I can’t get him to understand my point. Which I don’t think he does. He’s referring to a world without excess reserves in which people use cash as though they don’t have to first have inside money or as though mass cash withdrawals are a sustainable real world phenomenon. The whole thing looks kind of confused to me. Oh well, I’ll try to put something together on it to explain my position more clearly and I hope he reads it.

  • http://brown-blog-5.blogspot.com Tom Brown

    I made a comment to Paul to that effect, but it hasn’t been approved yet apparently. I used this logic:

    bank deposits = L + B + F – C

    L=loans, B=bank Tsy’s, F=Fed Tsy’s, C=cash, thus:

    stock of money = L + B + F
    while bank reserves are either 0 (no QE) or F – C (QE)

    Thus showing how misleading it is to say “Banks lend out reserves.” The important part of that sentence is “banks lend” the reserves have nothing to do with it (it being the public’s stock of money).

    We’ll see if that get’s approved or not!


  • http://brown-blog-5.blogspot.com Tom Brown

    Hahaha! :D

  • dctodd27

    It is possible to have a stellar fund that massively outperforms, yet still has a correlation over .9 to its benchmark. Correlation just isn’t as useful a metric as people think it is. It is not useful in identifying closet-indexers (or rather it may be, but it also can lump others who are not closet indexers in with those who are, nullifying the usefulness as an indicator).

    The most important factor in determining your future returns is the price you pay for what you buy. Period. This is even more important than fees. Therefore, look for managers who conservatively estimate the intrinsic value of the companies they buy, and who buy at large discounts to those estimates.

  • http://brown-blog-5.blogspot.com Tom Brown

    Right, again the important part of “loan out reserves” is “loan out”: the reserves (in this case cash) has nothing to do with it. Though I do think he was touching on the B (bank held Tsy debt) in my equation for the money stock though:

    public’s stock of money = L + B + F

    So, yeah, increasing B or F (Fed held Tsy debt) also increases the stock of money.

    But cash really has nothing to do with it! And the sentence is super misleading. They should just stop at “Bank lend.” The rest doesn’t add any information.

  • http://brown-blog-5.blogspot.com Tom Brown

    Should be “Banks lend.”

  • Geoff

    TB, your example #11 is pretty f-ing brilliant. I re-created it in Excel just so that I could play around with it myself. It not only shows that banks don’t lend reserves, but also that, no matter how you change the numbers, the public’s equity is always equal to the govt debt. Granted, this is a simple example that doesn’t include other financial assets like equities, or real assets like homes, but it does show the sector balance equation in action.

  • Richard Rosso

    Cullen is correct.

    1). Benchmarking. Two out of every five funds I investigate are benchmarked improperly.
    2). Closet indexing. Amazing how many mutual funds are index funds in fancier, costlier clothing.

    Did like the commentary on paying attention to a fund’s objective – also monitoring for when it appears the fund is not adhering to said objective. This monitoring process was very helpful to me during the financial crisis when a “conservative” ultra-short duration fixed income selection deviated dramatically and I was able to sell for clients before severe losses occurred.

  • http://brown-blog-5.blogspot.com/ Tom Brown

    Geoff, thanks! I’m glad you like it. In the comments, I have a number of “upgrades” I plan to do to it… trying to decide if it’s worth putting those in #11 or leaving #11 nice and simple, and making a new one. For example it’s easy to add

    Ut = unspent Tsy funds. Mus be > 0, but can be > T

    G = intra-governmental debt (SS admin, etc.)

    Ug = unspent intra-governmental Fed depostits (if ther is such a thing!

    D (or “I” not sure yet) = difference between L and the public’s bank deposits dependent on L: this represents the imbalance between the two that occurs when banks charge interest and fees, etc. This one can be both +/- … since there’s a possibility (however small!) that the banks owe more to the public than vice versa.

    Then there’s required reserves (which is more complicated, since it’s not an independent variable at all)… not sure how to deal with that, but adding it can make the case stronger because I can say “banks don’t lend out excess reserves” … I can say that now, but I have to admit I’m setting RR = 0% in my assumptions, so they’re all excess.

    Sound like a plan?

    Yes, I would like to animate that too… or have a spreadsheet version. An animated one would allow you to adjust each independent variable within it’s limits with sliders or something to the side, and then in each cell in the balance sheet I’d have a colored bar representing the quantity of each item (e.g. public’s bank deposits) but the bar would be divided into each contributing component (L, B, F, etc.).

    Sound like an ambitious plan? :D

  • Geoff

    Check your gmail email.

