Jordan Wainer: The Federal Highway Administration's Center for Innovative Finance Support. I'd like to welcome everyone to the fourth webinar in the P3 Value 2.0 Webinar Series. Today's webinar is on P3 project risk assessment. My name is Jordan Wainer. I am with the U.S. DOT's Volpe Center in Cambridge, Massachusetts, and today I will be moderating the webinar and facilitating our question and answer periods. I'd like to point out a few key features of our webinar room before we begin. On the top left side of your screen, you will find the audio call in information. If you are disconnected from our webinar at any time, please use that call in information to reconnect to our audio. Below the audio information is a list of attendees. Below the list of attendees is a box entitled, "Materials for Download," where you may access a PDF copy of today's presentation, as well as the exercise for next week. Simply select the file you want to download. Click download file, and follow the prompts on your screen. In the lower left corner is a chat box where you can submit questions to our presenters throughout the webinar. Our webinar is scheduled to run until 3:30 p.m. Eastern today, and we're recording today's webinar so that anyone unable to join us may review the material at a later time. All lines are muted right now, but we'll provide instructions to ask questions over the phone as time allows. And with that, I will turn it over to Patrick DeCorla-Souza. Patrick.
Patrick DeCorla-Souza: Thanks, Jordan, and welcome all of you to today's webinar, which is one of a series of webinars that we have been presenting on our new tool P3 value 2.0. Your instructors today are myself, Patrick DeCorla-Souza. I am the P3 Program Manager in FHWA's Center for Innovative Finance Support. And with me to present this webinar is Marcel Ham from IMG Rebel. And with him will be Wim Verdouw, who is the developer of this tool. Just to go over a few basic definitions, we've presented this in past webinars, but P3s stand for public-private partnerships. P3 value 2.0 is an educational tool that we've developed to help folks understand the value for money analysis. And listed below, you see all of the webinars that we have presented so far, and they are all available on our website. The recordings can be listened to at your convenience. Today we will be presenting the risk valuation or risk assessment webinar, and two weeks from today, we will present the financial viability assessment webinar. Today's webinar, we've got a lot of topics to cover, as you see from this list. First, we will talk about different types of risks that you will need to consider in a P3 project. We will then show you how risk assessment fits into the overall risk management process. Then we will get into the meat of this webinar, which is risk valuation. So we will show you how to-- how P3 value assesses risk. And in part three, we will talk about pure risk. And part four, we will talk about life cycle performance risk and revenue uncertainty. In part five, we will talk about how these risks can be transferred. What principles can be used to transfer these risks to the private sector? And then we will demonstrate the tool in part six. So hopefully, after this webinar, you will understand how to categorize P3 risks. You will understand the steps in the risk management process and where risk assessment fits in. You will understand the methods used to quantify and monetize risks or value for money analysis, as well as other types of analyses you might be doing. You will understand the risk allocation process. And finally, you will have an introduction to P3 value 2.0, and you should feel comfortable opening the tool and using it to do risk assessment. Let's first talk about categorizing P3 risks. Why would you want to do risk assessment? Now in the context of P3 value, the main reason we want to do risk assessment is we want to calculate the value of risks. That's an essential component of costs in your value for money analysis, and also in benefit cost analysis. And as we will show you in two weeks, it's also an essential component of financial viability analysis. Now you also want to do risk assessment in developing your draft RFP, the request for proposals for P3s, because you want to understand the magnitude of those risks and which risks might be better managed by a concessionaire. Risk assessment also helps you understand what kinds of prices might be appropriate in discussions with bidders where you have to negotiate who and how much will be paid to transfer certain risks. And finally, risk assessment helps you develop risk management plans. Risk assessment tells you which risks are the most important to consider, and which risks you need to pay attention to in order to address cost issues and delay issues.
Patrick DeCorla-Souza: So the impacts of risks can be of three types. First, costs are affected. As we said, risks can increase costs, and so you want to be aware of these impacts of risks on costs. They also have impacts on delays. Delays can increase costs by causing additional costs as well as inflation. And they have an effect of loss of revenue if the completion of the construction is delayed and the project is expecting total revenue, you would have an impact on revenues. Also, there is what is called a net present value effect. So when you delay a project's cost, actually the present value is reduced. And that affects your value for money analysis. Revenues, of course we said delays can affect the receipt of revenues. And there's other factors, other kinds of risks that affect revenues, you know, and these might be positive or negative. For example, a development that you were expecting, if it doesn't happen, you might face lower travel demand. On the other hand, if more than expected development occurs, you might get higher travel than you were expecting, and that would increase your revenue. So there can be both positive and negative aspects to risks, and the positive risks are called opportunities. So with regard to categories of risks, this slide shows you the different types of risks and helps us understand which types of risks are important in evaluation of P3s and in assessing P3 financial viability. It's important because some of these risks do not affect the actual P3 project financing. For example, right below on the second level, you see exogenous risks versus endogenous risks. Endogenous risks are risks related to the decision makers and perhaps, you know, them changing the scope of the project. All of these could happen before the project is actually let out for bid. But it's important in an overall project context that you be aware of these types of risks. Exogenous risks are those that are not within the control of the decision makers per se, and they might be divided into what we call process risks. For example, your National and Environmental Policy Act process, there are many decisions there and risks that could affect the scope of your project. But by the time we actually get to financial close, the only risks that we are concerned about in a P3 are what we call project risks. So we assume at this point at financial close, that's all the risk we are going to be concerned about, because that, those are the risks that the concessionaire is going to include in its estimate of the project costs in its bid. So the project risks are of two types. What are called systematic risks and non-systematic risks, and systematic risks are risks related to the economy at large. Things such as inflation, interest rates, growths in economic development. All of these are not necessarily within the control of the project managers. What they might be able to control are what are called non-systematic risks. And these non-systematic risks are of two types. First, pure risks which are sometimes called event risks. So they are risks that something may or may not happen. Whereas regular uncertainties are risks related to the estimation of pricing in volumes of material that affect your costs. And we'll go into more detail on all of these as we go along. So first, base variability or uncertainty in volume or pricing. That is something that, you know, we know we are going to need asphalt. We just don't know how much and what the price might be. And so that risk is called base cost variability or base variability. Pure risks are risks that may or may not happen. So an accident may or may not happen at the construction site, but there is a possibility that it might happen, and if it does happen, we would have impacts on costs and delays. The third category is life cycle performance risks. And these are over the entire life of the project, something that in conventional delivery, we don't-- we are not normally concerned with the project life, we are only concerned with risks up to project completion or completion of construction. So when you are talking about P3s, the concessionaire is responsible for the project for maybe 30, 40, 50 years, and is going to bear risks that occur over that entire period of time, such as coordinating contractors, events that might take place, inflation, and revenue risks, which is subject to, for example, travel demand that may or may not occur, depending on whether the economy of the particular metropolitan area or a region is positive or negative.
