Quick note: article shared by the Reserve Bank of India, the American Statistical Association, and written up in BloombergView. Tweeted here by the great Nassim Taleb. Polling statistics from Statistical Ideas topped 5 million reads in the past month and often read and shared by top advisors/surrogates from both sides.
With a quick glance, what does this map remind you of? You’d be forgiven if you felt it was last week’s presidential elections. 45 of the state colors (representing a stunning 94% of the U.S. population) aligns impeccably with the political party that won (i.e., red = Republicans, blue = Democrats). But instead what we are showing in the wide swath of red across the U.S. are those states that are relatively, economically starved regions and where most Americans live. The fact that just two basic economic variables can almost perfectly predict these election results therefore, provides you ample opportunity to see where the vulnerabilities for Hillary Clinton lay, when heading into this election season. It’s a good thing Ms. Clinton wasn’t purely evaluated based on these two economic variables, else (as shown here) she would have lost both the popular vote (by a wide 10 percentage points!), as well as the electoral vote (240 versus 298). There will be another chance for the Democrats at the election 4 years from now. Instead of getting grouchy and flaming up a riot on some street corner, now is a better time to learn things anew about how most Americans live, so that they are brought into the fold of policy, and political discourse.
The moral here is that for the plurality of Americans with less wealth accumulated, they expect somewhat greater job security in order to feel more satisfied. For the sake of putting on our hat of Americans first (politics second), let’s hope that over the next four years: the vast majority of Americans begin to see more satisfactory policies that positively influences their lives.
Last, we leave you with this:
With a quick glance, what does this map remind you of? You’d be forgiven if you felt it was last week’s presidential elections. 45 of the state colors (representing a stunning 94% of the U.S. population) aligns impeccably with the political party that won (i.e., red = Republicans, blue = Democrats). But instead what we are showing in the wide swath of red across the U.S. are those states that are relatively, economically starved regions and where most Americans live. The fact that just two basic economic variables can almost perfectly predict these election results therefore, provides you ample opportunity to see where the vulnerabilities for Hillary Clinton lay, when heading into this election season. It’s a good thing Ms. Clinton wasn’t purely evaluated based on these two economic variables, else (as shown here) she would have lost both the popular vote (by a wide 10 percentage points!), as well as the electoral vote (240 versus 298). There will be another chance for the Democrats at the election 4 years from now. Instead of getting grouchy and flaming up a riot on some street corner, now is a better time to learn things anew about how most Americans live, so that they are brought into the fold of policy, and political discourse.
There is a difference in
economics for the many states that voted for Donald Trump, versus those small,
dense pockets that voted for Hillary Clinton.
This is a nifty geospatial pattern tested across 51 states+D.C. and that endures in recent elections, though not an ideal forecasting model for the future. Republican states had a collective GDP per capita (chained to 2009
dollars) of $45k, which is fine relative to some other nations, but for Democrat states it was nearly 25% higher at $57k. And when
it came to unemployment, Republican states saw only a 2.7 percentage point
improvement since the last election, while Democrat states saw a 3.6 percentage
point improvement. One can attempt to
argue that Republican states may have had far greater economic strides in
the four years prior to the last election (between 2008-2012), a period where I
led multiple economic-related positions for the federal government. But of course this said argument is weakened
by the fact that our democratic president, Barack Obama was in fact re-elected
then in 2012 with a voting map that looked somewhat similar to what we saw last
week.
Using advanced machine learning
we were able to isolate two fundamental, economic variables (can be consused either individually, or combined). A splendidly compact model that in the end powerfully explains most of the voting differences. What are
these two joint economic signals that explained much of what occurred Election Day
2016?
1. First,
there were 20 states -plus Washington- with above average GDP per capita (above $50k). This also wistfully admits that this relevant economic
variable is very positively skewed, or highlighting inequality towards the hands of Democrat states. Of
those:
a. The
model predicts with 100% accuracy Hillary winning the 12 states -plus D.C.- where
the unemployment rate dropped beyond 1.8% since the last election
b. The
model near-perfectly predicts Trump winning the 8 states where the unemployment rate dropped
no more than 1.8% since the last election
2. Second,
there were 30 states with below average GDP per capita (below $50k). Of those:
a. The model
predicts with 100% accuracy Hillary winning the 2 states where the
unemployment rate dropped beyond a high 4.6% since the last election
b. The
model near-perfectly predicts Trump winning the 28 states where the unemployment rate dropped
no more than 4.6% since the last election. To the consternation of most popular pundits, a total of 6 states flipped and all were in just in this category; nearly a hundred electoral votes 100% accurately flipped to Trump from previously voting Obama (and can not be explained by other broad issues such as health or political-correctness culture).
The moral here is that for the plurality of Americans with less wealth accumulated, they expect somewhat greater job security in order to feel more satisfied. For the sake of putting on our hat of Americans first (politics second), let’s hope that over the next four years: the vast majority of Americans begin to see more satisfactory policies that positively influences their lives.
Last, we leave you with this:
I wonder if this would hold if you looked below state-level data to see who actually voted for Trump/Clinton. I wouldn't be surprised if Trump voters in Trump-voting states had higher incomes and higher employment than Clinton-voters, which would tend to undermine this conclusion?
ReplyDeleteHi Aaron, hope all is well. Don’t believe what you are stating is correct, and we also don’t have sufficient data to test it. Also it doesn’t matter to scrutinize things more disaggregated, when the election is based on electoral votes and not the popular vote! Thanks.
Delete