Home  /  Resources & support  /  FAQs  /  Time-series operators

Where can I find a description of the various time-series operators?

Title   Time-series operators
Author Allen McDowell, StataCorp


I’m puzzled by the command xtabond and how to interpret the following line:

        xtabond n l(0/1).w l(0/2).(k ys) yr1980-yr1984, lags(2)

The manual is unclear about this. I also don’t understand the output that is displayed. It contains symbols such as LD, L2D, and D1.

Since I suspect that these are time-series operators, I checked the manual for such information but was unable to find anything. Where can I find an explanation of these time-series operators?


To interpret the command, you need only understand that time-series operators accept both numlists and varlists (see [U] 11.4.4 Time-series varlists, [U] 11.4 varlists, and [U] 11.1.8 numlist).

For example, the term


means lags 0 through 1 of the variable w. That is, it means the variable w and the variable w lagged 1 period (i.e., w and l.w). The term

        l(0/2).(k ys)

means lags 0 through 2 of the variable k and the variable ys, or k, l.k, l2.k, ys, l.ys, and l2.ys.

In general, l.var or l1.var means the first lag of var; l2.var means the second lag of var; and in general, l#.var means the #th lag of var. There are also difference operators (using d.var), leads (f.var), and seasonal differences (s.var).

To summarize, the command above could be expanded to read

        xtabond n w l.w k l.k l2.k ys l.ys l2.ys  yr1980-yr1984, lags(2)

yr1980-yr1984 is Stata’s common shorthand for a varlist and is equivalent to yr1980 yr1981 yr1982 yr1983 yr1984. Finally, lags(2) means to include the first and second lag of the dependent variable in the model.