Energy > AG Speaking Energy > Got ITCs? How to Start Construction on Your Solar Project (Pre-IRS Guidance)
22 Mar '18

Solar developers are getting antsy about what is needed to show that they started construction on their 2020 (or later) projects by the end of 2019.

The answer is not clear in the absence of solar-specific guidance from the IRS. That said, there is a large amount of legislative history and IRS guidance in similar contexts (such as the old solar tax credit from 1978 and the wind production tax credit).

This article attempts to come up with some practical guidelines for planning in advance of the release of solar-specific IRS guidance (which we understand is expected to be issued by this summer).


Before 2015, the value of solar tax credits was adjusted based on the date that the project was placed in service. If a project was placed in service before 2017, it qualified for a 30 percent tax credit. If it was placed in service in 2017 or later, the credit was only 10 percent. Congress adopted a new framework in late 2015 that shifted the focus of the analysis to the date on which the construction of the project began.

Under the current framework, solar projects that are under construction by the end of 2019 qualify for the 30 percent investment tax credit. The credit dips to 26 percent for projects starting construction in 2020 and 22 percent for projects starting construction in 2021. Projects meeting these deadlines must still be placed in service by the end of 2023 to qualify for a credit above 10 percent.

The credit drops to a permanent 10 percent level for projects that begin construction in 2022 or later. Projects that begin construction before 2022, but are not placed in service until 2024 or later, are also limited to the 10 percent credit.

The IRS said in a 2016 notice that it anticipates issuing specific guidance on what it means to start construction of a solar project. We understand that the guidance project is under way, but it is not expected to be released until summer 2018 at the very earliest. One Treasury official said recently that the team is targeting June, but that tax reform implementation could derail timing.

The lack of guidance leaves solar developers who want to start planning ahead to 2019 in the lurch.

Fortunately, the start-of-construction concept is not completely foreign to the renewables space. The IRS and Congress have been addressing tax credit grandfathering issues since 1966 in a series of (largely consistent) legislative history and notices. The Treasury took the same approach under a cash grant program that was instituted to supplement the investment credit in 2009.

The current production tax credit guidance was modeled on that Treasury program, as well as on depreciation bonus rules that echo concepts from the tax credit grandfathering rules.

The IRS has issued six sets of guidance on this issue in the production tax credit context since 2013. It is a good bet that the IRS’s guidance for solar projects will look very similar to the production tax credit guidance that it has been refining over the past six years (and the IRS has said so informally). It is worth considering how the guidance, if left largely intact, would apply to a solar project.

The IRS has had a heck of a time implementing the wind rules. This is mostly because the very accomplished tax professionals at the IRS and the Treasury have not had years of project development and financing experience to be able to foresee some of the real-world issues that arise when tax policy and weird financing trends meet.

So, where does this leave us?

Two Pathways

The production tax credit and depreciation bonus rules use a dual-pathway approach to determine whether construction started by a given date.

One pathway requires “physical work of a significant nature” to begin. That physical work must then continue until completion. The other pathway requires the taxpayer to “incur” at least 5 percent of the total cost of the facility. As with the first method, the taxpayer must make continuous efforts to advance toward completion of the facility. Of the two, the continuous-efforts concept is theoretically easier to meet because it does not require continuous physical work.

Under the production tax credit guidance, any facility that is placed in service by the later of the end of the fourth calendar year after the year in which construction begins or December 31, 2018, will be automatically considered to have satisfied the continuous-construction or continuous-efforts requirement. This is true even if there is a long gap between when physical work starts in 2018 (or earlier) and when the work resumes in a later year. The IRS refers to this test as a “continuity safe harbor.”

The IRS and the Treasury are still considering whether solar should receive a four-year continuity safe harbor or whether it should be shorter. We have heard rumblings that two years could be on the table, but the concept is fluid at the moment.

The IRS said in 2016 that a taxpayer cannot buy more time under this continuity safe harbor by relying on the physical work test and the 5 percent test in alternating calendar years. For example, a taxpayer cannot rely on the physical work test in one year and then claim that it incurred at least 5 percent of the total cost of the facility in the next year. We assume that the IRS would take the same position in the solar context.

