Digital Core REIT

[specuvestor]

default 18 months after IPO... quite a coincidence

it's like an asset swap with Digital Realty with residual... wonder if Digital Osaka 3 or 11 Loyang Close has any sour taste in it My guess 2.5% is to mark up the pricing for whatever purpose

DCR had its 2nd largest tenant defaulting in mid 2023. The 2nd largest tenant occupied a total of 6 facilities (5 of them 100% and the 6th at 4%).

[swakoo]

By now, they should have seen that conditions on the SGX are much conducive to continue to rear a golden goose for its eggs than slay it for the meat.

For a reit of the same asset class, it's distribution yield is typically very much less when listed in Wall St than in SGX (in this DC case 2+% vs 7+%). Ie it's priced much higher in Wall St. That means the eggs laid in Wall St are worth much more. Perhaps if the goose has to be slaughtered for the meat, better to do it elsewhere?

Attached chart shows how the sponsor is doing vs the "baby" since the latter was "born".

[weijian]

By now, they should have seen that conditions on the SGX are much conducive to continue to rear a golden goose for its eggs than slay it for the meat.

For a reit of the same asset class, it's distribution yield is typically very much less when listed in Wall St than in SGX (in this DC case 2+% vs 7+%). Ie it's priced much higher in Wall St. That means the eggs laid in Wall St are worth much more. Perhaps if the goose has to be slaughtered for the meat, better to do it elsewhere?

Attached chart shows how the sponsor is doing vs the "baby" since the latter was "born".

hi swakoo,

I have not specifically looked into the parent (and don't intend to). Therefore, I wouldn't be able to effectively address the difference that you mentioned.

But what I can clarify is that there is a difference between distribution yield and cap rate.

A REIT's distribution yield is a function of net rental income after financial engineering.

A real estate's cap rate is a function of its net rental income. This is generally the figure used for property transactions (ie. weighing the goose's meat).

[weijian]

default 18 months after IPO... quite a coincidence

it's like an asset swap with Digital Realty with residual... wonder if Digital Osaka 3 or 11 Loyang Close has any sour taste in it My guess 2.5% is to mark up the pricing for whatever purpose

DCR had its 2nd largest tenant defaulting in mid 2023. The 2nd largest tenant occupied a total of 6 facilities (5 of them 100% and the 6th at 4%).

Hi specuvestor,

Investing is probabilistic after all. So sometimes a coincidence is just a coincidence. And sometimes a coincidence is probably too much of a coincidence.

The customer defaults at DCR had also affected the parent. But in this case, the parent's size was a factor of 25-30x bigger than DCR and hence the proportional impact was different. For example, the 2023 default (Cyxtera) accounted for 22% of DCR's annual revenue but only <2% of Digital Realty's income. So DCR had an arm cut off while it was just a small open wound for the parent.

And to be fair, Digital Realty did come in with a series of recovery actions like getting in a new capital partner (Brookfield Infrastructure Partners) to take over some of the obligations, buy back some of the data centers which suffered default and supported part of the rental loss through top-ups in the interim.

[specuvestor]

Hi Weijian

Yes that's one way to look at it mathematically but the other way is that the parent proactively "shared" the impact of "too much of a coincidence"

None of us here should bother but one way to verify is to see what other DC that the parent hold actually had the same counterparty default ie coincidentally all the assets that defaulted with this client was in DCREIT and none in the parent. That's a 6 out of 6 strike Maybe there's more than 1 lemon

Also interesting that they sold 39% of Linton Hall (but still consolidated at 51% ownership) when it "recently secured a 10-year agreement with an unnamed “investment grade global cloud service provider” for the whole facility" after existing sole tenancy ends 2026: https://www.datacenterdynamics.com/en/ne...s-in-asia/

Frankly I was wondering if there were any good reason they bothered to list in Singapore rather than US, except the US investors probably know their assets better

Many interesting points bought up by Dr Mak but I thought the structure is the most important part here.

