Category: Blogs

Inclusion of privately-held PV module suppliers within PV ModuleTech Bankability Ratings

During the recent series of articles on PV-Tech - outlining the drivers, methodology and validation of the new PV ModuleTech Bankability ratings for PV module suppliers – one of the key inputs required to generate the overall supplier bankability score (and rating class) was the financial strength of each company.

Indeed, to create a useable metric (bankability score, between zero and ten), it is necessary to know both the financial and manufacturing strengths of PV module suppliers. Therefore, it invariably raises questions as to how companies that are publically-traded and privately-held can be benchmarked.

This issue is of course part of a wider debate on how to benchmark all PV module suppliers from a financial standpoint, regardless of whether they are listed or not.

The topic is highly contentious within the PV industry: most of the PV module suppliers today are headquartered in China. In this country alone, PV module suppliers can be listed on Asian stock exchanges, be part of listed entities within which the solar operations (module supply or other sector related) is a minor business unit, or fall under the category of state-owned/controlled. A few of course are listed on New York stock exchanges, but this is becoming a less-desired option now compared to a few years ago.

However, the main questions I got, following the announcement of the four AA-Rated suppliers recently, went along the lines of: The four AA-Rated suppliers are all listed – did you not consider private companies? Others simply asked how private companies were dealt with within the analysis, partly out of curiosity, but also recognising that this is something no-one has addressed with any discipline or vigour until now.

This article describes how public-traded and privately-held PV module suppliers are benchmarked, as part of the overall PV ModuleTech Bankability analysis. In so doing, I will explain the methodology behind the approach, and show how the results were validated and tested. In particular, I will discuss error bounds associated with the new approach.

Let’s start with the obvious question though: Why bother at all?

Why explaining the methodology is key

The subject of benchmarking public and private companies is not new; anywhere, or at any time. Within the PV industry, due-diligence experts within certain financial circles will no doubt have their bespoke means of doing this. Credit ratings suppliers are also readily available to subscribe to (such as Dun & Bradstreet for example), but often they are relied upon only on a case-by-case basis as a means of minimizing risk associated with any payment terms or contract negotiations.

It is probably fair to say that, until now, no-one has proposed (or likely developed) any robust method to benchmark PV module suppliers, whether public or private. It is likely those that needed some kind of audit trail on private entities have relied upon accessing company accounts as required, ideally third-party verified.

Configuring a means of benchmarking any PV module supplier – at any time and with competitors that are public-traded or privately-held – is a whole new proposition; no doubt, many would suggest ideas but fall short by not disseminating any form of validation or gathering good source data in the first place.

Within the rest of this article, I will explain systematically how this challenge is addressed within the PV ModuleTech Bankability Ratings; why the issue turns out not to be as important as many fear; and finally, why having a practical means of getting up-to-date numbers from privately-held companies rather dominates the problem in the first instance.

Before diving into the details, I will now explain why the whole issue is not quite as critical today, especially when considering large-scale solar module supplier selection.

Framing the importance of privately-held PV module suppliers

In looking at the 14 PV module suppliers that have ‘A’ or ‘B’ ratings (i.e. fall outside the Speculative ‘C’ zone), only two of these companies are privately held today (Trina Solar and JA Solar). The rest are public, or part of parent companies that are listed (and by default act as module supplier unit guarantor). 

Recall that the full list of 14 companies will be revealed during the forthcoming PV ModuleTech 2019 event in Penang, Malaysia on 22-23 October 2019.

In fact, even the two companies (out of the top-14) that are the exception (Trina Solar and JA Solar) have a fully-audited multi-year track record of financial operations until the past 2-3 years when each company delisted: Trina until the end of 2016, JA Solar until the end of 2017. 

During the intervening time period, it would be fair to say that neither company has made any radical changes to its business operations or dominant source of revenues; nor have they diversified operations through any acquisitions or mergers. In short: how these companies operate today is likely to be very similar to the period before they delisted.

Therefore, even if there was no appetite to do any privately-held benchmarking, it would almost suffice to merely pro-rate the performance of Trina Solar and JA Solar for the current (intermediate) time-period (between delisting, up to the expected relisting in China/Asia in the near future), starting with all the historic filings that reside at the SEC.

The good news is that, in the case of these two companies, the pro-rated approach comes out within +/-5% of the method outlined below in the case of privately-held PV module supplier companies. I will likely explain this in a dedicated PV-Tech article in the coming weeks.

Putting privately-held suppliers into perspective

The vast majority of the privately-held PV module suppliers today simply don’t have the manufacturing strength to be ‘A’ or ‘B’ class rated, regardless of their finances. This was one of the fundamental drivers behind the PV ModuleTech Bankability Ratings: to be a top-performer, both manufacturing and financial strength have to be in place going back 12-24 months.

This was critical to differentiate between the companies that can supply to 300-400 MW utility sites, and those that barely produce this quantity over 12 months, or choose to sell in kW-volumes to residential rooftop markets globally without ever adding any new capacity, diversifying globally, or investing in R&D, for example.

Therefore, if the goal is simply to have a short-list of 10-20 PV module suppliers that are in the running for 100-MW-plus utility-scale projects today, it is possible to draw a line through about 95% of the PV module supplier universe, based entirely on manufacturing score (risk of supply).

However, there are still many strong reasons to develop a method to account for any privately-held company: there is no guarantee that the 14 companies currently A/B graded will remain listed in the future; to account for the 5-10 that fall into the border of the B/C grades; and to allow objective analysis and peer-based ranking of the 50-100 module suppliers that are often heard claiming to be on somebody’s ‘tier-1’ module supplier list. (This appears to be one of the most requested issues based on industry feedback to the new PV ModuleTech Bankability Ratings.)

Retaining the Altman Z starting point

I outlined in (previous articles on PV-Tech) why the Altman Z method (Altman, E.I., Journal of Finance, Vol. XXIII, No. 4, pp 589-609, 1968) was adopted as the starting point in the financial score analysis within the PV ModuleTech Bankability Ratings. Here is a quick summary again.

Most third-party industry firms – and plenty that have a form of bias that precludes them from being classed as independent – that have sought to benchmark PV module suppliers until now have used the Altman Z scores, confined to the 1968 ratios/coefficients/zones that were proposed by Altman over fifty years ago to assess the likelihood of bankruptcy for manufacturing companies with turnover above a certain threshold.

No assumption was made whether this is the best approach for PV module suppliers now, or in the past. The method was used simply because it has been, by far, the most frequently cited and adopted approach used across a wide range of sectors. Everyone identifies with it; the methodology is crystal-clear; and anyone can yield known data/scores from company filings, with limited accounting knowledge. The use of it avoids doubt, uncertainty and questioning; it offers a starting point that is credible and workable.

The way in which we took the Altman Z scores for public-listed PV module suppliers (or parent entities) and converted the scores (based on number of standard deviations from a mean value) to a PV-specific 0-10 scoring system, was covered recently in the article on (PV-Tech here). This is probably the first time this type of analysis has been proposed within the sector.

https://www.pv-tech.org/editors-blog/pv-tech-research-ranks-pv-module-suppliers-by-financial-health

The mapping to the new 0-10 band – and making it more aligned to PV industry operating zone terminology – was done entirely to have a means to then combine the financial scores with the manufacturing scores (also ranked 0-10) in the overall bankability calculation. It completed the quantitative nature of the studies, and allowed direct benchmarking to be done across a host of different areas.

The key thing to note though is as follows: the PV financial output in the PV ModuleTech Bankability studies is simply using the Altman Z calculations, then mapped to a 0-10 PV scoring band. Bankrupt companies score zero; chronic performing companies typically score below one. Any score above five (or 50%) is in what I refer to as the ‘Comfort Zone’ of PV module supplier operations, as seen looking at sector-specific data going back 5-10 years.

Therefore, when the challenge of benchmarking privately-held PV module suppliers was addressed, an obvious route was to align with the Altman approach, but look at a practical and approximate variant, using readily-obtainable financial data from private companies.

Reality-check on privately-held companies

Before I proceed to explaining the new approach and its validation, it is perhaps prudent to get some reality on what type of information one can expect to get from any privately-held organization – PV or otherwise – as this ultimately guides what any model can be based upon.

Simply put, if you can’t access full company accounts for any private company at any time, don’t set up a model that uses this as an essential pre-requisite. Don’t base any model also on a reporting requirement that is country-specific, even if there is a platform that one can go to in order to get (often well-outdated) company accounts. Don’t think for a second that private companies’ reporting is remotely in line with listed companies, or is even available covering the past 12-18 months (at best). And don’t use company-specific terminology (such as from a Dun & Bradstreet report for example) that is source-specific and not amenable to widespread benchmarking.

Therefore, the final approach adopted was guided by two key themes discussed above: first, equate with the public-listed specific Altman ratio-discriminant model; second, choose inputs that can be realistically obtained from companies (without relying upon a full audited set of accounts every three months dropping into your inbox!).

The new methodology adopted

It should be mentioned that there is an Altman equivalent for privately-held companies (see for example, Altman, E.I. (2000) in Handbook of Research Methods and Applications in Empirical Finance, Vol. 5). It retains the concept of summing terms based on liquidity, leverage, profitability, solvency and activity, but substitutes the working capital and market capitalization entries with alternate numbers/terms. It requires eight accounting terms to be known (compared to the public listed version that is based on seven terms).

It also creates different scoring values and zones, which makes any benchmarking even more challenging. In fact, the best that can be done is to use each model in isolation; not collectively.

To address the challenge of preserving the Altman ratio-based starting point, I opted to look at a modification of the public-listed Altman equation that reduced the terms and ratios to a minimum, while keeping the error bounds on the final financial score within certain acceptable bounds.

In essence, what is the minimum number of terms/ratios that allows getting scores to within +/- 10% of what they would be had all the terms/ratios been applied. For example, this allows us to decouple the market-cap issue, and not try to find any equivalent value for private companies (such as the book value of equity).

This was done by looking at all the Altman Z scores accumulated in-house for listed PV module suppliers (or parent entities), and identifying the significance of the terms (while also factoring in the type of data that that could be expected from private companies in practice, so they were not seen to be divulging too much confidential information to the outside world).

Importance of identifying different company profile groupings

In looking at all the PV module suppliers (and parent entities) that are publically-listed, there is of course a wide range of different business models in place. For example, you would never seek to find any commonality with SunPower and Yingli Green, or First Solar and BYD. 

Therefore, to establish any short-cut to reaching financial strength scores, it is first necessary to form test groups where selected companies have similar characteristics and can be correctly benchmarked or pro-rated.

As a prelude here, it should be pointed out that the fundamental driver is to benchmark PV module suppliers mainly in China that are not public-listed, such as JA Solar and Trina Solar; but also to account for a grouping of 5-10 other suppliers that may be perceived as competition for non-residential (in particular large-scale) business today outside China. Understanding this is key, as it helps increase the relevance and accuracy of the approach undertaken, and also avoids wasting time on PV module suppliers that are not in the mix for investor-driven projects.

It should be pointed out also that the requirement to broaden the scope for Japanese, South Korean or Taiwanese module suppliers is less important because virtually all PV module suppliers across these regions are part of larger public-listed organizations. The question of India is also an anomaly of sorts, with the financial health of Indian companies in general being somewhat unique; coupled with the fact that few Indian PV module suppliers have any appreciable market-share today outside India in terms of large-scale utility-based site deployment.

Returning then to the China challenge, the list of traceable public-listed PV module suppliers (or parent entities) was divided up into three different focus groups based on company profile (how much PV module supply dominates proceedings) and known operational performance (more on this topic below).

PV module supplier categories

In terms of splitting out the public-listed PV module suppliers, the first category was based upon current (or former) US-listed PV-sector revenue-dominant suppliers, and included JinkoSolar, Canadian Solar, JA Solar and Trina Solar.

The next group included Chinese-listed entities of which PV operations were a part (business unit or otherwise), such as Talesun, Risen, Astronergy and Jinergy. 
The final grouping was specific to Chinese PV module suppliers whose operations today were heavily constrained by ailing financial performance (including for example, Yingli Green).

Other companies were assessed but largely fell into one of the three categories above, or were clear outliers (extreme cases) and should not be used in any benchmarking exercise.

Keeping the goal at +/-10% equivalence to the scoring generated from the initial 5-ratio Altman Z approach, the number of ratios was able to be reduced from five to three for each grouping. (Note, the final three ratios chosen are the same for each grouping selected, not varied.)

Once this was done, the final coefficients for the three chosen ratios (noting that a scaling constant is essential now) were determined using a routine least-squares linear regression analysis, where the ‘residual’ is the difference between the original full 5-ratio Altman derived F score and the new reduced 3-ratio approach.

Therefore, unique weighting coefficients (and constants) are derived for the three different groupings, which should not come as any surprise, given the different operating models at play across the selected company groups chosen. Essentially, different ratios dominate when public-listed companies are performing well, compared to when they are in trouble.

With the three fixed numbers determined (two ratio coefficients and an additive constant) for each of the three ‘test’ groups, the only question was to determine the level of accuracy for the reduced-fit model when applied to the known dataset (public-listed PV module suppliers/parent-companies).

I will explain this more clearly below now, in reference to the figure shown also.

Validation of the model

Being able to show the validation in a simple graphical format was considered fundamental, as this allows anyone to see how accurate the approach is and make their mind up as to the approach used in the ratings system.

To convey this, I have plotted (on the x-axis) the original (full-analysis) Altman Z scores going back 3-4 years for each company, converted to the 0-10 (F) scoring band as explained previously; and on the y-axis, the equivalent 0-10 financial (F) score, using the new shortened variant (three ratios, with the best-fit coefficients/constant) that is to be used to normalize scores from privately-held companies.

The match with the shortened variant to the original Z score value is then a test of the approach validity. This is shown in the graphic below, where the 1:1 line-fit reference would represent 100% accuracy. How much the points deviate then from the 1:1 line-fit is a useful visual measure here of the approach accuracy/validity (without having to rely on statistical terminology for now).

Shown also in the scatter-plot are two dashed straight lines above and below the 1:1 fit. These show the upper/lower bounds at the +/-10% accuracy levels for any given value.

The conclusion here for now is that it is possible to use a reduced model (down to four values, excluding market-cap, and three ratios), that allows accuracy of the final financial score (0-10) within +/-10% accuracy; as long as companies are properly assigned to comparative groupings.

Red flags are rarely wrong

Indeed, the key is knowing which public-listed PV module supplier grouping to benchmark against, for any privately-held company that is likely to be competing globally for large-scale site deployment.

To do this, one does need good a-priori knowledge of the companies in question. This is probably a barrier to many third-party organizations that may have strong accounting or credit-benchmarking skills, but are not connected or knowledgeable enough of the sector. 

It also comes over in many of the legacy Altman Z comparisons done in the sector that these can lack a good dose of reality-checking, and can be comprised of entities no longer actually making solar modules!

Anyone embedded in day-to-day PV dealings will know intimately that there are many ‘signs’ that are indicative of how a PV module supplier is performing from an operational standpoint, without needing to be privy to any financial data!

In this regard, market-related activities tend to be most pronounced when PV module suppliers are going through difficulties, and are often accompanied by issues such as: lack of capacity expansions; reductions in headcount; litigation in respect of contract deliverables; cash-constrained enforcement limiting any technology upgrades to existing capacity; or unexpected and abrupt C-level resignations or forced removal from office.

However, in such extreme cases, it is somewhat futile to seek to benchmark these companies. The risk in engaging with this subgroup is considerable and should be clear to anyone looking at the sector. Or at lead, it should be…

Sadly, the PV industry appears to be highly adept in overlooking often blindingly obvious red-flags that hover around a number of PV module suppliers (many of whom have been included in ‘tier-1’ type listings in the past few years).