Five Percent Test

As stated above, the 5 percent path, or test, requires that a project owner incur at least 5 percent of the project’s cost prior to the construction start deadline. The “project’s cost” for this purpose means the cost of that portion of the facility that qualifies for five-year accelerated depreciation (basically the equipment necessary to generate electricity from solar power).

It is possible that the IRS and the Treasury could revisit the 5 percent threshold as it considers how the rules should apply to solar projects, although we have not heard anything to suggest that this is the case. Five percent is the threshold in the production tax credit context regardless of the size of the project, so, even though wind projects, for example, are larger than most solar projects, it is difficult to think of a compelling rationale for why the threshold should be different for solar.

A cost is incurred only when it is incurred under the taxpayer’s method of accounting. If it is a cash-basis taxpayer, then the cost is incurred when the cash goes out the door. Accrual basis taxpayers “accrue” costs only when all events that give rise to the liability have occurred, the amount owed can be determined with reasonable accuracy, and “economic performance” has occurred. Economic performance occurs when the item or service is delivered, title passes (along with risk of loss), or the item or service is accepted, with one exception.

The IRS rules do not define title, delivery or acceptance, so it is best to try to boil the concept down to either physical or constructive possession. Take physical possession. Take risk of loss (bear insurance costs). Take legal title and accept the item (through a bill of sale if possible).

We often see equipment “delivered” ex works, which means at the factory just after it is made. This is technically fine, but it puts more pressure on whether the buyer has constructive possession. The buyer should make sure the seller segregates the purchased equipment from other equipment at the factory, and the buyer should pay for transport (if possible) from the factory to the site or storage facility.

If a taxpayer cannot prove that it incurred costs under these rules, the taxpayer can look to a contractor’s costs. However, only costs incurred under a binding written contract with the taxpayer count. The binding written contract rules are discussed in detail below.

The one exception to the rules described above (for accrual-basis taxpayers) is called the “three and a half month” rule. This treats a payment as incurred on the date paid if the taxpayer takes physical or constructive possession within three and a half months of the first payment for the equipment. If more than one item is ordered, then all of the items must be provided within the three and a half month period.

It is important to note that the payment may not be funded by the seller of the equipment. This kind of arrangement calls into question whether the payment was real and whether the buyer truly incurred the costs. Related buyers and sellers should generally stay away from the three and a half month test to avoid getting caught up in this rule. Rather, the buyer (from a related seller) should take physical or constructive control of the equipment by the year-end construction start deadline.

During the cash grant era (generally 2009 to 2012), it was not uncommon to finance panel supply contracts that were used to grandfather projects with debt, and solar developers may find this to again be an attractive option as they think ahead to 2019 and 2020. One of the big issues is collateral. Should the lender get a lien on only the “magic” components, or the project as a whole? Are there even projects at this point that could be put up as collateral?  Is there a corporate guarantee as an alternative?

Also keep in mind that any payment made outside of the three and a half month window will not count. We often hear people say that they think that the rule is that you had to pay by the deadline and take delivery by mid-April of the next year. This is true only if you paid the entire purchase price at the end of December. For example, if you make monthly payments for equipment at the end of each month throughout 2019, you can include the December payment only if you do not receive the equipment until April 10, 2020. Any payments that were made between January and November 2019 do not count because they were outside of the three and a half month window.

Physical Work Test

If a solar developer cannot show that it incurred at least 5 percent of the project’s costs by the applicable deadline, it can show that it started construction by starting significant physical work.

The legislative history and guidance under the tax credit grandfathering rules have consistently said that work merely has to “start.”  The IRS and the Treasury have applied this concept in the production tax credit guidance as well. There is no reason to believe that the bar will be higher for solar. Congress intended for projects to be built. The rules merely require one to start the work.

Both on-site and off-site work can count.

The work can be performed by the taxpayer, by a contractor or by a subcontractor of the contractor. However, any work done by a contractor counts only if it is done under a binding contract that is in place before the work starts.

