As I mentioned, we have a prediction challenge every year in terms of how much money we're going to spend, because it's based on how many receipts we receive. For example, we are in the month of March right now. We're probably going to do about 35% to 40% of our expenditures for the year, and we will know between now and the early part of April how much money we're actually spending.
The good news is that our lapse as a department has reduced significantly. Last year, we lapsed $2.5 billion, and this year we are expected to lapse under $500 million. We're using predictive modelling to predict things, but for some major projects—for example in transit—we don't know when they're going to begin to send us requests for payment. They happened this year, and we've worked some of them in, which is why we're expecting to have about $10 billion in cash flow this year versus about $7 billion last year, but we won't know until probably January or February of next year the extent to which we are getting those requests to come forward.
This goes back to the point I made around programs. In terms of program money, you have to show the money in a certain year. If you don't spend it that year, then you go through re-profile requests to move the money to future years. It's not lost when it's not serving the need; it's simply an accounting treatment below the surface around when the money shows up in the public accounts. We do our best to predict it, but we don't control it. That is the key message that I would give.
Going back to Build Canada Homes, it's why the statutory authority works so well. CIB, you will note, doesn't lapse anything because CIB has a statutory allocation of money. If they don't spend the money this year, then it simply stays in their statutory envelope and they use it next year. For infrastructure, it's a much better way to account for the dollars, and it's why we've done that for Build Canada Homes and given them that same kind of capacity.