Monday, 12 August 2019

Effects of New Accounting Rule on Leases

Did you notice that in recent quarters, companies have been reporting better EBITDA (Earnings before Interest, Tax, Depreciation & Amortisation) and Free Cashflow figures? Do not be happy too soon, as the improvements could merely be due to a change in accounting rule for leases.

Before this year, companies that lease properties, equipment, etc. could choose to treat the leases as operating leases if they meet certain conditions and expense the rents as they fall due. There are no assets and liabilities on the balance sheet associated with these operating leases. This poses a problem when comparing against companies that own the properties and/or equipment. In reality, is a company that leases property very different from another that owns a leasehold property? In terms of the rights to use the property, the differences are small, but financially, the differences can be quite significant. On the balance sheet, such property-owning companies would have more assets and probably more loans to fund these assets. On the income statement, these companies would not have to pay any rent but would have higher depreciation and probably higher interest expenses.

Starting from this year, all companies have to adopt the Singapore Financial Reporting Standards (International) SFRS(I) 16 on Leases, which standardise the way companies report leases on their financial statements. Companies can no longer choose to treat their leases as operating leases and expense the rents. Companies have to treat their leases as financial leases and include the asset as a Right-of-Use (ROU) asset and a corresponding lease liability on the balance sheet. The lease liability is the present value of all future lease payments. At the start of the lease period, both the ROU asset and lease liability must match and balance out each other.

Companies only have the rights to use the ROU asset for the duration of the lease. Hence, at the end of the lease period, both the ROU asset and lease liability have to be reduced to zero. ROU asset is reduced to zero through depreciation. Lease liability is reduced to zero through amortisation. This is similar to loan repayment, in which the remaining loan amount each year is increased slightly by the interest expense, but reduced by a larger amount by the loan repayment. Both depreciation and amortisation affect the income and cashflow statements. It would be clearer to illustrate the changes using a company's actual financial statements.

The example used here is Hour Glass, which is a watch retailer that leases properties to operate its shops. Its business is generally stable, making year-to-year comparison valid. Furthermore, it separates out ROU assets and depreciation while most companies combine them with Property, Plant and Equipment (PPE), hence, allowing a clearer view of the effects of SFRS(I) 16. Companies should be encouraged to adopt the same practice.

Balance Sheet

Fig. 1 below shows the balance sheet for Hour Glass in 1Q2019.

Fig. 1: Balance Sheet

Hour Glass adopted SFRS(I) 16 starting from this Financial Year (FY), which began in Apr 2019. It added ROU assets of $113.8M and corresponding lease liabilities of $116.1M. The ROU assets are nearly twice as much as Property, Plant and Equipment (PPE) which amounts to $58.7M. This shows that the value of rental properties could be as much as or even more significant than the leasehold properties that the companies own. Note that not all companies report ROU assets separately; some companies report them together with PPE.

Income Statement

Fig. 2 below shows the income statement for 1Q2019.

Fig. 2: Balance Sheet

On the income statement, rental expenses are reduced significantly from $7.6M in 1Q2018 to $1.2M in 1Q2019. On the other hand, there is additional depreciation of $6.9M on the ROU assets. Finance cost is also higher by $0.5M to account for the interest expense incurred on the lease liabilities. Assuming all other factors remain constant, the net effect of these factors due to adoption of SFRS(I) 16 is a reduction of $1.1M in operating profit for Hour Glass in 1Q2019.

However, if you compute EBITDA, EBITDA computation excludes depreciation and interest, among others. Changes in depreciation and interest do not affect EBITDA. Hence, only the change in rental expenses affects EBITDA. Holding all other factors constant, EBITDA would have increased by $6.4M in 1Q2019. In 1Q2018, Hour Glass' EBITDA was $19.2M. The change in accounting rule for leases has increased Hour Glass' EBITDA by 33%!

Cashflow Statement

Fig. 3 below shows the cashflow statement for 1Q2019.

Fig. 3: Cashflow Statement

For Cashflow from Operations (CFO), depreciation from ROU assets is added back to the operating profit. For Cashflow from Financing (CFF), payment of lease liabilities, i.e. rental expenses, is included. 

Previously, rental expenses would have been deducted when computing the operating profit which is used as the starting point to compute the CFO. The rental expenses have been mostly replaced by depreciation of ROU assets, but in computing the CFO, the depreciation of ROU assets is added back. Hence, the net effect is an increase in CFO, by an amount equivalent to the payment of lease liabilities, which is now moved to CFF. Holding all other factors constant, the net effect is an increase in CFO by $6.1M, or 53% from 1Q2018!

