Tuesday, August 03, 2010

Capital Constraints at the Margin

Macroeconomic Resiliance thinks that new firm entry means that banking, at the sector level, is not capital constrained.
A popular line of argument blames the lack of bank lending despite the Fed’s extended ZIRP policy on the impaired capital position of the banking sector. For example, one of the central tenets of MMT is the thesis that “banks are capital constrained, not reserve constrained”. Understandably, commentators extrapolate from the importance of bank capital to argue that banks must be somehow recapitalised if the lending channel is to function properly as Michael Pettis does here.

The capital constraint that is an obvious empirical reality for individual banks’ does not imply that bank bailouts are the only way to prevent a collapse of the monetary transmission channel. Although individual banks are capital constrained, the argument that an impairment in capital will induce the bank to turn away profitable lending opportunities assumes that the bank is unable to attract a fresh injection of capital. Again, this is not far from the truth: As I have explained many times on this blog, banks are motivated to minimise capital and given the “liquidity” support extended to them by the central bank during the crisis, they are incentivised to turn away offers for recapitalisation and instead slowly recapitalise by borrowing from the central bank and lending out to low-risk ventures such as T-Bonds or AAA Bonds.

So the bank capital “limitation” that faces individual banks is real, in no small part due to the incestuous nature of their relationship with the central bank. But does this imply that the banking sector as a whole is capital constrained? The financial intermediation channel as a whole is capital constrained only if there is no entry of new firms into the banking sector despite the presence of profitable lending opportunities. Again this is empirically true but I would argue that changing this empirical reality is critical if we want to achieve a resilient financial system.
I would disagree. A bank knows what its cost of capital is. A bank should be able to estimate what the profitability of a new loan should be. Therefore, practically speaking, the constraint is whether the marginal profitability of the loan more than compensates for the marginal cost of capital that needs to be set aside in order to make that loan. So a bank many have plenty of capital but still not make loans if it feels they will not be profitable.


Blogger Ashwin said...

Hi - the blogger at Macroresilience here. I think we're making different points here. My point is that there is no constraint on the total amount of capital that is available to be employed in banking just as there is no constraint for any other sector of the economy. You're arguing that the total amount of capital that will be eventually employed will depend on the marginally profitable loan opportunities out there, which is not inconsistent with my statement.

My point is just that due to the crony capitalist nature of the banking system where small banks are so disadvantaged compared to the TBTF entities, there is an effective capital constraint on the sector which needs to be removed. Once this "potential capital" constraint is removed, atleast we can be better convinced that all marginally profitable loans are indeed being made.

Let me give an example - in the UK, the sector has become even more oligopolistic with the Lloyds-HBOS merger in the crisis and mortgage rates for example are similar to what they were before the crisis and terms (loan-to-value etc) are far more stringent. I don't deny that some of this adjustment is logical and reflects the obviously crazy pricing pre-2008 but some of it also reflects the banks taking advantage of their oligopolistic position.

To put it in your language, the banks are pricing loans far above marginal cost of capital because they can afford to do so in the absence of genuine competition from new entry especially.


1:32 AM  
Blogger winterspeak said...

MR: Welcome!

I agree that the TBTF advantage dramatically decreases their CoC, putting both new entrants and smaller competitors at a disadvantage. I also agree that this problem is worse now than it was in 2006/7

I'm not sure this gives them oligopolistic pricing power. Say they offer a 30yr fixed loan at 5%, when they could offer it for 3%. A local credit union offers it at 6%, and could offer it no lower than 5.5%

A consumer is still better off going to the high-priced oligopolist in this case as it's cheaper than the even higher-priced fair market participant (who also has a higher cost structure). This is not the traditional dynamic for oligopolistic high prices, it's the standard competitive model where the marginal firm sets the market price, and more efficient firms collect rents through lower cost structures.

11:45 AM  
Blogger JKH said...


Not surprisingly, I agree with you in general here, as we have discussed this many times in the past.

