Gaps frequency and seasonlity

[matina14]

Hello,

I'm using data from 01Jan1990 until 31Dec2011 for trading volume in stock market.

For example, in data editor it goes like this:

time volume
24dec1990: 26358
25dec1990: . --> holidays
26dec1990: . --> holidays
27dec1990: 51190
28dec1990: 79502 --> Friday
31dec1990: 60582 --> Sunday

Basically, there are two problems:
- when it's holidays (like 25th and 26th of December) the trading volume has a missing value
- when it's weekend, the time and trading volume are just omitted

In stock market, this means that there was no trading in these days (holidays and weekends).

I would like to see the frequency of these no trading gaps in a table. For example, how many times there was one day of no trading, two days of no trading etc
What command shall I use?

Also, how can I include a dummy variable to account for this seasonality problem?

Thanks.

[bukharin]

Here's one very basic solution to the first question. It might need to be generalised if you have eg panel data:

. list, clean

            date   volume  
  1.   24dec1990    26358  
  2.   25dec1990        .  
  3.   26dec1990        .  
  4.   27dec1990    51190  
  5.   28dec1990    79502  
  6.   31dec1990    60582  

. drop if missing(volume)
(2 observations deleted)

. sort date

. gen gap=date - date[_n-1] - 1
(1 missing value generated)

. list, clean

            date   volume   gap  
  1.   24dec1990    26358     .  
  2.   27dec1990    51190     2  
  3.   28dec1990    79502     0  
  4.   31dec1990    60582     2  

. tab gap

        gap |      Freq.     Percent        Cum.
------------+-----------------------------------
          0 |          1       33.33       33.33
          2 |          2       66.67      100.00
------------+-----------------------------------
      Total |          3      100.00

I am not sure why you call this a "seasonality problem"; using a dummy variable doesn't make much sense to me but it might be that I don't understand what you're trying to do. If you're doing a time series analysis you may wish to exclude the non-trading days, create your own time scale, and use that for your -tsset- command (see eg here).

[matina14]

Thanks bukharin. Your suggestion works fine.

I'm examining the relationship between stock returns and trading volume. As suggested from the available literature they adjust their data for seasonality; they include dummy variables for:
- day of the week
- gap of nontrading days (which captures the effect of holidays and weekends)
- for each month
- for each year
Basically that's why I consider it as a seasonality problem