Paul Krugman conveys an idea in words better than anyone else regarding the topic that is at the heart of modern banking: maturity transformation. In simple terms, individual savers deposit their money with the banks. These funds should be available on demand, because the savers can withdraw these on demand [hence the term demand deposit]. But the banks are allowed to loan these funds to borrowers for a specified amount of duration agreeable between the bank and the borrower. In this post, Krugman writes clearly:
Like a lot of people, my insights draw heavily on Diamond-Dybvig (pdf), one of those papers that just opens your mind to a wider reality. What DD argue is that there is a tension between the needs of individual savers — who want ready access to their funds in case a sudden need arises — and the requirements of productive investment, which requires sustained commitment of resources.
Prof. Krugman is talking about a dilemma, similar to Triffin Dilemma, but at a much more basic level. Triffin Dilemma deals with the consequences of using a single country's currency as the world reserve currency (which gives the single country an exorbitant privilege). Prof. Krugman recognizes that the same medium used by both the individual savers and the people requiring funds for productive investment leads to a conflict between both parties. If they both have different goals, why not use a different medium? [For the interested readers about a monetary system thesis where medium of exchange is decoupled from store of value, please consider this link.]
Prof. Krugman doesn't get to that part though, in his discussion. Note also that how he refers to requirements of productive investment but ignores borrowing for non-productive needs (such as for e.g., buying a house because prices and rising). Not all borrowing is of high quality and not all borrowing is done for the feasible term structure. A borrower wants to borrow at the maximum duration with the lowest interest rate possible, but a saver whose funds get lent may not accept the same terms. There is a yield spread in this case (bid and an offer, there is no such thing as a monolithic price) and market dynamics help figure out the duration/interest rate details. But if a central bank mucks with this market signal, what happens as a consequence? The market will eventually run out of qualified borrowers. There happens a disco-ordination of natural social interaction when the key signal of interest rate is messed up.
What if we do listen to what hard money advocates want? As Prof. Krugman rightly points out, there will be more panic crashes. The times when US and the world was on a fixed gold standard was very painful for a number of reasons. When credit expands beyond what can be meaningfully serviced by the economy, the credit levels naturally will contract. But under a fixed gold standard, credit is denominated in gold which causes the contraction to be extremely severe. There will be too much collateral damage. Even businesses that were operating under a reasonable profit margin and less debt would fail. This is what happened during Great Depression I.
America had no major financial panics other than in 1873, 1884, 1890, 1893, 1907, 1930, 1931, 1932, and 1933. Oh, wait. The truth is that returning to gold is an almost comically (and cosmically) bad idea.
In the same post on outlawing banks, Krugman talks about a task that banks engage in a normal fashion, borrowing short to lend long. He writes:
Any arrangement that borrows short and lends long, that offers investors claims that are liquid while using their funds to make illiquid investments is a bank in an economic sense — and is potentially subject to bank runs.
But there is another operation which is mostly unnoticed in mainstream economic circles and which essentially renders quantitative easing as a fuel added to the fire of deflation. The Fed could be unwittingly contributing to more deflation while engaging in an attempt to devalue the dollar. Professor Fekete has written several articles on risk free bond speculation, where the speculators are engaging in the same activity of borrowing short to lend long. It is clear that borrowing short to lend long is at least a dubious idea which can have negative long-term consequences.
Let's take an example and understand the duration mismatch problem in a concrete fashion. I want to borrow $250,000 for a 30 year loan. Assuming that the banks cannot create deposits to fulfill a loan without any reserves, will any one even lend me that sum for such a long duration? Even if I cannot find a willing lender who can accept the duration and the sum, will the interest rate be acceptable to me (the borrower) so that the deal can actually happen? Probably not, the interest rate charged may be exorbitant depending on available credit conditions (when there is a natural limit on credit creation). If a central bank mucks with the interest rates and keeps them artificially low for too long, may be I could borrow (but I shouldn't under normal circumstances). This has created a mispricing of risk which goes undetected initially but a fault nevertheless in the overall system.
Through this discussion one aspect becomes very clear: a person depositing funds at the bank should get to choose their time preference on those funds (and the fraction that should be available upon demand). Secondly, a honorable bank should match this duration for a qualified borrower. Vetting the borrower and matching the duration is the bank's job. Even prior to deciding the time preference of the saver's portion of loanable funds, the saver gets to make an arbitrage: what fraction to save and what fraction to leave as time deposits. What if the medium to save was completely different from the medium to transact/invest/loan etc.? Does that resolve this conflict which gets created between creditors and borrowers?