Using Illiquidity Measures to Analyze Thinly Traded Stocks

# Time series analysis on illiquid price data?


# Time series analysis on illiquid price data?












Say for example I have the following company in some specialized industry:

A - Company that is about to be listed in Exchange 1, i.e., no price history

B - Company that produce similar products as Company A, listed on Exchange 1 as well, however B has very thin volume and price could stay the same for weeks.

C - Similar to company A but listed on Exchange 2, again, thin trading volume

D - Similar to company B but listed on Exchange 2, also thin trading volume

For companies B, C and D, I have their historical EOD price for the past two years.

Exchange 1 and 2 are listed in different continents and there is very little correlation between the two, also, there is no index for this industrial sector. (However the price between company C&D and A&B should be correlated). Also, we can not assume the price time series is non-stationary as the products those company produce could be seasonal in nature.

I would like to figure out the "correct" market price for Company A before it is listed, based on the information above. And my results so far shows that each of the price time series that I have has a different ARIMA model.

Therefore, my question is how can I tackle these price data to start my analysis? Bearing in mind that those are all the data I have and can get.

## Answer by Quantopik (score 3, accepted)

https://quant.stackexchange.com/a/17557

Although I sincerely do not know the correct answer to your question because I never read about this kind of topic in particular and I agree for the most with @MattWolf, I found your question very interesting.

I suggest you to start by reading the literature on the market illiquidity proxies and the relative effect of market illiquidity on the stock market returns; some of my favourite authors about this topic are Amihud & Mendelson, that developed this research field with their seminal paper in 1986:

> Amihud, Yakov, and Haim Mendelson. "Asset pricing and the bid-ask spread." Journal of financial Economics 17.2 (1986): 223-249.

Browsing on google scholar who cited this paper, you can find other interesting papers about the topic.

As regards your question, particularly, you could examine how the illiquidity of a security B influences the illiquidity of security C, or vice versa (since the relative markets are on different continents, they pretty surely trade on different hours), since, if they are comparable, they should move in the same way and be correlated.

Said that, since it is proven that empirically this relationship holds:

$R_{t+1}$ = $\alpha$ + $\beta$*$ILLIQ_t$ + $\epsilon$

where:

- $R_{t+1}$: stock return at time $t+1$;

- $ILLIQ_t$: market illiquidity proxy;

- $\alpha$ and $\beta$: linear regression coefficients;

particularly when you measure the portfolio returns on the portfolio illiquidity.

You could be able to find a trading opportunity.

Alternatively, you could test a pair trading strategy based on security illiquidity; look, for instance, if there exist a granger causality between these two stocks.