The purpose of capital is to absorb unexpected losses due to risk.

Therefore, capital is a required structure of finance for risk taking.

And therefore, capital is a constraint on risk taking.

Almost all lending is a form of risk taking (e.g. lending to fiat currency issuers excluded, although even lending to a fiat currency issuer might include some interest rate risk which in turn requires capital, depending on the pricing structure of the deal).

So capital is a constraint on lending, not because it is lending per se, but because it is a form of risk taking.

At first, banks impose their own standards for capitalization. Then the regulators get involved, and their standards become superimposed.

The fact is that some banks, given regulatory constraints, then proceed to self-impose standards that are even more rigorous than regulatory minimums. E.g. this was the case through the crisis for all major Canadian banks.

Capital constraints translate in terms of both the quantum and pricing of risk, and the quantum and pricing of capital. These things must match up properly in order for banks to actually take on risk that they are considering in the proposal stage.

You point out very correctly (more or less) that the concept of “hurdle” is critical in the decision to take on risk.

So the constraint works there at the level of allocation of available capital. By definition, available capital is excess capital until it is deployed to take risk, whereupon it becomes utilized capital.

In addition to the constraint in terms of meeting the hurdle to utilize available capital, there is also the constraint in terms of the availability of excess capital per se – i.e. whether an individual bank in a position where it has the excess capital to even make the allocation choice for new projects.

Finally, the word “constraint” can apply in the aspect of a bank’s ability to generate new capital internally or externally.

The last two usages of the constraint idea were particularly impinging during the financial crisis.

So “constraint” is a multi-nuanced word when used to characterize bank capital.

However, I see little value in distinguishing the meaning of capital constraint in any truly fundamental way as it applies to the banking system versus an individual bank. And this is where it gets very interesting from an MMT perspective – because the appropriate and very meaningful distinction between systemic and specific dimensions when talking constraints is in the aspect of bank reserves rather than bank capital. The economics profession for the most part is apparently cursed by a failure to understand the difference in these two distinctions – i.e. the importance of the fallacy of composition in understanding the reserve system, versus its lesser role in understanding the nature of bank capital constraints, in my view.

6:04 PM  
Blogger winterspeak said...

JKH: Yes, and this is why I disagree with Macroresilience. I just don't think bank entry has much to do with the quantity of credit available right now.

And you can use the word "costraint" in many ways. During the day, a single bank might run up against a hard constraint if it makes loans right up to its capital requirements. It then cannot lend more unless it adds to its equity first. But longer term, a bank has time to do that, so how do you interpret the constraint?

I think marginal CoC is right as this creates the hurdle which gates individual loan decisions regardless of the capital level of a bank (so long as it has sufficient capital of course).

9:34 PM  
Blogger Ashwin said...

Winterspeak, JKH - this is very useful. Let me lay out my argument with some more clarity. My post aimed to challenge a popular meme which goes like this (Michael Pettis' post which I linked to in my post used this argument) : the banking system is undercapitalised due to loan losses and the household sector needs to pay in some manner to recapitalise the banks. So either you have the govt injecting equity into the banks or forcing them to raise equity in the markets, or you have the patented Greenspan-Bernanke method where rates are slashed and the incumbent banks just restrict themselves to "safe" lending to eventually recapitalise themselves. I think John Hempton had a post on why any bank no matter how insolvent can eventually recapitalise itself if its liabilities cost 0%. The argument goes that if the incumbent banks are not recapitalised, then credit to deserving and healthy borrowers will be restricted or more likely will be too expensive i.e. loans that would be made if the sector were adequately capitalised are not being made or are being made at exorbitant terms.

My point is simply this: the only reason why the above argument is empirically true is due to the significant barriers to entry in banking due to TBTF. How about this idea: Rather than shoring up the incumbents, we could have simply let them fail and put out a fast-track regulatory approval process for new "narrow banks". If there were sufficient profitable new lending opportunities out there, new players would have come in and filled the gap. The broader point is this: If entry and exit are free, then the amount of capital available to any industry is restricted only by the quantum of profitable investment opportunities in the industry. There is never a good reason to bail out an incumbent player, whether its Citibank or GM.

