What Happened in Minnesota

Adam Martin
The Book Aisle
Published in
25 min readJan 20, 2021


This is the sixth part in my 2020 election analysis and the last part that will focus on the Midwest. Minnesota is unique in that it’s the only state I’ll be covering in this series that Hillary Clinton won in 2016. Granted, the race there was much tighter in 2016 than in 2012, but this still spoke a lot about the state’s politics. Its association with the “Blue Wall”, coupled with its position as the epicenter of 2020’s BLM protest activity makes Minnesota an interesting state to discuss.

With that said, let’s dive in.

Big Picture

Joe Biden- 1,717,077 (52.40 percent, up 5.96 percent from Clinton in 2016)

Donald Trump- 1,484,065 (45.28 percent, up 0.36 percent from 2016)

FiveThirtyEight Projection- 53.7 for Biden (up 1.3), 44.6 for Trump (down 0.68)

RCP Average- 48 for Biden (down 4.4), 43.7 for Trump (down 1.58)

OurProgress Projection- 51.74 for Biden (down 0.64), 44.28 for Trump (down 1)

As of Election Day, there have been 156,852 COVID cases in Minnesota, ranking 22nd among the states. This isn’t too bad, considering that Minnesota is the 22nd largest state by population. Indeed, its incidence rate was only 2.8 percent as of Election Day, ranking 25th among the states. On the other hand, there have been 2,499 deaths (ranked 24th), setting the fatality rate at 1.6 percent (ranked 32nd). Meanwhile, while still damaging, Minnesota’s economic trajectory during the pandemic has been more mild than most of the country. Minnesota had the third smallest one-month increase in its unemployment of any state, going from 3.5 percent unemployment in March to 8.6 percent in April (5.1 percentage points). On the whole, the pandemic’s effects on Minnesota have been fairly modest, with average infection and fatality rates along with a relatively small spike to its unemployment rate.

Meanwhile, there have been 233 BLM protests in Minnesota as of Election Day, according to the US Crisis Monitor. Minnesota is unique in this regard, as it was the state where George Floyd, a black man, was killed while in police custody during Memorial Day weekend. This killing was the initial catalyst that sparked the wave of BLM protests across the United States that helped define the last few months of the campaign. And while there was substantial protest activity within the state, Minnesota only ranked 24th for number of protests which may seem surprising. But aside from only being 22nd in population, Minnesota is not well positioned demographically to be a top protest hub. It has the 24th largest black population by raw numbers (or 29th by percentage of the total population) and 22nd largest young adult population (or 41st by percentage). On the other hand, Minnesota does have the 18th highest amount of college graduates (and 11th by percentage). So while it does have a substantial college-educated population, the lack of racial diversity holds the state back from having more protest activity.

As for the exit polls, Biden retained 96 percent of Minnesota’s Clinton voters while picking up 54 percent of third-party and write-in voters and 62 percent of those that did not vote in 2016. Meanwhile, President Trump retained 92 percent of his voters from 2016. Interestingly, there were slightly more Trump voters in this sample than Clinton voters; however, Biden’s advantage with third-party voters and non-voters from 2016 helped him out tremendously.

For race, Biden won 51 percent of white voters, up from the 43 percent Clinton won in 2016, as well as 60 percent of Hispanic voters. Considering that Biden only won 41 percent of white voters nationwide and the fact that Minnesota’s electorate was 87 percent white, these numbers were critical for his performance here. Interestingly, though, Biden only won 77 percent of black voters (down from the 87 percent he received nationwide) and only 66 percent of nonwhite voters (compared to the 71 percent Clinton received in Minnesota four years ago). This is an unusual finding, given that the initial catalyst of the BLM protests occurred in Minneapolis; however, Biden still won these voters by strong margins.

Regarding educational attainment, Biden won 63 percent of college graduates (up from 53 percent for Clinton), including 62 percent of white college graduates (up from 50 percent for Clinton). Meanwhile, President Trump only received 53 percent of voters without a college degree (down from the 55 percent he got four years ago), but 57 percent of white voters without a college degree (still down from 60 percent four years ago).

As for age, Biden was pretty successful with multiple age brackets, although he did receive less support from older voters. He won 65 percent of voters in the 18–29 bracket, a 20-point improvement from Clinton in 2016, as well as 54 percent of voters in the 30–44 bracket, a 1-point improvement. Meanwhile, he edged Trump out in the 45–64 age bracket winning 50 percent of those voters (an 8-point improvement from Clinton) while being edged out by Trump among voters over 65, winning 48 percent (a 1-point loss from Clinton). Overall, Biden was more competitive with older voters in Minnesota than the rest of the nation while retaining his strong support with voters under 45.

