Off the Kuff Rotating Header Image

It was worse in Harris County during the freeze

Very interesting.

Harris County residents were far more likely to have lost electricity and water during February’s winter storm and blackout crisis than residents of other Texas counties, a survey by the University of Houston’s Hobby School of Public Affairs found.

The findings may help explain why Harris County residents account for a third of the almost 200 deaths so far attributed to the storm, while only accounting for 16 percent of the state’s population. Most froze to death in their homes or while exposed to the elements, succumbed to carbon monoxide poisoning or died when medical devices failed without electricity.

“During the week of the winter storm, Harris County residents were significantly more likely than other Texans to lose electrical power, lose internet service, lose access to drinkable water, be without running water, lose cell phone service, have food spoil, suffer economic damages, and experience difficulty finding a plumber,” the survey authors wrote.

Ninety-one percent of Harris County survey respondents said they lost power during the blackouts, compared to 64 percent of respondents from the other 212 counties on the state’s main power grid. Asked if they had lost water, 65 percent of Harris County residents said yes, compared to 44 percent of those in other counties. Thirty-eight percent of local respondents said they suffered burst pipes.

On average, Harris County respondents were without electricity for 49 total hours and 39 consecutive hours, confirming that the outages were not rotating as the grid operator, ERCOT, had hoped. CenterPoint Energy, the Houston area’s electricity distributor, said during the crisis it could not rotate blackouts because the drop in available power to distribute was so severe.

Perhaps unsurprisingly, 72 percent of Harris County respondents said they somewhat or strongly disagreed that the power outages were distributed in an equitable manner.


The survey also measured how Harris County respondents rated the performance of government officials and entities during the storm. President Joe Biden and County Judge Lina Hidalgo scored the highest, with more than 45 percent of respondents somewhat or strongly approving of their handling of the crisis.

Gov. Greg Abbott and state government as a whole were rated poorly, with about 21 percent somewhat or strongly approving of their performance. ERCOT polled the worst, with 78 percent of respondents somewhat or strongly disapproving of the power grid operator’s performance.

The survey found broad support among Harris County residents who identify as Republicans, Democrats and independents for a series of reforms. More than 70 percent of these respondents said they supported requiring the electric grid and natural gas pipelines to fully winterize and giving the Public Utilities Commission greater oversight over the electric grid.

Fifty-three percent of respondents, however, said they were unwilling to have higher utility bills to ensure the grid is better-prepared for severe weather.

The landing page for this poll is here. The data for Harris County is here and for the state as a whole is here. The poll was done via a webpanel, with a sample of 1500 adults in total, and an oversample of 513 adults in Harris County. Note that as before, the partisan makeup of the sample is more Democratic than it would be if we were talking about registered voters. In Harris County, it was 38% Democrats, 36% Independents, and 18% Republicans, and statewide it was 32% Dem, 30% Independent, and 25% Republican. I sent an inquiry about that, and was told that among those who reported voting in the 2020 election, Trump won by a 51-47 margin, not far off from the actual 52-46 spread. In other words, if there’s a Democratic skew it’s among the non-voters.

I say all that up front because there were approval ratings in the polls, at least for how the freeze was handled. For the statewide sample:

Governor Greg Abbott: Strong approve 15%, somewhat approve 13%, neutral 15%, somewhat disapprove 10%, strong disapprove 38%
Your County Judge: Strong approve 14%, somewhat approve 11%, neutral 26%, somewhat disapprove 7%, strong disapprove 16%
Your Mayor: Strong approve 14%, somewhat approve 14%, neutral 27%, somewhat disapprove 8%, strong disapprove 17%
President Joe Biden: Strong approve 21%, somewhat approve 11%, neutral 21%, somewhat disapprove 5%, strong disapprove 32%

And for Harris County:

Governor Greg Abbott: Strong approve 12%, somewhat approve 9%, neutral 17%, somewhat disapprove 9%, strong disapprove 47%
County Judge Lina Hidalgo: Strong approve 35%, somewhat approve 13%, neutral 19%, somewhat disapprove 6%, strong disapprove 18%
Your Mayor: Strong approve 27%, somewhat approve 19%, neutral 19%, somewhat disapprove 6%, strong disapprove 19%
President Joe Biden: Strong approve 36%, somewhat approve 13%, neutral 20%, somewhat disapprove 5%, strong disapprove 19%

Again, bear in mind the partisan breakdown of the sample. There’s a lot more to the polls, and a separate set of questions about lifting COVID-19 restrictions that I’ll write about separately, so go check it out.

Related Posts:


  1. League City peep says:

    Would be curious to see the breakdown by counties. While i think the focus on Harris County is spot-on, I wonder what the immediately-surrounding counties do to the numbers? i would think they make the bulls-eye a bit broader, even if still focused on the Houston area.
    I know for sure here in Galveston County, we also took very long power losses and had all the problems with groceries, water, plumbers.

