To make the DVIS work, you need the ‘congressional mandated algorithm’ - here it is.
“The bill divides the world into six regions and designates each a “high-admission” or “low-admission” region. High-admission regions are those that account for more than 1/6 of all family- and employer-sponsored visas and vice-versa for low-admission regions. The population of each region is calculated (excluding any “high-admission” country which has sent more than 50,000 migrants to the U.S. over the past five years), and the populations of all high- and low admission regions are added together separately. The population of each low-admission region is divided by the overall population of the low-admission regions, and the same is done for each high-admission region.
The DV Lottery allocates visas using the following formula:
For a low-admission region
(Regional population percentage) X (percentage of family- and employer sponsored visas that all high-admission regions account for)
For a high-admission region
(Regional population percentage) X (percentage of family- and employer sponsored visas that all low-admission regions account for)
As is evident from the formula, Schumer [US Senator from NY who set the formula] designed the DV Lottery formula to reflect changes in the levels and sources of immigration to the U.S. on an annual basis. The crux of the formula is that it allocates a percentage of visas to low-admission regions that is equal to the overall percentage of visas that high-admission regions account for in all other U.S. immigration programs. In order to ensure that the program does not benefit any single country disproportionately, no country can account for more than 7% of all Diversity Visas.”
Using DV2014 as the case study, shows how this works in practice - neglecting NA
The ’High’ regions populations:
AS 991m (86%)
SA 160m (14%)
The ’Low’:
AF 1100m (58%)
EU 751m (40 %)
OC 37m (1.97%)
These figures are estimates and would not match the exact ones used in DVIS – almost certainly the Bureau of Census annual figures. These change annually. [The accurate figures and annual changes are on the CIA web site – if you are into detailed analysis.]
All that is missing to complete the calculation is the single parameter that gives the percentage of sponsored visas from low admission regions. It does not seem to be publically available but it is trivial to calculate from the historic data. The definition allows it to move smoothly but it has been dead constant over the last four years – 0.81. That includes the 2012 episode. Indeed, if you exclude 2012, the factor stays exactly the same; but the standard deviation dramatically improves from 6% to 2%. So it is a pretty solid parameter. [You have to leave OC out of the fit; it is too small to impact.]
Given all that, the formula would, for DV2014, allocate visas to Regions as follows (first figure is total of 50k; second of 55k)
AF 23,644/26,008
AS 8,095/8,905
EU 16,156/17,771
SA 1,305/1,435
OC 801/881
Of course, these figures change annually. The first annual impact is from new additions/exclusions of countries. The second is from the differing growth rates among countries. (The change in ratios between AS/SA in 2014 is an example; Guatemala was added plus the population growths are different.) I have yet to find any secondary effects such as Country Caps impacting between regions –as opposed to within regions. Finally, the Bill allows swapping of quotas between regions if necessary; there is evidence of this happening – especially in the last quarter as Consulates’ capacities come under stress.
Finally, a mea culpa. Finding signals in lots of noisy data ain’t easy. Nat Silver’s book is essential encouragement! I had been badly diverted by the OC and SA selectees coming in exactly the same for 2012 and 2013; I now know that, though tantalisingly close [2,001 v 2,002!], this is just noise.I also now see that selectee/ allocation ratios are based on algorithms using historic take up rates –but at a Regional level. For example, for 2014 the selectee/allocation ratios appear to be
AF 2.4x
AS 2.6x
EU 2.6x
SA 3.2x
OC 5.2x
And on this basis, let me sign off for a month or so to head up the Rockies chasing avalanches.
Au revoir!
