Currency increase section upated through April 17, 2000, with January and February&March comments added.

My reasoning as to why the answer would be "no" seems to have proven out. I've written a short paper "The Banking System and Y2K: a Retrospective" as a followup to my "The Banking System: Will It Survive Y2K" paper presented at the
Y2K Before and After Conference at George Washington University in September, 1999.

Will Bank Runs Shut Down the Banking System?

You may have seen a newsletter ad proclaiming:

"A bank run like no other will bankrupt banks all over the world in 1999"

Let's consider how likely bank runs resulting from the Y2K problem are to shut down the system.
First, some background points:

1. Newsletter advertising tends to be overdramatic to attract attention and subscribers. (In this case, the newsletter editor does believe a widespread bank run to be at least very likely.)
2. The editor has a strong bias against fractional reserve banking. This may color his analysis.
3. The editor has been accused of trying, through the advertising and other means, to cause the bank runs he is predicting. There is some plausibility to this accusation since he believes the present banking system to be immoral and would like to see it gone.
4. The underlying premise of the forecast is correct: if all (or even a significant part) of depositors demanded their deposits in physical form in any short period, this would shut the banks down. [link to description of fractional reserve banking?]

The logic of the forecast of banking collapse is straightforward:
1. Banks are in trouble because of Year 2000 problems.
2. Depositors will become increasingly aware of the problems as 1999 progresses.
3. Not being willing to take a chance on their bank not being able to function in 2000, depositors will withdraw their funds in physical form.
4. As these withdrawals accelerate, the banks will not be able to meet them and thus will shut down.

There are thus three key elements to consider:
1. The extent to which banks are in trouble.
2. The degree of awareness among depositors.
3. The action taken by depositors.

I have covered the first point elsewhere on this site. Some banks will not be compliant by the end of 1999, but a substantial majority will. Not all non-compliant banks will fail in the technical sense.

Awareness of the Year 2000 problem was much slower to spread than aware programmers expected (see e.g. discussions of this on comp.software.year-2000). As awareness has spread, the implications for banks may not have been immediately apparent to many people. However, some surveys indicate that this may have changed. A survey conducted for the FFIEC by the Gallup organization and reported in June showed the percentage of Americans who had heard "a great deal" about Y2K had increased from 39 percent in December, 1998, to 52 percent when the survey was done in March and April of this year. Such surveys are helpful because it is perceptions that matter here. It is possible that the effect of the perception of bank problems could exceed the effects of actual problems.  This factor will need continued monitoring. (see below)

The action taken by depositors seems likely to be of manageable proportions because of:
   1. The concern among depositors is still limited. The Gallup survey reported that while 62 percent said they would probably or definitely take out some extra cash, only 14 percent said they would probably or definitely withdraw all deposits.
   2. The natural inertia that keeps even aware people from taking action. I expect that even among those who say they will definitely withdraw all deposits, not all actually will.
   3. Reassuring statements by banks and banking authorities are plentiful and will be believed by some depositors. Coverage by the FDIC is being emphasized. Examples include a brochure and statement stuffer made available by the FDIC as well as other material on its web page.
   4. Non-physical withdrawals by some depositors. Bank runs can take the form of adverse clearing balances as depositors shift funds to other banks. This was the case with the Continental Illinois failure. So long as some depositors think some banks are compliant, a portion, possibly substantial, of withdrawals will be in this form (though the Gallup survey indicated only 17 percent saying they would probably or definitely do this). The Federal Reserve can easily handle this type of run. It has established a Century Date Change Special Liquidity Facility as an additional means of helping banks cope with withdrawals.
   5. The Federal Reserve has been stockpiling currency in anticipation of increased withdrawals, increasing the normal holding of $150 billion to $200 billion. .
   6. In the extreme case, regulatory authorities will limit cash withdrawals from financial institutions.

I have begun tracking the rate of increase in currency in circulation.  There was only a very slight increase in the year over year rate of increase (from slightly under 8% to slightly over 8%) in late 1998 and early 1999 compared to most of 1998.  However, in March it began to exceed 9% and in May and June exceeded 10%. It  leveled off in this range, then dropped slightly in September, crept back up in early October. It fluctuated slightly around 10% into the first of November and has been creeping up some since, reaching almost 11% by the end of November,   around 12% in December, an jumping to 13 1/2% the last reported week. The rate peaked at over 15% the first week of January, 2000, which included the last few days of
December. This increase beginning in the spring and again in the past several weeks may reflect some additional withdrawals because of Y2K concerns, but is still far short of putting a strain on the system.  I've posted a chart of the rate of increase to give a visual picture and will try to update it every week.  (There is a two week lag in the availability of the data - latest posted is Apr. 17, 2000.)
(Additional January note: in seasonally adjusted dollar terms the currency supply went from $512.5B the week of Dec. 13 to $515.6B the week of Dec. 20 to $522.1B the week of Dec. 27 and $531.9B Jan. 3 (which included the last of December).  Because the figures are seasonally adjusted, this means there was about $19B withdrawn over the three weeks above the normal increase for this time of year.  I think this moderate rate of withdrawals reflects points 2 & 3 I made above. There was a slight decrease in the amount in circulation in the week of Jan. 10, about a $5.5B decrease the week of Jan 17, and a further decrease of about $2B through the week of Jan. 31, indicating that some of the withdrawals are starting to flow back in. )
(February note: There was a reduction of $2.5B the first half of February and a further slight decrease into March .)

In view of the above, I do not believe the banking system will shut down because of Y2K related bank runs. (This is not to say that it cannot happen. There is, in fact, no way to prove that it is impossible.) Nonetheless, preparing for your bank not being OK seems a wise precaution since at this point there seems no way to tell if it will be.   I have some suggestions for this.
(Given some of the comments I have received, I perhaps should note that my thesis in not that there will be no bank runs, but that bank runs will not shut down the banking system.)
(January note - so far there seem to almost no bank problems at all, so the precautions turned out to be unnecessary.  However, at the time I originally wrote about them, there was no way to know that this would be the case, so I think it was prudent to take precautions.)

I was surprised at the ad hominem nature of the editor's response to this page.   He did not refute the points I make above, but gave a list of assumptions he claims I am making, most of which I am not.  (I think I was especially annoyed at his questioning my motivation as resulting from my being at a state university since on email lists I have defended his motivation as not being just to sell newsletters or make money.   Being a tenured professor means that I can express whatever conclusions my research and study leads me to.)  

v 1.21 April 5, 2000