### Mutual Fund Cash Levels Adjusted For Inflation

Published September 3rd, 2009 in Market InternalsLast week we looked at the levels of cash and free credits being held in institutional and retail trading accounts in the US: Mutual Fund Cash Levels & NYSE Free Credits. I briefly touched on a research report on mutual fund cash levels by Jason Goepfert, who by the way, runs a great service at SentimenTrader. However, I wrote that:

Unfortunately, Goepfert’s research report does not take into consideration inflation or deflation but simply adjusts the level of mutual fund cash levels according to the 90 day T-Bill rate. I’ve sent him a message about this so hopefully when he’s back from vacation he can update it with this new twist thrown in.

Upon his return, Jason accepted my suggestion and lost no time in whipping up a new indicator which takes into account the added variable of inflation/deflation.

To be able to understand what this new chart is saying, it is helpful to go back to the award winning research report. In it, Jason argues that before we try to use mutual fund cash levels as an indicator, we need to adjust it for the prevailing interest rate environment. For example, in the 1980’s, with interest rates in the double digits, there was ample reward for sitting in cash. Stripping out this effect, therefore, is important because otherwise it is a distortion.

Using statistical modeling, we can determine how much cash *should* be held by assuming a certain level of interest (90 day T-Bill rate). After that, it is easy to compare the actual cash levels to this theoretical level to determine if mutual funds are overweight or underweight cash. Looked at this way, mutual fund cash levels are neutral, telling us that managers are neither overweight or underweight cash right now.

Alright, so what happens when we also take into account the effect of inflation or in today’s case deflation? In other words, the *real* rate of interest?

A completely different picture emerges. This isn’t surprising because we are currently experiencing real interest rates close to +6.5% - a far cry from the nominal rate of 90 day T-bills. **Here is the inflation/deflation adjusted chart of mutual fund cash levels:**

Well that changes everything!

The mutual fund cash level is no longer neutral. This new adjustment shows that on aggregate, mutual fund managers are holding about 1.5% *less* cash than the statistical model suggests. While that may seem a trifling difference, historically, it has been a tell for a topping market.

The data fits the market remarkably well. The only caveat is that as timing indicators go it is sloppy. The market may top now, a week from now or a bit longer. With the corollary that the amount of cash may decrease even further. But the message of this indicator is crystal clear: the market is top heavy.

If you are interested in further details, you can download Jason Goepfert’s research on mutual fund cash levels from the Free Trading Resource section. Check in the Charles H. Dow award folder. Needless to say, I highly recommend Jason’s work. Head over to his site and take him up on his no risk, 14 day free trial offer.

**Enjoyed this? Don't miss the next one, grab the feed or **

Excellent work!

Two comments. Firstly, Norman Fosback so totally DID the initial interest-rate adjustment, back in the 1970s and 80s, that he called it the FOSBACK INDEX for over a decade!

Finally he became a little sheepish about that, since it was, after all, just a single adjustment, and thereafter he called it the Mutual Fund Cash Ratio Index (see his classic book ‘Market Logic’). I subscribed to his newsletter until he sold all his publications, and semi-retired, so I followed this closely.

He considered it one of his most important indicators.

Second comment: don’t interest rates ALREADY somewhat adjust for inflation? Aren’t you causing Mr. Jason to “double-dip”, as it were?

Lastly, Fosback’s rationale was more along the lines of psychology–not the precise kind of metrics one uses in a decision to invest, say, in TIPS or Treasuries or CDs.

It was more an incorporation of the simple truth that a dubious manager, looking at an iffy market, and also looking at 11% yields on Treasuries (real OR nominal) is going to be sticky-fingered with his cash; while one looking at a yield of 1% is going to be nauseous and green and itchy-fingered to go invest in something.

Keep up the great work, I regularly absorb the thought provoking articles you provide.

Daniel

yeh, i was kind of amused to see that this was award winning work. It has been around for 30 years and I used it myself in my personal work. I used regression analysis a decade ago to come up with the equation

Expected Mutual Fund = 1.1 0.77*Tbill.

Fosback published it. I saw it in Ned Davis reports as well.

I probably need to calculate a new equation given that we now have a lot of new low interest rate data. I haven’t used it in awhile because,

1 It wasn’t helping me

2. It had a one month delay in obtaining the data

3. It seemed to have shifted, most likely due to investors having access to money market mutual funds themselves via Rydex and Pro funds. Not really sure.

I do agree with Jake, that the interest rate levels are more important than the inflation rate. The name of the game is a high return with as little risk as possible. Why would a fund manager be fully invested when he could get double digit returns during the 70-80s’s with no risk money market accounts. The managers that had the option for such, were going to use it Of course, they are going to carry 10 to 20% in cash during such periods. In my opinion, Inflation rates would simply be a proxy, highly correlated to interest rates.

The equation I came up with sometime ago and I’ll update if there is enough interest suggest that

Tbill Expected

Rate % Cash

00 01.10

05 03.85

10 08.80

15 12.65

For you mathematicians out there, it would probably make a lot of sense to use a linearly weighted regression approach so that more recent data would have more weight on establishing the relationship as time progresses and possibly look at using some weighted average of the yield curve.

