The fallacy of composition misleads you into thinking that you can infer the property of the whole by the property you can observe in its parts considered individually. This fallacy leads you to think that what is true for the parts must be true for the whole.
In economics, one of the most popular examples of the fallacy of composition is the “paradox of thrift,” popularized by Keynesian economics. If thrift is good for an individual, it should be good for the economy as a whole. Wrong, because if everyone saves more, this will bring down consumption, cause aggregate demand to fall, hamper any economic growth and, paradoxically, a rise in individuals’ thrift would lead to less saving on the whole and harm the economy. Policymakers believing in the paradox of thrift, particularly in times of recession, would look in horror at people saving more while trying to find their way out of the hole. Yet, despite what we can argue using curves that sometimes saving is equal to investment and sometimes is not, sound capital theory reminds us of the basic fact that an economy needs a boost in real saving for capital formation to bring about a real recovery (not a nominal or a statistical one). For those critics of Keynesian economics, the paradox of thrift is not an example of the fallacy of composition but one of bad economics.
Recently, the fallacy of composition has been also widely argued as a theoretical flaw in public policies that explains the current financial crisis. As the argument goes, regulators were misled since they assumed that recommending what seemed right for a single bank should be right for the whole banking sector. Supervisors were also misled by this same fallacy: checking individual banks’ soundness made them infer that the system was sound as a whole. This misconception left traditional regulators and supervisors, solely focused on the safety and soundness of individual banks, ill-equipped to prevent the systemic collapse seen at the end of 2008. Thus, a fresh conception of “macro-prudential supervision and regulation” is needed: a system-wide, top-down approach to regulation and supervision.
However, does this fallacy of composition really apply to the banking sector? If we look at other industries, a fallacy of composition does not seem to be the case. For example, if all individual fishing companies run their business in a safe and sound manner, we can infer that the fishing industry is safe and sound on the whole. The issue is, what do we mean by running an individual fishing company in a safe and sound manner. Even if it were profitable for an individual company to go carelessly about polluting waters, fishing out of season and in non-fishing areas, we could not call this a safe and sound way of running an individual business, so it would be unfounded to infer that the fishing industry is safe and sound on the whole.
In the banking industry something similar happens and the fallacy of composition might not be obvious at first sight. The two distinctive activities of banks are cash and payment services and credit intermediation. It is not easy to grasp how individual safe and sound “cash and payment agents” and individual safe and sound “credit intermediaries” can become a source of global systemic collapse when compounded as a whole.
Fierce critics of how banking operates today would argue that there is no such fallacy of composition, but a big misconception of what a sound individual bank should be. These critics would question whether we can call “sound” institutions that systematically borrow short-term and lend long-term, hold low capital, and intermingle socially-essential “utility” activities with highly risky trading and credit operations. Banks, for them, are so unsound in their present form that they can only operate if they are heavily propped up by the State. Without central banks’ ability to print as much money as needed, and in the absence of State-funded implicit and explicit deposit guarantees and State-promoted “too-big-to-fail” doctrines, modern banking industry simply could not exist.
Modern banking theory, however, would have it that there is little we can do. Banks have a social function of maturity transformation. They are expected “to allow individuals ready access to their money, while at the same time allowing most of that money to be invested in illiquid assets” (P. Krugman). For this Nobel Laureate the way to go with this intrinsically but economically desirably unstable banking industry is through “Government backing — the 21st -century version of deposit insurance — plus regulation so that the backed institutions don’t abuse the privilege”. Fierce critics would argue that maturity transformation is not a function for banks to do but for the society itself. Time preferences of savers and investment projects of entrepreneurs simply fall into banks’ hands. Their essential social function is to intermediate credit according to savers’ time preferences and borrowers’ funding needs.
Modern banking theory would continue saying that bank credit has to remain “flexible enough” to promote economic growth. Nobody questions the fact that bank credit is vital for capital accumulation that promotes economic growth. The question is when “flexible credit” stops promoting economic growth and starts creating a global financial meltdown. Our fierce critics, following Mises’ and Hayek’s business cycle theories, would argue that enough is enough when credit expansion does not follow the sustainable path marked by real savings. Cobden Center’s Senior Fellow Prof Huerta de Soto sets all of this out in great detail in his must-read book “Money, Bank Credit, and Economic Cycles”.
Milder critics, on the other hand, would also argue that there was no such fallacy of composition, but a systematic distortion of how to assess individual banks’ risk. They accept current unstable banking characteristics and State support, but question whether criteria to assess individual banks’ level of risk were correct. For them, regulators, supervisors and bankers simply failed to include a system-wide perspective in the individual assessment of risk. As a result, systemic risk was not the consequence of a fallacy of composition, but rather of poor assessment of individual banks’ risk, which failed to consider the fact that the collective behaviour of financial institutions was an important driver of risk.
No doubt there are good reasons to change many of the current approaches to banking regulation and supervision, but maybe the fallacy of composition is not one of them. And we should not be oblivious to the fact that in the financial system its opposite, ”fallacy of division”, would have even more damaging effects if policymakers were misled by assuming that, if the whole looks sound, all individual banks should be sound as well.
By your classification I’m a “fierce critice”, but without disagreeing with maturity mismatching in general though.
The beginning of the recession in 2008 was connected with the behaviour of commercial banks. But, I think it’s continuation now has much more to do with central banks and governments. But, there may still be some zombie banks out there hiding their losses.
I agree, Current.
The current trend appears to be that banks are writing back the loss provisions previously taken, but that loan growth remains weak by historical standards.
This suggests that banks’ losses have been and are continuing to be paid for by central banks’ money creation, but that behaviour has changed for now, with banks preferring to run much more cautious balance sheet positions.
I think banks and potential borrowers are very worried about future inflation and interest rates. They don’t trust the central banks to hit inflation targets anymore because of the incompetence they have revealed and the difficulty they have had using normal instruments to control the interest rate. So, making long term loans is difficult. Borrowers risk the problem of deflation making the purchasing power of repayments rise, whereas lender have a similar risk from inflation. At present we’re in a rare situation where both are likely.
Comments are closed.