When George Soros said “Every bubble has two components: an underlying trend that prevails in reality and a misconception relating to that trend” he went on to add “When a positive feedback develops between (that) trend and the misconception, a boom-bust process is set in motion.” That is, sentiment takes over in the last phase of any aggressive uptrend until the bubble is burst and results in a market crash. But he, like most other bubble commentators, fail to explain how that reality plays an important part in shaping that misconception itself. The reality or fundamentals of a bull market bubble often help determine the nature of the sentiment and that, in turn, has a knock-on effect on both fundamentals and how that sentiment is traded. It may intrigue some (but probably not many) to know there can be a positive feedback loop about a positive feedback loop. But what interests us at MatrixTrade, as a research company that trades and issues trading signals, is that this interaction help creates important tradeable milestones. In other words, there are important fundamental (not just economic) markers in an aggressive bull market or bubble that help us identify where we are in the trend so we can make money on the way up and on the way down. The purpose of this article is to show some of those fundamental milestones.
The comparison between the 1920s and 1980s bull market charts and the current uptrend is striking.
We have successfully made a great play of this fractal comparison for much of the uptrend not just from a technical perspective but also from the other core elements of the Matrix FITS approach (Fundamental, Intermarket, Technical and Sentiment). Today we demonstrate how the fundamentals in each of the 1920s 1980s and the 2010s are not only similar but have and should continue to shape the current bull market until the market can no longer sustain exorbitant PE values.
By fundamentals most people think of economics. But the fundamentals of, what we believe will be, the third largest stock bubble in 100 years are so significant that the reasons for and consequences of the uptrend that started in 2009 involve many aspects of the Great American society and indeed the world. The development of the 1920s, 1980s, and 2010s global bull markets can therefore be compared from more than just an economic and financial perspective. Not least there are also strong parallels between the development of financial market technology in these decades that helped fashion these uptrends and, probably more importantly, the subsequent reversals. All these combine to provide a check list of where the market is in the countdown to a top.
These milestones come in the form of theory (economic and financial), narrative (social, political, international and market which will be available in a later article) and the all-important data (available to paid subscribers as an appendix). The aim is not really to write a history book but make money in what may be known, in later years, as the Trump Bubble. To do that we need to understand what is really going on.
In the introduction to this series of seven articles on Milestones in Mania we outlined the now classic Kindleberger/Minsky model of Bubbles. They correctly identified the beginning of each bubble with a real cause. But the KM model, like most other models, then leave the fundamentals aside and focus more on the development of the bubble in terms of stock market sentiment. They therefore omitted to explain how the real cause serves to underpin and characterize the entire uptrend. The business model we have created could be thought of as a product cycle but that product or innovation has such an impact on the economy through pre-existing conditions, government policy and self-perpetuation that the product cycle becomes the bellwether for the entire economy and consequently the markets. More so than even the dot.com bubble of 2000. Rather than the Roaring Twenties, the Jazz era could be more rightfully be called the Mass Machine Age, the 1980s the Computer Age, and the 2010s? The all-pervasive Internet or Web 2.0 Age. In short, they are all technology bubbles.
THE MATRIX FUNDAMENTAL BUBBLE
There are essentially five fundamental stages to a Bubble that help us identify markers:
Innovation, Adversity, Universalisation, Saturation and Financialization.
Many would not associate this first phase with the Bubble itself as it refers to the previous period of ground breaking invention. But it warrants inclusion as it drives the later (current) bubble. What Kindleberger and Minsky describe as the ‘displacement phase’ is often incentivised by the final stage of the preceding boom and stock market rally. But the impact of that invention on the economy and stock market does not gain traction until the conditions are ripe for its development into an affordable mass market. Indeed we may well now be seeing the invention of product(s) that will drive the next bubble in 10-20 years time but not realise it.
1900s-1914: Affordable car (Model T 1905), fridges for home (1913) , television, and radio
1978-1981: Home computers
Mid 1990s: The Internet.
ADVERSITY: A period where real disposable incomes contract.
The time between invention and mass popularisation in the decades in question was interrupted by a period of economic adversity and stock market downturn. This adversity served to misallocate (and depress real) resources but in a way that later allows them to move far more freely into the development of the new technology as the economy improved. This period sees the stock market bottom and start a major bull market.
