- Speaker #0
Hello, welcome back to Papers with Backtest podcast. Today, we dive into another algo trading research paper.
- Speaker #1
Indeed. And this one looks at something maybe a bit different, overnight stock returns.
- Speaker #0
Exactly. The return from the close one day to the open the next.
- Speaker #1
And the big question is, can this tell us something about how investors specifically feel about a company?
- Speaker #0
Yeah, the core idea, and it builds on some previous work, is that maybe individual investor sentiment, they're... optimism or pessimism shows up in that after hours trading and gets baked into the overnight return.
- Speaker #1
It's an interesting angle trying to capture sentiment at the firm level, right? Not just broad market mood.
- Speaker #0
Right. So this paper, it's called Overnight Returns and Firm-Specific Investor Sentiment. And they're basically testing if this overnight return measure, well, if it actually acts like a sentiment indicator.
- Speaker #1
Makes sense. They needed to see if its characteristics line up with what you'd expect from sentiment.
- Speaker #0
So they use CRSP data. Looked at the periods from July 1992 through December 2013.
- Speaker #1
And 92 is key because that's when reliable open price data became available.
- Speaker #0
Got it. And for our deep dive today, we're really going to focus on the potential trading rules and the backtest results they found. OK,
- Speaker #1
so the paper breaks down into a few main areas. First, does this overnight return persist? You know, does a high return follow a high return in the short term? Then how does that link up with the type of company? Like, is it harder to value who owns it? Lots of institutions.
- Speaker #0
Okay, persistence and company characteristics.
- Speaker #1
And finally, what happens longer term? If a stock has really high overnight returns for a bit, does that reverse later on?
- Speaker #0
Contrarian possibilities there. Okay, let's start with that short-term persistence then. Right. How did they test it?
- Speaker #1
They used a weekly sorting method. So every single week, they'd rank all the stocks based on their total overnight return for that week.
- Speaker #0
Lowest to highest.
- Speaker #1
Exactly. And... divided them into 10 groups, deciles.
- Speaker #0
And what popped out?
- Speaker #1
Well, something quite noticeable for anyone looking at short-term moves, stops that were in the top 10% for overnight returns one week.
- Speaker #0
The highest performers.
- Speaker #1
Yeah. They tended to have a significantly better average overnight return the next week compared to the stocks in the bottom 10%.
- Speaker #0
How much better?
- Speaker #1
The difference was about 1.76 percentage points on average for that following week. Wow.
- Speaker #0
OK, 1.76% in a week is definitely not trivial.
- Speaker #1
No,
- Speaker #0
it's not. And you said it wasn't just a one-week phenomenon. It kind of lingered.
- Speaker #1
It did seem to. The effect diminished week by week, but it was still statistically significant for up to four weeks later.
- Speaker #0
So W plus 1, W plus 2, up to W plus 4.
- Speaker #1
That's right. So a big overnight jump seemed to signal a higher probability of more positive overnight returns, maybe smaller ones. for the next month or so.
- Speaker #0
Interesting. And did the average return generally go up across those 10 groups in the follow-up weeks?
- Speaker #1
Yes. They saw that generally the subsequent week's average overnight return increased as you move from the lowest initial decile to the highest.
- Speaker #0
Now, hold on. Could this just be like market mechanics, bid-ask-bounce or something?
- Speaker #1
That's a fair question. The researchers thought about that too. They re-ran the numbers using, quote, data, specifically the midpoint between the bid and ask prices.
- Speaker #0
To kind of smooth out the spread effect.
- Speaker #1
Exactly. And they found very similar results. So it strongly suggests that the bid-ask spread isn't the main thing driving this short-term continuation.
- Speaker #0
Okay, that's a crucial check. So how does this overnight persistence compare to just looking at regular close-to-close returns over the same weekly period? Ah,
- Speaker #1
good comparison. When they did the same decile sorting based on weekly close-to-close returns, the picture was much fuzzier.
- Speaker #0
No clear pattern.
- Speaker #1
Pretty much. No consistent monotonic increase across the deciles in the next week. And the differences between the top and bottom groups were smaller and, frankly, less reliable.
- Speaker #0
So it really points to something specific happening in that overnight close to open window.
- Speaker #1
It seems that way, yeah. Like that period is particularly sensitive to whatever's causing this persistence.
- Speaker #0
They did robustness checks, right? Made sure it wasn't just size or momentum or something else explaining it away.
