The usual reasons appear here: being bad at estimating how long common tasks will take. My first husband was astonishingly bad at this, and always assuming he could get done much, much faster than he ever had in the past. My oldest sister said that there were two things her computer science professors said would make her and her classmates highly desirable employees: typing accurately and fast, and being able to schedule accurately. Getting people to think about how long it took them to do something in the past helps them better predict how long it will take them in the future. Getting people to break down larger tasks into small components and building an estimate based on the smaller components also makes a big difference.
Jeff Conte is quoted as finding in numerous studies something that shows up in studies done by many people. He sorted by type A/type B personality types, but people have been doing this research with a lot of different markers, and there are groups of people who will slightly short a minute (believing a minute has passed in about 58 seconds) and people who will drastically exaggerate a minute (77 seconds). It's a little terrifying how poorly centered that curve is!
Others mentioned in the article say that age of employee's child is a good predictor for lateness (I bet special needs kids have an extra impact). ADHD and other mental health issues are also contributors. Having difficulty saying no/overbooking is another.
All of this stuff is captured in numerous self help books, along with advice on how to fix the problem. I ran across DeLonzor's book, which enjoys phenomenal reviews on Amazon.
She offers 7 cures, for 7 different varieties. I have some issues with her perspective on how to develop self-discipline, in that self-denial, imo, doesn't actually build self-discipline and it's pretty easy to imagine some of her suggestions as depleting that capacity (see this NYT article from 2011, h/t Korbett Miller). She also repeats the marshmallow study in the overly simplistic way, which Alfie Kohn does an excellent job of debunking here: http://www.edweek.org/ew/articles/2014/09/09/03kohn.h34.html.
But on the whole, Delonzor has a template for the self-help she is offering and it is a good one. Understand why you are late -- pay attention to the details. Design a way to counter that behavior in a tailored way. So for Cure #1 -- the Rationalizer -- the first exercise is to review your history of incidents and consequences especially the reactions of others. The second exercise is tracking future incidents on a calendar for a month. The third exercise (assuming #2 confirms that you really do have a problem) is to identify whether there are any changes you can make in your life to prevent recurrence. The second set of exercises is parallel, and involves the impact of lateness on others: talk to someone for another perspective, enlist someone to create incentives/disincentives (make me pay for dessert/the wine if I am late for dinner; compliment me if I show up on time or at least a lot less late than usual). Finally, notice how you feel when you start showing up. The third set of exercises involves self-talk/cognitive structures: quit with the denial and do some perspective taking on the person who is waiting for you; generally pay more attention to other people; apologize freely and quit making excuses.
So (in case it isn't obvious), the chapters first describe a cluster of symptoms or behaviors or mindset that leads to lateness, and then prescribes a set of relatively specific steps to take to move from lateness to punctuality, behaviorally. The first step is designed to move the person reading to acknowledge the problem as changeworthy. The second is designed to quantify the problem and design an intervention. And the third one is designed to ensure that the person who is making the changes perceives the effects of the change so that the change has a chance of becoming permanent.
Basically, competent habit change advice.
While this might sound like rationalization, I don't have a problem with tardiness in general. I do have some specific schedule problems. For example, the kids have swim lessons on Friday that are scheduled too tightly with T.'s arrival home from school for me to reliably get A. in to her lesson on time. This is not a problem except when R. works on Fridays, which he has been doing lately to make up for staying home on some snow days (shifting Friday to Monday, typically). Short of finding another teacher at a different pool to take my kids on, I'm pretty much gonna be late by at least 5-10 minutes when this happens, and maybe more if one of the kids throws a tantrum or Route 2 is stop-and-go. I used to have a horrible problem with Route 2 going down to Somerville for a Dutch lesson, but my instructor moved, which helped, and construction on Route 2 has passed the period of completely catastrophic and unpredictable 45 minute delays (in order to be sure I'd arrive on time, I was winding up arriving 20-25 minutes early, which is just awkward, and parking in Somerville being what it is, not actually easy to figure out something to do that fits into that time slot. You might think, hey, take the commuter rail, but the schedule to South Acton and back at that time of day did not work out at all). This book doesn't have any wisdom on the subject of specific schedule problems, which I harbor a suspicion other people have as well (and which is why I don't stress any more about people being late when it isn't habitual, or even when it _is_ habitual, if I understand and find legitimate the schedule conflict that is producing the problem). This book is aimed at that relative you have who reliably shows up to family functions 1-3 hours late (in this job market, if they were a coworker, they'd have lost their job, and if they are your friend, well, good luck to you), the old friend who is never less than 20 minutes late (and never, ever, ever early), etc. and whose explanation is banal (traffic when it wasn't any worse than usual, type of thing) or exasperating (I didn't have any clean clothes to wear, I took a nap and overslept, etc.)
I'm not sure I actually have any regular contact with anyone who behaves this way any more. Which is kind of amazing, now that I think about it, and probably means that I probably shouldn't push too hard in my efforts to become any more cold-hearted than I already am. But if you have a problem with chronic lateness, and it isn't driven by demands upon you as a caregiver, and assuming you aren't juggling multiple jobs none of which can be modified because you're barely making it as it is, you might find some useful ideas in here. Severe ADHD is beyond the scope of this book, as are issues such as agoraphobia and other global anxiety issues.
ETA: Here's the BI article that got me started on this:
Part of my interest is because I have an uncle who is just _infamous_ for this kind of lateness, and it didn't even seem to really bother him. I never understood it.
When I looked up the definitions of Type A/Type B, I was pretty unimpressed by them. I also hadn't realized just how much that typology was discredited some years ago by tobacco companies exploiting the Type A cardiac/smoking connection to justify non-cessation oriented interventions for the health risks of smoking (took me a minute even to understand the thought processes. I just went, whaa---? And then I remembered smoker-logic and just eyerolled and moved on).