 I want to talk about where to put the money. I want to talk about what are the best investments in biosecurity? Where do we get the best rate of return? Across which biosecurity measures? Pre-border, border, post-border. Local surveillance, containment, eradication. Where do we spread the money across those different measures? Across plants or animals or both? What plants? What animals? What do we do? How do we find out how to best use the dollars we have to handle a biosecurity issue to appropriately do a biosecurity measure with the best possible investment? And I want to make three points, just three points in this presentation. And the first one, something I feel strongly about lately, the first one is that in terms of allocating monies across biosecurity measures or different kinds of invasives or different kinds of established pests, benefit cost analysis is totally useless. It's of no value. Strong, tough talk from an economist. Very tough. Let me tell you why. Well, you know, this is standard in biosecurity. It's standard, in fact, across all kinds of government levels. There's a, for an example, a proposal to put a thousand new traps into Queensland to try and detect early the establishment of some exotic fruit fly to protect that wonderful horticultural industry that's in Queensland. Someone comes up with a cost of putting a thousand traps in place for early detection and an economist is called in to do the benefit. And the benefit is usually not that hard to calculate in many cases but can be difficult, particularly when the environment's involved but for an industry case relatively more so or more straightforward. The benefit is in terms of avoided losses. Right if you catch fruit flies early, you protect that industry from having significant damages. And you do a ratio of benefits to costs and it's always wonderful and positive. I've never, my 14 years of doing biosecurity, I've never seen a benefit cost ratio that's anything but positive. And on that basis, people invest in that activity, those thousand traps. But I want to convince you today that that's the wrong way to look at it. With that, there's also another practice that's very common and you'll see it a lot in ecological literature. Even in big universities here in Australia, they'll rank plan protection interventions, for example, by benefit cost ratios. You know, 20 weeds, for example. You rank the benefits relative to the costs across those 20 weeds and you start spending money on the one that has the highest benefit cost ratio until that one's serviced and you move down the line until the budget's exhausted. That's also totally wrong. It's a wrong thing to do. And the simple reason, I'm going to show you a little example in a minute, but the simple reason is that benefits and costs are not independent of scale of what you're doing in the activity. And this matters. In fact, it tells you that you can't compare benefit cost ratios across different kinds of things you care about unless the scale is comparable, which it never is. Go back to the case of a thousand traps in Queensland. The best way to approach that in terms of spending money is not to ask what the benefit cost ratio is. The best way to approach that is to ask what are the extra benefits from having additional traps to detect something early. What you care about, in other words, is the rate of return on activity, not the absolute benefit cost ratio. And for an economist, this is in a sense a pretty straightforward proposition. You want to put money where rates of return are largest and keep investing until rates of return, in fact, are equalized across all the activities you care about. So, for example, for those thousand traps to try and detect a fruit fly, what question would you ask? You'd say, well, if I put in not a thousand but two thousand, what's the extra cost of doing that? And then I'd compare it to the extra benefit, and what would that be? Well, with the extra traps, you would detect much earlier. And if you detect earlier, you can contain and hopefully eradicate before, in fact, causes trouble. You'd look at the extra benefits relative to the extra costs. And indeed, there's a natural trade-off here, and this is what I want you to understand. Natural trade-off for an economist. If I put in an extra thousand traps, I go to two thousand traps, I know I'll detect something quicker. Trap grids will get smaller, and I'll find flies quicker. But the cost of the program itself is now more expensive, because now I have two thousand traps instead of zero or a thousand traps. So you can detect early, but it's costly. If I take it from a thousand traps to five hundred traps, you won't detect as quickly, and the program itself is lower. And that's the trade-off the economist wants to exploit. That's what's meant by pursuing a rate of return rather than a benefit-cost ratio. Let me make point two. Rather than looking at benefit-cost, you should look at a portfolio rule that asks the question, where's the best rate of return if you can calculate it? In many cases you can. I'll show you real numbers in a minute. You want to ask yourself, what's the extra benefit relative to the extra cost? What are the ratio of those two values? Now, uncertainty matters. These are risk-adjusted. Rates of return have to account for all kinds of probability effects, but still the basic principle is there. And indeed, the portfolio approach tells you that if you pursue rates of return, in fact you'll always do better in terms of overall benefit-cost ratios. And again, I'll show you an example. One possible glitch in all of that is that the size of the budget matters. In fact, I'll show you that as my third point later. If you have a smaller or larger budget, the rankings in fact get much more severe. But here's the example. I just made this up. These are cooked numbers. These are just imaginary numbers. You're looking across different kinds of activities. Here's the portfolio. You're looking across prevention, surveillance, eradication, management of existing things and other kinds of management. Those are the biosecurity measures you care about at the moment. And I just put down a budget share arbitrarily of 10, 10, 10 and 60. In fact, I've specified benefit-cost ratios in a way that's consistent with my own priors. I think that's always better to invest more in prevention and surveillance than in management for very good reasons we can talk about. So I've specified some arbitrary benefit-cost ratios, and you can see prevention comes out best and they're ordered in decreasing range. So a standard benefit-cost approach would say, well, let's put all of our money into prevention until we exhaust that amount, that cost, and then move to surveillance, then eradication, and so on. You'd move down the order in terms of benefit-cost ratios. But instead, if you look at rates of return, you get a different outcome. You can see the way I've cooked the numbers. The rate of return on surveillance is higher than it is on prevention. Next slide. So here's the answer. If you were going to allocate appropriately, according to that portfolio allocation rule, in fact, you would have more or less equal