It's worth pointing out that depending on injection of capital into existing banks is tricky given the opacity of bank balance sheets and asset valuations in the midst of a crisis. Also we cannot depend on internal competition amongst incumbents except in a very slow manner as they all recapitalise via the Fed policy (We're seeing the first signs of this now as profit margins in lending and OTC derivatives are starting to shrink in the United States).

Winterspeak - let me take your example and restate it in my terms. Incumbent make loans to small biz at 5% when they could make it at 4% and a new entrant could make it at 4.5%. But there is no new entry because the new entrant knows that if it starts making loans at 4.5%, the TBTF incumbent can simply undercut it to say 4.25% and drive it out of business. End result - potential new entrant rationally decides to stay out. You can disagree with me that this is actually happening in the United States but it is the logical endgame as the sector keeps consolidating.

The counter-example to the US/UK is Germany where the solution to the TBTF problem is simply having a whole host of quasi-state guaranteed banks serving the small biz sector (the Mittelstand). Although I think such a system will ultimately end in tears, you can't debate that it works better than our crony capitalist system. Atleast all the subsidy provided by the state guarantee flow through back to the borrowers and depositors and bank profits are essentially non-existent given that even a small postal bank is close to state-guaranteed.

Can I also say that I seem to be using the word "constraint" in a fundamentally different meaning than your usage or the usage in MMT - I don't even disagree that weak loan demand may mean that even with entry of new banks, nothing changes. But atleast we'll know for sure that the argument of deserving borrowers being left high and dry is wrong and therefore, the Fed has nothing to offer in the current environment. This argument of small biz being ignored also gives Bernanke and Treasury an excuse to start all sorts of arbitrary lending programs, subsidies, guarantees etc.

2:54 AM  
Blogger Ashwin said...

Maybe it helps if I make the above argument more formal. Michael Woodford ( possibly the most cited and influential neoclassical monetary macroeconomist ) recently wrote a paper trying to incorporate financial intermediation into his macroeconomic framework. In this paper, he essentially moves away from his usual neo-Wicksellian approach to something that accepts a lot of the tenets of MMT. For one, he accepts that reserves are not a binding constraint on lending (for different reasons than the MMT logic but atleast the conclusion is right). But then to justify why monetary policy should be used to tackle "credit supply frictions", he introduces a key assumption on pg 11 that the "important limit on credit supply derives from the limited capital of intermediaries --- or, more fundamentally, the limited capital of the “natural buyers” of the debt of the ultimate borrowers. The market for the debt of the ultimate borrowers may be limited to a narrow class of “natural buyers” for any of a variety of reasons: special expertise may be required to evaluate such assets; other costs of market participation may be lower for certain investors; or the natural buyers may be less risk averse, or less uncertainty averse, or simply more optimistic about returns on the particular assets."

Essentially, I deny that there is any such limitation on bank capital - the concept of the "natural buyer" above is not empirically valid. The notion that the supply of equity capital is restricted because the potential equityholders are somehow different in their risk aversion or because they need expertise is just not true. The whole point of limited liability banking is that I can invest in a bank even though I know nothing about it and can hire the necessary expertise.

Incidentally, I haven't heard any commentary on Woodford's change of heart from the monetarists, probably because if you take his train of thought to its logical conclusion and if you tack on my point about capital, then monetary policy doesn't have much to offer right now!

4:13 AM  
Blogger JKH said...


Since governments can supply backstop capital every 80 years or so via TARP like arrangements, the only long run constraint for system capital is probably the explosion of the sun. That would be a constraint on central bank reserve provision as well.

5:17 AM  
Blogger winterspeak said...

MR: Yeah, but isn't that a CoC effect?