Geographically, Biden did very well in the cities, winning 68 percent of urban voters (up 13 points from Clinton in 2016). On the other hand, it appears that Biden underperformed elsewhere. President Trump won 56 percent of suburban voters, up from the 48 percent he received nationally and the 47 percent he received in Minnesota four years ago. Furthermore, Trump won 60 percent of rural voters in the state, up from the 58 percent he received in 2016. This indicates that while Biden was very successful in Minnesota, that successful was mostly limited to the cities.

Socioeconomically, only 7 percent of Minnesota voters said they have experienced severe financial hardship due to the pandemic (down from 16 percent nationwide), of which 66 percent voted for Biden (down from 69 percent nationwide). About 34 percent of voters said they have experienced moderate financial hardship (compared to 39 percent nationally), of which 56 percent voted for Biden (down from 59 percent nationally). And about 58 percent said they have not experienced any financial hardship (up from 44 percent nationwide), of which only 50 percent voted for President Trump (down from 60 percent nationwide). Regarding income brackets, Biden won 56 percent of voters making less than $50,000, up from the 45 percent Clinton received in 2016. Biden also won 51 percent of voters in the $50,000 to $99,999 bracket (up from 45 percent for Clinton) as well as 56 percent of voters making over $100,000 (up from 52 percent for Clinton). Overall, this demonstrates a broad socioeconomic appeal for Biden statewide, which hasn’t been seen in other Midwestern states, like Ohio or Iowa.

When asked about their most important issue, 32 percent of Minnesota voters said the economy (compared to 35 percent nationally), of which 88 percent voted for President Trump. Meanwhile, 19 percent said the pandemic was the most important issue (compared to 17 percent nationally) and only 14 percent said racial inequality (compared to 20 percent nationally), of which 94 percent and 90 percent voted for Biden respectively (compared to 81 percent and 92 percent nationally). Considering Minneapolis’s significance in the wave of BLM protest activity, it may be surprising that relatively few of its state’s voters cited racial inequality as their most important issue; however, it may be a reflection of the state’s demographics, which are overwhelmingly white. Finally, when asked what was more important to accomplish, 54 percent of voters said it was more important to contain the virus (compared to 52 percent nationally), of which 85 percent voted for Biden. On the other hand, 40 percent said it was more important to reopen the economy (compared to 42 percent nationally), of which 86 percent voted for Trump.

Now that we’ve gone over the exit polls, let’s run some regressions. First, here are various regression results for Biden’s county-level vote share.

For those who have read the previous parts in this series, there are a lot of familiar themes in these regression results. Similar to Michigan and Pennsylvania, there is no relationship between the COVID incidence rate and the Biden vote share, although there does seem to be a positive relationship with the September unemployment rate (B=2.13; 2.36). Furthermore, there doesn’t seem to be a relationship between BLM protest activity and the Biden vote share. This may appear surprising, given the pivotal role that the state played in generating the ongoing BLM movement; however, this finding is consistent with those in other states.

Aside from the pandemic, economic, and political factors, we continue to see a persistent education gap, where Biden enjoys greater support in counties with more college graduates (B=1.01; 1.04) and less support in counties with fewer high school graduates (B= -2.23; -2.38). Not to mention that Biden also enjoys greater support in counties with larger black populations (B=1.85; 0.24; 1.7), Native American populations (B=0.48), and Hispanic populations (B=1.04; 0.44; 0.93).

For the most part, similar trends hold here as well, such as the widening education gap as well as the widening gap with multiple racial and ethnic groups. The one highlight from this Figure is that the unemployment increase from March to April (B=0.35; 0.27; 0.33; 0.27) is positively associated with the change in Democratic vote share in all four regression models. This indicates that, at least in Minnesota, counties that experience a larger economic shock early in the pandemic increase their support for Biden compared to Clinton in the general election. It should be noted that this finding is held back by the fact that Minnesota saw one of the smallest unemployment increases of any state. Even so, this offers at least some evidence that Biden benefitted from the economic downturn.

With all that established, let’s look at each of the subsections in Minnesota.