  • http://brown-blog-5.blogspot.com/ Tom Brown

    Great, got it! Thanks Geoff. Unfortunately I did corrupt the example a little already by adding in unspent treasury funds (Ut), but I may remove that again and add it to a subsequent post instead.

  • http://alpholio.com Alpholio

    See the results of innovative benchmarking of funds at Alpholio.com.

  • Anon

    Beating the market by not following it ~5% of the time is extremely hard to do though – I’d say the correlation between a .9 correlation figure and closet indexing funds is much higher than .9, so it’s a good rule of thumb ;-)

  • Anon

    Cullen, I read a lot of Krugman commentary and in the vast majority of cases he is spot on. His skills in simplifying down arguments without losing the gist of it are unparalleled and refreshing.

    This was the first article of his iexcess s that seemed confused to me. I do think he is generous and intellectually open in replying to your point (he really didn’t have to!) – so I think you should try to come up with new arguments addressing Krugman’s apparent confusion specifically, instead of simply repeating past MMT arguments which he didn’t find convincing in the past.

    Allow me to offer a few thoughts and suggestions.

    Firstly, Krugman still appears to consider “bank reserves” in their historic form, which were a fixed percentage of deposits (mostly created through regular loans), where “excess reserves” were short term anomalies:


    Look at that data: excess reserves, prior the financial crisis, were trivial and were indeed anomalies. The correct resolution to past anomalies was probably indeed to “lend them out”. So Krugman’s intuitive response was correct, in historic context.

    But what happened in late 2008 was that the Fed effectively changed the meaning, definition and operative characteristics of bank reserves, including ‘excess reserves’.

    What happened? To fill in the growing void created by the collapsing shadow banking system, the Fed encouraged commercial banks to transfer most of their deposits to Fed accounts (mostly by paying interest on them). Those are not “reserves” anymore. The name “excess reserves” is actively misleading. Those funds are, in large part, all customer deposits.

    They are not “loanable funds” for the simple reason that they are mirror images of funds already lent out: they are the current results of all viable loans the economy can absorb.

    That they became “visible” in Fed accounts is not the result of some real, excess capacity of the economy that banks could and should “loan out” – they are just a new accounting technology that made them visible at the Fed as well. Nothing actually changed in the economy: the Fed and banks shuffled bits in computers.

    The Fed in essence created a national depository, at the commercial bank level. Such a new system, while useful for financial stability, does not in itself create a single new “loanable” dollar …

    And this is why saying that “excess reserves represent, in part, loanable funds” is not just misleading – even in context of a play model intended to simplify – but actively wrong and intellectually harmful.

    Likewise, claiming in the broader sense that “deposits arepresent loanable funds” and the whole “loanable funds market” is misleading and actively wrong/harmful because in a fractional reserve banking system where debt creates new money most of the “real economy” loans are created by banks loaning to private people or businesses “out of thin air”, conditional to ability to pay back the loan – not conditional on how many loans were already given out (=deposits at the bank).

  • Anon

    The sentence corrupted by autocorrect should read:

    “This was the first article of him in years that seemed confused to me.”

    (Sorry about that.)

  • http://brown-blog-5.blogspot.com Tom Brown

    Anon, you lost me with this:

    “the Fed encouraged commercial banks to transfer most of their deposits to Fed accounts (mostly by paying interest on them). Those are not “reserves” anymore. The name “excess reserves” is actively misleading. Those funds are, in large part, all customer deposits.”

    If you mean that bank deposits increased as excess reserve levels (ER) grew above 0 due to QE, I agree, but it’s not clear to me that’s what you’re saying. I illustrate what I’m talking about right here (case 2, not case 1):


    Reserves are all pretty much all the same, required or excess: Fed deposits held by banks. The Fed could drop the reserve requirement (RR) to 0% tomorrow and all the reserves would become “excess” and it wouldn’t change a thing except the volume of reserves moving on the inter-bank market to settle payments and transfer deposits (this volume would be diminished by something less than 10%).

    Here I demonstrate a simple example, one with RR = 0% and one with it at 10%:



  • http://brown-blog-5.blogspot.com Tom Brown

    BTW, I like your handle and avatar. :D

    … and one other thing. You could argue from a semantic position that reserves in the form of vault cash cease to be reserves when they leave the bank in a depositor’s pocket, but that’s a pretty weak argument. After all, cash does represent one of the three places reserves can go:


    However, my argument above:


    (and in my linked to Ex #11) demonstrates that the only thing meaningful about the statement “Banks can lend out their reserves.” from a money stock point of view is “Banks can lend.” The rest doesn’t matter, and is thus misleading regarding how reserves, excess or otherwise, can affect the stock of money (in the non-bank private sector).

