Popular metrics of aggregate market valuation, such as Wilshire Total Market Index to U.S. GDP, price to forward earnings ratio, price to book value ratio, price to cash flow ratio, cyclically adjusted price to earnings ratio (CAPE), the ratio of annual forward dividend to price (dividend yield), indicate the U.S. stock market is overvalued by between 10 per cent and 60 per cent.
With so many indicators flashing red, what is the likelihood that the stock market may crash?
Normally, it is difficult to determine whether a bubble exists until it is too late. But even if we know we are in a bubble, it is difficult to foresee when it will burst.
Bubble, bubble
In light of this, a recent study on the topic of bubbles and crashes got my attention. Researchers at Taiwan’s National Sun Yat-Sen University examined forty world markets covering 30 years of data and found that stock market crashes come in different forms and that their causes are different. In a paper titled “Rational or Irrational: A Comprehensive Study of Stock Market Crashes,” the authors find that behavioural factors and other market microstructure issues are more significant than macroeconomic factors in explaining stock market crashes. They use an investors’ sentiment measure from google searches to get a sense of herding in the markets as a proxy of behavioural biases. They also use market liquidity and volatility as a proxy for market microstructure issues and inflation, current account, growth rate in money supply, industrial production and the unemployment rate for macroeconomic factors.
They define a crash event by examining three-month cumulative returns vs. an estimated threshold that includes the standard deviation of the three-month cumulative returns. When the three-month cumulative returns are below this threshold, there is a crash event. The highest stock price in the three-month period of the crash event corresponds to the start of the crash. The end of the crash happens when the future monthly return in a given market exceeds the average monthly return in this market by one standard deviation.
In the forty markets examined, there were 681 crashes between 1985 and 2015. The average stock market decline was 25% and the average duration of the decline was 4.68 months. The impact of the crash is more severe in developing countries than developed countries. The number of crashes ranges from six in Vietnam to 25 in India, the cumulative stock market decline ranges from -9% in New Zealand to -39% in Ukraine, and the duration of decline ranges from 3.65 months in Colombia to 6.54 months in Portugal.
By comparison, NASDAQ and Dow Jones had 21 and 22 crashes, a cumulative stock market decline of -20% and -13%, and a duration of decline of 4.29 and 3.77 months, respectively.
The Taiwan-based authors find that behavioural factors are the most important determinants of the cumulative decline and duration of decline. The higher the irrational behavioural factor the greater the probability of a crash.
Liquidity and volatility, on the other hand, affect the severity of the crash and the speed of the decline — the greater the liquidity and the smaller the volatility the less severe the subsequent crash. Moreover, the probability of the crash is negatively related to liquidity and positively related to volatility. Macroeconomic factors do not do as well as behavioral and liquidity/volatility factors in explaining a crash.
Is the U.S. headed for a crash?
So, what are proxies for the indicators saying about the U.S. market right now?
First, behavioural metrics.
The Put to Call indicator stands at 0.76 currently, a multi-year low, as opposed to 1.36 in August 2011 and 1.15 in October 2008. It is an indicator of bullishness in the markets and may signify that the herd is becoming extremely exuberant.
In addition, the Google Search Volume Index (SVI) for the question “when to invest” showed an average metric of about 70 over the last two years vs. an average of 29 for the years between 2004 and 2014. With a peak of 100, SVI indicates that this question has been very popular in recent years, especially when compared to the historical experience.
The last two metrics taken together do indicate that the herd is becoming bullish regarding investing in the stock market, which, according to empirical evidence, increases the probability of a crash.
Second, a liquidity metric.
The NYSE annualized monthly volume to shares outstanding ratio, a proxy for liquidity, currently stands at 64%, a level which is about the average for the last five years, and much lower that the preceding five years over which it averaged about 95%. Although liquidity has not changed very much in recent years, this indicator is low from a historical context and signifies an increased probability of a crash vis a vis earlier years.
Third, a volatility metric.
The VIX (the Chicago Board Options Exchange Volatility) index is at the lowest level in years. It is currently at around 11 vs. 48 in August 2011 and 68 in October 2008. According to the research cited above, this indicator points against the possibility of a stock market crash.
So will there be a market crash? The proxy indicators based on the academic study referred to above give somewhat ambiguous signals about the possibility of a crash, despite the apparent overvaluation of the U.S. stock market.