Predicting future market returns using the Buffett indicator

Welcome to Finance and Fury.

The returns for shares over the last few years has not been good – especially in comparison to other assets – Is it still worth it to invest in shares? Afterall, the Australian Share index is sitting almost the same point as 18 months ago, whilst other assets are up over this period

1. To answer this question, Today – we will be looking at what is known as the buffet indicator – see how well this predicts the future returns of the markets
2. To start – clarify some things
   1. When talking about Shares – Monolith of the market – ASX 300, but this is made up of individual companies
3. Big disclaimer – historical data is no indicator of future performance – this episode is not intended as advice but as a general discussion around the usefulness of the Buffett indicator

Different assets go through cycles:

1. Property – In comparison to shares, this has done well over the last 5 years
   1. Australian house price index – 192 to 251, after declining from its peak of 273 over the last 12 months
   2. Gain of 31% over the last 5 years, but in two years from 2020 to 2022, it went up 42%
2. Gold – often has a negative correlation to shares – but gold has had a great 5 year period – in AUD it is up around 70% over this timeframe –
   1. Two reasons – high uncertainty, inflation – but also the currency to USD to AUD – as in USD it is up 52%
   2. But from 2012 to 2018, the returns were almost 0% over this whole time
3. Shares are similar, go through positive and negative periods

Good time to buy?

1. Looking back through history – Since 1900 – Average return of 13.21% p.a. – Very long term average – Includes Divs – Accumulation index
   1. 23 (19%) negative years and 99 (81%) positive years – takes on a distribution curve, but with a left hand skew around a return of 0% – this means there are many more numbers in the positive than negative
      1. 52 of the 122 years have had a return of 10% to 20% – but this does have outliers as well, like a negative return of between -50% to – 40% in 2008, then positive years of 60% to 70% in 1983 and 1975
   2. Past 5 years of returns, from 2018 to 2022 – there have been two negative years of returns so 40% which is statistically double what the historical data points to – also, one of those years returns was in the low single digits at 2%, but there were two good years in between – so the average returns over the last 5 years has been 8% p.a. and is up by around 43% cumulatively
   3. But it is rarer to get two negative returns in short timeframes
   4. Back to back negative returns – 5 times throughout 122 years – and within a 5 year window, was between the 1970s and 1980s
      23. 1973-1974 – -23.3%, -26.9% (-44% cumulative) – Gain of 441% before (10 positive, 2 negative years)
      24. 1981-1982 – -12.9%, -13.9% (-25% cumulative) – Gain of 584% before (6 positive years)
      25. But then in 1983, the market went up 67%, before having another negative year of -2%, then having two positive years of 44% and 52% respectively, before a -8% return in 1987
      26. Then the market went up by 18% and 17% prior to declining in 1990 by 18%
      27. So this period was volatile, with higher than normal negative returns – but what we haven’t seen are the same cumulative returns
      28. Example – take 1980 to 1994 – 15 years in total – 6 negative years or 40% of the time – you would think this may be a bad timeframe, but the cumulative returns even with all of these negative returns was 663%
      29. The last 15 years, taking 2008 to 2022 – 4 negative returns – but the cumulative return is only 110%
          1. The GFC is the outlier here, if we take 14 years from 2009 – it is 3 negative years, and 253% cumulative returns –
      30. But the GFC’s correction was as bad as it was due to the market being overvalued in 2007 – in 5 years the market went up 150%, compared to the last 5 years at 43%

This leads us into the Buffett indicator – as it is a measure of if markets are in a bubble, or fairly valued

5. The Buffett indicator is a valuation multiple used to assess how expensive or cheap a share market is at a given point in time.
   1. Call the Buffett indicator because it was popularised by Warren Buffett in 2001 – when he stated that it is “probably the best single measure of where valuations stand at any given moment”
   2. It is a simple calculations – looks at the total market cap of an index and divides this by the countries GDP – if the market cap (or total dollar size of the index) was $1 and the GDP was $1, then the Buffett indicator would be 100%, if the market cap was $1 and GDP was $0.5, then Buffett indicator would be 200%
   3. There is another expansion on this, where just instead of taking the GDP, you can include the total assets of a central bank – this can help to readjust for monetary expansion and is more important in the modern Central banking era