Legislative history suggests that this means that the contract must be binding by the applicable deadline, and that the focus of the inquiry is on only the contract between the taxpayer and first-level contractor (i.e., excluding any subcontracts). However, some tax counsel take a narrower view.

Some investors (and/or their counsel) require a binding contract for each subcontract, as well if a sub of a sub of a sub (for example) is the person doing the actual work. The legislative history (going all the way back to 1966) is clear that this is not required. The history says that the contract needs to be binding only with respect a distributor or middleman. Nevertheless, certain tax lawyers have required it recently because they are being asked to give exceptionally high-level tax opinions. Though the requirement is of questionable relevance, the thought is that it is safer to have binding contracts all the way down the chain than only at the taxpayer-contractor level. This kind of inquiry can run into practical roadblocks where contractors (or their subs) are reluctant to share confidential details of their manufacturing arrangements with tax equity.

Any components manufactured off-site cannot come from the manufacturer's inventory and cannot be equipment that the manufacturer normally holds in its inventory. Basically, real work has to occur that would not have occurred without the taxpayer placing the order. Here again, some conservative tax lawyers become concerned if the equipment does not appear “customized” to a particular project, as opposed to merely being custom-ordered for the taxpayer.

They want the equipment to be “bespoke” in the traditional sense. That is, it is not only made for the client, but it is made for the client’s project’s specific measurements.

If the noninventory rule were to apply to solar, it would be hard to find much equipment for which a solar project could claim that it started construction based on off-site manufacturing. Racking and carports designed specifically to fit a site are not normally held in inventory. The same should go for transformers as a general matter, but only very large projects have their own transformers. Work on a transformer that will be owned by the utility does not count.

Manufacturing work on solar panels or cells would not count, unless there is a major shift in policy to accommodate solar projects. They are clearly of a type normally held in inventory.

The physical work must be on equipment that is an integral part of the generating facility as opposed to transmission equipment or buildings. The IRS said in an internal memo in 2011 that all of the equipment at a substation used through the point where the electricity is stepped up to transmission voltage, plus equipment beyond the step-up transformer if the equipment is related to the functioning of the transformer or transfer equipment, is an integral part of the power generating activity and therefore qualified property.

The IRS guidance on production tax credit construction-start issues gives four examples of physical work that it thinks is “significant” enough to pass the test.

In the first example, a developer had a contractor excavate and install concrete pads for 20 percent of the turbines for his wind farm. The IRS subsequently clarified that the percentage in the example was not intended to suggest a minimum threshold.

Second, the production tax credit guidance says that “physical work on a custom-designed transformer that steps up the voltage of electricity produced at the facility to the voltage needed for transmission is physical work of a significant nature.” 

Starting construction on “string roads” (i.e., roads for equipment to operate and maintain the qualified facility) also constitutes physical work of a significant nature. In contrast, work for roads that are primarily used to access the site or that are primarily used for employee or visitor vehicles do not qualify.

Lastly, “onsite physical work of a significant nature begins (at a wind farm) with the beginning of the excavation for the foundation, the setting of anchor bolts into the ground, or the pouring of the concrete pads for the foundation.”

Many solar projects do not have foundations in the traditional sense. They might be ballasted on the roof of a commercial building, or they may merely be attached to racking on a residential building or vibrating a tracker into the desert.

It remains to be seen what level of on-site physical work investors will permit to count for these purposes. In any case, the options for solar appear much more limited, except in the case of larger utility-scale projects.

In the Treasury grant days, the investor community preferred the 5 percent safe harbor because the physical work test was largely unworkable. No one knew (or could get comfortable with a good deal of conservatism) what on-site work of a significant nature meant, except for the largest utility-scale projects. In those cases, we often saw investors (and the Treasury) count a combination of roadwork that permitted maintenance (and was not merely an access road), water well drilling, and/or transformer or substation construction.