I’ll put this on my todo list

darn it, should have read

Expected Mutual Fund Cash % = 1.1 0.77*(3 mt TBill yld)

system is not taking my sign.

should have read

Exp % Cash = 1.1 plus 0.77 times (3mt Tbill yld)

This idea that we are experiencing interest rates of 6.5% is absolute nonsense. 90d tbills yield 0.12% and forward looking inflation expectations vary from -0.75% on a 1 year towards 2% for 20 year. It is not hard to find odd numbers right now, but forward looking estimates are more meaningful and less dramatic. Looking at it another way, 5y index linked bonds in the US yield 1%, a multi-generational low. That is a real market and you should base your calculations on real tradeable assets, rather than some arbitrary headline-grabbing mis-calcualtion.

Paul, forward looking estimates may be less dramatic but they are actually less meaningful because they are simply estimates. Similar to how forward earnings estimates always look better. But as we painfully know, forward data can be revised over and over again to bring it in line with reality. So why not just look at the actual real data? So I’m not sure I follow you when on the one hand you want to base calculations on ‘real’ data and then you go ahead and use arbitrary numbers which are nothing more than somone’s estimate (!).

Babak,

my comment is that we are not buying the past and in the case of the stock market, we are attempting to buy a reasonable multiple of future earnings. It is difficult, but the analyst that are capable of making the best forward estimates will be the one’s taking the prom queen to the dance. But you are correct it is very difficult to do. Standard and Poor’s generally has estimates for earnings posted for the next four quarters and even they have missed terribly. Their current estimates are “under revision” and the truth is they probably don’t have a clue. Looking at past earnings is only useful as one piece of the puzzle in estimating future earnings. It is where you start. But never in the last 50 years, have past earnings had so little correlation to future earnings.

Babak, I don’t think you understood my point, and I take the blame for that. I will try to be more clear. I am talking about the use of “real interest rates”. There are 2 ways to get a measure of this:

1) take a nominal interest rate and subtract a measure of inflation. Two problems with this: a) the notion of comining interest rates (which have a forward looking component to them) is convenient but is meaningless from a economic point of view and b)the choice of which inflation measure to use is arbitrary: CPI, PPI, core CPI, average earnings, GDP deflator etc?

2)take a real interest rate from a traded instrument, such as index linked bond. This is a real market price, and is not arbitrary in any way.

To me, the second method is superior. If you take the real interest rate from these instruments, you find that real interest rates are at a generational low. That’s my point. I just can’t agree with your assertion that “we are currently experiencing real interest rates close to 6.5%”.

In deflationary periods, real interest rates take on a different dimension due to the fact that interest rates are bounded on the downside by zero and deflation is unbounded.

My other comment was that mutual fund managers incentive for investing a portion of their assets in interest bearing securities is the net return they can show investors via the interest rate return, rather than a real interest rate return

Yes, but these are not deflationary times, despite the assertion of many a blogger. Oil is up 70% from its Q4 low, gold is up 40%, copper up 40% etc. US medium term inflation expectations priced into bonds (the 5y/5y forward) now stands at 2.43%, back to pre-crisis levels. This measure stood at 0.5% back in Q4 when deflation fears were widespread and valid. That story is over. Inflation is normalising and since nominal interest rates are exceptionally low, that means that real interest rates are exceptionally low. Over the next few months, I think this cash in money market funds will gradually be invested in stocks (at the wrong time) as the acount holders realise that all this talk of deflation and high real interest is nothing more than idle chit-chat and scaremongering.

I think that I can agree with you, that deflation will not be the problem going forward that it was expected to be, but if you look at the CPI numbers, we had a bout of it in the recent 12 months.

We are to some extent in agreement. I think cash levels remain a bit over historic norms, especially if you go to the trouble to adjust the % of cash in mutual funds by some factor to account for very low interest rates.

Much like the 50’s, I think that it is very possible that we may have a decade where the cash levels stay above historic norms due to a shift in savings habits by consumers. I can’t really say with a high degree of confidence however.

Paul, I’d like to get a gauge of the inflation/deflation from TIPS but the data has been discontinued by the FedM. If you know of another way or source for the data, let me know.

Babak, I get it from Bloomberg. I use USGG5Y5Y Index to get the 5y5y forward (ie mkt’s expectation of average inflation from year 5 to year 10). Now 2.48; pre-crisi range 2.25-2.75%. The instrument is called “US 5yr 5yr forward breakeven” if you want to look it up elsewhere.

Supply and Demand moves prices, creates trends etc. If it were as easy as a formula then we’d all be rich. I don’t doubt that the equation does reveal some data about cash levels etc., but I’d be cautious about using any historical numbers to draw conclusions about the future. When was the last time we were $11T in debt and the US Government was the biggest buyer of its’ own debt ?

New Day - Uncharted Waters without a map……”Caution Will Robinson”