Analysis by Harry Enten, CNN

(CNN) -- The Commerce Department announced Monday the addition of a citizenship question to the 2020 Census in order to, they say, get the most accurate count of citizens at the most granular level.

Opponents of the move say the change will lead to an undercount of non-citizen residents here both legally and illegally who might fear answering the questions about their citizenship.

If there is an undercount, areas more favorable to Democrats will be the most adversely affected, with congressional apportionment and fund apportionment, which are Census-based, on the line.

According to 2016 American Community Survey, there are 13 million non-citizens living in states (and Washington DC) won by Democrat Hillary Clinton in 2016. There are only about 9.5 million non-citizens living in states won by Republican Donald Trump. In other words, about 58% of non-citizens live in Clinton states and 42% live in Trump states.

These percentages, however, undersell the political impact to blue states. The vast majority of the states to be impacted are blue states. Clinton-won states make up 76% of the states where there are more non-citizens than the average state to Trump's 24%.

The top five states by this metric are California, Texas, Nevada, New York and New Jersey. Non-citizens made up between 10% and 14% of each of these state's populations. Clinton carried all of these states, except Texas.

The discrepancy between the percentage of blue states impacted versus percentage of residents is because there are lot of non-citizens living (nearly 3 million) in Texas. Texas was a state Trump carried by only 9 points in 2016, seven less points than Mitt Romney in 2012.

In fact, there are very few safe Republican states where there are a lot of non-citizens. Only 5% of all the states with a percentage of non-citizens greater than the average state were won by Trump by more than 10 percentage points. On the other hand, 57% of all the states with an above average percentage of non-citizens are states that Clinton won by at least 10 percentage points.

The people who are the most likely to be hesitant to answer a Census form involving citizenship are those who are here illegally. Getting accurate measurements on this population is harder, but the Pew Research Center made a good attempt in 2014.

Of the 11 million people in the country illegally in 2014, according to Pew, 6 million were in states Clinton won to 5 million in states Trump won. The percentage difference is a bit closer than non-citizens overall, though there are still more immigrants here illegally in Clinton-won states at 54% to Trump-won states at 46%.

Again though, this likely undersells the number of Clinton states impacted because of how many immigrants here illegally reside in Texas. Clinton-won states make up 71% of states where immigrants who are here illegally make up a larger share of the state's population than the average to Trump's 29%. The top five states, according to Pew, were Nevada, Texas, California, New Jersey and Arizona. All of these states had about 5% to 7% of the total population being made up of residents here illegally. Trump did win two of these states (Arizona and Texas), though both of them by less than 10 points. Clinton easily took both California and New Jersey.

Indeed, the state most likely to be impacted by these questions is California. Not only is it in the top five for the percentage of non-citizens and residents who are here illegally that make up its the state's population, but the raw numbers are staggering. There are over 5 million non-citizens in California (or about 13.5% of the total population). There were about 2.4 million residents who are here illegally who live in California (or about 6% of the population). This could be why California is now suing the federal government to stop the citizenship question from being asked.

California and blue states in general could definitely lose seats in Congress too if non-citizens (here legally or illegally) are afraid to answer the Census. Let's say every non-citizen were afraid to answer Census. That's unlikely to happen, but it gives you a good idea of the stakes.

To calculate how many seats each state could lose if non-citizens decided not answer to the Census, I used the 2016 American Community Survey to determine how many residents there were in total and how many citizens there were in each state. I then plugged in these figures into a handy appointment calculator designed by the University of Michigan to determine how apportionment would differ in these different scenarios.

The 2016 American Community Survey suggests California could lose up to four seats in Congress in such a scenario. No other state would lose more than one. In total, Clinton-won states could lose up to three seats in Congress, while Trump-won states would gain three seats in Congress.

Now, the effect is more difficult to calculate if only residents here illegally would be afraid to answer the Census. It will also likely be smaller because there are far fewer immigrants here illegally than the overall number of non-citizens. California, though could still have had up to 2 fewer seats, applying the measurements from the Pew Research Center. Blue states would lose one or two seats at the expense of red states, depending on the exact calculation.

Correction: This post has been updated to reflect that there are about 9.5 million non-citizens living in states won by Trump.

The