Thoughts on Banking
By Arnold Kling
This post returns to the deep topic of banking theory. I am very unhappy with the state of banking theory in economics. People make very strong empirical claims about what they call the risk-taking process based on shallow thinking about that process. I think that Mencius Moldbug is giving the wrong answer, but at least he is asking the right question, which is more than I can say for, well, just about the entire economics profession.1. No information asymmetries
If you assume away information asymmetries, then I think you have two choices. You can read and absorb the thinking of Fischer Black; or you can build an ad hoc model that is fundamentally erroneous. Essentially everything in the economics journals and in textbooks is in the latter category.
Let us return to a world where investment projects are fruit trees,which take time to mature and which involve risk. The risk, which might be thought of as a disease that harms trees, does not strike every tree identically. There is also an asset (fruit?) that can be stored risk-free instantaneously.
An efficient portfolio of tree investments will have optimal risk-return characteristics for every investor, regardless of risk tolerance. That is the Capital Asset Pricing Model, which says that the most efficient way to dial risk up or down is to change your weighting of the riskless asset and the efficient portfolio in your personal portfolio. If you are particularly risk averse, you hold less of the market portfolio and more of the riskless asset.
Personal time horizons do not matter. The fact that I might want to save for retirement five years from now means nothing. The issue is my risk tolerance. If I can tolerate lots of risk, I put a lot of my savings into the market portfolio. If not, then I use the riskless asset. Also, I can start out with a lot in the market portfolio, but if the market goes down I can re-allocate in favor of the riskless asset. That way, I can limit my risk dynamically over time.
Banking is a simple service. The bank keeps track of everyone’s accounts. Customers who have accumulated more wealth than they have spent have positive balances. Customers who have spent more than they have accumulated have negative balances. I don’t think of the people with negative balances as over-stretched consumers (although they could be). I think of them as tree-planters.
Every instant, the bank calculates an interest rate on all of its accounts. Interest is paid in fruit. Customers who have positive balances receive interest. Customers who have negative balances pay interest. The bank gets a service fee from its customers to cover the cost of keeping the accounts straight. The bank itself takes no risk. It is just an accounting firm.
Let me elaborate on the last point. Suppose that I am a fruit-tree planter with a negative balance. My fruit tree, which does not yet bear fruit, nonetheless has a market value that exceeds my negative balance. As time passes, suppose that the prospects for my fruit tree worsen. The instant that the market value of my fruit tree threatens to drop below the value needed to cover my negative balance, the bank forecloses on my fruit tree. I am forced to sell my fruit tree in time to ensure that the bank never loses money.
In fact, my wealth is not going to be concentrated in one fruit tree. I am going to have shares in the market portfolio of fruit trees. When the market portfolio falls in value, the bank will force me to liquidate my shares. People who still have positive balances will not have to liquidate their portfolios.
Think of the economy as having two intermediaries. One is a mutual fund that assembles the market portfolio of shares in fruit trees. The other is a bank that handles people’s accounts. Customers take money out of the bank when they buy consumer goods or when they increase their investment in the market portfolio. Customers put money into the bank when they earn wages or when they liquidate shares in the market portfolio.
The economy is subject to real business cycles. As the saying goes, stuff happens. The fruit trees suffer a lot from disease. Everybody is poorer, with the biggest risk-takers suffering the most. The bank does not fail. But the value of shares in the mutual fund drops.
Fischer Black’s model ends here. If you ask why people don’t all hold the same market portfolio, the answer is “I don’t know'” If you ask why real-world banks operate in ways that allow them to fail, the answer is “I don’t know.” If you ask why people write long-term debt contracts, the answer is “I don’t know.”
Fischer Black would argue that all of these observed deviations from his model are second-order phenomena. To a first approximation, you can think of the economy as following a real-business-cycle model. In the late 1990’s, rational investors thought that we needed to increase fiber capacity on the Internet by a factor of 1000 within three years. It turned out they were “way off,” in the sense that it took several more years for Internet traffic to grow to the level that it was assumed to require in 2001. The result of that, and other risks that were rational ex ante, was a real business cycle slump from 2000-2002. Similarly, today we may suffer a loss of wealth because people made some calculations about the return on housing investment that, ex post, turned out poorly. Stuff happens. Nothing you can do about it.
Fischer Black’s model can be defended as an “as if” story. Yes, the real world differs in many ways. But perhaps these are not important, and you can treat the real world as if it were following Black’s model.
Instead, what economists tend to do is bring in arbitrary forms of financial intermediation ad hoc. Nobody asks why banks are structured so that they can fail. We just start with the assumption that they are so structured, and go from there.
Black might say that bank failures are the particular way that real business cycles manifest themselves in our economy. They are second-order phenomena. The first-order phenomena are the unavoidable risks that come from taking reasonable bets that work out badly due to chance.
2. Information Asymmetries
An alternative approach, which is quite difficult, is to try to start with information asymmetries. Some people know more than other people about some things. The pattern of financial intermediation is determined by these information asymmetries.
Recall that in Fischer Black’s world, there is no reason for unstable banks to emerge. The banks that should emerge should be banks that do not fail.