- Speaker #1
Oh, definitely. They controlled for standard factors. Market beta, from size, book to market. momentum, the persistence held up. Even when they looked within different groups, say, just within large cap stocks or just within high momentum stocks, they still found that stocks with higher overnight returns tended to have higher overnight returns the next week.
- Speaker #0
Okay. That makes the short-term findings seem pretty solid then. There's some kind of continuation happening there.
- Speaker #1
Seems like it.
- Speaker #0
All right. So let's connect this to those firm characteristics you mentioned. This could refine potential trading rules. First, the hard-to-value stocks. How did they define hard to value?
- Speaker #1
They used several proxies, things that usually indicate more uncertainty or less available information. Like what? Stock return volatility, more volatile, harder to value. Firm size, smaller firms are often trickier. Firm age younger means less history. Profitability. Less profitable can mean more uncertainty. OK. And also the earnings to price ratio. A low EP often implies high growth expectations, which are inherently harder to value accurately.
- Speaker #0
Makes sense. So five different ways to slice difficulty. How did they relate this back to the overnight returns?
- Speaker #1
They did a two step sort. First, each year they'd group stocks into quartiles based on one of those hard to value metrics, say volatility.
- Speaker #0
four groups from least to most volatile.
- Speaker #1
Exactly. Then within each of those four volatility groups, they did the weekly decile sort based on overnight returns, just like before.
- Speaker #0
I see. So sorting within sorts and what emerged was the persistence stronger or weaker for the hard to value ones?
- Speaker #1
Stronger, significantly stronger. And this was consistent across all five of their hard to value proxies.
- Speaker #0
Really? For all
- Speaker #1
Yeah. The difference in the next week's... W plus one overnight return between the top and bottom overnight deciles was always biggest in that quartile representing the most difficult to value stocks.
- Speaker #0
Can you give an example like with volatility or size?
- Speaker #1
Sure. For volatility, the top minus bottom decile difference in next week's overnight return was one point nine nine percentage points for the most volatile quartile compared to only one point zero four percentage points for the least volatile.
- Speaker #0
It's almost double.
- Speaker #1
It is. And for size, it was even more stark. 2.32 percentage points for the smallest, hardest to value quartile versus just 0.72 percentage points for the largest firms.
- Speaker #0
Wow. Okay. So the implication here for a trader might be focus short-term sentiment strategies on these harder to pin down stocks.
- Speaker #1
That's certainly what the results suggest. Sentiment seems to pack a bigger punch in the short run when fundamental value is more ambiguous.
- Speaker #0
Okay. Now the other characteristic. Institutional ownership. What was the idea there?
- Speaker #1
The hypothesis was that sentiment, especially the kind potentially driving overnight returns, is more associated with individual investors rather than large institutions.
- Speaker #0
Because institutions are maybe more fundamentals driven, less swayed by short term noise?
- Speaker #1
That's the general thinking. So they predicted the overnight persistence effect would be weaker in stocks with high institutional ownership.
- Speaker #0
And did they test that the same way? Sorting by ownership level first?
- Speaker #1
Yes, similar approach. They sorted stocks into four groups based on the percentage of shares held by institutions from lowest to highest I.O.
- Speaker #0
And then the weekly overnight return deciles within each I.O. group.
- Speaker #1
Precisely.
- Speaker #0
And the result. Did high I.O. dampen the effect?
- Speaker #1
It did. The persistence, that next week return difference between the top and bottom overnight deciles, it systematically decreased as institutional ownership increased.
- Speaker #0
How much of a decrease?
- Speaker #1
For Week W Plus One. The difference was 2.36 percentage points for the lowest IO quartile, but it fell to 1.07 percentage points for the highest IO quartile.
- Speaker #0
Again, quite a significant drop. More than halved. Yeah.
- Speaker #1
It really supports the idea that this overnight phenomenon is more strongly linked to segments of the market where institutions aren't the dominant players.
- Speaker #0
So another potential filter for a short-term strategy. Perhaps favor stocks with lower institutional holdings if you're playing this sentiment persistence.
- Speaker #1
That could be a logical conclusion from these findings, yes. High I.O. stocks seem less susceptible to this particular effect.
- Speaker #0
Okay, that covers the short-term and how firm type matters. What about the flip side, the longer-term picture? You mentioned potential reversals.
- Speaker #1
Right. So does excessive sentiment, as may be proxied by these overnight returns, lead to longer-term corrections?
- Speaker #0
A classic sentiment story. How did they investigate this?
- Speaker #1
They switched gears a bit. Instead of weekly source, they formed portfolios monthly, specifically every December.
- Speaker #0
Why December?