Free Cashflow (FCF) is generally computed as CFO minus capex. Since CFO increased by $6.1M, FCF would have increased by an equivalent amount. Assuming no change in capex, FCF for Hour Glass in 1Q2019 would have increased by 68% simply from a change in accounting rule!

Conclusion

The above discussion summarises the key changes arising from the change in accounting rule for leases. There are other changes required to reconcile the balance sheet items for each company. For more details, readers are advised to read the companies' financial statments.

Also note that all the improvements in EBITDA and FCF compared to the previous FY mentioned above are also because companies do not have to restate the financial statements for the previous FY. Effectively, the financial statements for this FY are not comparable to that of the previous FY. Do not be too happy when you see an increase in EBITDA and FCF figures this year!

P.S. I am vested in Hour Glass.


See related blog posts:

Sunday, 28 July 2019

Does SIA's 3.03% Bond Have Sufficient Margin of Safety?

Following up from last week's blog post on Will Temasek Bail Out SIA Bondholders In Event of Default?, here is the analysis on SIA's 5-year 3.03% bond based on Benjamin Graham's criteria as described in The Lost Art of Bond Investment. Surprisingly, the bond is not as strong as I initially thought based on a simple Debt-to-Equity ratio check. Below are the computation of the earnings coverage and stock value ratio based on SIA's latest financial statements for Financial Year 18/19 ending in Mar 2019.

Earnings Coverage
 
Profit before tax = $868.6M
Adjusted for:
- Deduct: Share of profits from joint ventures = $23.2M
- Add: Share of losses from associates = $97.4M
- Add: Rental on leased aircraft = $679.7M
- Add: Finance cost = $116.1M
Total earnings available for covering fixed charges = $1,738.6M


Current finance cost = $116.1M
Adjusted for:
- Add: Rental on leased aircraft = $679.7M
- Add: Interest for $600M @ 3.16% MTN 007 series = 3.16% x $600M

= $19.0M
- Add: Interest for $750M @ 3.03% MTN 001 series = 3.03% x $750M

= $22.7M
Total finance cost = $837.5M


Earning Coverage = $1,738.6M / $837.5M

= 2.08

Several adjustments were made to compute the total earnings available for covering fixed charges and total finance cost. The more unusual adjustment in SIA's case involves adding the rental of leased aircraft to both figures. This is because SIA leases aircraft in addition to buying them. The leases range from 6 to 12 years and cannot be cancelled, although there are options for early termination for up to 2 years before the original lease expiry. The leased aircraft do not appear on the balance sheet.

On the other hand, for aircraft that SIA owns by borrowing money from the banks, the aircraft appear as a asset and the loan appears as a liability on the balance sheet. On the income statement, there is no rental required, but SIA incurs depreciation and interest on the loan.

Hence, for the leased aircraft, even though SIA does not incur a finance cost, the rental is effectively a fixed charge that SIA has to cover, as the leases cannot be cancelled and SIA needs the aircraft to continue its operations. A similar adjustment needs to be made for the balance sheet, as discussed in the section later.

The current finance cost is also adjusted for interest on a $600M 3.16% Medium Term Note (MTN) issued on 26 Oct 2018 and the $750M 3.03% bond issued on 29 Mar 2019 which is the subject of this blog post. Although both bonds already appear on the balance sheet as at 31 Mar 2019, their first interest payments will only be made in FY19/20, Hence, their interest payments need to be added to compute the actual finance cost.

Based on the above adjusted figures, the earnings coverage is computed to be 2.08 times, which is below the minimum average earnings coverage of 3 times for industrial companies.

Stock Value Ratio

No. of shares = 1,199.9M
Share price = $9.62
Market value of shares = $11,542.6M


Current borrowings = $6,654.4M
Adjusted for:
- Add: Liability for rental aircraft = $2,200.0M
Total borrowings = $8,854.4M


Stock value ratio = $11,542.6M / $8,854.4M

= 1.30

As mentioned above, adjustments need to be made to the balance sheet for the rental aircraft. The rental aircraft is effectively an asset for SIA. Correspondingly, there should be a liability to account for the loan that SIA would have borrowed to purchase the aircraft outright. In fact, International Financial Reporting Standard (IFRS) 16 on Leases came into effective starting from Jan 2019 that requires companies like SIA to account for leased assets on their balance sheets. SIA has disclosed in its Annual Report that the assets will be increased by $1.7B while liabilities will be increased by $2.2B.

Considering the $2.2B increase in total borrowings, the stock value ratio is computed to be 1.30, which is higher than the minimum stock value ratio of 1.0 for industrial companies.