It is amazing, when you think about it, that just about every economist has an opinion about how to prevent bank failures, and yet not one of those opinions is based on a theory of why banks that can fail emerge in the first place. They just start with the assumption that we have banks structured as they are, not as Fischer Black would predict them to be.
By thinking about information asymmetries, I hope to explain the mergence of the phenomenon of banks that can fail. Only then can one make justifiable recommendations about what would be a safer banking system.
In fact, my guess is that our financial system is as complex as it is mostly because information has become increasingly specialized. To anticipate conclusions before doing any analysis, my intuition is that the financial system cannot be “fixed” from the center for the same reason that the system is so complex in the first place. Information is too diffuse to be managed centrally. Trying to plan the financial system is as futile as Russia’s attempts to plan agriculture and manufacturing under Communism. But again, that is my intuition rushing ahead of the analysis.
My guess is that in our fruit tree economy, we will gradually add more and more financial intermediaries, each specializing in different types of information. The information pattern will in turn determine the pattern of securities that is traded.
For example, suppose that we start with an entrepreneur who knows more than most people about growing fruit trees in a particular location. That entrepreneur will want to plant fruit trees in that location. The entrepreneur needs funds. The entrepreneur could sell shares in his business, but nobody else knows what the shares are worth. To save on information costs, the entrepreneur chooses to borrow funds and pay interest. The lenders do not have to know exactly what the fruit trees are worth. All they need to know is that they are worth more than enough to pay back the debt. They are presumably worth more because the entrepreneur put in a lot of his own funds, as well as time, in addition to the borrowed money.
Over time, lenders start to compete with one another. A lender realizes that by identifying low-risk fruit-tree entrepreneurs, it can lend at a slightly lower interest rate than competitors are offering but make a more reliable profit. So the lender develops some capability for analyzing fruit tree businesses, although the lender will never know as much as the individual entrepreneur.
As the lender develops information about specific entrepreneurs, the relationship between that lender and those entrepreneurs starts to matter. If the entrepreneur has to switch lenders, the entrepreneur will face a higher interest rate, because a new lender will know less about the entrepreneur’s business. In other words, it is costly for the entrepreneur to continuously search for a new lender. Instead, the entrepreneur will prefer a long-term contract with a lender who understands the entrepreneur’s business. With a long-term contract, the lender and the entrepreneur share in the benefits of the lender’s familiarity with the entrepreneur.
The lender’s sources of funds do not necessarily want to engage in a long-term contract. The lender would prefer to be supplied with long-term debt, but savers may prefer to keep short-term deposits with the lender. Thus, the lender acts like a real-world bank, not a Fischer Black bank. If the particular fruit tree projects turn out badly, the lender may fail. Moreover, if customers believe that the fruit tree projects that this bank is backing are not going to do well, then customers may start a run on this bank.
A genuine fruit tree failure is a solvency crisis. A rumored fruit tree failure is a liquidity crisis. Given the information asymmetries, either type of crisis is possible. Deposit insurance can greatly reduce the risk of a liquidity crisis, but it cannot prevent a solvency crisis.
Government-run deposit insurance acts like an asset on the bank’s books that can be trusted by depositors. There are other possible assets. Private insurance could be one such asset. Reserves could be another asset. Highly-marketable loans (loans where the bank’s inside information is not terribly costly for other banks to verify) could be another asset that would reassure depositors, because such liquid assets can be tapped in order to stem a bank run.
As we are learning today, there are many types of financial intermediaries that are important. Some of them, such as Fannie Mae and Freddie Mac, suffered the equivalent of bank runs. In addition, some of them may be insolvent. It is probably not practical for government to insure the liabilities of all of these intermediaries. On the other hand, government officials fear the consequences of not insuring their liabilities.
The rules of the financial system used to be: certain bank liabilities are insured (deposits up to $100 K); other corporate liabilities typically uninsured; some liabilities unclear (Freddie Mac and Fannie Mae; large deposits or debt issued by banks). Under the new rules, a larger set of liabilities is insured.
I think that one consequence of that will be to reduce the competitive viability of intermediaries that issue uninsured liabilities. All of a sudden, you find yourself playing in a game against competitors who have a tremendous potential advantage, playing by new and unknown rules. Until things get sorted out, any intermediary that wants to survive, which includes every corporation in America, has to be searching for an angle to get under the government-insured tent.
Once again, I am rushing ahead to real-world analysis, when the real world is much more complicated than my simple model. I think that the key to making sensible policy recommendations is understanding how the complex financial system emerged. Only then will one be able to make sensible guesses about the consequences of policy actions.
Right now, as Lawrence H. White puts it, the popular belief is as follows:
1: Banking and financial markets are inherently unstable. 2: Government intervention into banking and financial markets can only stabilize (never destabilize).
This is what people want to believe. My guess is that economists who cater to that belief will have credibility. Those who doubt that belief will not.
Just as those who describe the current crisis is a villains/victims/heroes story have credibility. Someone like myself, who looks at securitization as a phenomenon that emerged from arbitrary capital requirements, is not going to be heard. There is no market for my beliefs (readers of this blog excepted, of course).