Solidly Democratic Counties

Cook (northeast corner of state, on Lake Superior and Canadian border)- Biden Hold

Biden- 2,496 (65.58 percent, up 9.26), Trump- 1,203 (31.61 percent, down 2.44)

OurProgress Projection- 62.99 for Biden (down 2.59), 37.01 for Trump (up 5.40)

Hennepin (Minneapolis city proper and surrounding communities)- Biden Hold

Biden- 532,623 (70.46 percent, up 7.32), Trump- 205,973 (27.25 percent, down 0.96)

OurProgress Projection- 70.13 for Biden (down 0.33), 29.87 for Trump (up 2.62)

Ramsey (St. Paul city proper and surrounding communities)- Biden Hold

Biden- 211,620 (71.50 percent, up 6.26), Trump- 77,376 (26.14 percent, up 0.04)

OurProgress Projection- 72.32 for Biden (up 0.72), 27.68 for Trump (up 1.54)

First, there are three solidly Democratic counties. Two of these, Hennepin and Ramsey, comprise the Twin Cities, the state’s main population center. Meanwhile, Cook seems to be the opposite: a small, rural county tucked away in the corner of the state, heavily forested and dotted with lakes. But for those who have read my two Wisconsin articles, you’ll spot some similarities with this subsection. Namely, Cook County is home to the Grand Portage Indian Reservation, which plays a key role in the county’s politics. Granted, the county is mostly white; however at 8.6 percent, Cook has a considerably larger Native American population than most of the country.

Overall, these counties have a history of supporting Democratic candidates. Cook has backed the Democratic candidate in eight of the last nine presidential elections (with the exception being in 2000), Hennepin has remained blue in every election since 1976, and Ramsey has voted Democrat in every election since 1928. In 2012, President Obama won this subsection with 63.5 percent of the vote. Four years later, Clinton still performed considerably well in this subsection. In fact, she actually improved upon Obama’s performance with 63.7 percent, although lower voter turnout did result in her receiving 2,784 fewer votes overall.

Going into 2020, Biden had many advantages in this subsection, helped by the fact that Clinton pulled her weight in this subsection. That, coupled with favorable demographics, allowed him to win the subsection with 70.7 percent of the vote, translating to 138,615 more votes than Clinton. He also built upon Clinton’s performance in individual counties, adding 6.3 percent in Ramsey, 7.3 percent in Hennepin, and 9.3 percent in Cook.

Demographically, this subsection is more diverse than the rest of the state. 27.9 percent of the population is nonwhite and 7.1 percent is Hispanic. In addition to black residents comprising 13.5 percent, this subsection includes large pockets of other racial groups, including Asian Americans (who make up 15.3 percent of Ramsey’s population) and Native Americans (who, as mentioned above, make up over 8 percent of Cook’s population). And regarding educational attainment, this subsection is very highly educated. All three counties are in the top ten for highest percentage of adults holding at least a bachelor’s degree, with Hennepin ranking 1st in the state (at 49.2 percent), Cook ranking 5th (at 42.4 percent), and Ramsey ranking 6th (at 41.9 percent).

Socioeconomically, this subsection performs pretty well, but has considerable inequality. By median household income, Hennepin and Ramsey are in the top half of the state, with Hennepin ranking 9th (at $76,067) and Ramsey ranking 23rd (at $62,937). Meanwhile, Cook is ranked 58th (at $54,465), placing it in the bottom half of the state. As for poverty, the situation in this subsection is reversed. On one hand, Cook is ranked at 36th lowest (with a rate of 9.6 percent). But on the other hand, poverty is a more substantial problem in the Twin Cities, with Hennepin in the middle of the pack (at 10.3 percent) and Ramsey having the sixth highest poverty rate in the state (at 14.6 percent). And for unemployment, this subsection fared well for the most part. Cook had the tenth lowest unemployment rate in August 2019 at just 2.3 percent, Hennepin was ranked 39th at 2.5 percent, and Ramsey was ranked at 57th at 2.7 percent. Overall, while this subsection is fairly prosperous, but inequality is a massive problem, especially in the Twin Cities. Hennepin and Ramsey has the second and third highest Gini Indices in the state (at 0.487 and 0.466), a reflection of their high median household income and high poverty rates.

As of Election Day, there have been 54,152 COVID cases in this subsection, setting the incidence rate at about 3 percent (slightly above the Minnesota average). Almost all of these cases have occurred in either Hennepin or Ramsey, with only 27 having occurred in Cook (setting its incidence rate at just 0.5 percent). On the other hand, there have been 1,390 deaths as of Election Day, setting the fatality rate at 2.6 percent (above the Minnesota average). All of these deaths have occurred in either Hennepin or Ramsey; Cook didn’t have a single COVID death as of Election Day. Regarding timing, there was a large spike between the early months and summer months in this subsection followed by a smaller increase from the summer to the early fall months. And for unemployment, there was a stronger impact in this subsection compared to the rest of the state, going from 2.8 percent in March to 9.1 percent in April (a 6.3-point increase). While infections were extremely low in Cook, this county took the largest hit, going from 5.2 percent to 15.3 percent (10.1 points). Meanwhile, the impact in the Twin Cities was more mild, even though infections were more widespread. And while there has been some recovery in the subsequent months, the unemployment rate in the subsection was still at 6.7 percent as of September. Cook had the largest recovery, getting its unemployment rate down to 4.7 percent, below its March rate.