Quantitative Assessment

Thus, based on the above figures, SIA's 5-year 3.03% bond does not meet the earnings coverage criterion but meets the stock value ratio criterion. Based on Benjamin Graham's criteria, the bond does not have sufficient margin of safety.

Other Considerations

As mention in my blog post on Will Temasek Bail Out SIA Bondholders In Event of Default? last week, I believe Temasek will come to the rescue of SIA bondholders in the event that SIA could not pay interest and/or redeem the bond.

P.S. I am vested in SIA's 3.03% bond.


See related blog posts:

Sunday, 21 July 2019

Will Temasek Bail Out SIA Bondholders In Event of Default?

As regular readers would know, I stopped blogging for a year. I did not just stopped blogging; I also stopped monitoring performance of my portfolio and analysing shares and bonds in detail. So when SIA launched its retail 3.03% bond in Mar this year, I did not analyse it in detail as I would typically do using Benjamin Graham's method (see The Lost Art of Bond Investment for details) and simply bought it. I took a glance at its financial statements, carried out a simple Debt-to-Equity check and concluded that its debt obligations were not too excessive. Most importantly, I relied upon the assumption that SIA's parent, Temasek, would bail out bondholders in full in event of a default. This was not the only time that I relied on similar assumptions when buying bonds. I did the same when I bought Fraser Property's 3.65% bond (see Does FCL's 3.65% Bond Have Sufficient Margin of Safety? for more info).

Now that I am back to blogging (and thinking about financial issues), it is worth diving deeper to examine whether the assumption is rock solid and could be relied upon. Do note that this is not entirely a hypothetical question, as the airlines industry is a highly competitive one. Bankruptcies are not uncommon. Past examples include American Airlines, Delta Airways, Northwest Airlines, etc. going into Chapter 11 protection.

So, will Temasek come to the rescue of bondholders in the event SIA defaults on the bond? To answer this question, we need to first understand the background of SIA and Temasek. So, the first question is: will Temasek rescue SIA? The answer must be a resounding yes. SIA is the national airline and the pride of the nation. When our political leaders go overseas for conferences, they fly with SIA. It is difficult to imagine our political leaders flying on some other countries' national airlines. Thus, yes, Temasek will come to the rescue of SIA.

However, rescuing SIA the company is not the same as rescuing SIA bondholders. While SIA the company is a strategic national asset, SIA bondholders (and shareholders) are not. Furthermore, the ultimate shareholders of Temasek are Singaporeans. Bailing out SIA and its bondholders with Temasek's money is akin to using taxpayers' money to do so. It will be politically difficult to use taxpayers' money to bail out bondholders in full. The most likely scenario is that the settlement with bondholders will be at arm's length basis on normal commercial terms, i.e. the outcome would be similar to another company that is not owned by Temasek. In other words, bondholders will suffer some capital losses. This is not unlike the case of Hyflux preference shares and perpetual capital securities. Thus, from this perspective, SIA bondholders should not expect Temasek to bail them out in full.

On the other hand, SIA is not the only company that Temasek owns. Temasek owns a lot of companies in its portfolio. Temasek has triple-A ratings from both Moody's and S&P. Not all the companies in its portfolio has similar ratings based on their own merits. When companies with lower ratings borrow money, lenders take into consideration the fact that the company is majority owned by Temasek and offer a lower interest rate compared to a company that does not have Temasek as its backing. In the event that Temasek does not step in to bail out bondholders in full, credit markets will take note and will not offer lower interest rates to Temasek-owned companies in future. Each company has to pay an interest rate that is commensurate with its own credit rating. In other words, Temasek will have to bear higher interest payments across most of its subsidiaries. This long-term economic cost might outweigh the short-term political cost and prompt Temasek to bail out bondholders in its subsidiaries in full.

Finally, we should also note that Temasek, although it does not play an active role in day-to-day management of the subsidiaries' operations, is not a sleeping partner either. Long before trouble happens and hits the headlines, Temasek would have done something to avert it. For example, Tiger Airways, SIA's budget airline, had been losing money for 4 out of 6 years since its listing on SGX in Jan 2010. It carried out 3 rights issues over the same period. It also issued a 2% perpetual capital convertible securities (PCCS) in Apr 2013. If the trend were to continue, Tiger Airways would probably have failed and defaulted on the PCCS. In Nov 2015, SIA announced that it would take over Tiger Airways. It also redeemed all outstanding PCCS upon successful takeover of the company.

In conclusion, I had been lazy in analysing the SIA 3.03% bond when I bought it. But I guess I still can rely on the assumption that Temasek would step in and bail bondholders out in the event of default by SIA. Nevertheless, I should do my homework and analyse whether the SIA bond has sufficient margin of safety according to Benjamin Graham's method.


