Tagged: data science

Time and Sentiment

Time matters.

We’ve gone 42 trading days with the ten-point Market Structure Sentiment index, our proprietary measure of the propensity of algorithms to lift or lower prices, over 5.0. That’s a growth-at-a-reasonable-price (GARP) market since Oct 17.

It’s by no means the longest. More on that in a moment.

The market seems impervious to fault lines as we move into year-ending options-expirations tomorrow through Friday, and index rebalances, and portfolio window-dressing.

It’s nothing like either 2010, the Year of the Flash Crash, or 2012, the Year of the Glitch. I wrote an editorial for IR Magazine on those, a retrospective ahead of 2020 on the last decade of market structure.

The market’s capacity to relentlessly rise through a corporate earnings recession (we’ve had three quarters of falling comparative profits), trade disputes, Presidential impeachment, on it goes, shows how both the IR profession and investors need different data in the arsenal to understand how stocks are valued.

Two other terrific IR Magazine pieces highlight the value of data to IR. We’re not alone! Reporter Tim Human describes how AI Alpha Labs uses deep-learning to help investors understand how to achieve better returns – something IR must know.

Oliver Schutzmann of Iridium Advisors says future IR will be data science – because that’s how the market works.

Speaking of data, ours for the S&P 500 show Active money the past year was the lead buying behavior 7% of the time. The worst day was Dec 19, 2018, with just one company earning what we call a new “Rational Price” from Active Investment. The best day came a month later on Jan 18, 2019, when 26% of the index had new Rational Prices.

Add selling, and Active money leads the S&P 500 behaviorally about 14% of the time.

Back to Sentiment, over the 42 days where the market has gone up, up, up on positive Sentiment, Active money led buying just 6% of the time. Exchange Traded Funds have led, our data indicate, 67% of the time, directly or indirectly.

Data points like these are requisites of the IR job the next decade. And measuring changing behavioral trends will be essential to understanding stock prices. Case in point, I saw a client’s data yesterday where Fast Traders, automated machines creating price-changes, were responsible for a 40% 2019 decline in equity value.

Machines don’t know what you do. But they can trigger negative and positive chains of events divergent from fundamentals.

In this case, trading machines blistered shares on consecutive earnings reports, which in turn pushed Passive money to leave because volatility created tracking errors, which caused market cap to fall out of the Russell 1000 – which is 95% of market value today.

Active money was incapable of overcoming the overwhelming force of Passive behavior.

Back to Sentiment and Time, before the market tipped over in the fourth quarter of 2018 Sentiment stayed around 5.0 or higher for 66 days. When it finally dipped below 4.0, a Sentiment bottom, the market was cast into tumult.

Part of the trouble was delayed portfolio rebalances. When markets go up, investors put off aligning positions to models. When the turn comes, there’s a scramble that compounds the consequence.

We’re hitting one of those crucial points this week, delayed rebalances colliding with options-expirations, index-rebalances and year-end.

We may see nothing. After all, it’s only been 42 days. Sentiment went 45 days at 5.0 or better this spring on the great January rally. There were 42 positive days from June 11 to Aug 8, too.

My final thought in the final fresh blog for 2019 (we’ll do retrospectives to finish out the year) is a monetary one. Karen reminds me that monetary policy clears a room like a fart – so you can’t talk about it often.

But it’s another data point driving market behavior. As the Federal Reserve has turned accommodative again – that is, it’s shifted from shrinking its balance sheet and raising rates to expanding assets and lowering rates – we’ve seen a corresponding fall in shorting and derivatives-leverage in stock-trading.

In fact, a steep drop for shorting coincided with a sudden spike in what’s called the Fed Funds overnight rate. Remember that? Happened in September. The rate the Fed had set near 1.5% exploded to 10% as the market ran out of cash.

Ever since it’s been troubled, the Fed Funds market. The Fed keeps injecting tens of billions of dollars into it. That market is meant to provide banks with temporary liquidity to process payrolls, taxes, credit-card payments, transaction-settlements and so on.

What if ETFs are using cash to collateralize transactions (rather than actual stocks) at the same time rising consumer debt is beginning to weigh on bank receipts and liquidity?

If that’s at all the case, time matters. We’re not worried about it – just watchful. But with so vast a part of market volume tracing to ETFs, and Sentiment getting long in the tooth, we’re cautiously wary as this fantastic trading year ends.