Australian versus American equities: past, present and future

Overview

Is “the ability to grow your capital materially using a portfolio of Australian shares ... much more difficult than using international shares, US shares in particular”? Have “US shares outperformed literally everything in literally every timeframe,” and should this “be enough to have (Australian) financial advisers heading for the New York Stock Exchange and the NASDAQ”?

Is it true that “the results aren’t even close – whether it’s 5 years, 20 years, or 30 years, the US indices, when comparing for growth ... have destroyed the Aussie markets over time”? Above all, is there “literally not one statistic that will support AU shares broadly being a better current or prospective growth engine than US shares, not one”? In It’s too hard to genuinely grow a portfolio of only Aussie shares (28 October), and in his responses to comments to that article, Sebastian Ferrando makes these assertions.

They’re ALL demonstrably false, and in this article I’ll show conclusively that they’re untrue. I have no reason to believe that Ferrando’s trying to deceive others. Clearly, however, he’s convinced himself that what’s false is true – and on that basis has repeatedly peddled incorrect claims.

I’ve comprehensively and rigorously compared Australian to American equities over the past half-century. From this comparison emerge key facts and risks; this article summarises them. Have Aussie stocks generally underperformed? Bearing in mind these facts and risks, they plainly haven’t. Moreover, I’ll show why, over the next five years, the odds favour Australian equities to outperform American stocks.

The most important conclusion of my analysis, which I’ll elaborate in the final section, is that conservative Australian investors should take heart. Since 1999, Leithner & Company has focussed overwhelmingly upon Australian securities and almost completely eschewed American stocks.

My analysis reaffirms systematically what conservative Australian investors have long known intuitively: from the point of view of long-term returns, this choice is eminently sensible and entirely justifiable.

Five Key Errors and Two Overlooked Risks

Ferrando overlooks four crucial sets of facts (and one factually-based inference):

1. The $A-$US exchange rate: the more the $A depreciates against the $US, the higher, other things equal, are the returns of American assets denominated in $A. My analysis quantifies and removes this “exchange rate effect” – and finds that, depending upon the interval, up to 87% of the apparent outperformance of American over Australian equities thereby disappears.
2. The S&P 500’s most recent returns are unrepresentative: its most recent 12-month, five-year, 10-year and 20-year CAGRs are much higher than their long-term averages. These CAGRs are normally distributed; hence future ones are likely to revert – that is, fall – to their means.
3. The volatility of returns: over short-term, medium-term and long-term intervals, the All Ords’ returns have been less volatile than the S&P 500’s. By incorporating these returns’ volatility into my analysis, I demonstrate that, over periods of 30 years and less, the S&P 500’s “outperformance” of the All Ords is insignificant. Indeed, in simulations over periods of 20 years and less, American equities underperform Australian equities 40-50% of the time.
4. American equities’ huge margin expansion since the GFC: when I value the Australian and American indexes at sensible (by long-term historical standards) CAPE ratios, the remaining – that is, net of the exchange rate effect – apparent outperformance of American over Australian equities disappears.
5. The All Ords’ prospects are reasonable, but the S&P 500’s are poor: very high CAPEs, such as those which exist today in the U.S., don’t portend imminent crashes. Over the past half-century, however, they’ve augured poor (or worse) returns.

Ferrando’s apparently unaware of these key facts and this inference; he therefore overlooks two major risks:

1. An appreciation of the $A relative to the $US will, others things equal, reduce the returns (denominated in $A) of Australians’ investments in the U.S.
2. If it’s not different this time – that is, if American equities’ medium- and long-term returns revert to their long-term averages – then the rest of their apparent outperformance since the GFC (i.e., the portion which the exchange rate effect didn’t create) will disappear.

Australian Stocks Haven’t Underperformed; the $A Has

Using data compiled by the RBA, Figure 1 plots the $A- $US exchange rate on the last day of each month since July 1969. In December 1983, the Hawke government “floated” the dollar. Before then, monetary authorities utilised several types of fixed or managed exchange rate systems. From November 1976 until the float, for example, they managed the $A via a “crawling peg.” Under this system, each morning a committee of Treasury bureaucrats and the RBA’s Governor set its exchange rate.

In July 1969, $A1.00 was worth $US1.11. Since then, the mean rate has been $A1.00 = $US0.86. Throughout the 1970s and during the early-1980s, and again from September 2009 to November 2014, the rate was above-average; at all other times – and for more than a decade to September 2025 – it’s been below-average.

Between February and August 1974, the $A sharply appreciated to $US1.49. For the next 26.5 years, however, it depreciated virtually without interruption: by March 2001 it fell to its all time low of $US0.49. That was a total depreciation of ($0.49 - $1.49) ÷ $1.49 = 67% and a compound annual growth rate (CAGR) of -4.1% per year.

Figure 1: $A-$US Exchange Rate, Monthly, July 1969-September 2025

Then occurred the $A’s only extended period of appreciation: by April 2011 the rate was $US1.09. That was a total increase of 122% and a CAGR of 8.6% per year. Finally, in that month depreciation resumed and has continued virtually without interruption: in September 2025, $A1.00 = $US0.66. This has been a total depreciation of 39% and a CAGR of -3.7% per year for more than 14 years.

From its peak in 1974, the $A has depreciated vis-à-vis the $US by more than one-half (56%) and at a compound rate of -1.6% per year for 51 years.

