Galileo Japan Trust

The fat lady has sung

Final distribution for GJT

GJT is a gift that keeps on giving. Earlier in the year, I got a nice surprise from GJT when they announced one last distribution which was to be paid in October. So yesterday I received the final 2 cent distribution from GJT. All in all, shareholders have received the following capital distributions:

The share price was $1.77 when I first blogged about it in July 2015 so the cash profit is $1.13 (64% return). Not bad for a two year-ish investment.

Galileo Japan Trust

It’s a wrap

Sakura Sogo REIT successfully listed on the Tokyo Stock Exchange which brings the curtain down on GJT. As the sale proceeds from the property portfolio in Yen has been converted into Australian dollars at an exchange rate of A$1:¥78.08, shareholders are expected to get a special distribution between $2.68 and $2.69 and an ordinary distribution of between 3 and 4 cents.

I did a quick post mortem on the numbers based on what was disclosed in the initial information memorandum in February 2016 and discovered that some unexpected cost were incurred. Back then, management forecasted the special distribution to be $2.65 based on an exchange rate of A$1: ¥82. Assuming management forecast in Feb 16 were accurate, the favourable exchange rate movement since February should have resulted in a special distribution of $2.78 (8-9 cents higher than what was recently announced).

My analysis indicates that management’s February forecast didn’t include the following expenses:

  • Earthquake repair works in April 2016 after Kunamoto earthquake
  • Capital expenditure incurred for 6 months to June 2016
  • Unfavourable movement in interest rate swap in the Japanese TK business
  • Unamortised debt costs as at 30 June 2016 written off

Some of these were obviously unforeseeable (earthquake and interest rate movements) but I wonder whether the other costs should have been foreseen?

Luckily, the appreciation in the Yen more than offset these unforeseen costs. If the Yen had stayed at A$1: ¥82, my calculations show that the special distribution would have been $2.59. This would have resulted in a distribution about 2-3% lower than expected.

For long term holders, it doesn’t really matter but for short term traders it emphasis the lesson of allowing for some error in management’s forecast even for situations like this where the income and expenses are relatively easy to forecast.

Galileo Japan Trust

High five GJT holders

GJT has worked out very well for me and I’m particularly pleased with the result. I’m happy not only because the share price is up but more so because the opportunity to close the discount to book value was clear to me when I invested. That has finally occurred with the masterstroke by management to list the portfolio in Japan.

Management has done a great job in delivering shareholder returns post the 2013 recapitalisation. I’ve calculated the IRR post recapitalisation below to be in excess of 30% using last Friday’s closing share price of $2.44.


Shareholders have already passed the resolution for the IPO to proceed and management has advised that the expected capital return to investors would be $2.65 due in August/ September 2016. So should I sell my shares at Friday’s closing share price of $2.44 or wait for the capital return? The risks being that the IPO may not proceed and/ or the Yen falling against the Australian dollar (shrinking the Yen sale proceeds).

The IPO risks

Let me first say I do not have any information on the progress of the IPO process. So I’ve tried to assess the likelihood of the IPO’s appeal to the Japanese investor by analysing the potential dividend yield that an investor in the new Japanese entity (Newco) could achieve based on historical data. REITS are essentially yield plays and figuring out the potential yield relative to existing Japanese REITS would give a sense of the attractiveness of the IPO.

This analysis was done by assuming a geared and ungeared structure for Newco based on past proceeds received by GJT from the TK business in Japan and assuming that the portfolio value is ¥57.8 billion as indicated by management. No details on Newco structure was provided so I’m assuming the cost will be similar to the existing TK business.

IPO div yield

As shown above, the potential dividend yield range is 3.8% to 6.4% depending on the level of gearing.  My sense is that Newco’s ultimate gearing would be a bit lower than the current 59% D/EV gearing to enable the dividend yield to be around 5%. I’ve selected a handful of Japanese office focus REITS trading on the Tokyo Stock Exchange to get a sense of the dividend yields, these are shown below.