Patrick DeCorla-Souza: So I will just talk a little bit about valuing base variability. And we've got entire sections later on where we will focus on pure risks and life cycle performance risks. So base variability is as we said, it can be above, or the actual cost can be above or below what you estimated. Now P3 value accounts for this base variability by using a percentage, and it is a positive percentage. And the reason is that we don't necessarily consider averages when you're looking at financial viability, for example, or setting aside sufficient funds in our budget. We try to look at something that might be what we call a P70 value. And a P70 value simply means that 70 percent of the time, our cost estimate will be at or below the P70 value. So that still means that 30 percent of the time, we run the risk of the costs exceeding P70. So the percentage that we provide in P3 value reflects basically the fact that if you want to be fairly confident, that is at the P70 level, that we will have sufficient funds, we need to increase our cost estimate by that percentage. So you will, for example, put in a value, whatever you think the P70 value might be above a P50 value. In this case, in the example it shows you 20 percent. If the base construction cost is $200 million, you multiply that by 20 percent, and you find the increment of cost that you need to put into your cost estimate of $40 million simply by multiplying 20 percent by $200 million. This slide shows you the difference between the life cycle performance risk and the pure risks and base variability. Below the line, you see that the SPV or the concessionaire employs various types of subcontractors for design build, for maintenance, for operating the facility, and tries to push down as much of the risk as possible to the subcontractors. And the risks that they are able to push down are the pure risks and the base variability risks, because those are risks that those contractors are most capable and you know, most able to manage. Above the dotted line or dashed line, you have the remaining risks that are not pushed down to contractors-- subcontractors. And those risk are what we call life cycle performance risks, inflation, interest rate, performance of the facility itself, integration of the subcontractors, integration of phases, and of course, toll revenue risks, which is, it tends to be one of the biggest risks. And who holds these risks are the financiers, the equity providers and the lenders, bond holders, banks, et cetera. And the way they get reimbursed for those risks is through the rates of return that they command. For example, the equity rate of return might be, let's say, 12 to 15 percent. The higher percent of return relative to a risk-free rate, which might be, for example, the rate that the public agency might be able to borrow funds. That difference reflects basically these systematic and, you know, risks which are long-term performance risks and revenue risks. Similarly, you have the lenders and they may require a higher interest rate if they perceive more risks. And that premium that they require over and above the risk-free rate reflects the life cycle performance risks or revenue risks that they see in the project.
Patrick DeCorla-Souza: This slide shows you that depending on where you are in the project development, the percentage of set aside or risks can be higher or lower. So early in the project you don't have a lot of information and so you do need to set aside much higher amounts to account for base variability and pure risks. And as it shows you here, there are some risks that you can identify and some pure risks that you just aren't able to identify and you have to account for those. And those are all accounted for in contingencies, whereas base variability is accounted for through allowances. This slide shows you the same type of situation, but instead of just looking at the project development phase that simply stands development right through construction, we are now looking at the entire project's life. So what you see is yes, there is a drop in risk until the project is completed, but you still have a lot of risk. Particularly on a toll project, what you see here is the ramp up phase. And until that phase is completed, which can take several years, there is still a lot of risk that revenues may not be what we had anticipated. Now after the traffic stabilizes at this point on, we are fairly certain about the receipt of revenues, and there's a lot less risk. It's important to understand this risk profile because depending on when you come in as an investor, so initially the developers require really high rate of return because they have a high level of risk. If I'm a pension fund and I don't like to hold a lot of risk but I am willing to take less risk, I might come in after the project revenues have stabilized, and so since the risk is lower, I will also be getting a much lower rate of return on my investment.
Patrick DeCorla-Souza: So we are ready to go to the question.
Jordan Wainer: Okay. The first question is, all project risks are pushed down to subcontractors of this concessionaire, true or false?
Jordan Wainer: And Patrick, we had a request from someone in the audience for you to speak a little louder if possible.
Patrick DeCorla-Souza: Okay.
Jordan Wainer: So I'll give everyone just a couple more seconds to answer.
Jordan Wainer: Okay. So we-- everyone said false.
Patrick DeCorla-Souza: Yeah, and that's true, because as we indicated, there are life cycle performance risks and revenue uncertainty, which are held by the concessionaire.
Patrick DeCorla-Souza: We're ready to-- Are there any questions at this point?
Jordan Wainer: There are no questions in the Chat pod at this time. As a reminder, you can submit questions through the Chat pod at any time throughout the webinar.
Patrick DeCorla-Souza: Okay. So let's move on. So we're ready to talk about where risk assessment fits in the overall risk management process. And this graphic shows you the risk management process. Basically, you identify risks, then you assess them, then you try to figure out how to manage them or how do you respond to these risks. And one of the ways you might respond is actually allocating that risk to a P3 concessionaire, so that's in step four. And then step five is you monitor all of the risks to make sure that, you know, that your management is working and that you've identified all the risks and no more further action needs to be taken. So let's talk about each step now. And risk identification, as we said, there can be two types of risks, either threats, which means an increase in costs or loss of revenue, and opportunities, which could be reduction in cost of an increase in revenue. And how do you identify these risks? Normally we do that through risk workshops. And you hire a facilitator who brings in subject matter experts. They look at the project, try to identify where the risks might be, and probably start with an existing previously implemented projects where all the risks have already been listed, starting from their then build up and try to see which of those risks apply to the particular project they are looking at. So they might use a checklist, for example, and then based on that, develop a risk register where they put in a lot of information about each risk, each of the risks that they identify. But in the risk assessment phase, now we are trying to qualitatively first assess each risk. How probable is the risk and what are the potential consequences? And we look and try to figure out a scale. Is it low, medium or high? Probability. Is it low, medium or high impact on cost or a schedule, et cetera? So we first do a qualitative information. After we've done-- I mean, a qualitative assessment. After we've done that, we can then go ahead and use that information to actually quantify and monetize each risk. So the probability and consequence information can help us monetize or put value on those risks.
