The system for taking account of goods in transit is necessary for industry supply to work properly. The idea is that an industry will not demand that more goods be put in transit if, having regard to the goods already in transit and when they are likely to arrive, the industry will be adequately stocked.
For example, if an industry demands 10 units of cargo per game month, and units of cargo take one half of a game month to arrive, then an industry will demand that more units be put into transit when it has 5 or fewer units left, assuming none are already in transit. If there are already 5 units in transit, the industry will only demand that more cargo be put in transit if there are fewer than 5 units in storage.
This system does not affect the overall demand of industry, or the profitability of lines (in the long term). What would happen without it (and what used to happen before it was implemented) was that an industry would keep demanding that goods be put in transit so long as its storage was not full. Lots of goods would then be put in transit, and, eventually, the first of them would arrive (all the time more and more being put in transit). When the input store would get full, no more goods would be demanded. By this stage, there might be a huge amount at stops or on vehicles at various stages along the route. Goods would accumulate eventually in the desintation stop and transfer over to the input store whenever the input store became empty, such that the industry would demand no more goods for a very long time. This would have the result that the player's network would end up being used extremely intensively for a short period, and then not at all for a long period. This effect would become more pronounced the longer that the lead time between supplier and consumer was, and could become extreme in cases with a long lead time (e.g., industries with thousands of times the input store accumulating in the destination station, not demanding any goods for many, many years).
Thus, without the system of taking account goods in transit, industry flows would be much more, rather than less, intermittent.
As to the starting capital, this is not yet balanced: what I am aiming for is a system in which players need to borrow money to start, so can borrow the right amount to make any given sort of network work effectively, providing that they can keep up the interest and capital repayments. However, there is a long queue of higher priority balancing tasks (and UI improvements, such as allowing manually selected prices to be automatically cloned and propagated, as well as bug fixes) that need to be done first. Anyone who would like to work on coding the cost of capital feature would be welcome to do so, as this would greatly speed the development to a more balanced state.