Running the Risk Premiums

What should we expect from Mr. Market? The answer's always in doubt, but strategic-minded investors should run through the numbers anyway.

It's hard to overrate the value of taking a hard look at investing assumptions. By continually putting an expected price on risk, we become better investors. There are no crystal balls, but the next best thing is improving our skills in the art and science of discounting the future as a tool for enhancing return…maybe.

As an example, in the next issue of The Beta Investment Report we forecast an expected equilibrium risk premium for the market portfolio of 2.5%. We arrive at that forecast from summing up the individual forecasts for each of the major asset classes based on their market-cap-informed share of the total portfolio. Overall, it's a long-term prediction based on an equilibrium assumption and, we believe, a reasonable benchmark forecast, for reasons we discuss in some detail in the newsletter. But we can't be sure that our forecast is accurate and so we need to stress test our assumptions a bit. Here's a very brief illustration of the basic concept, including some of the give and take that keeps us forever wondering about what's coming.

If we assume a 3.0% risk-free rate for 3-month T-bills, our 2.5% risk premium forecast for the market portfolio becomes a 5.5% annualized total return outlook. But that's nominal. We also need an inflation forecast for estimating our real, or inflation-adjusted return.

Let's be generous and assume that inflation will be modest 2.0% going forward. That's a bit higher than the market's 1.8% inflation expectation, based on the spread between the yields on the nominal and inflation-indexed 10-year Treasury Notes. Yet 2.0% inflation is also below the 3.0% inflation that prevailed during the 20 years of the Great Moderation through 2008.

Putting it all together, expecting 2.0% inflation and a 5.5% nominal total return for the market portfolio leaves us with a 3.5% real annualized total return. But let's remember that forecasting is difficult, especially about the future, as the old saw goes. As such, we should assume that our predictions are subject to error. The question is how much error?

If inflation ends up at 3.0% instead of 2.0%, our real return falls to 2.5%. At 5.5% inflation, our real return is zero.

But perhaps actual inflation will be lower than our expectations. Or perhaps our risk premium projection of 2.5% will be higher. Meanwhile, a confident investor might project a higher return by skewing his asset allocation to, say, U.S. stocks because he thinks returns there will exceed the market portfolio's expected risk premium.

Then again, let's consider another alternative. We could buy a 20-year TIPS and lock in the real return of 2.42%, based on last night's close. That's nearly as high as our market portfolio risk premium forecast, which is a nominal return. Why is the TIPS market offering what appears to be a high real return? Are TIPS prices too low? Or is our risk premium outlook too low?

There are no definitive answers if we're looking forward. But until we start plugging in the numbers and making assumptions, the difficult business of making investment decisions is that much tougher. In a perfect world, we could simply extrapolate historical returns into the future. That would make our job as investors rather easy. Alas, it's not quite so simple. Historical returns are a volatile lot and it's not clear which historical period applies to the immediate future that awaits.

What should we do? Before we can begin making judgments about asset allocation, we need a neutral (or minimally biased) reference point for assessing the market outlook. In short, we need an estimate of equilibrium risk premiums for individual asset classes and the portfolio overall. From this foundation, we can amend the market-inspired asset allocation based on our risk preferences and expectations.

Estimates of equilibrium risk premium offer a basis for adjusting the asset allocation, if at all. Expecting that future returns will deviate from the implied equilibrium predictions in the short-to-medium term suggests changing the passive market-cap asset allocation. For instance, let's say that an investor is quite a bit more bullish on U.S. stocks for the next 3-5 years compared with the assumption in the long-run equilibrium risk premium forecast. Adapting that view into an asset allocation strategy translates into raising the U.S. equity weight over the market-cap weight; a more bearish perspective calls for a lower-than-average market-cap weight.

Most investors should probably accept the market-cap weight; in practice, almost no one does. Then again, that probably explains why the market portfolio's track record looks pretty good over time relative to actual portfolio results. To understand why, we need to make some assumptions beyond risk and return by also considering taxes and trading costs.

Beating the market is hard, in part because making accurate predictions is difficult. But even forecasts that are reasonable are in danger of being irrelevant, or worse, once we factor in the cost of trading and paying the government, say, one-third of any profits. That implies that we need more than reasonably good predictions to overcome the frictions that harass actual portfolios; we probably need stellar forecasts. No wonder that alpha's so scarce.