Why does this matter? Movements of exchange rates can significantly impact the returns (denominated in $A) of foreign investments. The total returns of such investments derive ultimately from two factors: the total return of the investment denominated in the foreign currency, and the change of the $A’s rate of exchange with that currency.

Accordingly, a change of the foreign currency’s price relative to the $A can either increase or decrease your overall return denominated in $A (see, for example, “How the fall in the Aussie dollar impacts investors,” Morningstar, 20 October 2023).

Two sets of key points follow. Firstly, the lower the $A falls relative to the $US, the fewer $US-denominated assets you can buy after you convert your $A into $US. However, while you hold these assets, the more the $A depreciates vis-à-vis the $US the higher are your returns when expressed in $A. Conversely, the higher the $A rises relative to the $US, the more you can invest in $US-denominated assets after you convert them from $A – and the lower are your returns when you express them in $A.

It’s thus possible, given a sufficiently big depreciation of the $A relative to the $US, that a loss in $US can become a gain in $A. It’s also possible – we’ll shortly see a stark example – that, given a sufficiently big appreciation of the $A relative to the $US, a gain denominated in $US can become a loss in $A.

The general and sharp depreciation of the $A relative to the $US over the past half-century has thus greatly benefitted Australians who’ve invested in American stock (and bond, real estate, etc.) markets and held these investments for decades.

Figure 2, which plots the investments of $100 in a portfolio of stocks which perfectly mimicked (1) the All Ordinaries Index and (2) the Standard & Poor’s 500 Index, quantifies this key point. I’ve plotted the investment in the S&P 500 under three scenarios: 

1. what’s actually occurred (“Fluctuating $A”): each month, I’ve calculated the value of the portfolio in $A at that month’s prevailing rate of exchange;
2. counterfactual #1 (“$A Constant since 1974”): what if, continuously since January 1974, the $A had neither appreciated nor depreciated, i.e., remained constant at $A1.00 = $US1.00?
3. counterfactual #2 (“$A Constant since 2011”): what if, from January 1974 to December 2010, the actual exchange rate prevailed; but then, beginning in January 2011, remained constant the then-prevailing rate of $A1.00=$US1.00? 

(A crucial point regarding the two “counterfactuals:” what matters is not the level at which I set a constant exchange rate, but that it remains invariant over time.)

Figure 2: Values of Investments of $100 in the All Ords and S&P 500, CPI-Adjusted and Including Dividends, January 1974-September 2025  

Two sets of results emerge from Figure 2. Firstly, apart from the Dot Com Bubble the S&P’s outperformance before the GFC was minimal. Secondly, there’s a clear – and considerable – “exchange rate effect.”

Given the actual rates prevailing since January 1974, an investment of $A100 in a portfolio that perfected mimicked the S&P 500, including dividends and ignoring tax and costs of transactions, etc., and net of consumer price inflation, would have grown to $A8,929 in September 2025. That greatly exceeds an equivalent investment in the All Ordinaries Index: by September 2025 it would have grown to just $A2,308.

How much of the outperformance of the S&P 500 can we attribute to the generally depreciating exchange rate? Assuming a constant rate since 1974, by September of this year the investment would have grown to $A3,996. That’s just 55% (i.e., ($A8,929 - $A3,996) ÷$A8,929) of the actual amount.

Accordingly, by this measure the “exchange rate effect” accounts for almost one-half (i.e., 100% - 55% = 45%) of the Australian investor’s total return from her investment in the S&P 500 over these 51 years.

The “counterfactual” that the exchange rate has remained fixed since January 2011 – that is, that no depreciation has occurred since then – also prunes returns. Under this assumption, the investment grows from $A100 in January 1974 to $A5,895 in September 2025. That’s just two-thirds of the actual total, i.e., this assumption reduces the total return by one-third.

Figure 3 quantifies a key reality and resultant significant risk: adverse movements of exchange rates can convert foreign investment gains (denominated in $US) into losses (denominated in $A).

Figure 3: CPI-Adjusted Total Returns (CAGRs), All Ordinaries and S&P 500 Index, Three Intervals, January 1974-September 2025

Recall from Figure 1 that between August 1974 and March 2001 the $A depreciated virtually without interruption: indeed, by the latter month it had fallen to its all time low of $US0.49. That was a total depreciation of ($0.49 - $1.49) ÷ $1.49 = 67% and a compound annual growth rate of -4.1% per year.

This cumulatively huge depreciation boosted the returns (in $A) of American stocks. From August 1974 to March 2001, the All Ords CPI-adjusted total return was 8.7% per year; the S&P 500’s, however, was 14.2%. Much of this excess return is attributable to the depreciation: assuming a constant rate of exchange, the S&P’s return was 9.5%.

At this juncture, we can anticipate an objection: “what do I care that the depreciation of the $A-$US exchange rate has boosted the S&P 500’s returns expressed in $A? What do I care that an “investment” in American stocks necessarily entails a tacit speculation in the $A-$US exchange rate?”

The big problem – and the resultant significant risk – is: what the exchange rate giveth, it can also taketh away.

Recall as well from Figure 1 that from March 2001 to April 2011 occurred the $A’s only extended period of appreciation: in the latter month, $A1.00 was worth $US1.09. That was a total increase of 122% from March 2001 and a CAGR of 8.6% per year. During this interval, as Figure 3 shows, the All Ords’ CAGR was 5.7% per year – and the S&P 500’s was -6.3%. At a constant rate of exchange, the S&P 500’s CAGR was 0.6% per year.