    This also reinforces, for me anyway, the ghost of an idea that the lack of fire/reaction by those in Austin to these issues is at least partially attributed to the fact that it was focused on (the) Houston (area). Whether intentional or not, it seems those ‘Libs in Houston’ got to suffer and they are not particularly bothered by it. Would they be more reactive if it had been a more Republican space that had been similarly focused in impact?
    Am I really seeing ghosts?

  2. Jason Hochman says:

    Yes, that makes sense. Harris County is primarily Houston. And Houston is run by a regime that is corporate developer friendly and allows building willy nilly with no concern for infrastructure or esthetic design, or any kind of planning. I’m sure that the new construction was part of the reason why the Harris County infrastructure was so weak. And the city and county government knew about it, but didn’t fully explain it to the people. I planned to spend the week of cold weather enjoying some real winter in Houston, by baking ginger men, cooking a pot of lentil soup, popping corn on the stove top, and making hot chocolate, while reading and watching some movies. But it was not to be.

    I am still waiting on the Mayor of Houston to apologize to the State Governor. He accused the State Governor of invoking the Maskless Mandate to deflect from the infrastructure failure of the freeze. As we now see, the numbers for Texas have remained just fine, for one month of the Maskless Mandate, while the Mask States have seen a fourth surge. This is just a fact. And it is in spite of the fact that Texas had 100 migrants who were tested positive released by the Biden regime. And that Texas had spring break in South Padre and Galveston. The Mayor of HOusotn needs to man up, and eat his giant plant of crow and ask the Governor to accept his apology for accusing the Governor of deflecting.

    Dimbulb Beto also was twittering a panic stricken statement about how the Governor sentenced us all to an agonizing death. Hopefully, if Beto runs for governor, the people will remember this. The Governor is a better scientist than the Mayor and Beto, whose science experience is probably a Young Einstein chemistry set in the corner of the garage littered with What A Burger wrappers.

  3. C.L. says:

    Do tell, Dr. Hochman, how exactly a plethora of new construction in Houston ultimately resulted in ERCOT turning off the power to homes in Harris County. This ought to be good !

  4. Jason Hochman says:

    C L, if I hadn’t made it clear, the out of control new construction is an extra strain on the infrastructure. My AT&T goes out every now and then. And the reason is cable cut. I see wires dangling and on the ground all the time. There is a utility pole on W. 18th, and it has some hook up that is just hanging on the ground, I trip over it all the time.

    So two reasons: sloppy, unregulated, out of control construction damaging infrastructure, and, increased customers without improving the infrastructure.

  5. Kibitzer says:


    One might question how meaningful any job approval ratings for local leaders — such as mayor and county judge — are in relation to 2021 winter storm management.

    Granted, the UH survey covered much more ground, but understandably the focus on OTK is on poll data on political leadership.

    One obvious problem (for formation of event-specific public opinion on local-leader performance) is that the TV/broadband was down for so many people, which just left cell phone (or radio) as a news source, with associated battery charge/life issues. So, overall, there was much less opportunity to learn *in real time* what county judge and mayor were up to even for people that pay attention to local pols.

    Therefore, it could be expected that survey responses regarding job performance (if any at all) would be influenced by what the local leaders said *before* the power went out (such as Hidalgo’s warning and Hurricane severity-rating equivalency), not their actual crisis management during the blackout. And for those who received emergency text alerts, who gets credit for them? (If you didn’t receive them, you wouldn’t know whether Lina Hidalgo caused them to be sent).

    Depending on *when* survey responses were provided, such responses could also have been influenced (and distorted differentially) by media coverage *after* the crisis. In other words, respondents may have formed an opinion looking back based on information that they had since received, including post-crisis statements by politicians and commentators (such as Gov. Abbott blaming ERCOT, and not PUC or electricity companies).


    Another problem is even more substantive. Assessing crisis management performance assumes that the crisis manager in question has the ability to do something meaningful. There must at least be some implied standard or notion as to what should have been done. If this condition is not satisfied, public opinion data on job performance will necessarily reflect something else, such as diffuse support for the local leader in question more generally as distinguished from a crisis-event-specific evaluation. A meaningful basis for comparison would be the relative job approval of the various local and statewide leaders under normal conditions, or an array of other issue areas, including COVID-19 response and vaccine rollout. But the caveat regarding uneven access to real-time information about what leaders were doing and saying would still apply.

    And one constraining fact here was the universal lack of information on when the power supply would be back on to satisfy demand without forced blackouts.

    The folks at ERCOT didn’t know either and didn’t even have a reporting system in place that would allow them to assess in real time the reasons for power plant outages, and how soon they could start up or run at full capacity. They later had to do a post-mortem survey, and even the results of that one will be of dubious reliability and validity as companies have reasons to not be forthcoming.