Looking at some numbers

1. Australia GDP Growth – The GDP has grown over the last 10 years – Current Annual GDP: $1.6 trillion US dollars or $2.5tr in AUD – but this was $780bn in 2003
   12. This is a growth of around 12.3% p.a. over the last decade
   13. note this growth rate includes the effect of price inflation – so it is the nominal growth rate and note the real GDP growth rate
2. Share Market Cap – The data comes from the World Bank – measuring the ASX 200 index
   2. They do some normalising with this to re-adjust some figures – but it is around $2.4tr AUD, but from $760bn in 2003
   3. Similar sort of growth, at 12.2% p.a.
3. Historical Total Assets of Central Bank – This includes treasury notes, bonds purchased, loans to commercial banks
   3. $603bn now but was $74bn in 2003 – biggest jump happened from 2020 to 2022 where it went from $180bn to $640bn, a growth of 3.5 times – but still, averaging out the growth it has been 23.3% p.a. over the last decade
4. Putting these all together
   99. Looking first at the Total Market Cap over GDP (%) – The current ratio of total market cap over GDP for Australia is 99.75% – it is almost one to one
       153. Over the last 20 year – the peak was 153.89% in October 2007; the recent 20 low was 74.65% Feb of 2009.
       154. The mean over this time period is 106.87% 
   100. When including the Total assets of the Central bank – the current Ratio for Australia is 80.03%
        141. Again, this reached its peak of 141.12% in October 2007; the recent 20 low was 68.07% in Feb of 2009
   101. What do these numbers mean – with a ratio of 99.75% and 80.03%
        1. These ratios indicate where the market is sitting versus valuations – ranging from significantly undervalued to significantly overvalued. But what is considered under or over valued changes between the TMC/GDP versus if you include total assets of the central bank
        2. For just the TMC/GDP
           1. Ratio of <75% = Significantly undervalued, 75% to 95% = mostly undervalued, 95% to 115% are fair value, anything above this is overvalued
           2. With a ratio of 99.75% = Fair value
        3. For the TMC/GDP + CB
           1. Ratio of <65% = Significantly undervalued, 65% to 85% = mostly undervalued, 85% to 105% are fair value, anything above this is overvalued = so including the CB assets, this decreases the classification of the valuation by around 10%
           2. With a ratio of 80% = Mostly undervalued

How well does this metric work when making investment decisions?

1. The aim of using the Buffett indicator is to tell if a market is over or undervalued –
   1. if it is undervalued, there is a higher chance that the returns will be higher than average due to the market cap reverting to its mean
   2. If it is overvalued, then the returns will likely be lower than average due to the market cap declining to return to the mean
2. The returns are not predicting the returns over the next year or two, but has a 8 to 10 years timeframe, looking at the average annualised returns above or below the average returns
   106. TMC / GDP ratio has 20 year mean of 106.87%.
   107. TMC / (GDP + Total Assets of Central Bank) ratio has a 20 year mean of 96.15%.
3. Expected future annual return has a higher chance of being above the average return of around 10% p.a. – under one model it is looking at between 10.5% to 12% annualised over the next decade
4. Historically, how has this performed in guessing the returns:
5. When looking at the data over the last 20 years, there is a correlation however, the actual returns do differ –
   1. Using this model can largely predict the trend in the share market as the actual returns have tracked the direction of the predicted returns – remember that this isn’t a comparison between yearly performance, but the annualised return over a timeframe of around 8 years
      1. So if the predicted returns are 10% and the actual returns are 9%, this doesn’t mean that this was one single years’ worth of returns but what was experiences when taking the sum of all the returns over this time and annualising them
   2. But whilst the direction or trends can match, the actual returns have historically been lower than what was predicted – by a range of 1% to 2% – so if the modified returns are predicting 12%, then the actual returns over the next decade may be closer to 10%
      1. These are all predictions – long term predictions at that – so next year may be negative 20%, before having strong performances to recoup the losses – investing in shares is a long term endeavour to the volatility
   3. Where it can go wrong – External factors
      1. Emotions – fear or greed can make the market lag or lead the indicators – people over buying or over selling compared to valuations leads to deviations
         1. This is why the Buffett indicator can be useful to show when markets are likely to underperform when they are overvalued, or outperform when they are undervalued
         2. But how the share market is likely to perform, or should perform, is different to how it actually performs
      2. Central banks and monetary responses – changes in the money base can skew the numbers as well
         1. If additional liquidity is introduced, or cheap money is around, and stays that way for a while, the market cap can become overvalued – the opposite is true when the money base is contracted

Summary – I find this indicator useful in telling where the valuations of the share market at large are sitting

7. But by no means should it be used as the sole basis for making an investment decision
8. Under the Buffett indicator, excluding CB assets, the Australian share market is fairly valued, and based around this sole indicator, over the next decade expected to return pretty normal returns compared to historical returns
9. Once this is modified to include CB assets, the market is undervalued – but only slightly
10. Investing is a long term endeavour to build wealth – this indicator is helpful to indicate if the market is overvalued, and if it has a higher chance to underperform, or outperform if undervalued compared to the historical averages
    1. At the moment, it isn’t significantly undervalued, so average returns are likely – but anything can happen in markets in the short term
    2. If CBS go on a cheap money spree for the next 20 years, average returns could be better, if they keep tightening rates and squeeze the financial system, lower returns can come from this
    3. Also – remember that this is looking at averages over decades – the returns year in year out will fluctuate, such as what we discussed – earlier in the episode – its not like it will constantly be 10% p.a. with no other volatility