William J. Bernstein
Of Markets, Economies, and Populations
In the winter issue, "How Much Pie Can You Buy," we followed the bread crumbs from economic growth to rises in stock price. We found that from 1929 to 2000, domestic GDP grew by 6.63% per year. It would have been nice if all this growth had filtered directly through to share prices. Alas, it did not. During the same period, aggregate corporate profits grew by slightly less (about 6.42%), but the real damage came from the investment bankers, who issued so many new shares that per-share corporate earnings grew by an even lower 4.98%. Fortunately, the gods smiled on shareholders in this periodmultiples expanded and stock prices rose by 5.87% per year.
Some suggest that per-capita GDP is a more useful indicator of stock prices than gross GDP and, in fact, this may be the case: from 1929 to 2000, per-capita GDP rose by 5.38%. It was a better guide of per-share earnings growth than raw GDP growth. Unfortunately, weíre looking at only one data point here, the U.S. market. We'd be more informed on the issue if we could examine the impact of GDP and population growth on share prices across a number of nations. Thanks to a happy coincidence, we are now in a position to do so.
In 1999, William Goetzmann and Philippe Jorion published in the Journal of Finance a seminal study of stock returns around the planet, entitled "Global Stock Returns in the Twentieth Century." Their key parameter was annualized real price rise (that is, dividends not included, inflation factored in) of many national markets in U.S. dollars.
Almost two years later, Angus Maddison published his landmark The World Economy: A Millennial Perspective. This tome, together with his earlier work, Monitoring the World Economy, provided real gross and per-capita GDP (again, expressed in real U.S. dollars) as well as population figures for virtually the entire globe, over the same period studied by Goetzmann and Jorion. In the table below, Iíve summarized the relevant data from both sources, including only those nations with more than a 35-year history of stock returns:
Per Cap. GDP
By combining the data from both Maddison and Goetzmann/Jorion, we can study the effect of these three factors on global stock prices. The story is fascinating, if complex, and the trail well worth following. But keep in mind, we are in fact looking at only two variables, since the three variables (gross GDP, per-capita GDP, and population) are mathematically related:Per-Capita GDP = GDP/Population
Further, the growth of these three parameters are roughly arithmetically related:Per-Capita GDP Growth ~ GDP Growth Ė Population Growth
Therefore, if you know two variables, you automatically know the third. For simplicityís sake, weíll initially consider only population and gross GDP. Itís obvious that as GDP grows, so should stock price; there ought to be a direct positive relationship between a nationís GDP and stock-price rise. Unfortunately, there is not. Below, Iíve plotted stock-price growth versus GDP growth for the 31 nations in the Goetzmann/Jorion database with a greater than 35-year track record:
As you can see, there is a weak negative relationship between the two, albeit statistically insignificant. Whatís going on here? The next step is to plot stock-price growth versus population growth:
Here we see a negative correlation that is highly statistically significant. In other words, population growth appears to be bad for stock prices.
The mystery of the above apparent irrelevancy of GDP is solved when one simultaneously regresses stock-price growth against both population and GDP growth. When the two are looked at together, GDP is seen to exert a positive effect on share prices after all. This is corroborated by plotting per-capita GDP growth versus stock-price growth:
Per-capita GDP very nicely captures both of these factors, with a rise in gross GDP helping the numerator and tiny rises in population helping the denominator. In fact, the bivariate regression shows that population growth is about twice as important as GDP growth. This suggests a trick that should capture even more of the price return: per-capita GDP growth minus population growth. Indeed, it turns out that this simple metric explains fully 55% of national equity returns:
Finally, one can specifically examine the "leakage" between the growth of GDP and the growth of stock prices. For many nations, the amount of GDP growth that escapes before arriving in the hands of shareholders is huge. For example, in Pakistan, GDP rose at an annualized rate of 5.73% while stock prices fell at an annualized rate of 1.77%; there was thus 7.50% of leakage between GDP growth and stock prices. In a few countries, notably the U.S. and Sweden, the discrepancy was actually negative due to rising valuations. Again, "leakage" is a good measure of property and shareholder rights. Its relationship to population growth is nothing short of phenomenal, with an R squared of 0.63:
These results shed a small amount of light on the sources of equity returns. GDP rises are good for stock prices only when they come from increases in individual productivity, as measured by per-capita GDP; they are bad when caused predominantly by population growth. The classic case of the latter is Pakistan, which, believe it or not, had the second highest gross GDP growth rate in our sample, but also one of the highest population growth rates. Needless to say, the Karachi Stock Exchange has not been a happy place the past four decades.
It seems unlikely that population growth itself is the cause of poor stock returns, but rather a marker for the underlying social and cultural conditions that attend high birth rates and disrupt corporate profits: a dirigiste government, with its concomitant disrespect for private property, lack of functioning capital markets, and intrusion of religious thinking into areas better served by Western rationalism. For example, Israel had the second highest population growth in the sample, but since this was largely due to immigration and not a high birth rate, it offered reasonable stock returns.
Is the above paradigm useful for predicting future returns? The key question here is, can we divine the demographic transition that accompanies successful industrialization? At one end of the spectrum, the traditional developed nations should continue to experience agreeable real returns, as should emerging Western countries like Hungary, Poland, Taiwan, and Singapore. At the other end of the spectrum are nations like Pakistan and Peru, where a demographic transition seems remote. And in between are nations like Indonesia, Thailand, and Mexico, which may or may not be in the early stages of demographic transition.
At the end of the day, though, a central paradox remains: Raw economics explains only a small part of equity returns. Factors well outside the ken of the financial economist are far more important and likely less predictable.
Copyright © 2002, William J. Bernstein. All rights reserved.
The right to download, store and/or output any material
on this Web site is granted for viewing use only. Material may not
be reproduced in any form without the express written permission of William
J. Bernstein. Reproduction or editing by any means, mechanical or electronic,
in whole or in part, without the express written permission of William J.
Bernstein is strictly prohibited. Please read the disclaimer.
The right to download, store and/or output any material on this Web site is granted for viewing use only. Material may not be reproduced in any form without the express written permission of William J. Bernstein. Reproduction or editing by any means, mechanical or electronic, in whole or in part, without the express written permission of William J. Bernstein is strictly prohibited. Please read the disclaimer.