Compared to Australian REITS, the dividend yields on the Japanese REITS are quite low and this is probably driven by the current negative interest rate environment in Japan. Even on an ungeared basis, Newco’s dividend yield has the potential to surpass the above average dividend yield. Therefore, from a dividend yield perspective, I believe Newco has the ability to make an attractive IPO proposition to Japanese investors.

Exchange rate risks

In the event the IPO proceeds, management has forecast a capital return of $2.65 based an exchange rate of A$1:¥82. I have charted the AUD:JPY movement from January 2014 to present.


Some observations I would like to make on the exchange rate:

  • In my original post, I calculated the average 20 year rate (to Jan 2015) to be A$1: ¥81
  • The A$:¥ rate appears to be in a downtrend (Australian dollar falling) since November 2014
  • Based on the above period, there appears to be some support at the A$1: ¥80 level and the current rate of A$1:¥82 is close to that support level

Currency movements are terribly hard to predict. However, to get some understanding of the extreme scenarios, I have analysed the above trading data to work out the maximum depreciation and appreciation over a 100 day trading cycle. I choose 100 days to be representative of the trading period from now to the receipt of the IPO proceeds. Barring any “black swan” events, this trading period which covers 471 observations of 100 day trading should provide some clues to the potential extreme outcomes.

100 day AUDJPY

Over this period the maximum A$:¥ depreciation and appreciation over 100 days was -12.7% and 7.4% respectively. Applying the maximum depreciation and appreciation to the current rate of A$1:¥82 gives us a potential IPO proceeds range of $2.46 to $3.03 per share.

scenario - cap return

The shares last traded at $2.44 (8 August) and management has announced a 8.7 cents dividend payment in August so the theoretically undiscounted ex-dividend price is $2.35. This share price factors in a probability of an IPO success so we can take a further step in our analysis by including this assumption.

scenario - cap return2

Based on the pre-IPO announcement share price, I have worked out the potential uplift based on the base, high and low cases and assuming 80% probability of success and a pre-IPO announcement share price of $2. Basically if the IPO does not proceed, the share price falls back to $2. I believe 80% probability is reasonable given 1) Newco dividend yield will be attractive and 2) the current MD and CEO (Neil Werrett) will have an interest in the asset manager for Newco and this is an incentive for ensuring the IPO goes through.

In addition, I’ve calculated that the implied IPO probability based on current share price (and base case) is 55% which is too conservative in my view.

I’m not advocating a buy at current share price but as an existing holder of GJT, the share price is too low to consider selling now. I may consider selling  if the share price gets closer to $2.60 (pre-dividend) with the exchange rate still sitting around A$1:¥82.



Galileo Japan Trust

Land of the rising GJT…

Galileo Japan Trust FY2015 results

GJT announced its FY2015 results last week.  All in all, it was a strong set of results from the Company with a nice surprise from the rental revisions.  The key positive highlights are:

  1. In the recapitalisation memo in late 2013, management forecasted the annualised pro-forma net profit to be A$18.2 million (assuming A$1: ¥88) and the recurring net profit for FY2015 was $17.8 million (average exchange rate A$1: ¥95).  Even with a higher Australian dollar, they managed to perform in-line with the pro-forma.
  1. GJT to pay out dividend of $0.149 in FY2015, also in-line with its expectations and is 42% higher than $0.105 paid in FY2014.
  1. Weighted average cost of debt falls from 2.7% (FY2014) to 1.9% (FY2015) as a result of an early refinance of Eurobonds in Oct 2014. GJT paid $68 million including accrued interest to acquire all the Eurobonds.
  1. Upward rent revisions on nine office tenancies were successfully negotiated resulting in an average increase over previous passing rent of approximately 13.4%.
  1. The buyback of 1,059,250 units at an average price of $1.79 (below NAV) was completed.