Patrick DeCorla-Souza: In the risk response planning phase, we try to figure out how to minimize the impact of the risk should it occur, or how to reduce the possibility that the risk might actually occur. So there are several strategies. You try to avoid the risk. You figure out strategies to mitigate the risk if it does occur. One response strategy is to try to transfer it to someone else that can manage it better. And in this case, you know, it might be a design build subcontractor or a contractor or the concessionaire. And finally, if no party is able to either reduce the possibility of occurrence of the risk, or able to reduce its impact should it occur, then you simply have to accept the risk so that it could be a strategy. Risk allocation, and we'll go through this in a separate session in this webinar, involves figuring out which party is best able to manage the risk, and then accordingly, either transferring it to the concessionaire or retaining it within the agency. Sometimes an option might be to share the risk if neither party is able to manage that risk. Risk monitoring is trying to constantly evaluate to see how we are doing with regards to management of these risks. And an important thing is to understand the P3 agreement. Because in the P3 agreement, you have provisions where you have either transferred a specific risk or the agency has retained it. So it's important to understand that, because if you don't, you run the danger of taking back a transferred risk by getting involved in a particular risk. And finally, of course, by monitoring risks, you might find out you have not identified all of the risks, and you might have to re-go through the cycle again to develop and update your risk management plans and response strategies.
Patrick DeCorla-Souza: Are we ready for the next question?
Jordan Wainer: So the next question is, the probability and potential consequences of a specific risk must be quantified in order to estimate its cost impact. True or false?
Jordan Wainer: I'll give everyone just a couple more seconds to answer.
Jordan Wainer: Okay. We have 77 percent of people say true and 23 percent say false.
Patrick DeCorla Souza: Well, it is true because you do need information on probability and potential consequence or impact in order to be able to identify risk. So can we go to the questions and see if there are any questions.
Jordan Wainer: There are no questions in the Chat pod right now.
Patrick DeCorla Souza: Okay. Let's move on. Part three is pure risk assessment. And as I indicated, there were three risk categories. Base variability, pure risks and life cycle performance risks. So in this part, we are going to talk about pure risks.
Patrick DeCorla-Souza: The first step is to do a qualitative assessment, and in this type of assessment, you're just basically trying to rate risks between low to high. So in this slide, you see those risks that are rated either very low or low. You can, if they are very insignificant, you can discard them and not worry about actually quantifying and valuing them. The others, you might aggregate into a combined risk and try to figure out a value to assess to those risks. The remaining risks are medium, high or very high in rating, and those are the ones you would want to do a quantitative assessment. And in that quantitative assessment, you are looking at the probability of that each of the risks occurring, and then considering the potential impacts if the risks do occur. And by using some kind of either a formula-based calculation, or a Monte Carlo simulation, you then try to figure out what the aggregate impact of all of these risks. The medium to high risk, the aggregate risk, they're all combined and you come up with an aggregate cost, aggregate delay or aggregate revenue impact. So we will show you how P3 value goes about doing the quantitative step. But first, let's talk about the qualitative step. And here you see in the qualitative valuation, you simply rate the possibility of the risk occurring from very low to very high, and you put it into a table. Now in this table or matrix, what you see is the probability scale is shown here, so above 70 percent probability and from 0 to 5 is the lowest probability scale. And along the X axis you have the cost consequence. So more than 25 percent impact on cost versus less than 1 percent. So you can categorize each risk in one of these cells based on their probability of occurrence and the scale of impact. And what you see is you can then look at the ones that are really high in both respects, which is all of the red ones, red cells. Their probability is high as well as their impact is high, so those are the risks we need to pay a lot of attention to. And then you have somewhat lower probabilities of occurrence with somewhat lower impact and those are rated medium. And then the low category are ones that you can either ignore or aggregate into a combined risk. And you can do that also for schedule. And this is you can see here, this is schedule delay. And in P3 value, we basically combine the schedule impacts into the cost impacts, so you don't do a delay impact separately simply to simplify the process. So in quantitative assessment, we now actually come up with an actual percentage. You know, 35 percent, 40 percent, 45 percent. So we have to actually come up with an actual probability percentage. We have to come up with the dollar amount instead of just saying percent above and below. You know, cost overrun or delay. We have to come up with the number of days. So much more detailed evaluation in quantitative assessment. But just doing that is not enough. To get the value of the risk, we have to do some calculations. And here in P3 value, we show you one way of doing that. Some very simple methods simply take the probability and multiply it by the impact and come up with a value. In P3 value, we're a little more sophisticated in our formula. We use a distribution because we figure you may not have, you know, a single value for the impact. You might have a range. And so that range might have a distribution, so we consider that. So in the case of a uniform distribution, basically a uniform distribution simply means that all the probabilities are equal. So it's equally probable that if the risk occurs, you might have an impact of $2 million dollars, as well as its equally probable that you might have an impact of $4 million dollars. So in order to value the risk, you simply take the percentage probability, so in this case it was 10 percent, and we multiply it by the average of these two extremes, $2 million and $4 million. So the formula is simply 2 plus 4 divided by 2, which is what this half is here, and we come up with the value of the risk. And so that's the way P3 value handles a pure risk, which has a uniform distribution.
