What are the main factors and drivers that repairing and maintaining is often significantly more expensive than buying the good new?
Whether repairing or replacing is cheaper varies a lot between situations.
You wouldn't buy a new bike or car because of a flat tire.
Depending on the hole and your DIY abilities, replacing the tube or tire or patching are the usual ways to deal with this.
But putting in a replacement zipper into a well-worn jeans is a different situation as the repair starts by almost completely disassembling the jeans. So even a professional jeans sewing expert may need considerably more time to replace the zipper than sewing a completely new jeans. If then the expected increase in remaining lifetime isn't that much because of general wear, replacing the whole jeans may be the way to go.
In addition, if the zipper is never failing before the jeans is thrown away due to wear and tear, it was an overengineered waste of ressources.
Of course, two products with the same functionality may be more or less repair friendly, thus potentially shifting the rational, economic and ecological decision boundaries.
There are also often inherent trade-offs for the producer that affect repair friendliness:
repair friendly vs. ruggedness: consider a mobile phone, gluing the case and soldering the battery vs. screwing on a lid over replaceable battery. The latter is clearly more repair friendly, but without additional effort, the former may be more rugged against falling, water and contact issues of the battery.
repair friendly vs. ease of assembling: consider some household machine with the case being plastic and screwed vs. clicked on. The former is repair friendly, but the latter is more easily assembled in the factory.
spare parts: repair friendly design makes sense only if spare parts are produced and sold for the envisioned (enhanced) lifetime of the product.
This can be quite costly. Not only in terms of money, but there are likely to be also ecological trade-offs (which may improve with small series manufacturing/production on demand improving).
spare parts vs. product improvement: Supplying spare parts is easier/cheaper if your product stays the same over longer time/many units. While many changes in products are largely motivated by marketing without providing better functionality, there is also an imoprtant tradeoff here in case of improved functionality.
E.g. is it better to stick to the current fridge model in order to supply spare parts for longer, allowing for repair and increased lifetime - or is it better to discontinue the model and move on to a more energy saving next model?
repair-friendly design: (incl. using mostly standard parts). A product that has a given functionality and is as repair friendly as possible may not only be more expensive to produce, but already the development is probably more expensive as not-repair-friendly ideas have to be improved upon.
Of course, if repair-friendly design becomes more common, standard solutions will emerge and the state of the art in this respect will improve.
However, I'm told that universities around here do not even offer courses in safe design (as in product safety) for engineering students - despite safe design being a legal requirement for decades now.
Safety vs. ease of disassmbling: Where I am, certain types of safety measures must be attached in a manner that they cannot easily be disassembled or otherwise put out of action.
Also, the click-on case in the 2nd example may better from a product safety point of view if it cannot easily be taken off.
Legal risk: (I'm in the EU) An additional thought may be (depending on legislation) that e.g. such cases that are not easily disassembled may give the manufacturer a bit more legal safety wrt. accidents during repair attempts: even if the manual says "do not attempt to repair", a case that is trivially taken off would be a foreseeable misuse and thus have the producer liable for safety/accidents.
Also, I just learned in a seminar about product safety legislation and in contrast to data services where GPPR over here make decided statements against vendor lockin, allowing repairs or merely handing out detailed plans of the product that make the users think they may repair it may place the manufacturer at higher risk in case of accidents.
This doesn't mean repair friendly cannot be done with acceptable legal risk, but it does mean that releasing (existing) information that would come handy for repair comes at surprisingly high cost for the producer: there is basically no possibility to say take this info at own risk - if released, it basically has to be in the form of a repair handbook that is safe for the average customer (or repair person).
(I've been interested in DIY labware projects for some time - and after that product safety regulation seminar I understand quite a bit better where the high prices for some not-too-complicated labware machinery comes from: the existing regulations are extremely unfavorable to anything that isn't a mass product)
Costs and consumers shopping for low prices. Most "stuff" is made in massive quantities in factories where process and other types of engineers have refined the steps to produce an item. Each person on the floor needs to know only one small bit of the operation. Alternatively, to repair something, a capable service man needs to know all aspects of the item. Such knowledgeable service people are expensive to train, equip and require relatively high pay. So repairs cost more than manufacture of a new item generally.
It is a question of man hours used with technology against man ours (your time used in making repairs) without technology. If you worked in a factory you could make a hundred items in the time it would take you to repair one item. Even with a household sewing machine, you couldn't keep up with the speed of factory output. Everyone who works for a wage is in some way contributing to factory output. So any time that you spend repairing anything is going to be less productive than the time you spend at work. To make repairing "Pay" you need to corner a niche, and repair things that other people value and want to keep so much that they are willing to pay a bit more for them. Only that way can you make your time spent repairing, equal in value to time spent earning a wage.
Another element that other answers haven't touched on is Economic Production Quantity.
For many components in products, for example the metal casing of a washing machine, or any plastic injection moulded part, the machinery required to produce them takes a significant amount of set up time. This set up involves getting the mould out of storeage, installing it in the press or injection moulding machine, checking it all works.
The material that parts are made from could be only available by the tonne, sufficient to make a few thousand of components. It might be impossible to buy the material to make just one unit.
So production runs for some parts will be several thousand. Because the set up costs and material costs are spread out over several thousand, then one part might only cost $1. But if you were to 3D print, or otherwise just make one of that part it would cost $20.
If you were an artist making one discrete product, you might be willing to pay $20 for each component and £2,000 for the finished piece. But to mass produce a product in quantities of thousands then a lot of cost can be taken out due to manufacturing in bulk.
Another example is in logistics. In a factory a storeman will be given a list of parts to build a batch of products, they will wander round the warehouse locations grabbing each part. It takes approximately the same amount of time to visit each location whether the batch size is one unit or a hundred units. He still has to get paid, but if he's doing a batch of a hundred then the cost is spread out more thinly.