The Demise of the Demise of Donald J Trump

Without having looked back at this years posts, I am aware there are at least several elections I have not properly recapped and analysed my predictions. I intend to redress this towards the end of the year with a post considering each election covered by this blog. And there are several elections ahead that I probably will not cover, but I will at least allow myself space to do so. Switzerland will hold a referendum on November 27 regarding the phase-out of nuclear power plants. Austria had a two-stage election for president this year, with the second controversial round being annulled and rescheduled for December 4. And the Gaza Strip could still have its on-again, off-again municipal elections this year.

Despite this, I think the 2016 US presidential election deserves some specific analysis. A full breakdown of my predictions and their accuracy will occur at the end of the year (we are, after all, still awaiting the results of Michigan, pending a possible recount). This will include a discussion of an error I should, in the interest of not hiding my mistakes, point out at the earliest opportunity (i.e. now). Maine and Nebraska were stated to split their electoral college votes proportionally during my predictions. This is not the case; both states deliver two electoral college seats (representing their senators) on a winner-takes-all basis, and the rest by first-past-the-post in their congressional districts (representing their members of the house).

What I will be looking at here, instead, will be the large scale result that surprised me as much as the other pollsters relying on US polling data: the victory of Donald Trump. In the minds of some, this is a result of racist, sexist and generally hateful voters. For others this is a rejection of politics as usual and a desire for change. For yet more, this is a far stronger emotion--revenge upon those who have made America not-great. I'll be comparing these three proposed rationalisations against the results to see which fit the data.

RACISM

There are many metrics that can be used to estimate which US states are the most racist. The most popular have been by analysis of twitter posts and google searches (the latter having first been used in psephological analysis of Obama's electoral prospects), however these are unreliable for certain reasons (including access and use of internet due to varying demographics between states and the fact that many terms flagged as offensive have been reclaimed by minorities and used in a positive context). Instead I'll be using the FBI Hate Crimes Statistics 2015, which while also flawed (such as relying on crimes being reported) is somewhat more rigorous.

STATE	POPULATION			AGENCIES			POPULATION REPORTED %	RACIAL HATE CRIMES
	TOTAL	EXAMINED	% EXAMINED	TOTAL	REPORTING	% REPORTING		REPORTED	CALCULATED	PER MILLION
Alabama	4,858,979	1,252,146	25.8%	34	3	8.8%	2.3%	8	352	72.41
Alaska	738,432	734,820	99.5%	33	4	12.1%	12.1%	7	58	78.59
Arizona	6,828,065	6,622,880	97.0%	101	21	20.8%	20.2%	162	803	117.64
Arkansas	2,978,204	2,754,543	92.5%	279	4	1.4%	1.3%	3	226	75.97
California	39,144,818	39,137,326	100.0%	730	213	29.2%	29.2%	427	1464	37.39
Colorado	5,456,574	5,445,853	99.8%	234	42	17.9%	17.9%	65	363	66.50
Connecticut	3,590,886	3,399,068	94.7%	95	44	46.3%	43.8%	62	141	39.38
Delaware	945,934	945,934	100.0%	60	7	11.7%	11.7%	9	77	81.55
D. C.	672,228	672,228	100.0%	2	2	100.0%	100.0%	23	23	34.21
Florida	20,271,272	5,356,877	26.4%	38	36	94.7%	25.0%	44	176	8.67
Georgia	10,214,860	7,991,234	78.2%	473	7	1.5%	1.2%	32	2764	270.58
Hawaii	1,431,603	0	0.0%	0	0	0.0%	0.0%	0	N/A	N/A
Idaho	1,654,930	1,654,475	100.0%	112	19	17.0%	17.0%	14	83	49.88
Illinois	12,859,995	12,501,008	97.2%	741	43	5.8%	5.6%	59	1046	81.33
Indiana	6,619,680	3,224,755	48.7%	168	18	10.7%	5.2%	43	824	124.45
Iowa	3,123,899	3,105,094	99.4%	237	5	2.1%	2.1%	3	143	45.80
Kansas	2,911,641	2,741,323	94.2%	345	34	9.9%	9.3%	46	496	170.27
Kentucky	4,425,092	4,402,368	99.5%	403	83	20.6%	20.5%	113	551	124.63
Louisiana	4,670,724	3,711,824	79.5%	148	15	10.1%	8.1%	22	273	58.48
Maine	1,329,328	1,329,328	100.0%	184	14	7.6%	7.6%	16	210	158.19
Maryland	6,006,401	6,006,401	100.0%	154	11	7.1%	7.1%	22	308	51.28
Massachusetts	6,794,422	6,566,279	96.6%	342	85	24.9%	24.0%	198	824	121.33
Michigan	9,922,576	9,834,270	99.1%	617	127	20.6%	20.4%	198	971	97.81
Minnesota	5,489,594	5,218,435	95.1%	319	27	8.5%	8.0%	58	721	131.32
Mississippi	2,992,333	763,830	25.5%	43	0	0.0%	0.0%	0	N/A	N/A
Missouri	6,083,672	6,079,483	99.9%	628	28	4.5%	4.5%	70	1571	258.25
Montana	1,032,949	1,023,807	99.1%	101	13	12.9%	12.8%	28	219	212.48
Nebraska	1,896,190	1,821,196	96.0%	227	3	1.3%	1.3%	10	788	415.48
Nevada	2,890,845	2,890,845	100.0%	53	6	11.3%	11.3%	36	318	110.00
New Hamp.	1,330,608	1,267,715	95.3%	168	9	5.4%	5.1%	8	157	117.80
New Jersey	8,958,013	8,956,395	100.0%	508	123	24.2%	24.2%	169	698	77.93
New Mexico	2,085,109	574,972	27.6%	18	2	11.1%	3.1%	8	261	125.22
New York	19,795,791	19,766,342	99.9%	575	60	10.4%	10.4%	136	1305	65.94
North Carolina	10,042,802	10,041,690	100.0%	532	52	9.8%	9.8%	106	1085	108.00
North Dakota	756,927	756,927	100.0%	112	19	17.0%	17.0%	29	171	225.84
Ohio	11,613,423	9,781,677	84.2%	595	109	18.3%	15.4%	309	2003	172.44
Oklahoma	3,911,338	3,896,985	99.6%	351	29	8.3%	8.2%	27	328	83.86
Oregon	4,028,977	1,671,416	41.5%	130	16	12.3%	5.1%	41	803	199.31
Pennsylvania	12,802,503	12,550,581	98.0%	1,436	26	1.8%	1.8%	40	2254	176.03
Rhode Island	1,056,298	1,056,298	100.0%	49	8	16.3%	16.3%	6	37	34.79
South Carolina	4,896,146	4,826,241	98.6%	436	40	9.2%	9.0%	44	487	99.37
South Dakota	858,469	782,152	91.1%	121	9	7.4%	6.8%	9	133	154.70
Tennessee	6,600,299	6,600,299	100.0%	463	61	13.2%	13.2%	157	1192	180.55
Texas	27,469,114	27,390,337	99.7%	1,026	62	6.0%	6.0%	107	1776	64.65
Utah	2,995,919	2,966,781	99.0%	133	22	16.5%	16.4%	29	177	59.09
Vermont	626,042	626,042	100.0%	90	5	5.6%	5.6%	5	90	143.76
Virginia	8,382,993	8,380,278	100.0%	414	58	14.0%	14.0%	108	771	91.99
Washington	7,170,351	7,163,444	99.9%	256	79	30.9%	30.8%	160	519	72.38
West Virginia	1,844,128	1,494,503	81.0%	229	12	5.2%	4.2%	28	659	357.53
Wisconsin	5,771,337	5,580,752	96.7%	395	25	6.3%	6.1%	27	441	76.44
Wyoming	586,107	564,577	96.3%	57	1	1.8%	1.7%	2	118	201.92

