The man who would be prophet dare not miss.

kw: book reviews, nonfiction, economics, demographics, forecasting

This book took a week to read, and a couple days to think over. In The Next Great Bubble Boom: How to Profit from the Greatest Boom in History: 2005-2009, written in 2002, Harry S. Dent, Jr. forecasts a boom in the last half of the decade, followed by a significant recession, or even a depression, of about 14 years' duration. He rides on his credentials as the one who forecast the economic ride we had in the 1990s.

So is he right this time? Or is it as Churchill said, "Even a fool is right sometimes." Author Dent's main thrust is the demographic changes that America and much of the world will go through as the Baby Boom generation finishes its climb to the peak of economic power, then goes into retirement, followed by the smaller, and thus less powerful, echo boom, the X generation.

I am certain there will be a significant impact. Whether the reduction in the need for goods and services will be greater or less than the reduction of workers available is another question. In other words, we know demand is going to reduce, probably from 2010 until the early 2020s. Will the job market shrink more than the working population, leading to rampant unemployment? Will the working population instead contract the more rapidly, leading to a surplus of job and business opportunity?

The demographics are pretty clear. There are many studies of the expected population trajectory for most countries, and of the economic rise and decline of typical individuals as they age into the work force, buy houses, raise families, retire, and "fade away." The use we make of such data requires a more rigorous mathematical basis than it receives at the hands of economic forecasters.

For example, the author shows that for the average American, spending is the greatest around age 42. Then he takes the birth curve of the 1940s to 1960s, projects forward 42 years, and shows it as an economic projection. However, that curve has a rather broad peak, and is skewed: rapid rise, slower fall with aging.

To properly combine a spending curve with a changing population curve, the proper mathematical tool is convolution, AKA cross-correlation. Only a convolution of spending with age, and the numbers at each age over time, will properly predict an economic trajectory.

An extremely simplified model shows the principle. This chart shows a population bubble: an otherwise steady birthrate of 10 (thousands, millions, whatever) per time period (years, decades...), rises to 25, then drops again to 10. We'll call this the Birth Function.

This second chart shows economic effectiveness (on whatever scale you'd prefer to measure it) as it rises then falls with age. Supposing these time periods are decades: earning "enough to care about" begins in the 20s, rises to a peak in the 50s, then falls to "too little to care about" in the 80s...on average. We can call this the Spend Function.

You may of course get more complicated schemes, with differing trajectories (and impact levels!) for rich, middle, and poor "classes." Regardless, the homogeneous economics-with-age Spend Function function is convolved with the Birth Function to produce our result:

This chart shows that the convolution is not as sharply peaked as the two functions that produced it. The time scale is different also; this Economic Impact Function covers about an amount of time equal to the width of the Birth Function and the width of the Spend Function.

The Birth function and Spend Function both fall from a peak to the baseline (10 in the first case, zero in the second) in three time periods. Their total width is the same, six periods.

The total width of the Economic Impact Function, the convolution, is twelve time periods. However, because it is more 'curvy', it really has a significant difference only over about eight time periods, so the fall from the peak to the effective end covers about four periods.

Now, given the many other factors that affect a national market over various periods of time, there is a lot of noise masking any signal we'd like to discern. Specifically, if the month-to-month variation in a market signal is a few percent (10% is common for the DJIA, for example, from any month to the next), it can take a long time to pick out a long-term trend.

Thus, we may expect the US economy to turn downward around the end of 2009, and it very well may. However, it may have market peaks and valleys that make it hard to determine the fall has really begun for several months. Major market slides take a couple years to work out. Only in retrospect can we say, "The downturn was such-and-such a date."

Harry Dent's book is interesting, informative, and gave me a lot to think about. I like his ideas, though I think the analyses view some noise as though it were a signal. That's where I'd put Elliott Waves, on which Dent spends a chapter. Bottom line: the Boomers and the Echo Boom can be expected to make the world economy ring like a bell for about a century to come. When you expect a downturn, how do you prepare? Best way known: start a business that caters to those you expect to have money anyway.

My grandfather, once a salesman, later a piano tuner, began renting and leasing used and repaired pianos in the 1930s. He figured almost anyone would go for having a piano if it only cost them a dollar or two a month. Everyone needs entertainment, and radios then cost a lot. He was right, and continuing piano rental receipts provided my grandparents a very comfortable retirement, including two vacation homes.

Harry Dent makes me think. I like that.