The appreciation of the $A thus changed a small gain into a large loss. Cumulatively over this period, in $A terms Australian holders of a portfolio which mimicked the S&P 500 collapsed 48%, that is, (1.0 – 0.063)^10.1 = 0.518.

Finally, recall from Figure 1 that the $A’s depreciation resumed in April 2011: by September 2025, the $A was worth $US0.66. That’s a total depreciation of 39% and a CAGR of -3.7% per year. Figure 3 quantifies the effects. They corroborate the GFC effect: since April 2011 the All Ords’ CAGR has been 5.9% per year and the S&P 500’s 10.9%; assuming a constant rate of exchange, the S&P 500’s CAGR remains much the same.

Since 2011, in other words, the depreciation of the exchange rate hasn’t boosted the S&P 500’s results denominated in $A; something else – which I’ll shortly specify and quantify – has been at work.

Although by smaller extents, under the two exchange rate “counterfactuals” the S&P 500 appears to outperform the All Ordinaries. Equally, we can’t assume that the American index’s outperformance will continue indefinitely into the future.

Specifically, nobody can reasonably assume that the $A will continue to depreciate vis-à-vis the $US; most of all, we can’t assume that it’ll continue to devalue at the same rate as hitherto.

Hence the risk: a sharp and sustained depreciation of the $A from its current rate of exchange (say, to $US0.45) would please Australia’s exporters but concern consumers and the RBA (the former will eventually pay higher prices for imports, which would place upward pressure upon consumer price inflation; as a result, the latter would face rising pressure to lift its Overnight Cash Rate). A sharp and sustained appreciation, on the other hand, particularly if the S&P 500 generates mediocre results (never mind hefty losses) would decimate Australian investors in American equities.

That’s the risk; what’s the probability that it occurs? I don’t know; nor does anybody else. What’s certain – and what Figure 3 demonstrates – is that it’s happened before, and that its consequences were severe. It’s imprudent tacitly to assume that it can’t occur again.

The S&P’s Most Recent Returns Are Unrepresentative

Ferrando commits a second major blunder: he’s (1) apparently unaware that American equities’ most recent returns are by historical standards exceptional, and (2) assumes that such returns will continue. Hard data and basic logic imply that they won’t.

I’ve calculated the total, CPI-adjusted returns of the All Ordinaries and S&P 500 indexes (the latter includes the assumption of a constant rate of exchange since 1974) over ALL short-term (12-month) periods, that is, for January 1974-January 1975, February 1974-February 1975, ... , and September 2024-2025. I’ve also computed corresponding CAGRs and associated statistics for ALL medium-term (five-year), long-term (ten-year), and very long-term (20-year and beyond) periods since January 1974.

Table 1 summarises the results; three are most noteworthy. Firstly, they extend and elaborate the fact that the S&P 500 hasn’t outperformed the All Ords; instead, the $A has underperformed the $US.

Table 1: Short, Medium, Long and Very Long-Term Total Returns (CPI-Adjusted CAGRs), January 1974-September 2025

Net of this effect (in other words, assuming an invariant exchange rate), over all intervals the Ords’ average CAGR either exceeds (up to five years) or differs little from (10-50 years) the S&P 500’s.

Secondly, not only are the All Ords’ CAGRs mostly comparable to the S&P 500’s: over most periods they fluctuate less – and often much less. That is, their standard deviations are typically lower. If faith is confidence in things unseen, then volatility is risk made visible.

Finally, whether they’re actual (fluctuating $A) or counterfactual (constant rate of exchange), the S&P 500’s most recent returns over periods up to 20 years greatly exceed – and are thus unrepresentative of – their averages.

Over the 12 months to September 2025, for example, its total return has been 20.7%; that’s almost double the average 12-month return (11.2%). Over all other periods from five to 20 years the same is true: the most recent CAGR exceeds its mean.

Because these returns are approximately random, we can infer that the S&P 500’s future returns, whatever their duration, will revert – that is, decrease – towards their modal (that is, their most common and thus most likely) values, i.e., their long-term means – which differ little from the All Ords’.

Alleged Outperformance at the Cost of Higher Risk

A third major flaw undermines Ferrando’s claims: he ignores the variability of returns, particularly the greater – and often much greater – variability of the S&P 500’s compared to the All Ords’ over intervals of 20 years and less (recall Table 1). To see the implications, let’s first acknowledge that, at first glance, over all intervals beyond the short-term (12 months), the means of the S&P 500’s “fluctuating $A” returns in Table 1 are higher than the All Ords’. However, over intervals of up to 20 the standard deviations of the S&P 500’s returns are also much larger than the All Ords’.

The S&P 500’s average returns are higher than the All Ords’ but they also fluctuate more – and thus, by conventional financial standards, are more risky. In other words, the S&P 500 seems to offer the benefit of higher returns at the cost of higher risk.

To appreciate the implications, let’s also anticipate and address another objection: “never mind the hypothetical CAGRs in the S&P 500’s ‘constant $A’ column in Table 1: comparing actual results (that is, the CAGRs in the S&P 500’s ‘fluctuating $A’ column) to the All Ords’ CAGRs, the American index’s average CAGR exceeds its Australian counterpart’s over all time intervals.” “Extrapolating these differences five years, ten years and more into the future,” the objector might infer, “the result – that is, the S&P 500’s total outperformance – becomes massive.”