With the exception of the 2015 state populations, derived from the US Census Bureau's population estimates (XLSX, CSV), all data is taken from the FBI report (specifically tables 12 and 13) or calculated therefrom:

• Population: % examined ('Population: examined'/'Population: total' x 100%)
• the proportion of the population covered by the FBI's research 
• Agencies: % reporting ('Agencies: reporting'/'Agencies: total' x 100%)
• the proportion of affiliated agencies (which cover the population examined) providing data
• Population reported % ('Population: % examined' x 'Agencies: % reporting')
• the estimated proportion of the population for which data is available
• does not account for variation in agency size
• Racial hate crimes: calculated ('Racial hate crimes: reported' x 100%/'population reported %')
• number of hate crimes in each state, extrapolated from the estimated proportion of the population for which data is available
• does not account for variation in reporting practices
• Racial hate crimes: per million ('Racial hate crimes: calculated'/'Population total' x 1,000,000)
• number of hate crimes calculated to have occurred, per million people residing in the state

On these numbers, the most racist states are Nebraska, West Virginia, Georgia, Missouri and North Dakota. Data is not fully available for Hawaii and Mississippi. This data differs from the previously linked data, although West Virginia is also over-represented in anti-black tweets and racist google searches.

STATE	HATE CRIMES PER MILLION	TRUMP SUPPORT
Alabama	72.41	62.9%
Alaska	78.59	52.9%
Arizona	117.64	49.5%
Arkansas	75.97	60.4%
California	37.39	33.2%
Colorado	66.50	44.4%
Connecticut	39.38	41.7%
Delaware	81.55	41.9%
District of Columbia	34.21	4.1%
Florida	8.67	49.1%
Georgia	270.58	51.3%
Hawaii	N/A	30.1%
Idaho	49.88	59.2%
Illinois	81.33	39.4%
Indiana	124.45	57.2%
Iowa	45.80	51.8%
Kansas	170.27	57.2%
Kentucky	124.63	62.5%
Louisiana	58.48	58.1%
Maine	158.19	45.2%
Maryland	51.28	60.5%
Massachusetts	121.33	33.5%
Michigan	97.81	47.6%
Minnesota	131.32	45.4%
Mississippi	N/A	39.7%
Missouri	258.25	57.1%
Montana	212.48	56.5%
Nebraska	415.48	60.3%
Nevada	110.00	45.5%
New Hampshire	117.80	47.3%
New Jersey	77.93	41.8%
New Mexico	125.22	40.0%
New York	65.94	37.5%
North Carolina	108.00	50.5%
North Dakota	225.84	64.1%
Ohio	172.44	52.1%
Oklahoma	83.86	65.3%
Oregon	199.31	41.1%
Pennsylvania	176.03	48.8%
Rhode Island	34.79	39.8%
South Carolina	99.37	54.9%
South Dakota	154.70	61.5%
Tennessee	180.55	61.1%
Texas	64.65	52.6%
Utah	59.09	46.6%
Vermont	143.76	32.6%
Virginia	91.99	45.0%
Washington	72.38	38.3%
West Virginia	357.53	68.7%
Wisconsin	76.44	47.9%
Wyoming	201.92	70.1%
Pearson's Correlation	0.41

While there is a low correlation between primary votes for Trump ~(+.41) there are some obvious outliers; in particular the least racist state by our metrics is Florida (which many may disagree with) yet this was famously won by Trump (though by less than 50% of the popular vote).

SEXISM

Sexism has largely been studies by the same means as racism in the past, particularly through twitter posts. Since the FBI also records hate crimes motivated by gender, it is possible to perform the same analysis as above. However, there is such limited data that this does not provide much useful data:

STATE	POPULATION			AGENCIES			POPULATION REPORTED	RACIAL HATE CRIMES
	TOTAL	EXAMINED	% EXAMINED	TOTAL	REPORTING	% REPORTING		REPORTED	CALCULATED	PER MILLION
Alabama	4,858,979	1,252,146	25.8%	34	3	8.8%	2.3%	0	0	0.00
Alaska	738,432	734,820	99.5%	33	4	12.1%	12.1%	0	0	0.00
Arizona	6,828,065	6,622,880	97.0%	101	21	20.8%	20.2%	0	0	0.00
Arkansas	2,978,204	2,754,543	92.5%	279	4	1.4%	1.3%	0	0	0.00
California	39,144,818	39,137,326	100.0%	730	213	29.2%	29.2%	1	3	0.09
Colorado	5,456,574	5,445,853	99.8%	234	42	17.9%	17.9%	0	0	0.00
Connecticut	3,590,886	3,399,068	94.7%	95	44	46.3%	43.8%	0	0	0.00
Delaware	945,934	945,934	100.0%	60	7	11.7%	11.7%	0	0	0.00
D.C.	672,228	672,228	100.0%	2	2	100.0%	100.0%	0	0	0.00
Florida	20,271,272	5,356,877	26.4%	38	36	94.7%	25.0%	0	0	0.00
Georgia	10,214,860	7,991,234	78.2%	473	7	1.5%	1.2%	0	0	0.00
Hawaii	1,431,603	0	0.0%	0	0	0.0%	0.0%	0	N/A	N/A
Idaho	1,654,930	1,654,475	100.0%	112	19	17.0%	17.0%	0	0	0.00
Illinois	12,859,995	12,501,008	97.2%	741	43	5.8%	5.6%	1	18	1.38
Indiana	6,619,680	3,224,755	48.7%	168	18	10.7%	5.2%	1	19	2.89
Iowa	3,123,899	3,105,094	99.4%	237	5	2.1%	2.1%	0	0	0.00
Kansas	2,911,641	2,741,323	94.2%	345	34	9.9%	9.3%	1	11	3.70
Kentucky	4,425,092	4,402,368	99.5%	403	83	20.6%	20.5%	0	0	0.00
Louisiana	4,670,724	3,711,824	79.5%	148	15	10.1%	8.1%	0	0	0.00
Maine	1,329,328	1,329,328	100.0%	184	14	7.6%	7.6%	0	0	0.00
Maryland	6,006,401	6,006,401	100.0%	154	11	7.1%	7.1%	0	0	0.00
Massachusetts	6,794,422	6,566,279	96.6%	342	85	24.9%	24.0%	15	62	9.19
Michigan	9,922,576	9,834,270	99.1%	617	127	20.6%	20.4%	1	5	0.49
Minnesota	5,489,594	5,218,435	95.1%	319	27	8.5%	8.0%	0	0	0.00
Mississippi	2,992,333	763,830	25.5%	43	0	0.0%	0.0%	0	N/A	N/A
Missouri	6,083,672	6,079,483	99.9%	628	28	4.5%	4.5%	0	0	0.00
Montana	1,032,949	1,023,807	99.1%	101	13	12.9%	12.8%	0	0	0.00
Nebraska	1,896,190	1,821,196	96.0%	227	3	1.3%	1.3%	0	0	0.00
Nevada	2,890,845	2,890,845	100.0%	53	6	11.3%	11.3%	0	0	0.00
New Hamp.	1,330,608	1,267,715	95.3%	168	9	5.4%	5.1%	0	0	0.00
New Jersey	8,958,013	8,956,395	100.0%	508	123	24.2%	24.2%	0	0	0.00
New Mexico	2,085,109	574,972	27.6%	18	2	11.1%	3.1%	0	0	0.00
New York	19,795,791	19,766,342	99.9%	575	60	10.4%	10.4%	0	0	0.00
North Carolina	10,042,802	10,041,690	100.0%	532	52	9.8%	9.8%	0	0	0.00
North Dakota	756,927	756,927	100.0%	112	19	17.0%	17.0%	0	0	0.00
Ohio	11,613,423	9,781,677	84.2%	595	109	18.3%	15.4%	0	0	0.00
Oklahoma	3,911,338	3,896,985	99.6%	351	29	8.3%	8.2%	0	0	0.00
Oregon	4,028,977	1,671,416	41.5%	130	16	12.3%	5.1%	0	0	0.00
Pennsylvania	12,802,503	12,550,581	98.0%	1,436	26	1.8%	1.8%	0	0	0.00
Rhode Island	1,056,298	1,056,298	100.0%	49	8	16.3%	16.3%	0	0	0.00
South Carolina	4,896,146	4,826,241	98.6%	436	40	9.2%	9.0%	0	0	0.00
South Dakota	858,469	782,152	91.1%	121	9	7.4%	6.8%	0	0	0.00
Tennessee	6,600,299	6,600,299	100.0%	463	61	13.2%	13.2%	1	8	1.15
Texas	27,469,114	27,390,337	99.7%	1,026	62	6.0%	6.0%	1	17	0.60
Utah	2,995,919	2,966,781	99.0%	133	22	16.5%	16.4%	0	0	0.00
Vermont	626,042	626,042	100.0%	90	5	5.6%	5.6%	0	0	0.00
Virginia	8,382,993	8,380,278	100.0%	414	58	14.0%	14.0%	0	0	0.00
Washington	7,170,351	7,163,444	99.9%	256	79	30.9%	30.8%	3	10	1.36
West Virginia	1,844,128	1,494,503	81.0%	229	12	5.2%	4.2%	0	0	0.00
Wisconsin	5,771,337	5,580,752	96.7%	395	25	6.3%	6.1%	1	16	2.83
Wyoming	586,107	564,577	96.3%	57	1	1.8%	1.7%	0	0	0.00

For the curious, there is a negligible negative correlation between this limited data and support for Trump (~-0.15).

Instead, I'll be using the rankings produced by WalletHub, which has combined a number of metrics more effectively than I could. For the large number of people that do not understand that the actual enemy of feminism is not men but the enforcement of gender roles, and that metrics that represent male disadvantage (e.g. incarceration rates and duration, length of working hours, child custody rates, life span etc.) are a direct result of these roles which, although disadvantaging men in these instances, have historically been enforced to limit female empowerment, allow me to point out that these metrics have been combined into this study even though the data is billed as regarding "women's equality". In other words, even though the majority of Trump's rhetoric relevant to the claim that Trump's support was supported by sexism has been directed against women, the inclusion of data highlighting examples of male disadvantage is relevant as it indicates more traditional, gender-role enforcing attitudes in the relevant states.

Interestingly, the correlation is again negative, but low at ~0.30:

STATE	WALLETHUB SCORE	TRUMP SUPPORT
Alabama	50.61	62.9%
Alaska	72.46	52.9%
Arizona	57.31	49.5%
Arkansas	51.57	60.4%
California	69.1	33.2%
Colorado	57.37	44.4%
Connecticut	51.46	41.7%
Delaware	54.22	41.9%
District of Columbia	N/A	4.1%
Florida	55.73	49.1%
Georgia	44.46	51.3%
Hawaii	81.67	30.1%
Idaho	50.29	59.2%
Illinois	58.59	39.4%
Indiana	56.35	57.2%
Iowa	58.37	51.8%
Kansas	53.92	57.2%
Kentucky	47.49	62.5%
Louisiana	45.45	58.1%
Maine	69.49	45.2%
Maryland	65.09	60.5%
Massachusetts	58.65	33.5%
Michigan	55.5	47.6%
Minnesota	66.1	45.4%
Mississippi	46.89	39.7%
Missouri	59.22	57.1%
Montana	51.58	56.5%
Nebraska	57.86	60.3%
Nevada	58.05	45.5%
New Hampshire	66.4	47.3%
New Jersey	44.97	41.8%
New Mexico	58.33	40.0%
New York	63.32	37.5%
North Carolina	50.42	50.5%
North Dakota	61.37	64.1%
Ohio	48.83	52.1%
Oklahoma	48.3	65.3%
Oregon	60.18	41.1%
Pennsylvania	45.15	48.8%
Rhode Island	49.08	39.8%
South Carolina	45.8	54.9%
South Dakota	57.86	61.5%
Tennessee	52.8	61.1%
Texas	49.77	52.6%
Utah	33.7	46.6%
Vermont	68.16	32.6%
Virginia	48.67	45.0%
Washington	61.93	38.3%
West Virginia	59.72	68.7%
Wisconsin	62.75	47.9%
Wyoming	57.11	70.1%
Pearson's Correlation	-0.30

POLITICAL DISCONTENT

Measuring voter discontent is difficult in the current election. Normally these votes emerge as protest votes or failure to vote. A higher 3rd party vote, for example, may indicate dissatisfaction with either candidate but (ignoring that dissatisfaction in this election was high in part due to Trump's candidacy) there is little to be learned by comparing Trump support with 3rd party results. Does a positive correlation indicate Trump is popular where voters are discontent? Or does a negative correlation indicate that he is absorbing the discontent vote?

