 Well good afternoon. It's a pleasure to be here. I want to talk to you about the solar PV industry, my take on what's happening in that industry and where it's going. It's obviously a topic that we read a lot about in the papers these days. We in particular have heard about companies like Apple, Kaiser Permanente, also Stanford University, not a corporation, but still a large electricity consumer signing big utility-scale deals in recent weeks and months. People like Al Gore are on record now to say that they believe solar has actually reached cost parity with other traditional electricity sources. And so I guess that all of that sort of suggests that the subtitle that I chose here is solar becoming red hot. And I basically would like to tell you sort of the economic analysis that we have been doing at the business school. I've been doing this with two colleagues and I'm going to sort of weave in two research projects that we have been doing in that domain as part of this conversation. So one thing I think we can all agree on and that is simply growth in solar capacity installations as you see here over the last 15 years. It has been spectacular by any measure just in terms of the growth rates that have been sustained over the time period. As you may be able to see depending on how far in the back you are, the Europeans early on had the majority of those installations. They still account for the majority but that curve is now sort of leveling out. The Asia-Pacific part is growing rapidly. China in particular and on top of it the Americas largely the U.S. is finally getting into that market or into that industry. Here's a breakdown by country and perhaps most remarkable Germany has been sholdering a large part of the cumulative burden that's on the right up to till the end of 2014. Countries like China now are catching up and if you see in terms of the additions in 2014 alone China accounts for almost one third of its total cumulative installations. In other words China being of course an important producer of solar PV modules has in recent years also become a consumer and has installed these at a very fast rate. So if I look at sort of the U.S. and ask what has been behind the rapid growth in the U.S. market alone I think there are a variety of factors and part of what we're trying to do in these studies is to disentangle and quantify the different effects that have contributed to this very rapid growth here. There's first and foremost technological improvements and cost reductions at the level of the PV modules but also other parts of the hardware have been coming down in cost rapidly. And then the next part of this has been changes really in the way the industry and businesses have been conducting those deployments. The industry is probably becoming more vertically integrated say in the U.S. companies like Sunpower which traditionally used to be a manufacturer of solar panels now integrating downward into the developing side also. And companies like SolarCity which traditionally have been in the installation business integrating back upwards and now also becoming a module manufacturer. So that seems to be at least for some major U.S. players a trend towards more vertical integration. Companies like SolarCity also have been pioneers in the way they have financed these deployments, these investments, the way they have presented packages to investors be they individual homeowners or companies and perhaps the probably the most significant change there the securitization of lease claims that go hand in hand with these deployments has arguably lowered the cost of capital for these companies considerably and I'll in a moment sort of show you how important that has been in terms of the overall calculations. Finally and that's the part that I'm probably going to spend the most time on here in the next 30 or so minutes. Federal policy and also state level policy at the federal level the investment tax credit that was given to solar about five to six years ago in the amount of 30 percent that will not is not scheduled to stay there and that's an issue that we're going to pay particular attention to but also other forms of tax incentives in the form of accelerated depreciation. In addition at the state level many states have enacted renewable portfolio standards which have also given a boost to solar but also other forms of renewables and that is probably among all the lists that are all the items that I'm listing here the smallest item but should not be forgotten if we're trying to explain the rapid growth in this industry over the last five or six years. By the way feel free to interrupt anytime if I'd be happy to make this a dialogue all along. So in terms of questions the first one probably to ask in my mind is how competitive a solar PV today and to assess that we this is part of a study that we have done for the Department of Energy recently we have looked at the three main segments of this industry utility scale commercial and residential and the numbers vary quite a bit and we've also picked five sample states that you see listed there on the slide those cumulatively account for probably close to 70 percent of all solar deployments in the U.S. in the first place. The next question then after we've answered the first one is how important is the change in the tax code that is anticipated in early 2017 when the investment tax credit is currently scheduled to go from 30 to 10 percent how much of a difference would that make and then what I would like to sketch at the end is basically a proposal that we have worked out recently as part of this Department of Energy study that I mentioned which tries to sketch an alternative to this step down that is currently envisioned