Finally, as of Election Day, there have been 137 BLM protests in this subsection. As the site of George Floyd’s killing in police custody, Hennepin had the most protests at 74 while the neighboring Ramsey had 61. Conversely, Cook only had 2 protests. As stated above, there’s are considerable pockets of black residents in Hennepin and Ramsey, although not more than the national average. Even so, the proximity to a catalyst as large as Floyd’s death was a major driving force for protest activity in the Twin Cities. The efforts were also bolstered by the large share of college graduates in the area as well as the average, but still sizable base of young adults.

Swing Counties

Carlton (Duluth metropolitan area, on Lake Superior and Wisconsin border)- Biden Hold

Biden- 10,098 (49.58 percent, up 3.12), Trump- 9,791 (48.07 percent, up 3.26)

OurProgress Projection- 58.47 for Biden (up 8.89), 41.53 for Trump (down 6.55)

Dakota (Twin Cities metropolitan area, on Wisconsin border)- Biden Hold

Biden- 146,155 (55.73 percent, up 7.42), Trump- 109,638 (41.81 percent, down 1.84)

OurProgress Projection- 54.34 for Biden (down 1.39), 45.66 for Trump (up 3.85)

Lake (northeastern part of state, on Lake Superior and Canadian border)- Biden Hold

Biden- 3,647 (50.64 percent, up 3.45), Trump- 3,393 (47.11 percent, up 2.15)

OurProgress Projection- 58.10 for Biden (up 7.46), 41.90 for Trump (down 5.21)

Olmsted (Rochester city proper and surrounding communities)- Biden Hold

Biden- 49,491 (54.16 percent, up 8.90), Trump- 39,692 (43.43 percent, down 1.08)

OurProgress Projection- 52.35 for Biden (down 1.81), 47.65 for Trump (up 4.22)

Saint Louis (Duluth city proper and surrounding communities)- Biden Hold

Biden- 67,704 (56.64 percent, up 5.26), Trump- 49,017 (41.01 percent, up 1.31)

OurProgress Projection- 62.18 for Biden (up 5.54), 37.82 for Trump (down 3.19)

Washington (Twin Cities metropolitan area, on Wisconsin border)- Biden Hold

Biden- 89,165 (53.46 percent, up 6.95), Trump- 73,764 (44.23 percent, down 0.44)

OurProgress Projection- 52.41 for Biden (down 1.05), 47.59 for Trump (up 3.36)

Next, there are six swing counties which are concentrated in the northern and southern half of the state. The counties in the southern half are mostly metropolitan, consisting of Twin Cities suburbs as well as Rochester, a medium-sized city. The counties in the northern half are more rural, although they also include Duluth, another medium-sized city. But geography isn’t the only dividing line between these two halves. Similar to Cook, the counties in the northern half of the state have a very long tradition of supporting Democrats, having back all the Party’s candidates since at least 1936. While Trump was more competitive than other Republicans in these counties, he couldn’t do well enough to break that tradition. Meanwhile, the counties in the southern half have either alternated between Democrat and Republican or, in the case of Olmsted, have historically been Republican until relatively recently. Obama flipped these three counties in 2008, however, and they’ve managed to stay in that column ever since.

In 2012, President Obama won all six of these counties with 53.1 percent of the vote. Based on these results, the three northern counties were classified as solidly Democratic following 2012. But, as expected, President Trump made the region more competitive. What’s interesting about 2016 is that while Clinton only received 48 percent of the subsection’s vote, she still managed to hold all six counties. Saint Louis was the only county she carried by over 50 percent while in the others, third-party candidates received enough votes to deny either Clinton or Trump a majority. Geographically, it’s clear that Trump’s increased support came mostly from the northern counties. While Clinton only sustained minor losses in the more urban southern counties, she lost more than 10 points from President Obama in each of the northern counties (which, again, were classified as solidly Democratic just four years earlier). As will be discussed below, demographic and socioeconomic differences between the two counties can explain this discrepancy. Overall, Clinton’s performance translated to a net loss of 31,694 votes in the subsection.