Two proxy measures of discontent present themselves: 3rd party voting and non-voting in 2012 (inherently assuming this discontent is not the result of the last 4 years and was present in that data) and 3rd party voting in each state's other federal races this November. Both are problematic, of course. The former assumes little or no change in almost half a decade. The latter ignores the possible swell of Trump voters who turned out to vote (and who normally would not) due to the fervent support the candidate produced but didn't care about other races and therefore deliberately "wasted" their vote. I have gone with the former for a number of reasons, including the nature of Senate elections that mean not all states have a candidate for us to consider, and the frequent absence of any third party candidate in many HoR seats.

STATE	2012 TURNOUT	TRUMP SUPPORT		STATE	3RD PARTY	TRUMP SUPPORT
Alabama	58.60%	62.9%		Alabama	1.09%	62.9%
Alaska	58.70%	52.9%		Alaska	4.39%	52.9%
Arizona	52.60%	49.5%		Arizona	1.76%	49.5%
Arkansas	50.70%	60.4%		Arkansas	2.55%	60.4%
California	55.10%	33.2%		California	2.64%	33.2%
Colorado	69.90%	44.4%		Colorado	2.38%	44.4%
Connecticut	61.30%	41.7%		Connecticut	1.21%	41.7%
Delaware	62.30%	41.9%		Delaware	1.41%	41.9%
District of Columbia	61.50%	4.1%		District of Columbia	1.81%	4.1%
Florida	62.80%	49.1%		Florida	0.86%	49.1%
Georgia	59.00%	51.3%		Georgia	1.22%	51.3%
Hawaii	44.20%	30.1%		Hawaii	1.61%	30.1%
Idaho	59.80%	59.2%		Idaho	2.85%	59.2%
Illinois	58.90%	39.4%		Illinois	1.67%	39.4%
Indiana	55.20%	57.2%		Indiana	1.94%	57.2%
Iowa	70.30%	51.8%		Iowa	1.83%	51.8%
Kansas	56.90%	57.2%		Kansas	2.30%	57.2%
Kentucky	55.70%	62.5%		Kentucky	1.71%	62.5%
Louisiana	60.20%	58.1%		Louisiana	1.64%	58.1%
Maine	68.20%	45.2%		Maine	2.75%	45.2%
Maryland	66.60%	60.5%		Maryland	2.13%	60.5%
Massachusetts	65.90%	33.5%		Massachusetts	1.84%	33.5%
Michigan	64.70%	47.6%		Michigan	1.08%	47.6%
Minnesota	76.00%	45.4%		Minnesota	2.39%	45.4%
Mississippi	59.30%	39.7%		Mississippi	0.92%	39.7%
Missouri	62.20%	57.1%		Missouri	1.86%	57.1%
Montana	62.50%	56.5%		Montana	2.95%	56.5%
Nebraska	60.30%	60.3%		Nebraska	2.17%	60.3%
Nevada	56.40%	45.5%		Nevada	1.96%	45.5%
New Hampshire	70.20%	47.3%		New Hampshire	1.62%	47.3%
New Jersey	61.50%	41.8%		New Jersey	1.03%	41.8%
New Mexico	54.60%	40.0%		New Mexico	4.17%	40.0%
New York	53.10%	37.5%		New York	1.48%	37.5%
North Carolina	64.80%	50.5%		North Carolina	1.26%	50.5%
North Dakota	59.80%	64.1%		North Dakota	2.99%	64.1%
Ohio	64.50%	52.1%		Ohio	1.64%	52.1%
Oklahoma	49.20%	65.3%		Oklahoma	0.00%	65.3%
Oregon	63.10%	41.1%		Oregon	3.61%	41.1%
Pennsylvania	59.50%	48.8%		Pennsylvania	1.44%	48.8%
Rhode Island	58.00%	39.8%		Rhode Island	2.06%	39.8%
South Carolina	56.30%	54.9%		South Carolina	1.35%	54.9%
South Dakota	59.30%	61.5%		South Dakota	2.24%	61.5%
Tennessee	51.90%	61.1%		Tennessee	1.44%	61.1%
Texas	49.60%	52.6%		Texas	1.45%	52.6%
Utah	55.50%	46.6%		Utah	2.46%	46.6%
Vermont	60.70%	32.6%		Vermont	2.46%	32.6%
Virginia	66.10%	45.0%		Virginia	1.56%	45.0%
Washington	64.80%	38.3%		Washington	2.55%	38.3%
West Virginia	46.30%	68.7%		West Virginia	2.16%	68.7%
Wisconsin	72.90%	47.9%		Wisconsin	1.28%	47.9%
Wyoming	58.60%	70.1%		Wyoming	3.54%	70.1%
Pearson's Correlation	-0.17			Pearson's Correlation	0.03

Both non-voting and 3rd-party voting in 2012 show little correlation to support for Trump in 2016, though both nominally in the direction that would suggest Trump served as a vessel for protest votes: a slight negative correlation with 2012 voter turnout and slight positive correlation with 2012 3rd party results. For voter turnout, I'm using the eligible voter highest office data from the United States Electoral Project.

REVENGE OF ANTIGLOBALISTS

This last rationalisation is the one that is often cited in the media as the rise of "angry white men". This rationalisation, though largely ignored in the earlier campaign, is now the dominant focus of media attention. This rationalisation is often portrayed as white, middle-aged men--the pale, stale males--enraged at the rule of a Black President, furious at the prospect of a female one and livid at the shift of American culture towards one of inclusiveness and multiculturalism. However, there's another side to this view. One in which, despite their privilege, these men have faced genuine hardships. One in which their anger is not directed in violence against the non-white, non-male population which we considered earlier. This is a view that, in my opinion, was best explained by documentary filmmaker Michael Moore in clarity I could not hope to equal.

