On Reducing Rush Hours
How modifying three social systems can lead to easier, greener living
In discussing global warming, everyone talks about systemic change; few propose ways to actually do it. Here I present the results of some research I’ve engaged in for the last few months, which are initial steps into the kind of societal systems re-thinking we will need to engage in to soften the blow of climate change.
The Short Version: Traffic robs us of all that is holy. It’s what happens when our road system fails to handle the difference between peak demand and off-peak usage. One reason we have congestion, whether part of the weekday or the weekend rush, is that our society generally operates around a standard schedule. What’s interesting is that there is an analogous problem in our electrical grids. Both types of congestion are two specific instances of the problem of peak demand, which stems from two things: lots of people living together in the same area, and a lack of smart coordination.
One way to reduce peak demand across different social systems is by redividing the time and space in which we collectively organize our schedules.
In terms of redividing time, we can reduce weekend crowds by having people volunteer to cycle their weekends between Friday-Saturday, Saturday-Sunday, and Sunday-Monday. And by explicitly offsetting when people are expected to work their “9–5” in 10 or 15-minute increments, we can reduce daily rush hours.
In terms of re-dividing space, offsetting when cities start their days via a series of half-hour or 15-minute wide “micro time zones” would likely see the reduction of peak demand on the electrical grid — my initial analysis of 15-minute electricity load profile data supports this conclusion. While further research is needed, the discussion and implementation of these systemic schedule adjustments have the potential to form a grassroots climate change effort which we all can partake in. Lastly, in terms of implementation, college campuses may be an ideal microcosm to test out these ideas.
The Longer, Illustrated Version:
1. Introduction
Traffic is one of the most salient, bone-grinding realities of city life, common to every urban center on earth. We’re all aware that road congestion isn’t just confined to a daily rush-hour — weekend traffic plagues city dwellers as well.
Bay Area residents, for instance, often drive up to Napa, Sonoma, or Yosemite in the summer and Lake Tahoe in the winter for weekend trips. Not only do they find their destinations crowded, the traffic coming back is often terrible. Though the trip from San Francisco to Lake Tahoe takes roughly 3.5 hours without traffic, it can take up to 12 hours with traffic.
Daily rush hour and weekend traffic are instances of “peak-demand congestion” in our road system; which, on an abstract level, is an issue with every societal system. A similar thing happens in our electrical grids, our water and sewage networks, our public transit systems, our beach and museum parking lots, etc. Most people utilize societal systems in similar manners at similar times, which means the systems in question have to be engineered to handle the large disparity between peak and non-peak usage.
2. Examples of Peak Demand in Various Systems
Consider the daily demand for parking spaces in a parking lot. Here is a graph of the average 24-hour percentage utilization of downtown Santa Monica parking lots in 2018, broken down by Weekend and Weekday:
Common sense, experience, and this graph tell us two things. First, the Santa Monica parking system works when it has the capacity to handle its versions of rush hour(s). Second, the city is simultaneously incentivized to have enough capacity to handle most of the crowd on most crowded days, while avoiding building too much extra capacity.
For an example of “rush hour” in the electrical system, here is a graph of the demand on the California power grid on 4/29/2019. Once again, note the difference between the highest and lowest point:
In Britain, one type of peak-demand congestion on the electrical grid is called the “TV Pickup.” It happens when a large number of people watch the same TV show or sports event and use the commercial break to open the refrigerator and make tea.
The reason peak demand is an issue is that the most expensive electricity is purchased during times of high demand from coal- and gas-fired “peaker power plants.”
The equivalent of rush-hour congestion affects more than just municipal road and electrical systems. We are all familiar with long lines at grocery stores after work or on Sunday afternoons, or what happens when several international flights are all scheduled to leave within five minutes of each other and the line at customs becomes really, terribly long. As of 2012, Americans spent roughly 37 billion hours waiting in lines per year. In some sense, congestion is one of the least desirable aspects of living in a densely populated area.
What’s interesting is that one reason Peak Demand happens is that we were all born into a system built around a) one schedule, and b) the same clock. We think this arrangement works, and so we grin and bear it, even though the former assumption is not strictly necessary and better solutions are possible.
