Monday, 9 October 2017

The Accounting for Hyflux's Water Treatment Plants

Hyflux develops and operates various types of water treatment plants for municipals and industrial companies in concessions of 20 to 30 years. It has a strong order book of $3,187M as at Dec 2016, of which $1,916M is related to future revenue from the operations & maintenance (O&M) of the water treatment plants over the concession periods. Fig. 1 below shows Hyflux's order book.

Fig. 1: Hyflux's order book

Yet, despite the strong O&M order book, Hyflux lost money in recent quarters from the operations of the Tuaspring Integrated Water and Power Project. In 1H2017, Hyflux lost $25.4M, of which $47.9M was due to Tuaspring. The reason given for Tuaspring's loss is the continuing weak power market for Singapore and losses are expected for the next 2 quarters. Accounting-wise, why did Tuaspring lose money, despite it having a 25-year concession from the Public Utilities Board (PUB) to sell water and excess power from the plant?

Firstly, some background on Tuaspring. Tuaspring is a Design-Build-Own-Operate-Transfer (generically called BOT) project with a 25-year concession from PUB. The plant generates desalinated water and power, some of which is used for the operation of the plant. The excess power can be sold to the National Electricity Market. During the concession period, Hyflux will receive guaranteed minimum payments from PUB. It will also receive additional revenue from the sale of water and excess power. At the end of the concession period, the plant will be transferred to PUB's ownership.

Traditionally, when companies build plants, there is no revenue and there is investment in the Property, Plant and Equipment (PPE) account in the balance sheet during the construction phase. When the plant is operational, the company generates revenue from the sale of goods/ services produced by the plant and depreciates the PPE until the plant reaches the end of its economic life. 

Due to the nature of the BOT arrangement and concession payments from its water treatment plants, Hyflux adopts Financial Reporting Standard (FRS) 112. Under FRS 112, there is no investment in the PPE account during the construction phase. Instead, all the projected payments during the concession period are discounted to the present value and considered as financial receivables (for guaranteed payments) and intangible assets (for non-guaranteed payments). During the construction phase, these 2 accounts are progressively increased in the balance sheet instead of the PPE account.

Correspondingly, in Hyflux's cashflow statements, negative cashflows show up in "changes in financial receivables/ intangible assets arising from service concession arrangements (SCA)", which we do not usually find in other companies' cashflow statements. Traditionally, it is the PPE line under the Cashflow from Investing that is negative during plant construction. In the case of Hyflux, it is these 2 numbers in Cashflow from Operations that are negative.

Fig. 2: Hyflux's Cashflow Statement

There is another major difference between Hyflux and other companies. Traditionally, when a company builds the plant, it is recorded on the balance sheet at cost, i.e. the cost that the company pays for the plant. However, Hyflux does not pay other companies to construct its plants; it builds them itself. Hence, it is able to recognise a construction revenue for the plants during the construction phase. Since concession grantors like PUB do not pay for the construction of the plant, the construction revenue is the financial receivables and the intangible assets arising from SCA described above. Note from the earlier discussion that the financial receivables and SCA intangible assets are computed as the present value of the projected payments over the concession period. Thus, by the time the plant is completed, all the projected payments over the 20- to 30-year concession period would have been largely accounted for.

When the plant is completed and operational, Hyflux generates revenue from the guaranteed payments from PUB and the sale of water and excess power. On the other hand, it has to amortise the financial receivables and SCA intangible assets over the concession period. Thus, if the actual payments match the projected payments, what Hyflux earns during the operational phase is the interest from the financial receivables and SCA intangible assets, based on the discount rate used to compute the present value. If the actual payments exceed the projected payments, Hyflux earns more profit. Conversely, if the actual payments fall below the projected payments, Hyflux earns less profit or loses money. In the case of Tuaspring, due to the weak power market, actual payments came in below projected payments and Hyflux lost money on it.

Thus, contrary to conventional wisdom, the 20- to 30-year operational phase of a water treatment plant may not be the most profitable phase for Hyflux. A lot will depend on whether the actual payments exceed the projected payments. In contrast, the construction phase of the water treatment plant can be quite profitable.

In FY2016 financial results, Hyflux reported a net profit of $4.8M. It also mentioned that profits from Engineering, Procurement and Construction (EPC) were substantially wiped out by losses from Tuaspring, which amounted to $113.2M. This shows that the EPC profits can be quite substantial.

Fig. 3: Earnings Review for FY2016

Thus, based on current understanding, Hyflux might not generate the most profits from operating the water treatment plants during the 20- to 30-year concession periods. It probably generates more profits from constructing them.

Just a disclaimer, this post is not a recommendation for anyone to buy or sell Hyflux's shares or perps. Please read the accounting policies in Hyflux's annual report and do your own due diligence.

See related blog posts:


  1. Hi,

    Your research is very thorough and I came to a similar conclusion independently prior to this post.

    Have you heard of Benjamin graham book value approach towards valuation of stocks? I will love to see if you have any insight on it and how it applies to Hyflux.

    1. Thks for your compliments.

      I'm not good at using Benjamin Graham's valuation approach. Also, after analysing Hyflux, I prefer to stay away from it.