In this view, all of this anger (if we accept that it is distinct from the racism and sexism analysed above) comes from financial hardship and the social effects it has had. Despite the disparity in real living conditions between White and Non-White on average, White men are angrier than Black and feel as though they were made a promise that was then ripped away from them. What matters isn't necessarily that these individuals are doing it tough, but that they feel they are doing it tough, or tougher than they are owed. For a long time America has persisted on the lie that if you work hard and deserve success you can achieve it. The dark corollary to this assertion, unchallenged and patriotically insisted upon, is that if you don't have success it's because you don't deserve it. And when that lack of success falls on minorities, that rationalisation follows easily. It also deprived any impetus for support to the underprivileged. even with minimum welfare, no socialised healthcare and poverty conditions that rival those found in third-world countries, it was easy to feel no obligation for government support. Now that poverty has come for the angry white men too; the ones who dreamed of better things, and the unfairness is revealed. But enough leftist socialist communist bleeding-heart liberal wishy-washy nonsense from me. How does Trump support correlate with job loss and redundancy from, among other things, competition in international manufacture?

Comparisons with state poverty rates and unemployment were negligible. The strongest correlation so far, at ~-0.57, demonstrates a moderate negative correlation between per capita income and support for Trump.

STATE	TRUMP SUPPORT	POVERTY LEVEL	UNEMPLOYMENT	INCOME PER CAPITA
Alabama	62.9%	18.5%	5.4%	$23,606
Alaska	52.9%	10.3%	6.9%	$33,062
Arizona	49.5%	17.4%	5.5%	$25,715
Arkansas	60.4%	19.1%	4.0%	$22,883
California	33.2%	15.3%	5.5%	$30,441
Colorado	44.4%	11.5%	3.6%	$32,357
Connecticut	41.7%	10.5%	5.4%	$39,373
Delaware	41.9%	12.4%	4.3%	$30,488
District of Columbia	4.1%	17.3%	6.1%	$45,877
Florida	49.1%	15.7%	4.7%	$26,582
Georgia	51.3%	17.0%	5.1%	$25,615
Hawaii	30.1%	10.6%	3.3%	$29,736
Idaho	59.2%	15.1%	3.8%	$23,938
Illinois	39.4%	13.6%	5.5%	$30,417
Indiana	57.2%	14.5%	4.5%	$25,140
Iowa	51.8%	12.2%	4.2%	$28,361
Kansas	57.2%	13.0%	4.4%	$27,870
Kentucky	62.5%	18.5%	5.0%	$23,684
Louisiana	58.1%	19.6%	6.4%	$24,800
Maine	45.2%	13.4%	4.1%	$27,978
Maryland	60.5%	9.7%	4.2%	$36,338
Massachusetts	33.5%	11.5%	3.6%	$36,593
Michigan	47.6%	15.8%	4.6%	$26,613
Minnesota	45.4%	10.2%	4.0%	$32,638
Mississippi	39.7%	22.0%	6.0%	$21,036
Missouri	57.1%	14.8%	5.2%	$26,126
Montana	56.5%	14.6%	4.3%	$25,989
Nebraska	60.3%	12.6%	3.2%	$27,446
Nevada	45.5%	14.7%	5.8%	$25,773
New Hampshire	47.3%	8.2%	2.9%	$34,691
New Jersey	41.8%	10.8%	5.3%	$37,288
New Mexico	40.0%	20.4%	6.7%	$23,683
New York	37.5%	15.4%	5.0%	$33,095
North Carolina	50.5%	16.4%	4.7%	$25,774
North Dakota	64.1%	11.0%	3.0%	$33,071
Ohio	52.1%	14.8%	4.8%	$26,937
Oklahoma	65.3%	16.1%	5.2%	$25,229
Oregon	41.1%	15.4%	5.5%	$27,646
Pennsylvania	48.8%	13.2%	5.7%	$29,220
Rhode Island	39.8%	13.9%	5.6%	$30,830
South Carolina	54.9%	16.6%	4.9%	$24,596
South Dakota	61.5%	13.7%	2.9%	$26,959
Tennessee	61.1%	16.7%	4.6%	$24,922
Texas	52.6%	15.9%	4.8%	$27,125
Utah	46.6%	11.3%	3.4%	$24,877
Vermont	32.6%	10.2%	3.3%	$29,178
Virginia	45.0%	11.2%	4.0%	$34,052
Washington	38.3%	12.2%	5.6%	$31,841
West Virginia	68.7%	17.9%	5.8%	$22,714
Wisconsin	47.9%	12.1%	4.1%	$28,213
Wyoming	70.1%	11.1%	5.3%	$29,698
Pearson's Correlation		0.11	-0.13	-0.57

But all of these metrics are too blunt. Measures of general hardship in a state will generally include a large proportion of the Non-White population due to the historical inequality with which these groups were treated. Data from the US Census Bureau allows a state by state breakdown of poverty, unemployment and income for "white alone":

 