in the law and replaces it by a gradual phase down and I want to just show you some numbers to give you a sense of how that might work and how it would in our minds alter the path that the industry in the U.S. is likely to take if current law is not changed. Okay so first question what's a good metric here in all of this the one that we're using is a traditional one which has been discussed a lot and also criticized in certain quarters we think it's still a good workhorse the levelized cost of electricity you may have heard the term basically looks at is a soup to nuts concept where you're looking at the entire investment from the initial installation of the system you're operating it for the next 25 years and you try to take into consideration all the costs and you're backing out from it a revenue figure that's what an investor would need to obtain on average on average is the operative word here per kilowatt hour in order to break even on this investment in terms of components there are basically fixed operating cost on an annual basis those tend to be relatively small think of just maintenance of the facility the big ticket item is the unit cost of capacity where you're taking the initial systems price and levelizing it or anuitizing it over the lifespan of this investment and the factor that we're going to pay particular attention to if we want to look at the importance of tax policy in this area is what's called the tax factor and that takes basically everything into consideration related to income taxes investment tax credits and the like that's all sort of baked into that third factor so in terms of as I said we're trying to do this by applications in different regions of the US and also in terms of different applications that is the size of the solar facilities so for PV module cost it's pretty much the same across the different applications but once you get to the so-called balance of systems cost everything other than the modules themselves you will have some variation both depending on geography and and the on the application itself big factor for solar of course is simply the amount of sunshine so the insulation factor that makes a large difference and finally the whole question of what's the applicable cost of capital that goes into these lifecycle cost calculations that also varies in our mind by application and also a little bit by geography we're working sort of with numbers there in the range of 6.5 to 9 percent and I'll show you some numbers later on in terms of how sensitive things are to those specifications so here is our reading our current reading sort of of the tea leaves where we are in 2016 so we're going slightly forward looking to say the middle or the end of 2016 and the numbers you have see in the tables if you could focus sort of for a moment on the first column for each of the four applications utility commercial and residential and for utility we're distinguishing both silicon and thin film so say pick california for a moment we have a number of at the moment or 2016 of about 6.9 cents per kilowatt hour as this levelized or lifecycle cost now to some people that may sort of seem a little high we have already seen better numbers but that takes into consideration here our numbers here that we're basically doing a broad average so we're not taking sort of the perfect location within california we're also ignoring any state level incentives in the in these numbers so that in a state like california we tend we have a tendency to make the numbers even lower even better than what we included in our calculations perhaps most importantly if you now take this to alternatives say a natural gas type facility we're pretty close our numbers suggest even at the very low natural gas prices that we're seeing at the moment this is a close horse race perhaps the number for natural gas depending on how you calculate comes in somewhere in the range of 6.4 to 6.7 cents so solar from that perspective is really catching up very very rapidly what we have in the second column are comparison prices so not necessarily a natural gas power plant but question in california if you peg it to wholesale electricity prices that's in the second column there and if you go over to commercial and residential those comparisons those benchmarks change of course because we're looking at essentially what a business that is contemplating to put solar panels on its warehouse rooftops would have to pay if it worked with the local utility and again this these are broad averages this does in the present form not take into consideration that these electricity rates vary a great deal throughout the day there are ways to adjust an account for that but we'll have to leave that for another day so to speak and finally there are the numbers in red almost everywhere and this is basically now a levelized cost calculation in which we are anticipating that by 2017 unless congress changes its mind the investment tax credit were to go from 30 to 10 percent so those subscripts 30 represents 30 percent and 10 means 10 percent in the third column here what's striking in my mind is now sort of talk about solar catching on and being really close to being competitive in many of the applications is reversed just about across the board so all the red numbers the red numbers are not the ones necessarily in the third column but the ones that are indicating that we are above the comparison price in other words the investor that's looking trying to earn a normal return on their investment relative to the current comparison prices would find solar unattractive at that point in time and you see the only application that we have been able to identify is that commercial in california at 14 cents per