Seven Paw-prints of the Bear


For decades, investors have sought out methods to detect oncoming bear markets. With this current bull market now in its 5th year the subject has become even more topical – “Has the bull market still got legs?” is a question pondered every day by millions of investors. In this research note, we cover seven of ten reliable Bear Market warning signs we use and how they can be combined into a simple stock market exposure allocation strategy.

But first, let’s define a “Bear Market”. A Bear market is a market condition in which the prices of securities are falling, and widespread pessimism causes the negative sentiment to be self-sustaining. As investors anticipate losses in a bear market and selling continues, pessimism only grows. Although figures can vary, for many, a downturn of 20% or more in multiple broad market indexes over at least a two-month period, is considered an entry into a bear market. In most instances (but not all) an economic recession will accompany (even trigger) a  Bear market. A Bear market should not be confused with a correction, which is a short-term trend that has a duration of less than two months. Although the Crash of ’87 was a major correction, it technically was not a bear market since the stock markets recovered shortly afterwards.

Our endeavor in this exercise is what we term “Investment Market Timing” which is to avoid major, extended stock market corrections, most often accompanied by recessions, in order to reduce brutal buy-and-hold draw-downs and outperform the buy & hold strategy on an absolute and risk-adjusted return basis, deploying as few trading transactions as possible.

Using robust models to detect for probability of oncoming economic recession is thus one method to avoid bear markets, but although this will allow you to dodge the worst damage, a recession rarely encompasses all of and overlaps perfectly with, the entire stock market correction. In general, the use of econometric models will allow you to match closely with some stock  market peaks (see “Recession : Just how much warning is useful anyway?“) but will sorely lag with the ideal  stock market re-entry points in many cases, since stock markets rebound way before the recession is over. Furthermore, in many instances a recession will not even lead to a bear market correction and in even more instances, a bear market correction will not lead to a recession! There is the well known jibe on Wall Street: “The Stock market has predicted 9 of the last 5 recessions” to illustrate this point.  Nonetheless, a decent long-run  investment return sans the gut-wrenching  draw-downs of most bear markets can be achieved through recession probability monitoring, as described in our Recession Forecasting Ensemble & Investment Timing research note.

You can make substantial improvements in your Bear Market detection methodology, and your subsequent stock market investment timing endeavors, using factors derived directly from the stock market itself. These include, but are not limited to, rate of change, moving averages and other technical indicators, market breadth indicators and even seasonal factors. We deploy ten of these, seven of which we will demonstrate in this research note. Before we continue, it is important to understand that major stock market peaks that precede bear markets take a long time to form. In fact, major stock market top formations are a rather drawn-out process taking up to 12 months in some cases. Major stock market top-detection techniques are diametrically opposed to the techniques used for detecting major stock market bottoms, which are violent rapid reversals that catch most people by surprise. To the astute observer, warnings of impending bear markets thus start making themselves known way before the bear market correction ensues.

We will examine bear market detection using seven methods drawn from a diverse array of indicators that have robust long-term (multi-decade, multi-business cycle) track records. The reason we deploy a diverse array of methods is that no single technical, econometric, breadth, sentiment or seasonal indicator is infallible and 100% correct 100% of the time. The deployment of a battery of indicators into a consensus model reduces the risks of over-reliance on any single factor. The consensus composite we build from these seven factors is dubbed the “Composite Market Health Index, or CMHI for short. For each of the charts we show below, we take the relevant indicator and standardize them to have a mean of zero and a variance of 1. This allows us to combine them into an equally weighted composite later.

As the stock markets approach their major peak, more and more of their individual constituent stocks fall into their own little bear markets. Eventually with only a few large cap stocks left propping up the market rise, the overall index has to succumb to the weight of gravity of its constituent falling stocks and the bear market correction ensues. It is a remarkably little-known fact that on the day of the peak, less than 10% of an indexes’ constituent stocks will be printing new highs. Thus one way to detect the onset of a bear market is to track the percentage of shares in the index that are printing new long term highs. Many people prefer to use the 52-week highs to do this, but we deploy our own measurement based on less than 52 weeks and subtract new lows from new highs to achieve more timely results.

As the stock markets approach their ultimate top, shares get more and more expensive forcing savvy investors to deploy extreme selectivity in individual share selection to minimize their risks of overexposure to over-valued shares. Fewer and fewer stocks get chased by a growing pool of investors and the stock markets gains are driven by fewer and fewer stocks. We can measure for this by tracking the percentage of constituent stocks that are still in medium to long-term up-trends.  When more stocks are in down-trends than are in up-trends for an extended period and the index is still rising, then extreme selectively is in play and we have our market top warning. You can use the % of stocks trading above their 200 day moving average as a quick and dirty calculation here but the metric we use is more optimized for timeliness and produces far fewer false positives.


Two indicators we deploy are derived directly from the SP-500 price action itself. One is a rate of change oscillator and the other is an oscillator derived from a set of dual crossover moving averages. Both of these have been optimized to provide stellar market timing performance since the post war period, allowing you to outperform the buy & hold by a factor of 1.4-to-1 with half the stock market exposure (being 50% in cash.) They are shown below since 1990 and are deadly accurate in detecting major bear markets, but unlike the breadth indicators are subject to more false positives and worse, heavily delayed re-entries. The 50-day and 200-day moving average of the SP-500 is a quick and dirty method you can use for this metric.