See related blog posts:

Sunday, 7 July 2019

Are Private Equity Bonds Better Than Corporate Bonds?

When Astrea V 3.85% bond was launched, some investors remarked that it is better than some of the recently issued corporate bonds, such as SIA 3.03% bonds. What are the differences between Private Equity (PE) bonds and corporate bonds, and are PE bonds really better than corporate bonds?

It is difficult to compare Astrea bonds with, say, SIA bonds, since their nature of business are different. To make the comparison between PE bonds and corporate bonds more meaningful, let us consider a hypothetical bond issued by Azalea Asset Management, which is the sponsor of the Astrea III/IV/V bonds. Fig. 1 below shows the corporate structure of Azalea.

Fig. 1: Azalea Corporate Structure

The assets of Azalea are the 3 Astrea companies issuing the Astrea III/IV/V bonds and owning the underlying portfolios of PE funds. Azalea probably has some other income-generating assets, such as the investment management company shown in Fig. 1 above, plus some other unlisted PE funds. Thus, the nature of business of Azalea and Astrea III/IV/V companies are similar. What would be the differences between the hypothetical Azalea corporate bond and Astrea III/IV/V PE bonds? Note that the PE bonds are not limited to the Class A-1 bonds which are open to retail investors. There are also Class A-2, B and C bonds.

The first key difference would be the security of the bonds over the assets. Astrea PE bonds are secured against the PE funds in the respective Astrea companies, whereas Azalea corporate bond would be unsecured. In the hypothetical scenario where the Astrea PE bonds default, Astrea bondholders could force the respective Astrea companies to liquidate their PE funds and return money to the bondholders. However, in the event that the liquidation proceeds are insufficient to redeem the bonds, bondholders have no recourse to Azalea, or to the other Astrea companies. For example, if Astrea III bonds were to default, Astrea III bondholders have no rights to the assets of Azalea, Astrea IV and Astrea V companies. The assets of each company are ring-fenced and could only be used to service the bonds issued by the respective company.

Similarly, in the hypothetical scenario where the Azalea corporate bond defaults, Azalea bondholders have no claims over the PE funds held in the 3 Astrea companies. Nevertheless, they could force Azalea to sell off the Astrea companies together with their portfolio of PE funds and PE bonds. However, they could not force Azalea to break up the Astrea companies, sell off their PE funds, redeem the Astrea PE bonds, and return excess cash to Azalea to pay off the corporate bondholders (Note: it might be possible to do so for other project/ asset-level bonds, but the terms of Astrea PE bonds do not allow for early liquidation of assets and redemption of bonds). In other words, regardless of what happens to Azalea, Astrea PE bondholders will not be affected. 

So does it mean that Astrea PE bonds, which are secured against the PE funds of the respective Astrea companies, are better than Azalea corporate bonds which are unsecured? Not necessarily. The key factor is the quality of the assets that are securing the bonds. If the assets are of high quality, the PE bonds have good collaterals. Conversely, if the assets are of low quality, the collaterals would be useless. Remember, Astrea PE bondholders have no recourse to Azalea and the other Astrea companies. They can only count on the assets in their respective Astrea companies to pay interest and redeem the bonds. 

Although Azalea corporate bond is unsecured, if the Astrea companies are generating good cashflows for Azalea, it does not matter whether the bond is secured or not. In a hypothetical scenario where one of the Astrea companies have poor assets whereas the other Astrea companies have good assets, it might be better to hold the unsecured Azalea corporate bond than the secured but troubled Astrea PE bond. So, quality of assets is key in determining whether secured or unsecured bonds are better.

The second difference is that Azalea could have other income-generating assets and businesses besides the 3 Astrea companies. In Fig. 1 above, it has an investment management subsidiary to manage the investments in PE funds for the Astrea companies in return for a fee. It could also have other PE funds that are outside the Astrea companies. So, for Azalea corporate bonds, there could be other sources of operating cashflows, whereas for Astrea PE bonds, the only source of cashflows is the PE funds in the respective Astrea companies.

The third difference is that besides receiving cashflows from the Astrea companies to redeem the Azalea corporate bond, Azalea could refinance the bond through bank borrowings, new corporate bonds, shareholder loans from Temasek, or even private share placements and IPO! Being a corporate bond, there are many avenues to refinance it. Astrea PE bonds do not have such avenues. To reiterate, Astrea PE bondholders can only count on the assets in the respective Astrea companies. If the assets are good, PE bondholders will get the promised returns. If the assets are poor, they will suffer some losses.