Patrick DeCorla-Souza: In the case of a triangular distribution, which has a little peak, and here you see, you know, so it's a much lower probability at the two extremes, but a higher probability in the center. And it can be a little bit skewed where your highest probability might be skewed to the right or to the left. So what P3 value does is looks at the minimum impact, looks at the maximum impact, and then looks at the most likely impact, which is this peak over here, which is a, you know, occurring, let's say at 3.5. Let's say that would be 2.0 here, that might be 4, and this is 3.5 occurring here. And then uses a formula that takes into consideration this higher peaking of the probability. And so it takes 2 plus 3.5 plus 4, divides it by 3 to come up with the average. And then again multiplies it by 10 percent. So what you have that's different for the same numbers, 2 and 4 at the extremes, is your value is a little higher simply because of this skewed distribution that you get with this, at least in this case.
Patrick DeCorla-Souza: So now if you do that for each individual risk, you are going to get a most likely value. Now how can we put it all together to aggregate all of these pure risks to come up with a total? So we use something called central limit theorem. And you know, another way to do this is to do a Monte Carlo simulation by throwing all of those risks into a Monte Carlo simulation program. But in P3 value, we do everything using a formula, and the reason is we are able to reduce the run time of the model by using a formula-based approach. So we use a central limit theorem, and I'll show you how P3 value does the calculations. So here is the risk one as a probability of 20 percent. The most likely impact is $10 million dollars. So you simply multiply probability by most likely and come up with a most likely value. And that's the simplest of approach. P3 value goes a little more by looking at the distribution and says that may not necessarily be the only value, because it could be 20 percent lower or it could be 50 percent higher. And these are, of course, information you would get from your subject matter experts. So with this information now about the minimum and maximum value, we can calculate what those values might be. So 20 percent lower would give us $1.6 million dollars, and 50 percent higher gives us $3 million dollars. So now we are in a position to calculate a different mean value, which takes into consideration this range over here. And the range is, since this is a uniform distribution, it's simply averaging these two values here, which gives us, you know, exactly in the middle, you add $1.6 to $3 million and divide it by 2 and that's what you get here. So if you add up-- Now we've simply shown here uniform distributions for all of the risks-- they are all uniform-- to keep it simple so you can use this simple formula. And we, you know, we can simply add all of these and we get the mean value. Now, you know, as I indicated earlier, you don't want to simply use a mean value in your estimate because that really means that about half the time, your estimate is going to be wrong. You know, too low. So you want to have a cushion or instead of a P50 value, you want to have a P70. And so you can do that. You can get a P70 value. You have to go into a little more complication here by calculating a variance. And, you know, our P3 value concept guide shows you how to calculate these. But here's the simple formula for a uniform distribution. So you can use that formula to calculate the variance. And then you add up all the variances of all of the risks, and you get a total variance. And then by taking the square root of that total variance, you get a standard deviation. Now by adding the standard or the deviation-- by using the standard deviation within this formula in Excel you can calculate a P70 value. And so now that we know the distribution, because of the standard deviation, we can simply calculate. Use the norm INV function in P3 value, and it's an Excel function. And we come up with this extra amount that we would need to set aside for these risks. So it's instead of a mean value, we now have a much more sophisticated P70 value that we can use in our calculations. Monte Carlo simulation does the same thing, but it uses simulation, so it takes all of the risks, and each of the risks has a probability distribution. Monte Carlo simulation simply chooses from each risk for each run of the model. It chooses a particular value from the distribution. It looks at the possibility or the occurrence, the probability of occurrence. And does several runs to then come up with a distribution. So it comes up with an aggregate distribution like this, and a cumulative distribution. And if we want to know what the P70 value is, we can simply read off of it from this, looking at it on the left axis here, the Y axis, read and read down, and we know what the P70 value is that we need to consider. We're ready to go to the questions. I'm sorry, to the poll question.
Jordan Wainer: Okay. So the question is, the aggregate impact of pure risks on cost may be estimated either by using a formula-based method, or by using the Monte Carlo simulation, true or false?
Jordan Wainer: I'll give everyone just a couple more seconds to answer.
Jordan Wainer: Okay. And everyone said true.
Patrick DeCorla Souza: Yes, and that is correct. These are two different ways. Our P3 value 1.0 actually used the Monte Carlo simulation. And now we are using formula-based to speed up the computations in the model. All right. Are there any questions?
Jordan Wainer: There are no questions in the Chat pod right now.
Patrick DeCorla Souza: Okay. So we're ready to move on. Marcel.