Aside from rather questionable due-diligence (assuming this even took place), the reason that distressed and technically-bankrupt PV module suppliers still seem to sell product for large-scale utility projects must surely come down to one simple fact: price (ASP). 

Related to this is product availability (limited pipeline owing to sales inefficiency), and perhaps payment terms that end up being hugely in favour of the buyer who ends up concluding that they have nothing to lose.

But in these cases, it is likely someone will indeed lose out over the next 25 years of desired site yield and targeted performance ratios, even if the original buyer of the modules has long since cashed in on their up-front investment and exposure to risk.

Applying to the private PV module supplier universe

In summary, just four data-points (two from the balance sheet, two from the income statement) allows for fairly good benchmarking of private companies. In fact, it gets somewhat easier if the company is mostly a pure-play module supplier (revenues dominated by PV module sales). In this case, revenues themselves are of course easy to estimate (normally to within +/-10%) based on annual module shipments and blended ASP sector trends.

Ultimately however, the driver is not to rank the 100-plus companies with sub-100-MW annual module shipment volumes (normally confined to their local markets and rooftops), alongside any 5-10 GW global utility-scale module supplier. This comparison benefits no-one at all, is highly misleading, and incorrectly normalizes companies based purely on them being able to produce/source/ship a PV module.

Going forward, I will be looking at validating further the new approach when required, and also doing any changes that increase its accuracy (or indeed flagging any one-off outliers that are simply extreme values and have no means of aligning, such as chronic-performing near-bankrupt companies).

Forthcoming PV ModuleTech Bankability findings

During the next couple of months, expect to read more on PV-Tech about the new PV ModuleTech Bankability Ratings. The goal for now is simply to explain clearly how to rate, rank and assign classification to PV module suppliers, and to provide validation and discussion at every stage of the process.

I will soon be addressing one of the other frequent questions received in the past few weeks: How many so-called ‘tier-1’ companies are unbankable? This question is of course not as black-and-white as many suspect. I will explain more on this topic shortly on PV-Tech, but very quickly for now…

Over the past ten-years-plus, there have been various ‘tier-1’ type listings generated by different organizations, some of the first being muted on PV-Tech back in 2010. Some are widely circulated on social media platforms; others are held in-house and rarely seen. 

More than 50 companies today can be heard claiming to be ‘tier-1’, and within this list, there is a wide range of company strengths and weaknesses.

For example, can one really consider companies like Risen Energy and Winaico (both on ‘tier-1-type’ lists) to be remotely comparable? Therefore, one can understand why there is so much confusion about this somewhat redundant terminology; more on this shortly with supporting analysis and why indeed it may be time for the industry as a whole to finally consign this classification to the history books. 

The next major company-name disclosure from the PV ModuleTech Bankability Ratings will occur during my presentation at the forthcoming PV ModuleTech 2019 meeting in Penang, Malaysia on 22-23 October 2019. Details on how to attend this event can be found here.

Read the entire story

Inclusion of privately-held PV module suppliers within PV ModuleTech Bankability Ratings

During the recent series of articles on PV-Tech - outlining the drivers, methodology and validation of the new PV ModuleTech Bankability ratings for PV module suppliers – one of the key inputs required to generate the overall supplier bankability score (and rating class) was the financial strength of each company.

Indeed, to create a useable metric (bankability score, between zero and ten), it is necessary to know both the financial and manufacturing strengths of PV module suppliers. Therefore, it invariably raises questions as to how companies that are publically-traded and privately-held can be benchmarked.

This issue is of course part of a wider debate on how to benchmark all PV module suppliers from a financial standpoint, regardless of whether they are listed or not.

The topic is highly contentious within the PV industry: most of the PV module suppliers today are headquartered in China. In this country alone, PV module suppliers can be listed on Asian stock exchanges, be part of listed entities within which the solar operations (module supply or other sector related) is a minor business unit, or fall under the category of state-owned/controlled. A few of course are listed on New York stock exchanges, but this is becoming a less-desired option now compared to a few years ago.

However, the main questions I got, following the announcement of the four AA-Rated suppliers recently, went along the lines of: The four AA-Rated suppliers are all listed – did you not consider private companies? Others simply asked how private companies were dealt with within the analysis, partly out of curiosity, but also recognising that this is something no-one has addressed with any discipline or vigour until now.

This article describes how public-traded and privately-held PV module suppliers are benchmarked, as part of the overall PV ModuleTech Bankability analysis. In so doing, I will explain the methodology behind the approach, and show how the results were validated and tested. In particular, I will discuss error bounds associated with the new approach.

Let’s start with the obvious question though: Why bother at all?

Why explaining the methodology is key

The subject of benchmarking public and private companies is not new; anywhere, or at any time. Within the PV industry, due-diligence experts within certain financial circles will no doubt have their bespoke means of doing this. Credit ratings suppliers are also readily available to subscribe to (such as Dun & Bradstreet for example), but often they are relied upon only on a case-by-case basis as a means of minimizing risk associated with any payment terms or contract negotiations.

It is probably fair to say that, until now, no-one has proposed (or likely developed) any robust method to benchmark PV module suppliers, whether public or private. It is likely those that needed some kind of audit trail on private entities have relied upon accessing company accounts as required, ideally third-party verified.

Configuring a means of benchmarking any PV module supplier – at any time and with competitors that are public-traded or privately-held – is a whole new proposition; no doubt, many would suggest ideas but fall short by not disseminating any form of validation or gathering good source data in the first place.

Within the rest of this article, I will explain systematically how this challenge is addressed within the PV ModuleTech Bankability Ratings; why the issue turns out not to be as important as many fear; and finally, why having a practical means of getting up-to-date numbers from privately-held companies rather dominates the problem in the first instance.

Before diving into the details, I will now explain why the whole issue is not quite as critical today, especially when considering large-scale solar module supplier selection.

Framing the importance of privately-held PV module suppliers

In looking at the 14 PV module suppliers that have ‘A’ or ‘B’ ratings (i.e. fall outside the Speculative ‘C’ zone), only two of these companies are privately held today (Trina Solar and JA Solar). The rest are public, or part of parent companies that are listed (and by default act as module supplier unit guarantor). 

Recall that the full list of 14 companies will be revealed during the forthcoming PV ModuleTech 2019 event in Penang, Malaysia on 22-23 October 2019.

In fact, even the two companies (out of the top-14) that are the exception (Trina Solar and JA Solar) have a fully-audited multi-year track record of financial operations until the past 2-3 years when each company delisted: Trina until the end of 2016, JA Solar until the end of 2017. 

During the intervening time period, it would be fair to say that neither company has made any radical changes to its business operations or dominant source of revenues; nor have they diversified operations through any acquisitions or mergers. In short: how these companies operate today is likely to be very similar to the period before they delisted.

Therefore, even if there was no appetite to do any privately-held benchmarking, it would almost suffice to merely pro-rate the performance of Trina Solar and JA Solar for the current (intermediate) time-period (between delisting, up to the expected relisting in China/Asia in the near future), starting with all the historic filings that reside at the SEC.

The good news is that, in the case of these two companies, the pro-rated approach comes out within +/-5% of the method outlined below in the case of privately-held PV module supplier companies. I will likely explain this in a dedicated PV-Tech article in the coming weeks.

Putting privately-held suppliers into perspective

The vast majority of the privately-held PV module suppliers today simply don’t have the manufacturing strength to be ‘A’ or ‘B’ class rated, regardless of their finances. This was one of the fundamental drivers behind the PV ModuleTech Bankability Ratings: to be a top-performer, both manufacturing and financial strength have to be in place going back 12-24 months.

This was critical to differentiate between the companies that can supply to 300-400 MW utility sites, and those that barely produce this quantity over 12 months, or choose to sell in kW-volumes to residential rooftop markets globally without ever adding any new capacity, diversifying globally, or investing in R&D, for example.

Therefore, if the goal is simply to have a short-list of 10-20 PV module suppliers that are in the running for 100-MW-plus utility-scale projects today, it is possible to draw a line through about 95% of the PV module supplier universe, based entirely on manufacturing score (risk of supply).

However, there are still many strong reasons to develop a method to account for any privately-held company: there is no guarantee that the 14 companies currently A/B graded will remain listed in the future; to account for the 5-10 that fall into the border of the B/C grades; and to allow objective analysis and peer-based ranking of the 50-100 module suppliers that are often heard claiming to be on somebody’s ‘tier-1’ module supplier list. (This appears to be one of the most requested issues based on industry feedback to the new PV ModuleTech Bankability Ratings.)

Retaining the Altman Z starting point

I outlined in (previous articles on PV-Tech) why the Altman Z method (Altman, E.I., Journal of Finance, Vol. XXIII, No. 4, pp 589-609, 1968) was adopted as the starting point in the financial score analysis within the PV ModuleTech Bankability Ratings. Here is a quick summary again.

Most third-party industry firms – and plenty that have a form of bias that precludes them from being classed as independent – that have sought to benchmark PV module suppliers until now have used the Altman Z scores, confined to the 1968 ratios/coefficients/zones that were proposed by Altman over fifty years ago to assess the likelihood of bankruptcy for manufacturing companies with turnover above a certain threshold.

No assumption was made whether this is the best approach for PV module suppliers now, or in the past. The method was used simply because it has been, by far, the most frequently cited and adopted approach used across a wide range of sectors. Everyone identifies with it; the methodology is crystal-clear; and anyone can yield known data/scores from company filings, with limited accounting knowledge. The use of it avoids doubt, uncertainty and questioning; it offers a starting point that is credible and workable.

The way in which we took the Altman Z scores for public-listed PV module suppliers (or parent entities) and converted the scores (based on number of standard deviations from a mean value) to a PV-specific 0-10 scoring system, was covered recently in the article on (PV-Tech here). This is probably the first time this type of analysis has been proposed within the sector.

https://www.pv-tech.org/editors-blog/pv-tech-research-ranks-pv-module-suppliers-by-financial-health

The mapping to the new 0-10 band – and making it more aligned to PV industry operating zone terminology – was done entirely to have a means to then combine the financial scores with the manufacturing scores (also ranked 0-10) in the overall bankability calculation. It completed the quantitative nature of the studies, and allowed direct benchmarking to be done across a host of different areas.

The key thing to note though is as follows: the PV financial output in the PV ModuleTech Bankability studies is simply using the Altman Z calculations, then mapped to a 0-10 PV scoring band. Bankrupt companies score zero; chronic performing companies typically score below one. Any score above five (or 50%) is in what I refer to as the ‘Comfort Zone’ of PV module supplier operations, as seen looking at sector-specific data going back 5-10 years.

Therefore, when the challenge of benchmarking privately-held PV module suppliers was addressed, an obvious route was to align with the Altman approach, but look at a practical and approximate variant, using readily-obtainable financial data from private companies.

Reality-check on privately-held companies

Before I proceed to explaining the new approach and its validation, it is perhaps prudent to get some reality on what type of information one can expect to get from any privately-held organization – PV or otherwise – as this ultimately guides what any model can be based upon.

Simply put, if you can’t access full company accounts for any private company at any time, don’t set up a model that uses this as an essential pre-requisite. Don’t base any model also on a reporting requirement that is country-specific, even if there is a platform that one can go to in order to get (often well-outdated) company accounts. Don’t think for a second that private companies’ reporting is remotely in line with listed companies, or is even available covering the past 12-18 months (at best). And don’t use company-specific terminology (such as from a Dun & Bradstreet report for example) that is source-specific and not amenable to widespread benchmarking.

Therefore, the final approach adopted was guided by two key themes discussed above: first, equate with the public-listed specific Altman ratio-discriminant model; second, choose inputs that can be realistically obtained from companies (without relying upon a full audited set of accounts every three months dropping into your inbox!).

The new methodology adopted

It should be mentioned that there is an Altman equivalent for privately-held companies (see for example, Altman, E.I. (2000) in Handbook of Research Methods and Applications in Empirical Finance, Vol. 5). It retains the concept of summing terms based on liquidity, leverage, profitability, solvency and activity, but substitutes the working capital and market capitalization entries with alternate numbers/terms. It requires eight accounting terms to be known (compared to the public listed version that is based on seven terms).

It also creates different scoring values and zones, which makes any benchmarking even more challenging. In fact, the best that can be done is to use each model in isolation; not collectively.

To address the challenge of preserving the Altman ratio-based starting point, I opted to look at a modification of the public-listed Altman equation that reduced the terms and ratios to a minimum, while keeping the error bounds on the final financial score within certain acceptable bounds.

In essence, what is the minimum number of terms/ratios that allows getting scores to within +/- 10% of what they would be had all the terms/ratios been applied. For example, this allows us to decouple the market-cap issue, and not try to find any equivalent value for private companies (such as the book value of equity).

This was done by looking at all the Altman Z scores accumulated in-house for listed PV module suppliers (or parent entities), and identifying the significance of the terms (while also factoring in the type of data that that could be expected from private companies in practice, so they were not seen to be divulging too much confidential information to the outside world).

Importance of identifying different company profile groupings

In looking at all the PV module suppliers (and parent entities) that are publically-listed, there is of course a wide range of different business models in place. For example, you would never seek to find any commonality with SunPower and Yingli Green, or First Solar and BYD. 

Therefore, to establish any short-cut to reaching financial strength scores, it is first necessary to form test groups where selected companies have similar characteristics and can be correctly benchmarked or pro-rated.

As a prelude here, it should be pointed out that the fundamental driver is to benchmark PV module suppliers mainly in China that are not public-listed, such as JA Solar and Trina Solar; but also to account for a grouping of 5-10 other suppliers that may be perceived as competition for non-residential (in particular large-scale) business today outside China. Understanding this is key, as it helps increase the relevance and accuracy of the approach undertaken, and also avoids wasting time on PV module suppliers that are not in the mix for investor-driven projects.

It should be pointed out also that the requirement to broaden the scope for Japanese, South Korean or Taiwanese module suppliers is less important because virtually all PV module suppliers across these regions are part of larger public-listed organizations. The question of India is also an anomaly of sorts, with the financial health of Indian companies in general being somewhat unique; coupled with the fact that few Indian PV module suppliers have any appreciable market-share today outside India in terms of large-scale utility-based site deployment.

Returning then to the China challenge, the list of traceable public-listed PV module suppliers (or parent entities) was divided up into three different focus groups based on company profile (how much PV module supply dominates proceedings) and known operational performance (more on this topic below).

PV module supplier categories

In terms of splitting out the public-listed PV module suppliers, the first category was based upon current (or former) US-listed PV-sector revenue-dominant suppliers, and included JinkoSolar, Canadian Solar, JA Solar and Trina Solar.

The next group included Chinese-listed entities of which PV operations were a part (business unit or otherwise), such as Talesun, Risen, Astronergy and Jinergy. 
The final grouping was specific to Chinese PV module suppliers whose operations today were heavily constrained by ailing financial performance (including for example, Yingli Green).

Other companies were assessed but largely fell into one of the three categories above, or were clear outliers (extreme cases) and should not be used in any benchmarking exercise.

Keeping the goal at +/-10% equivalence to the scoring generated from the initial 5-ratio Altman Z approach, the number of ratios was able to be reduced from five to three for each grouping. (Note, the final three ratios chosen are the same for each grouping selected, not varied.)

Once this was done, the final coefficients for the three chosen ratios (noting that a scaling constant is essential now) were determined using a routine least-squares linear regression analysis, where the ‘residual’ is the difference between the original full 5-ratio Altman derived F score and the new reduced 3-ratio approach.

Therefore, unique weighting coefficients (and constants) are derived for the three different groupings, which should not come as any surprise, given the different operating models at play across the selected company groups chosen. Essentially, different ratios dominate when public-listed companies are performing well, compared to when they are in trouble.

With the three fixed numbers determined (two ratio coefficients and an additive constant) for each of the three ‘test’ groups, the only question was to determine the level of accuracy for the reduced-fit model when applied to the known dataset (public-listed PV module suppliers/parent-companies).

