The Bank for International Settlements has published an excellent working paper on the effects of population age structure on home prices. Here's how the author summarizes his findings (emphasis added):
The results suggest that global asset prices are likely to face substantial demographic headwinds in the next forty years. The theory is straightforward: House prices are determined jointly with financial asset prices. Hence, if house prices face headwinds, so should financial asset prices. Using the estimated coefficients and global population forecasts, asset prices would face headwinds up to a full percentage point. This headwind is substantial, but based on historical returns would not imply real asset price declines...
Though the results do not imply absolute real price declines, they suggest that in the next forty years house prices in advanced economies will face a more difficult environment than in the past forty years.
One of the most important aspects of the study, and of most studies of age structure's effects on economic and financial market outcomes, is the illumination of the impacts that baby boom generations have had in different countries. For example, here's a graphic from the paper showing the impact on real housing values over the past thirty years (appreciation or depreciation in basis points per year), and as forecasted for the next thirty, in the major English speaking countries (the projections are even more pessimistic for many non-English speaking countries):
We can't overstate how important it is for savers, investors, retirees, business leaders, and policymakers to grasp the dynamics implied by a chart like this, as it calls into question each and every assumption about future nominal asset returns based on the last thirty years of data -- assumptions that are widely used and abused by the financial services industry, in our experience!
Human beings are innately backward looking when making predictions. That's OK if we are diligent about investigating the hows and the whys of past experience, and assessing how the relevant factors are likely to behave in the future (and just as importantly, being attentive and responsive to the many factors and possible outcomes that we are as yet unaware of). But it's rather dangerous when we just scratch the surface of historical data in order to produce statistical claims (usually for marketing purposes in our industry) about the future. Garbage in, garbage out. Unfortunately, our industry will continue to create and peddle garbage as long as clients are willing to be over charged for it.
In the matter at hand, the essential fact is that most of our prevailing expectations about economic and market outcomes (and prudent public policy) were formed in recent decades, as baby boomers moved through their early and middle adult years. Those expectations could be turned upside down as boomer generations head into retirement and old age in many countries during the decades ahead. Any long term investment process that ignores this secular dynamic is likely to create more downside risk than intended.
A few caveats on the paper's implications, the last three from the BIS author:
- As touched on in our recent interview with Diane Macunovich, there are many different ways to slice and dice demographic data, and different methods can produce very different results.
- Other non-demographic factors have clearly played a role in the past, sometimes much stronger than age structure effects.
- Technology and lifestyle changes -- for example, higher typical retirement ages driven by technology and/or concerns about transfer payments -- are hard to predict and account for.
- Predicting age structure effects forty years into the future is hazardous, and long term forecasting has a dismal track record.
That last point is also extremely relevant to debates over national deficits and debt. Apparently it's accepted wisdom at the current BIS that future social benefits will have to be cut dramatically in advanced nations:
In the government sector ageing related entitlement spending is currently set on an unsustainable path as Cecchetti et al. (2010) shows. Hence, lowering old age government benefits seems to be inevitable. Consequently, the next old generation might have to run down their assets in old age more aggressively than previous old generations as their private and public entitlements would be much less generous. This would exacerbate the negative impact of ageing on asset prices.
That doesn't really make sense though. If transfer payments to retirees are cut in such a way that existing financial assets have to be drawn down more aggressively, then all else equal, you're going to see asset deflation and rising pessimism. If powerful enough to be deflationary, then spending new money into existence would be an appropriate response -- which would mean (ignoring valid political arguments over how and where government spending should occur) that the transfer payments should never have been cut.
So this passage and the paper it's based on look like yet more hand wringing over cursory statistics by the debt phobes. But as we noted above, cursory statistics are garbage, and that's true whether we're talking about financial services, economics, or public policy. For a more complete fisking of the Cecchetti et al paper, see here.