Having said the above, being able to borrow money is a double-edged sword. While borrowings could help to refinance the Azalea corporate bond, Azalea could also run the risk of borrowing too much money and jeopardise its ability to pay interest to and/or redeem the Azalea corporate bond if banks decide that Azalea's credit risk is too high. For the Astrea PE bonds, such risks have been mitigated. The terms of Astrea PE bonds prohibit the Astrea companies to borrow money other than to issue the different classes of bonds at inception, as well as to meet capital calls and cover bond interest payment shortfalls. The last 2 conditions are actually safeguards for the PE bondholders (see Understanding the Safeguards of Astrea IV 4.35% Bonds for more info).

In conclusion, PE bonds are not necessarily better or worse than corporate bonds. The key words are: quality of assets securing the PE bonds. This is unlike preference shares and perpetual capital securities, which are inherently inferior to stocks and bonds (see Prefs and Perps are Generally Inferior to Stocks and Bonds as an Investment Form for more info).


See related blog posts:

Sunday, 30 June 2019

A Shrinking Balance Sheet for Bonds in Mature PE Funds

In another few more days, the Class A-1 bonds of Astrea III, the first wholesale Private Equity (PE) bond listed in Singapore, will be redeemed as scheduled. What could we learn from the 3-year existence of this bond, which could provide some useful insights on the behaviour of Astrea IV and V bonds?

Astrea III publishes annual reports, which document the cashflows received, performance of its underlying PE investments, outlook for PE investments as well as the usual income statements and balance sheets. In the 3 years of its existence, the cashflows of Astrea III from its investments in PE funds are shown in Fig. 1 below.

Fig. 1: 3-Year Cashflows of Astrea III

Over the 3 years, Astrea III has received total distributions of USD952M, capital calls of USD176M, resulting in net distributions of USD776M. At inception, the weighted average age of the PE funds which Astrea III invested into is 6 years. The cashflows are typical of investments in mature PE funds, as shown in Fig. 2 below.

Fig. 2: Typical Cashflows of PE Investments

Moving forward, Astrea III will likely see less distributions, as indicated in Fig. 2. Already, the investments made by the underlying PE funds into companies are showing signs of ageing, as shown in Fig. 3.

Fig. 3: Holding Period of Underlying PE Investments

The average holding period of the underlying PE funds' investments into companies has increased from 4.0 years in 2016, to 4.5 years in 2017 and 5.2 years in 2018. There are 2 opposing reasons why PE funds hold onto their investments for longer than usual -- it could be to extract more value from a good company, or it could be the company could not deliver as promised and the PE fund has difficulty selling it for a good price. If it is the first reason, it is a good thing for PE bond investors. But if it is the second reason, PE bond investors will have reasons to be worried.

What happened to the distributions received by Astrea III over the 3 years? Fig. 4 below shows the balance sheet for the Financial Years ending in Mar 2017, 2018 and 2019.

Fig. 4: Astrea III's Balance Sheets

The first thing to notice is the investments in PE funds (yellow bar) have been shrinking, from USD1,070M in 2017 to USD904M in 2018 and finally to USD739M in 2019. This is due to the net distributions of USD568M from the PE funds from 2017 till 2019, offset by fair value gains (i.e. capital gains) in the PE funds of USD236M over the same period.

A portion of the net distributions went to increase the cash account (grey bar), which increased from $203M in 2017 to around USD340M in both 2018 and 2019. Another portion went to pay interest of USD21M to bondholders in 2018 and 2019. The remaining distributions were used to pay the sponsor in the form of repayment of shareholder loans and dividends to shareholder. From 2017 till 2019, a total of USD385M was paid to the sponsor/ shareholder. As a result, the sponsor equity (orange bar) has been shrinking.

Because of the distributions and payments to sponsor, the asset base has been shrinking. On the other hand, there is relatively little cashflow used to pay down the bonds (blue bar), as they have not matured. If the trend continues, bond holders might end up holding onto a shrinking asset base of ageing PE funds with declining distributions while the sponsor gets all its capital back (see Where Do Astrea Bonds Stand Along PE Fund Lifecycle? for more info on PE lifecycle and its impact on cashflows). By the time the bonds mature, there might be little cashflows left to redeem the bonds. Remember, once the cash leaves Astrea III by way of repayment of shareholder loans and/or dividend to shareholder, bond holders have no recourse to the sponsor, Azalea, or Temasek.

Thus, one key risk for bond investors in mature PE funds is when the fund is in the midst of the harvesting period, the sponsor gets most of the money while bond investors get only the interest payment. When the harvesting dries up, bond investors do not get sufficient cashflows to redeem the bonds while the sponsor already gets all its capital back.