3. Re(de)fining the Problem
First, a word about why we have the system we’re in. One way of phrasing the main reason we’re all on the same schedule is so that we can each take the same time off as people we want to see. Another phrasing is that the services, institutions, and spaces we use or work at (banks, restaurants, dance-halls, typewriter-repair shops, pastry shops, etc.) have all been built in reaction to each other’s schedules. But given that these things are run or attended by people, in a way, this second phrasing uses different words to say the same thing as the first phrasing.
For the purposes of illustration, allow me to drastically oversimplify things. Let’s pretend that my social network solely consists of five people very close to me — my parents, one romantic partner, and one best friend. In the following visualization, I’m person “0.”
I want all to have the same time off as my five connections, and so we all coordinate to work the same hours. Great, so far, no problem. The issue arises when we extend this courtesy to each of my five people’s five connections.
In this simplified society, where every person strictly relates to exactly five other people, we can quickly get millions of people distributed across thousands of square miles who all seemingly have to be on the same schedule. If one person wants to change their schedule, then the entire network, locked together, will be forced to change at once. The sheer size of the network, however, ensures no one will ever change on their own, and the lockstep synchronicity across the entire network is the main cause of rush hours.
As I’ve mentioned, reducing congestion and rush hours involve finding ways to relax the lock-step-ness of the network.
To put it visually, rush hours could be reduced if there was a way of demarcating groups (indicated by the green and blue lines) and assigning them to different “schedules.” In the chart below, all people to the left of the green line have flex-jobs working as waiters and take Monday and Tuesday off. Maybe the people to the right of the blue line all work from 7:30 to 3:30 and help smooth out rush hours.
4. Weekly Traffic Reduction Via Shuffled Weekends
Two hundred years ago, the official western weekend was just one day long: people had Sunday off so that they could go to church. Unofficially, many (including Benjamin Franklin) also skipped work the next day too, and called the practice “keeping St. Monday.”
From a business owner’s point of view, perhaps what really matters is that their employees work an average of five days a week, not necessarily five days every week. Assuming this is true, then it is possible to shuffle the weekend around so that traffic and both work-week and weekend congestion is reduced. However, we still want to preserve one of the main function of the weekends, which is having the same time off as other people.
Consider the following cyclical schedule pattern: a six-day workweek and a two-day weekend, followed by a four-day workweek and a two-day weekend, and then a five-day work-week and a two-day weekend, followed by a six-day work-week all over again, etc.
For the sake of argument, let’s say half of a theoretical city volunteers to start on the six-day workweek, and the other half volunteers to start on the five-day workweek. Both groups wind up with the same two days off every three weekends, and on the other two, they overlap by a day. This rotating overlap allows everyone in each group the opportunity to see their friends.
With this scheme, both groups always get Sunday off, and wind up with the same two days off every three weekends. More importantly, two out of every three Mondays and two out of every three Fridays see 50% fewer people driving to work. Theoretically, half of everyone’s weekend could be half as crowded, two-thirds of the time.
From the point of view of retailers, restaurants, theaters, theme parks, shopping malls, and other destinations, the weekend effectively expands: they all become less congested over the 3-day split weekend, but they do not see a reduction in total volume of people who visit them. From an individual’s point of view, these places are still popular, but not bone-crushingly crowded. Weekend and weekday traffic is reduced, commutes are easier, and everyone has more free time and peace of mind. In fact, this plan is actually more advantageous if some part of the population retained the traditional Monday through Friday workweek.
There’s also no need for membership in the groups to be rigidly defined, wherein once you volunteer for one of the schedules you can’t switch to the other shuffled schedule, or the normal M-F one. Instead, people can hop around at the end of each three-week period: what matters is that we offset our schedules just enough.
True, working around people’s needs and obligations (like their children’s school’s schedules) will be hurdles. Yet if alternating schools or grades were to be on one of these two schedules, dropping kids off at school might be less fraught.
Once the assumption about weekends has been seen for what it is, the next question concerns the optimal set of schedules and days off. Using a dot (·) to represent non-work days, in addition to the 6··4··5 and 5··6··4 repeating schedules, a third group could follow a 4···5··6· repeating schedule. This one is a little more extreme, but would preserve the “Sundays off” component and further reduce weekend congestion. Lay these on top of the current 5··5·· schedule all 9–5ers work, and now there are four potential schedules. Each additional schedule helps smooth out congestion. [1]
5. Reducing Daily Rush-hour Via Multiple Shifts
In the previous section, I explored what possibilities are open to us if we let go of the assumption that weekends must be tied to the days named Saturday and Sunday. Yet while shuffled weekends are meant to alleviate general congestion on certain days, they don’t specifically address daily rush hour.