STATE	TRUMP SUPPORT	POVERTY LEVEL	UNEMPLOYMENT	INCOME PER CAPITA
Alabama	62.9%	13.7%	5.5%	$28,235
Alaska	52.9%	6.6%	5.9%	$41,133
Arizona	49.5%	15.2%	6.0%	$29,115
Arkansas	60.4%	15.5%	4.9%	$25,864
California	33.2%	14.1%	6.8%	$35,523
Colorado	44.4%	10.4%	4.8%	$35,676
Connecticut	41.7%	7.9%	5.6%	$44,134
Delaware	41.9%	9.6%	5.5%	$34,522
District of Columbia	4.1%	7.1%	3.1%	$81,474
Florida	49.1%	13.6%	6.0%	$30,582
Georgia	51.3%	12.5%	5.2%	$31,583
Hawaii	30.1%	10.1%	5.3%	$40,010
Idaho	59.2%	14.4%	5.1%	$24,935
Illinois	39.4%	10.1%	5.2%	$35,754
Indiana	57.2%	12.2%	5.0%	$27,907
Iowa	51.8%	10.7%	3.6%	$29,973
Kansas	57.2%	11.2%	4.2%	$30,533
Kentucky	62.5%	17.1%	6.1%	$25,942
Louisiana	58.1%	13.2%	5.5%	$30,792
Maine	45.2%	12.7%	5.1%	$29,112
Maryland	60.5%	7.1%	4.2%	$44,016
Massachusetts	33.5%	9.2%	5.2%	$41,556
Michigan	47.6%	12.3%	5.5%	$30,171
Minnesota	45.4%	7.8%	3.5%	$35,930
Mississippi	39.7%	13.6%	5.8%	$26,281
Missouri	57.1%	12.5%	4.6%	$29,155
Montana	56.5%	12.7%	3.9%	$29,352
Nebraska	60.3%	10.6%	2.8%	$30,708
Nevada	45.5%	12.8%	7.4%	$30,382
New Hampshire	47.3%	7.9%	4.1%	$36,589
New Jersey	41.8%	8.2%	5.6%	$41,743
New Mexico	40.0%	17.6%	6.4%	$26,722
New York	37.5%	11.3%	5.1%	$40,483
North Carolina	50.5%	12.7%	5.4%	$30,430
North Dakota	64.1%	9.0%	2.1%	$35,895
Ohio	52.1%	11.6%	5.2%	$30,077
Oklahoma	65.3%	13.2%	4.7%	$28,919
Oregon	41.1%	14.5%	6.5%	$30,269
Pennsylvania	48.8%	10.3%	5.2%	$32,865
Rhode Island	39.8%	10.9%	5.2%	$34,749
South Carolina	54.9%	12.0%	5.6%	$29,899
South Dakota	61.5%	9.6%	2.9%	$30,461
Tennessee	61.1%	14.5%	5.2%	$28,338
Texas	52.6%	14.8%	5.2%	$30,067
Utah	46.6%	10.0%	3.7%	$27,132
Vermont	32.6%	9.7%	3.7%	$31,747
Virginia	45.0%	9.1%	4.5%	$38,687
Washington	38.3%	11.0%	5.6%	$35,872
West Virginia	68.7%	17.4%	7.0%	$23,840
Wisconsin	47.9%	9.6%	3.5%	$31,605
Wyoming	70.1%	10.8%	4.6%	$32,735
Pearson's Correlation		0.33	-0.07	-0.66

Focusing solely on white poverty levels sees the correlation in the data more than double to a point where it is only eclipsed by the racism and per-capita income metrics. However, at 0.33, this is only slightly stronger than the negative correlation between sexism and Trump support at 0.30, and no one is suggesting that Trump won on the back of feminist support.

Using only the White unemployment rates, the correlation becomes even more negligible.

However, there is (by psephological standards) a high negative correlation between white median wages and support for Trump. And while it's true that this merely points out the existence of a certain stereotype--the working class, White Trump voter--understanding voter demographics is important to understanding why Trump got elected contrary to all conventional wisdom. This was a section of the community that turned out to vote, and felt the need to stand up for something.

CONCLUSION

While much of this data operates as a proxy for some other social issue, such as racism or anger at economic leadership, It is important to realise that there are several reasons for Trump's support. A candidate does not get elected by a single demographic. Not solely by a racist element, or by disillusioned working class voters, or conspiracy theorists who think Hillary Clinton orchestrated World War I from her Russian space-dreadnaught. Of the various reasons studied two comparatively strong correlations emerged.

For those who like to argue Trump supporters are not racist, it is worth noting that states prone to racial violence tend to be more pro-Trump. This should hardly need pointing out for a candidate endorsed by the Klan, but some not-insignificant part of Trump's supporter base comes from a xenophobia that has not been given a political vent as prominent or blatant as Trump before.

On the other hand, for those who, on the other hand, argue that all Trump supporters are hateful racists, it is noteworthy that being from a poor state, and particularly a state with a poor White population, is a better indicator of support for Trump than being from a state with more racially motivated Hate Crimes; some might, of course, argue that voting for Trump in and of itself should be a hate crime.

It is hard for some people to understand how anyone could support a candidate who has voiced the opinions Trump has and not inherently be a biggot. To these people I would suggest the voters need not agree with Trump's statements, only to rationalise them as 'speaking his mind', 'not what he meant', 'locker-room talk' and so forth. People who voted for Trump committed to one aspect or another of his personality or platform--as did Clinton supporters--and forgave the many flaws of their preferred candidate while exaggerating the other's.

There are, of course, many other factors--Trump's ability to seize the zeitgeist, Clinton's many issues with engaging voters, the vast free publicity news outlets offered Trump while underestimating his potential, Trump's ability to give vague and often contradictory promises worded so as to let each listener take what they liked and disregard the rest and so forth. But there are several reasons voters latched on to Trump, some good and some less so.

TL;DR: There is some evidence that Trump had support due to his racist statements. He also appealed to a class of disenfranchised voter with legitimate (though arguably misdirected) grievances against the current system to the extend that they'd happily burn the whole thing down. There would have been other reasons as well, but the portrayal of Trump supporters as rampant racists or as righteous rebels are both partially correct, and partially incorrect.

UPDATE!

Hope this works out for you, America.

The Demise of Donald J Trump

A bold prediction? I don't think so. The media has a bad habit of calling every election "too close to call" in the final days, regardless of whether or not that's technically true. To wring any hint of a nail-biter out of this contest one has to look at the polling data so superficially it takes genuine effort.

But before I get to this analysis, a few quick observations: despite being November the weather is looking not at all terrible in most of the US. In a non-compulsory voting nation, this means more of the less-enthused voters will turn out to vote (or more accurately, rain will cause these voters to not vote). Based on pure energy levels, I think the unenthused voter demographic is broadly not Trump's demographic. The main areas of rain seem to focus on the states around Texas, and those around Washington State, reasonably safe seats for the Republicans and Democrats respectively where weather will have a less drastic impact than a more marginal state.

The other thing I'd like to point out is that whoever gets elected will be unpopular. This means the usual mid-term trend of votes away from the President's party is likely to be particularly pronounced, and either candidate may find themselves unable to achieve very much at all after their first two years.

And now...

The Unpredictable Candidate

Most conventional wisdom in this campaign has gone out the window. On conventional wisdom, Trump should have lost right at the start of the primaries. Since that point there have been 100 more that should have ended his campaign. Trump defies most understanding about American politics. As such, traditional methods of prediction are likely to prove unreliable. Trump has driven more than a few Republicans to the Democratic side this election, and enthused demographics that have not traditionally had the faith, patience or trust in politics to vote before. Historical data about state voting patterns has been used in this post, but only in a minor way when no other solution presented itself.

For the most part this post attempts to predict the results of this irregular election without recourse to historical data. Instead, we're entirely using this year's polling data. A brief glance at the national polls gives a result of 46.6 for Clinton and 44.8 for Trump: a mere 1.8 percentage point gap between the candidates. But to call this election close ignores the realities of the Electoral College.