I will explain this more clearly below now, in reference to the figure shown also.

Validation of the model

Being able to show the validation in a simple graphical format was considered fundamental, as this allows anyone to see how accurate the approach is and make their mind up as to the approach used in the ratings system.

To convey this, I have plotted (on the x-axis) the original (full-analysis) Altman Z scores going back 3-4 years for each company, converted to the 0-10 (F) scoring band as explained previously; and on the y-axis, the equivalent 0-10 financial (F) score, using the new shortened variant (three ratios, with the best-fit coefficients/constant) that is to be used to normalize scores from privately-held companies.

The match with the shortened variant to the original Z score value is then a test of the approach validity. This is shown in the graphic below, where the 1:1 line-fit reference would represent 100% accuracy. How much the points deviate then from the 1:1 line-fit is a useful visual measure here of the approach accuracy/validity (without having to rely on statistical terminology for now).

Shown also in the scatter-plot are two dashed straight lines above and below the 1:1 fit. These show the upper/lower bounds at the +/-10% accuracy levels for any given value.

The conclusion here for now is that it is possible to use a reduced model (down to four values, excluding market-cap, and three ratios), that allows accuracy of the final financial score (0-10) within +/-10% accuracy; as long as companies are properly assigned to comparative groupings.

Red flags are rarely wrong

Indeed, the key is knowing which public-listed PV module supplier grouping to benchmark against, for any privately-held company that is likely to be competing globally for large-scale site deployment.

To do this, one does need good a-priori knowledge of the companies in question. This is probably a barrier to many third-party organizations that may have strong accounting or credit-benchmarking skills, but are not connected or knowledgeable enough of the sector. 

It also comes over in many of the legacy Altman Z comparisons done in the sector that these can lack a good dose of reality-checking, and can be comprised of entities no longer actually making solar modules!

Anyone embedded in day-to-day PV dealings will know intimately that there are many ‘signs’ that are indicative of how a PV module supplier is performing from an operational standpoint, without needing to be privy to any financial data!

In this regard, market-related activities tend to be most pronounced when PV module suppliers are going through difficulties, and are often accompanied by issues such as: lack of capacity expansions; reductions in headcount; litigation in respect of contract deliverables; cash-constrained enforcement limiting any technology upgrades to existing capacity; or unexpected and abrupt C-level resignations or forced removal from office.

However, in such extreme cases, it is somewhat futile to seek to benchmark these companies. The risk in engaging with this subgroup is considerable and should be clear to anyone looking at the sector. Or at lead, it should be…

Sadly, the PV industry appears to be highly adept in overlooking often blindingly obvious red-flags that hover around a number of PV module suppliers (many of whom have been included in ‘tier-1’ type listings in the past few years).

Aside from rather questionable due-diligence (assuming this even took place), the reason that distressed and technically-bankrupt PV module suppliers still seem to sell product for large-scale utility projects must surely come down to one simple fact: price (ASP). 

Related to this is product availability (limited pipeline owing to sales inefficiency), and perhaps payment terms that end up being hugely in favour of the buyer who ends up concluding that they have nothing to lose.

But in these cases, it is likely someone will indeed lose out over the next 25 years of desired site yield and targeted performance ratios, even if the original buyer of the modules has long since cashed in on their up-front investment and exposure to risk.

Applying to the private PV module supplier universe

In summary, just four data-points (two from the balance sheet, two from the income statement) allows for fairly good benchmarking of private companies. In fact, it gets somewhat easier if the company is mostly a pure-play module supplier (revenues dominated by PV module sales). In this case, revenues themselves are of course easy to estimate (normally to within +/-10%) based on annual module shipments and blended ASP sector trends.

Ultimately however, the driver is not to rank the 100-plus companies with sub-100-MW annual module shipment volumes (normally confined to their local markets and rooftops), alongside any 5-10 GW global utility-scale module supplier. This comparison benefits no-one at all, is highly misleading, and incorrectly normalizes companies based purely on them being able to produce/source/ship a PV module.

Going forward, I will be looking at validating further the new approach when required, and also doing any changes that increase its accuracy (or indeed flagging any one-off outliers that are simply extreme values and have no means of aligning, such as chronic-performing near-bankrupt companies).

Forthcoming PV ModuleTech Bankability findings

During the next couple of months, expect to read more on PV-Tech about the new PV ModuleTech Bankability Ratings. The goal for now is simply to explain clearly how to rate, rank and assign classification to PV module suppliers, and to provide validation and discussion at every stage of the process.

I will soon be addressing one of the other frequent questions received in the past few weeks: How many so-called ‘tier-1’ companies are unbankable? This question is of course not as black-and-white as many suspect. I will explain more on this topic shortly on PV-Tech, but very quickly for now…

Over the past ten-years-plus, there have been various ‘tier-1’ type listings generated by different organizations, some of the first being muted on PV-Tech back in 2010. Some are widely circulated on social media platforms; others are held in-house and rarely seen. 

More than 50 companies today can be heard claiming to be ‘tier-1’, and within this list, there is a wide range of company strengths and weaknesses.

For example, can one really consider companies like Risen Energy and Winaico (both on ‘tier-1-type’ lists) to be remotely comparable? Therefore, one can understand why there is so much confusion about this somewhat redundant terminology; more on this shortly with supporting analysis and why indeed it may be time for the industry as a whole to finally consign this classification to the history books. 

The next major company-name disclosure from the PV ModuleTech Bankability Ratings will occur during my presentation at the forthcoming PV ModuleTech 2019 meeting in Penang, Malaysia on 22-23 October 2019. Details on how to attend this event can be found here.

Read the entire story

Jinko, First Solar, LONGi and Canadian Solar revealed as top AA-Rated bankable PV module suppliers

The PV-Tech market research team can now reveal that the four solar PV module suppliers meeting AA-Rated bankability status are JinkoSolar, First Solar, LONGi Solar and Canadian Solar.

In the first quarterly release of the PV ModuleTech Bankability Ratings, effective Q3’19, no module suppliers qualify in the highest AAA-Rated band, with just four companies (mentioned above) meeting the next-highest rating of AA.

The company names were revealed this week during a set of two webinars I delivered, explaining the methodology, validation and output underpinning the new PV ModuleTech Bankability Ratings system that scores/grades and benchmarks any PV module supplier in the industry today.

This article provides further details on the number of companies meeting ‘A’ and ‘B’ grade status; the two most important grading categories for bankable PV module suppliers to commercial, industrial and utility PV sites. I will also show historic trends of the four leading AA-Rated companies forming the key results from the Q3’19 bankability rankings generated by the PV-Tech research team.

Why generate a Bankability Ratings system for the PV industry now?

For those that dialled in and listened to either of the webinars I delivered this week, much was spoken about the need for a credible, fully-independent, transparent and professional ranking system to differentiate between the hundreds of companies selling PV modules today; and also within the high-number (40-50) of PV module supplier currently claiming to be on someone’s tier-1-type membership grouping.

Occasionally, during the series of six articles I posted on PV-Tech between 30 June and 8 August 2019 (starting here, and ending here), I alluded to the chronic need for the PV industry to mature from the somewhat academic approach of ranking and benchmarking PV module suppliers. As the stakes get higher (the more the industry moves to 100MW-plus utility site deployment as the norm), the requirement to differentiate between suppliers becomes even more paramount.

During the webinars, I pointed out that the role of such benchmarking platforms should be firmly to allow investors, banks, project developers and EPCs to short-list potential suppliers during the initial stage of the overall module supplier selection process. This should provide clear guidance as to the financial and manufacturing (supply in particular) strengths and weaknesses of potential module suppliers, and crucially, how they benchmark relative to one another today, and over the past 12-18 months.

How banks and lenders (and of course developers and EPCs) generate their short-lists is their business, but few have sufficient (and independent) means of doing this to everyone’s satisfaction, or in keeping track of changes to their select group of PV module suppliers as changes occur within the sector.

It has alarmed me more than once over the years to scan through lists of bank’s ‘approved’ PV module suppliers, only to learn that their lists had not been changed for over two years, and that many on the lists were technically bankrupt and going through Chinese-style asset reassignment to a new owner-entity.

The short-listing – or whitelist generation – phase ultimately sets the basis for proper module supplier due-diligence and final choice for site delivery. The due-diligence phase is where the IE’s, auditors, reliability testing and certification bodies enter the scene. This grouping typically performs steps here to a very high standard, and many in fact are able (by virtue of business volumes accrued over the years) to generate their own ‘rankings’ or ‘top-performer’ lists. However, these are based entirely on work these bodies are contracted to do at any given time.

Therefore, like any competitive analysis, the key issue is how and when to use the output best, rather than to extend optimistically the scope and expectations from any single part of the whole module-selection process.

In the webinars, I used the example of Yingli Green to show how a company, that had been on several tier-1-type or top-performer listings until 2016, had in fact declined to our speculative risk grade category in 2014, and had been in the highest risk-category (C-Rated) since 2017. While Yingli Green was used as a case-study here, there are countless other module suppliers that have been known to shout loudly about being a tier-1 module supplier, but were in fact technically insolvent and exhibiting chronic financial results.

It seems incredible that the PV industry has been operating in this manner for the past 10 years or so, and in particular when solar moved into asset-class territory, and out of the cottage-industry residential FIT-frenzy that stimulated its initial commercial growth phase (where bankability was barely relevant).

Ultimately then, the driver for the PV ModuleTech Bankability Ratings was to provide full and transparent benchmarking of any PV module supplier (selling to commercial, industrial and utility segments globally); without any bias, prejudice, naïve assumptions or loaded pre-conditions of entry into a media-hyped membership grouping.

Fundamental to this entire process is being able to correctly balance the contributions from module supply (manufacturing) and corporate solvency (financial). It requires a broad range of knowledge across technology and production issues and supply globally, in addition to understanding exactly how the whole upstream and downstream segments operate in practice; where the key drivers are profitability, financial health, asset returns and risk. Framed in this way, it is perhaps not a surprise why professionally produced company rankings and benchmarking have been somewhat alien concepts until now within the industry. Hopefully, this will change going forward.

More explanation of using Altman Z score for PV module suppliers

I used the webinar platform also to explain more on the use of the Altman Z model (Altman, E.I., Journal of Finance, Vol. XXIII, No. 4, pp 589-609, 1968.) for PV module suppliers. It was discussed in length during the six-part series of articles recently on PV-Tech, but it remains critical that this part of any credible bankability analysis is done properly (and fully explained).

It should be remembered that the Altman Z score method was proposed over 50 years ago, to predict the likelihood of manufacturing companies (above a pre-set annual turnover threshold) going bankrupt in the near future.

The model has stood the test of time, and is so well known and accepted across different sectors that using this as a starting point is probably still the best place to begin for the PV industry. It allows everyone to know clearly what the starting parameters are (namely the five accounting ratios employed within the Altman Z model).

To back this up, many third-party observers to the PV industry routinely cite Altman Z scores for publically-listed PV companies. However, there is one major problem here: the output shown does not seem to be aligned to how the PV industry actually operates: or else almost every company would be bankrupt by now!

In fact, simply applying the 1968 equations of Altman to the PV industry of 2019 can and does create more questions than answers. Anyone that has read perspective-based studies by Altman and others – looking at the applicability of the Altman Z model – will know that the key thing is to understand the numbers in the context of the industry in question. This is obvious, but it does require the user to actually know how the PV industry works in the first place. This may be part of the problem.

As an example, let me state now where uneducated-use of Altman Z data creates problems in the PV industry today.

In the past few weeks alone, I have come across a couple of statements issued in the context of tier-1 and top-performer analyses from various third-party bodies. Essentially – at the same time – it was stated that only three PV module suppliers were not at risk of bankruptcy today, and then there was a list of 40-50 module suppliers that were assigned tier-1 status.

Quod erat demonstrandum: virtually all tier-1 companies are at risk of bankruptcy in the next couple of years. Judge for yourself if this truly makes sense? Or if it is remotely useful for anyone today trying to decide which module suppliers to short-list for site selection?

During our research phase in understanding the operating results of PV module suppliers (going back 5-10 years), it became clear that the use of the Altman Z ratios was not the actual problem. The mistake related entirely to an incorrect interpretation of the numeric values generated and a misunderstanding of PV module supplier operations and the overall sector.

How we converted the Altman Z scores to solar module financial strength scores (in a 0 to 10 scale) was covered in detail within part five of the PV-Tech article series.

It is worth pointing out that starting from Altman Z scores – and doing statistical analysis to convert these to the 0-10 scale – is done solely to preserve the starting point being the Altman Z approach that has widely accepted.

A more correct method would be to do a PV-specific regression or multivariate discriminant analysis based purely on ratios derived from PV module suppliers during the past 5-10 years. It is likely no-one has even attempted this, but may well be something to consider in the future. This would almost certainly have to discriminate between non-Chinese and Chinese headquartered companies, with a further differentiation between Chinese (listed on US stock exchanges) and state-owned/controlled Chinese entities within which is a subsidiary solar module operation operating as a dedicated business unit.

It may seem I am somewhat labouring the point; but the fact that tier 1 lists have routinely been comprised of companies that either have red-flags in abundance or are operating close to technical bankruptcy, should serve to highlight why correct assessment matters.

The four PV module suppliers with top-performing AA-rating status in Q3’19

It would be fair to say that most of the people that dialled into the webinars this week were doing so to learn which four PV module suppliers were in the top-performing AA-Rating band today.

The graphic below reveals this set of ‘most-bankable’ PV module suppliers today: JinkoSolar, First Solar, LONGi Solar and Canadian Solar.

No PV module supplier scores the highest rating of AAA, and in fact, rarely over the past 10 years, has any PV module supplier been in this band. More on this below.

A total of eight companies have ‘A’ rated status (four AA-Rated as outlined above, and four others with A-Rating). Interestingly, only six companies fall into the three ‘B’ rating categories. Therefore, just 14 PV module suppliers (out of a pool of several hundred in the industry today) have PV ModuleTech Bankability Ratings of B or higher.

All other PV module suppliers (several hundred in total) fall into the ‘Speculative’ zone (CCC, CC, and C), and are therefore assigned as risky propositions when large-scale solar site deployment is under consideration. Normally, the companies in ‘C’ bands either have limited capacity availability, have failed to keep shipment volumes (to non-residential segments) up with overall industry growth rates, or are in poor financial health; or for many it is a combination of both.

To show some further detail on the four AA-Rated top-performers today, the webinar talked through different graphics for these suppliers; this was done to support the overall methodology validation. A version of this is shown in the graphic below now.

I decided that looking at either quarterly ratings (going back 12 quarters from today), or year-end ratings (covering six years including forecasted year-end 2019 ratings), was most useful during the webinar discussion. The choice of annual or quarterly trending for each company depended on which option best showed changes over the defined period. In reference to the graphic above, we can interpret the bankability scores and ratings for each as follows.

  • JinkoSolar: The graphic above shows that Jinko is the only PV module supplier to have AA-Ratings for the past 12 consecutive quarters. It confirms what is seen in the market, especially when one considers how the company has been so effective in gaining market-share globally in the regions that truly matter for utility-scale solar.
  • LONGi Solar: The trending here by year confirms the company as the high-growth supplier of the past few years, moving from CCC-Rated to AA-Rated bands. Key here is the growth trajectory relative to the other AA-Rated companies, suggesting that 2020 is going to be the year that LONGi finally reaches widespread global brand recognition as a PV module supplier.
  • First Solar: The choice of annual data here is done to highlight that First Solar is the only PV module supplier to have been AAA-Rated during the past six years. The cyclic trending is directly related to the operational reset that has been occurring at the company during the past few years, in reprioritizing its manufacturing module supply strategy, and in making massive investments into new Series 6 panel facilities. Assuming profitability targets are hit with regards Series 6 production, it is possible that First Solar could return to AAA-Rated territory during 2020.
  • Canadian Solar: When looking at Canadian Solar’s strategy, it is clear that the company has managed to effectively pursue a dual manufacturing/downstream business within the industry that most other have desired but failed at. This comes over in the graphic above (upstream specific of course) that shows an almost straight-line scoring output (conveyed also in annual figures). It confirms that Canadian Solar has managed to keep its module output at premium-bankable levels, with overall corporate operations benefiting from the timely sale of short-term owned assets (flipping).