The same principle of shuffling schedules, but on a different scale, can help us out here. The following “shifts” of work, or of banking hours etc. could be culturally accepted: 8:45–4:45, 9:00–5:00, 9:15–5:15, 9:30–5:30, and so on through to 10:30–6:30.
In practice, not everyone working full time follows a strict “9–5” schedule. If Germany or other countries get serious about implementing a 32-hour workweek, then they have more latitude to shuffle “shifts,” further reducing traffic and congestion.
6. Empirical Research On Traffic Demand Management
The strategies of shuffled weekends and multiple shifts can help smooth out the intensity of daily traffic. (Click here for an overview of traffic demand management). The empirical research I’ve reviewed has found that both ideas work.
Specifically, in 1970, peak demand reduction was successfully implemented by the Port Authority of New York and New Jersey, and various measures of transportation congestion (people in subway stations, the peak volume of people waiting for elevators,etc.) were reduced. [2]
Both of the strategies are often wrapped up in what researchers call a “compressed workweek,” (CW) and CWs have been found to improve commutes and reduced fuel consumption. [3,4] Note that CWs usually just extend everyone’s weekends to be three days long.
A limited version of the shuffled weekends idea was implemented for public workers in the Phillippines for April and May of 2002. While individuals reported reduced commute times, unfortunately, the proper data was not collected, leaving the evidence in the “promising, but needs further study” category. [5]
Lastly, in 1988, the city of Honolulu experimented with a limited, “two-shift” version of this idea and found that commute times were reduced, on average by a range of -4.6 to 14.33 %. Those that voluntarily chose to shift when they were expected to be work liked the program; and those that were forced to participate, surprise, did not like it. [6]
Given the promising empirical evidence, my conclusion is that no one has implemented these ideas widely enough — usually, it is an individual forward-thinking company or factory, rather and local society. In addition, while researchers have looked at transportation-related metrics, to my knowledge, no one has quantified how people’s overall experience of practical city living has been improved. (As I mentioned above, a college campus or a small town might be an ideal place to test these ideas out, one can look at grocery store lines, the amount of time it takes people to get service at the post office, etc.)
Many people and individual companies utilize non-traditional scheduling to avoid rush-hour traffic; it has been successfully implemented in a wide variety of places. Societally speaking, it is already being done implicitly. It seems like we stand to gain from making this sort of thing explicit, which leads me to two concluding thoughts.
First. To my knowledge, no research team has analyzed or simulated these ideas thoroughly enough, and the widespread adoption of GPS-enabled smartphones offers a tantalizing avenue through which the advantage of these proposals can be quantified.
Second. The initial evidence indicates that these schedule proposals should be part of how people see the world part of the common vocabulary. Instead of a shuffled schedule just being a quirky thing done by a progressive employer, it will have maximum benefit if such schedules become as commonplace as our current conception of “the weekend.”
If you think about it, western society has names for days of the week and months of the year; other cultures have names for hours and names for the years themselves. What’s missing is commonly used names for various patterns of work/weekend days and working-hour shifts.
6. Reducing Electrical Grid Peak Demand Via Micro Time Zones
So far, I’ve proposed that we re-think how we divide our societal conception of time as a way to break up the lock-step synchronicity that contributes to congestion in our traffic systems. But as I mentioned in the introduction, there are systems other than traffic negatively affected by peak demand, and one that is especially pertinent is the electrical grid.
While shuffled weekends and work-hours will likely have secondary effects of reducing carbon emissions if implemented, reducing our need for power plants will have a bigger effect. As I mentioned, one way to achieve this is to revisit our notions of how we divide the physical space in which we all keep the same schedule, or more simply what we call time zones.
To jog your memory, here’s another chart illustrating the demand on the Californian electrical grid demand, this time from April 4th, 2019:
In words, this graph says that at the end of the day, just about everyone in California comes home from work or school, and turns on the lights, cooks, and watches TV at roughly the same time. This is despite the fact that the sun does not set at the same time across the state: Lake Tahoe — at California’s eastern border in the north part of the state — is west of Los Angeles.