Comments

[dybydx]

i wouldnt count on the 2% figure. the official inflation figures are often understated. for this reason TIPS are being criticized as having negative yield.

for example, gold over the last 38 yrs showed appreciation of 8.6% per yr.

2009 May 29 01:58 PM

[Prudent Man CFA]

It is naive and a fools game to use a lagging metric like beta to measure risk. The only thing beta measures is what the volatility of price was around a moving average of the mean. Sounds nice but means nothing. There are a myriad of metrics and fundamentals that can change the beta in a NY minute so it is statistically insignificant.

Using these computer driven measurements to determine risk is a lazy persons analysis with the pay back a lazy person gets. The complexities of risk are dynamic, forever changing in both description and effect. One must be continuously aggressive, flexible and adaptable in their work if one wants to remain in the game for fifty years. No two numbers or variables tell the same thing in investing.

(I explained this to my MBA quant professor forty years ago as he was in love with beta and he took me down from an A to B. That was a stunning example to me of higher education search for the truth, as long it agrees with their ignorance.)

2009 May 29 02:03 PM

[Prudent Man CFA]

On May 29 01:58 PM dybydx wrote:

> i wouldnt count on the 2% figure. the official inflation figures
> are often understated. for this reason TIPS are being criticized
> as having negative yield.
>
> for example, gold over the last 38 yrs showed appreciation of 8.6%
> per yr.

Nothing like picking data points to prove your point. What was the compounded annual rate return the past ten years? If you rolled futures what was the expenses? If you bought gold outright what were the carrying costs, e.g., insurance, storage, shipping, etc.?

Your are right about 2% inflation. The rate is too low for our debt and too high if the economy completely tanks. But, interest rates will go up regardless of inflation because of supply and demand by governments world wide.

2009 May 29 02:11 PM

[Prudent Man CFA]

The author assumes that we have to be invested in equities or anything but cash, for that matter.

This assumption is a myth sold by Wall Street and has little basis in reason. Equity mutual funds have sold this idea for decades and look at their returns. Those on the top one year are on the bottom the next - elevator funds.

Many periods in this country's history "investors" would have been safer and financially better off holding cash. Most Medal of Honor winners are dead so who wants to be a hero?

2009 May 29 02:18 PM

[Larry House]

Interesting anaylsis, James. As you aptly bring out, this is a most difficult time to make assumptions about risk and portfolio expectations. "Expert" opinions are everywhere. Doug Kass says he is neither bear nor bull right now. Art Cashin says we are in a very dangerous period. It is hard to have a conviction one way or another. It is a time to be well diversified. If anyone has an edge on this market and dares put all eggs in one basket, I would like to hear that person's views.

2009 May 29 02:27 PM

[dcb]

I left my MBA program when I knew shit was going to hit the fan alt may. I then did a valuation for him which if accurate would imply a risk premium that would cause the market to fall another 40% after Lehamn. we fought for days. I mentioned many things like rate of rise in unemployement, misery index. all these things hadn't happened at any times other than very severe recessions. I wrote him back after the week of march 9th and told him now that the market had fallen 44% since then he could get back in. He never answered.

Look if the beta's and risk premium people had assigned was correct, then all these banks wouldn't have lost their shirt. the proof that you are right is in the crisis itself. But it is good to understand the models. If we end up delevering, I think the market will end up very stable and regular again. credit of course screws thing up.

On May 29 02:03 PM Prudent Man CFA wrote:

> It is naive and a fools game to use a lagging metric like beta to
> measure risk. The only thing beta measures is what the volatility
> of price was around a moving average of the mean. Sounds nice but
> means nothing. There are a myriad of metrics and fundamentals that
> can change the beta in a NY minute so it is statistically insignificant.
>
>
> Using these computer driven measurements to determine risk is a lazy
> persons analysis with the pay back a lazy person gets. The complexities
> of risk are dynamic, forever changing in both description and effect.
> One must be continuously aggressive, flexible and adaptable in their
> work if one wants to remain in the game for fifty years. No two
> numbers or variables tell the same thing in investing.
>
> (I explained this to my MBA quant professor forty years ago as he
> was in love with beta and he took me down from an A to B. That was
> a stunning example to me of higher education search for the truth,
> as long it agrees with their ignorance.)

2009 May 29 02:53 PM