Why does teffects psmatch number of observations not match number of treated + number of controls with 1 match per treated?

[Nicholas Poggioli]

My data have 200 treated units and 9,800 untreated units. When I run teffects psmatch using default options, the results show

• Number of observations = 10,000
• Matches: requested = 1
• min = 1
• max = 1

If I have 200 treated units, and each unit is matched with one control unit, shouldn't the number of observations be 400?

I am confused because I cannot tell how many observations are being used in the estimation of the average treatment effect.