In response, let’s note that, assuming a fluctuating exchange rate, over all intervals up to 20 years, the standard deviation of the S&P 500’s outperformance of the All Ords exceeds its mean; in plain English, its average outperformance is statistically insignificantly different from zero. Assuming a constant rate, the same point applies to intervals up to 40 years (Table 2).

Table 2: Descriptive Statistics, Two Indexes, Two Assumptions, Seven Intervals

So what about this objection? The extrapolation of statistically insignificant outperformance five, ten and more years into the future is imprudent, hazardous and unreliable.

Firstly, that’s because it doesn’t take into consideration the variability of the two indexes’ returns – and particularly the bigger fluctuations of the S&P 500’s vis-à-vis the Ords’ returns. Moreover, such extrapolation tacitly assumes that underlying trends (such as the depreciation of the $A and rising CAPE ratio) will continue. Clearly, however, and as we’ve seen, trends can change – or reverse. This objection reflects a profound and irreconcilable difference between prudent investors and aggressive speculators.

Investors prudently regress to the mean; in sharp contrast, speculators overconfidently – and thus aggressively – extrapolate trends (whether real or, as is often the case, imagined).

The S&P 500 Has Underperformed the All Ords since 1900

At this point, critics would likely remain unmoved: “but what about the S&P 500’s actual outperformance in Table 2 over 50-year intervals? According to your own results, its outperformance is statistically significant.” In response to this objection, I’d say two sets of things. Firstly, bear in mind that I’ve analysed ALL 12-month, 60-month, etc., intervals since January 1974. Accordingly, as intervals have lengthened sample sizes have dwindled (Table 2).

The results of statistical tests depend, among other things, upon sample size and effect size. All else equal, and assuming in our case – and for Ferrando’s benefit – that the null hypothesis is false (that is, that S&P 500 outperforms the All Ords), larger samples provide more statistical power and precision; they thereby make it easier to detect effects. Inferences from smaller samples, on the other hand, are more likely to be unduly influenced by outliers, have lower statistical power to detect true effects, and are thereby more likely to produce misleading results. While larger samples are generally more reliable, excessively large samples may detect statistically significant but in practical terms meaningless differences.

These trade-offs raise our confidence that the S&P 500 doesn’t outperform the All Ords over periods of 30 years and less.

The possible problem with the 50-year CAGRs isn’t just small sample size: it’s their end points, which begin in January 2024 and conclude in September 2025. All of them are recent – and as we’ve already seen, returns with recent end points have abnormally high CAGRs.

This first set of responses, I strongly suspect, won’t satisfy critics. Hence my second set: why stop at fifty years? It’s easy to ascertain Australian stocks’ performance vis-à-vis their American counterparts over the longest interval over which data are available – that is, since 1900.

Elroy Dimson (Chairman of the Centre for Endowment Asset Management at Cambridge Judge Business School, and Emeritus Professor of Finance at London Business School), Paul Marsh (Emeritus Professor of Finance at London Business School) and Mike Staunton (Director of the London Share Price Database, a research resource of London Business School) have done so – and much more.

Figure 4 plots a portion of Table 1 (p. 16) of Credit Suisse’s Global Investment Returns Yearbook 2023 Summary Edition, which Dimson, Marsh and Staunton compiled. It plots the rank-ordered, total, constant-currency returns of the world’s five best-performing equity markets from 1900 to 1922; it also destroys this objection.

Figure 4: Top-Five Developed Nations’ CPI-Adjusted, Common-Currency Equity Returns (CAGRs), 1900-2022

A critic might try, but obviously can’t have it all three ways: he can’t simultaneously (1) deny the 12-month to 30-year results in Table 2, (2) laud the S&P 500's 50-year CAGR in Table 2, but (3) ignore the 120-plus year CAGRs in Figure 4.

Over this interval of more than 120 years, Australian stocks haven’t merely outperformed American stocks (CAGRs of 6.7% per year versus 6.4% per year respectively): they’ve outperformed every other country’s equities, and thus rank #1 in the world. Moreover, Australian stocks’ super-long returns aren’t just higher than American stocks’: their volatility (standard deviation of 17.4%) is lower than the volatility of American returns (19.9%).

The weight of reliable and very long-term evidence thus crushes this objection: since 1900, Australian stocks have generated slightly higher returns, at lower risk, than American stocks.

It’s likely (I don’t have access to Dimson et al.’s raw data) that this outperformance is, in statistical and practical terms, insignificant – that is, American and Australian stocks’ returns and volatility over this interval are effectively the same. 

If so, that continues to comprehensively disconfirm Ferrando’s core claim.

Performance versus Volatility; Statistical versus Practical Significance

Two crucial questions arise from the analyses I summarised in Table 1 and Table 2. Firstly, are the differences between the All Ords and S&P 500, which are mostly statistically insignificant, actually significant in practical terms? Secondly, what happens when we incorporate the fluctuations of both indexes’ returns – and particularly the S&P 500’s more variable returns – explicitly into the analysis?

To answer these questions, I’ve conducted seven (one for each interval from 12 months to 50 years) simple Monte Carlo simulations. These methods, sometimes called experiments, are a form of algorithmic analysis. They undertake large (in my case, 10,000 per interval) numbers of repeated random samples from probability distributions with given parameters; their outputs enable us to assess risks and opportunities under different assumptions and scenarios.

These simulations are well-suited to situations in which we must consider outcomes whose likelihoods vary. Monte Carlo simulations therefore provide a valuable but seldom-utilised tool in finance and investment; in particular, they’re an ideal way to quantify and assess stock markets’ possible and likely returns.