One of the main reasons for this is because (in America), railroad companies were responsible for both adopting a standardized clock, as well as getting cities to shift into standardized time zones. On November 18th, 1883, “600 railroad lines dropped the 53 arbitrary times they had been using and… defined the times in each of the four new time zones — Eastern, Central, Mountain, and Pacific. Most major American cities followed the railroads’ lead, adopting zone time for their own uses…” [7] Three-and-a-half decades later, the US federal government officially recognized the four time zones (with some modifications) and signed them into law.
One effect of the adoption of time zones is that electrical peak-demand is a bigger problem than it could otherwise be, another is that living near the western edge of a time zone can negatively affect your health. Mere “railroad convenience” is a historically meaningful but increasingly irrelevant reason for having millions of people keep similar business hours. [8] One solution to peak-demand reduction is thus to get people (and thus factories, etc.) in different cities to shift their schedules as if they were in a micro-timezone (which I notate as “μ-timezone” for short).
One reason the southwest coast of America has peaking power plants is that San Francisco, Sacramento, Las Vegas, and Los Angeles all keep the same business hours, even though there is a large east-west disparity between these regions. The same is true on the American eastern seaboard, wherein Atlanta and Boston keep the same socially constructed business hours.
A similar solution has been proposed for the country of India, which currently according to one official time zone. Dr. Amlan Chakrabarti, a professor of electrical engineering, has developed models which suggest that if three one-hour-wide time zones were to be introduced to the Indian Subcontinent, the entire Indian electrical grid would become 5% more efficient. [9,10]
But if the question is one of peak-demand reduction and not railroad timekeeping, why stop at time-zones that are an hour wide? If you were to divide each hour time-zone by four, or get people to shift their schedules by 15–45 minutes the effect would be to further smooth out the impact of the “rush-hour equivalent” (i.e. peak demand) on the electrical grid. Importantly, the clocks don’t need to be shifted; shifting when stores and services open as if the clocks have shifted would amount to the same thing from the power grid's point of view.
One way to implement this solution in California is to use the following boundaries as guides to drawing 15-minute μ-timezone offsets:
Also, you’d want to link these μ-timezones up with states to the north and Mexico to the south, and while you are at it, you might as well do the whole US:
I’ve done some initial modeling the effects of micro-timezones on peak demand, based on data from 475 customers from a midwest public utility in 15-minute increments.[11] I analyzed the day of highest demand out of the entire year of the dataset, and I simulated assigning equal populations to four different μ-Timezones, 1000 different times. Here are the ensuing load profiles:
After simulating assigning customers to different timezones 3000 times across the whole year, the data supports two conclusions: first, peak demand is indeed reduced, and second, four 15-minute μ-timezones are better than two 30 minute μ-timezones. In the four μ-timezones scenario, I found the average net reduction of peak 15-minute interval to be roughly 1390 kilowatt-hours, for an average of 2.9 kwh per customer.
Currently, it is impossible to say to what degree this generalizes — if the average reduction per customer holds, then when scaled out to millions of customers, the savings could be vast. However, data is hard to come by, and assumes that behavior is the main driver of peak demand, rather than something like the temperature. At the same time, there may be more benefits to smoothing out the “ramp” up to the top of the demand peak, my understanding is that a smoother ramp means less congestion in the high voltage transportation wires and less wear-and-tear on the electricity infrastructure.
These results, the graph above, and Professor Chakrabarti’s work all indicate that the question is not whether μ-timezones will reduce peak demand, but by how much. The next steps for this idea probably involve a researcher or academic lab getting more data and scaling up the analysis. Given that such a system would likely see fewer power plants built, electricity could be cheaper for consumers, and ultimately more profitable for power utilities.
6. Conclusion
Peak demand problems affect every municipality of a certain size on the planet earth. Even if we were to colonize Mars (say, in response to climate change), traffic, congestion and peak demand issues would still be present provided we keep the same scheduling traditions. Conversely, once we truly solve these issues in one place, we will know how to solve them just about everywhere.
In my mind, there are three steps towards the larger goal of getting local cultures to shift and shuffle their schedules.