  • http://orcamgroup.com Cullen Roche

    Thanks Anon. This is obviously an important discussion and one I’d like to put some serious thought into before responding. I really appreciate your insights.

  • http://rldinvestments.com/Articles/NFL2013Preview/DallasCowboys.html Ryan

    You have to do your own investigations these days, recent history has taught us that. There is so much outright lying and deceit to the public that if you don’t do your own research you’re putting yourself at risk. It’s an unfortunate truth of our society, but the bright side is it should make people more aware of their own finances. That is the one positive thing that we can all take away form this, more knowledge. If we blindly follow someone else because they have an impressive job title, or are licensed by the state, we know what can happen. If people spent as much time and energy planning and preparing for their retirement as they do when they go clothes shopping we wouldn’t have nearly as many people in bad financial spots. It’s all about where your priorities lie.

  • Anon

    I mean right now there’s around 2 trillion dollars held by commercial banks in Fed accounts (TOTRESNS data series on Fred2).

    Only a small portion of that are ‘required reserves’.

    Total checkable deposits at commercial banks (see TCD data series) are at 1.4 trillion dollars today.

    So to me this tells that all (liquid) deposits moved to Fed accounts – a central national depository. They are not “excess”, they are not “reserves” (they just happen to be accounted at Fed banks, earning interest) and they are not “loanable” either.

    They are in essence accounts transferred to the Fed, mostly resulting from QE bond sales. They are bond/deposit substitutes not controlled by banks in any “loanable” fashion – so they cannot spur economic activity either.

    There’s (at least) one issue I don’t understand: if banks can transfer deposits to Fed accounts and earn 0.25% interest there, why don’t they game the system via “fake” loans which create unused deposits which they transfer to the Fed, earning 0.25%, then are “paid back” and thus destroyed – and they keep the risk free 0.25%? Is there some mechanism that prevents this?

    I.e. why don’t banks create money and transfer it to the Fed to earn interest?

  • http://brown-blog-5.blogspot.com/ Tom Brown

    Anon you write:

    I’m with you on some of that. You haven’t convinced me that the term excess reserves is a bad one. I’m with you on the QE part.

    However, I’ll address your last paragraph. This is where I think you have a misunderstanding. Banks CAN’T transfer the deposits they hold as liabilities on their balance sheets to “Fed accounts.”

    Bank deposits are liabilities to the banks. Bank Fed deposits (1 or 2 components of reserves… the other being vault cash) are assets to banks. Soi what’s to prevent that scheme you outlined? Well there’s no way to make it happen…there’s no way to make the transfer. I’ll demonstrate. Take a careful look at the following example:


    Do you see how A loans money to x out of thin air? Now that requires A to hold some reserves, which it must borrow from the CB. Say x is in on your scam: how does that do the bank any good? How can this liability be transferred to the bank’s Fed deposit? Sketch it out for me on balance sheets if you have an answer, because I don’t see it.

  • Anon

    How about this – we don’t even need the Fed: say Evil Bank creates an off shore entity called Tricky Business.

    Evil Bank pretends to like Tricky Business’s business idea and loans it 1 billion dollars “out of thin air”, for a loan duration of 1 year. It ties up 100 million dollars in mandatory reserves, for one year.

    Tricky Business uses the freshly printed 1 billion dollars to buy 1 year treasury bonds. Those bonds mature in a year, earning 25 million dollars plus 1 billion dollars in principal.

    The loan is paid back to Evil Bank, Tricky Business closes shop. 100 million dollars in mandatory reserves are now free again.

    Evil Bank earned 25 million dollars on 100 million dollars, risk free.

    Where is the flaw in my thinking?

  • Anon

    To reply to my question: the flaw is that the moment Tricky Business buys the bonds, the money leaves Evil Bank who has to come up with 900 million in reserves to facilitate the transfer.

    So banks can only create money “out of thin air”, as long as it is used between customers of that bank. The moment it’s transferred to another bank the costs for the lending bank increase significantly, in terms of having to come up with full matching reserves. Right?

  • http://brown-blog-5.blogspot.com/ Tom Brown

    I like your response to your own question. Very nice. ;D

    I also like “Evil Bank” and “Tricky Business” Ha!

  • Nils

    Name five.