In essence, I’ve 

1. created a simple mathematical model of the processes (the S&P 500’s CAGR relative to the All Ords’ over each interval) I wish to analyse;
2. represented these processes not as single values (that is, an index’s mean CAGR and its standard deviation over a given interval) but as a probability distribution which has a limitless number of possible results; and
3. run a very large number (10,000) of simulations of each distribution.

These simulations enable us not just to assess the variability of the S&P 500’s performance, but also to estimate the probability that it outperforms the All Ords. On this basis, investors can better understand risks and make more informed decisions.

As a first step, I produced 10,000 simulated observations from a probability distribution whose mean is 8.4% and whose standard deviation is 17.4% (i.e., is identical to the All Ords’ CPI-adjusted, total 12-month return since January 1974). I then did the same for a distribution whose mean is 11.2% and whose standard deviation is 19.5% (in other words, is identical to the S&P 500’s CPI-adjusted, total 12-month return).

For each of these pairs of 10,000 simulated observations, I then (1) subtracted the All Ords’ return from the S&P 500’s; (2) calculated the mean and standard deviation of this series of 10,000 simulated relative performances, and (3) calculated the percentage of simulations whose value is greater than 0.0% (i.e., in which the S&P 500 outperformed the All Ords).

Table 2 summarised the 609 12-month returns for each index which have occurred; based upon these actual results’ parameters, each Monte Carlo simulation generated 10,000 results which could have occurred. I’ve also conducted corresponding simulations for intervals of five years, 10 years, ... , and 50 years. That’s 10,000 simulations per interval, and thus 10,000 × 7 = 70,000 simulated observations.

Table 3 summarises the results. They incorporate the volatility of both indexes’ CAGR into the analysis – and thereby refute Ferrando’s assertions.

Three results are paramount. Firstly, comparing the simulated outperformance under the two assumptions, it’s clear that the “exchange rate effect” is considerable. Over 12-month and five-year intervals and assuming a fluctuating $A, the S&P 500’s estimated mean outperformance is more than 2% per year. Assuming a constant rate of exchange, however, outperformance becomes underperformance of 0.3% per year. Over 10-year intervals the reduction of outperformance is 62% (i.e., (0.8% - 2.1%) ÷ 2.1%). Over intervals of 10-50 years, the average reduction of outperformance is 60%.

Table 3: S&P 500’s Performance Relative to the All Ordinaries – Results of Monte Carlo Experiments

Secondly, comparing the means and standard deviations of the S&P 500’s outperformance, it’s clear that only over 40-year and 50-year intervals, and only under a floating exchange rate, is it statistically significantly greater than 0.0%. Thirdly, comparing the probabilities of underperformance and bearing in mind a crucial caveat (see the next paragraph), only over intervals of 30 years or more do the odds of outperformance shift decisively in the S&P 500’s favour.

We can anticipate the next objection: “look at the columns you’ve highlighted in red. Over intervals of 40 and 50 years, American stocks almost always outperform Australian stocks.” This could be correct; but it’s more likely, particularly over 50-year intervals, that it’s questionable. It derives not just from small sample sizes but from CAGRs whose end points are very recent (recall Table 3 and Figure 4).

Clearly, as the size of the samples generating the simulated means and standard deviations increases – and as we move from longer to shorter intervals – the S&P 500’s probability of underperformance rapidly rises.

Hence my two conclusions from Table 3: 

1. removing the strong exchange rate effect and taking the volatility of returns into consideration, over no timeframe from 12 months to 30 years has the S&P 500 significantly outperformed the All Ords;
2. in practical terms, regardless of the assumption (fluctuating or constant exchange rate) and over intervals of 20 years and less, on a CPI-adjusted, total return basis the S&P 500’s outperformance of the All Ords is effectively a coin toss: it underperforms 40-50% of the time.

Two more general conclusions emerge. Firstly, properly incorporating the $A-$US exchange rate and the variability of returns into the analysis, there’s no reason to believe that the S&P 500 outperforms the All Ordinaries index over intervals of 30 years or less. Secondly, and bearing in mind small underlying sample sizes, it’s possible – but hardly certain – that the S&P 500 has outperformed over periods of 40-50 years.

But even if it has, rigorous research from respected academics concludes that American equities have underperformed Australian stocks over a period of more than 120 years (Figure 4).

The All Ords’ CAPE Is High – and the S&P 500’s Is Very High

Ferrando commits a fourth major blunder: he ignores the valuations – and relative valuations – of the All Ordinaries and Standard & Poor’s 500 indexes. Above all, he ignores recent valuations. Figure 5, which plots their cyclically-adjusted price-to-earnings (“CAPE”) ratios since June 1947, corrects this grave weakness.

Figure 5: Cyclically-Adjusted PE Ratios, All Ordinaries and Standard & Poor’s 500 Indexes, Monthly, June 1947-September 2025

Originally calculated with American data, CAPE has been extended to markets in more than a dozen other countries. Reputable research has demonstrated that the same relation between CAPE and future equity returns exists in every equity market so far examined; moreover, in these markets CAPE reliably estimates future market returns over 5-10 year periods (see, for example, Robert D. Arnott et al., “King of the Mountain: The Shiller P/E and Macroeconomic Conditions,” The Journal of Portfolio Management (vol. 44, no. 1, Fall 2017, pp. 55-68).