An unfavourable result is that occupancy levels have slightly decreased to 96% (FY2015) from 98.8% (FY2014), details are:

  • Office portfolio occupancy rates: 99.1% (FY2014) vs 97.9% (FY2015)
  • Retail/ mixed portfolio occupancy rates: 98.9% (FY2014) vs 98.5% (FY2015)
  • Residential occupancy rates: 97.7% (FY2014) vs 96.4% (FY2015)

This fall in occupancy is mainly due to the lease cancellation of the single tenant (Tesco) at Funabashi Hi Tech industrial property in Oct 2014 and has remained vacant since.

With regards to the dividends in FY2016, the Company has entered into a forward contract to buy $9.166 million at an exchange rate of A$1:¥87.59.  Based on the current units outstanding of 105.4 million units, this shows that management is confident of paying $0.087 for HY2016; this is a 6% increase over the final dividend of $0.082 paid in FY2015.  Extrapolating this 6% increase over the entire year implies a dividend of $0.157 for FY2016, which in turns implies a 9% forward yield based on current unit price of $1.74.

The dividend yield is very attractive (FY2015 dividend yield – 8.6% and forecast FY2016 – 9%) in the current low interest environment.  Given twin pressures of low commodity prices and record low interest rates, it is unlikely that the Australian dollar will see a material appreciation.  However, although the Yen also has depreciated due to Abenomics, the current low energy and commodity prices should give its economy and currency a tailwind to move higher.

Net asset value has increased from my last post of $2.15 due to slight revision in property values and a fall in the Australian dollar. Nonetheless, using the old net asset value still provides a reasonable margin of safety at current share prices.

Overall, a strong result from GJT.

Galileo Japan Trust

Probably the cheapest REIT in town …

Galileo Japan Trust

Back in 1990, Japanese wealth was being supercharged by one of the greatest asset bubbles in financial history.  Land prices in Tokyo reached such astronomical levels that it was famously said that the imperial palace in Tokyo was worth more than all the real estate in California! What goes up must come down and the real estate bubble popped in 1991. By 2004 residential property in Tokyo was worth less than 15% of their 1991 peak. Here is a graph of commercial real estate values for Japan’s six largest cities from 1971 to 2014 which illustrates the size of the bubble.

Japan land prices

After the bubble popped, Japanese commercial property values fell rapidly to a low in 2004 and have gone nowhere in last 10 years.  However, there have been some recent reports that suggest that the property market is slowly recovering thanks to higher housing demand and the reflationary policies of Prime Minister Shinzo Abe. For an Australian investor, an easy way to get exposure to the Japanese property market is by investing in a Japan focused property group. Two such groups exist on the ASX: Galileo Japan Trust (GJT) and Astro Japan Property (AJP).  I like GJT better.

GJT owns the property portfolio through the Tokumei Kumiai (TK) investment structure.  The TK structure is similar to a silent partnership structure.  Under Japanese law this is a contractual relationship between an investor(s) and a TK Operator where the investor(s) provides capital to the TK Operator who acquires the assets and manages the business venture.  The investor(s) have neither control over the TK Operator nor the business venture but instead has a contractual claim against the business income and assets.

These schemes are set up for tax purposes and the net effect is that GJT will be entitled to 98.5% of the equity (TK Operator provided 1.5% of equity from its own account) and 97% of the profit/ loss of the TK business (TK Operator is entitled to the remaining profit/ loss).

  • Portfolio

GJT controls a mix property portfolio made up of offices, residential, retail and mixed use properties located mostly in the Tokyo and Greater Tokyo area along with a few properties in Osaka.  A snapshot of its properties from HY2015 shareholder’s presentation:

Property portfolio

As shown in the table below, properties values have been falling from 2010 to 2013 whilst gross property yield have been rising.  Occupancy levels have been steady over the last 4 years and close to 100% occupied.