Correlating historic trends forms part of the overall validation process of the PV ModuleTech Bankability Ratings methodology. Any PV module supplier, present or in the past, has to be validated in this manner, in order to confirm that the scoring system, and rating bands, are what is seen within the sector. As we reveal companies in the A, and ‘B’ bands, over the coming months, we will show that bankability scores and rating categories for all other PV module suppliers can be understood in this way.

PV ModuleTech 2019 conference becomes go-to bankability forum for the industry

PV-Tech’s annual PV ModuleTech conference was originally configured to provide developers, EPCs and utility-scale investors the opportunity to understand factors relating to module quality, reliability and performance.

With the release of the PV ModuleTech Bankability Ratings, the event now becomes a must-attend platform, where module suppliers will be addressing specifically how their offerings can maintain or elevate them to AA-Rated status or better.

The latest agenda for the forthcoming PV ModuleTech 2019 event was released in the past few days, and can be viewed here. It shows that three of the four AA-Rated module suppliers, highlighted above, will indeed be explaining many of the factors that are implicit in their top-performer rating grades.

Complementing these talks are most of the leading third-party bodies that conduct due-diligence phases on short-listed or selected module suppliers for large-scale project globally; IE’s, auditors, reliability testing and certification bodies. Speakers are confirmed from companies such as PV Evolution Labs, RETC, PI Berlin, STS, Clean Energy Associates, and Kiwa Asia.

The whole event will in fact start with a 45-minute presentation I will give, where all companies occupying bankability ratings in the ‘A’ and ‘B’ categories will be previewed, including benchmarking guides that will assist developers and EPCs guidance going forward.

To attend the PV ModuleTech 2019 event in Penang on 22-23 October 2019, please follow the registration link here.

Read the entire story

Top-4 most bankable PV module suppliers to be revealed on PV-Tech webinars this week

The leading four PV module suppliers, having PV ModuleTech bankability ratings in the highest A-rated category, are set to be revealed on a webinar series to be presented by PV-Tech on 21 and 22 August 2019.

Details on how to register for one the webinars can be found though the link here. The two webinars cover the same material/slide content, but are timed to allow dial-in participation across western and Asian time-zones. Therefore, it is only necessary to register for just one of the webinar sessions.

The new PV ModuleTech Bankability Ratings list

During the past few weeks, the new PV ModuleTech Bankability Ratings methodology has been outlined clearly across a series of six PV-Tech articles that explained how PV module suppliers can be graded (from the top AAA-rated to the lowest/highest-risk C-rated). The final ratings system overview can be found on the final of the six articles here, with links to each of the series features highlighted at the bottom of this webpage also.

The new PV-Tech Ratings system is the first industry analysis that combines each company’s track-record in large-scale global shipments, with its financial health, on a rolling quarterly basis. The analysis uses data collected over 10 years at PV-Tech, across a wealth of manufacturing and financial inputs; these are all covered in the series of six articles on PV-Tech recently.

In contrast to all other tier-based or top/leading-supplier related tables and lists disseminated throughout the industry over the past few decades, the PV ModuleTech Bankability Ratings system uses statistical analysis and modelling, carefully validated against each company’s historic and current status within large-scale PV module deployment.

The driver for the new ratings system has been from downstream PV module users and investors who have been constantly confused about which module suppliers were truly bankable, being able to supply volumes with confidence and having a balance-sheet that reduced the risk of imminent bankruptcy or in-house manufacturing re-organization.

This PV ModuleTech Bankability Ratings system finally allows project developers, EPCs, site investors and asset owners to understand the key investment differences across the range of PV module suppliers bidding to supply to commercial, industrial and utility-scale PV solar sites globally. It is ideal for competitive benchmarking, and shows the strengths and weaknesses of each PV module supplier from each of the key manufacturing and financial perspectives. It is perfect for short-listing potential suppliers, prior to factory audits and reliability tests that are essential to meet specific investor requirements.

Only four PV module suppliers meet AA-rated qualifications

During the webinars this week, I will reveal the leading – and most bankable – PV module suppliers to the industry; and why only a select group of companies (even from among the 40-50 companies currently featuring on tier-type tables) have both the track-record and financial stability to be considered as key contenders for some of the largest utility PV plants currently in planning or construction globally.

The webinars will not only reveal these four companies, but also show historic ratings over the past 3-5 years, on a quarterly and year-end basis. The use of historic checks has turned out to be one of most important validation steps within our entire bankability analysis done in-house at PV-Tech; it allows forward-looking conclusions to be reached on all PV module suppliers, indicative of any ‘red-flags’ that may be forthcoming from a manufacturing or financial standpoint.

The webinars will also show that no company today meets the top-performer AAA-rated grade, and that this has rarely been obtained by any PV module supplier in the past. This is not too much of a surprise however, as I will explain during the webinars, and is in part arising from a still-fragmented landscape where the market-leaders command typically a 10% market-share of module supply; and where some 200-plus companies fight over business globally. It is also arising from the rather precarious financial health of companies that have been overly-dependent on revenue streams from module sales that have been impacted regularly by ASP declines well above cost-reduction measures implemented internally.

The highest ratings grade achieved by a PV module supplier today is AA-rated, and there are only four companies within this top-performer category only. The webinars will therefore focus on why these four companies have top bankability status within the industry, and what can be learned from this in terms of the chasing pack – most of whom are at best CCC-rated or worse (higher-risk).

The graphic below has the names of the top-4 rated PV module suppliers anonymized. They will be revealed for the first time in the webinars, where the historic values are shown to make perfect sense given the status of each company by quarter over the past three years.

PV ModuleTech 2019 to provide further details on suppliers’ ratings status

The forthcoming PV ModuleTech 2019 conference in Penang, Malaysia on 22-23 October 2019 will see many of the AA-rated and A-rated companies presenting and in attendance. This event will start with a 45-minute talk I will deliver, specific to the PV ModuleTech Bankability Ratings. To register to attend PV ModuleTech 2019, go to the link here.

Read the entire story

R&D spending analysis of 21 PV manufacturers

PV Tech has been covering analysis of R&D expenditures of PV manufacturers for over a decade. This blog looks at some of the key trends, which are being detailed in full in the forthcoming edition of technical journal Photovoltaics International.

This is the second sequential year that R&D spending exceeded the US$1.0 billion level, although down slightly from US$1,086.86 million (US$1.08 billion) cumulative annual R&D spending of these 21 key PV manufacturers, in 2017. 

Spending pattern divergence

It is interesting to note that R&D spending doubled over a five-year period, growing from 2013 (US$504 million) to 2017 ((US$1.08 billion). Almost all 21 PV manufacturers under analysis were public in 2013 and all had reached that status by 2014. 

However, a continued trend since 2014 has been the growing number of companies that lowered R&D spending, compared to those increasing spending on a year-on-year basis.

Only two companies lowered R&D spending in 2014, while the crossover point was reached in 2017, when 10 companies reduced spending compared to the previous year. This trend continued in 2018, as for the first time the number of companies lowering R&D spending (11) exceeded the number (9) increasing spending on this front.  

There are also two companies (First Solar and Yingli Green) that have lowered spending for four consecutive years since 2015. 

In the 2018 analysis, two companies (Eging PV and Hareon Solar) have lowered spending for three consecutive years, while two companies (ZJ Sunflower, Wuxi Suntech and ) had lowered spending for two consecutive years.

The spending pattern divergence is primarily driven by the financial condition of some of the companies, such as Yingli Green, Hareon Solar, SunPower and others in the past. 

However, the growing number of companies reducing spending in 2018, is also due to the weaker downstream PV market in China, after the Chinese government announced the ‘531 New Deal’ that put a halt to utility-scale and DG markets, as installations were viewed to have far exceeded plans and the market was subsequently overheating. 

It is also interesting to note that due to this trend, only two companies since 2012 (LONGi Group and Zhongli Talesun) have consistently increased R&D spending, year-on-year.

Other companies that increased spending in 2018, included: JinkoSolar, Canadian Solar, SunPower, Tongwei, Hanergy Thin Film, URE, TZS and Comtec. 

The chart below covers the last five years of annual R&D spending of the 21 key PV manufacturers.

The chart shows a group of five companies (First Solar, LONGi Group, Hanergy Thin Film, SunPower and GCL Group) are clearly separated from the pack by a minimum of over US$100 million in cumulative R&D spending over the last five years. 

Despite First Solar and SunPower dropping in the annual rankings, the changes over a five-year period are less pronounced for First Solar, which remains the cumulative R&D spending leader.

These companies are in the high US$500 million spending range through to the low US$400 million spending range, over the last five years. 

However, SunPower’s position dropped two places in the 2018 rankings and has also been replaced by LONGi Group and Hanergy Thin Film in the last five year period. However, GCL Group was closing in fast in SunPower until a significant reduction in R&D spending took place in 2018. 

LONGi Group and Hanergy Thin Film have been two of the three fastest growing companies in terms of R&D spending, notably in the last three years as shown in the chart.

The chart also highlights that three companies (Zhongli Talesun, TZS and Tongwei Group) have formed a second strong group accelerating R&D spending in the last four of five years. 
Zhongli Talesun, TZS and Tongwei Group have an R&D spending range between the very high US$200 million level to mid US$250 million level. 

Below TZS, things also look interesting as the low levels of spending by Yingli Green in the last three years highlight its declining position in the rankings, while Hareon Solar collapsed.

That means the accelerated R&D spending by Risen Energy, JinkoSolar and Canadian solar in the last two years underscores the ability to move ahead of Yingli Green very soon. 

However, it also indicates that they remain a significant distance behind the second leading pack of Zhongli Talesun, TZS and Tongwei Group. Despite the potential to climb slowly up the ranking, primarily because otherS are falling by the wayside, there is every chance the gap to the second group will widen, locking the two major SMSL members in lower middle range rankings. 

This may also be exacerbated by the expected return of two other SMSL members, Trina Solar and JA Solar into the R&D analysis in 2019, as they return to Chinese stock markets. 

As for those companies below Canadian Solar, five years of R&D spending have mainly highlighted the chasm to the lead and secondary leadership group that increasingly looks insurmountable. 

On a side note, we always get questions over the fact that we have previously not published data and charts related to companies’ R&D expenditure as a percentage of revenue. In the past, a key reason was the almost universal rule that companies’ R&D expenditure as a percentage of revenue lay in the 0.8%-to-3% range. The exception had always been First Solar and Sunpower with higher percentages. 

Plotting that in charts would have just shown a very thick slightly wavering line with two occasional spikes. Basically, not much use at all.

However, there is more variability today, so we have shown two examples of R&D expenditure as a percentage of revenue.

 

In the first sample chart, we have included First Solar and Sunpower to represent the historical high-end of R&D expenditure as a percentage of revenue as well as the inclusion of the two major SMSL’s (JinkoSolar and Canadian Solar) that had been perennial laggards in total annual R&D spending. We have also included a typical example of a relatively small PV manufacturer in the form of Eging PV. 

This selection of companies is a good representation of the historical highs and lows of R&D expenditure as a percentage of revenue.

A key takeaway is that proprietary technology used by First Solar and Sunpower requires, compared to the others, much higher R&D expenditure as a percentage of revenue.

However, being laggards in total annual R&D expenditure as well as a percentage of revenue has historically had little negative impact on JinkoSolar and Canadian Solar, as both have become the largest crystalline PV module manufacturers in the world today. 

In the second sample chart below, we have included three major, China-based integrated PV manufacturers, LONGi Group, TZS and GCL Group and arguably the most closely matched from a business model perspective. 

The main deviation here is that GCL Group can be deemed as the historical major incumbent and has been the largest company in the PV industry by revenue and scale in polysilicon and multicrystalline wafer capacity for many years. 

LONGi Group and TZS have become fast growing companies that have strong R&D spending regimes coupled to strong revenue growth. Indeed, in 2018 both companies R&D expenditure as a percentage of revenue declined at almost the same rates but the reality was that both companies’ total revenue significantly increased over the previous year, while R&D spending increased but clearly at a slower pace than revenue. 

In contrast, GCL Group reported markedly lower revenue in 2018, compared to the previous year. GCL Group companies significantly cut R&D spending, year-on-year, due to financial constraints, causing R&D expenditure as a percentage of revenue to decline. 

Therefore, it could be argued that emerging major players displaying high R&D spending in the last five years have gained significant market share against an historical incumbent. However, GCL has been an investor in TZS as well as SunPower, muddying the waters for a clear-cut comparison. 

Despite continued upheaval in the companies being tracked and those untracked that will surely occur in 2019 as well, it remains somewhat remarkable that over US$1.0 billion was allocated in R&D expenditure in 2018.

Read the entire story

First PV module supplier bankability ratings tool created by PV Tech research team

This article represents the concluding part of a six-part series on PV-Tech over the past couple of weeks, introducing new methodology to allow leading PV module suppliers to be categorized, ranked and short-listed by manufacturing and financial strength metrics; ultimately providing an investor-risk (or bankability) profile of PV module suppliers for non-residential end-market selection.

Collectively, the six articles on PV-Tech.org provide transparency into the methodology used to assign investment risk to PV module suppliers selling to commercial, industrial and utility segments of the industry. The full dataset captures research findings by PV-Tech going back more than 10 years.

The first five articles explained the full methodology behind the two key metrics/scores that are essential to ultimately rank PV module suppliers by way of a bankability ratings tool: manufacturing and financial health.

This article completes the jigsaw. It is the final part that allows readers to understand why a PV module supplier needs to have strong health scores for both manufacturing and financial operations to be bankable, at any given time.

The ongoing research and methodology tracking the bankability rankings of PV module suppliers will be explained in greater detail during forthcoming webinars I will deliver over 21-22 August 2019 (register here), and my opening talk at the PV ModuleTech 2019 conference in Penang, Malaysia on 22-23 October 2019.

Methodology overview

The entire methodology has been founded on the basis that PV module supplier bankability (B), is a function of the supplier’s manufacturing (M) and financial (F) health scores as:

where k is a scaling factor that maps bankability scores to a 0-10 band, m and n are power coefficients derived from regression analysis, and i is a variable that is module-supplier and time-period specific.

The first four articles covered the three key components (supply, capacity, and technology) within the manufacturing health score/ranking, M, variable; and how these components are combined to yield the M values for each company by quarter.

The previous article (number five in the series) focused on the financial health score/ranking, F, variable.

This article now deals with the end-goal of the whole study: forming a scoring system to grade PV module suppliers by bankability health scores between 0 and 10, and how these can be interpreted through a new dynamic bankability ratings tool that is directly analogous to the credit ratings system widely in use today to assess the creditworthiness and risk profile of governments and corporate entities.

By mapping PV module suppliers into clearly-defined, letter-graded categories (ranging from AAA-rated to C-rated), the new PV ModuleTech Bankability tool should finally provide a high-quality, independent and validated means of shortlisting potential module suppliers for commercial, industrial and utility site selection.

Bankability strength (B) score methodology

Let’s return now to the equation that underpins the entire PV module bankability study:

First, I will talk through why the relationship here is actually quite intuitive; in particular, in terms of the bankability score (B) directly scaling (proportional) to powers of each of the manufacturing (M) and financial (F) health scores.

