I've been working on this model for a while, and at this point I'm using it as a way of thinking about market level as affected by fiscal (and monetary) policy. Without further ado:
Briefly, as of Friday last week, I see 6.9% as a realistic expectation for the ten year rate of return investing in the S&P 500.
How It Works
The items with the gray background are factual inputs. E10 is 10 year average earnings for the S&P 500 index, derived from Professor Shiller's spreadsheet. The beginning dividend comes from the same source.
The items with the light blue background are assumptions. I used 2% for real GDP growth, and 2% for PCE inflation excluding food and energy. The numbers are long term projections from the SEP as published at intervals by the FOMC. CBO numbers are similar.
Financial Engineering reflects the idea that management can grow earnings faster than the combination of GDP and Inflation by means of buybacks and M&A activity.
The Fed Model Cutoff reflects the observation that when the 10 Year Treasury rate (GS10) is 5% or higher, PE multiples on the market tend to conform to a version of the Fed Model. With the 10 year over 5%, CAPE is projected to equal 1/(GS10-.002). That is, with GS10 at 6%, .06 -.02 = .04, then 1/.04 = 25.
This cutoff is derived from data prior to QE, and there is a possibility that it will go lower if the market comes to expect GS10 to stay under 5% indefinitely.
With GS10 under 5%, the model computes CAPE as 1/((GS10 X -1.45) + .1023). So, for 4%: .04 X -1.45 = -.058 + .1023 = .0443. Then 1/.0443 = 22.6. That's how CAPE behaved until QE. With GS10 under 5%, CAPE declined as GS10 declined. Counterintuitive, it reflects the flight to safety when the economy does poorly.
GS10 is estimated as follows: .32 GDP + 1.87 PCE +.0027. So, for 2% and 2%, (.32 X .02) + (1.87 X .02) + .0023 = .0465, or 4.7%. The equation here was developed by applying multiple linear equation software to real GDP, PCE inflation and GS10 for the period 1/1/1988 to the present, with R2 of .69.
The underlying mechanics are: real GDP growth and PCE inflation drive GS10, which drives CAPE. Beginning CAPE is a fact, CAPE at the end of the 10 year period is a projection. The estimated rate of return for the period assumes E10 and the dividend increase annually by the sum of real GDP, PCE inflation and Financial Engineering. The investor buys the index, receives dividends for 10 years, and winds up owning the index at the projected CAPE.
Inflation plays a large role. Low values are toxic to market returns. This fact explains a lot of FED behavior: they're afraid of deflation, and asset inflation is as important to them as unemployment and price stability.
PCE inflation is more stable than CPI, and provides better correlation with GS10. PCE inflation hit a minimum of 1.2% during and after the financial crisis. If it returns to that level and stays there, this model calls for another lost decade.
The Fed Model Cutoff, if reduced, would give very high CAPE multiples. For example, referring to the spreadsheet values at the beginning of article, if the cutoff were reduced to 4%, ending CAPE would be 37.7, driving investment returns to 9.5% annually. How sweet would that be?
What Ifs (Ceteris Paribus)
If fiscal policy were able to drive real GDP to 3%, then market returns of 9.3% would be a realistic expectation. For 4%, 9.5% is realistic. I think the market expected this type of result after the election.
Inflation at 2.5% doesn't help matters and would surely draw attention from the Fed. Inflation at 1.5% takes nominal returns down to 3.3%. At 1.2% you get another lost decade. 2% seems to be ideal. The current value is 1.7%, and the trend is up.
The Fed Model cutoff is a sleeper, as mentioned earlier. Extreme values of CAPE have been seen in Japan. In my opinion, Japan is typified by a homogeneous population and a strong work/savings ethic. The upsurge in divisiveness in the US, together with a weak work/savings ethic, argues against such outcomes here.
I Want My 9%
Investors like the idea of long term returns in the 9% area. Due to uncertainties engendered by the current occupant of the Oval Office, I really would like to receive that kind of return to compensate for political risk. According to the model here, the S&P would need to go down to 1,955 to make that a realistic expectation.
That's 19% below the recent high water mark. A full-blown correction. I have put on a small hedge, and plan to leave the bulk of my holdings untouched, ride the dip out, if it occurs.
Caveats and Reservations
Nothing in this model has an R2 of more than 0.71. Results become less predictable when PCE inflation is below 2%, or GS10 below 5%: that is, right now. We're in uncharted territory.
The overall impression is kind of a Rube Goldberg device, lots of wheels, gears and levers.
The model develops GS10 values that aren't consistent with current reality. In effect, it expects reversion to something in the 4% to 5% area. The Fed has bought up a lot of treasuries. The supply and demand implications of this action, or the unwinding thereof, are unknown. It sure would be nice if the unwind would result in a reversion to the mean here.
I'm a retail investor, talking shop, not giving advice or making prognostications. I'm looking for comments and criticism on the thinking in this article, hopefully to clarify my thinking.
Disclosure: I am/we are long SPY.
Additional disclosure: I'm long the S&P 500 by means of the Vanguard Index Fund. I have a small hedge: long SPY puts, deep in the money, distant expiration.