Why is it that native speakers of Japanese have a hard time pronouncing "l"? Whenever a western word contains "L" I see that they pronounce it "ru", "ra", "ro", "ri", or "re".
If you ask a Japanese person to say a word like renraku fast, and then gradually ask them to say it more and more slowly, you will notice that what initially sounded like an r becomes an l as they slow down (usually earlier on for women). So the claim that l and r don't exist is simply wrong -- they both do, but as variants (allophones) of the same sound (phoneme).
If you tell them that they first said an r and then an l, you will surprise the heck out of them as they usually have no idea about this and will probably be unable to tell the difference. The reason is fairly simple: it's the same sound in Japanese. The only difference between the Japanese l and the Japanese r -- not the English r, which is quite different -- is mostly the duration of the contact between the tongue and the roof of the mouth and while this feature allows the two sounds to be distinguished in European languages, it is NOT a distinguishing feature in Japanese.
Similarly, English speakers fail to distinguish many sounds. Some languages have nasal vowels (French, Portuguese, Polish, etc.): when the soft palate is lowered to allow air to pass through the nose as a vowel is produced, it changes a vowel to a nasal vowel and this is a distinguishing feature -- but not in English. So in a word like bank, the a is nasal because it is following by a "ng" sound, and if you ask English speakers what vowel is it, they'll say it's an a, while for a speaker of Portuguese, it's a nasal vowel, not an "a". Another example is how English t, p and k sounds are aspirated (there is a puff of air) at the beginning of words or between vowels, but not after an s, for instance. English speakers usually have no idea about this, but in other languages, such as Korean, aspirated and unaspirated t's are two very different consonants. So if you want some insight into how it feels to not tell l and r apart, ask yourself why you, as an English speaker (if you are), can't distinguish nasality or aspiration.
So the first problem is that Japanese people have a hard time hearing whether an l or an r is being produced because they are two possible realizations of the single phoneme to them, and consequently, they have a very difficult time remembering which is which. You'll find the same thing with z/dz. Ask a Japanese if you should say zettai (with a z) and they'll say yes. Ask again if it's dzettai (with dz) and they'll also say yes. Ask which they prefer and they'll tell you they only heard one pronunciation, even though you clearly said z in one case and dz in the other.
To make matters worse, they start out learning English using the wrong kind of flapped r (as in Spanish or Italian), while the English r is not flapped, but is rather a retroflex r where the tongue is curled back and doesn't touch any part of the mouth. To make matters even worse, the way English words are borrowed into Japanese interferes with their perception of English words, ie. they are spelled in a way that doesn't distinguish l and r (love=rabu), final r's are turned into a's (computer=kompyuutaa), etc. They really need to stop using katakana transcriptions. Note also that the English r involves lip-rounding, but since this isn't used in Japanese -- and they start learning to pronounce English without ever using lip-rounding -- they fail to see this as an important feature of English r.
The second problem is that even if they remember which is which, production is very difficult because the brain is used to treating both sounds as one, and they usually spent a fair bit of their lives as learners of English in a context where you simply said l and r the same and that's how everyone did it. It's a bit like relearning to walk after an injury.
So, while it's true that some Japanese people have a difficult time saying the English r, most can usually make l and r in isolation once they've understood how the sounds work. The biggest problem is that they have a hard time telling l and r apart when hearing it in speech, remembering whether a given word actually has an l or an r (and they often misspell words containing these sounds), and then producing either sound with regularity is ultimately the most difficult exercise.
One final note: to make it easier to explain, I left out some details. For instance, the English l is sometimes curled and English l and r are sometimes devoiced (compare ray and pray). These are small details that also make the sounds harder to produce accurately.
Same reason English native speakers have problems with the Spanish /r/ (both versions) They don't exist in the language. The Japanese ら、れ、etc is a combination of /r/ and /l/ (it's actually quite similar to the soft Spanish /r/) It is possible for Japanese people who have been studying the language for a long time and have traveled abroad to pronounce them correctly, but for a beginner, or somebody who just studied the basics (in a very deficient way) at high school, it's probably difficult. Same happens with English natives learning Spanish, and vice versa. And obviously, it's also difficult for us to pronounce correctly the Japanese ら、れ、etc (among others)
The problem with L and R is that niether of them exist in Japanese. The Japanese sound is more of a cross between the English R and L, so it's very difficult to distinguish the two, hence Engrish. A proper hard R is actually just as difficult to pronounce as an L for Japanese speakers, and the hardest words to pronounce are those with both sounds (for example, parallel).
It might be hard to understand from an English speaker's viewpoint, because the difference is as clear as day, but go ahead and try to distinguish the Chinese X and SH
There is a very good article in the Scientific American Mind magazine: http://www.scientificamerican.com/magazine/mind/2014/07-01/ (pages 24-25).
The authors also talk about an experiment that was developed at Carnegie Melon University, which teaches native Japanese speakers to learn the difference between L and R sounds with reinforcement learning.
There is also a free app which utilses the same reinforcement learning technique discovered by the scientist at Carnegie Melon University at: http://www.lirril.co
Because [l] simply isn't in the native phonetic inventory of Japanese. The lack of phones in your native phonetic inventory is the reason for most pronunciation difficulties.
Just to add an observation, though:
When trying to explain the [l] sound to native Japanese speakers, I usually tell them that it's like an [n] sound, except you should point your tongue a bit to allow air to escape on both sides of it. But it seems that this can be harder than it sounds. And thinking about it, no Japanese sound requires you to point your tongue. Doing things with your tongue you're not used to requires muscle training, and native Japanese speakers would probably lack this training.
The answer here summarises my feelings on the matter as someone who's grown up with both languages:
I'd strenuate the use of the English "L" noise when attempting Japanese words, I've never really understood why learning systems adopted "R's" when "L's" seem closer in sound.