The First World War and aftermath both delayed but later assisted the mass production of machine technology for home use. OPEC inspired inflation and recession in the 1970s similarly slowed the development of the computer into the mass home product of the 1980s. But the death of inflation and economies of scale soon made computer prices affordable. And although the internet was well established at the of the 20th century, it is only since the sub-prime crash of 2008 that the internet has become all-pervasive through affordable hand held devices and mainly free social media.
Universalisation: Mass production of a good or service to a point where most are able and do buy it.
As the US economy came out of these periods of adversity, spare capacity in the economy facilitated the relative speedy development of innovations previously available only to a few. As they become mass markets , the stock rally gains momentum and the economy starts to grow rapidly. But more than that, the actual technology has in all three cases helped spur even greater productivity gains. The universalisation period typically produces the strongest economic growth in the entire bubble.
1920s: The post WW1 economy developed quickly as new technologies became cheap when mass manufacturing methods, eg Ford’s assembly line, were perfected. The infrastructure for the masses (paved roads, household electrical and telephone wiring) was in place. Companies active in new technologies were deemed to have immense growth potential. Although motor vehicles, radio, movies, the telephone, and significantly, stock tickers remote from the exchange floor had been invented in the nineteenth century, and available to the privileged few before World War I, these technologies only came in to widespread usage by the masses in the 1920s. It was the first era of mass communications, vital to the creation of market volatility. Notably these new industries themselves were public joint stock companies offering unlimited promise. Radio in particular saw stratospheric valuations at various times in the decade.
1980s: The personal computer revolutionized the way people did business even if they only used it as a word processor. Demand was huge. Companies involved were stock market darlings and, as usually happens, spawned a whole set of new satellite industries (software, maintenance, supply chains). A similar but more limited effect applied to the other great 1980s technologies, video recorders and multi-channel (satellite) broadcasting.
2010s: The internet had already changed the world, and by the start of this decade, virtually everyone was connected. As collaborative technology and software was developed, social media exploded. The internet became indispensable in daily life, as whole categories of consumer life (music, maps, newspapers, banking) became applications on a smartphone. Winners in the space, such as AAPL, GOOG and FB, exploded in value.
The mass manufacturing boom of the 1920s, the computer boom of the 1980s or the all-pervasive internet (Web 2.0) recovery would all probably have happened without the help of business friendly and expansionist governments and arguably more flexible economies. But probably not with the same speed. In the 1920s income tax was steadily cut as GDP expanded from $73Bn in 1922 to $97Bn in 1928. Rates hovered between 3.5% and 4% during the period. In the 1980s the overriding challenge was clearly to reduce stagflation and after that interest rates, and this was largely achieved. Enterprise was stimulated by tax cuts. It is not surprising the market continued to rise on these measures. The last decade has been dominated by the need not just to recover from the 2008 crash but avoid its recurrence. In complete contrast to the 1980s, deflation was the problem. With relatively low inflation and interest rates, quantitative easing was successfully introduced as a means of supporting and reflating the economy. Fiscal policy has remained relatively neutral until the arrival of Donald Trump who promises a far more business friendly fiscal expansion.
The finance sector in each decade also enjoyed a similar experience. Markets generally (not just exchanges) benefited from the new technology making them more accessible and provided financial institutions with the means and incentive to attract more investors to the market place and drive the stock market higher. The opposite of the Perfect Storm. Not quite Serendipity more like BuytheDipity.
Saturation : The slowdown in growth caused by many more companies joining a new industry.
As profits in these key markets attract new entrants they approach saturation point in the later but not final stages of the stock market uptrend. Unless there is sufficient spare capacity in the economy or significant productivity gains through economies of scale then the growth of this dominant sector and the economy starts to slow. Some of the early entrants disappear as they fail to keep up with the pace of innovation in a highly competitive market place. This is an early warning sign that the fundamental uptrend is beginning to deteriorate and lacks the breadth to be sustained. This is reflected by a decline in the growth rate and most notably a steady decline in earnings growth in all three decades.
1920s: There were allegedly 100 automobile companies in Detroit in 1900 but by the second half of the 1920s GM, Ford and Chrysler had 70% of all. The industry had become characterized more by economies of scale than innovation.
1980s: Previously dominant mainframe computer companies like Digital Equipment Corporation and Honeywell lost out. For example, in the bubble year before the 1987 crash when Intel shares rose over 300%, Honeywell shares only managed a 10% rise.