First, the types of schedule shifts proposed above need to be more thoroughly explored by researchers, economists, traffic engineers, and policy wonks. Questions that should be answered include: How much will different rush hours be reduced by these solutions in major cities? What specific environmental benefits will these plans likely entail? How much easier will it be to commute across town? How will the price of electricity change? Are there any unintended consequences that can be foreseen? If all of the proposals I’ve outlined here are implemented, what kinds of scheduling changes would that mean for people in each zip code? I will be pursuing the answers to these questions with the resources I have available to me; in the coming days and weeks I will be seeking out academic partnerships to further these studies.
Second, a practical thing all of us can do to concretely combat climate change is to start introducing these ideas to each other, and thus into our culture. We’ll all need to discuss these proposals, see if we could adapt to them, and work to convince our bosses, school principals, boards of directors, city officials, and so on, to see them through.
Questions worth discussing include: If everyone in your town agreed to change when the business day was supposed to start, could you see yourself going along with that change? What would it take for you to work a shuffled weekend schedule? How many people you relate to would also need to shift theirs?
Shifting schedules has the potential to reduce traffic and the price of electricity; the tradeoff is that coordinating meetings and social gatherings might take slightly more planning. While getting large numbers of people to change their behavior is non-trivial, these peak demand reduction proposals have the potential to form a grassroots climate-change movement all of us can partake in.
In the meantime, never sit in traffic, look at your electricity bill, or have global warming cross your mind and avoid thinking about how things could be better again.
____________
Endnotes / Citations:
[1] Keen readers will also recognize that there is yet another possible schedule, based on the same 4,6,5 structure: 4··6··5. To ensure that each schedule always gets Sunday off, this would need to be offset by one day and start on Tuesday.
Readers may also notice a connection between how increasing the number of schedules helps to better smooth out congestion (by distributing the crowd to different days) and the shift to from Just Intonation to Equal Temperament tuning systems (where the ‘error’ is distributed equally to all keys.)
[2] O’Malley, Brendan W. “Work Schedule Changes to Reduce Peak Transportation Demand.” Special Report-Transportation Research Board, National Research Council 153 (1974): 166.
[3] Hung, Rudy. “Using compressed workweeks to reduce work commuting.” Transportation Research Part A: Policy and Practice 30.1 (1996): 11–19.
[4] Percoco, Marco. “The impact of working time on fuel consumption and CO2 emissions of public fleets: Evidence from a policy experiment.” Transport Policy 71 (2018): 126–129.
[5] Sundo, Marloe B., and Satoshi Fujii. “The effects of a compressed working week on commuters’ daily activity patterns.” Transportation Research Part A: Policy and Practice39.10 (2005): 835–848.
[6] Staggered work hours for traffic management: A case study (Vol. 1280). Giuliano, G., & Golob, T. F. (1990). Institute of Transportation Studies, University of California, Irvine.
[7] The Invention of Railroad Time, Ian R. Barkty, Railroad History, Volume 138, Page 13.
[8] As the railroads grew to connect America, they found adjusting to each city’s local time to be a hassle. In addition to making it easier for themselves and for travelers, another reason they adopted “standard railroad time” was due to efforts of a small group of people. The group includes Charles Dowd, Cleveland “Father of the Weather Bureau” Abbe, who wanted people to make weather and astronomical observations at the exact same time in different parts of the country; Sanford Flemming, a Canadian railway engineer who saw the advantages of standard time, and William F. Allen, the secretary of the General Time Convention who urged the railroad companies to “settle this question among ourselves, and not entrust it to the infinite wisdom of the … State legislatures.” The Adoption of Standard Time, Ian R. Barkty, Technology and Culture, Volume 30, Issue 1, Pages 25–56.
[9] Carbon Emission Savings by Reduction in Cycling Operations of Power Plants, A. Chakrabarti and C. K. Chanda, 5th International Conference on Advances in Energy Research, Book of Abstracts.
[10] Reducing Peak Demand Via Time Zone Divisions, Amlan Chakrabarti, Journal of The Institution of Engineers (India): Series B, Volume 95, Issue 3, Pages 219–230.
[11] Props to Vote Solar for making this data available to me.
Lastly, a huge thanks to Karoun Kasraie, Sharon Liu, Sachu Constantine, Jamie Levhic, Malini Kannan, and Dr. Amlan Chakrabarti for their feedback on these ideas.
