What “real” (ie, after inflation) rate of return should you expect to earn on your savings over the long-run? The Governor of the Reserve Bank of Australia, Glenn Stevens, offered an unsolicited answer to this thorny question in not one, but two, speeches delivered during the past week.
Central bank governors don't typically wander out in public and proffer specific point estimates for expected through-the-cycle returns generated by diversified investment portfolios. So a related ponderable is why Stevens would engage in such an exercise.
One possible answer is that he was seeking to condition community beliefs. Now this is a valuable undertaking. The next step in Stevens’s argument, which he did not articulate, is that if investment returns in the future are much lower than they were in the past, households will need to do more saving in order to fund their retirements, and adequately mitigate the rather perverse “longevity risk” that retirement brings.
In his optimistic “Glass Half Full” speech last week, which came hot on the heels of Australia’s exceptionally healthy GDP and labour market data (to say nothing of three rate cuts in two months that have radically reduced the real returns earned by savers), Governor Stevens argued that household wealth “will surely resume an upward track, sooner or later.”
Now if you want to be negative, this contention is debatable, and the experience of Japanese savers is a sobering counterfactual. But let’s assume Governor Stevens is on the right track, and that there are credible differences--population growth chief amongst them—that argue against a convergence in Australia and Japan’s economic destinies.
Governor Stevens’ key message was that the returns savers realise in the future will probably be less than half of what they have earned over the last 10 to 15 years. More exactly, he said, “When [wealth does start rising again]…it is unlikely to be at 6 or 7 per cent per year in real, per capita terms. I would guess that over the long term, something more like 3 per cent would be nearer the mark.”
If Governor Stevens is right, savers should budget on total net nominal asset price growth of about 6% per annum--ie, adding in expected future inflation of about 3% per annum--across a well-diversified portfolio comprising, as it historically has, exposures to cash, shares and property.
For the avoidance of any doubt, Governor Stevens believes this is a big deal: “I think this is a profoundly important point and worth emphasising. The decade or more up to about 2007 was unusual. It would be quite surprising, really, if the same trends – persistent strong increases in asset values, very strong growth in per capita consumption, increasing leverage, little or no saving from current income – were to re-emerge any time soon.”
Another way to define asset price growth expectations is to look at sustainable, or “steady-state”, rates of economic growth. One school of thought says that no individual sector can out-grow the overall economy for any lengthy period of time otherwise it will eventually become bigger than the economy itself.
Since 1974, Australia’s real (after inflation) economic growth rate has been 3.1% per annum. But underlying this figure is quite a bit of population growth. As the population expands, we tend to produce more. If we, therefore, strip out population growth, we find that real per capita economic growth has only been about 1.7% per annum.
As I have shown before, Australia’s central bank has a strong statistical bias to overshooting, rather than undershooting, the mid-point of its “thick” 2-3% inflation target. Over the last decade, core inflation has averaged about 3% per annum. And there is mounting evidence that inflation targets are becoming less important to central banks, including Australia’s, as they wilt under community pressure to foster growth.
Let’s nevertheless assume that future inflation remains contained at about 3% per annum. Combining this with our real per capita economic growth estimate of 1.7% per annum gives us future nominal growth of about 4.7% per annum.
Observe that both these numbers are significantly lower than the figures offered by Governor Stevens: he gave us a real asset price growth estimate of 3% per annum, which in my nominal terms would be closer to 6.0% per annum.
What might explain this disconnect? In all likelihood, the fact that the household sector is a major supplier of riskier “equity” capital for both listed companies (ie, shares) and property. Who, then, provides the non-equity, or debt, capital? The banks and bond markets.
When thinking about investments, people forget that the share market is a leveraged play on the Australian economy. That is, most listed companies fund themselves with a combination of debt from the banks or bond markets, and equity capital from shareholders.
The same underappreciated insight applies to housing. Over the last 25 to 30 years, Australian residential real estate has generated a total net return (including both capital gains and rents after transaction costs but before taxes) of about 10% per annum.
Yet gearing across the entire housing market is about 30%, and for those property investors with some debt (as opposed to those that have paid it off), gearing rises to around 45%. Thus the actual returns to equity investors in housing would be significantly higher than the 10% figure I quoted above. This makes intuitive sense because, like equities, an individual property has very high return volatility of about 15% per annum.
So irrespective of whether you are an investor in shares of Australian listed companies, which in the non-bank sector have average gearing of about 25% (note the banks are leveraged about 20 times), or in residential real estate that offers even more attractive leverage of up to 90% at lower rates of interest and longer terms, your “equity” returns (and risks) are going to be greater than the blended equity plus debt returns reflected in my aggregated 4.7% per annum asset price growth estimate. This blended return is like averaging the returns that both the bank and you get.
And this analysis probably helps explain the difference between real per capita economic growth of 1.7% per annum and real household asset price appreciation (in per capita terms) of more than 3.0% per annum.