2010s: Because of the crazy valuations of the 1998-2000 tech bubble (353% in 17 months), there was a huge saturation clear out in the previous decade leaving space for the formation of almost accidental monopolies such as FB GOOG and NFLX. Even AAPL is growing inexorably proving that smartphones are not saturated yet.
This saturation point sows the seeds of the final leg in the upcycle
This last typically most aggressive phase of the bubble is driven more by greed and sentiment than fundamental reasons. And exactly where we are now. The gulf between reality and expectation in this last phase typically grows wider and wider. The gap between expectation reflected by stock market prices and what Donald Trump can deliver almost seems to get bigger each day.
Both the KM and Rodrigue models highlight how some of the smarter earlier investors and pioneers crystallize the stock market value of their companies by selling out, often either to start a new business or acquire more. Both the 1920s and 1980s witnessed a relative explosion in merger and acquisition activity.
1920s: When Walter P. Chrysler bought the famous Dodge car company in 1928, it was said “he rescued a failing business which was barely meeting its payrolls”
1980s: This was the era of (often hostile) M&A and leveraged buyouts (LBOs) without any necessary synergies to create conglomerates (common ownership of businesses in diverse sectors). However these were not generally in computer technology. For example IBM only made two acquisitions in the 1980s (compared to 70 in the 2010s and even three in the 1920s)
2010s: The main realization was FB, the largest IPO of the decade in 2012. The company was only eight years old and is now the fifth largest company in the world. The other notable financialisation was LKND, bought by MSFT in June 2016 for a 47% premium on the traded price, and most recently SNAP, which launched in 2017 at a larger market cap than American Airlines or Hilton Hotels, showing that tech enthusiasm is undiminished.
The financialization phase, by no means, entails the end of the new technology. Rather it has reached a maturity where its growth is more dependent on acquisition then further development of the technology itself. Partly as a consequence, an important tell-tale feature of the final financialization rally in stocks is resurgent inflation.
This is in stark contrast to the implosion in dividend yields in the final leg that marks the very last stage of the bull market and the widest gulf between perception and reality.
The 1920s, 1980s, and the current decade show a remarkable similarity that we have tried to evidence with this model. The reason the dot.com bubble of 2000 does not belong in this direct comparison is threefold. Firstly, the exorbitant rise in stocks and PEs was reflected mainly in the NASDAQ. The other indices did rally and reverse but not on a bubble scale (although it also finished with rising stock values, interest rates and Dollar). Secondly the dot.com bubble in NASDAQ reflected the first innovation phase that helped create the current and second much larger and more pervasive bubble. Thirdly the social, political and international narratives for 2000 are, partly as a consequence. sufficiently different to those of the 1920s, 1980s and the current decade. These will be the subject of a further article. They also contain markers by which we can judge our position within this tremendous bull market.
The more we can quantify these milestones, the easier they are to trade. This quantification is important as the nearer we get to the top the more commentators will cite a whole series of arguments to show this time is different. Super low interest rates do partly explain why our targets for PE and CAPE have always been more aggressive than many would have believed. Spare capacity is always hard to gauge and growth rates may be lower (certainly compared to 1920s) but it is the change in growth rates and the change in tell-tale signs of full capacity eg inflation, that are important even in a now de-industrialised economy with a rapidly changing labor market structure. Yes, Donald Trump could do a 1953 Eisenhower and prove us wrong by swelling capacity and productivity in a sustained rally that never returns to current levels. But it is highly unlikely. And fortunately, we can rely on an important data set to prove us right or wrong before we commit either way.
So when people tell you this time is different based on price rather than reality, ask them for their evidence. The evidence we provide in our fundamental data set for these periods (in an appendix for subscribers) hopefully demonstrates that this time is no different to the 1980s nor the 1920s.
The evidence suggests, just like before, dividend yields indicate when the last leg has started - confirmed by Productivity, corporate earnings that the top is close, that inflation is the trigger to get out of longs, GDP the sell signal and yields confirmation to sell more. And fed policy? That you have missed the trade of the decade. How one exactly interprets and converts the evidence into a generational trade will be the subject of our last article.
The following charts, grouped into Economy, Company and Inflation sections, reveal a strong similarity between the 1920s, 1980s, and 2010s and show that the S&P is closer but not yet at a top. Although we believe the eventual reversal is months (six months) and possibly quite a distance away (20-30%!) we thankfully do not have to commit to that. The existence of certain conditions prior to and indeed at the top could allow us to fine tune a trading strategy based on the fundamentals alone. These strongly suggest that Trump uptrend is not sustainable and that its very foundation – the expectations of inflation and growth – are the very triggers for the Trump crash.