- Speaker #1
Likely just to have a consistent annual rebalancing point. They ranked stocks based on their average daily overnight return over that entire month of December.
- Speaker #0
Okay. Average for the whole month? Then DeSiles again.
- Speaker #1
DeSiles again. Lowest average to highest average.
- Speaker #0
And the strategy?
- Speaker #1
A long, short approach. Go long the bottom decile stocks with the lowest average overnight returns that month, suggesting maybe pessimistic sentiment. And go short the top decile stocks with the highest average returns, maybe overly optimistic sentiment.
- Speaker #0
And hold for how long?
- Speaker #1
They held these long short portfolios for the next 12 months.
- Speaker #0
OK, a longer term contrarian bet against the prior month's extreme overnight movers. What did the back test show?
- Speaker #1
It showed a significant positive abnormal return. After adjusting for the usual risk factors, market, size, value, momentum, they used a five-factor model. The strategy generated an average monthly alpha of 0.62 percentage points.
- Speaker #0
0.62% per month. That's true. Yeah. That's substantial.
- Speaker #1
It adds up. Annually, that's around 7.4% alpha.
- Speaker #0
Wow. So It really suggests that stocks getting pushed up hard overnight, perhaps on sentiment, tend to underperform significantly over the following year and vice versa.
- Speaker #1
Exactly. It provides evidence for a longer term reversal based on this overnight sentiment proxy, a potential contrarian strategy.
- Speaker #0
And did this reversal effect also vary with how hard the stocks were to value? Did sentiment matter more there in the long run, too?
- Speaker #1
Yes. Interestingly, it did. When they looked at the performance of this long short strategy within those different hard to value subgroups. Uh-huh. The positive abnormal returns were generally stronger and more consistently significant for the stocks deemed most difficult to value.
- Speaker #0
So not only is the short-term persistence stronger in those stocks, but the eventual reversal seems more pronounced, too.
- Speaker #1
That's what the data indicated. It reinforces the idea that sentiment effects, both the initial momentum and the subsequent correction, are amplified when fundamentals are less certain.
- Speaker #0
Okay. They also briefly touched on earnings announcements, didn't they? Just as another way to show this measure matters.
- Speaker #1
Yeah, it was more illustrative. They found that the level of pre-announcement overnight returns, basically, the sentiment leading into the announcement affected how the price reacted to the earnings news itself.
- Speaker #0
How so?
- Speaker #1
If sentiment, high overnight returns, was already optimistic before the announcement, the positive price reaction to the actual earnings report tended to be weaker. It's like... Some of the good news or perhaps excessive optimism was already priced in via that overnight sentiment.
- Speaker #0
I see. So the overnight return isn't just predicting future returns. It's also conditioning how the market reacts to new information.
- Speaker #1
Exactly. It demonstrates that this firm specific sentiment measure has tangible impacts on market dynamics. OK,
- Speaker #0
let's try to summarize the key takeaways here, especially for algo traders listening. What should they be thinking about?
- Speaker #1
Well, first, these overnight returns aren't just random noise. They show short term persistence. especially over the next week or so.
- Speaker #0
Potential momentum signals there.
- Speaker #1
Right. Second, this persistence seems stronger in specific types of stocks, those that are harder to value based on various metrics and those with lower institutional ownership.
- Speaker #0
So maybe fertile ground for short-term sentiment strategies in those segments.
- Speaker #1
Potentially, yeah. But then there's the third point, the longer-term reversal. High short-term overnight returns seem to predict longer-term underperformance.
- Speaker #0
Suggesting contrarian opportunities if you have a longer horizon.
- Speaker #1
Precisely. It highlights that tension between short-term sentiment continuation and longer-term mean reversion.
- Speaker #0
And fundamentally, the paper makes a case that the overnight return can serve as a useful quantifiable proxy for firm-specific investor sentiment.
- Speaker #1
It does. It gives you a number to potentially work with rather than just a vague notion of mood.
- Speaker #0
So the final thought for you, the listener, might be, how could you incorporate a measure like this? Thinking about overnight action, maybe combined with factors like valuation difficulty or institutional presence.
- Speaker #1
Yeah. How might it enhance existing strategies or could it form the basis of a new one? Balancing that short term momentum against the longer term reversal is likely key. It's definitely food for thought.
- Speaker #0
Thank you for tuning in to Papers with Backtest podcast. We hope today's episode gave you useful insights. Join us next time as we break down more research. And for more papers and backtests, find us at https.paperswithbacktest.com. Happy trading!
