The last article in this series on First Solar (NASDAQ:FSLR) looked at their earnings guidance and results for 2013 and first quarter. The article suggested that First Solar is consistently conservative on guidance and that 2014 would be no different. The question that was unanswered was what should investors expect and what would be the bandwidth of potential earnings for First Solar in 2014. The current analyst estimates for Q2 earnings per share (EPS) fall between $0.21 and $1.04, averaging $0.37, and for the full year, fall between $2.46 and $3.27, averaging $2.70. The full-year average analyst estimate of falls within the high and low of the guidance provided by First Solar of $2.40 to $2.80 per share. The analysis here attempts build a more accurate view of possible outcomes for the second quarter and full-year earnings and what to look for in First Solar's second quarter results when they are announced.
Reviewing the SEC 10K and 10Q provides an outline of the key projects that are in progress for 2014. Table 1 compiles data from both the 2013 and 2014 SEC reports. The table outlines the key projects, the state of completion, revenue recognition at the end of 2013 and first quarter 2014 as well as the expected completion dates for these projects. This provides a framework for forecasting First Solar's results for second quarter and full-year 2014.
Source: 2013 10K report, 2014 10Q report
The following sections of the article go into further detail on how the elements of the forecast were estimated and overviews the rationale for the key assumptions. In summary, the approach was to group projects and estimate their revenue and margins based on the attributes of the group. For instance, legacy projects which had recognized revenue and profit over a number of quarters were split out into two groups (Group 1 & 2 - see Table 2 for groupings). Their margins were based on their suggested historical run rate (See Table 3 for details). New or existing projects with little historical evidence of their margins were split into three groups (Group 3 to 5). These projects had their margins varied to assess the impact to First Solar's results. Finally, Group 6 contained revenue that was generated from other sources such as module only sales, Operation and Maintenance (O&M) contracts and small project revenue. Margins for modules are reported in detail in the 10Q SEC reports. So their margins were based on first quarter margins of 7.9%. Revenue for module only sales was modeled at about 8% of total revenues as this is roughly the average of the last 4 quarters. Since there is no detailed backlog or reporting of O&M or smaller project revenue, both the revenue and margin of this group was varied. A sensitivity analysis was then run which resulted in various views of First Solar's potential revenue and gross margin.
Once the overall gross margin dollars were calculated, selling, administration expense and tax were deducted to come up with a bottom line profit number and EPS. These amounts were based on the mid-point of First Solar's Analyst Day guidance. This was $375M on the selling and administration expenses and a tax rate of 17.5%. This was divided by the expected diluted share count of approximately 102M shares. This resulted in various views of First Solar's EPS under various project margin and revenue assumptions.
The reader that is not interested in further details of the forecasting assumptions can skip the next sections and jump to the section titled "Results of Sensitivity Analysis".
Assessing Key Factors and Assumptions in the Forecast
There are 3 key factors that need to be assessed to do a forecast for the second quarter and full-year revenue. First is how much of the project is expected to be completed throughout 2014, second, how much revenue is expected to be recognized on each project and third, what are the margins on those revenues.
Percentage Completed in 2014
In order to assess the first question the rate of completion can be looked at for those projects already in progress. So for instance, Desert Sunlight went from 88% complete at the end of 2013 to 96% complete by the end of the first quarter of 2014 which would suggest this project should complete in second quarter. In the case of other projects which haven't started, an estimate can be made assuming projects will progress on a straight line basis to completion and assuming they will be complete at the end of the year they are scheduled to be delivered. In other words the project progress and estimated revenue is spread evenly over the number of quarters to the end of the project. This basic methodology is applied to estimate the percentage of completion for each project.
Revenue and Margin Estimates for 2014
The assumptions and rationale for estimating the project revenues and margins is outlined in Table 2 & 3. As mentioned earlier, projects were broken up into 6 groups. A key assumption is that project revenue is estimated based on the size of the project in watts times a factor of $2.60/watt(ac). So for example, a 50MW project would be estimated to be worth approximately $130M ($2.60/watt X 50MW). The estimated revenue would then be calculated based on the percentage of the project that was completed in the period. So in this example if the project was 5% complete at the end of first quarter and estimated to be 15% complete by the end of second quarter then the project would have completed an additional 10% in second quarter and $13M would be the estimated revenue recognized in the quarter (10% X $130M). The $2.60/watt factor was determined by reviewing several data points which indicated this was a reasonable estimate (see Table 3 for details).
