You need to remember if you do this then the block chain size will increase and so will the required bandwidth. So what will happen is that it will take more resources to secure the network by node operators. Eventually the resources will be too much so node operators would stop providing the service.
The block size and reward would be reduced by 2x (or multiplied by 0.5), and halving time amount of blocks (to halving still happen every 4 years, despise faster blocks), this to not increase the speed the blockchain increases and not change the reward dinamics.
you do know that average 10 minutes vs average 5 minutes makes no difference in the real world
people have already done loads of surveys,
the results are that whether a block is produced per 5 or 10 minutes, it doesnt then open up any new utility..
for instance physical retail checkouts will still see people waiting too long for a confirm before they can leave a store with their groceries that they still wont use bitcoin at a grocery supermarket checkout even with an average 5 minute confirm model
secondly without addressing the spam/bloat issues. it will still be the spam/bloaters getting first priority so genuine bitcoiners end up waiting for the next block anyway
thirdly although it does not harm the nodes in regards to block data propogation. by having blocks being hashed by mining asics of mining pools. those miners have half the time to hash, meaning it causes the hashrate to be halved. which is weaker security than a 10 minute block average.
also for pools to work making block templates after a solved block appears, there will be more instances of "empty block" due to how mining pools send out templates to miners in the initial few seconds of previous block solve and then slowly fill blocks per update request over the next few minutes. due to the fact if all pools are regimented to average 5minutes, it would cause more 'lucky' blocks found in less time. meaning more empty block chances
also due to the lower hashrate and difficulty due to hashes per 300sec instead of hashes per 600sec.. people end up instead of waiting 6 confirms for large amounts will wait 12 confirms for same large value amounts. and 2 confirms instead of 1 confirm for same smaller amounts.. due to security metrics of economics
thus your idea has no advantages