Consider the following. To be bankable, a PV module supplier must first have manufacturing strength (for example through a strong demonstrated track-record of utility-scale shipments, having the capacity strength to sustain future shipment levels at the GW-level, and investing into capex and R&D). Additionally, the company must have demonstrated financial stability and health status over a prolonged period of time.

The two conditions must be met simultaneously. Being a shipment leader, but having poor financial standing, is not a condition that can ever meet leading-bankability status. Conversely, no matter how strong the company’s books are, a lack of product availability (reflected in a low manufacturing health score) ultimately renders the company as ‘unbankable’ (carrying excessive risk).

Weakness in either of the two variables (M and F) therefore precludes commercial-deployment bankability status, for any PV module supplier.

By default, the bankability score (B) must be derived from the product of the two terms, as opposed to any weighted summation. In short, no matter how strong manufacturing or finances are, risk-mitigation during investor due-diligence (supplier short-listing) cannot be compensated by just one of the M or F factors being at the upper end of its respective 0-10 score bands.

Once this premise is accepted, the challenge is simply to identify the scaling constant (k), and the power factors (m and n), within the above equation; and test this against different module suppliers, going back as far as ten years ago if needed.

Solving the bankability equation

The solution to the constants (k, m, n) turned out to be relatively straightforward, when looking at different observables derived from various PV module suppliers during the past few years, in addition to the current industry landscape today.

The constants should be time invariant, meaning that the equation can be applied to any company, at any point in time, and should always be a true reflection of that company’s bankability status, as seen by the industry.

The solution starts by considering the anchor points of the bankability, B, scoring band, from the lowest bankability score (zero) to the maximum (ten). The lower bound is self-explanatory:

The conditions governing the upper band turned out to be slightly more complicated however. In theory, one would expect the maximum bankability score to be obtained as follows:

While this is theoretically possible, it is practically unattainable. If the coefficients are set using this boundary condition then few, if any, PV module suppliers would ever achieve a bankability score above 5/10. (Think of this like some kind of real-work reality-check. Being almost-perfect at two different things, at exactly the same time, never actually happens!)

To address this, it is necessary to remove any one-off outliers (extreme values) in the datasets for each of the M and F scores. In reality, it turns out that the issue is mainly one of removing F score outliers (as opposed to M values). This is now explained in more detail.

When looking at PV module suppliers to the industry over the past decade, leading companies rarely exceed a market-share of more than 10% of shipments. Indeed, to account for 90% of global annual shipments (confined simply to the commercial, industrial and utility segments), it is necessary to consider almost fifty different module suppliers. Moving to 95% coverage demands almost 100 different module suppliers to be analysed.

Therefore, the M scores are relatively bunched today within a band 1-8 (maximum, M=10). As such, there are no clear anomalies (or extreme score outliers) in the M category, especially when accumulating data for 95% of global shipments (about 100 module suppliers).

The variance is revealed however in the F score, with this being emphasized because finance dominates over manufacturing in final bankability ratings of PV module suppliers. (More on this below during the discussion of the power coefficients, m and n.)

The model is therefore adjusted to account for potential outliers, as it pertains to setting the maximum value for B, as shown in the equation above.

The solution to this is to introduce percentiles, which effectively allow the removal of the extreme outliers, with the maximum value of B now governed by the following conditions:

Here, Mv and Fv are the percentile values of M and F across a total or Nm and Nf data entries over a trailing three-year period (t3y), and Pm and Pf are the input percentiles for M and F.

If the percentile for either M or F is set to 100%, this is equivalent to simply using the maximum value seen by any company for either M or F during the past three years. As discussed above, this is not reality; for the F score to make sense in the overall bankability calculation, establishing the F percentile (number of extreme outliners to remove) is essential.

Once the percentile values are derived, the next issue is to set the ratio of the power coefficients, n and m. The solution is achieved relatively quickly by recognizing that financial health is always more important than manufacturing health (when dealing with supplier bankability).

Once n and m are derived, the solution to the scaling coefficient, k, is shown to be:

This completes the methodology overview, and how to generate the three constants that are required to calculate PV module supplier bankability scores (in a 0-10 band), given knowledge of the respective company’s manufacturing and financial scores, at any given quarter-end.

Introducing PV module supplier bankability terms & definitions

Before looking at the bankability scores for a sample group of GW-level PV module suppliers, it is essential that the final part of the model is explained. This involves putting classification, terminology and description to the numbers generated above.

The bankability scores (0 lowest, 10 highest) fall into three pre-assigned grade categories; premium, second-tier, and speculative. The naming here is done such that module suppliers with bankability scores in the range 5-10 are placed in the upper, premium grade; in contrast, the lowest performers (scores between 0 and 2) are in the speculative grade. The middle grade is called second-tier, and has the companies with scores in the range 2-5.

Each of the three grades (premium, second-tier, and speculative) has three different rating grades (or simply ratings). For example, premium includes AAA, AA, and A ratings. This is shown clearly in the image below.

Therefore, the highest performers in the sector are referred to as AAA-rated (read as triple-A-rated). PV module suppliers with this rating label are likely to be setting benchmarks for all other PV module suppliers, across a wide range of operational metrics; not to mention getting a modest ASP premium through having a brand-value delta.

We elected to use letter-designation ratings, analogous to those issued by credit ratings agencies (such as Standard & Poor’s that can be seen here) when they assess the creditworthiness of entities (governments or companies, for example).

Credit ratings affect the likelihood of a company being approved for a loan. By analogy then, PV bankability ratings should affect module suppliers’ chances of being considered for any volume supply contract.

This is a key point; bankability scores/ratings are best used to form shortlists of companies as potential suppliers; third-party agencies then tick the final boxes (factory auditing, pre-shipment inspection, certification, full independent engineer due diligence, reliability testing, inspection, etc.); buyers also of course have price offers from different suppliers, and run their own ROI projections. Having a means however of sanity-checking the companies behind the offers, or knowing which half-dozen to focus on most, is hugely important.

Another good analogy to credit rating nomenclature arises when noting that credit takes time to build up, and as such, high scores/ratings reflect the entity’s track-record (to repay loans). Similarly, the module bankability ratings have a strong supply track-record element built in, with the leading companies being the ones that have the strongest brand ‘value’ in the industry at any given time, built up over time.

Indeed, the grade category ‘premium’ was so named because module suppliers at the upper end of this category are normally the ones with premium brand, resulting in premium ASPs compared to the rest of the industry.

Complementing the image above (explaining credit rating grades), the nine bankability ratings are better understood through a brief description of each one. While simply a two-line summary in most cases, it nonetheless gives a quick check on the generic characteristics of the companies that share unique rating grade lettering.

This is shown in the image below. (Over time, the precise wording will certainly evolve, to be as succinct and accurate as possible.)

Bankability sample group study

During each stage of the six-article series, every effort has been made to validate the model; in particular, for the five scores across the manufacturing and financial health analyses, asking the question “does this match with what is seen in the industry, now and at any point in time during the past 3-5 years?”

Validation of the bankability ratings is undoubtedly the most important part of this exercise, prior to releasing (in the public domain) the findings as they relate to specific companies; and indeed ahead of periodic updates going forward.

By default largely, most articles posted on PV-Tech regarding the PV ModuleTech Bankability ratings over the next six months will feature aspects of ratings validation. In particular, every quarter when the bankability scores are updated for the PV module suppliers, validation must be done.

One of the initial validation checks done on the bankability scores and supplier ratings grades is now discussed and reviewed. Here, eight leading PV module suppliers were chosen, satisfying the following criteria:

• Annual module shipment levels >1 GW for the calendar year 2018
• Listing within Top-10 PV module suppliers by volume, for one or more of the past three calendar years
• (Specific to Chinese-based module suppliers) >10% annual module shipment volumes to overseas markets (non-China) during the past three years
• Claim of being in any third-party agency list, such as Tier 1, etc.
• Brand awareness across decision-making parties overseeing selection of large-scale utility solar site builds in excess of 50 MW

All suppliers satisfied at least four out of the five conditions, with many of them comfortably meeting all five.

The choice of the criteria above was important as part of our initial validation checks. Often, when a host of companies each claims the same thing (industry top-performer, top-10 supplier, tier-1 status, etc.), it is easy to conclude erroneously that there is little to choose between group members. Indeed, this is one of factors that prompted the bankability methodology to be developed in the first place.

There is an extra caveat introduced in the selection process. As opposed to choosing several companies with similar profiles today (in terms of module supply tactics and strategy), the eight companies were selected because they were considered to have different strengths and weaknesses (from the buyer’s perspective).

The graphic below shows the calculated bankability scores for the eight different module suppliers, over a ten quarter period to the end of Q1’19. The module supplier names are anonymized here: the main issue is to check validity at this point, not analyse companies per se.
 

The graphic shown above turns out to be incredibly informative. It confirms – with crystal clarity – that the bankability scores/grades of companies (all producing and shipping GW-plus of solar modules each year) can indeed range from AA-rated to the lowest-grade and highest-risk category at the bottom.

It goes some way to debunk the myth that simply being in some random top-10 or tier-based table carries any real substance (other than membership of a group), or that being a GW-status module producer is something of value in today’s PV industry.

Over the next few months, graphs similar to the one above (colour-coded, with trended bankability scores and rating grades) will be commonplace in articles on PV-Tech, where specific companies are discussed and reviewed. In addition, since the ranking system methodology is ultimately of most use to module buyers/investors, I will provide clarity on how accurate various aspects of the model turn out to be, from a predictive forecasting perspective.

Getting more details about PV ModuleTech Bankability ratings

So far, during the two-week period where I have outlined the full methodology behind the PV ModuleTech Bankability rating system, I have been making a note of the most common questions that have come my way. It seems fitting to list these quickly now:

• Which module suppliers are the most ‘bankable’?
• Where does “my company” rank compared to Jinko?
• How many top-10 or tier-listed companies today are ‘unbankable’?

Over the coming months, hopefully these questions will be discussed in more detail. In the meantime, there are a couple of dates to note for the diary, when I will be speaking more on the new PV ModuleTech Bankability ratings and the suppliers featuring within the premium grade categories.

I will deliver online webinars over 21-22 August 2019 (register to watch here), and give the 45-minute opening talk at the forthcoming PV ModuleTech 2019 conference in Penang, Malaysia on 22-23 October 2019.
In particular, during the forthcoming webinar presentations on 21-22 August 2019, I will reveal for the first time which PV module suppliers fall into the highest PV ModuleTech Bankability ratings grade today!
 

PV Tech’s bankability analysis series links are below

Part 1. PV-Tech research set to reveal investment grades for global PV module suppliers

Part 2. PV-Tech research reveals how to assess PV module suppliers’ capacity claims

Part 3. PV-Tech research establishes technology-leadership scorecard for top-100 module suppliers

Part 4. PV-Tech research reveals ranking tool for manufacturing strength of global module suppliers

Part 5. PV-Tech research ranks PV module suppliers by financial health

Part 6. First PV module supplier bankability ratings tool created by PV Tech research team

Read the entire story

PV-Tech research ranks PV module suppliers by financial health

This article continues our series of features introducing new methodology that allows leading PV module producers to be categorised, ranked and short-listed by manufacturing and financial strength metrics; ultimately providing an investor-risk (or bankability) profile of PV module suppliers for non-residential end-market selection.

This is the fifth of six articles on PV-Tech.org providing full transparency on the methodology used to assign investment risk to PV module suppliers selling to commercial, industrial and utility segments of the industry. The full dataset captures research findings by PV-Tech going back more than 10 years.

The first four articles focused on the manufacturing health of PV module suppliers, with the final ranking model explained in the fourth article of the series, PV-Tech research reveals ranking tool for manufacturing strength of global module suppliers, earlier this week.

This article explains the research conducted by PV-Tech to allow PV module suppliers to be grouped into three zones, through looking at the finances of companies in the sector over the past decade.

The ongoing research and methodology tracking the investment rankings of PV module suppliers will be explained in greater detail during forthcoming webinars I will deliver over 21-22 August 2019 (register here), and my opening talk at the forthcoming PV ModuleTech 2019 conference in Penang, Malaysia on 22-23 October 2019.

Methodology overview

Previous articles revealed the basic relationship between module supplier bankability (B), manufacturing (M) and financial (F) scores as:

where k is a scaling factor that maps bankability scores to a 0-10 band, m and n are power coefficients derived from regression analysis, and i is a variable that is module-supplier and time-period specific. The manufacturing health score/ranking, M – and its constituent elements – were the subject of the first four articles of the series.

This article focuses on the financial score, F, needed to establish the final bankability score/rating, B.

Financial strength (F) score methodology

Most of the PV industry’s downstream community (project developers, EPC’s, and asset owners, to name just three examples) have, for many years, lacked any credible means of understanding and benchmarking PV module suppliers, with more concerns typically raised about the financial health of the companies than the state of its manufacturing output.

Some PV module suppliers (mainly the few that are currently publically-traded on US stock exchanges) tend to be scrutinised the most, but largely from an investor standpoint. For this small grouping, the financial health coverage of these companies by certain financial analysts is often of the highest standard.

However, the trend in recent years has been for leading Chinese PV module suppliers to delist from US stock exchanges. Scanning across the PV module suppliers that comprise the top-100 in the sector today, most are obviously Chinese entities operating within larger (often state-owned or controlled) organizations. As such, there has been a lack of benchmarking analysis performed, not simply regarding the financial health of the PV module suppliers (or parent organization), but across the whole range of financial and manufacturing metrics that come together to complete the bankability picture.

Therefore, it comes as no real surprise that buyers of PV modules (or the institutions that provide the financing) have been operating in a state of constant confusion. Various third-party observers have sought to create temporary solutions to this, including Top-10 shipment tables (typically of mixed quality), or tier-based listings (often with a binary yes/no inclusion and no scoring mechanism, or featuring a broad mix of module suppliers that rarely compete with one another in practice).

When it comes to benchmarking the financial health of PV module suppliers, a common technique that has been routinely applied over the past decade is the model originally developed in 1968 by Robert Altman (Altman, E.I., Journal of Finance, Vol. XXIII, No. 4, pp 589-609, 1968) as a measure of financial distress relative to potential corporate bankruptcy. Indeed, some PV module suppliers themselves have used this type of benchmarking to convey financial superiority against competitors, most often when a graphic produced by a third-party organization paints them in a positive light at the specific point the benchmarking analysis was undertaken.

The use of Altman methodology is not in question; it has been one of the most widely used tools over the past few decades to generate metrics that can track a company’s financial health (sector and turnover specific). The problem so far when applied to the PV industry (in the case of PV module suppliers in particular) is the lack of qualifiers or checks on the validity of the scoring system, and comparison with historic and current due-diligence checks on the companies in question.

The remainder of this article will explain how it is possible to retain the integrity of the Altman scoring system, but adapt this through new methodology for PV module suppliers, in a way that more accurately tracks and benchmarks the financial health of companies operating in this space.

Mapping Altman data for PV module suppliers to a new scoring system

The use Altman’s Z Score model is widespread across different sectors, and, as discussed above, the PV industry is no exception. The model was developed by Altman more than 50 years ago and was originally designed to help predict the likelihood of public-listed companies (with revenues above a certain threshold) – manufacturing specific – going bankrupt. Crucially however, the output is also a means of assessing financial strength (by default, least likely to go bankrupt during a forward-looking period).

First, however, a very quick overview of the statistical model approach. A number (Z Score) is generated for each company based on the weighted sum (linear combination) of key accounting ratios. Multiple ratios are used (multivariate in nature). The choice of ratios– and their weightings – are derived by classifying the observations (output) into different, mutually-exclusive (discriminated), qualitative, a priori groupings (safe, grey, distress zones). This explains why the statistical technique is referred to as multivariate discriminant analysis, or MDA.