Also mentioned earlier, module revenue was based its historical share of the overall revenue mix for the last 4 quarters and first quarter margins of 7.9% was assumed in all cases. It should also be noted that there are several large projects which are near completion which haven't been sold (e.g. Solar Gen 2). When these projects are sold, all the revenue will be recognized in the same quarter. In this analysis it was assumed that the three projects listed in this category will be sold by year end. Lastly, "Other" revenues for smaller projects as well as O&M were varied across several scenarios.
The details on margins assumptions are outlined in Table 3. Group 4, 5 and 6c) are key groups as these are new projects with no historical evidence of their potential margin and little significant price information. It is recommended the reader refer to Table 3 for more detail on the margin assumptions in this analysis. Lastly, one key factor to note in Table 3 is the conversion factor for converting watts(ac) to watts(dc). The conversion factor used is 1.3 (i.e. 1.3 watts(dc) converts to 1.0 watts(ac)). It is important to note whether the watts are (dc) or (ac) as numbers quoted here change depending on context of the discussion.
A sensitivity analysis was run against a number of project margin scenarios for Groups 4, 5 and 6c). In addition, scenarios were run for different levels of "Other" revenues from projects and revenue streams such as O&M contracts. Table 4 provides an example of second quarter and full-year forecast revenue and total margin dollars under a scenario of 20% margin for the variable group. Second quarter revenues were estimated assuming the same year-to-year growth as first quarter and the full-year revenue is based on the mid-point of First Solar's revenue guidance.
Note: $ for revenue and margins in millions
Results of Sensitivity Analysis
The results of the analysis was then converted into an EPS estimate and provided in Table 5 and 6. Table 5 provides the results for the second quarter EPS estimate. In this analysis the revenue for second quarter was varied between the highs and lows of current analyst revenue estimates by varying the level of "Other" revenue. The mid-point reflects second quarter revenue growing at the same rate as they did in first quarter (25%). As can be seen, if the growth rate for second quarter is similar to first quarter the outcomes for EPS suggest that First Solar is likely to be under current analyst estimates which range from $0.21 to $1.04 and average $0.37.
Table 6 provides the sensitivity analysis on the full-year EPS which is much more positive. In this analysis the high, mid and low revenue reflect the high, mid and low of First Solar's current revenue guidance. The revenue was varied by adjusting the "Other" revenue. Even at the lowest end margin on new projects of 17.5% the EPS is $3.10. As outlined in the next section, this analysis suggests the range will be in the $3.45 to $4.01 range with the mid-point of $3.73 considered likely to be the closest to the actual outcome. This quite a bit higher than the average analyst EPS estimate of $2.70 and even above the range of analyst estimate between $2.46 and $3.27.
The analysis here suggests that second quarter average EPS estimate is overstated by analysts. The prime difference is that analyst average estimates for revenue come in at $796M, which would be a growth rate of about 53% year to year. This would be well above first quarter's growth rate of 25.8% as well as the high end of the growth rate that First Solar bas projected for the year of approximately 21%. Implicitly analysts are projecting margins of approximately 17.5% vs. 16.5% for the modeling done here. The likely difference is that Macho Springs has been assumed at zero margin here given its aggressive pricing and this analysis has taken a conservative approach. However, if analysts are correct and margins are higher it would suggest that First Solar's has the ability to deliver low cost power purchase agreements (PPA's) more profitably than assumed in this analysis. This would have potential upside for the balance of their portfolio in terms of overall EPS for the year as well as First Solar's competitiveness in the marketplace in general. Other potential upside to the second quarter EPS could come from additional revenue from O&M or other projects that have been underestimated here. One significant item would be if Solar Gen 2 had sold in the quarter - although that appears unlikely at this point given the lack of any announcements (Solar Gen 2 could add as much as $1.00 to the EPS in second quarter)
While second quarter appears low, full-year EPS appears likely to exceed analyst estimates and First Solar's guidance. This is not completely surprising as the previous article (linked above) pointed out that First Solar has a history of being very conservative on guidance. Overall, after reflecting on the data, it seems likely that EPS for 2014 will fall in the range of $3.45 to $4.01 with the mid-point of $3.73 a reasonable estimate. However, this would be up to 55% higher than either First Solar's current guidance or analyst estimates. So the question is why is there such a large difference.