kilowatt hour would still look like a goodbye but everything else would be as we put it here under water okay perhaps one word on sort of sensitivities of these calculations refer to this sort of as a spider graph here in the center of my little grid I have zero zero that those are if you want to the baseline calculations and now you ask well if you vary one or several of these factors in particular in the blue line the capacity factor largely driven by location and or efficiency of the cells what impact does that have on the levelized cost of electricity for a solar system and just to put that in perspective if you had a 50 cut in your capacity factor because you're moving into a less favorable location in the world talked about Germany earlier Germany's insulation factor at least if you go sufficiently far north in Germany is probably about one half of what it is here in california you're pretty much doubling your levelized cost of electricity so that's how sensitive solar and solar's competitiveness is to location on the yellow golden line or the red line here the impact of both the system price so that's basically what we're talking about when we look at reductions in module prices or balance of systems costs how sensitive that is in the red line cost of capital again all under percentage basis and you see all of them have very very rapid impact on these levelized cost calculations so things are quite sensitive with respect to these three factors if you want to sort of try to forecast where things are going in the future okay let me do for a moment sort of the two yes go ahead that depends on the application so we have four utilities we have put in a cut at weighted average cost of capital of seven percent and for residentials we have gone all the way up to eight and a half to nine percent so it depends on the application and it's also a little tailored to the state so I want to talk a little bit about the dynamics of these system price costs and the two main components I mentioned earlier being the modules and the balance of systems this chart here is probably known to many of you it's sometimes referred to as Moore's law of the solar PV industry it was I saw it here first in a seminar I guess was one of Sally's energy talks on the Monday afternoon in 2011 Dick Swanson the founder of Sunpower presented it and what you see on the axis here is cumulative total production measured in megawatts of solar PV modules and over on the other axis you have price both of these are on the logarithmic scale rather than the actual prices so if you had these in terms of actual numbers you would have a very steep convex curve what's remarkable about this is it has held up for as you see they are close to 30 years and at the seminar in 2011 the questions was well that's really great but this can't possibly go on like this this industry must run out of steam pretty soon this is a very impressive pace of cost reductions over 30 years but obviously sooner or later you know you're going to hit a wall there this also for for people in in business schools the nice thing about this curve is it corresponds almost too well to be true to what's known as an 80% learning curve that is with every doubling of cumulative production you cut your cost by 20% so it's only 80% of what it was before and again if somebody shows you that graph you would think you know this is this can't be true that it comes out that cleanly there's a little hump there in red commonly attributed to a shortage of polysilicon around those years but then the industry also got back pretty much to the historic trend line with remarkable accuracy so again the line the the conclusion at the time was in 2011 well this can't go on guess what not only did it go on but it got even faster and as many of you know so what I have in this graph here is simply I haven't done anything other than extrapolated in red this 80% learning curve beyond what we had in 2011 when Swanson gave the talk and also tracked the average sales supply prices in blue since then and of course the remarkable thing is starting in 2011 while this was a pretty close fit all along the bottom seems to be falling out here and that coincidental of course was a new wave of manufacturers most of them from China entering the industry and generally the industry also being perceived as hurting most of the players being highly unprofitable during those years okay so the interesting thing then in our minds is to try to disentangle well what's going on here is to what extent are the observed prices a consequence of excess capacity which may have been created for whatever reason we can speculate and to what extent is it does it actually reflect sustained cost reductions because there's also a reason to believe that the innovations have gone on during those years so trying to disentangle those two effects is one of the things we were trying to do in one of the studies and this is with one of my colleagues at the business school and at the Schleyer-Taylor Center here on Susahu what we have done is basically now gone and looked at the financial statements of about a dozen companies in the industry most of them from China relatively pure players in this market and have basically looked at annual reports to try to get at cost information and then from the cost information try to back out what prices should have been would have been in this industry had it been in equilibrium so it's basically sort of an industrial organization type study where you're backing prices out from the underlying cost structure of course as researchers we're a little limited in terms of what we can do because we have to rely largely on the financial statements and some industry association data