CAPE hasn’t escaped criticism. On the one hand, Jeremy Siegel, a professor at the Wharton School at the University of Pennsylvania and author of Stocks for the Long Run (McGraw-Hill, 1994), has stated that it “has served as one of the best forecasting models for long-term future stock returns” (see “The Shiller CAPE Ratio: A New Look,” Financial Analysts Journal, vol. 72, no. 3, May 2016). On the other hand, he reckons that Standard & Poor’s (which produce the data upon which recent readings of CAPE rely) hasn’t take into account changes in accounting standards since the 1990s; as a result, CAPE’s estimate of earnings is understated – and thus its inferences about future returns are overly pessimistic.

In response, in 2017 Shiller and a colleague (see below) devised a variation of CAPE which considered Siegel’s criticism.

The criticism failed: this variation continued to project that “expected returns should be low for the next six to 10 years based on historical” trends (see also “Siegel vs. Shiller: Is the Stock Market Overvalued?” Knowledge at Wharton, 18 September 2018).

CAPE has also been the subject of a large amount of uninformed criticism – seemingly from people who simply reject its implications. In “The Many Colours of CAPE” (Yale ICF Working Paper No. 2018-22), Shiller and Farouk Jivraj of Imperial College London investigated “the efficacy and validity of CAPE from several different perspectives.” Specifically, they found that “CAPE consistently displays economic and statistical significance far better than any of its peers.” They also explored “alternative constructions of CAPE ... (and found) that original CAPE is still best when comprehensively and fairly reviewing the other proxies ... ”

In plain English, never mind its critics (informed and uninformed): CAPE is, logically and empirically, superior to any alternative method of valuation – including those advocated by its informed critics (its uninformed critics typically offer no alternative).

Other studies (I won’t bother to cite them; interested and energetic investors can easily locate them) have corroborated this key conclusion. I’m unaware of any dispassionate and rigorous analysis – as opposed to idle and biased opinion from people unfamiliar with this literature and with an axe to grind – which challenges (never mind overturns) it.

Since June 1947 the All Ords’ CAPE has averaged 14.8 and its standard deviation is 4.6; the S&P 500’s has averaged 20.3 and its standard deviation is 8.3. As a whole since the Second World War, Australian equities have been more moderately priced – and their valuation has fluctuated less – than their American counterparts.

Nonetheless, before the GFC and with the notable exception of the Dot Com Bubble, the All Ords’ CAPE broadly tracked the S&P 500’s. Since the GFC, however, that’s changed dramatically.

Figure 6: Values of Investments of $100 in the All Ords and S&P 500, Including Dividends and CPI-Adjusted, Counterfactual “A,” January 1974-September 2025  

What if the S&P 500’s CAPE ratio hadn’t risen dramatically since the GFC? What if, for example, beginning in January 2008, when its CAPE (23.5) was much the same as the All Ords’ (23.0), the American CAPE henceforth equaled the Australian one? Figure 6 plots the answer. Given a fluctuating exchange rate, $A100 invested in the S&P 500 index in January 1974 grew to $A5,057 in September 2025. Assuming a constant rate of exchange, however, it grows to just $A2,263 – which is slightly LESS than the All Ordinaries ($A2,308).

Net of the (1) exchange-rate effect and (2) post-GFC CAPE effect, the outperformance of the S&P 500 disappears.

Alternatively, what if, beginning in June 2008, the S&P 500’s CAPE (20.9) henceforth equaled its average since January 1974 (20.3)? Figure 7 plots the answer. Given a fluctuating exchange rate, $A100 invested in the S&P 500 index in January 1974 grew to $A4,726 in September 2025. Assuming a constant rate of exchange, however, it grows to just $A2,115 – which, again, is slightly LESS than the All Ordinaries ($A2,308).

Figure 7: Values of Investments of $100 in the All Ords and S&P 500, Including Dividends and CPI-Adjusted, Counterfactual “B,” January 1974-September 2025  

It bears emphasis: net of (1) the exchange-rate effect and (2) CAPE’s post-GFC expansion, the S&P 500’s outperformance of the All Ords evaporates.

Multiple expansion occurs when a company’s or a market index’s valuation increases without a corresponding rise in its earnings. Under these conditions, market participants are willing to pay more per each $1 of earnings. Expansion occurs when the perception of a company or market index improves, often as a result of factors such as rising investor confidence, etc.

The conclusion is as simple as it is inescapable: multiple expansion is a key driver of investment gains – and multiple contraction is a key driver of investment losses.

Why the Ords Is Likely to Match or Outperform the S&P 500 over the Next Five Years

At this point, it’s easy to anticipate another objection: “why should I care about the huge rise of the S&P 500’s CAPE ratio since the GFC? For this very reason, Australian speculators in American stocks have recently outperformed Australian investors in Aussie equities!” Besides, and as Ferrando responded to a comment to his article, there’s “literally not one statistic that will support AU shares broadly being a better current or prospective growth engine than US shares, not one.”

Table 1, Table 2 and Table 3 demonstrated that Ferrando’s latter assertion is false. So does Figure 8.

“There is no way,” noted the Royal Swedish Academy of Sciences on 14 October 2013, “to predict the price of stocks and bonds over the next few days or weeks.” (Indeed, as I’ve demonstrated in various articles, it’s usually not possible over upcoming periods of 12 months). “But it is quite possible,” it continued, “to foresee the broad course of (returns) over longer periods, such as the next ... (five years and beyond) ...”