Reviewing analyst guidance their implicit assumption is that margins for the year will average about 18.6%. Backing out first quarter margins from analyst estimates implies that their average margin for the period from second quarter to the end of year is in the 16% range. Assuming that they made similar assumptions on module revenue and margins as reflected here it would suggest that they have projected new project margins will average 17.5% - which isn't unreasonable given First Solar's guidance. However, there are a number of reasons to believe that margins on projects are not likely to be this low.
First, reported module production in 2012 and 2013 was reported by First Solar as 1875MW and 1600MW respectively. Using the SEC reports we can estimate the amount of MW for power systems and consequently the approximate $/watt that First Solar was recognizing on their power system sales. For 2012 approximately 72% of the modules went to power system sales (utility scale projects etc.) which drove $3.0B, this amounts to approximately $2.24/watt(dc). In 2013 approximately 68% of the modules went to power system sales resulting in $2.9B of revenue which translates to about $2.06/watt(dc). As already noted, the backlog opened the year at $2.78/watt (dc) ($7.5B backlog of 2.7GW(dc)). Granted the $2.78/watt backlog also includes other revenues such as O&M contracts which would not be directly tied to module backlog. However, the difference between the backlog average of $2.78/watt(dc) and 2013 average of $2.06/watt(dc) would allow for nearly $2 billion in non-module related backlog before the current backlog declined below the $2.06/watt recognized in 2013. This provides significant buffer for other revenues that may not be tied to module backlog.
In addition as noted in Table 3, AGL is a relatively recent contract sold for approximately $2.70/watt(ac) or the equivalent of about $2.08/watt(dc) which is actually slightly higher than the derived revenue per watt for power system revenue in 2013 of $2.06/watt(dc). The $2.06/watt(dc) average in 2013 drove a margin of 26% in overall power system margins. Recognizing these estimates may have some variability due to a number of factors, there is still quite a bit of buffer for the estimates to be wrong. For example, First Solar indicated in its Analyst Day meeting that costs for utility scale installations was running below $1.59/watt which is consistent with what was reported its SEC reports. Assuming this cost profile would apply to AGL, even a 2% reduction would imply margins greater than 22% - and as noted in the Analyst Day meeting First Solar is working to reduce their costs by 10% per year through 2017. So AGL pricing also suggests that margins may be higher than what analysts are using in their current modeling.
The final clue as to First Solar's margins is contained in their 2015 guidance which is for earnings of $4.50 to $6.00 per share on $3.8B to $4.3B of revenue. Based on this revenue and EPS it would suggest margins would need to be in the order of 21% or higher in order to meet these EPS estimates. Based on their charts from their analyst day presentation, about 80% of this profit is already booked in their current backlog. Many of the projects listed here span a number of years. Accounting principles effectively requires they be booked based on the expected margins across the project lifespan. This suggests that either guidance on margins for 2014 is too low or 2015 are too high. Historical precedent suggests that First Solar's 2014 guidance is too low (see previous article).
So after reviewing these data points there appears to be significant evidence in the backlog data, current project pricing and First Solar's own 2015 guidance that margins are higher than what First Solar is reflecting in their 2014 guidance as well as what analyst's are projecting in their 2014 EPS estimates. So as previously indicated the conclusion of this analysis is that full-year 2014 margins will likely fall between 20% and 24% and that EPS will be between $3.45 and $4.01 with the mid-point of $3.73 considered the best estimate based on the data presented here.
Overall, this analysis builds on the previous article and supports the conclusion that First Solar's full-year 2014 EPS is likely to exceed both First Solar's guidance and analyst's full-year estimates. While second quarter EPS may be less than what analysts are projecting, First Solar will likely be able to increase full-year guidance. First Solar's first-quarter margins of 24.9% easily beat out many of its key competitors such as Canadian Solar (NASDAQ:CSIQ) at 14.7%, Trina Solar (NYSE:TSL) at 20.6%, SunPower (NASDAQ:SPWR) at 23.5% and Yingli (NYSE:YGE) estimated to be about 16.5%. In a previous article, "Why First Solar Remains a Buy", First Solar's valuation using current guidance is in the area of $89 to $128 per share. If First Solar does exceed their full-year guidance this would further argue for a higher valuation. Bottom line, for investors with a long-term view, First Solar remains a buy.
Disclosure: The author is long FSLR. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it. The author has no business relationship with any company whose stock is mentioned in this article.