so it would be nicer if we had sort of really inside cost data but we didn't have that at least not for the dozen or so companies looked at so what we get at the end of this day is basically the green line and the green line is what we term economically sustainable prices these are the prices that we should have seen if firms in the if this industry had been in an equilibrium and firms would have made a normal profit and once again they didn't but what's in my mind sort of noteworthy about it is there is really a pretty close hug here between the blue and the green line so average sales prices were roughly in line with what we infer the cost structure of these firms to be but things start to diverge really in 2011 and if you now think about sort of the historic trend line the actually observed prices and the inferred economically sustainable prices the distance between the green and the green and the blue line that becomes ultimately sort of our measure of the impact of excess capacity because if things had sort of state in balance and the cost structure would have sort of been where it is then we should have seen prices at the green line so that's sort of our measure of excess there you can use this type of analysis now also to go forward and sort of say okay so given this trend line and costs if we now extrapolate and basically sort of go back to ask the question where are we going to be you know two or three years down the road so our study ends in 2013 and pretty soon we'll be able to update it when the 2014 numbers are all in what we're finding actually what again surprised us a little bit here the so-called 80% learning curve is alive and well so any worries that have been expressed that you know this industry would hit a wall and would not be able to maintain this pace of innovations we don't see that in the numbers actually if anything at all it's slightly faster than the 80% we're arriving at a 78% learning curve so you would be reducing your cost by more than 20% with every doubling so so far at least as far as modules are concerned there is no reason to believe that things are going to slow down that of course is before we're talking about any of the breakthroughs that people are waiting for sort of with other technologies this is just all focused on crystalline PV okay last point then balance of systems here we have less in data but we've been sort of looking around basically at various industry association reports conduct the same interviews try to get a sense of where in this in those remaining part balance of systems cost used to be the not so interesting part of the overall systems prices because it was small in comparison to modules but of course that has changed now balance of systems accounts for more than half and if you're going into the residential market substantially more than one half mentioned vertical integration before but you know companies like solar city around here and a couple of others I think have shown basically a number of measures that really just in terms of bringing that balance of system costs down have been rather powerful and so in forecasting things we're willing to work with roughly a 5% in cost improvement on average each year for the time frame that we're going to that I'm going to project out here in a moment okay so where does all of this leave us cost reductions likely to continue at the pace that I had sort of indicated for modules and balance of systems costs now bring in the role of the federal government and tax policy in all of this unless there is a change in washington this investment tax credit would drop from 30 to 10 percent however it's also intended to stay there at 10 percent without any sunset provision without any termination so that in our minds sort of really raises the question is that a reasonable way to go where you the industry would experience this sharp step down and I showed you earlier the numbers in terms of what this would do to the levelized cost calculations and the competitiveness but then you would also continue with indefinite tax support really beyond that at the 10 percent level and 10 percent plus the accelerate the depreciation benefits still is a very substantial type of subsidy in this case so another way to slice this in my mind this is the proposal that been working doing with another colleague here at the business school is to actually say well rather than have this sharp one-time step down replace it by a gradual phase down and to the extent that some kind of quid pro quo is needed here to get a deal in congress also say well we're not going to do this forever we're only going to do this for 10 years and then they if you want to in return for to give the industry a little bit more early on we would be saying then even the 10 percent investment tax credit good would go away so here is a illustration of what we have in mind just conceptually this is when the ITC is supposed to drop from 30 to 10 percent and correspondingly the levelized cost would shoot up the numbers I showed you in the table earlier so we have sort of an in-between regime for the next four years 2017 through 2020 that would then bring with it a first bump but much smaller than the one we're talking about and since there is sort of some headroom lift in the numbers that I showed you early on that in our mind would not be a cliff but it would just be if you want to a small bump in the road and then you would do the same thing a second time in 2020 for all the way through December of 2024 at which point then you would revert back to absolutely no subsidies at that point in time so the question then is how are these numbers calculated that give rise to the red line one idea here in all of this is if you