Fortunately for Class A (A-1 and A-2) bond holders of Astrea III, there are safeguards in place to ensure that the above scenario does not happen. Every 6 months, Astrea III has to set aside some cash in reserves accounts which can only be used to redeem the Class A bonds. The reserves accounts totalled USD161M, USD224M and USD258M in 2017, 2018 and 2019 respectively. These reserves accounts are sufficient to redeem the Class A-1 bonds which will mature in the next few days. The total amount of Class A-1 bonds is SGD228M (approximately USD170M).

Another safeguard that Astrea III put in place is the Loan-to-Value (LTV) ratio should not exceed certain thresholds ranging from 20% to 45%. If these thresholds were exceeded, Astrea III has to divert more cashflows to the reserves accounts. Based on Fig. 4, the LTV ratio of Astrea III is 27%, 18% and 24% in 2017, 2018 and 2019 respectively.

Had there been no such safeguards to set aside cash during the harvesting period, the balance sheet would have been worse for bondholders. Fig. 5 below shows the balance sheet had the reserves accounts been paid out to sponsor.

Fig. 5: Balance Sheet of Astrea III Excluding Reserves Accounts

The LTV ratios would have been 42%, 43% and 58% in 2017, 2018 and 2019 respectively. In particular, bond holders would end up being a larger supplier of capital than the sponsor while still not getting the bulk of the distributions from the PE investments!

In conclusion, if you are a bond investor in mature PE funds, be prepared to see a shrinking asset base while most of the distributions go to the sponsor. Make sure you have safeguards in place to ensure that a portion of the distributions are set aside for the sole purpose of redeeming the bonds!

Sunday, 23 June 2019

Where Do Astrea Bonds Stand Along PE Fund Lifecycle?

Now that the allocation for Astrea V 3.85% bonds is released, everyone can go back to their regular activities and forget about it. However, if you, like myself, hope to get a piece of the equity portion of Private Equity (PE) investments one day, then we should continue to learn and understand more about PE. Today's discussion is on the typical lifecycle of a PE fund and where do the 3 Astrea bonds stand along the lifecycle. This has major implications on the risks of the bonds, as we shall discuss later.

A PE fund usually has a lifecycle of around 10 years and comprises 4 phases. These 4 phases usually overlap. See Fig. 1 below (source: The Ultimate Guide to Private Equity).

Fig. 1: Lifecycle of PE Fund

In the first phase, the General Partner (GP) who runs the PE fund, will source for new investors and get their commitment to provide capital as and when needed to invest into promising companies. When capital commitments reach the target fund size, the fund will close. 

In the second phase, the GP will source for promising companies to invest into. This is usually called the Investment Period, as the fund is busy investing into companies. During this phase, the fund can invest in any new companies so long as they satisfy the criteria in the investment mandate.

After buying into the companies, the fund will work with the companies' management to enhance their operations such that they can achieve a higher valuation than what the fund paid for. The typical duration that a fund holds onto a company is about 5 years, but could exceed 10 years for an under-performing company.

When the time is ripe, the fund will seek to exit the company, either by IPO, sale to another company, or even sale to another PE fund. This is known as the Harvesting phase. During this period, the PE fund is not allowed to invest in new companies, but is allowed to invest additional money into existing companies, known as follow-on investments.

Finally, before the stated lifespan of the fund is up, the GP has to exit all remaining companies, return all proceeds to investors and dissolve the fund. The 10-year lifespan is not a fixed timeline, as GPs can request for extensions of 1-2 years so that they do not need to carry out a fire sale of the remaining companies.

Based on the above lifecycle of a PE fund, investors of the PE fund (known as Limited Partners) will contribute capital in the early years of the fund before receiving distributions in the later years when the PE fund exits its investments. The cumulative cash inflow over the lifecycle of the PE fund resembles a J-Curve. See Fig. 2 below.

Fig. 2: PE J-Curve

For an investor in a PE fund, where the fund is along its lifespan matters. In the early years, there is significant capital outlay as the PE fund invests into new companies. There are also significant risks in streamlining the company for greater efficiency. In addition, the company might struggle under the usually significant debt load placed upon it by the PE fund. Not all of the companies will prosper and make money for the PE fund. 

In the closing years of the PE fund, although there is no further capital outlay, the cash distribution from the PE fund is declining as it exits more and more companies. Also, whatever companies remaining in the PE fund might be under-performing and might not be worth much. If a company is attractive to other buyers, it probably would have been sold before the 10-year lifespan of the PE fund is up and not get left behind in the fund.