The output (Z Score) is a number, and this places the company into one of the three zones; safe (>2.99), grey (1.81-2.99) and distress (<1.81), when using the original model version that applied to listed manufacturing companies. Note that the model used is absent of any constant numerical term that would shift the boundary condition between the distress zone and the non-distressed zones (grey/safe) to zero. The Z Scores represent the number of standard deviations above or below the mean of the original sample set (explaining why scores can be positive or negative).

It is simple and straightforward to generate Altman Z Scores for PV module suppliers (or corporate holding entities). The challenge is how to best interpret and understand them in context.

Indeed, it is hard to imagine that the exact 50-year old zonal categories should be instantly applicable to the PV industry of 2019, which – like any sector – is subject to its own set of dynamics, and is prone to fluctuations arising from rapid cycles in demand, policy, pricing and competition; and furthermore, is a sector that is dominated now by Chinese conglomerates (fully or part state-owned in many occasions) of which the PV module operations can be a very small part.

However, our goal here is firmly to start with the integrity and common-place acceptance of the Altman model; and then to adapt the scores to have direct meaning within the PV industry (as seen over the past decade).

The model proposed has two steps, starting from the Altman Z Scores and ending up with financial health scores (F) that rank companies between 0-10 across new zones (score bands) that align with data observed over the past decade in the PV sector (for leading PV module suppliers).

The basis of the model is illustrated pictorially in the graphic below:

The first stage of the model involves gathering Altman Z Scores for PV module suppliers or parent companies (module warranty ‘guarantors’) where appropriate. The data collection stage here uses quarterly-reported information, as opposed to simply relying on annual (reporting period specific) information only; more on this distinction shortly.

This is where traditional approaches have basically stopped, and chosen to categorize Z Scores of PV module companies within the legacy zones. However, in doing this, typically more than 50% of the top-20 module suppliers (at any given time during the past 10 years) have been scoring at levels that would imply imminent bankruptcy. Furthermore, often the analyses have been limited to one data point per year, and are therefore highly sensitive to one particular accounting ratio that has an extreme value at that particular point in time.

The next step is to set new Altman Z Score limits that represent the ten-year upper and lower values of PV module suppliers to the industry, shown in the centre image of graphic above by the terms Best-in-Class (PV-BiC) for the upper value, and Technically Bankrupt (PV-TB) for the lower value. The technically bankrupt value is taken at the point of ‘no-return’ for the company in question, driven mainly by the fact that some of the Chinese-listed companies have tended to go into a prolonged state of financial distress as they seek to restructure operations (normally offloading assets to other existing or newly-formed entities/vehicles).

Next, it is necessary to adjust the Altman bands (safety, grey, distress) to ones aligned with PV. Here, I have retained a basic 3-level traffic-light colour-coding (green, amber, red) but have named the bands as comfort zone (green), zone of uncertainty (amber), and distressed zone (red). The renaming here is a minor cosmetic adjustment to assign more accurate descriptions to PV module suppliers’ business models: the main change relates to altering the Altman score values that were applied to the boundaries of the safety/grey (2.99) and grey/distress (1.81) zones.

So far, we have not done any adjustment to the Altman Z Scores themselves, other than rework the three zones and place upper and lower bounds, as shown by the mapping of the left image above to the centre one. The renaming of the mid band to zone of uncertainty is done because this terms more accurately sums up what has been seen in the last decade in the sector. PV module suppliers in this band can be seen to either recover operations (move up to the comfort zone) or descend rapidly (ultimately becoming unbankable for project deployment).

The next step to the model turns out to be useful as a visual tool in the final F score output analysis (simple 0 to 10 scoring), but is essential in being able to combine financial health (F) with manufacturing health (M), in order to derive the final bankability factor (B); and ultimately to allocate a risk designation to provide a guide to the investor community when looking at PV module suppliers for site selection.

This involves assigning PV financial scores (F) in the 0-10 band, with F scores above 5 all falling into our comfort zone, and the other zones in the 0-5 range.
Mapping Altman Z Scores (now labelled simply as A for the analysis) to our new PV financial health scores (F) is not linear in nature (in part due to the change in zone ranges and limits imposed above), but can be done relatively easy through a few simple mathematical procedures, as outlined below now.

The mapping process can be expressed as:

The key term above (in the middle) is best mapped through a polynomial of order n (coefficients given by the β terms).

The best-fit solution to the polynomial is now determined by approximating the upper and lower values of F (10 and 0) to successive local minima/maxima, mapping the boundary data sets (distressed/uncertainty and uncertainty/comfort), and finally reducing to a set of simulations equations that can be readily solved.

This now allows us to take any Altman Z Score of a PV module supplier, and map it to a PV-specific 0-10 scale, where anything above 5 is representative of a PV module company operating financially in an industry comfort zone.

A final correction turns out to be needed when looking at the overall analysis, to remove any one-off accounting issues and smooth out seasonal lumpiness that tends to characterize PV module suppliers whose corporate entities have a strong weighting to PV related business. To account for this, trailing twelve months (ttm) averages are taken, based on the previous four calendar year quarters, at each quarter-end.

While several of the Altman based values are cumulative in nature (and therefore largely invariant to doing quarterly or annual scoring calculations), Z Scores can swing from quarter to quarter, or when simply taking single datapoints from annual filings. One key entry here relates to company valuation (market cap) and share prices used to derive this at any given time. It turns out that it is preferential to have more frequent Z Score refreshes (quarterly at a minimum) than rely simply on a Z Score once a year, when using the market cap at this one point in time.

The ttm averaging process can be done either before or after the mapping step. I have chosen to implement this once the F scores have been derived from the quarterly Z Scores.

In this way, the resulting F scores show a clearer long-term trend, and would appear to prevent the often-cited industry practice of taking one-off Altman Z Scores of companies and rushing to quick conclusions that are not representative of long-term factors.

To convey this point, consider the graphic below where two multi-GW suppliers to utility-scale solar over the past 5-10 years are analysed using the new model. We have confined the output here to seven quarters up to end Q1’19, simply to best visualize the issues discussed above.

Each of the time-period specific case-studies in the graphic above serves to illustrate how taking the Altman Z Scores at face-value (and purely based on quarterly datapoints) can lead to erratic – and sometimes erroneous – conclusions regarding the financial health of PV module supplier companies (or parent guarantor entities).

While the above analysis has been confined largely to the mechanics behind the mapping process, the ultimate validity can only come from looking at the results from an extensive dataset of leading PV module suppliers going back in time, and assessing if the current scores reflect what is happening in the market in 2019. This is now reviewed in the final section of the article.

Manufacturing strength (M) score output

The following graph shows a subset of leading companies, where we have chosen mainly GW-scale PV module suppliers (during the past few years) that have been featuring prominently on the bankability short-lists of various lenders. In addition, one module supplier was chosen that was suspected to be in the zone of uncertainty for the past couple of years.

With all of the companies we selected here (except for the one selected that was known to have areas of concern) falling into the comfort zone region (and many just hovering above the uncertainty zone), this would appear to be an excellent validation of the scoring system we intent to use as part of the bankability ratings, to be explained in the final part of the blog series this week on PV-Tech.

Furthermore, F scores are shown to follow more graduated trending, compared to the lumpy output that can arise from benchmarking companies purely on Altman Z Scores at a specific point in time.

In the graph above, we have chosen to highlight/label LONGi Solar, as one of the few companies whose F score has been consistently in the mid-range of the comfort zone band. Again, knowledge of the company in question here is another validation that the model is tracking real-world industry activity.

Another key scenario we had to factor in was the case of a company whose F scores started in the comfort zone, moved downward to the zone of uncertainty, and then fell into the distressed zone. This is clearly an extremely important leading indicator within the PV industry, and in part forms the basis of what has been the most frequently cited concern of the investment community, normally along the lines of: “How do I know the module supplier I have chosen is not going to go bankrupt?”

This issue is of course why so many people gravitate to the scoring system and zonal categories created by Altman in the first place, and apply these to the PV module suppliers or the parent entity as the guarantor.

It would obviously be great if life was so simple that bankruptcy could be predicted, but this is not why the F scoring system for PV module suppliers has been established. It is firmly to track and benchmark the financial health (on a 0-10 scale) of PV module suppliers, within the overall bankability rating system.

Ultimately, companies failing to emerge from the distressed zone should simply be assigned a risk parameter that feeds into any due-diligence exercise, and likely eliminates the company in question from any short-listing in the first place.

Whether the company goes bankrupt is immaterial. Everything is about risk and being able to benchmark this for any initial short-list of PV module suppliers under consideration. Indeed, anyone looking at the PV industry over the past decade knows that bankruptcy is far from a binary transition process.

Previewing the next part of the article series

The final article (part six) will explain how the manufacturing strength (M) and financial strength (F) values are combined to form an overall bankability/risk score for PV module suppliers (B), offering the first fully-researched benchmarking tool for investors, developers, EPCs, and asset owners of global solar PV sites today.

Attend PV ModuleTech 2019 to hear the first presentation on the findings

The full results of the overall study will be released by the PV-Tech market research team before the end of August, with the key findings presented, explained and discussed in the 45 minute opening talk I will be giving at the forthcoming PV ModuleTech 2019 event in Penang on 22-23 October 2019.

PV Tech’s bankability analysis series links are below

Part 1. PV-Tech research set to reveal investment grades for global PV module suppliers

Part 2. PV-Tech research reveals how to assess PV module suppliers’ capacity claims

Part 3. PV-Tech research establishes technology-leadership scorecard for top-100 module suppliers

Part 4. PV-Tech research reveals ranking tool for manufacturing strength of global module suppliers

Part 5. PV-Tech research ranks PV module suppliers by financial health

Part 6. First PV module supplier bankability ratings tool created by PV Tech research team

Read the entire story

PV-Tech research reveals ranking tool for manufacturing strength of global module suppliers

This article continues our series of features introducing new methodology that allows leading PV module producers to be categorized, ranked and short-listed by manufacturing and financial strength metrics; ultimately providing an investor-risk (or bankability) profile of PV module suppliers for non-residential end-market selection.

This is the fourth of six articles on PV-Tech.org providing full transparency on the methodology used to assign investment risk to PV module suppliers selling to commercial, industrial and utility segments of the industry. The full dataset captures research findings by PV-Tech going back more than 10 years.

The first article, PV-Tech research set to reveal investment grades for global PV module suppliers, introduced the research methodology, focusing on the supply strength ranking of PV module suppliers. The second feature, PV-Tech research reveals how to assess PV module suppliers’ capacity claims, explained how these companies can be ranked based on their capacity strength. This was followed by the feature titled PV-Tech research establishes technology-leadership scorecard for top-100 module suppliers that examined technology leadership.

Collectively, these articles covered the three key contributions to the overall manufacturing health score of PV module suppliers. How these parts are combined to form the final manufacturing health output is the subject is this, the fourth article in the series.

The ongoing research and methodology tracking the investment rankings of PV module suppliers will be explained in greater detail during my opening talk at the forthcoming PV ModuleTech 2019 conference in Penang, Malaysia on 22-23 October 2019.

Methodology overview

Previous articles revealed the basic relationship between module supplier bankability (B), manufacturing (M) and financial (F) scores as:

where k is a scaling factor that maps bankability scores to a 0-10 band, m and n are power coefficients derived from regression analysis, and i is a variable that is module-supplier and time-period specific. The manufacturing health score/ranking is expressed as:

where a, b, and c are factor-dependent weightings, scaled to generate manufacturing health scores for each company by quarter (i) in a 0-10 band; S, C and T are the module supplier shipment, capacity and technology ratios introduced above; p, q and r represent power factors derived from regression analysis.

With the S, C, and T scores explained in the first three articles, we now focus on how these are combined to form the M values for each company at each quarter-end.

Manufacturing strength (M) score methodology

The analysis considers the dependence of the three manufacturing variables (supply, capacity and technology), treating the contributions as independent quantitative factors, and finding solutions to the coefficients (a, b and c) and the powers (p, q, and r) shown above.

Of course, in practice, there is correlation between almost everything related to a specific company’s operations (across the full range of inputs within any financial and manufacturing analysis), but the choice of discrete supply, capacity and technology factors within the manufacturing score analysis does have resonance with PV module suppliers’ modus operandi, as discussed now.

Module supply levels are driven by in-house capacity to a large degree, but almost every c-Si module supplier operates with some degree of flexibility where third-party outsourcing is used. For some PV module suppliers, third-party supply dominates module shipments for example.

This was always a feature for Chinese based module suppliers in the past (although rarely disseminated in words), but moved to a new level when Europe and the US imposed tariffs on Chinese-produced modules. This in turn stimulated a multi-GW capacity allocation across various OEM suppliers in Vietnam (and to a lesser degree, the whole Southeast Asia region). Japanese companies also use this model frequently.

This was the principal reason for decoupling the supply (S) and capacity (C) terms. In short, module supply can vary significantly when compared to each company’s in-house capacity or, moreover, capacity conversion (utilization) rates.

Next, the degree of correlation between the technology factor (T) and other manufacturing terms meant that the decision to decouple the dependence of technology from other variables was more obvious. Capex and R&D investments should certainly be treated in isolation for a host of factors in the PV industry. Here are a few reasons now.

Many companies (especially new entrants seeking to commercialize a disruptive technology type) often burn capex and R&D spending for several years, but fail to translate this into any meaningful shipment volumes. Also, Chinese companies have taken capex spending on c-Si additions to incredibly low levels today, and are also prone to have effective in-house capacity increments through absorbing mothballed or zombie-based facilities (zero cash spend) after restructuring of bankrupt entities. Also, there are large variances in R&D allocations, depending simply on the location of corporate headquarters. Most Chinese companies for example spend a fraction of a percentage point on R&D every year; by contrast, western PV companies tend to allocate 3-5% of annual revenues (or higher) to R&D spending, regardless of capex/industry cycles.

To keep the statistical overview to a minimum, I will bypass the workings, and move straight to the validation stages.

One way to understand the dependence of the S, C and T variables is to consider the final model accuracy (goodness-of-fit) for the supply, capacity and technology terms. This is shown in the figure below (two upper graphs and the lower-left one). For each of these graphs, the value of S, C and T is plotted (x-axis) against the original qualitative entries for each company’s M scores (y-axis), with the sold line fit based on the final terms a.Sp, b.Cq, and c.Tr, scaled to the 0-10 scoring band.

The scatter plot variation across the S, C, and T graphs goes some way to backing up the initial discussion in terms of the dependence of S, C and T on the observable, M: strong dependence of the first term (supply, S); low correlation arising from the technology term, T. Essentially, the closer the scatter points are to the line fit, the stronger the dependence.

Basically, the profile of the curves, in each of the S, C, and T plots above, drives the power factor determination for the variables. The coefficients are then derived by combining the power dependency of each variable with the corresponding data fit accuracy and the 0-10 scaling inputs (to ensure all M values fall into this range). Collectively, the coefficients and the power factors form the overall weighting of each variable, S, C, and T. (It should be pointed out that the overall weightings by default must be positive in value. There are many case-studies in the PV industry to support this point.)

The final graphic above (lower-right) provides the last check on the analysis (validation). The fit between the original qualitative M values (observable, y-axis) should ultimately be as close to a 1:1 linear fit, when calculating M using the modelled equation for M (after all coefficients and factors are determined), plotted on the a-axis.

The next major validation of the analysis involves comparing company-specific scores over a multi-year period until now, and seeing if the scores fully capture the manufacturing strengths as seen by the industry at large.

Manufacturing strength (M) score output

Across all the articles so far in the bankability study/methodology, we have sought to show validation of the approach by looking at the overall scores derived from the different metrics used. Supply, capacity and technology graphics to this end were discussed in the previous three articles of the series. Different companies were selected to validate the analysis, at each stage.

The graph below follows this validation process, by highlighting some of the major PV module suppliers to the industry today, and seeing how their manufacturing strength (M) scores have varied over the past five years.