  • Anon

    Note the consequence of matching reserves being necessary to transfer funds to another bank: it means that for “local” use of money, banks create money out of thin air.

    For “global” – or even just national, interregional uses of money, banks are very much reserve constrained: they can only create new loans if they have (or could easily borrow) the reserves.

    And that is not just a 10% required reserve, it’s the full 100% reserve that is required to truly create new, transferrable money.

    So they need money to create new (transferrable) loans.

    Doesn’t this weaken the MMR (and MMT) arguments? Doesn’t this reintroduce the “loanable funds market”, in form of a “loanable reserves” market?

  • Anon

    To put it differently: reserves at the Fed are “mirror images” of money-equivalent deposits at banks: such as bonds that the Fed bought.

    Every single Fed reserve dollar originated from some real, liquid financial asset – not from “thin air”.

    So we could consider those trillions of reserve dollars as the “loanable funds”.

    New loans that create dollars out of “thin air” have to be backed by such reserves – I.e. by “loanable funds”, if they ever get transferred to another bank.

    Only money that never leaves the originating bank (I.e. only flows within that single bank) does not have to be backed by 100% matching reserves.

  • http://brown-blog-5.blogspot.com/ Tom Brown

    I don’t think this makes banks reserve constrained as long as the Fed continues to offer infinite reserves when needed and also continues to target the overnight rate (federal funds rate or FFR). Now with ER > 0 (like now) the FFT is forced down to the same value as IOR. But when ER = 0 (like prior to 2008) in order to control the FFR at a level above IOR (IOR was 0 then), the Fed had to add or subtract reserves (essentially providing the banks w/ exactly what was required… no more no less) using OMOs. In other words, w/ the Fed targeting the FFR and promising to facilitate payment clearing and deposit transfers and supply cash as needed (so our ATMs don’t run dry… which you’ve probably noticed they never do), then reserves don’t cause a constraint. If the Fed were to suddenly say “That’s it: no more reserves. Make do with what’s there!” prior to 2008, it would have been a disaster: ATMs would run dry, payment clearing stops, and the overnight rate would zoom out of control. Now if they did that… not as big a deal, since the system is flooded with cheap reserves… more than the banks need… anyway. But eventually it would cause problems.

  • Anon

    But if that is true, what keeps small and mid size banks from looting Fed reserves via creating huge, questionable loan portfolios and looting via bankruptcy? It’s easy to create a bank.

    I submit that they don’t because they cannot, they are reserve constrained.

    They can only create loans by attracting transfer deposits, or capital, or loan guarantees from larger banks (which are equivalent to capital to a large degree).

    Take for example the Fed discount window reserve lending facility. It’s not open ended: AFAICS as a bank you have to pledge real securies such as government bonds or (worst case) GSE mortgage bonds.

    In 2008 indeed the Fed gave open ended guarantees to the largest banks (after Lehman), but that was transitory and arguably justified. I don’t see where the Fed gives “easy money” currently. Do you?

  • http://brown-blog-5.blogspot.com/ Tom Brown

    Anon, you wrote:

    “But if that is true, what keeps small and mid size banks from looting Fed reserves via creating huge, questionable loan portfolios and looting via bankruptcy?”

    Funny you should use the term “looting.” Have you read this?


    It was written in the aftermath of the S&L meltdown of the mid-to-late 80s and early 90s, but I think it might apply more than ever today. Sure, what you describes is fraud and it’s criminal (though prosecutions have been sorely lacking in the past decade).

    Banks are capital constrained! … not reserve constrained. Most of the reserve “loans” from the Fed are though repos, not the discount window. A bank can buy a T-bond by creating a TT&L account (out of thin air) for the gov. Sure, when Tsy transfers that account to their TGA, then the bank needs to come up w/ the reserves, but it has a nice T-bond to repo. When the Tsy spends (and they’ll spend every $) those reserves end right back in the banking system, available to be loaned out on the interbank market.

    Here’s and example of capital constraints (which are constraints on BOTH sides of the banks’ balance sheets) and how it differs from reserve requirements (which only deals with demand deposits):


    Capital constraints are real lending constraints because they fold loan assets directly into the constraint. Read John Carney’s article I link to there at the top as well. It’s good.