This is our second favorite bear-market indicator as shown with our “Select Paw-Print” logo! It is said that volume precedes price in the stock markets and this is certainly the case when we examine advancing volume less declining volume on a daily basis. This data appears wild and random on a daily basis but when appropriately smoothed and manipulated provides a rock solid indicator of internal supply and demand and thus liquidity. When advancing volume exceeds declining volume, demand exceeds supply and we are in a net liquidity inflow situation. When declining volume exceeds advancing volume, then supply  exceeds demand and we are in a net liquidity outflow situation. Extended periods of liquidity outflows are sure fire warning flags of an impending bear market. You can use the McClellan Summation Index (for volume)  as a quick and dirty calculation here but the method we use, although similar, is optimized for timely detection of large bear markets, fewer false positives and speedy re-entry on new bull markets. In fact this is one of the better non-econometric tools we have for flagging the end of recessions in a timely manner. Your first warning with this breadth metric is when it prints lower highs when the stock market is printing higher highs. This divergence is the first sign that the bull market is reaching its sunset stage.

This is our favorite bear-market indicator, also shown with our “Select Paw-Print” logo! This is a weekly composite of over 50 weekly time series grouped into five major sub-composites, namely Corporate Bonds, Credit Spreads, Term Spreads, Unemployment data and an economically sensitive stock/commodity  price index. We cover two versions of this for our subscribers every Thursday, dubbed WLIr (includes revisable components) and WLInr (includes only non-revisable items.) Its primary task is to gauge future U.S economic growth and monitor probability of economic recession (when it drops below zero) and is thus not as targeted directly at stock market timing as all the other metrics we have discussed. However it is here to ensure that if the risks of recession do escalate dramatically that we are pulled out the market if the other metrics are in a weak state. Your quick and dirty option here is to use the ECRI Weekly leading Index, but as is well documented over the last 18 months this has suffered with false positives lately. As we warned in the beginning you need to remind yourself that using this in isolation is not recommenced since not every bear market is accompanied by a recession and vice versa.

Remarkably, there is indeed a multi-annual cyclically persistent seasonality that exists to isolates certain favorable times to be in the stock markets and certain unfavorable times that seem to be  haunted by bear markets, crashes and recessions. We cover this extensively with a practical “Active Investing” methodology in our “Seasonal Methodology for Market Timing” research note. We have adapted this seasonality model for a more sedate version more suited to the avoidance of major bear markets since 1929. Whilst a convincing bulk of recessions and bear markets occur during the “BEARISH” seasons, with the remainder occurring during the “NEUTRAL” seasons, we gain more useful insight from the fact that a recession or bear market has never started in a FAVORABLE season!

We can now build a consensus model that at each point in time counts how many of the seven “paw prints” discussed above are below zero and flagging “Bear market”.

We modify this consensus vote with the formula: DIFFUSION = (3 LESS number of paw-prints below zero)/3. 

When the Diffusion is 1 it means all 7 indicators are above zero (very bullish) and when it is 0 it means 3 of the components are below zero and flagging a bear market. A diffusion of -1 means at least 6 indicators are flagging a bear market. You can use the Diffusion readings above zero to match your equity allocation to bear-market risk, such as : +1 = 100% stock market exposure, 0.67=67% stock-market exposure, 0.33= 33% stock market exposure and 0=zero stock market exposure (cash). We also combine all the seven indicators together into an equally weighted “Composite Market Health Index” or CMHI for short:

The CMHI composite coupled with its associated Diffusion (consensus vote) provide a means to gauge Bear market risk and to adjust your stock market exposure accordingly. Alternatively, an excellent timing model can be deployed that moves into cash when both the CMHI and Diffusion are below zero and remains in the stock market otherwise. Both these methods are not perfect – you can see they both flagged one to two near-misses in each bull market. But at the end of the day, after just one full business cycle, you will most certainly beat the buy and hold on an absolute and risk adjusted return basis. The occasional exits that don’t land up being bear markets are merely “insurance policies” that we don’t get caught trapped in the bear cave where all the damage happens.

Despite the fact that the CMHI ranges between -1 and +1, its long-term linear regression trend line is at the +0.3 level. It makes sense that its regression mean is above zero, since the stock market, and hence all the CMHI components, is going up more than it is going down. The bull markets are gradual and long and the bear markets are sharp and short. The saying “The bull climbs the stairs and the bear comes down the elevator” comes to mind.

If we examine the behavior of the CMHI around its long-term linear regression trend level 0f +0.3, as opposed to around the zero pivot point, we get an indicator that is extremely useful for timely warning of bear markets in U.S stocks:

About RecessionALERT

Dwaine has a Bachelor of Science (BSc Hons) university degree majoring in computer science, math & statistics and is a full-time trader and investor. His passion for numbers and keen research & analytic ability has helped grow RecessionALERT into a company used by hundreds of hedge funds, brokerage firms and financial advisers around the world.

, ,