How much of those reserves is from tarp funds and how much is from banks moving their assets to a safer place? That would considerable affect your assessment of the dollar.
 

So you're extrapolating that number from a 3 month period?

In other words, when all is said in done, by next year that figure could actually be higher.
 

Your calcs seem correct assuming that the reserves compound, but do they? Do reserves grow linearly instead? That is, 700 billion this quarter, 700 billion next quarter, etc...then annually, you are looking at (700*4 + 100)/100 * 100 - 100, which is 2800%
 

Greg,

I'm not predicting that reserves in Sept 09 will be 4096x higher than in Sept 08, but the definition of "annualized growth rate" is what it is.
 

Bob,

The way I do annualized rate of a quarterly growth rate is as follows. I take the end month (Month3) and subtract the start month (Month1) then divide by the start month.

This gives the growth rate for the three month period. So if you want to annualize this (on a simple basis without compoundng), your three month growth rate is 1/4 of the annual rate, so you multply by 4 to get your answer.

In the case you state, it is

800-100=700

700/100= 700%

700% * 4= 2,800%
 

Robert,
When you get percentages that high compounding it makes a huge difference. You can't approximate it by ignoring the compounding factor like you can with smaller percentages.
 

Bryan,

The problem is that this type of growth generally does not last for a full year, unless you are Zimbabwe. So it is terribly misleading to compound the event as though it is going to go on for a year.

The whole idea of annualizing in these cases is to compare apples to apples period. If CPI is coming in at 12%, then I know that 2800% annualized inflation is huge, but to then jump to the "forecast" of compounded 12 months of this distorts the picture.
 

To repeat, everybody, I did not forecast that the Fed would increase reserves by a factor of 4096 from Sept - Sept. I was merely trying to illustrate how much the Fed pumped in during the last three months. I used the same formula for the ECB's 25% figure, so in that respect it was definitely apples to apples.

Wenzel, I get what you're saying, but by the same token unless we're Zimbabwe, presumably Bernanke isn't going to add $700 billion in reserves every three months, for the next 9 months. So your technique is "wildly misleading" too, right?

I'm falling back on the fact that my formula is the mathematical definition of an annualized rate of growth.

More generally, in finance most things grow exponentially, not linearly.

Now if there is a convention on these things, and everybody reading the newsletter will assume I'm doing it Wenzel's way, then I should have done it that way and clarified "(without compounding)" or something.

Last point in my defense: My technique can yield ludicrous results if you shrink the period. E.g. if you just look at the hour in which $1 billion hits the reserve balance, then you would get "annualized" growth rates in the kajillions.

But I took a 3-month period, which is pretty long. And bank reserves grew by 700%+ in 3 months, aka 400,000% at an annualized rate. :)
 

Wenzel: Now I'm really worried. When you were reporting annualized M2 growth rates on your blog, I thought you were doing it my way! That means inflation of money stock is a lot higher than I had thought, from your figures. :)
 

Surely currency traders know all of this? Perhaps they don't expect the increase in reserves to increase the actual money supply nearly as significantly?
 

What is the difference between the Reserves, the monetary base, and M1? I know currency is at least included in M1, is it included in the monetary base too?
 

Bob,

I don't have any major disagreements with you on this as long as everyone knows how you are calculating your growth rates, the people that come to your blog should have the acumen to understand your point.

But, I'm trying to think if there ever has been a period over the last couple of decades where the annualized compounding of a three month reserve growth has ever matched the actual annual growth rate over that period. (I doubt it.)

Compounding I think is implying activity that is not likely going on (There is not that kind of smoothness to Fed actions), whereas if I take a three month period and multiply by 4,I am saying something like "Look this growth rate is equivalent over 12 months to x%"

BUT in the end I think what is really going on here is that the Feds incredible bursts of huge reserve injections is breaking down back of envelope type calculations.

In short, it's Bernanke's fault :)
 

Out of curiousity, I clicked on the "Compound Annual Rate of Change" graph on FRED. They also show a spike to $400,000%, so the calculation isn't wrong.
 

The link in my previous comment will take you to a "Chart not available" page. Just change the start year to something like 2000 and you'll see the spike.