Now in 2020, Biden retained all six counties by receiving 54.9 percent of the vote, slightly better than President Obama’s 2012 performance. Once again, geography is telling. On one hand, Biden only made modest improvements in the northern counties that Trump had made a major dent in four years ago. While he elevated Saint Louis (the one northern county with a notable city) back into the solidly Democratic column, he only got 50.6 percent in Lake and 49.6 percent in Carlton. To that end, Biden didn’t do much to stave off the Trump support there. But on the other hand, Biden made massive gains in the larger, more metropolitan southern counties. In Washington and Dakota, both consisting of Twin Cities suburbs, Biden increased Clinton’s vote share by 6.9 percent and 7.4 percent respectively, while elevating Dakota into the solidly Democratic column. Meanwhile in Olmsted, he increased Clinton’s vote share by 8.9 percent. These trends are consistent with those seen in other states, where Biden made major gains in suburban counties even while lagging in more rural and working class areas.

Demographically, this subsection is 14.1 percent nonwhite, 5.5 percent black, and 5.2 percent Hispanic. Once again, there is a clear divide between the northern and southern counties. For the most part, the southern counties are more diverse, each being in double digits for nonwhite residents. While blacks make up a decent chunk of this population, the southern counties also have substantial shares of Asian American and multiracial residents. On the other hand, the northern counties are more white and have fewer residents of these other demographic groups; however, Carlton and Saint Louis have notable pockets of Native American residents, a reflection of their proximity to Native American reservations. And regarding educational attainment, all six counties are in the top half of the state and five of them are in the top twenty; however, the geographical divide persists. The three southern counties are not only in the top ten, but in each of them, college graduates account for over 40 percent of their populations. Meanwhile, the three northern counties are below 30 percent, but are still in the top half, with Lake ranking 12th (at 29.8 percent), Saint Louis ranking 16th (at 28.7 percent), and Carlton ranking 39th (at 22.5 percent).

Socioeconomically, this subsection does pretty well, although there does seem to be a geographic split. By median household income, three counties are in the top in the state, with Washington ranked 3rd (at $95,840), Dakota ranked 5th (at $86,403), and Olmsted ranked 10th (at $76,001). All three of these counties are in the southern half of the state, with two of them in the Twin Cities metropolitan area. Meanwhile, the three counties in the northern half of the state still hold up, although they’re not as prosperous. Lake ranked 28th (at $61,015), Carlton ranked 34th (at $60,412), and Saint Louis ranked 61st (at $53,833). Regarding poverty, this geographic split holds. On one hand, Washington, Dakota, and Olmsted rank pretty low for poverty, all of which have a rate under 8 percent. On the other hand, Lake, Carlton, and Saint Louis all have poverty rates above 8 percent, with Saint Louis having the fifth highest rate in the entire state (at 14.7 percent). And for unemployment, this subsection spans the spectrum of outcomes for the state, although the rate was pretty low. On one end, Olmsted had the lowest unemployment rate in August 2019 at just 2.1 percent. And on the other end, Carlton had the highest unemployment rate in August 2019 at 3.2 percent. Overall, while this subsection performs pretty well socioeconomically, there is a clear split between the more affluent southern half and the less well-off northern half.

As of Election Day, there have been 26,657 COVID cases in this subsection, setting the incidence rate at 2.4 percent (slightly below the state average). While the incidence rate appears to be slightly higher in the southern counties than the more remote northern counties, there isn’t a substantial difference between the two. On the other hand, there have been 324 deaths as of Election Day, setting the fatality rate to just 1.2 percent. Regarding timing, there has been a steady increase in cases in the months since the start of the pandemic. This increase over time is reflected in all six counties. And for unemployment, the impact has reflected that seen in the rest of the state, going from 3.2 percent in March to 8.9 percent in April. There doesn’t seem to be any major differences between the two regions.

Finally, as of Election Day, there have only been 43 BLM protests in this subsection. Interestingly, over half of these protests have occurred in Saint Louis which, despite being home to a medium-sized city in Duluth, doesn’t appear to be a traditional hub for intense protest activity. Meanwhile, there have only been 10 protests between Dakota and Washington and only 6 in Olmsted. While this subsection does have a substantial share of college graduates and young adults, the lack of racial diversity can explain the lack of urgency for intense protest activity.