Each state has two senators and at least one member of the House, making a minimum of three representatives. (Washington DC has no senators and thus only has one elected representative, but for the purposes of the electoral college counts as having three). These equate to seats in the electoral college with several states having the bare minimum three, while more populous states like California sail well into the double digits. These seats are filled by votes during the election and, in turn, elect the president. Every state except Nebraska and Main operate on a winner-takes-all system. Therefore the national polling is irrelevant. A candidate can win election with just 270 seats in the college, or the 11 largest states combined. We need a state-by-state approach.

Here is a table of all the state polls found at this source:



Unfortunately there is no polling data from Mississippi, Hawaii and Washington D.C., so these states (and federal district) are dealt with at the end.

Now, to calculate each state's natural lean and project this on the national polling we need to work out how each poll diverges from the national norm. For consistency of data, I've averaged the state polls on each day rather than the national polls which may not align with the state polling dates. This does raise some issues where a lot of polling in red or blue states in a cluster might skew the average, however.

To smooth out outliers and deal with days that have few polls conducted, I've also produced a 5-day average that averages the polls of each day, the two proceeding days and the two following days. (Note that this is different to the average of the relevant daily averages as some days have more polls than others).

This data allows us to measure the degree to which each state poll diverges from an average that (hopefully) follows the rise and fall of candidate popularity throughout the campaign.

Next we need to refine this bulk of data into an assessment of each state's bias. The two measures I'll be using are each state's most recent divergence, and each state's average divergence:

 These divergences (which in truth are measured in percentage points, not percentages) can then be applied to national polling data (currently 51% once normalised so that Trump + Clinton = 100%) to emulate the state's natural lean and allow us to count the electoral college votes (EC) that each candidate can count on.

(Strictly speaking, Maine and Nebraska will divide their EC votes between the candidates based on their respective results on election day. However, I'd expect Maine to divide 2/2 and Nebraska 3/2 in favor of Trump on these numbers. This comes to Trump 5, Clinton 4 which is the same as just giving all of Maine to Clinton and Nebraska to Trump)

All I have done is award the states that a candidate wins on all four measures to them. Yet already Clinton has 266 of the required 270. Since the smallest state has 3 votes, Trump draws on the loss of any undecided states at this point, and Clinton can win outright with any one of these states except North Dakota or Alaska. Realistically Clinton will get one of these and win.

Increased Confidence

Now, this data all but gives Clinton the presidency off the bat, but it'd be foolhardy to claim a win for Clinton on this untested method alone. Interestingly, though, ALL of these predictions agree with the latest state polling except New Hampshire, worth a mere four votes. Everything I look at makes me view these allocated seats as the "safe bets", and however you cut the cake Clinton is over 260/270. Trump, by comparison, can't break 180 on these declared seats and needs to pretty well sweep every tossup to win.

Let's test his chances in each of the seats in doubt:

Alaska (3): With 3 votes, Alaska isn't crucial for Trump. And by not crucial, I mean he can still draw if he loses it, or if he somehow wins New Hampshire.
Historically Alaska should be a republican win, but Trump has alienated enough Alaskans, it seems, for Clinton to be ahead in the latest polls. Still, he wins 2/3 of the non-even predictions I've made, so I'll tentatively give him this state. [REP]


District of Columbia (3): Again, 3 votes is something Trump can draw without, but he can't really afford a draw. If he loses DC (and he will, as you will see) he needs to win New Hampshire to have the slightest chance at winning.

Why am i certain Clinton wins DC when we have no district poling data? History. Yes, I know this is an atypical election, but LOOK AT THOSE NUMBERS! There are world dictators with worse reported support than that! The Democrats' worst result in recent history is 85.2% to the Republicans' high of 9.3%.

Assuming my call on New Hampshire stands up, DC puts Clinton at a draw before any other state on this list is called. [DEM]


Florida (29): At 29 votes Florida is an auto-win for Clinton if she gets it. Florida is always an important seat, due to having the equal-third most EC votes and having a mixed voting record (unlike #1 California and equal #3 New York, which are reliably Democrat, and #2 Texas which is firmly Republican). This is the kind of seat, therefore, where history is of no use to us. Instead, latest polling has Clinton ahead in Florida, and she wins 3/4 of our metrics. I'm calling Florida democrat this year, and Clinton president. [DEM]

Georgia (16): 16 votes. Auto-win for Clinton. The flipside of Florida, Georgia is 3/4 pro-Trump in my calculations and Trump leads the latest poll. Unlike Florida, Georgia's history is more solid -- pro-Republican. Georgia goes to Trump in my view. [REP]

Hawaii (4): If my predictions are correct and Clinton wins Hawaii as well, she reaches 270 and claims the presidency. Trump can stave this off by winning New Hampshire and every other state on this list. Otherwise Hawaii is a must-win.

There's no state data to fall back on, but Hawaii has a strong Democrat history, and the lowest white population in the US which is where Trump gets most of his support. Hawaii for Clinton, putting all Trump's hopes on New Hampshire, but realistically, that won's save him. [DEM]

Mississippi (6): If Trump loses Mississippi's 6 votes he needs to win New Hampshire and every other state on this list. If he can pull that off, he wins by a single vote.


Without any state data, we're forced to rely on history. Still, Mississippi is a reliable Republican seat. I don't have too many doubts about handing it to Trump at this point. [REP]

New Hampshire (4): We've assumed Clinton wins this one based on our data, but this is the only state where our system disagrees with the latest polling. I'm sticking to my guns, partly because I need to pick a side and want to back my method, but partly because that latest poll looks like it may be an aberration. If Trump can win this seat, however, it gives him breathing space to lose Alaska, Hawaii or North Dakota as well as his inevitable loss in DC. [DEM]

North Carolina (15): at 15 EC votes, Clinton wins if she carries North Carolina.


Trump claims 2/3 of our calculations above. Clinton is ahead in the latest polling, but NC is historically GOP. Under the old system where I'd get 5% of the predictions as tossups, this'd be one of them (rounding up to 3 tossups I'd also claim New Hampshire and Alaska). But since we don't keep tally any more I should at least take a punt. Eeny Meeny Miny Trump. [REP]


North Dakota (3): North Dakota is small. If Trump loses this and doesn't win New Hampshire, it's a tie at best. If he loses NH or any other state on this list, he's done for. History and the latest polls suggest I should break this tied data in Trump's favour, and I have no better indicator. [REP]

Ohio (18): Ohio's 18 votes alone will give Clinton the presidency if she wins it. 50:50 on out calculations, and an equally mixed history make this an edge seat. Clinton leads the latest polls, and that's all I have to go on. [DEM]

That leaves us with this as a prediction:

TL; DR: Clinton wins with 320 votes, 50 more than required, and thus a victory with 59.5% of the EC vote. So not a landslide, but for Trump to win this requires one of a handful of very specific and unlikely situations. DC is a guaranteed loss, meaning he has to win everything else on the tossup list (Alaska, DC, Florida, Georgia, Hawaii, Mississippi, New Hampshire, North Carolina, North Dakota and Ohio) to win.

