1: \begin{abstract}

2: We introduce a class of two-parameter discrete dispersion models, obtained

3: by combining convolution with a factorial tilting operation, similar to

4: exponential dispersion models which combine convolution and exponential

5: tilting. The equidispersed Poisson model has a special place in this

6: approach, whereas several overdispersed discrete distributions, such as the

7: Neyman Type A, Pólya-Aeppli, negative binomial and Poisson-inverse Gaussian,

8: turn out to be Poisson-Tweedie factorial dispersion models with power

9: dispersion functions, analogous to ordinary Tweedie exponential dispersion

10: models with power variance functions. Using the factorial cumulant

11: generating function as tool, we introduce a dilation operation as a discrete

12: analogue of scaling, generalizing binomial thinning. The Poisson-Tweedie

13: factorial dispersion models are closed under dilation, which in turn leads

14: to a Poisson-Tweedie asymptotic framework where Poisson-Tweedie models

15: appear as dilation limits. This unifies many discrete convergence results

16: and leads to Poisson and Hermite convergence results, similar to the law of

17: large numbers and the central limit theorem, respectively. The dilation

18: operator also leads to a duality transformation which in some cases

19: transforms overdispersion into underdispersion and vice-versa. Many of the

20: results have multivariate analogues, and in particular we consider a class

21: of multivariate Poisson-Tweedie models, a multivariate notion of over- and

22: underdispersion, and a multivariate zero-inflation index.\medskip\ \newline

23: \emph{Keywords: }factorial cumulant generating function; factorial tilting

24: family; infinite dilatability; multivariate discrete distribution;

25: over-/underdispersion; Poisson-Tweedie mixture \medskip\ \newline

26: \emph{Mathematics Subject Classification:} 60E10; 62E20; 62H05

27: \end{abstract}