Marcel Ham: Yes. Hi. In the next section, we are going to talk about the assessment of the life cycle performance risk and revenue uncertainty adjustments. Those are the categories of risks that we haven't discussed so far. And they're also typically a little more challenging to assess, and also to value. The reason being that in conventional contracts, long-term-- risks with a long-term nature are typically not considered. So it's sort of easy to overlook these risks, either on the cost side, the life cycle performance risk, or on the revenue side, the revenue uncertainty adjustment. The first-- the first one is the life cycle performance risks. And as Patrick already explained, the type of risks we are talking about are related to the coordination between different subcontractors. As Patrick showed this work chart for a typical P3 structure with the special purpose vehicle and the subcontractors. And so typically there's many subcontractors involved in maintenance and operations. Sometimes even multiple maintenance contracts over time. And so there can be coordination issues between contracts and also between contractors. And there can be interface issues. What we are talking about here is that we are, we want to understand what the risks associated with a long-term performance guarantee are, because that's what we are trying to do in a P3 contract, is to ensure long-term performance. And of course, we should be considering a similar approach in a conventional delivery, because also under conventional delivery, the agency wants to ensure a certain quality level over a longer period of time. Now there's all kind of risks related to ensuring that-- that longer term performance risk, and that's why we qualify or name this risk life cycle performance risk. We know that these risks exist, but typically they are not recognizable in conventional contract cost estimates. And sometimes, they are even very hard to find in budgets. Because these risk are typically something you run into and then deal with. Whereas now, under a P3 contract, we are forcing the bidder, and later the contractor, to value those risks over a longer period of time. Now there's in the value for money assessment, and also in the PDBCA. We recognize two ways to value those risks. The first one is in a cash flow. So what we're trying to-- what the even the P3 value 2.0 toolkit offers the option to enter and input that is going to be a flat fee over the life of the contract, representing the life cycle performance risk. The question of course is, where do I get the information to value those risks. And that is very challenging. So maybe some agencies do have information on life cycle performance risks over time. And in those cases, we can simply enter a value that is going to be a flat payment over the life of the project. In many cases, that kind of information is not available. And then, there is an alternative approach, and that is a more market-based approach. Looking at the P3 financing conditions as a proxy for the value of the life cycle performance risk, the reasoning is that most, as Patrick explained in part two of this presentation, is that most risks are being-- that are being transferred to a P3 contractor are then back to back transferred to subcontractors. Now that is not true for life cycle performance risks, which means that the special purpose vehicle retains those life cycle performance risks, and also retains the revenue risk that we're going to talk about later. Now that means that if the SBV retains those risks, then the financiers of an SBV are pricing these risks in their financing conditions. That is exactly the kind of information that we'll be using to value the life cycle performance risk. How we do this is following the next approach. What we do is we calculate the net present value of all the cost cash flows in the project, excluding financing. Once using a market-based weighted average cost of capital. The weighted average cost of capital represents as we talked about in previous webinars, represents-- it's sort of the weighted average of all financing costs for the special purpose vehicle. We are using that WACC, weighted average cost of capital, that incorporates or reflects the full risk profile of an availability payment P3, to calculate the net present value in the first iteration. Then there's a second iteration, and now we calculate the NPV using a project risk free discount rate. And the difference between the two can be calculated too, which is then an NPV, a net present value. And the difference between the two represents the risk premium that is also incorporated in the weighted average cost of capital, or in the financing conditions. So under A, we calculate the NPV including all the risk premium in the financing conditions. Under B, we calculate the NPV, not incorporating the risk premium. And the difference between the two is the risk premium. Now we are first calculating that as a difference in net present value, so the results of this exercise is a net present value. In the next step, we can also spread out that net present value as a cash flow which then becomes sort of an insurance premium or a virtual insurance premium. Another way to look at this whole approach is by calculating the shadow bid in two different ways. So if we take fully the private perspective, and then we try to answer the question, what does the availability payment need to be in order to make a certain return and to meet all the financing conditions, as we first do that on the basis of a fully loaded weighted average cost of capital, then the result will be a certain availability payment. If we then do the same exercise again for a weighted average cost of capital or a discount rate that is not incorporating those risks, then the availability payments, using that approach, will of course be much lower. So in other words, the bidder who would not be requiring any risk premium in its financing conditions would require a much lower availability payment. Now the difference between those two availability payments can also be recognized as the value of the life cycle performance risks. Or, in other words, the value of the risk profile of the special purpose vehicle that enters into an availability payment P3 contract.
Marcel Ham: On the next slide, we are showing this in the graph. So what you're seeing here in orange is below the time axis, the negative cash flows related to the project, the investment costs, the maintenance costs in those. So the major maintenance costs. And in the purple bars at the start of the time axis, we see two NPVs. One NPV using the market-based WACC, and a higher NPV using the project risk free discount rate. And the difference between the two is the value of the risks, which can then also be expressed as a risk premium, which is the flat purple bar at the top. I'm looking at the Chat box and I see a question related to the project risk free discount rate. That's a good question. And the way we typically do this is to use the boring rate of the agency we're talking about. Because the boring rate of the agency does not include specific project risks, but only the overall risk profile of the agency. And for example, that can be a state. So the boring rate of the state does not include any risk premiums related to the project. And it typically is also based on a full guarantee from the agency that the loan will be repaid. In other words, the risk profile is very, very low from a financing perspective, but it's not zero. There is still a small risk premium representing the creditworthiness of the agency itself. But it does not include any project risks. I hope that answers the question. So this is our approach to value life cycle performance risk. The other category that we want to talk about is revenue uncertainty. Of course, if we are considering a long-term contract and in this case, even a revenue toll P3, then there are uncertainties related to the revenues. There are two ways to value those uncertainties. One is to apply a haircut to the revenue projections. So typically in the preparation of a project, there is a T&R forecast, there is a traffic and revenue forecast. And a professional traffic and revenue forecast has a probability analysis in it. Meaning that the T&R study does not only indicate what the P50 or expected value of the revenues is, but also what the revenues would be under other probability levels. If such probability analysis is available, then the agency can decide upon an acceptable probability level. For example, a P70 or a P80, and use that as a starting point. Or actually, to use that as a revenue projection that then of course incorporates a revenue uncertainty adjustment, because a P70 or a P80 level is a much lower revenue projection than a P50 level would be. If such probability analysis is not available, then it means that we need to make a haircut on the revenue projections based on previous projects. That's one approach. The other approach is to use an approach that is similar to the approach that we've just discussed for the life cycle performance risks, and that is to use the financing conditions as a proxy to determine the value of revenue uncertainty. And let me briefly explain how that works. Again, this is very similar to the approach that we just discussed, because now we are considering all the cash flows in the project, both positive and negative, both revenue and cost cash flows, and we do again to NPV, net present value calculations. One time, we were using the market-based WACC, weighted average cost of capital, that includes all the risks, including the revenue risk. And we calculate the same-- No, not the same-- We calculate another NPV using a project risk free discount rate. So in the first NPV at calculation, the risk profile of the project, including the revenue risk, is being acknowledged through the financing costs. In the second NPV, it is not acknowledged. And the difference between the two represents the risk profile for the special purpose entity or a special purpose vehicle. However, in a toll revenue deal, that risk profile does not only include the revenue risk, but also the life cycle performance risk, as we just discussed. So in an availability payment deal, the SPV financing conditions would only represent the life cycle performance risks, whereas in a P3 toll concession, the SPV financing conditions would represent both a revenue risk and a life cycle performance risk. Now, if we use the approach that we just discussed for determining the life cycle performance risk, then we already have an NPV for that. And the NPV that we just calculated is the sum of both revenue uncertainty and the life cycle performance risk. So, if we take this, the last NPV and subtract from that the NPV representing the life cycle performance risk, we are exactly where we want to be, because we have calculated the NPV for the revenue risk. So this is taking a very similar approach, as we just discussed for a life cycle performance risk. And we are using market-based information. And so such information should be available on the basis of past P3 transactions. In this case, of course, it's important that we find finance-- representative financing conditions for a toll concession deal, not for an availability payment contract. And the way that looks like in a graph is like this. In orange, we see the same investment costs, maintenance costs, the major maintenance costs. Now we are adding the toll revenues north of the time axis. And again, we do two NPV calculations. One is with the market-based WACC for a toll concession. And one is with the NPV-- One is with the project risk free discount rate. And the difference is the NPV of the full risk profile of the SPV, which then includes both the life cycle performance risk and the revenue uncertainty adjustment.