Thus, if the PE J-Curve shown in Fig. 2 is accurate, the sweet spot is around the 5th year of the PE fund. In addition, exits should be timed before the 10-year lifespan is up.

In Jun 2016, Azalea first introduced PE bonds to the market with Astrea III bonds. 2 years later, it launched Astrea IV bonds and just last week, it launched Astrea V bonds. The weighted average age of funds at the time of launch and scheduled call date of the 3 Astrea Class A-1 bonds are as follow.

Bond Wt Age of Fund Scheduled Call (Yrs)
Astrea III 6 3
Astrea IV 7 5
Astrea V 5.4 5

Their lifespans superimposed onto the PE J-Curve are shown in Fig. 3 below.

Fig. 3: Lifespan of Astrea Bonds Relative to PE Fund Lifecycle

Astrea III bonds (blue line) will last from the 6th to 9th year of the underlying PE funds that Astrea III invested into. This is the sweet spot that we discussed earlier -- enter around the 5th year and exit before the 10th year. It is probably the safest of the 3 Astrea bonds. In fact, it has just been announced that this bond would be redeemed as scheduled in 2 weeks' time.

Astrea IV bonds (purple line) will last from the 7th to 12th year of the underlying PE funds. The PE funds are generating a lot of cash currently, with total net distributions (after deducting capital calls) of USD243M in Astrea IV's first year of existence. This is equivalent to 22% of its portfolio value at the time of IPO. However, as shown in Fig. 3, the amount of distributions is expected to decline moving forward.

Astrea V bonds (brown line) will last from the 5th to 10th year of the underlying PE funds. Compared to Astrea IV bonds, there is a little more risks initially, as the companies in the PE funds are less mature and less ready to be exited. Furthermore, there are higher capital calls expected. But if these risks in the initial years can be overcome, Astrea V bonds will have less risks towards the end compared to Astrea IV bonds, as cashflows towards the tail-end would not have declined as much.

When you look at the risks of Astrea IV and Astrea V bonds, some of the safeguards that Azalea put in place for the bonds start to make a lot of sense. Let's talk about Astrea V bonds first, as it is simpler. As mentioned, one of the key risks for it is higher capital calls. This is mitigated by the capital call facility that allows Astrea V to borrow money from the banks to meet the capital calls.

For Astrea IV bonds, the key risk is not receiving sufficient distributions towards the tail-end to redeem the bonds in full. Although there is liquidity facility to meet interest payments in the event of shortfalls in distributions, there is no equivalent liquidity facility that Astrea IV can borrow from to redeem the bonds. Money to redeem the bonds can only come from the distributions from the underlying PE funds. However, towards the tail-end, such distributions are declining and might not be sufficient to redeem the SGD242M bonds in full.

In fact, it would be unfair to bondholders if it were to happen, considering that in the initial years of the bond, the underlying PE funds are generating a lot of cash, most of which are passed through to the sponsor. For the first year of Astrea IV, out of total net distributions of USD243M, all classes of bondholders (Class A-1, A-2 and B) were paid only USD26M in interest. Other expenses during the same period totalled USD7M. If there were no safeguards in place, the remaining USD210M (equivalent to 86% of net distribution of USD243M) and another USD15M in existing cash would flow to the sponsor. This money can leave Astrea IV as dividends to shareholder and repayment of shareholder loans instead of being retained within Astrea IV. Once the money leaves Astrea IV, bondholders have no recourse to Azalea as the shareholder/ sponsor to redeem the bonds in full. Recall that the bond is not guaranteed by either Azalea, or its parent, Temasek?

To prevent this scenario from happening, Azalea has put in place a safeguard that part of the distributions have to be set aside in reserves accounts for the sole purpose of redeeming the bonds. For the first year of Astrea IV, a total of USD80M has been set aside. Hence, the distributions flowing to the sponsor is reduced from USD225M to USD145M. This ensures that there will be sufficient cash to redeem the bonds when they mature, even though distributions from the underlying PE funds are declining.

In conclusion, PE bonds are not a simple matter of buy-and-forget. Investors need to understand where they stand along the lifecycle of PE funds and also what investor protection they have. Azalea has done a good job protecting investors from losses, but investors still need to protect themselves by learning more about PE investments.


See related blog posts:

Sunday, 16 June 2019

Astrea V 3.85% Bonds – Understanding What You Are Buying Into

It has been exactly a year since I last blogged. My last blog post was on Astrea IV 4.35% bonds. Coincidentally, Astrea's management, Azalea, has recently launched the IPO for Astrea V 3.85% bonds. One year has passed. What do I think about Astrea bonds?

