Market Uncertainty Cycles
“For uncertain matters there are no calculable probabilities whatsoever. We simply do not know.”
- John Maynard Keynes, A Treatise on Probability 1921
The Market Uncertainty Cycle is where the four systems coincide at a time scale that most affects our perception of investments. It’s similar to the way the daily weather report affects our perceptions more directly than projections of climate change. It’s where we can take the most action.
My original market state indicator was called the “Market Risk Indicator” or “MRI.” But that was for marketing purposes. What we’re really talking about is market uncertainty rather than risk. Some may ask what the difference is. In the same year as Keynes, Frank Knight in his 1921 book Risk, Uncertainty and Profit distinguished between risk, which was quantifiable and could be insured (or hedged) vs. uncertainty which cannot. Risk is the possibility that your house burns down. Insurance companies can estimate that probability (because there are so many homeowners) and sell you insurance to hedge against that possibility while still making a profit. Uncertainty is something like the likelihood of nuclear war. Since there’s never been a nuclear war (though bombs were dropped in World War 2), the probability of nuclear war cannot be calculated based on any objective information. It is an uncertain event. In investment finance risk has been translated as volatility, or standard deviation, of returns which is easily measured. Uncertainty, on the other hand are possible events without probabilities such as the start or duration of a bear or bull market.
In the case of the market uncertainty we’re discussing here, it is a change of state from a resilient, stable environment where standard capital market theory assumptions hold, to a fragile state where they do not. Uncertainty is more than volatility though volatility is a symptom of uncertainty.
The Two States
I cover this more in "Stable vs. Unstable Markets" and "How Tail Risk Changes over the Market Cycle" , but basically the markets have two states: a low uncertainty, resilient state and a high uncertainty, fragile state. But how are the two states different? Why do they exist?
In the resilient state most standard capital market theory assumptions hold:
Low and stable volatility,
The normal distribution well describes market risk (extreme moves are rare),
Stable correlations across and within asset classes,
Liquidity is high, and
Risk is rewarded with return
In the Fractal Market Hypothesis, this corresponds to periods where investment horizons are diversified. The stable characteristics exist because investors, having many different investment horizons, interpret information differently. A panic at the short horizon is a buying opportunity for those with a long horizon. The diversity of investment horizons assures there is ample liquidity for both buyers and sellers because there is always someone willing to take the other side of the trade.
In the fragile state things become unstable and most capital market assumptions lose validity:
Volatility is high and unstable,
Fat-tails develop in the return distribution as extreme events become common.
The Stable Paretian distribution rather than the normal distribution defines market returns, and variance is undefined.
Correlations become unstable,
Within asset class correlations increase.
In most cases, across allocation correlations decrease.
Liquidity is low, though trading volume can be high, and
Risk and return appear unrelated.
The fragile state corresponds to those periods in the Fractal Market Hypothesis where the long term becomes uncertain, so long term investors shorten their investment horizon. With a more uniform market investment horizon investors interpret information in much the same way. So good news can cause a buying stampede, while bad news produces an avalanche of sales. Accordingly, trading is mostly on one side of the market. So liquidity dries up causing high volatility. But the most significant practical difference is the last bullet. Taking more risk does not necessarily mean you will receive more return. Unlike traditional capital market theory which assumes investors are “rational,” in the fragile state risk and return empirically are unrelated.
Resilient and Fragile States: 1988 - 2018
Source: MSCI, St Louis Fed, FMCR Analytics
Risk and return over the two states for various asset classes.
“Stable vs. Unstable Markets,” “How Tail Risk Changes,” and "Risk Cascades" (in the Newsletter section) provide empirical evidence of these claims. The existence of two states also explains the weak evidence supporting the Capital Asset Pricing Model (CAPM) in numerous research papers going back to the 1960s. CAPM assumes that risk is rewarded with return and many measure risk by a stock’s beta, or sensitivity to the market as a whole. Most studies show that realized excess return is lower than the return predicted by beta. But these studies use a long time period and inadvertently average the two states together assuming that the average is the most likely outcome. Unfortunately the average is only what investors pass through transitioning from the resilient to the fragile state. They don’t spend much time there. Using the average is like the old joke, “If my head’s in the oven and my feet are in the freezer, on average I feel fine.”
