Predicting the future: 27-July-2010

I am an engineer by training.   It is in my blood to try to engineer a investment solution that gives good upside performance while structurally limiting risk to reasonable levels (e.g., no greater than the upside opportunity).   A few years ago I concluded that I had not figured out a way to do this, and that it is probably impossible.

For example highly rated bonds, usually not considered the riskiest of investments, are sensitive to prevailing interest rates.  AGG, a bond ETF is currently yielding around 3.7% annualized interest.  Its duration, a term that defines the average time until maturity for the bonds in the fund is around 4.    The duration metric quantifies how sensitive a bond investment is to interest rate fluctuations. In general, the price of a bond will drop by the duration amount in percent if applicable interest rates (intermediate time frame in this example) go up by one percent.   So, if interest rates go up by one percent, which is almost certain to happen, the bond fund like AGG will lose 4%.  Interest rate increases of several interest points could easily happen.   AGG’s  3.7% interest rate starts looking like an inadequate return given the risk.   Of course you could start adding things like protective puts, but those add commissions and costs that reduce your upside, and are typically very expensive if you are trying to completely limit your risk.

Some investment advisors might promote their asset allocation approach as not requiring predictions about the future, but their strategies are full of predictions (e.g., long term growth of stocks,  interest rates going down when equities go down, stocks in emerging markets growing faster than USA based stocks).

So realistically our only way to make better than CD level returns is to make good predictions about the future.   If you think interest rates are going to stay low for the next year, then AGG is not a bad investment–especially when compared to CDs and the like.  There are a quite a few things that can be predicted beyond just interest rates (e.g, short term stock direction, stock gains over 10 year periods, volatility, volatility skew,  “black swan” events, earnings reports, commodity prices, inflation rates, correlation between stocks, correlations between stocks and their sectors).

I am certainly not saying that we should forget strategies that limit/manage risk–I’m quite fond of those.     I’m just saying that managing risk is only part of the story, we also need to be right with our predictions most of the time.

So what are my predictions?   In general I tend to be pretty good at the longer term predictions,  usually not patient enough with the middle term predictions (e.g., 3 months),  not bad at the week level, and no better than a coin toss intra-day.

My 12 month predictions:

  • No double dip recession
  • Bull market starting back up again in the fall

My 3 month prediction:

  • Sideways market.  S&P highs lows near 1020, highs near 1140

The remainder of this week:

  • 20% chance upside breakout for S&P > 1120
  • 50% chance sideways movement the rest of the week, closing in 1090 to 1120 range
  • 30 % chance fallback into the 1070 range


First posted on

Leave a Comment