Marcel Ham: So how does P3-- P3 value 2.0 uses or actually offers the same approaches to the life cycle performance risk valuation and revenue uncertainty valuation? In both cases, the two approaches are being offered in P3 value 2.0. Now if we look at all the risks and uncertainties that we discussed in this session, then from base variability, pure risk, life cycle performance risk and revenue risk, and we try to understand how that is incorporated in the value for money assessment, and the PDBCA assessment, then it looks like this. In the value for money public sector comparator, all the risks are being considered, and explicitly priced in a cash flow, following the methods that we just discussed. In the value for money P3 side, or shadow bid, base variability and pure risk are being explicitly valued. Life cycle performance risk and revenue risk are not explicitly valued, but the financing conditions for the SBV are being considered. So instead of separately calculating a cash flow representing the life cycle performance risk and revenue risk, we are simply or maybe not so simply considering the financing costs, the representative financing costs for the P3 bidder. And those are being incorporated to calculate a P3 shadow bid for the P3 cash flows. And that those then include, implicitly include too the life cycle performance risks and risks and revenue risks. In the PDBCA, the same base variability and pure risks are being considered. And in the PDBCA are both on the PHC side and the P3 side, the life cycle performance risk is being-- is explicitly added as a cash flow. This is because the financing is-- the financing costs are not being considered in the PDBCA, which means that the P3 side of the PDBCA looks different from the P3 side from-- for a PFF. Since revenue risk is not considered-- or revenues are not considered altogether in a PDBCA, because again in the PCA, a toll revenue is a transfer. All we are interested in is travel time savings and all the other societal costs and benefits. So for the same reason, revenue risk is not being considered in the PDBCA. Okay, this is a lot, of course. And it's also pretty challenging. There are some-- these methods themselves are, may come across as challenging. But if you've done it once or twice, then it's-- then it is sort of, it is doable. But risk assessment per se is-- it can be pretty challenging. And the most important challenges we discussed on this slide. Starting from the top, estimating risk impacts and probabilities in itself can be very challenging. What is the impact that we should be expecting from a certain risk, if it happens? And what is the probability of it happening? Also, attitude at accounting for a correlation among risks. What if risks are not correlated at all? Or what if they are correlated? Or what if they are partially correlated? How to account for that? Of course, there's ways to deal with this, especially in more sophisticated probability analysis risk packages. But it's, of course, the challenge then is to determine what the correlation is between the individual risks. Another challenge is to account for an identified risk. So many risk assessments start with the risk identification, a long list of identified risks. But those are only the risks that are being identified. And of course, we know from experience that not all risks are eventually being identified. So there's many risks that simply happen. They are somehow part of the risk profile of a project, but they are unknown until the moment that they happen. So how to account for those? The risk assessment guidebook provides some more guidance on how to deal with that. Avoiding double counting of risks is another one. We want to avoid that, of course. Using all those different methods, we should be, and also the different valuation techniques, we want to be careful not to double count. Of course, if we follow the methods that we've been discussing today, we are pretty sure that we are not double counting, but you know, it's sort of very easy to run into troubles here, which is why double counting is a very, very common challenge in risk assessments and risk valuation. Now also, another challenge is how to deal with low probability or low consequence risks and also what if there is many of them. Then another one is how to account for a procurement phase risk, because typically, in our value for money assessment, we are most interested in those risks that happen after the closing of the project contract clause and/or commercial clause and financial clause. But of course, before that, there is the whole preparation and procurement of the project, and there are certain risks related to that too, and those are unique to it. Those can be unique to P3s. And those can be pretty relevant in very early stage value for money assessment, but not in later stage value for money assessments. And then of course, lastly, there's also a bias. There can be bias. Sometimes people believe that a value for money assessment is just a method to justify already made decisions. But it shouldn't be like that. Value for money adjustment and PDBCA should be a structured approach to do a fair comparison of delivery models, but sometimes the evaluators can be biased towards one delivery model. With that, I think we're ready to move on to the question.
Jordan Wainer: Okay. The question is the WACC for a toll concession includes a risk premium that accounts for revenue risk as well as for life cycle performance risks, true or false?
Jordan Wainer: So I'll give everyone a couple more seconds to answer.
Jordan Wainer: Okay. This one was pretty even. Fifty-seven percent said true, and 43 percent said false.
Marcel Ham: Okay. So I think that means that the majority is right. Okay, this is a lot of information, and the method itself is also a little more complicated than the slightly more straightforward other risk valuation techniques that we've been discussing. But as we discussed in this part of the session, the risk premium that is included in the financing costs in the WACC for an availability payment P3 only represents the life cycle performance risks. And the risk premium represented in the weighted average cost of capital for a toll concession P3 represents both revenue risk and the life cycle performance risk. Because this is really looking at the special purpose vehicle, the special purpose vehicle cannot transfer the life cycle performance risk nor the revenue risk to its subcontractors typically. And that means that those risks are being priced in the financing conditions for the special purpose vehicle. So the answer was yes, the WACC represents both. Any further questions on this section?
Jordan Wainer: There is one question in the Chat pod. The question is, how do you get project risk free discount rate?
Marcel Ham: Yeah, I tried to already address that when we were on the topic.
Jordan Wainer: Okay. Then there are no further questions right now.