“These findings,” the Academy concluded, “which might seem both surprising and contradictory, were made ... by this year’s (winners of the Bank of Sweden Prize in Economic Sciences in Memory of Alfred Nobel), Eugene Fama, Lars Peter Hansen and Robert Shiller.”

Following Shiller’s general logic, for each month since January 1974 I (1) sorted American data by the current month’s CAPE and (2) calculated the total CPI-adjusted return (CAGR) during the previous five years and the next five years. I then (3) divided these data into five equal (by number of observations) segments. Quintile #1 contains the 20% of observations with the lowest CAPEs, Quintile #2 the next 20%, ... and Quintile #5 the 20% of observations with the highest CAPEs. Finally, I computed (4) each quintile’s average five-year CAGRs for both the previous and subsequent five years. I then repeated the exercise with the Australian data. Figure 8 summarises the results.

Figure 8: CPI-Adjusted Total Returns, by Quintile of CAPE, January 1974-September 2025

The key result couldn’t be more fundamental: valuations matter.

Specifically, CAPE matters. stocks’ current valuations drive their subsequent returns. In both countries since 1974, the HIGHER has been the CAPE during a given month, the LOWER has been the CPI-adjusted total return over the next five years.

For enthusiasts of American stocks, it gets even worse. In September 2025, the S&P 500’s CAPE rose to 39.5 – the highest since September 2000. This does NOT imply that a crash is imminent; however, and as Figure 8 shows, it DOES signal that it’s reasonable to expect poor returns over the next five years. Indeed, since 1974 whenever the S&P 500’s CAPE has risen above 35, its subsequent five-year CAGR has been -4.1% per year.

That’s a total, CPI-adjusted LOSS of (1 – 0.041)⁵ = 19% over the next five years. In contrast, owners of Australian equities can expect a total, CPI-adjusted return of ca. 3.5% per year. That’s a GAIN of ca. (1 + 0.035)⁵ = 19%.

At this point, it’s easy to anticipate yet another objection: “exclude the ‘Magnificent 7’ and the S&P 500’s margin expansion since the GFC greatly diminishes.” Figure 9 plots the current (September) CAPE ratios of each Mag7 stock. All are above (Apple, at 41.1, albeit marginally) the S&P 500’s CAPE (39.5); the Mag7 average, boosted by Tesla and Nvidia, is 134.

Figure 9: CAPE Ratios, “Magnificent 7” Stocks, September 2025

What’s the S&P 493’s – that is, minus the Mag7’s – current CAPE? Michael Lebowitz (“CAPE Valuations: Does Nvidia Overstate Its Ominous Warning?” Realinvestmentadvice.com, 30 October) writes: “if we strip Nvidia out of the CAPE calculation, the CAPE for the remaining S&P 500 will fall by nearly 3 points ... (and) if we take all (of) the Magnificent Seven stocks out of the calculation, CAPE will decline from 41 to 33.”

A CAPE ratio of 33 falls towards the top of the highest (#5) quintile of the S&P 500’s CAPEs in Figure 8. Since 1975, when a given month’s CAPE has fallen within this quartile, the S&P 500 has generated an average, total, CPI-adjusted return of 2.5% per year over the next five years. As Figure 8 shows, that’s no higher than the All Ords’ prospective return.

Including the Mag7, the most plausible inference is that during the next five years the S&P 500 will generate a loss – and thereby greatly underperform the All Ords. Excluding them, the most credible inference is that the two indexes will generate identical returns – that is, the S&P 500 won’t outperform.

Conclusions

Daniel Kahneman, a co-recipient in 2002 of the Bank of Sweden Prize in Economic Sciences in Memory of Alfred Nobel, regarded overconfidence as the “most significant of the cognitive biases.” It’s dangerous because it reliably produces poor – and worse – decisions in everything from investment and speculation to international relations.

In that respect, Sebastian Ferrando is remarkable. Seldom have I encountered somebody who’s so overconfident, so oblivious to (or, if he’s aware of it, unconcerned by) his confirmation bias – and, as a result, is so consistently and greatly mistaken. He’s exceptional, but he’s hardly alone: like Chris Bowen, Matt Kean and many others, his purpose is to protect his cherished beliefs no matter what; he therefore ignores and dismisses any and all disconfirming evidence.

It’s vital to reiterate what I’ve already stated: I have no reason to believe that Ferrando’s deliberately trying to deceive others. Clearly, however, he’s convinced himself that what’s demonstrably false is somehow true – and on that basis has repeatedly peddled incorrect claims.

Is “the ability to grow your capital materially using a portfolio of Australian shares ... much more difficult than using international shares, U.S. shares in particular”? I’ve shown that, over intervals ranging from 12 months to 122 years, it hasn’t been. 

Have “US shares outperformed literally everything in literally every timeframe”? Quite the contrary: removing the exchange rate effect and incorporating the volatility of returns into the analysis, I’ve shown that over no timeframe from 12 months to 30 years has the S&P 500 statistically significantly outperformed the All Ordinaries. And in a practical sense, results of Monte Carlo simulations demonstrate that in no timeframe of 20 years or less would it reliably outperform.

Should this “be enough to have (Australian) financial advisers heading for the New York Stock Exchange and the NASDAQ”? Not if they’re aware – as most of them probably are but Ferrando doesn’t seem to be – of valuation, volatility and exchange-rate risks.