1st table

In 2014, GJT sold two non-core properties:

  1. Doshoumachi in Osaka for ¥800 million which is 14% above its carrying value; and
  2. Lion Square in greater Tokyo for ¥2.385 billion which is a whopping 32.5% above its carrying value.

The results of these transactions may suggest that GJT’s valuation is on the conservative side or a result of them selling their best properties.

  • Financials

I’ve re-calculated the earnings to the unit holders from the reported income statements and removed some of the one-off and non-cash items to give sense of the recurring cash profits attributable to the unit holders.  The pro-forma FY2014 earnings were previously provided by management during the re-capitalisation exercise in the 1st half of FY2014. The recapitalisation exercise significantly reduced debt levels and this enabled GJT to revert back to its focus of managing its portfolio and paying distributions.

2nd table

A few observations I would like to point out:

  1. Management fees as a percentage of average portfolio value are broadly in-line with the fee schedule (0.65% management fee and 0.1% responsible entity fee –reinstated after the re-capitalisation).
  2. One of the reasons why rental income in Australian dollars has fallen from FY2012 – FY2014 is due to the Australian dollar appreciating against the Yen.
  3. One of the key reasons why EBIT has been declining from 2010 was due to the previous strategy of selling off properties to repay debt. As costs did not fall in tandem with the loss rental revenue, earnings declined.
  4. Interest rate on the outstanding debt has also been falling. Based on the HY2015 presentation, the current effective rate is 3.3% p.a. (the above 3.5% was calculated based on finance cost and average debt balance).
  5. Moving forward, without any significant changes to the portfolio, I expect the cash earnings to be in line with management’s pro-forma guidance. Based on the results in  HY2015, GLT is on track to meet this guidance.

Financial Position

3rd table

As at 31 December 2014, GJT had $350.4 million in outstanding debt made up of senior debt (83%) and mezzanine debt (junior and senior) making up the remainder of the debt.  The loans are all due to mature in October 2018 and interest rates are mostly fixed from a fixed for floating interest rate swap facility.  As at HY2015, gearing (debt/ portfolio value) is 62.5% and interest coverage ratio (net profit before interest/ interest) is about 2.5x.  In my opinion, the gearing level is not excessive for a REIT with stable cash flows.

As at 31 December 2014, the book value per unit is $2.15 (which I consider to be its current intrinsic value).

  • Dividend yield and price to book value relative to other REITS

After the recapitalisation exercise, GJT was able to pay distributions and have paid the following distributions:

  • FY2014: 10.5 cents per share
  • FY2015: 14.9 cents per share (estimated)

To get a sense of the recurring level of dividend based on the current portfolio, I estimate that the recurring earnings (assuming exchange rate remains at these levels) to be $17.2 million based on the pro-forma earnings adjusted for the loss of income from the sale of the two properties in FY2014 (they contributed to approximately 6% of the portfolio).  I understand that capital expenditure for maintenance and additions are circa $2 million a year, which leaves free cash for distributions around $15 million or 14 cents per share.  So I believe the current distribution level is sustainable (assuming no major drop in the ¥).

Comparison to other REITS (constituents of the REIT index) on the ASX:

4th table

Based on the above metrics (dividend yield and price to book value) it appears that GJT is the cheapest of this group.  There are 40 plus property groups trading on the ASX and some have a higher dividend yield GJT but I couldn’t find one that had both a superior yield and a lower price to book value than GJT.  In August 2014 GJT announced a share buyback program (maximum buyback lesser of $5.0 million or 10% of its capital).  So far it has bought back about $1.8 million worth of units at prices below the book value per share – great for existing shareholders.

  • Margin of safety

I purchased the shares last year at $1.55 which provided a margin of safety of around 28%.  The margin of safety today (calculated based on the book value per share and current share price) has reduced to about 17%. Whether or not a margin of safety of 17% is adequate depends on your risk adverseness but for me personally, I would probably wait for another 5% pull back from current levels before taking the plunge.  I think with the current Greek problems, there is a real possibility of a drop to that level.