Marcel Ham: Okay. Then we can move on to risk allocation. So if we look at the different delivery types, then of course the risk allocation is also different. And that's basically the main point of P3s in the first place, is to have a very different risk allocation from more conventional contract types. And I would say that even it is at the core of the delivery type discussions is the extent to which there is a risk transfer. If we start from the top design builds in the design build, the design risk and the construction risks are being transferred and the other risks, financing, O&M-related, and also traffic and revenue are not being transferred to the contractor but retained by the agency. At the other extreme, we have the toll concession in which all risks being transferred to the private sector. And all, is not really all because in the P3 agreements there is typically a very sophisticated risk allocation mechanism. But maybe the starting point for P3 transactions is that all risks are being transferred to the contractor. But then again, there is always certain risks that are better retained by the government, which then leads to some level of sharing of risks or even the agency retaining risks. Examples of risks that are being retained by the government are, for example, change in scope. So if the government wants to change the scope, then of course that risk of the government changing the scope in the future cannot be transferred to the contractor. The same is true for right-of-way acquisition that is typically considered to be at typical government risk. Archeological findings is another one. Now other risks can be shared with the contractor. Typically, delays in the permitting can be shared. There's typically a more sophisticated risk allocation mechanism on the permitting. So the contractor needs to timely apply for permits and then any delays in the procedures are typically accepted by the contracting authority. And so there are many sharing mechanisms typically through delays of completion date and also compensation by the government. The question that we really want to answer in order to get to the optimal risk allocation, the questions in order to get to the optimal risk allocation is first of all, which party is best able to control the likelihood of the risk occurring? Then also, which party is best able to control the impact of the risk? And those are not the same, so the likelihood is different from the impact, of course. And then if there is no obvious best allocation on the basis of those principles, then the third one can be okay, which party is best able to absorb the risk at the lowest cost, if both are not so easy to control. And that should lead-- those three questions should lead to the optimal risk allocation between the public agency, the contracting authority, and the concessionaire. Now the concessionaire itself, or the SPV, as we just discussed, does not necessarily retain all those risks, but pushes them down to the subcontractors. And that is especially true for construction risk, operation risk and also maintenance risk. As we just discussed, the coordination and long-term performance risks are typically retained by the SPV. Now the overall idea of a P3 and also of a value for money is that if the risk allocation is efficient, that then also leads to an overall lower risk valuation and a lower valuation of contingencies. So in the comparison between a conventional delivery and a P3 delivery, the overall value of risk should be lower, because more efficient risk allocation should also lead to a lower risk valuation. An example of a risk allocation for a random tunnel project. What we are seeing here is that the typical public risks are being retained by the contracting authority, political risks. In this case, also traffic and revenue. That is not always a given, so there is also many toll concessions in which the revenue risk is being transferred or there is also many mechanisms to share those risks. Right away, it's typically retained, and in this project, it's also retained by the government. Procurement risk, of course, is retained by the government, and change in law is too, to a large extent. Typically, there are certain changes in law that can be transferred to the private sector. All the DBFOM-related risks are being transferred to the private sector. And then there are certain risks that both the private-- the concessionaire and the contracting authority can contribute to in terms of the management of the risk, in which case it is helpful to share those risks. In this case, it's planning and permitting, but it's also relocation of utilities. Force majeure, there's, that's a very common principle to share the consequences of a force majeure or, in other words, for each party to bear its own force majeure risks, and the same is true for geotechnical risks. And that brings us to the question for risk allocation.
Jordan Wainer: The poll question is true or false, public agency's goal in risk allocation to be to transfer all the risks to the private partner in a P3?
Jordan Wainer: We'll give everyone just a couple more seconds.
Jordan Wainer: And the majority of you said false.
Marcel Ham: Yeah. And that's correct. So the goal of a P3 is of course not to transfer all risks to the private sector. It's a common sort-- Okay. So if agencies are not so experienced in P3s, then the first starting point is typically to try and transfer all risks to the private sector, to then come to the conclusion that can be very inefficient, and the bids can be very unattractive if we do that. So it's then the challenge becomes to come to a much more balanced risk allocation, which is also much more efficient and leads to better value for money. And with that, I want to hand over to Wim Verdouw, who is going to discuss the P3 value 2.0 model on risk assessment with you a little more.
Wim Verdouw: Thank you. And Jordan, if you can help us share the screen. Thank you.
Wim Verdouw: All right. So as you have already seen it a few times now, this is the window that you find when you open the model. We're going through the train navigator, and I'm going directly into the risk assessments. Again, as before, the first slides are the inputs, and below here you find the outputs. The key inputs for this webinar is the risk input sheet. Let's go there directly. And in the risk input sheet, we'll first discuss the pure risk. At the very top here, you see a probability level for a pure risk analysis. So if there's a 70 percent, that means that there's a 70 percent probability that the value that we calculate here will not be exceeded in reality. And of course, the higher the percentage, then the higher the risk value, and the lower the chance that the actual risk value will exceed the calculated risk value. To walk you very quickly through an example of what the calculation would look like, let's look at the engineering and construction risks, here on row 17. We're saying the probability of the, or the likelihood of occurrence of this risk is 50 percent. With the impact, the most likely impact being $15 million dollars. Now the most likely risk value is simply the multiplication of this 50 percent times the $15 million dollars. And of course, it's as we've seen before, the PSC and the delayed PSC or the conventional delivery and the delayed conventional delivery are exactly the same. So this column M and column N seem to pick up the values from the PSC inputs. Then we get to the P3. As Marcel just explained, if there indeed is a better risk management then this can lead to a reduction in risk costs, and therefore the most likely input effect here if we assume a 10 percent risk efficiency is $13.5 million, as opposed to $15 million under the PSC. If we don't want to input the percentage but directly input a risk impact, we can also provide that here in column Q. The same way this is calculated for the PSC, here at the P3, we calculate the most likely risk value by multiplying the most likely impact and the probability in column J. The next key element for a P3 is to determine how much of the risk is transferred. Is it 90 percent or 80 percent or 50 percent? And this goes back to the discussion that Marcel just did on whether it's efficient to transfer all or only part of certain risks. Based on that, we have to split between 90 percent in this case going to the P3 concessionaire and 10 percent being retained by the agency. And the last columns here tell us something about the distribution of the risks, and these, this distribution applies to all three delivery models. The PSC, the delayed PSC and the P3 option. And in this case, we're saying that we know what the most likely impact was, and the minimum value is 25 percent lower than that, and the maximum value is 50 percent higher than that. And we also need to specify whether they're triangular or a uniform distribution. And once we have all these inputs, then we can go back to the calculation that Patrick presented earlier to calculate the actual value of an individual risk. And using the central theorem <laughs>--
Patrick DeCorla Souza: Limits.