Ferrando alleges: “the results aren’t even close – whether it’s 5 years, 20 years, or 30 years, the US indices ... have destroyed the Aussie markets over time.” Above all, there’s “literally not one statistic that will support AU shares broadly being a better current or prospective growth engine than US shares, not one.”

My analysis, drawing upon methods from Robert Shiller (who, in turn, elaborated and formalised insights from Benjamin Graham) has comprehensively disproved these assertions.

Ferraro’s confirmation bias has blinded him: he’s urging that Australians “invest” (“speculate” is more apt) in American shares at a time when (1) they’re very highly valued and (2) the $A-$US exchange rate is well below its long-term average.

He’s thereby unaware of two particularly salient risks: that (a) a reversion to less extreme valuations or (b) an appreciation of the exchange rate occurs. Each event could generate significant losses. I’ve demonstrated that these things have happened before; he can’t credibly imply that they can’t recur.

Ferrando’s also apparently unaware that a simultaneous occurrence of a hefty appreciation of the $A (as occurred from 2001 to 2011) and mediocre returns from the S&P 500 (as occurred from the height of the Dot Com Bubble to the wake of the Global Financial Crisis) would devastate the returns of Australians who own American equities.

As I’ve also shown, this has happened before. What’s the probability of a recurrence? Neither I nor anybody else can plausibly claim to know. What’s certain, however, is that this confluence of risks can’t simply be ignored.

My Results Won’t Surprise Conservative, Long-Term Australian Investors

Whether they’re American or Australian, what long-term return can equity investors reasonably expect? On 22 March 2019, The Sarasota Herald-Tribune quoted Warren Buffett: “stocks are a decent way to make (a long-term average CAGR of) 6-7% annually.” The historical returns of major indexes in both Australia and the U.S. over various intervals (Table 1) corroborate this expectation. Buffett therefore cautioned: “anyone who expects to make 15% (long-term) ... is living in a dream world.”

“Now, maybe you’d like to argue a different case,” Buffett challenged bulls at the apex of the Dot Com Bubble (“Mr Buffett on the Stock Market,” Forbes, 22 November 1999). “Fair enough. But give me your assumptions. If you think the American public is going to make 12% a year in stocks, ... you have to say, for example, ‘Well, that’s because I expect GDP to grow at 10% a year, dividends to add two percentage points to returns, and interest rates to stay at a constant level.’ Or you’ve got to rearrange these key variables in some other manner.”

Buffett concluded: “the Tinker Bell approach – clap if you believe – just won’t cut it.”

Conservative, Long-Term Investors in Australian Equities: Take Heart!

Ferrando’s confirmation bias generates the overconfidence which produces his false assertions. But he’s not my concern.

My concern, as a director of Leithner & Company, Education Partner of the Australian Shareholders Association and contributor to Livewiremarkets, is conservative, long-term Australian investors.

On the one hand, as a rule they’re diligent, knowledgeable and sensible. Accordingly, their principles are usually sound. On the other, the demands upon their time are great – and they’re constantly and relentlessly bombarded with what, frankly, is rubbish. To put it charitably, as Shiller did in Irrational Exuberance (Princeton University Press, 3^rd ed., 2015), “unknown to most investors is the troubling lack of credibility in the quality of research being done on the stock market ... Some of this so-called research often seems no more rigorous than the reading of the tea leaves.”

In Dividends aren’t a bane – they’re a boon (20 November 2023), for example, I demonstrated – from first principles and with 150 years of valid and reliable data, and never mind a prominent fund manager’s opinion – that dividends have long been and remain a pillar of conservative, long-term, successful investing.

On this basis, we can anticipate – and dismiss – a final objection. I’ve analysed total returns including dividends (but exclude franking credits from Australian dividends; accordingly, my calculations of Australian total returns will for many investors be conservative under-estimates).

Ferrando, in contrast, discounts and even denigrates dividends. That, he alleges, is because “when comparing for growth, (American stocks) have destroyed the Aussie markets over time.” Is he unaware of Buffett’s view – notice that Buffett includes dividends in his estimate of future returns – and Shiller’s and others’ research, or does he simply ignore them?

“The reliable return attributable to dividends,” concluded Shiller in Irrational Exuberance, “not the less predictable portion arising from capital gains, is the main reason why stocks have on average been such good investments historically.”

Of course, conservative, long-term investors knew this already: my purpose has been to reassure them, and to give them the confidence to uphold their sound principles in the face of distracting nonsense. So-called financial experts, first do no harm! Properly considering the risks, Australian equities have held – and, I’ve cautiously inferred, will continue to hold – their own against their American counterparts.

Obviously, risks are unavoidable and therefore success is hardly guaranteed; equally, for conservative, long-term Australian investors long-term rewards will tend to outweigh short- and medium-term risks.

Hence focussing almost exclusively to Australian equities, and eschewing American stocks, as Leithner & Company has done since 1999, is sensible and thus justifiable.

So too – whether on the basis of your own extensive research or with the assistance of a competent advisor, and with proper appreciation of exchange rate, volatility, valuation and other risks – can be the allocation of some portion of your capital to foreign (including American) assets.

But succumbing to the confirmation bias of Sebastian Ferrando, and his resultant overconfident and mistaken convictions about American versus Australian stocks, is unjustifiable – and, for conservative Australian investors under current conditions, imprudent.

Click LIKE so that Livewire knows that you want more of this type of content. Click FOLLOW on my profile for notification when my next wire appears.