The market is up as consumer sentiment comes in higher than expected!

The market is down over war worries!

New low in unemployment boosts market outlook as it closes higher!


We've all seen headlines like these.

They follow a predictable pattern: market is up/down as X occurs.


X could be an assassination, a meeting of the Federal Reserve board, a new low in bond yields, updated inflation numbers, a bill gets killed or passed, an election, and on and on it goes.

Whatever X is, the authors subtly link the event with the day's closing price, but they're very clever!

They never explicitly say X caused the price to rise or fall, simply that the market was up or down at the same time that X occurred - if you infer causation (which seems to be the intent), then that's on you (to be honest, a headline saying "STOCKS UP!" would be far less interesting).

The fact is, we like a good story.

We tell ourselves stories all the time about why our investments went up or down. It's a way to justify our (usually poor) financial choices to ourselves (and comes in handy when explaining it to our significant others, "Sorry honey, I would have gotten you that gift you wanted, but who could have known the housing market would crash?").

This isn't to say there are no economic laws or these events don't cause other effects in the economy and markets in the long run, rather we quickly turn to these events to explain market gyrations because it feels good, even though they have no discernible effect when they happen.

What? Trading on news is meaningless noise?

Yes. And we can back that up.

Endogenous and Exogenous Factors in the Market

What moves the market?

Well, the market moves the market.

Famed trader (and either arch-villain or kind-hearted philanthropist depending on your political persuasion) George Soros, made this point when describing his theory of reflexivity. The market isn't just an efficient computation of fundamental value - the prices influence the fundamentals.

This creates a self-referential feedback loop in market prices leading to endogenous shocks (endogenous means it comes from within, exogenous from the outside - even with all the letters after my name I still think back to high-school chemistry and endothermic vs exothermic reactions to keep these straight).

Traders who base their moves on the news are betting that exogenous information is what leads to price changes. These are the true believers in the Efficient Market Hypothesis.

We have a hypothesis - the market moves primarily based on internal dynamics, not external events which change the fundamental value of the companies.

So let's test this.

News vs. Market Structure for Market Moves

Jean-Philippe Bouchard tested this particular hypothesis by mapping news coming across his Bloomberg terminal to high-frequency price movements in the market. If news came through and moved a ticker, then we have evidence that the market is news-driven. If the market moved in the absence of news, then we don't get supporting evidence.

As Bouchard discovered, while news does move the market, it's much less frequent or important than the market structure between buyers and sellers themselves.

Take a look at the plot below.


Bouchard plotted the cumulative probability of a price jump against the magnitude for the market structure (endogenous affects) and news (exogenous affects). They're on a log-log scale and both fit a nice power law.

What we can see is that the market moves a lot without any news being associated with it. This makes up most of the small moves in the market, shown in the top left of the plot.

As you move to the right of the plot, you get larger and larger moves. Notice too that the news line is below the endogenous jumps line. Probability is on the y-axis, so this shows that even in these cases where there's a large price jump, it's still more likely to happen without any corresponding piece of news that can be linked to it!

Just to be clear, there's evidence that news flow has some effect on market movements, but what this result shows is that it has less of an effect than the reflexive action of the market itself.

This result stands at odds with the Efficient Market Hypothesis, which states that all public knowledge is already priced into the market. If that were true, then only these exogenous news events would be moving the market.

Markets aren't Looking so "Efficient" After All

Bouchard isn't alone in his conclusion. In a similar study looking at one and five-minute bars from 1982-1999, Yale professor R.C. Fair concluded, "many large price changes correspond to no obvious events."

The Santa Fe Institute - an academic organization that focuses on complexity science - published results from an agent-based stock market simulation that widely agreed with Bouchard's data. They found these feedback loops and endogenous effects dominated their model. This is significant because this agent-based approach goes against many prevailing assumptions of economic and financial orthodoxy regarding efficient markets and rational actors.

It looks like the markets aren't quite what we were led to believe them to be...

Ok, great. So how do we use this information when trading?

Simplicity and Complexity

It has long been known that simple rules can lead to complex behavior. For example, play with Conway's addictive Game of Life and see how a single, simple rule, can evolve over a complex tapestry and produce surprising results.

These simple rules yield complex adaptive systems that are next to impossible to predict.

As a trader, rather than meeting the complexity of the market with an equally complex model, why not try something simple instead?

Research in many areas has shown that simple rules often out perform complex models. It does so in forecasting competitions, asset allocation models, and developing business strategies in highly competitive markets. Simple rules can perform well in trading algorithms too.

There are many reasons these simple models perform so well in different domains. One of the most compelling reasons comes from the sensitivity of parameter estimates for more complicated methods.

Complicated models tend to have a variety of parameters that they take as inputs. Maybe they have estimations of volatility, liquidity, correlations, and so forth. Each of these estimates has some error around them, and they aren't stable over time (non-stationarity is one of the most frustrating facts of financial markets), which makes them very difficult to pin down. These small errors can compound and lead a sophisticated model wildly off-course!

On the other hand, simple systems, like trend following, for example, don't rely on these estimates. They rely on simple, mechanical rules that are designed to capture market moves while taking on small risks. They're "hunting for outliers" wherever (and why-ever) they may occur.

Stocks Finish Higher Mind

Next time you see a headline claiming the S&P 500 or the Dow Jones was up or down "as" some event happened, hopefully you'll see through the subtle implication that we can easily find cause and effect relations between the news and stock prices. For better or worse, the world is much more complicated than that.

Although prediction is hard - especially when it's about the future - we can produce profitable trading strategies from a simple set of rules. And this is exactly what we built Raposa to do.

We want to provide traders and investors the ability to design and test trading algorithms to find a profitable trading system that they can easily deploy, all with no coding experience needed.

We're building the future of retail trading. Check out our free demo today!