Wim Verdouw: The central-- a limit theorem, sorry, we can calculate the overall value of the cost under P3 and PSC. Now we've seen these risks now for construction periods. We have the same thing for the operations periods from row 30 to 38, and this allows us to calculate the pure risks. Next, we have a number of inputs for the base variability, these are in row 45 to 47, which can be distinguished between preconstruction costs, construction costs and O&M costs. And of course, we have different inputs for the PSC and for the P3. So in this case, we're saying that the PSC and the P3 are the exactly the same base variability, but if we would like to differentiate, that would also be possible. And the last risk inputs that are key here are for the life cycle performance risks and the revenue uncertainty adjustment. As Marcel explained, there's always two ways to calculate both. For the life cycle performance risk, you can either have an input value which is in row 53, is an input value for the total concession bid. Or instead, you can calculate it by selecting a zero in row 52, and in that case, we use the financing conditions of the P3 to determine the life cycle performance risks. Similarly, for the revenue uncertainty adjustment, we can input in cell 59 whether or not we want to calculate it based on the financing conditions, or use the input and if it's an input, it's a simple percentage of the overall revenues. There's one more key element at the input, which is in the input FIN sheet, and that is the difference between the WACC of an AP and a toll concession-- sorry, an availability payment WACC and a toll concession WA. With all these inputs, the model calculates the various risks. And I'm just showing you here the risks for the value for money. There's a very similar sheet right next to it, which is for the benefit cost analysis. And as you will see here, the risks are listed by construction risks, pure risks, pure risks for operations, phase availability for pre-construction, construction and O&M, and below, the life cycle performance risks and the revenue uncertainty and adjustment. And as you will see under P3, of course the life cycle performance risk is already included in the financing cost, and so therefore it doesn't need to be calculated. Whereas for the conventional delivery and the transferred risk on the P3, these need to be calculated. And with that, I would like to hand it back to Patrick to wrap up the webinar.
Patrick DeCorla-Souza: All right. Thanks, Wim. Are there any questions for Wim before we move forward?
Jordan Wainer: There was one question. I think it's more for Marcel, but it's the question is, if we may go back to the Monte Carlo versus formula estimation of risks, do you have a recommendation for what is better to use? Does it vary based on type of project?
Patrick DeCorla-Souza: Marcel?
Patrick DeCorla-Souza: Are you on mute?
Marcel Ham: I was on mute. I'm sorry. I think this goes back to Patrick's part of the presentation for the valuation of regular uncertainty and pure risks too. There's not one hard and fast rule, I would say. I would say that the central limits theorem is a little easier to apply, but if you have the information to do a Monte Carlo simulation, and also too all the inputs to do a Monte Carlo simulation, then I think that would be in general that would be preferable. Then still, of course, there's a lot of decisions that need to be made on correlations between the different inputs. And in my experience, that's the hardest part. So typically, we should be careful not to get into all kind of false precision, because that's what can happen. But I would say in general, if you have the information, then a Monte Carlo simulation is helpful also because you can better visually see what's going on, and see what the uncertainties are.
Patrick DeCorla-Souza: All right. Thanks, Marcel. So let's move on. Of course, this is what Marcel just presented, the P3 value 2.0 and we've gone through these slides on previous webinars, so we won't go through them again. Let's go to the summary, and just to review what we covered. We explained the different types of risks. Where risk assessment fits in the risk management process. And between Marcel and myself, we talked about how to value each of these risks, starting with base variability, and then pure risks, life cycle performance risk and revenue uncertainty. Marcel explained the process used to allocate risks to the private sector, and Wim explained how P3 value uses all of these concepts to calculate risk for the model. Now there's a lot more information that I would encourage you to consult. First of all, I invite you to download the spreadsheet from our website, P3-VALUE 2.0 Tool. Also on the website in the P3 Toolkit section, you will see a user guide on risk assessment, as well as a risk assessment primer and guidebook. So the user guide of course I meant to say, was actually the P3 value user guide, which includes the concepts that Marcel was explaining, though which I understand are quite complicated using market-based conditions to quantify and value risks. And those explain in much more detail and thoroughly in the user guide. So please consult the user guide to get more information on that. Now we will have more instruction on this particular risk assessment topic. Next Monday, we will review the exercise which you can download. And that is available on the left-hand side in the materials for download, there's an exercise that you can do over these next few days and we will review it on next Monday. If you complete the exercise and send it to me, you will get a certificate. You know, so follow the instructions in the exercise to send the assignment to me. Of course, you can always contact me if you have any questions on the exercise. And of course, next Monday we have the information to connect to the webinar down below on the slide. We will also be sending you along with the recording, we will-- of this webinar, we will send you the information to connect to the exercise webinar next Monday. This is the Office of Innovative Program Deliveries website, and the specific P3 website from which you can download the tool and all of the guides that I spoke about. And here's my contact information. So if there are any questions, we can-- you can pick up the phone at this time and ask them. Just unmute yourself hitting *6.
Patrick DeCorla-Souza: Okay, hearing no questions. I want to thank my colleagues Marcel and Wim Verdouw for helping me present this webinar, and I want to thank the moderator, Jordan Wainer, for helping with the moderation. And I look forward to you sending in your exercise and joining us next Monday at 12:30 p.m. It won't be at 2:00, it will be at 12:30 instead of 2:00. So with that, thank you all for your participation in the webinar, and see you next Monday.