Obama-Trump Counties

Beltrami (Bemidji city proper and surrounding communities)- Trump Hold

Biden- 11,426 (47.24 percent, up 6.95), Trump- 12,188 (50.39 percent, up 0.38)

OurProgress Projection- 52.33 for Biden (up 5.09), 47.67 for Trump (down 2.72)

Blue Earth (Mankato city proper and surrounding communities)- Biden Flip

Biden- 18,330 (50.84 percent, up 7.90), Trump- 16,731 (46.41 percent, down 0.23)

OurProgress Projection- 53.47 for Biden (up 2.63), 46.53 for Trump (up 0.12)

Chippewa (western part of state)- Trump Hold

Biden- 2,226 (33.67 percent, up 1.88), Trump- 4,250 (64.29 percent, up 3.79)

OurProgress Projection- 45.60 for Biden (up 11.93), 54.40 for Trump (down 8.89)

Clay (Fargo, North Dakota metropolitan area)- Biden Flip

Biden- 16,357 (50.74 percent, up 6.62), Trump- 15,043 (46.67 percent, up 0.59)

OurProgress Projection- 52.82 for Biden (up 2.08), 47.18 for Trump (up 0.51)

Fillmore (Rochester metropolitan area, on Iowa border)- Trump Hold

Biden- 4,551 (37.48 percent, up 2.46), Trump- 7,301 (60.14 percent, up 3.41)

OurProgress Projection- 47.82 for Biden (up 10.34), 52.18 for Trump (down 7.96)

Freeborn (southern part of state, on Iowa border)- Trump Hold

Biden- 6,889 (40.96 percent, up 3.33), Trump- 9,578 (56.95 percent, up 2.08)

OurProgress Projection- 51.45 for Biden (up 10.49), 48.55 for Trump (down 8.40)

Houston (southeastern corner of state, on Iowa and Wisconsin borders)- Trump Hold

Biden- 4,853 (42.42 percent, up 3.33), Trump- 6,334 (55.37 percent, up 2.41)

OurProgress Projection- 48.74 for Biden (up 6.32), 51.26 for Trump (down 4.11)

Itasca (northern part of state)- Trump Hold

Biden- 10,786 (40.61 percent, up 2.86), Trump- 15,239 (57.37 percent, up 3.27)

OurProgress Projection- 49.94 for Biden (up 9.33), 50.06 for Trump (down 7.31)

Kittson (on Canadian and North Dakota borders)- Trump Hold

Biden- 1,006 (38.12 percent, up 3.61), Trump- 1,546 (58.58 percent, up 2.02)

OurProgress Projection- 49.62 for Biden (up 11.50), 50.38 for Trump (down 8.20)

Koochiching (northern part of state, on Canadian border)- Trump Hold

Biden- 2,659 (38.41 percent, up 2.17), Trump- 4,131 (59.68 percent, up 3.59)

OurProgress Projection- 49.75 for Biden (up 11.34), 50.25 for Trump (down 9.43)

Lac qui Parle (western part of state, on South Dakota border)- Trump Hold

Biden- 1,446 (35.79 percent, up 2.06), Trump- 2,528 (62.57 percent, up 3.10)

OurProgress Projection- 46.46 for Biden (up 10.67), 53.54 for Trump (down 9.03)

Mahnomen (northwestern part of state)- Trump Hold

Biden- 1,112 (48.26 percent, up 5.93), Trump- 1,142 (49.57 percent, up 4.46)

OurProgress Projection- 56.42 for Biden (up 8.16), 43.58 for Trump (down 5.99)

Mower (southern part of state, on Iowa border)- Trump Hold

Biden- 8,899 (46 percent, up 4), Trump- 10,025 (51.82 percent, up 2.03)

OurProgress Projection- 55.88 for Biden (up 9.88), 44.12 for Trump (down 7.70)

Nicollet (Mankato metropolitan area)- Biden Flip

Biden- 9,622 (50.31 percent, up 6.73), Trump- 9,018 (47.15 percent, up 0.53)

OurProgress Projection- 53.04 for Biden (up 2.73), 46.96 for Trump (down 0.19)

Norman (northwestern part of state, on North Dakota border)- Trump Hold

Biden- 1,404 (40.80 percent, up 2.04), Trump- 1,953 (56.76 percent, up 4.66)

OurProgress Projection- 51.08 for Biden (up 10.28), 48.92 for Trump (down 7.84)

Rice (Twin Cities metropolitan area)- Trump Hold

Biden- 17,402 (48.76 percent, up 4.26), Trump- 17,464 (48.94 percent, up 1.38)

OurProgress Projection- 52.40 for Biden (up 3.64), 47.60 for Trump (down 1.34)