shares in prevention and surveillance. You wouldn't dump everything into prevention. You'd find equal shares, and indeed, the eradication share, in fact, would still be fairly large, even though its original benefit-cost ratio is small. Much different than just allocating by benefits and costs. And ultimately, if you do it that way, as I said, the overall benefit-cost ratio, this value is in fact largest. And that's what causes confusion. The best way to get the overall benefit-cost ratio to be as large as possible is not to look at benefit-cost ratios, but to look at extra benefits relative to extra costs and always invest in those things that have the highest rate of return. Not just in biosecurity, but I think across governments. Certainly true in my home country, where everything has to have a benefit-cost ratio before you can move. Well, that really leads to a lot of sort of misallocations that are substantial. And indeed, in this little example, you can see the shares are in fact much different than what you would have imagined by just using that simple benefit-cost rule. Now, can this be implemented practically? Well, we did a pilot study. Sebber did a pilot study for the National Biosecurity Committee and also to assist the department. We did a pilot study to see whether or not we could allocate monies. We started with $26 million and asked the question, could we allocate monies in a way that was appropriate across different measures and different kinds of invasives or pests? And the ones that we started with, it's a pilot study, so we're just taking a small set. We started by looking at things we've been doing for decades anyway, hawkweed, fruit fly, papaya fruit fly, foot and mouth disease, and red imported fire end. In fact, let me be more specific here. We look at the case of how would you invest across biosecurity measures to control hawkweed in Victoria in terms of a surveillance eradication campaign. For foot and mouth disease, we asked the question, is it appropriate to do active surveillance on top of the existing passive surveillance system? And we used bulk milk testing as a protocol, trying to find an FMD virus early in a dairy herd. For red imported fire ends, we looked at the Queensland case. Foot and mouth disease was also in Victoria, by the way. For Arefa, we looked at Queensland and looked at its surveillance effort. And for papaya fruit fly, we looked at the tourist straight, what Gary was just talking about. Do we have enough traps in the tourist straight and in Northern Queensland to assure, ensure that we're in fact acting optimally, to in fact maximize the rate of return? And we looked across these four activities, eradication and containment, active surveillance, three activities, and prevention in this exercise, eradication and containment, we're lumped together. So there's some sophisticated modeling behind the scenes here. There are spread density models and all kinds of parameter values that go with this exercise to see whether or not you could find something optimal, the optimal number of traps to find fruit flies without spending too much on a trapping program to find them early and protect the, you know, the fruit industry. Is there an amount that should be spent to further contain and eradicate rifa? Should we be looking for hawkweed more aggressively in Victoria? Should we be doing an active surveillance regime for foot and mouth disease? And indeed, if you just did benefit cost ratios, FMD is always the winner, always will have the highest benefit to cost ratio. Here are the results, their mean value, pilot study and this is what you get. Just talk about them for a minute. The dollar amounts were spending 26 million on top of the existing system. So there's already passes surveillance in place, there's border protection, there's all kinds of things in place. What I want you just to see is not so much to focus on the dollar amounts, but to see that in fact the optimal spread of monies in fact is quite varied. It's a cross-prevention surveillance and eradication. It's not lumped into one thing or another. With FMD, quite a bit in prevention, that's border measures to prevent FMD, of course, from arriving in Australia. But nothing on active surveillance which surprises my animal health friends which I'm sympathetic toward. It tells you that the cost of bulk milk testing is too large relative to all the other parameter values that are of interest. Given the cost of the test, $32 per test, the arrival rate which is assumed to be one in 50 years, which may be too large but the arrival rate of one in 50 years and all the other things that go with a wonderful passive surveillance system behind animal health, it doesn't pay to spend on active surveillance. It does pay, if you look at the 282, it does pay to hold into an account and a dollar amount each year. It's actually quite a bit of money in the event there is an outbreak. That's what the economics would tell you. It's a good return. REFA, you can see, there's still a good deal to invest in REFA and Fruit Fly that's good news for Sarah and Fruit Fly in this exercise in fact gets a healthy dose of prevention and active surveillance. Indeed, if you drill down and just look at the surveillance exercise on Fruit Fly in Queensland, it's badly under invested according to this work. We should have many more traps, try and catch things much earlier in the Torres Strait or in Northern Queensland to try and detect those flies early. You get a real mix of results here that cost-benefit analysis in fact couldn't even come close to. Final point, and then I think I'm done. Final point, what happens if you have a fixed budget? Instead of 26 million, you have something less. Well, the rates of return here in fact vary considerably. You don't pick up FMD prevention in fact until quite late in the profile until about 7 million dollars are being spent. First expenditures go on things like Fruit Fly in particular which have the highest rate of return in this approach. So when the budget's constrained it's almost always constrained in fact the choices become very severe. Some things are out and some are in and where you put the first dollar matters and that's the basic bottom line here. This is a principle that tells you don't look at absolute amounts of benefits to cost but where the extra dollar is going. If you're spending an extra dollar and the extra benefit that goes with that expenditure is greater it's a good investment. It's a good rate of return and that's how you allocate. Difficult to do but conceivable in principle you can do it. You can start to think more about how these things might play out and indeed if you look at this generally we've been looking at it for a while. It's quite clear that prevention and surveillance become much more important than you would imagine. The under investment in Fruit Fly traps in Queensland is a good example but it's better to catch things early in many cases without spending too much on trying to catch things early because there is that trade-off than it is to manage once they're here. That's the economics of biosecurity. Benefit cost ratios awful. Keep that in mind. Thank you very much.