This diary is part of a continuing series. For previous installments, covering Iowa and the Dakotas, follow the links.
The basic concept is to imagine how congressional boundaries might look if America followed the Canadian system, where provinces can gain new ridings as a result of redistribution but cannot lose them.
Mid-western rust belt states have seen slow population growth in general, but Michigan has been particularly badly hit. The decline of the auto industry and the emptying out of Detroit mean that the state actually lost population between 2000 and 2010 and thus its delegation shrank to 14 members. Until 1983, however, Michigan actually had 19 congressional districts.
In a 19-district Michigan, each congressman would represent slightly over 520,000 people, around 185,000 fewer than at present. So how would this affect the balance of the congressional map?
The current Michigan map is a Republican gerrymander, but one which operates within specific limits. Redistricting in Michigan is guided by a statute mandating that division of counties and towns should be minimised and in particular that two different districts should not both cross the same county border. However, it is possible to remain within these guidelines whilst still drawing a strong partisan gerrymander, and indeed in the last two rounds of redistricting state Republicans have done so. Moreover, the language of the statute has been found to be unenforceable and hence even where it prevents Republican gerrymanders, it can simply be ignored (as Republicans did with MI-11 and MI-14 this time.)
If we assume that state Republicans attempted to gerrymander a 19-district map, what limits would they have faced? The answer, it turns out, is relatively few. Over-concentration of Democratic votes in major urban centres mean that they can be corralled in a few super-safe districts, allowing Republicans to dominate the bulk of the seats in a state where Obama beat Romney by 10 points and McCain by 16.
I've drawn what ought to be a relatively comfortable 13-6 map in favour of Republicans, with 3 VRA districts and only one district likely to be seriously contested outside wave years. No account has been taken of incumbents, because obviously if there were 19 districts the delegation would be different and trying to account for this introduces too many tendentious hypotheticals.
Whilst Obama won 15 of 19 districts in 2008, no fewer than 8 of them flipped back to Romney in 2012. McCain did so badly in the state in 2008 that his results are probably below the Republican congressional floor. Seven Michigan Republicans currently represent districts Obama carried in 2008 and we can therefore assume that the Romney figures give a better indication of underlying partisanship. All the Republican districts saw a hefty swing-back in 2012, whereas Democratic districts either stood still or drifted further leftwards.
Before going any further, I ought to justify the assumption of 3 VRA districts. This would equate to 15.8% of districts, whereas Michigan's black population is only 14.2%. Nevertheless, VRA districts aren't supposed to equate to a state's overall minority population, but rather should be drawn when concentrations of populations allow them to be drawn whilst satisfying the Gingles criteria. This is relatively easy to do in the Metro Detroit area and I suspect that with 19 districts, the Bush Sr. administration would have insisted on a 3rd VRA district in the early 1990s, following which the prohibition of retrogression would have sustained it to the present day.
All three VRA districts are more than 50% black by VAP, which slightly disadvantages Republicans by forcing heavily Democratic but overwhelmingly white areas of Downriver into the 9th district. However, prior to the striking down of Virginia's congressional map, it was generally assumed that an effective VRA district had to meet this metric and I've therefore retained it anyway.
You'll note that 2012 voting figures for MI-15, MI-18 and MI-19 are missing. Although I have voting figures for Detroit in 2012 by precinct, I haven't found any concordance of them to the voting districts in Dave's Redistricting App and therefore can't assign them to my hypothesised districts. But Republicans aren't going to win black-majority districts in Metro Detroit, so that shouldn't matter too much. All percentages are as a share of the two-party vote.
A brief district-by-district run through follows:
MI-1 (Blue) - reduced size means that this is more of an Upper Peninsula district than in real life, though even so it only accounts for 60% of the population. The ancestral Democratic roots of the UP might make this a little more competitive than the real MI-1, but those roots are fading and in terms of presidential vote this matches up pretty well to the district Dan Benishek holds down. Safe R
MI-2 (Green) - a district made up of Muskegon and the north-west of the Lower Peninsula ought to be swingy. Which is why this district stops just outside Muskegon and instead takes enough reliably red central Michigan turf to prevent real competition. There's enough room to spare for it to also pick up the college town of Mount Pleasant. Safe R
MI-3 (Purple) - Muskegon and two-thirds of Grand Rapids provide a fairly strong Democratic floor here. On the other hand, monolithically Republican Ottawa County sticks a clear ceiling on the prospects for team blue. Yes, Obama did win here, but any reasonably competent local Republican ought to lock this down with ease.
Safe R
MI-4 (Red) - from Grand Rapids this district extends west past Lansing and into Shiawassee County. This is one of only four districts McCain won and as such should be considered absolutely safe. Safe R
MI-5 (Gold) - the counties along Saginaw Bay has historically been fairly good for Democrats, whereas Midland is notably more Republican than cities of similar size elsewhere in Michigan. These two elements more or less cancel each other out, with rural areas of central Michigan tipping the balance inexorably in favour of Republicans. Safe R
MI-6 (Teal) - with only half a million people to play with, diluting Lansing with Livingston County no longer works. On the other hand, with five extra congressional districts to play with, it's easy to connect all central Michigan's largest cities into a single baconmander without having to double-cross any county boundaries. From Lansing, this seat snakes around Shiawassee County and into Genesee County, where it picks up Flint, its bluest suburbs and reliably Democratic townships in the north-west of the county. A finger then reaches north to Saginaw to produce the most Democratic district possible outside Detroit. This would probably suit a Flint Democrat, although Lansing mayor Virg Bernero might also fancy his chances here.