Swift (western part of state)- Trump Hold

Biden- 1,784 (34.35 percent, up 0.75), Trump- 3,316 (63.86 percent, up 4.57)

OurProgress Projection- 49.05 (up 14.70), 50.95 for Trump (down 12.91)

Traverse (on North and South Dakota borders)- Trump Hold

Biden- 661 (35.46 percent, up 0.43), Trump- 1,172 (62.88 percent, up 4.57)

OurProgress Projection- 47.73 for Biden (up 12.27), 52.27 for Trump (down 10.61)

Winona (southeastern part of state, on Wisconsin border)- Biden Flip

Biden- 13,333 (49.07 percent, up 5.48), Trump- 13,227 (48.68 percent, up 2.19)

OurProgress Projection- 54.99 for Biden (up 5.92), 45.01 for Trump (down 3.67)

Next, there are 19 counties that flipped from President Obama in 2012 to Trump in 2016. Many of these counties are rural and tend to hug borders with other states (and even Canada in some cases), but there are also counties that include medium-sized cities along with several suburbs. Historically, many of these counties became reliably Democratic in 1988 following the two terms of Ronald Reagan, although some have been Democratic longer than that (Itasca and Swift, for example, had been consistently Democratic in almost every election since 1932). While many of them also occasionally supported Republicans, such as Reagan and George W. Bush, for the most part they were reliably Democratic before 2016.

In 2012, President Obama won all 19 counties with 53.7 percent of the vote and four of these counties were classified as solidly Democratic. Four years later, however, Clinton suffered massive losses in this subsection, only receiving 40.9 percent of the vote, translating to a net loss of 36,417 votes. On average, Clinton lost 14.6 percent in each county compared to President Obama in 2012. Clinton only kept her losses to single digits in three counties (Rice, Clay, and Nicollet) and in Swift, she lost over 20 percent from Obama. Indeed, this demonstrated to be a tremendous loss.

Going into 2020, the OurProgress model expected Biden to win back ten of these counties. In the end, however, Biden only recaptured four while receiving 45.9 percent of the vote. While this is a notable improvement from Clinton in 2016, it also doesn’t reach the same peak that President Obama attained in 2012. On average, Biden improved upon Clinton’s vote share by about 3.8 percent in each county. On the other hand, he did not underperform Clinton in any of these counties and in six counties, he added more than 5 percent onto Clinton’s performance, including the four that he flipped back.

While there are counties within the Twin Cities metropolitan area in this subsection, the four counties that Biden won were mostly in and around medium-sized cities, such as Mankato and Fargo, North Dakota. Mankato, mostly located in Blue Earth, is home to Minnesota State University, which has over 17,000 enrolled students and over 1,800 employees. To that end, that area has a base of support for Democrats such as Joe Biden. Meanwhile, while the main city is located in North Dakota, many of its suburbs are located in Clay County, Minnesota. Given its emergence as an innovative small city and home to food processing, technology, and healthcare, the area has attracted both residents that to voted for Trump in 2016 but also willing to flip to Biden in 2020. And finally, there’s Winona, a fairly small county in the southern part of the state.

Demographically, only 10.4 percent of the population is nonwhite, only 2.9 percent is black, and 5.1 percent is Hispanic. There are notable pockets of other groups, such as Beltrami, where Native Americans make up 22 percent of the population, but for the most part, the subsection is mostly white throughout. And regarding educational attainment, this subsection spans the spectrum of outcomes for the state. The four counties that Biden won have the most college graduates in the subsection and are in the top fifteen of the state. Blue Earth has the most in the subsection (with 33.6 percent), followed by Clay (with 33.5 percent), Nicollet (with 32.6 percent), and Winona (with 29.8 percent). On the other hand, Mahnomen has the fewest college graduates in the entire state, at just 12.2 percent.

Socioeconomically, this subsection spans the entire spectrum of outcomes for the state of Minnesota. For median household income, Rice has the highest in the subsection and ranks 15th in the state (at $67,024). Nicollet and Clay are not far behind, ranking 17th (at $66,209) and 19th (at $65,914) respectively. On the other hand, Mahnomen has the lowest value in the entire state at just $43,162. Koochiching and Beltrami join it in the bottom five, with values at $46,281 and $47,073 respectively. As for the poverty rate, Houston has the lowest in the subsection and ranks 22nd lowest in the state (at 8.5 percent). Nicollet and Chippewa are not far behind, ranking 25th (at 8.7 percent) and 26th (at 8.8 percent) respectively. On the other hand, the four counties with the highest poverty in the state are all in this subsection. Mahnomen has the highest rate (at 20.9 percent), followed directly by Beltrami (at 18.4 percent), Blue Earth (at 15.1 percent), and Koochiching (at 14.9 percent). And for unemployment, Nicollet had the fourth lowest unemployment rate in the state in August 2019 at just 2.2 percent with Fillmore not far behind at 2.2 percent. On the other hand, Koochiching had the highest unemployment rate in the entire state in August 2019, coming in at 5 percent. Meanwhile, Itasca and Mahnomen weren’t far behind, with rates of 4.1 percent and 3.4 percent respectively.