Given the differing characteristics between the two states, it would be useful to know when we are transitioning from one to the other. But that alone is not enough. The nature of the low and high uncertainty states are determined by the longer cycles of financial leverage and inflation levels, particularly when we are in the high uncertainty state.
The Market Uncertainty State Indicator (MUSI)
Unfortunately there is no way of knowing exactly when the shift in uncertainty happens even after the fact. We can only estimate it. The MUSI is based upon the idea that uncertainty changes for fundamental reasons. Those risks include credit, real rates, real growth, trading liquidity, inflation, extreme valuation, and financial leverage. If one or two of these risks become high, but the others are OK, then the markets can absorb shocks, like a spike in oil prices for instance. Markets are resilient to shocks. However if a large number of these risks become high, the market environment reaches a critical mass and when one goes off they all go off causing a risk cascade. That’s why and how markets become fragile. The MUSI measures these risks and as they increase the market state becomes more uncertain, and more fragile. Some details are in the “Risk Cascades” paper though there have been some changes. If you read the Constructing Indicators post you’ll note the similarity between the MUSI and how ancient people anticipated seasonal change to the weather.
The MUSI has five states. The first three relate to market uncertainty alone:
A resilient state when fundamental risks are low,
A transition state when there is a mixture, and
A turbulent state when most risks are high.
But there are two more states that link to the longer cycles of inflation levels and the financial instability caused by excessive leverage:
A fragile state when market risks are high, and financial leverage is high but inflation levels are low, and
A critical state where market risk, financial leverage and inflation levels are high.
Inflation and financial instability risk play the same role that the “z” axis causes for the Lorenz Attractor or El Nino does for climatology. They shift the market cycle to different regimes.
Financial leverage risk comes from the Financial Instability Regime Indicator (FIRI). In the turbulent state risk can be high, but returns can also be high. In the fragile state, high financial leverage combined with high market uncertainty causes a “Minsky moment” where deleveraging can cause a financial crisis. Inflation level risk comes from the long-term Inflation Level Regime Indicator (ILRI). While high readings are rare, given the long period of the inflation regime, when all three indicators are at high risk then not only is a “Minsky Moment” possible, but the nature of bonds as a haven asset changes as well. The only place to hide is cash, and sometimes gold.
Each risk indicator is composed of a number of sub-indicators. In the newsletter section I will go into more detail.
The “Risk Cascades” paper gives empirical evidence of the efficacy of this approach. I do have one caveat. The MUSI measures when uncertainty rises for fundamental reasons. There can, however, be exogenous, non-fundamental shocks. Those include geo-political events (such as war or a revolution) and natural disasters (such as earthquakes, tsunamis, and pandemics). The MUSI can’t predict such events. No one can. They are rare but should be kept in mind.
The market uncertainty cycle averages four years. In “Chaos and Order” and “Fractal Market Analysis” I use a fractal methodology called rescaled range, or R/S analysis that among other things is able to determine a non-periodic cycle length. A periodic cycle like, a sine wave, is easily confirmed by standard methods such as Fourier analysis. However, the market is a non-linear, chaotic system so Fourier analysis will produce inconclusive results. In “Fractal Market Analysis” I show that R/S analysis can find the average length of a non-periodic cycle even with a high level of noise. I also find that the US stock market has an average non-periodic cycle of four years. Whether you sample data at daily, weekly or monthly intervals the answer is always four years. For other developed markets I find a similar four to five year cycle. For the US economy, the average cycle length is a slightly longer five years. For many of the indicators in the MUSI this four to five year cycle is an important part of their construction. It also means that the market uncertainty cycle is of medium-term length compared to the inflation and financial instability regimes which are much longer. So the market uncertainty cycle moves within the inflation and financial instability regimes, but they also affect one another.
Culmination
The MUSI is represents the culmination of market climatology since it incorporates market, inflation and financial uncertainty at the time scale that is most useful to us. In addition markets are where most of us have a direct interest. Unfortunately, no one and no indicator can predict the future with certainty. But by looking at all four indicators we can have an idea of where we are and where we’re going. That’s what we need to make decisions.