Safe D
MI-7 (Grey) - Kalamazoo is the largest city here, but it's stripped of many of its suburbs and combined with reliably Republican turf south of Grand Rapids. Even if this were to be lost in a presidential year (and it's unlikely), Democratic reliance on student turnout would easily flip this back in the next mid-term. Safe R
MI-8 (Slate Blue) - no district stretching from Lake Michigan to Jackson is going to be pretty, but that isn't the point. A cleaner district would be redder, but as it is Republicans ought to have few troubles here anyway. Safe R
MI-9 (Cyan) - the only swing district on this map begins in the reliably Republican southern tier of Michigan counties before heading east to swingy Lenawee and Monroe and curving up to grab parts of the Downriver region of Wayne County. The result is a district that Obama won in both 2008 and 2012. Nevertheless, an R+1 district is far from hopeless and in mid-term years Republicans would definitely be favoured here. They'd also be helped by the finger of MI-10 into Lenawee, which carves out not just Adrian but also Tim Huelskamp's home, and thus opens the way for a more centrist Republican. Lean R
MI-10 (Magenta) - this district has tentacles stretching in all directions throughout southern Michigan, but the core is Livingston County, which ought to be able to outweigh any other element of the district. Obama's narrow 2008 victory notwithstanding, this therefore ought to favour Republicans even in most wave years. Safe R
MI-11 (Chartreuse) - this Oakland-Genesee district is made up of the most Republican parts of these two generally Democratic counties. The major danger to Republicans is that its lack of geographic cohesiveness leads to an Oakland representative who neglects Genesee at the expense of Troy and Bloomington Hills. Assuming half-decent constituent service, however, Republicans will win going away in this district. Safe R
MI-12 (Cornflower blue) - this is predominantly a Thumb district, but in addition there's a sliver of Oakland County and a number of Saginaw's suburbs. These comprise most of the Democratic strength here, but it's still a district Obama lost twice and wouldn't feature on any target lists. Safe R
MI-13 (Dark salmon) - this is a relatively compact Macomb district, although in the south-east it crosses over into Wayne to add the Grosse Pointes (mostly for VRA purposes.) A mixture of swingy territory and Republican strongholds, this is the second most Republican district on the map. Safe R
MI-14 (Olive) - By and large this map attempts to ensure Democrats represent only heavily minority districts, denying a platform for white Democrats with statewide potential. Nevertheless, Macomb and Oakland just contain too many working-class white Democrats for a Michigan equivalent of the Delaymander to be possible. In general terms this is a compacted version of the MI-9 currently represented by Sander Levin and although it's the least Democratic of the seats this map concedes, it's nevertheless well beyond the reach of the Republicans. Safe D
MI-15 (Orange) - the first of the VRA districts is the most suburban. Although it contains north-west Detroit at one end and Pontiac at the other, the centre of gravity is provided by the black middle-class city of Southfield and rapidly diversifying Farmington. White liberals in West Bloomfield and Pleasant Ridge are also vote-sinked here, but the district's representative would probably be a black politician from southern Oakland (and if not, the long-standing Detroit congressman who'd hold it instead would certainly be succeeded by a representative from Oakland.) Safe D
MI-16 (Lime) - the VRA is used to justify a double-cross of the Wayne-Oakland county border. Traditionally Republican parts of north-west Wayne are combined with implacably Republican western Oakland, ensuring that even if the former diversify, this district will continue to hold firm. McCain won by more than 3 points here, which might not sound like much but makes this the most Republican district on the map. Safe R
MI-17 (Dark blue) - as well as a white working-class Macomb-Oakland district, any Republican Michigan map needs an Ann Arbor/Ypsilanti vote-sink. This one reaches eastward to grab moderately Democratic and overwhelmingly white parts of western Wayne and then meanders into the Washtenaw countryside. It'd be possible to pack it further by grabbing more of western Wayne instead, but only at the expense of preventing MI-18 from remaining black-majority. It's still safe as houses, though.
Safe D
MI-18 (Yellow) - this VRA district snakes south-west from western Detroit to pick up the city's blacker suburbs, such as Redford Township, Inkster and Romulus. Once it runs out of these, it gets up to population by adding the least overwhelmingly white cities and townships it can. Safe D
MI-19 (Yellow/Green) - the bulk of Detroit is found here, both in terms of area and population. The balance is provided Dearborn and by racially-mixed cities in the northern parts of Downriver. Safe D
Obviously this is an unlikely hypothetical, but it does demonstrate the extent to which Democratic congressional troubles in Michigan are to a large extent a problem of geography. It's easy to combine 80% Obama cities with their 60% Obama inner suburbs, whilst overwhelming 55% Obama outer suburbs in a sea of 45% Obama countryside. Remove the 3 VRA districts and you'd actually have a state Romney narrowly won. Whilst undoing the Republican gerrymander would help a little, in the end any scheme focusing on the integrity of administrative sub-units is going to disadvantage Michigan Democrats.
Comments? Suggestions? Criticism?