In reference to the above graph, I have highlighted four leading PV module suppliers (all top-10 by module shipment volumes in 2018/2019 to the non-residential end-market segments). Let’s check and validate the analysis for these four companies now.

JinkoSolar’s manufacturing strength growth has indeed been a carefully-considered ramp of in-house wafer, cell and module capacities, with frequent capex allocations and an uptick over the time period to R&D spending. The prudent choice of capacity located within China and across Southeast Asia has also been key to maximizing the value/strength of its manufacturing base at any given time. The net result is evident in the graphic by viewing the gradual Y/Y manufacturing strength score growth, ultimately confirmed by the delta between JinkoSolar’s end-2019 score and the rest of the industry.

LONGi Solar’s manufacturing scores over the past five years confirms the rapid manufacturing strength growth (scoring approximately 1/10 in 2014 to about 6/10 in 2019), and the upward trending in particular during 2019 that is setting up the company to be the second strongest module supplier from a manufacturing perspective, going into 2020.

First Solar’s manufacturing strength cycle is entirely consistent with the company’s manufacturing operations over the past five years, and has been discussed during previous articles within this series. The uptick in manufacturing strength scores from 2017 is in contrast to many of the other multi-GW (c-Si) module suppliers that are having to reset manufacturing operations in 2019 (as shown by the range of downward trend lines in the upper half of the graphic above). The uptick in manufacturing strength scores from First Solar during 2017-2019 is consistent with the company being largely sold-out today for its manufacturing output for the next couple of years, and reveals that the model here is also very good for predictive forecasting.

Finally, I have chosen to look at Risen Energy in the graphic above. Risen is one of several Chinese-headquartered PV module suppliers that has been operating the past few years with multi-GW module supply levels, but somewhat deprioritizing manufacturing investments against module supply volume levels (in-house and third-party blended) and downstream investment/EPC activity. The manufacturing score trend for Risen reflects this accurately, with modest long-term performance, aligned with manufacturing in-house kept at levels needed to support its downstream drive, but not subject to any major swing Y/Y.

We have gone through this type of anecdotal validation for more than 50 of the leading PV module suppliers today. In each case, the score at any given time (year-end, quarter-end) for the company’s manufacturing health (from 0 to 10) not only confirms our previous assessment of the company in question, but also provides new insights from a benchmarking perspective.

At this point in the series of articles, we have now developed a robust methodology to allow us to score all PV module suppliers by their manufacturing strength. We will return to using the M scores in the final article, when pulling together ultimate bankability scores for the PV module suppliers.

Previewing the next part of the article series

Two more articles will be released in the coming days, bringing the six-article set to a close. The next article (number five) will look at how PV module suppliers can be ranked purely on financial strength (the F values in our analysis).

The final article will explain how the manufacturing strength (M) and financial strength (F) values are combined to form an overall bankability/risk score for PV module suppliers, offering the first fully-researched benchmarking tool for investors, developers, EPCs, and asset owners of global solar PV sites today.

Attend PV ModuleTech 2019 to hear the first presentation on the findings

The full results of the overall study will be released by the PV-Tech market research team before the end of August, with the key findings presented, explained and discussed in the 45 minute opening talk I will be giving at the forthcoming PV ModuleTech 2019 event in Penang on 22-23 October 2019.

PV Tech’s bankability analysis series links are below

Part 1. PV-Tech research set to reveal investment grades for global PV module suppliers

Part 2. PV-Tech research reveals how to assess PV module suppliers’ capacity claims

Part 3. PV-Tech research establishes technology-leadership scorecard for top-100 module suppliers

Part 4. PV-Tech research reveals ranking tool for manufacturing strength of global module suppliers

Part 5. PV-Tech research ranks PV module suppliers by financial health

Part 6. First PV module supplier bankability ratings tool created by PV Tech research team

Read the entire story

PV-Tech research establishes technology-leadership scorecard for top-100 module suppliers

This article continues our series of features introducing new methodology that allows leading PV module producers to be categorised, ranked and short-listed by manufacturing and financial strength metrics; ultimately providing an investor-risk (or bankability) profile of PV module suppliers for non-residential end-market selection.

This is the third of six articles on PV-Tech.org that will provide full transparency on the methodology used to assign investment risk to PV module suppliers selling to commercial, industrial and utility segments of the industry. The full dataset captures research findings by PV-Tech going back more than 10 years.

The first article, PV-Tech research set to reveal investment grades for global PV module suppliers, introduced the research methodology, focusing on the supply strength ranking of PV module suppliers. The second feature, PV-Tech research reveals how to assess PV module suppliers’ capacity claims, explained how these companies can be ranked based on their capacity strength.

The third part of the series here outlines PV technology leadership among the top 100 global module suppliers, forming the final part of the inputs that allow manufacturing strength to be fully understood for each company.

The output from the overall analysis – accumulated by the PV-Tech research team over the past five years in particular – will form a key part of my opening talk at the forthcoming PV ModuleTech 2019 conference in Penang, Malaysia on 22-23 October 2019.

The first article introduced the basic relationship between module supplier bankability (B), manufacturing (M) and financial (F) scores as:

where k is a scaling factor that maps bankability scores to a 0-10 band, m and n are power coefficients derived from regression analysis, and i is a variable that is module-supplier and time-period specific. The manufacturing health score/ranking is expressed as:

where a, b, and c are factor-dependent weightings, scaled to generate manufacturing health scores for each company by quarter (i) in a 0-10 band; S, C and T are the module supplier shipment, capacity and technology ratios introduced above; p, q and r represent power factors derived from regression analysis.

This article focuses on the technology score (T) and how this is derived.

Manufacturing technology (T) strength score methodology

The manufacturing technology factor (T) ranks PV module suppliers, by looking at investments into capital expenditure (capex) and research and development (R&D).

Extensive efforts are taken by PV module suppliers to create a perception with customers of technical differentiation and leadership; often this has the appearance of brand positioning, rather than supported by any major performance benefits, or matched by investments into capex (in particular for line upgrades) and R&D.

However, today in the CIU segment – with the exception of First Solar (by virtue of being technology-differentiated) – the non-residential segments of global demand are dominated by p-type c-Si modules, with most companies having near-identical module specifications across their datasheet sets. Performance differentiation tends to come out mainly during reliability and pre-shipment testing stages.

As such, it is inappropriate (and largely misleading) to rank module suppliers based purely on product specifications. This conclusion is further justified by looking at cell and module developments over the past decade, and the ability of new entrants into PV manufacturing to replicate production lines and process flows of the current state-of-the-art p-type manufacturers, and claim technology-equivalence very quickly.

The solar PV industry is certainly not like adjacent technology segments such as semiconductor and flat-panel displays, where a small group of companies control technology roadmaps and market-share, in part arising from major capital expenditure and R&D spending that prohibits the marketplace from becoming crowded.

However, despite these observations, there is a direct relationship between technology investment levels (capex and R&D spending) and sustained module supply leadership to non-residential PV segments. 

Indeed, major shifts in allocations to capex and R&D (either through increased investments or dramatic cuts) still tend to be a leading indicator of changes 12-18 months out across both shipment volumes and operational performance.

The market-leaders also tend to have meaningful investments (across both capex and R&D) over extended time periods, and especially during industry downturn cycles that are normally pre-empted by sharp declines in total capex across the sector.

However, there remains a sizeable group of companies (almost exclusive to China) that operates with practically no R&D investments of note, and that appears to be able to add new capacity volumes at capex levels a fraction of those typical of the sector as a whole. In this respect, simply rationalizing capex investments across the top 100 PV module suppliers can be further complicated when distressed capacity from insolvent entities gets transferred to new or existing manufacturers on a zero-cash basis.

In order to derive technology strength (T) scores (across capex and R&D terms), each of these terms needs to be isolated for all the module suppliers by quarter. This turns out to be highly time-intensive and challenging, and has taken the PV-Tech research team almost five years to establish a robust methodology here.

Very few companies issue PV-specific capex and R&D numbers any more. Those that do tend to have many different contributions to capex (PV and other segment spending allocations, PV project site acquisition and related downstream spending line-items) or lump R&D into whole-company reporting data. Furthermore, no company today segments capex by ingot/wafer/cell/module, unless of course they are pure-play at a single part of the value-chain.

For the capex part of the analysis, we focus only on the cell and module stages of the value-chain, and remove polysilicon, ingot and wafer capex for any companies that have backward-integrated capacities in-house. The rationale for limiting capex to only cell and module stages is similar to that outlined in the previous blog where we outlined the factors driving PV module supplier capacity (C) values. We went through various iterations of capex segmentation, checking at each time the results of the analysis against the perceived market brand/positioning/bankability (the qualitative dependent variable).

The clear outcome from this was not to confine capex for PV module suppliers at the module assembly stage only, but to include company spending on both cell and module stages. This turns out to be extremely powerful in the context of the study, because of the trend of cell-dominant players (mainly Chinese companies) using this as a stepping-stone to having global module supply aspirations.

This has happened frequently over the past 20 years, going back to the move by JA Solar from pure-play cell maker to global module brand supplier status. Other similar tactics have been deployed with success by LONGi Solar (prioritising the move from wafer-leader to cell production and module supply driven), and GCL (through GCL-Poly and GCL-SI affiliate operations).

The fact that multi-GW cell producers in China have been following internal mandates to play on the global (overseas) stage at the module supply level should not come as too much of a surprise. 

Indeed, today, the climate with China as a whole is only supportive of companies seeking to sell the China technology-leadership brand outside the country, and the solar sector is prime for this type of behaviour.

Therefore, module suppliers that have recently expanded upstream (cell to cell/module)  – and have established multi-GW cell production status (and are already embedded in the cell supply chain to the leading PV module suppliers) – are likely to be among the next wave of companies from China what will evolve into global module brands going forward. More on this later in the article. For now, the key takeaway from a capex standpoint is that is it critical to look at both cell and module capex by company, when benchmarking company technology ranking scores.

The analysis starts by going through all module suppliers, isolating total PV capex (sometimes called manufacturing capex) by quarter. The next step is to remove PV capex allocations to polysilicon, ingot or wafer stages, where necessary, leaving cell and module contributions; Capex(CM).

The quarterly capex data for each module supplier is then simply the sum of cell and module investments. No factor weightings are applied to cell and module contributions here, due to capex into cell and module capacity generally being equally advantageous in terms of technology-related health of any specific module supplier organization.

Capex is included across facilities, maintenance, upgrades and new production lines; in reality today, site costs for leading c-Si makers (especially when new sites are in China or Southeast Asia) are very low and often partially gratis on somewhat barter terms, when investing in new countries or new Chinese provinces.

For reference, when dealing with thin-film capex, it is necessary to normalize (or derate) capex allocations due to the much higher spending in recent years when adding GW-levels of new thin-film capacity. Derating factors are therefore applied (and adjusted annually) to thin-film capex contributions to allow direct comparison with typical values of leading Chinese c-Si module suppliers when adding GW-levels of new cell and module capacity. (If this is not done, the capex levels by First Solar, for example, would be excessively high and not aligned with new capacity levels ultimately coming online.)

For each module supplier (i), the respective quarterly cell/module capex values are then converted into t24m sums (previous eight quarter totals, at the end of any given quarter). This is essential when ranking the companies, as capex on a quarterly basis tends to be lumpy in nature. In addition, return on capex investments in the PV industry can be anywhere from six months to three years, depending on the company/country in question.

Capex scores (in the range 0-10) for each module supplier (by quarter) are then established by analyzing the data distribution and normalizing each quarter (u) for correct benchmarking purposes. 

In fact, capex must be normalized in this way, in order to rank companies regardless of where the industry as a whole is on capex cycles: capex is by nature cyclic, and is normally one of the first leading indicators of a pending market downturn (as related to listed company valuations and profitability). Therefore, companies investing during downturns see higher scores, regardless of the total capex levels at the time.

The analysis of R&D spending by the module suppliers follows the same methodology as capex, discussed above. In contrast however to capex allocations being restricted to the cell and module stages of the value-chain, R&D entries are based on total PV spending, with the exclusion only of polysilicon contributions; R&D(PV).

This involves quarterly PV R&D spending being assigned to each module supplier, t24m values being determined at the end of each quarter, and scores being converted to a 0-10 scale based on normalization each quarter (v). Similar to the discussion on capex investments during downturns, priority is given to companies investing in R&D.

To establish the final technology-based quarterly score (T) by module supplier (i) for any given quarter, the two scores (capex and R&D spending) are combined through applying weightings (prioritized to capex), denoted by the t coefficients below. The last step is again to normalize each quarter to a 0-10 band, to standardize each of the S, C, and T contributions for the overall manufacturing health/strength parameter, M, through the quarterly coefficients k below.

The final expression then for the technology strength value for each company by quarter, can be written as:

Manufacturing technology (T) strength score output

Similar to the analysis covered in the previous articles within this series, the manufacturing technology (T) study can be adapted to look at only capex and R&D, or through further manipulation, capex across the different manufacturing zones discussed in the capacity feature.

The graphic below shows PV module supplier technology scores from the top 100 companies in the industry today, with a few highlighted again to convey key trends arising from the analysis.

Capex and R&D spending are of course accounting terms, and investments are mostly connected to company turnover, profitability and investment-related financial metrics. During our research phase in building up the final module supplier bankability whitelist for the investment community, we explored the option of amalgamating capex and R&D within the financial part of the study, and isolating ratios such as return-on-capital-employed or return-on-invested-capital. This ultimately proved overcomplicated, and detracted from having a single universally-recognized financial health metric that ensured full audit trail understanding.

Other reasons for keeping capex and R&D spending within the manufacturing health section are explained in more detail when looking at the overall manufacturing analysis (M); the subject of the fourth article in the series.

The discussion on this now is in reference to the graphic shown above. Clearly, strong technology scores are more likely to come from companies that have healthy financial health status. This in part explains the segmentation of SunPower and Shunfeng in the graphic above, compared to the three other companies highlighted here: Tongwei, First Solar and LONGi Solar.

In fact, analysis alone of the technology strength score is also a strong leading indicator for future company success (or declines) when it comes to module supply levels and ultimate investment risk. This should come as no surprise, since capex and R&D investments are both investment-based and are placeholders for increased productivity, higher-performing product, etc. When time allows, it is our intention to return to the technology analysis here, and explore correlations over the past 10 years to see how true this hypothesis is.

However, simply looking at the trends between 2014 and 2019 of the highlighted companies above suggests that this term could be incredibly useful in isolation to investors seeking to project future investment risk for module makers during short-listing and final due-diligence ahead of supplier selection.

LONGi Solar’s upward trend, to become the clear technology-leader in our analysis (from 2017) is a prime example in this regard. The company was still setting out its transition from leading-wafer supplier to global module entrant during 2015-2017, and it is only in the past couple of years that LONGi has been added to the global bankable brand short-list of most-suitable suppliers.

First Solar’s cycle above is also fully consistent with the 2016 inflection point in the graphic coinciding with the start of the Series 6 expansion plans that will have the greatest impact on the company’s expected upside growth (on many counts) during 2020 in particular.

Finally, returning to the issue I discussed before about the value of extending the capex analysis of PV module suppliers to be inclusive of both cell and module capex investments, the inclusion of Tongwei above is another clear sign that the technology analysis alone could be one of the most important means of forecasting company-specific investment risk levels going forward.

Tongwei today is the leading cell producer globally, and the first to make a business out of being a 10-GW-plus annual producer; and with the same aggressive capacity expansion (and market-share growth) model that was seen several years ago by LONGi (from wafer supply standpoint). Tongwei’s move to module-status is lost in the news only because any GW-based plans pale into insignificance compared to the 10-20GW cell capacity growth strategy being unfolded.

Clearly, the overall manufacturing health score for PV module suppliers has other (and more important) contributions coming in particular from the supply (S) terms that is driven mainly by t24m module shipment globally to commercial, industrial and utility applications. In this respect, the overall (module supplier) score for Tongwei would be reset based on this factor, but as soon as module supply increases – and overseas projects are supplied to – the immediate bankability of the company as a credible module supplier becomes very real (assuming of course financial health is in place).

Previewing the next part of the article series

The next article in this series will focus on the overall manufacturing health score (M), the contributions of the individual supply (S), capacity (C), and technology (T) scores, the results of the regression analysis, and validation by way of comparing the results of each company between 2013 and 2019.

Ultimately, for companies to be ranked in the top categories of bankability for module supply (lowest investment risk) for large-scale solar deployment, both manufacturing and financial health status must be in place at the same time, and shown to be stable over periods longer than just the trailing quarter under investigation.

Attend PV ModuleTech 2019 to hear the first presentation on the findings

The full results of the overall study will be released by the PV-Tech market research team before the end of August, with the key findings presented, explained and discussed in the 45 minute opening talk I will be giving at the forthcoming PV ModuleTech 2019 event in Penang on 22-23 October 2019.

PV Tech’s bankability analysis series links are below

Part 1. PV-Tech research set to reveal investment grades for global PV module suppliers

Part 2. PV-Tech research reveals how to assess PV module suppliers’ capacity claims

Part 3. PV-Tech research establishes technology-leadership scorecard for top-100 module suppliers

Part 4. PV-Tech research reveals ranking tool for manufacturing strength of global module suppliers

Part 5. PV-Tech research ranks PV module suppliers by financial health

Part 6. First PV module supplier bankability ratings tool created by PV Tech research team

Read the entire story

PV-Tech research reveals how to assess PV module suppliers’ capacity claims

This article continues our series of features introducing new methodology that allows leading PV module producers to be categorised, ranked and short-listed by manufacturing and financial strength metrics; ultimately providing an investor-risk (or bankability) profile of bankable module suppliers for non-residential end-market selection.

This is the second of six articles on PV-Tech.org that will provide full transparency on the methodology used to assign investment risk to PV module suppliers selling to commercial, industrial and utility segments of the industry. The full dataset captures research findings by PV-Tech going back more than 10 years.

The first article, PV-Tech research set to reveal investment grades for global PV module suppliers, introduced the research methodology, focusing on the supply strength ranking of PV module suppliers. This (second) feature focuses on the capacity factor used within the bankability rankings study.

The output from the overall analysis – accumulated by the PV-Tech research team over the past five years in particular – will form a key part of my opening talk at the forthcoming PV ModuleTech 2019 conference in Penang, Malaysia on 22-23 October 2019.

Methodology overview

The first article introduced the basic relationship between module supplier bankability (B), manufacturing (M) and financial (F) scores as:

where k is a scaling factor that maps bankability scores to a 0-10 band, m and n are power coefficients derived from regression analysis, and i is a variable that is module-supplier and time-period specific. The manufacturing health score/ranking is expressed as:

where a, b, and c are factor-dependent weightings, scaled to generate manufacturing health scores for each company by quarter (i) in a 0-10 band; S, C and T are the module supplier shipment, capacity and technology ratios introduced above; p, q and r represent power factors derived from regression analysis.

This article focuses on the capacity score (C) and how this is derived.

Manufacturing capacity (C) strength score methodology

The manufacturing capacity factor (C) ranks PV module suppliers, by looking at in-house cell and module effective quarterly capacities across different global PV manufacturing zones, and factoring in the access these manufacturing zones have at any given time to global module supply end-markets.

This type of analysis turns out to be incredibly insightful, and explains why the common practice of PV industry observers to consider one capacity number (often based on unsubstantiated nameplate single-entry data points) is both misleading and inappropriate within a changeable trade-barrier influenced global landscape where origin-of-manufacture is of the utmost importance.

First, I will explain some of the key issues related to capacity within the PV industry today.

Excluding thin-film PV technologies, all c-Si based module suppliers operate with different levels of backward-integration capacity, across cells, wafers and ingots. While various Chinese companies have subsidiary operations that produce polysilicon, no company today in the PV industry operates with a full value-chain model where every component is made in-house.

Across the ingot-to-module stages, the most critical parts in terms of module supply are cell and module production. Wafer supply has now become a China-centric commoditised offering, and crucially this part of the value-chain has been exempt from trade-related origin-of-manufacturing. Indeed, with more than 95% of c-Si wafers produced today within China, there is even less prospect of wafer supply being incorporated into any meaningful tariff-related policy.

Moreover, such trade-related duties have typically focused on the cell and module segments of the value-chain. Therefore, in assessing company-specific capacity-based strength metrics, it is the cell and module stages that are important to evaluate. This becomes further justified when recalling that module specifications are mostly driven by cell performance and quality.

In addition to the need to have high levels of in-house cell and module capacities, most Chinese c-Si module suppliers have routinely relied upon strong third-party outsourcing of cells and modules.

At the two extremes of the Chinese module supply practice are the fully-integrated in-house supply-constrained model, and the so-called ‘fabless’ alternative.
Leading multi-GW module suppliers – that adhere to using only in-house produced cells and modules – are the exception within the PV industry today. Indeed, the practice of relying on third-party companies for production has only increased in recent years, with Southeast Asia based companies often being called upon when shipping modules without prohibitive duties to the US (and until recently, to Europe).

By default, the only multi-GW thin-film producer (First Solar) is mandated to use in-house product, as a result of being technology-differentiated; this is an exception to the rule today, with the company being the only truly-differentiated alternative to non-residential (commercial, industrial and utility, or CIU) applications.

The fabless model – where all manufacturing is outsourced – remains popular within adjacent technology sectors (in particular, the semiconductor industry), but has been largely ineffective until now within the PV industry. The only company that sought to pursue a cell/module fabless model was SunEdison several years ago.

Other companies (including several Japanese module suppliers) did shift to strong outsourcing in an attempt to stay competitive (quasi-fabless), but such efforts were largely short-lived. In reality, a host of factors has prevented the fabless model working in the PV industry, including single-digit production margin constraints, and the need to quickly adjust to market dynamics resulting from technology and tariff related issues.

In assessing the relative strengths of module suppliers, in terms of manufacturing capacity, it proved necessary to fully understand how much effective cell and module capacity was owned by each company, and across which manufacturing zones globally. In particular, the levels of cell and module capacity by zone turns out to be crucial in assessing which end-markets are on offer through in-house cell and module production. The growth of cell and module capacities across Southeast Asia in recent years illustrates this point succinctly.

The analysis of manufacturing capacity strength (C) starts by splitting each company’s effective cell and module capacities across eight pertinent manufacturing zones globally: China, Taiwan, India, Japan, Southeast Asia, the US, Europe and the Rest of the World (RoW). These locations are chosen in part from a legacy manufacturing standpoint (in particular Japan), and crucially because trade-related import barriers tend to differentiate between cell and modules produced and shipped from these areas (origin-of-manufacture).

This type of segmentation is also important because ultimately the strength of in-house capacity depends on the served addressable market (SAM) available; namely which end-markets are absent of prohibitive import conditions at any given time.

This has been most pronounced in the case of China over the past decade. As such, it can be concluded a Chinese company having multi-GW of in-house cell and module capacity only in China sees a lower SAM for its factory output, compared to a competitor that has domestic and overseas manufacturing capability. While this alone is a simplistic case-study, the reality is a rapidly evolving global landscape that needs a robust methodology constructed in order to deal with changes by manufacturing zone and regional end-market supply (export and import).

The first part of the analysis here therefore requires the effective quarterly cell and module capacities (Cap) by quarter, for the top-100 module suppliers globally, to be segmented into each of the eight manufacturing zones (p=1…8), as outlined above.

For reference, effective capacity refers to the available/ramped capacity and its maximum productivity levels if operated 24/7. Very few fabs operate under these conditions in the PV industry, with only First Solar having a consistent track-record of fab productivity in the 95-100% range over a multi-year time period. The key issue here though is to differentiate between erroneous and misleading capacity figures that are all too commonly used within the PV industry, such as nameplate capacity or ‘available’ capacity (which is often no more than an ambitious summation of in-house and third-party capacities that can be called upon if needed).

Effective capacities in general go up and down every quarter, due to efficiency/power improvements at the module level, technology upgrades, debottlenecking, routine maintenance, or temporary factory mothballing. 

A key indicator of capacity definition used within the industry by any company/observer can be understood quickly, by noting that effective capacity figures are different every quarter: erroneous nameplate or ‘available’ capacity figures are often quoted to the nearest 100MW or GW and don’t change, by comparison. 

Another guide may come from related utilization rates cited that exceed 100%, reflecting inaccurate capacity allocations: capacity conversion is a more accurate means of quoting utilization rates in practice.

With the eight segmented module capacities (Cap) by manufacturing zone location established, the next stage is to determine how much effective in-house cell capacity is available to each of the module suppliers in these zones. 

This is important as it allows us to differentiate between modules produced by any company (in any zone) using in-house cells (IHC) or third-party cells (TPC). As discussed above, a major part of module quality, performance and reliability can be traced back to the origin of cell manufacturing; additionally of course, module trade-barriers routinely extend to cell component origin-of-manufacture.
The resulting module capacity (Cap) value by company (i) by manufacturing zone (p=1…8) can therefore be expressed as:

where the c coefficients are weighting factors that depend on whether module capacity uses in-house cells made in the same manufacturing zone, Cap(IHC), or by third-party cell producers, Cap(TPC).

This clearly promotes the strength of module suppliers that use in-house made cells only, produced local to module assembly activity. This is entirely consistent with how the industry operates today, and is a key issue for any investor-led due-diligence process as it pertains to module quality and bill-of-materials integrity.

The weighting factors, c, are qualitative data entries by nature, and can be adjusted by quarter or by manufacturing zone depending on how important in-house vertical integration of cells and modules is. The precise relative weighting between the factors turns out to be somewhat secondary within the overall bankability studies, and as such it is not essential to overcomplicate this part of the analysis, other than to have a means of differentiating between IHC and TPC supply-chains.

For reference, First Solar’s manufacturing is split up into cell and module capacities, although a single thin-film line incorporates the equivalent of c-Si cell/module stages. By default therefore, all First Solar product is cell/module matched, as it is for other thin-film makers in general.

The next stage of the analysis is the most important and valuable part of the overall capacity strength factor studies, because it introduces the impact of trade (export) restrictions on modules produced within any of the eight (p=1…8) manufacturing zones shipped to any of the six (j=1…6) end-market regions (Reg) that were introduced in the first part of the article series before.

Simply put, the value of having module (and cell) capacity in any part of the world is only as useful as the SAM available at any given time, factoring in trade-barriers that tend to be somewhat binary in nature when it comes to market accessibility (either the end-market region is ‘open’ or ‘closed’ with limited scope for any middle-ground).

One of the most insightful example of this relates to Chinese cell/module capacity (one of our eight manufacturing zones) and shipments into Europe (one of our six end-market regions). Prior to the establishment of the minimum import pricing (MIP) constraints imposed by the European Union on Chinese imports, Europe was fully accessible to Chinese produced modules. Once MIP was imposed, shipments from China to Europe collapsed to near-zero. Then when the MIP was removed, Europe than became fully-accessible again to Chinese produce.

In order to restate module capacity by company/quarter within the eight manufacturing zones globally, each capacity value (obtained through the summed term above) is multiplied by an end-market ‘access-related’ factor that is both manufacturing region and end-market specific.

To do this, the module sum factor (above) for each module supplier is multiplied by a quarterly-variable term based on combining the total quarterly CIU demand (Dem’) (for each of six end-market segments (j=1…6) for shipments) with a qualitative access percentage term (Access) that defines the availability of end-market j for module production in manufacturing zone p at any given point.

For example, returning to the Chinese module capacity example above, where the manufacturing zone is China, then China-specific access percentage terms would be 100% for China (naturally), and near-100% for regions such as Europe (today), Japan and most of the RoW sub-segments. By contrast, percentage levels would be very low for shipments to the US market, and fluctuating for supply to the Indian market.
The pro-rated regional contributions for each manufacturing zone are finally scaled by dividing by the total global CIU market demand in each quarter. This overall scaling factor can be expressed as:

Therefore, this type of analysis not only adjusts module capacity by manufacturing zone, it also scales the size of served end-market by the importance of each region, by looking at the ratio of the demand (CIU) from that region and the total CIU demand each quarter.

The steps above turned out to among the most insightful within the overall study, in building up the manufacturing strength of PV module suppliers in the industry. This analysis clearly takes capacity assessment (previously largely misunderstood and erroneously presented) to a new level of scrutiny, and finally allows for capacity to be valued based on where the product is made, how much incorporates in-house and local cell supply, and which end-market is being targeted; for all module suppliers, by quarter.

The final capacity score (C) of each module supplier is then simply the sum of the scores derived for all eight manufacturing zones, by quarter. The full equation can be written finally as:

where k is a variable quarterly scaling factor, to map capacity scores into a 1-10 band; again based on distribution and standard deviation checks done by quarter.

It should also be pointed out that the capacity analysis here is confined to quarter-only data points, and not any trailing for forward-looking time periods (as was entirely valid for the supply/shipment analysis before). This is done because capacity strength is an instantaneous variable (has a specific value at any given moment in time) that is entirely dependent on regional trade-access conditions.

Manufacturing capacity (C) strength score output

Similar to the analysis covered in the manufacturing supply (S) analysis outlined in part 1 of this series, the manufacturing capacity (C) study yields a vast quantity of benchmarking for different PV module suppliers, when isolating manufacturing zones and end-market shipment regions over different time periods.

For now however, we look at the final C values for PV module suppliers, choosing to show year-end values for simplicity, although the analysis of course tracks scores by quarter.

The graphic below captures PV module supplier scores from the top 100 companies in the industry today, with a few highlighted again to convey key trends arising from the analysis.

While the highlighted companies show a range of different fortunes for PV module suppliers’ capacity strength factors, the most interesting ones to discuss are those of Canadian Solar and Hanwha Q CELLS. Each of these companies has maintained capacity effectiveness by having a flexible strategy that allows modules made in different manufacturing zones to be prioritised at different times, depending on which end-markets are favourable to origin-of-manufacture. 

The case of Hanwha Q CELLS is perhaps the most robust in this regard, with the company able to adjust product availability from China, Korea, Malaysia (and now the US) as and when trade conditions apply; only companies with strong balance sheets can purse this strategy for any meaningful length of time.

Previewing the next part of the article series

The next article in this series will focus on the last term in the manufacturing health score analysis; technology (T). This allows us to incorporate R&D spending (confined to PV operations) and capex (restricted to cell and module stages) for each of the 100-plus module suppliers under review, again analysed by quarter.

The conclusions will be shown during the article to reveal the unique way in which R&D spending and capex (the hallmarks of technology leadership in other technology sectors) impact on PV manufacturing strength and module bankability rankings.

Attend PV ModuleTech 2019 to hear the first presentation on the findings

The full results of the overall study will be released by the PV-Tech market research team before the end of August, with the key findings presented, explained and discussed in the 45 minute opening talk I will be giving at the forthcoming PV ModuleTech 2019 event in Penang on 22-23 October 2019.

PV Tech’s bankability analysis series links are below

Part 1. PV-Tech research set to reveal investment grades for global PV module suppliers

Part 2. PV-Tech research reveals how to assess PV module suppliers’ capacity claims

Part 3. PV-Tech research establishes technology-leadership scorecard for top-100 module suppliers

Part 4. PV-Tech research reveals ranking tool for manufacturing strength of global module suppliers

Part 5. PV-Tech research ranks PV module suppliers by financial health

Part 6. First PV module supplier bankability ratings tool created by PV Tech research team

Read the entire story

Copyright 2019 Solar Media Ltd. All rights reserved