The problem with the comparison to wind for on-site physical work is that wind turbine foundations are typically 70-80 feet in diameter and six to eight feet deep. You should consider what work you can do on site or off that is qualitatively significant. While work does not have to be quantitatively significant as a technical matter, investors will require it to be so, absent clear IRS guidance to the contrary. The more work, the better.

Physical work counts only if it is done under a binding contract that is in place before the work starts.

A contract is binding only if it is enforceable under state law against the taxpayer or a predecessor and does not limit damages to a specified amount. For this purpose, a contractual provision that limits damages to at least 5 percent of the total contract price will not be treated as limiting damages to a specified amount.

In 2013, a group of wind generators, lenders and tax equity investors asked the Treasury to draw clear lines about how much work had to be done on turbine excavations, roads, transformers or other major components for a project to be considered under construction in 2013 (the deadline at that time). The IRS national office staff stated repeatedly on an informal basis that there was no definite minimum threshold for work. The analysis is based on a qualitative analysis of the work. That is, was the work on some material aspect of the project?  The IRS then issued a new notice that confirmed the position, saying, “[a]ssuming the work performed is of a significant nature, there is no fixed minimum amount of work or monetary or percentage threshold required to satisfy the Physical Work Test.” 

This does not mean that investors will not impose their own minimum-spend threshold based on internal preferences.

For example, we are only recently seeing the larger tax equity investors accept the position that off-site transformer fabrication is physical work of a significant nature, even though it has been specifically permitted under IRS guidance for several years. The unease is largely due to the fact that it may cost only $200,000 to begin the fabrication of a transformer that will be part of a $300,000,000 project. Investors are coming around on the issue, but each has its own quirks as to how it thinks about the concept and what kinds of support it needs to see.

The IRS national office staff has reiterated (again informally) this view, saying that it is not troubled by a low level of work as long as the work is of a significant nature. The government is protected by the requirement to show continuous work if the project is not placed in service by a certain date.

If work finishes after the later of December 31, 2018 or the end of the four-year “continuous work” presumption (applicable to wind and not yet clear whether it is applicable to solar), the IRS will closely scrutinize the work to ensure that it was continuous. The IRS has listed a number of tasks (generally outside of the developer’s control, like severe weather and financing issues) that are permissible disruptions in continuous work. It is unclear whether a disruption could include, for these purposes, the fact that a transmission upgrade is not expected to be made within the continuity safe harbor.

At the end of the day, the IRS developed the continuity safe harbor discussed above because the continuity requirements are unworkable in practice. No one knows what they mean, really.

Single Project vs. Blocks

Under the wind tax credit guidance, multiple turbines that are operated as a single, integrated project are treated as a single facility for purposes of testing when construction started. We expect a similar rule for solar.

The question of whether multiple turbines should be treated as a single facility depends on the facts. Facts that point to a single facility are that one company owns the entire project, all of the electricity is sold under a single power contract, all of the electricity moves to the grid through a single substation and intertie, the entire project is financed under a single loan agreement, all the turbines are on contiguous sites, and all of the equipment is procured under a single turbine supply agreement.

A similar multiple “block” concept would presumably apply to solar projects as well. In the Treasury grant context, solar projects could be broken down into blocks that include one or more strings of panels and an inverter. The rationale is that, like a wind turbine, a string of panels and an inverter can operate independent of other strings of panels, in theory. Whether multiple strings of panels and inverters would be treated as a single block would likely depend on similar facts as those described above.

Larger projects are often completed in phases that begin in different years. In the wind context, the best practice is typically to assume that each phase needs to meet the start of construction rules independently. It is somewhat unusual for these kinds of projects to have enough commonality between phases for financiers to be comfortable that each phase is part of a single project. The same rationale should apply to large-scale solar projects.

Solar projects located on separate buildings should be analyzed independently from each other. We often see this in the context of solar projects on college or school campuses.

Key Observations for Solar

  • The 5 percent test (technically, but not economically) is the easiest path for solar projects.

It is almost certain to be included in the solar guidance, it is the most objective test, and it will be the easiest path to “sell” to investors and their tax counsel.