If you read last year's blog post on Would I Invest in Astrea IV 4.35% Bonds?, you would know that I was not too keen on Astrea IV 4.35% bonds. A large part of the reasons had to do with Private Equity (PE) bonds being a new asset class and there was too little time to properly analyse whether it would be a good investment. Given the time constraint, I relied on whatever understanding I had about fund of funds and leveraged buyout funds and concluded that I would not be applying for the IPO.

A week later, after the IPO had closed, I had more time to look at the structure of the Astrea IV bond and acknowledged that it could be a safe one, but only because of all the credit enhancement safeguards put in place. See Understanding the Safeguards of Astrea IV 4.35% Bonds for more info.

Thus, when the IPO for Astrea V 3.85% bonds was launched this week, the first thing I checked was whether it has similar safeguards as Astrea IV 4.35% bonds. It has. Still, it is necessary to re-iterate that being PE bonds, Astrea bonds are not traditional bonds and it is important to understand the risks of the underlying assets. Below is a summary of the risks that I am aware of.

Understated Loan-to-Value Ratio in Fund of Funds

Astrea V bonds invest in 38 PE funds run by independent PE fund managers. Its stated Loan-to-Value (LTV) ratio for the Class A bonds (comprising Class A-1 and Class A-2 bonds which have equal seniority) is 34.8%. This means that for Class A bonds to start losing money, the value of the underlying investments has to drop by 65.2%. However, the underlying PE funds have their own debts and these debts are not considered when computing the LTV ratio of 34.8% for Astrea V bonds. The true LTV ratio after considering the debts in the underlying PE funds (i.e. look-though basis) is likely to be much higher. This ratio matters. See Would I Invest in Astrea IV 4.35% Bonds? for an example.

High Leverage Used by Buyout Funds

80% of the Astrea V investments are in buyout funds. As discussed in Would I Invest in Astrea IV 4.35% Bonds?, buyout funds use a lot of debts when acquiring companies. Typical debts is in the region of 6-7 times Earnings Before Interest, Tax, Depreciation and Amortisation (EBITDA). High debts at the underlying PE funds, couple with a Fund of Fund structure, underestimates the true, look-through LTV ratio of the Astrea bonds.

Assurance of Net Asset Value

Astrea V has a portfolio value of USD1,342M. This is an important figure that is used to compute the LTV ratio. After the debacle of the Hyflux preference shares and perpetual securities, it became clear that asset values should not be taken at face value. Hyflux's main asset, Tuaspring Integrated Water and Power Project, which has a stated Net Asset Value (NAV) of $902M as at end of Financial Year 2017, could not be sold at close to book value. Given that PE investments are illiquid assets, what is the assurance that the portfolio value of Astrea V is really as stated?

This question was posted during the Astrea Investor Day in Jan 2019 and during the public roadshow on Astrea V bonds conducted with SGX Academy on Saturday. Azalea's management replied that the NAV of PE funds is checked by reputable auditors. In addition, there are secondary markets where PE funds are traded. The value at which they are traded is close to the NAV reported by the PE fund managers. Furthermore, when the PE investments are disposed of, Azalea cross-checks the sale value against the reported NAV. In most cases, the sale value exceeds the reported NAV.

PE in a Potential Bubble

PE investments have generated better returns than public equities in the last 20 years. This has resulted in a lot of funds flowing into PE investments, and increased competition between PE fund managers to find good deals. This has led to assets being purchased at higher prices. At the same time, the debts used by buyout funds to acquire companies has been on the rise. At some point in time, the PE boom will probably end, potentially leading to falls in NAV. See Bain & Company's report on Private Equity: Still Booming, but Is the Cycle Near Its End? for more info.

At Saturday's roadshow, Azalea replied that this is also a good time for selling assets in the PE funds that Astrea V has already invested in, which will result in cashflows coming back to the Astrea bonds. Furthermore, as most of the PE fund managers have a lot of experience running PE funds, they believe that the PE fund managers will be able to navigate the environment.

While I agree that this is a good time for selling assets in PE funds, this also means that the high asset prices are reflected in the portfolio value of Astrea's investments. In the event that asset prices correct, Astrea's portfolio value will also decline. This will lead to a rise in the LTV ratio, but there is a safeguard in place if the LTV ratio exceeds 50%.

Conclusion

Although I believe Astrea IV and V bonds to be fairly safe for retail investors, I cannot emphasize enough that the reason this is so is because of the credit enhancement safeguards that Azalea painstakingly put in place. Also, for investors interested to buy Astrea bonds, please understand what you are buying into.


See related blog posts: