If you read the first source [0] in the article, you will find the following refutation in the first page:
<<Consider the possibility that quarterly company earn-
ings announcements in the United States (outside reg-
ular trading hours) are responsible for the top row of
Figure 1 [15]. You can falsify this attempted explanation
by removing the returns around earnings announcements
and noting that this does not meaningfully change the
plots (as in Section 4.1 of Ref. [5]).>>
Not really, because in an efficient market one would expect prices to reflect expected dividend payouts. Also, it would not explain why the data from the Chinese market follows the opposite trend.
I'm not saying there's a big conspiracy here. There's most likely some simple explanation, perhaps involving the different algorithm used in exchanges to match buy/sell orders during opening and closing, compared to the rest of the day. I wish I had some more time to dig deeper into this, because at face value it's surprising.
If stock A is going to pay a dividend of $1 per share, you are suggesting that the price of the stock is going to be $1 higher than it would have been if there was no dividend? Ok, that's fine. But, then you actually get the $1 paid to you.
If they are including dividends in the returns, then they are talking about total returns, and the formula for a total return is (price + dividend) / (previous price).
If you ever look closely at the claims that the market goes up by an average of 7-10% yearly, you'll see that it always includes dividend reinvestment. Returns are way lower if you don't include the dividends. If you are cash them out as they come, they call that the income return on your investment.
EDIT - I'll also point out that the US returns, according to that source study, are coming from CRSP. CRSP publishes US data according to their own methodology. Chinese data is not coming from them, and there is no reason to believe that it is using the same methodology. Apples v Apples, you know.
6 comments
[ 3.9 ms ] story [ 30.3 ms ] threadNews that happens when the markets are open is generally unexpected, which is most frequently bad.
<<Consider the possibility that quarterly company earn- ings announcements in the United States (outside reg- ular trading hours) are responsible for the top row of Figure 1 [15]. You can falsify this attempted explanation by removing the returns around earnings announcements and noting that this does not meaningfully change the plots (as in Section 4.1 of Ref. [5]).>>
[0] https://arxiv.org/abs/2010.01727
> The Night return includes split-adjusted dividends as reported on CRSP.
So, uh, that's a pretty big reason for night returns being higher.
[1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1004081
I'm not saying there's a big conspiracy here. There's most likely some simple explanation, perhaps involving the different algorithm used in exchanges to match buy/sell orders during opening and closing, compared to the rest of the day. I wish I had some more time to dig deeper into this, because at face value it's surprising.
If stock A is going to pay a dividend of $1 per share, you are suggesting that the price of the stock is going to be $1 higher than it would have been if there was no dividend? Ok, that's fine. But, then you actually get the $1 paid to you.
If they are including dividends in the returns, then they are talking about total returns, and the formula for a total return is (price + dividend) / (previous price).
If you ever look closely at the claims that the market goes up by an average of 7-10% yearly, you'll see that it always includes dividend reinvestment. Returns are way lower if you don't include the dividends. If you are cash them out as they come, they call that the income return on your investment.
EDIT - I'll also point out that the US returns, according to that source study, are coming from CRSP. CRSP publishes US data according to their own methodology. Chinese data is not coming from them, and there is no reason to believe that it is using the same methodology. Apples v Apples, you know.