Analytics Vidhya
Published in

Analytics Vidhya

Finance Theory — Market Micro Structure — Non-synchronous Trading

In this article, I will try to explain another big cost of trading caused by non-synchronous trading. You can see my previous article about market microstructure and bid ask spread on medium here.

What is non-synchronous trading?

Well, let’s try to understand why synchronicity matters. When you are sitting in a car and staring at other running cars wheels, I believe that you have seen for a short period of time a car is moving forward but its wheels seems rotating backwards. Or in the old movies you have seen the film actually going backward. Or you must have seen a gif with a bird flying but its wings barely flapping. Or you must have seen a gif with a chopper in it, but the propeller seems spinning very slowly or not spinning at all.

The reason why we don’t see the propeller moving is that the frequency of its rotation matches the frequency of this gif, i.e. synchronous.

When looking at the stocks, some stocks will be traded by people until 4 pm everyday but some stocks won’t be traded or very thinly traded after 11:00 am. However, trading will resume the next morning. This kind of non-synchronicity among different securities’ trading activities is referred to as non-synchronous trading.

Why does it matter?

One problem is that it creates inflated standard deviations. Suppose that the last price of stock A today is recorded at 3:00 pm and the last price the next day is recorded at 4:00 pm, then from 3:00 pm today to 4:00 pm is 25 hours of trading. There will be, on average, more price actions for these kind of stocks. When we estimate volatilities of stock returns, we will over estimate the volatilities.

Here is a screen shot of a 1999 academic paper by Greg Kadlec and Doug Patterson. Both are from Virginia Tech.

As you can see, most large stocks last trading time will be after 3:45 pm, but 20% of the small stocks do not trade and only about 40% of them trades until after 3:45 pm. You may wanna argue that this data is too old — pre decimalization, but this kind of phenomena is still important after year 2000.

Another problem is that it will create spurious correlation that is not tradable. If stock A stop trading at 3:00 pm, and if stock B that is highly positively correlated with stock A increases its price from 3:00 pm to 4:00 pm, then stock A’s price should be very likely to increase tomorrow because that information is not incorporated in price. The problem is that it can be observed in the data, but it is not tradable, because nobody is willing to trade stock A after 3:00 pm. So even if there is observed correlation, it is possible that part of this correlation is not tradable.

What to do?

You will have to live with it. If you want to avoid it in the algorithm, you should focus on large company stocks or liquid ETFs. If the stocks you want to trade has some non-synchronicity issue, then in your back test, try to use the 12:00 price or 10:00 price. Quite some cross-sectional trading rules won’t work if you take the mid day price in the algorithm.


I hope this article is helpful. Market microstructure is especially important to small stocks. The data we play with usually don’t have the microstructure information needed. This is one of the reasons why institutional investors do not trade anything that is smaller than medium firms in the market (unless it is an index fund).



Analytics Vidhya is a community of Analytics and Data Science professionals. We are building the next-gen data science ecosystem

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store