As of Election Day, there have been 15,562 COVID cases in this subsection, setting the incidence rate at 2.9 percent (slightly above the state average). There doesn’t seem to be a substantial difference in incidence between the counties that Biden won and the ones that President Trump retained. On the other hand, there have been 172 deaths, setting the fatality rate at 1.1 percent (below the state average). Regarding timing, there has been a steady increase in cases during the pandemic, from the early months to the summer to the early fall months. And economically, the subsection sees trends similar to those statewide, going from 3.8 percent unemployment in March to 7.3 percent in April. Several counties such as Kittson and Traverse saw minimal change to their unemployment rates; however, no county saw a tremendous spike above the subsection average. Ultimately, the subsection has largely recovered in the months since the start of the pandemic, getting its unemployment down to 4.2 percent by September.

Finally, as of Election Day, there have only been 13 BLM protests in the subsection. Many counties only had one protest while ten counties had no protests at all. Given the lack of racial diversity in the subsection, this makes sense. But even in the counties that Biden won, which also have more college graduates, protest activity was very minimal. And overall, the lack of substantial pockets of young adults further limits the amount of protest activity experienced in the subsection.


As with the other states, there are a few reclassifications for counties based on how their voting behavior changed from 2016 to 2020. These changes are reflected below.

Anoka (Solidly Republican to Swing)

Beltrami (Obama-Trump to Swing)

Blue Earth (Obama-Trump to Biden)

Carver (Solidly Republican to Swing)

Clay (Obama-Trump to Biden)

Dakota (Swing to Solidly Democratic)

Mahnomen (Obama-Trump to Swing)

Mower (Obama-Trump to Swing)

Nicollet (Obama-Trump to Biden)

Rice (Obama-Trump to Swing)

Saint Louis (Swing to Solidly Democratic)

Scott (Solidly Republican to Swing)

Winona (Obama-Trump to Biden)

Most of these counties have already been discussed, especially with the four Obama-Trump counties that Biden flipped. In addition, Biden made several of the Obama-Trump and traditionally Republican counties more competitive while shoring up support in certain swing counties by shifting them into the solidly Democratic column.

The most notable feature of this map is the concentration of newly classified swing and solidly Democratic counties in the Twin Cities metropolitan area, such as Anoka, Carver, Rice, and Scott. This observation reinforces Biden’s continued improvement in suburban counties outside major cities. While these suburban counties aren’t as locked in Democratic as those of other major cities, Biden at least made these areas more competitive, an indication of his nationwide performance.


On the whole, Minnesota didn’t seem like an exciting state to cover in this series. After all, it went to Clinton in 2016 and through that built-in Trump-proof support, Biden was unsurprisingly able to turn Minnesota into a solidly Democratic state this time around. Nor does Minnesota hold a historical distinction as a bellwether, like Ohio. Despite this, I covered this state for several reasons.

One was to round out my coverage of the Midwest, a region that I consider Minnesota an integral part of. But the more compelling stemmed from the state’s critical role in sparking the wave of BLM protests across the country. I wanted to explore how such protest activity affected the election here, especially in a predominantly white state. I think through this discussion, Minnesota became an interesting microcosm of trends seen across the country.

Particularly, BLM protests are not affecting voting behavior in a significant way, other than reinforcing partisan leanings that are already prevalent. For Minnesota, the protests were disproportionately concentrated in the Twin Cities themselves, an already solidly Democratic area. But once one gets out of the Cities and into the suburbs, protest activity dropped significantly (and it’s even worse when you leave the metropolitan area altogether). Yet despite this, Biden did well in the places he was expected to do well in, such as medium-sized cities, the northeast corner of the state, and in the suburbs. To that end, Biden succeeded due to factors independent of proximity to protest events.

So that will do it for the fourth part in this series. If you enjoyed this, please like and follow the Book Aisle. Also share this article on Facebook, LinkedIn, and other social media platforms.