Making a Comeback

Mea Culpa (Iterum)

So a lot has happened since my last post but, as usual, I have not found the time to do anything preemptive for the various electoral shenanigans at home and abroad. I missed both of the territorial elections, which is disappointing because they get rather limited coverage as it is and there is a lot to talk about--particularly in the ACT which has several nice modifications to their electoral system.

Then, further abroad, there has been the usual endless train of elections--Palau, for example, elected it's President and both houses of parliament yesterday. Sure, most of these I wouldn't have commented on except in the presence of abundant leisure time anyhow, but there were some cool things that I might have gone out of my way to examine. Last month had a lot of multi-stage elections (Czech Republic, Georgia and Lithuania). Azerbaijan had a constitutional referendum in September, and I firmly believe Azerbaijan is the most enjoyable nation to say aloud. And, most realistically in terms of me actually covering something, Iceland recently elected its independence party to minority lead in the Althing (which is also fun to say) with 21 of the 63 seats, more than its two next closes rivals combined (Left-Greens and Pirates with 10 seats each). This may possibly pave the way towards an Icelandic referendum on leaving the EU which (in the tradition of the Grexit and Brexit) will probably be called Icexit because that's the nomenclaturial hell we appear to live in.

Also, when this guy has six seats in your parliament you know you had a good election.

So I missed a lot of history. Again. But even if I did neglect two territorial elections in my own country, there is one election in 2016 I cannot possibly overlook. Okay, two, but the Australian Federal Election has come and gone, which leaves...

The United States Presidential Election 2016

(Or: No, really guys. Joke's over. Where are the real candidates?)

And when you're talking about the US Presidential Election, how could you not be of Donald J Trump? With less than a week until a combination of voluntary voting and the monstrosity that is first-past-the-post polling conspire to elect the most despised president of the Unites States in recorded history, (though Salon offers and interesting counterpoint), Trump is ahead in the polls today for the first time in a long while. Such a simplistic polling analysis is far from telling of the likely electoral result, however, due to the variation of polling state-by-state and the nature of the electoral college system. A detailed prediction that factors these in will appear on this blog at some point in the coming week (I promise) though history suggests that may well be on the night before the election.

There is, however, an important caveat to be made regarding any predictions for this election: Trump has repeatedly defied conventional wisdom and expert analysis by far wiser minds than mine. Anything could happen.

In particular, I want to look at Trump's polling results and their resilience against gaffes that would eliminate any other candidate.

Anatomy of a Trumpism

Politicians can get away with a lot of things. Embezzlement, corruption, war, sexual assault, even rigging elections. The one thing that was assumed they couldn't survive--at least in a democracy--was upsetting the voters. Since becoming the de facto nominee of the Republican Party on May 4, Trump has (among many less caustic (mis?)statements like forgetting the date of the September 11 attacks) insulted many significant demographics in this election. A short list is provided here:

Date Reported	Demographic Targeted	Demographic Size
03/10/16	PTSD Victims	0.8%
21/09/16	African Americans	12.6%
02/08/16	Women*	50.8%
06/07/16	Jews	0.2%
03/06/16	African Americans	12.6%
31/05/16	Latinos	16.3%
11/05/16	Non-Whites	36.3%
* Certain comments about a Purple Heart Medal may have put American veterans off side here as well.

All of these were sourced from The Atlantic's 'Gaffe Track', where a fuller list can be found, dating back further and covering all candidates. These were, in my opinion, the statements most likely to put large demographics off from voting for Trump. Only the most recent three of these, however, have occurred since Trump became the official Republican nominee and polling started
standardising.

Well, sort of standardising. House biases remained, of course, and sample size varied. We discussed previously the important distinction between phone and internet polling methods. Sample sizes varied, and so did the duration of polling. In short, trying to produce a detailed map of Trump's polling is as futile as it would be inaccurate. And since this blog is all about futile and inaccurate statistical analysis, I figured I'd do just that by averaging all the polls released each day since Trump became the official nominee. (Some days had no polls released, these were skipped on the inaccurate (but not futile!!! :D) assumption that his polling remained on a smooth gradient over these days).

The red line is an average of the blue line for two days in either direction, to remove some of the wild variation.

Next, I highlighted the seven day periods following each statement noted above:

And from this we can determine... absolutely nothing. The blue line remains as erratic as ever, and there is no obvious pattern in the red, either. No consistent period of lost votes regained over time, nor surge in support from "I want a politician who speaks their mind" types. This may be because we only have three sample periods to consider (the first may be an aberration, for example, and a longer analysis will see a general pattern of Trump ending the week lower). But there's not much we can do about that, and no real way of extrapolating any pattern.

But my general observation would be this: Trump isn't necessarily resilient. He doesn't bounce back from his gaffes, because he never suffers any significant loss of support to begin with. It's not that supporters forgive Trump's statements, its that they are unaffected by them. Part of this is probably the polarising nature of this campaign. Most Trump supporters aren't pro-Trump, they're anti-Clinton (and vice versa) and nothing Trump could say will dissuade the bulk of his supporters from opposing his Democratic opponent over poorly configured email servers and Benghazi.

A further explanation might be appropriated from Clinton's response to her own campaign issues regarding the FBI's un-closeing of their investigation into her emails. Clinton believes (or at least hopes) that most people had already made up their minds about this issue before the FBI reopened the case. The same is probably true of Trump supporters regarding the Donald's often hateful statements. There's no one left in the Trump camp who would abandon the candidate over such speeches because, quite simply, they left during the primaries and never came back.

In Australian politics we've had a lot of elections that seem to hinge on a candidate not losing their support. 2007 was Rudd's to lose (which he didn't), 2010 was Abbott's to lose (which he did), 2013 was the same (this time he won) and 2014 had the Turnbull honeymoon period so long as he didn't stuff it up. USA 2016 is not like these. Trump is probably incapable of losing supporters, short of foul weather and the usual variations of the US voluntary system. Assuming his current lead is indicative of a Trump win looming (and I personally am not convinced this is the case), Trump cannot stuff this up and hand the victory to Clinton. The latter, instead, must make a comeback and claim the election on her own merits.