  • http://brown-blog-5.blogspot.com Tom Brown

    Anaon, one other thing I forgot to mention: That is that reserve settlements take place in blocks, not instantaneously (from what I understand). This is the system I’ve hear described in more detail by Nick Rowe (who’s Canadian, and described how the BoC operates), but I believe it’s similar in the US. What I mean is that instead of like I show in my simple exmples, with person x transferring his deposit from Bank A to B, and thus A having to come up with the full amount of x’s deposit in reserves by the end of the day to cover the overdraft… instead what happens, is there’s almost as much activity taking place in the other direction from B to A (maybe person y, who banks at B, buys something from person z who banks at A). Thus the NET amount of reserves that need to be provided is a lot less. You can imagine that if traffic from A to B was exactly equal to that from B to A, actually no reserves would be required from the Fed.

  • Anon

    If you look at current U.S. bank capital levels:


    They are well above 10% – while the current capital requirements are 3%. (Even with Basel II/III they would be 6%.)

    So U.S. banks are not capital constrained currently. They could expand lending – if they wanted to.

    Regarding settlement in blocks: the batch netting at the end of the day does not change my argument: a bank, if it wants to create a significant new loan portfolio, created at a higher rate than other banks, faces a likely net “outflow” of those newly created above average funds.

    In other words: creating loans at about the same rate as other banks has minimal new capital and reserve costs, because other banks expand with a similar rate and it nets out, statistically.

    But if a bank expands loans faster, it sees an almost 100% cost due to reserve requirements on transfer deposits that flow out.

    So the banking industry is in some sort of game theory dilemma: first movers of loan expansion get “punished” via outflow of transfer funds.

    OTOH if for example a small bank attracts a 1 billion dollars time deposit from Warren Buffet, it can create 1 billion dollars of new loans (with a similar term structure).

    So the “loanable funds” market exists due to 100% reserve requirements on outflowing deposits (which your examples showed as well). To create loans, banks need capital (which are permanent loanable funds) or time deposits (which are time limited loanable funds). It can not create money “out of thin air” if that money can leave it.

    A bank might still decide to loan below available “loanable funds” if the loans are too risky at the margin: that is the situation today more or less.

    There are only two actors who can create new money in significant quantities while the economy is shrinking: the Fed and the Treasury.

    Banks can shuffle money around, but cannot expand it significantly beyond the current rate of expansion – due to having to come up with 100% matching funds when above-average deposits move out of the bank.

    So the “loanable funds” market exists and is a constraint to lending if a bank tries to go beyond the average loan rate. At or below the natural rate of loan expansion a bank indeed is not reserve constrained, as MMR/MMT claims.

    It fact this property of money cteation magnifies recessions: when the average loan expansion rate is negative, a bank that merely keeps the loan rate at zero will already face a (net) outflow. So the rational response is to shrink the loan rate.

    Regarding QE, it indeed has no effect on loan creation – other than the secondary channel of giving banks 0.25% interest on excess reserves (which they don’t pass on to customers AFAICS). This gives banks 2.5 billion per year per trillion dollar reserves. Not a significant effect macro economically.

    Does this line of argument make sense to you?

  • http://brown-blog-5.blogspot.com Tom Brown

    Anon, you write:

    “There are only two actors who can create new money in significant quantities while the economy is shrinking: the Fed and the Treasury.”

    “while it’s shrinking” … that seems to be the key phrase here. What’s cause and effect?

    Yes, I read through your post pretty quick, but I get the gist I think. But what about at the ZLB (where we are)? The cost of reserves is very low — if it were truely 0% (IOR = 0), would they even care if they had depositors? They’d charge them all for the privilege of banking… and thus drive a few out anyway.

    Regarding the capital requirements, Basal III is tougher than what we have now, I understand. Also, I picked up the example 10% CAR threshold from this:


    I thought there were a series of thresholds for both Basel II & III.

    Plus there’s CAMELS:


    Another thing to consider is that with just a few super large commercial banks, there’s a significant chance right away that loans stay within the bank (and thus few reserves are needed): not more than 50% perhaps, but if we’ve got 5 major banks, then 20% would be my first guess (w/o seeing any data).

    You’re making an interesting argument, but it sounds like a mathematical one, which is best done with some actual data and math. Not that I don’t trust you, but I always get a little queasy when word arguments go much beyond 0th order.

  • http://brown-blog-5.blogspot.com Tom Brown

    So with my final point there, what does a bank care if the FFR is truly 0% if depositors leave it? It has free replacement funds. Now there’s an interest rate risk of course. Is that what you’re getting at?

    If it’s not concerned with an interest rate risk, it looses one pain in the ass “free” liability (the depositors, with their various demands and need for ATM machines and bank cards, and tellers etc) and replaces them with nice clean 0% funds from the Fed, which accepts as collateral their MBSs: no need for expensive vaults or ATMs, etc.