think about it on a in terms of dollars per watt the subsidy of course is very relatively high for residential systems which have the highest system price so one as part of this proposal that we have been evaluating and calculating doing the numbers on the suggestion would be to give investors a choice between a reduced percentage investment tax credit which we're putting initially at 20 rather than 10% or a lump sum amount and you see the initial lump sum amount was suggesting there is 40 cents per watt the idea being a targeted incentives so that at the end of the day the residential systems would opt for the lump sum 40 cents while commercial and utility scale would still stick with the percentage based ITC next question then is well where does the 40 cents come from we didn't quite pull it out of thin air but we did a calculation basically along the lines of if you compare solar to a fossil fuel alternative and you put a price on carbon and you now do a calculation of the avoided carbon emissions because you're doing solar as opposed to burning natural gas and you evaluate that over the lifetime of the investment 40 cents actually pops out as the number here that would basically can think of this as sort of a negative carbon tax as a subsidy that you're offering the solar investments the nice thing is so that's how we calculated the numbers if you now run this through and evaluated over those periods you are of course encountering those bumps at two points in time but I would say relative to the sharp step down scenario that I showed early on those bumps are relatively small and would keep solar in many of the applications either competitive or pretty close to it I like the concept of your your your lump sum payment is that in the same form as the grant was where it you wouldn't have to have the tax equity would basically be a grant you don't have to go give away half your margin to a tax equity provider yes so I think there are certain similarities whether you would still at the end of the day I think we have calculated it's in such a fashion that you still would need the tax appetite but that's something you know one could consider that it would basically be targeted to people who don't need that would just be a lump sum amount it would be a lot more valuable that way oh I believe that opportunity yeah yeah so bottom line here is this would be those numbers would be sufficient to keep things either at competitive levels or very close again in most of the states New Jersey is one application one exception to what I said but the other four states and most of the applications this would work and it would have in my mind obvious advantages in terms of the impact on the industry so let me conclude then if there is no change in the tax code we are predicting at least for the US industry what we call a cliff here and when I started out this whole talk by the question is solar PV becoming red hot yes it is red hot at the moment but probably partly in anticipation of the fact that the tax code is very is likely to change at least unless something happens if it were to happen I think there would be an undesirable boom and bust cycle for the next year and a half that in my mind should be avoided what our numbers basically speak to is what it would take and again you have to sort of take my word for it or look at our report to see that this alternative phase down scenario would go a long way towards smoothing this path into the future and the interesting thing of course is beyond 2024 if the trend lines continue we project that beyond another 10 years even without any incentives in the tax code solar would be able to stand on its own feet in most of these applications okay this is only for the US predictions are dangerous in particular when they concern the future I'll go out on a limb here and say I'll go and say within the next 10 to 15 years regardless of what happens in the US and what happens to the US tax code I think prospects look pretty bright for solar PV compared to the current technologies we're aware of there are indirect subsidies that you didn't mention actually from the california eyes a point of view solar is eligible intermittent resource that means that they have to buy it no matter it stabilizes system or not so they actually have somebody has to pay for spinning reserve to keep it intact on the residential side it is basically you can it's there is feeding tariff but to maintain stability of the system you either load the grid with a large variation or you have to buy a battery so you can think about it as a cost of the battery and this this subsidy though it exists there are several processes that in the sense of that are aimed at reducing it can you comment on anything about that will it change your analysis so the numbers say since you mentioned one form let's look at residential installations for a moment the numbers that we have presented and studied here definitely do assume that there is net metering and that if you have solar panels on your rooftop you can buy or sell an unlimited quantity at any point in time at the rate at which you are being charged by the by the utility your residential rate now that as we know is something that's currently being debated the utilities would like to get to different types of structures if that went away I mean if I sort of look at sensitivity if you want to think of this as a form of a subsidy and the sensitivity of the numbers if you took that away the residential numbers would shoot up dramatically that is if if you did not have the ability at peak times to sell back to the grid at exactly the same rate that you're being charged at other times I don't know is that I just picked one aspect very welcome