The 5 percent test also has the added benefit that, if you start construction with the safe harbor, you merely have to move development forward continuously by incurring costs, rather than continue with physical work on a continuous basis. The physical work test requires continuous physical work going forward.

  • The 5 percent test has practical difficulties that must be balanced.

The 5 percent test is likely the more expensive path. The alternatives that involve merely “starting” work on qualitatively significant equipment can have a much lower barrier to entry. The developer of a $200 million project would need to incur $10 million under the 5 percent test. Alternatively, a developer can start construction on a transformer (with a roughly $1 million price tag), paying less than 10 percent of the purchase price prior to the construction start deadline.

The 5 percent test also brings with it the risk that equipment purchased for grandfathering purposes becomes outdated before a project can be deployed. We saw solar panels and inverters purchased for these purposes in connection with the Treasury grant program where project buyers were not interested in the grandfathered equipment. In some cases, it was due to the preference for “new” models of equipment. In others, it was because construction or permitting standards have evolved past the equipment purchased in prior years.

  • The most practical spend to include in a 5 percent safe harbor approach is to buy solar panels or inverters.

As a technical matter, any costs incurred with respect to the project’s solar generation equipment counts. But, if you look to costs other than the central generating equipment itself (accounting, legal, permitting, design, etc.), you run the risk that an investor will require part or all of the line items to be allocated to noneligible costs. Much of the allocation procedure is done by feel, meaning that there is more than one “right” way to make the allocation. The possibility of different approaches reduces the benefit of using the 5 percent test, which should be objective certainty.

  • For equipment purchased for the 5 percent test, make sure that you are very comfortable with either (1) taking physical or constructive delivery by the deadline, or (2) paying for the equipment with your own money or a loan you take out from a third party (e.g., not the seller of the equipment).

If you pay for only the equipment by the deadline, make sure that the equipment is delivered within three and a half months of the first payment. Make sure that you check with your accountants on whether your tax methods of accounting permit you to use the three and a half month rule. A new taxpayer (partnership or corporation) can use the three and a half month rule, but this method’s use needs to be consistent going forward.

Make sure that there are no obligations to give the equipment back if the purchased equipment cannot be deployed.

Make sure that any “delivery” off site gives the buyer the right to visit the equipment and segregates it from equipment that is intended for other buyers or projects. The buyer also should have legal title, pay for insurance, and pay for further shipping or installation. Do not rely merely on legal title transfer alone, even if people try to tell you that’s what the words in the guidance say.

  • As for the physical work test, solar projects are harmed by rules that require work to occur on only items not held in inventory.

There just are not a large number of custom components for a solar project. If you have custom racking, carports or a transformer (for larger projects), you may be able to start work on those items. Everything else is generally held in inventory.

  • Confirm that the contract was binding and written when the work started (and each subcontract was binding and written as well).

Make sure that the contracts under which the work is performed are binding for state (or country) law purposes, that they state the specs of the item being purchased and that they do not permit either party to get out of the deal willy-nilly.

If you have a termination option, the best course is to have the buyer pay damages of at least 5 percent of the purchase price plus (ideally) the value of equipment built to date. Make sure that any termination language you add to a purchase agreement form does not conflict with the form’s existing termination or refund language.

If you adjust the terms of the contract after it is signed, run the changes by your tax lawyer. A change of the price by 10 percent or less should not be a problem, but some tax lawyers worry about certain changes, notwithstanding the fact that the price may not have changed significantly.

Do not include a right to suspend work at the buyer’s convenience. If you do include such a right, make sure that there is a date by which work must restart, or termination will be presumed, triggering damages.

  • Unlike wind projects, which may have hundreds of large excavations and miles of new roads on a site, distributed generation projects (and even some large utility-scale projects) might not have any excavations and may be placed on a single roof.

This means that the only option for some solar projects to show that construction commenced may be to use the 5 percent safe harbor or to start construction of equipment off site. Because off-site construction can be counted only if it is on specially ordered equipment, this likely would eliminate any physical work prospects for most of the solar projects being built. That leaves these developers to rely on the 5 percent safe harbor, which, while viable, is vastly more expensive.