  • Risk Factors

 An investment in GJT is not without risks and the key risks I see with GJT is:

  1. A decline in the ¥ (GJT doesn’t hedge against movements in A$:¥)
    The chart below shows the A$:¥ over a 20 year period.  The average rate over the period is $1:¥81 and the A$ is current trading at the ¥94 level.  So the A$ is trading above its average historical level. I don’t know where the currency is heading to but I believe currency movements are cyclical and mean reverting. The quantitative easing (“QE”) policies in Japan which started in December 2012 have driven the Yen down by 2.3% and 19.3% against the Australian dollar and US dollar respectively.  If recent US experience is anything to go by, then the Yen should spike when the Bank of Japan starts to signal a taper in the QE program. 5th table
  2. Continued deflation in the Japanese economy
    I have no idea whether a deflation environment will further drive down property prices but land price data shows that commercial property prices in major urban areas have grown by 0.1% and 3.6% in 2013 and 2014 respectively. Obviously, a risk exists that deflation may take hold and drive prices down again.  However, intuitively speaking, if Japan’s inflation rate is persistently lower than Australia’s inflation then all things being equal, Purchase Price Parity theory states that over the long run, the Yen should appreciate against the Australian dollar.
  3. Increasing maintenance expenditure (Capex)
    All of the properties in the portfolio were completed between the periods 1986 to 2005 with the majority built from the late 80s to the mid 90s.  So the buildings are not too old and average capital expenditure from FY2009 to FY2013 has been around $3 million a year. This is somewhat less tangible but I believe the Japanese culture of “taking pride and respecting their work place” should mean on average buildings in Japan are better maintained.
  4. Natural disasters
    Japan sits at the meeting point of three tectonic plates and is therefore very prone to earthquakes.  There have been 69 earthquakes greater than a 7.0 magnitude that have occurred over the last 4oo years in Japan.  As a consequence, Japan has one of the strictest building codes in the world and after the 1978 Miyagi earthquake, comprehensive seismic requirements in building standards were introduced in 1981.  All of the properties in the portfolio have been constructed after 1981 and none of properties in the portfolio is currently insured for earthquake damage.  GJT’s policy is to procure earthquake disaster insurance for properties only where the PML is in excess of 15% (none at the moment).  I understand that this is due to the cost of earthquake insurance.
    Earthquakes are nearly impossible to predict and the probability of losses on the portfolio is beyond my capability to assess. So my strategy is too not put all my eggs into one basket and to exit the investment when the share price reaches its intrinsic value (this is not a hold forever type investment).
  • Summary – Key reasons to like GJT

    1. The current dividend yield is attractive (dividends are tax deferred as well) and sustainable (if ¥ holds up). So my strategy is to collect a nice dividend while waiting for the share price to align with the intrinsic value.
    2. Recent properties sales may potentially indicate that the portfolio’s carrying value is below market value.
    3. It provides a hedge (not perfect) in the event the Australian dollar declines. Falling commodity prices and the RBA intent to drive the dollar down worries me.
    4. Tokyo will host the 2020 Olympic games.  This could be a catalyst for a boom in property prices as experience by many past host cities.

So why is the market pricing GJT at a discount to its peers? The most obvious answer to me is due to size, liquidity, GJT not being a constituent of the S&P ASX200 A-REIT index and market’s expectation of a continued decline in the Yen (which I’ve covered in the risk para).

  • Size – GJT’s market capitalisation is less than $200 million compared with most of the REITS mentioned above which have market capitalisation in excess of $1 billion (National Storage market cap is about $500 million).
  • Liquidity – GJT shares are not actively traded so liquidity is low. For example, 11,700 shares was traded today and I hardly see more than 100,000 shares a day being traded.
  • S&P ASX200 A-REIT index – GJT is not a constituent of this index probably due to the size and liquidity issues and therefore unlikely to get the attention of the large institutional fund managers.