    What am I missing?

    And some bank out there is going to have a bunch of ER they want to dump, so they’ll take less collateral (perhaps?) for the chance to earn 0.01% on it.

  • http://brown-blog-5.blogspot.com Tom Brown

    Anon, I’m not saying your argument is wrong… but these thought experiments are tough to work out.

    If one bank is loaning $ at a higher rate, sure, it expects there to be a net outflow initially I guess… but even then, this will ensure a glut of reserves at all the rest of the banks… which does them no good (if IOR = 0), so why not loan it back to a bank that can use it? Unless they’re worried that the expanding bank is taking too much risk.. what’s the downside?… Sure under normal conditions (ER = 0), reserve loans are probably more expensive than deposit liabilities, but if you’re aggressively adding higher yielding loans, I don’t see the problem… unless risk gets away from you. Plus eventually those deposits will come back. Why wouldn’t they? None of that really hurts your capital position which is just an abstraction anyway: it’s not like capital = money… it’s a relationship on the balance sheet. In fact you add certain liabilities when calculating Tier 2 (subordinated debts).

    I think risk is the big player here. In 2009 the banks were looking around at each other knowing that if the others were lying as much as they were about their asset quality, then there was no reason to trust anyone.

    Plus as a creditor, you’re taken care of 1st should the bank become insolvent.. well ahead of the shareholders anyway.

    So I see your thought experiment, but I’m not convinced that you can really describe it as a reserve constraint. Maybe the heightened levels of distrust are way more important. Reserves needs bring that to light.

    In this line of thinking, you’re definitely challenging my usual macro simplifications: to just imagine there’s a single commercial bank…. with lots of highly internally competitive regional branches perhaps… but just one in the end. (it makes life so much simpler to be able to make that assumption). Now if all the branches don’t trust each other’s book-keeping anymore, then things grind to a halt!

  • http://brown-blog-5.blogspot.com Tom Brown

    Here’s another place I saw 10% CAR:


    So, not the greatest source perhaps!… (I noticed they’re missing one there). But I’ve seen other’s use a similar figure (John Carney).

  • Anon

    I’m not trying to make (or hide) any subtler point – I’m simply not convinced (yet?) that the “loanable funds market” does not exist – even near zero (short term) central bank interest rates.

    Wrt. your “above average loan funds will flow back” argument: in most cases they will only flow back in proportion to the bank’s size. All banks are small relative to the banking sector at large.

    Wrt. your “but banks can borrow some of the other banks’ excess reserves”: yes, exactly, and they will (have to) borrow about 100% of the loan’s value. I.e. a 100% reserve mechanism, the loan was not created out of thin air in the end – it was in essence a reallocation of existing reserves.

    So “excess reserves” and the size of M0 matters. Right now they don’t constrain lending because there’s so much of them in a deleveraging economy, but after the Fed stops injecting reserves they’ll become a factor in banking again and banks will again attempt to attract deposits for their reserve effects, not to attract all the other transactions around a deposit account.

    Can you see any blatant flaw in my line of argument? I think it’s all fundamentally backed up by your banking examples, so we don’t have any microeconomic disagreements AFAICS.

  • http://brown-blog-5.blogspot.com/ Tom Brown

    Anon: it seem to me the “out of thin air” thing is real. Again I go back to my thought experiment of treating the banking system as just one big bank with lots of internal competition between branches. How does this bank buy anything from the private sector? It credits banks deposits. Sure it can use cash too, but mostly it just credits bank deposits and every time it does that it really truly does just create money out of thin air EXACTLY like the Fed does, when it credits Fed deposits. Sure, our megabank has to worry more about risk and solvency, but in the end that’s what it does: whether or not it’s buying employee time (salaries), paying shareholder dividends, paying it’s electric bill or rent or for office supplies or if it’s buying securities or loan documents from a borrower (i.e. it’s loaning money), or buying the cheap liability that is a deposit (by paying a pittance for depositor’s interest)… all that buying is the same: crediting bank deposits: i.e. creating money of out thin air. When viewed int that light, the movement of reserves doesn’t really matter.

    Same goes for how our mega-bank gets paid: it debits bank deposits: interest, feeds, fines, points, whatever: it’s all paid to the bank by allowing the bank to debit bank deposits (i.e. destroy money: a money which is by definition a liability to our one big bank). Of course it can alternatively accept cash payments too, but again those are a rarity (except of course for trading existing bank deposits for cash or vice versa, which happens all the time).

    I’m not convinced that view is fundamentally wrong.