GDP is the clearest and strongest of all the economic indicators and demonstrates the three phases of the bubble: The universalisation period which witnesses the strongest and fastest growth. The Saturation period where economic grown starts to fade and which is a warning that the economy/rally is no longer sustainable. And the financialization period which sees a sudden resurgence in growth. As this soon as the growth rate turns down that has previously indicated the market is at a top. (Note 1987 flattened once before the top).
Productivity growth fluctuates in a fairly constant range in the 1980s and 2010s (data not available for the 1920s). But, if 1987 is any guide at all, the next surge in productivity coincides with the very final rally.
Similar to GDP, earnings growth tapers off after the initial ramp until a final flurry. Again, as these flatten out it or turn down, this suggests the stock market is very close to a top.
Dividend yields remain in a range for the period until the last very leg of the uptrend when they capitulate. This is strong evidence that we are yet to see the aggressive climax. Other than there is no warning of the top.
Although both PE and CAPE are now above the minimum 25 threshold for a top, both the 1920s and 1980s showed a reasonably proportionate momentum in the last year and, combined still with a relatively low interest rate environment, suggest record levels for PE of above 30 and CAPE above 35 before the market turns.
INFLATION / RATES
Inflation is the second strongest indicator and reveals a close similarity to the last 5 years into both the 1929 and 1987 top. The current upturn in inflation is the last but suggests a period of flattening out for three months before it rallies as an indicator of the very final rally. It almost appears that inflation is the trigger for the reversal.
The non-rate nature of quantitative easing makes comparison difficult. And although rates turned back up in the last two years (suggesting a top by December) the extent of the rate is impossible to gauge given the lack of any previous scale. Suffice it to say, the next rate cut indicates we have already seen the top for a 30%+ decline in stocks.
The 1980s US Note market and the current are surprisingly similar given the very different preceding decades. However, as in previous inflation indicators, yields turn higher 12-18 months prior to the top. In the 1929 they starting back down ten months before the top and coincides with our view of an August-September top.
Milestones in Sequence
The relative consistency between these different sets of varying indicators for the three decades is significant. Firstly, they demonstrate a strong interrelationship between economic variables necessary for any economic model to perform. Secondly, despite their inter-relationship, the fact that different indicators are consistent with the same outcome, ie blowout and crash, increases the probability of that outcome and the position of any one indicator in that process. This is important particularly when we create a sequence of indicators with specific values.
The following table shows the sequence, priority and values of specific indicators in order to position the market within the last stages of the 1987 and 1929 style blowout.
This is not a precise science and our projected values could be significantly wrong. But it is the sequence and trend of these indicators that is of paramount importance here. In other words, it is necessary for each condition to be satisfied before the next can be met and the bubble burst. Should dividend yields collapse, and even PEs hit target this could still be a sustainable bull market until inflation and growth flatten or turn lower. This places greater emphasis on monitoring these two pillars of Trumpflation to determine the last stage of the uptrend. It also highlights the importance of leads, lags and reporting delays.
Lead Lags and Delays
It is all very well going back through history to determine a sequence, conditions and values if we cannot apply them in practice. All our economic data charts correspond to the period they are measuring and not when they were reported. This is one reason why we have excluded employment as an indicator (also data is insufficient in the 1920s) as it is a notorious lagger with respect to the economy and the stock market. Similarly economic growth, a key driver and variable in all 3 bubbles is measured quarterly and, from a trading perspective significantly, in arrears. However, we are fortunate in that leading indicators for growth such as PMIs represent a reasonably good current proxy as evidenced by the current close similarity to PMIs in the 1980s. In 1987 PMIs recorded (triggered?) a notable drop from the top and then made one further peak before the actual crash. An example of prices leading the fundamentals.
We believe strongly in the efficacy of this model and supporting data. But we do not have to rely on this evidence alone. At Matrix we analyse the varying (FITS) combination of Fundamentals, Intermarket and Sentiment drivers to determine our trading (signal) strategies. It is how these are now combining that makes us confident that we are still some way off the top of this bubble. Their interaction and resulting strategy will be the subject of our last article in this series.