Welcome to the hotseat. We've prepared a guide if you'd like to read more about how it works.

A little help with the Px (t) function

+2 votes
asked Mar 23, 2016 in BUS 3018F - Models by asilmotala (2,380 points)

So I am a little confused about the function for the number of censored lives aged x at time t and how exposure to risk is estimated using the census data and the trapezoidal rule. Since we assume the function continuous with time, does that mean it increases and decreases according to deaths, withdrawals and new entries? If so, is it possible to get a probability qx greater than 1, if we experience a large amount of deaths and use the trapezoidal rule? (Under the binomial model)

commented Mar 31, 2016 by simon_rigby (4,220 points)

I think this is theoretically possible, say if \(P_x(t) = 100\), \(P_x(t+1) = 1000\) and \(d_x = 600\). But this would be very peculiar and, if it happened, you may have bigger problems!

commented Apr 12, 2016 by Andrew_Soane (480 points)
Also, when we are using the census method for estimating the exposed to risk, we normally assume that migration is ignored (ie. only deaths would decrease or population). If we were to not ignore migration then we would need to change our formula for the exposed to risk.
commented Apr 13, 2016 by asilmotala (2,380 points)

Thanks! That makes a lot of sense :)

Please log in or register to answer this question.