First off, senators were required to own land, hence they praised land highly. It was a cultural thing. The ideal Roman was a yeoman farmer, and senators were expected to be the ideal. Hence you get the legendary Cincinnatus who becomes dictator then returns to his farm immediately after handling the threat to Rome.

Wealth was kept all kinds of ways, but it's hard to talk about a typical Roman because we don't have enough information. Wealthy Romans tended to have tons of properties. They would have a house in Rome, a house just outside of Rome, and some other properties in Italy (Campania was popular) and perhaps the provinces. One Late Antique aristocrat, Melania the Younger had land in Italy, Spain, Africa, and the Levant. Apparently she caused a small financial panic when she tried to sell it, to give you an indication of her worth. She is the subject of a hagiography, the Life of Melania the Younger.

There certainly was a risk of theft. There is a Greek court case (4th c. BCE) where Lysias is prosecuting a thief who stole the gold from his house. A good way to prevent this was to by land or lend your money or invest in commerce. Of course, the richer you were, the more people you could have around your house to protect your money. There were essentially banks, they tend to be men instead of institutions. Crassus was basically a bank. There has been a lot of interest in Athenian lending that has revealed that Greeks were practically using fiat money (E. Cohen researches this). You could also store your money at a temple. They would give you a receipt and use the money in the meantime, just like a bank might. You could even use the receipt you got from that temple to get the money from someone else. Xenophon did this during his retreat through Persia. While This happened a lot in the Greek world, I don't know whether it happened in the Roman. I haven't seen any indications in the usual places.

Hope this helps.

I've answered this question before, so I'm just going to pull from there. http://www.reddit.com/r/AskHistorians/comments/2e0wt7/roman_historians_where_did_marcus_licinius/cjv70h3

If you're asking if Crassus had huge vaults filled with gold, the answer is probably no. Crassus's utilization of his money was largely investment, particularly in land, of which he owned enormous amounts of. The Roman banking system, even in this time, fairly advanced, and should he ever need to have cash on hand in significant quantities he could have obtained it. However, gold does nothing really in the Ancient World to increase wealth; it is a commodity which stores value, nothing more. You want to turn that into something which makes you more money. In the ancient world, that was agricultural land and the money-lending business. In terms of large cash transactions, these were going to be done on paper; basically a money order. Those times when he needed to actually access a large stack of currency (such as when he had to raise, outfit, and pay his Legions for the attack on Parthia) we can assume that it would have come out of the accounts of his personal bankers that he had stored, or, more likely in my opinion, been loans against his assets. I would compare it to modern day 'Billionaires'. How many have stacks of hundreds of millions of dollars sitting around? Besides the obvious security risk, this currency does absolutely nothing for them in terms of creation of wealth, and if there was one thing Crassus was particularly enamored with, it was the creation of more wealth for himself. But I must be clear here: most large transactions in the Late Republic were not conducted with wagon loads of coin. If you wanted to buy a building, a parcel of land, or give someone a loan, almost the entirety of the transaction would consist of paper being exchanged between private bankers and public servants, all of which was meticulously recorded. If you purchased a plot of land, you would pay the previous owner with a slip from a trusted banker, who would 'transfer' the funds from the accounts of the purchaser to the accounts of the previous owner. When I discuss this, often times the question comes up of "What happens if everyone wanted to withdraw their accounts in hard currency" and the answer is that clearly most bankers weren't keeping millions of denarii on hand to pay out to their account holders, and point to the 1930s as a possible answer to that question. On the other hand, unlike then, the Roman banking system was largely dominated by the rich in terms of creditors, and not that many average Roman citizens were going to have significant amounts of cash deposited with any of these bankers to really make it that big of a risk, and the predispoition for Romans to invest cash into agricultural land was paramount. Think again back to modern rich; how many sit on millions in a bank rather than investing it? EDIT: If you're interested in some more information of Roman banking methods, this short article should give you some more: http://penelope.uchicago.edu/Thayer/E/Roman/Texts/secondary/SMIGRA*/Argentarii.html