It's like saying "you cannot open any new supermarkets because the incumbent supermarkets are more efficient than a new supermarket can be, and will just undercut their prices and drive the entrants out. So the entrants don't bother coming in."

TBTF means you win. The cost of credit is still set by the market, but it's by the marginal lender with the highest CoC. The inframarginal lender uses his lower operating cost (CoC) to collect rent and recapitalize, not to gain marketshare by lower pricing.

7:08 AM  
Blogger JKH said...


I think I’m slowly getting the drift of what you’re saying.

Divide the banking system into two parts – the existing banks and the new entrants.

Given the calamitous economic and financial environment of late, existing banks have to deal with two major categories of uncertainty:

a) The interpretation of risk in their existing book of business

b) The interpretation of the adequacy of their existing capital positions, particularly in light of regulatory uncertainty (Basel, etc.) and then in the context of an unusually broad range of scenarios to be considered for risk and the effect those scenarios have on capital adequacy

I would describe that frontal assault of uncertainty as contributing to the importance of the capital constraint, as a general concept, on exiting banks. As mentioned earlier, there are many gradations in the conceptual hierarchy of the meaning of capital constraint. Risk and uncertainty mean that bankers are simply going to be more cautious in how they allocate capital to new business, given the backlog of issues they must deal with in their existing books. In some cases, it means existing banks are going to continue to generate capital internally without deploying much of it into new risk.

New entrants don’t have to deal with that legacy issue of risk and the uncertainty of capital requirements for an existing book of business. They are free to deploy without that burden. It’s very clean.

... continued

9:42 AM  
Blogger JKH said...

Continued ...

Looking back, the existing banking system could have in theory been transformed into its own “macro new entrant” with a major wash out of capital and credit losses through short term nationalization and recapitalization. I think this was always an extremely naive idea, for reasons I won’t go into here. But had it been implemented, the issue of capital constraints in the sense of legacy asset risk uncertainty and capital shortage would have been largely washed away as a structural dilemma.

The fact is that the existing banks have been recapitalizing through both external equity raises and internal generation, trying to reach some sort of equilibrium with the dangerous economy they face, the capital requirements for their existing books, and the capital requirements to support new risk taking. The very idea of recapitalization is a manifestation of the idea of capital as a constraint at the most general level.

I don’t think the distinction between capital for existing banks and capital for new business is somehow an indicator of the difference between system capital constraints and individual bank capital constraints. The distinction is in the capital requirement for existing on-the-books risk, whether at the individual or system level. It is more than likely that an existing bank has every bit as much access to new equity raise capital as a new entrant. The cost of capital may be higher to the degree that an existing book of business increases the business scope of uncertainty related both to the measurement of risk and the allocation of capita to support that risk.

I remember that Hempton post. I did not think it was perfect, but it was great in the sense that he hit on a basic aspect of banking that few bloggers would be inclined to recognize. One of the consequences of blogosphere focus on banking has been that the dimension of marked to market risk has been blown way out of proportion to its proper balanced place in the analysis of banking. I think this is because bloggers and their commenters fancy themselves to have the insights of great traders, because great traders presumably are known for quick thinking. It is a competition of quick thinking as much as anything. The global financial crisis has in large part been a crisis in marked to market accounting. Hempton saw through that. He knew that banks have a franchise income capability that transcends marked to market risk over the long term. So it is with the entire banking system and the entire economy. Bloggers are two eager to declare the banking system insolvent, without understanding that internally generated capital has been the core means for recovering from bad risk experiences for banks since banks were banks. It takes time to recover capital that way. New entrants don’t face that problem.

Is that getting any closer to your point?

9:43 AM  
Blogger Ashwin said...

JKH - Yes, that is pretty much my point. The problem with the slow recapitalisation is this: The large banks have taken advantage of the Fed's unwarranted generosity over the Greenspan-Bernanke era to push the envelope further and further in terms of risk which means that the recap takes longer and longer in each subsequent crisis. I don't see why the economy needs to wait for this process.

11:10 AM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home