    As for cutting power to residences, and restoring it, that was controlled by ERCOT at the systemic level (i.e., the allocation of mandated load shed to CenterPoint service territory) and by CenterPoint in Houston (and local utilities elsewhere) as to which neighborhoods in its territory to take offline. It doesn’t appear that local leaders had any ability to have input, at least not in the Houston area. (Other cities such as Austin and San Antonio, for example, have municipally owned utilities).

    Much of the affected public probably did not know how these blackouts were imposed and may have blamed their chosen retail power providers (since we have competitive retail market in Houston), or the storm itself. After all, when the power goes out, it cannot be known whether the cause is a downed line (a natural event triggering a circuit breaker) or a deliberate shut-off in the control room or at a sub-station).

    Affected residential consumers would have learned about the role of ERCOT only after the fact/blackout. Very few had even heard about ERCOT before, not to mention its role of keeping the electrical grid in balance, and wouldn’t therefore have a basis for an opinion.

    As for the knock-on water supply and quality issues, it could be said that the City bore some responsibility to the extent back-up equipment to operate pumps failed, but again: there was no real-time information available to people (and therefore to survey respondents) as to why no water was coming out of their particular faucet (even with internet available). Nor could the residence- or neighborhood-specific cause be determined looking back. It would therefore even be more difficult to allocate blame/responsibility with water supply issues — and assess performance in handling the restoration/recovery — compared to the electricity supply interruption.

    This is not to say that such online questionnaires cannot serve some useful purpose, but the findings should be taken with a grain of salt, and put in a proper theoretical framework. Particularly when any lessons are to be drawn from such data.


    Also of relevance, as pointed out by Kuff, is the sample frame. The voting public is a subset of registered voters, which is a subset of population legally eligible to vote, which is a subset of resident population (which also includes those below voting age, noncitizens, and citizens deprived of the right to vote).

    As is typical for surveys, young people under 18 were excluded. – Is this justified when the sample frame is neither actual voters nor vote-eligible citizens? Shouldn’t 16-17 year olds be represented in the sample? If for no other reason that many of this demographic will have the right to vote come 2022.

    Note that the electoral dimension was within the scope of the survey:

    “Harris County residents also were asked what type of impact the winter storm would have on their 2022 vote decision. The options were that it would have no impact since they would not be voting in 2022, that it would not be a factor in their vote decision, that it would be one of many factors in their vote decision, and it would be a very important factor in their vote decision.”

    Also: On-line surveys do not include off-line segment and people not inclined to participate in such surveys. The latter group’s attitudes may be distinct from those willing to do so irrespective of the usual demographic attributes for which some statistical corrections might be implemented (such as re-weighting or selective deletion of cases from the initial nonrandom sample to match known distributions in the population about which inferences are to be made).

    So, to the extent public opinion should play a legitimate role, whether in the electricity system policy debate or in forthcoming elections, or both, some thought should be given about the proper sample frame, i.e. which segment of the population the sample statistics represent.


    Arguably, the most valuable information here is the respondents’ report of their actual personal experience with electricity outage and water service interruption/degrade, including the research finding that matters were worse in Harris County, compared to the state as a whole.

    Those parts of the survey instrument went to matters of fact and sensory experience, not just of abstract perception/attitude, emotional response, or conjecture in the absence of pertinent information and lack of underlying knowledge of how the electricity system operates.

    But that’s perhaps stuff for another blog post … and additional ruminations from the OTK Kommentariat.

  6. Bill Daniels says:


    I always enjoy your posts, but may I suggest a TL;DR Cliff’s Notes version at the bottom of the more lengthy posts?

  7. Bill Daniels says:

    Answering the question of why Harris County suffered more than it’s share of power outages…..hey, ERCOT’s appointees were probably Republicans, and Harris County’s leadership is now firmly Democrat. Don’t think there was not some satisfaction taken in putting the brunt of the blackouts on Democrat led areas. What better way to get the citizenry upset with Sly and Dora?

  8. Bill Daniels says:


    “Affected residential consumers would have learned about the role of ERCOT only after the fact/blackout. Very few had even heard about ERCOT before….”

    I’m going to take exception to this supposition, Wolf. Literally every Summer, on the hottest days of the year, we hear from ERCOT asking people to conserve, because usage is peaking. I would be surprised if more than 10% of the public does not know about ERCOT, or have some sort of understanding about what they do.

  9. Tpain says:

    Dora???!!! Really? You want to go there with that? I guess you and your buddies are playing some sort of drinking game and this is a twofer.

  10. […] little while ago I blogged about the recent UH Hobby Center poll regarding the winter freeze and blackouts and responses to […]