DIGITAL CORE REIT: ANOTHER FOREIGN REIT ON A BUMPY RIDE

We pointed out a number of unique features and concerns about DCREIT. First, its sponsor and manager, Digital Realty Trust, is listed on NYSE and was the sixth largest publicly traded US REIT. It is the first time that a SGX-listed REIT has a sponsor and manager which is another (listed) REIT.

The sponsor carved out 10 freehold data centres in US and Canada out of a global portfolio of 290 facilities in 26 countries to form the initial portfolio for the listing of DCREIT. We questioned how the 10 data centres were selected and valued.

We concluded our writeup with this statement: “Digital Core REIT has a financially strong sponsor listed on the largest stock exchange in the world. However, the sponsor has its own unitholders to account to, and the interests of those unitholders may not necessarily be the same as the unitholders of Digital Core REIT here.”

https://governanceforstakeholders.com/20...umpy-ride/

[swakoo]

But what I can clarify is that there is a difference between distribution yield and cap rate.

Hi weijian, ya know the distinction between the 2. The distribution yield is a very rough reflection on cost of equity. It can be easily and quickly extracted, so have used it as a rough quick and dirty benchmark for difference in acquisition prowess of assets between the 2 reits. And the gap is huge.

[weijian]

None of us here should bother but one way to verify is to see what other DC that the parent hold actually had the same counterparty default ie coincidentally all the assets that defaulted with this client was in DCREIT and none in the parent. That's a 6 out of 6 strike Maybe there's more than 1 lemon

hi specuvestor,

At point of bankruptcy, Cyxtera had leased ~60 data centers. 15 of them were with Digital Realty (DR) and 6 of them with DCR. Closer to home, Keppel DC Reit was impacted too with 1 data center 100% leased to Cyxtera but Keppel DC Reit dealt with the UK subsidiary and hence was not affected at material time.

Therefore, both parent DR and DCR were affected and due to size and hit rate, the latter was impacted ~22% of annual revenue and DR only <2%. Of course one doesn't need to be a statistician to appreciate the statistical significance

[weijian]

But what I can clarify is that there is a difference between distribution yield and cap rate.

Hi weijian, ya know the distinction between the 2. The distribution yield is a very rough reflection on cost of equity. It can be easily and quickly extracted, so have used it as a rough quick and dirty benchmark for difference in acquisition prowess of assets between the 2 reits. And the gap is huge.

Hi swakoo,

Unlike DCR which only owns data center assets, parent Digital Realty (DR) also actively develops and operates data centers. This is 1 fundamental difference which probably explains much of the difference in valuations between both. This is exacerbated by the fact that data center infrastructure are the hottest asset class right now, attracting most attention and fund flows.

DCR IPO-ed in late 2021 at ~1.04x P/B and a target forecasted distribution yield of ~4.75%. DR's P/B has always been comfortably at >1.7x P/B and above, but US GAAP records its operating data centers as "historical cost minus depreciation", ie. very similar to how IFRS records PPE in our part of the world. So the comparison is not apple to apple as DCR's data center assets are at fair value.

Now back to the question of where to keep the goose, whether is it to slay or keep it for laying eggs. Pre-covid, there is a reason why we see many condos in Cambodia or Iskandar been sold in Spore. And move a few years back, they even got a few ex EPL players to advertise US/UK land banking.

Some assets are simply meant to be sold.

[swakoo]

Unlike DCR which only owns data center assets, parent Digital Realty (DR) also actively develops and operates data centers. This is 1 fundamental difference which probably explains much of the difference in valuations between both. This is exacerbated by the fact that data center infrastructure are the hottest asset class right now, attracting most attention and fund flows.

DCR IPO-ed in late 2021 at ~1.04x P/B and a target forecasted distribution yield of ~4.75%. DR's P/B has always been comfortably at >1.7x P/B and above, but US GAAP records its operating data centers as "historical cost minus depreciation", ie. very similar to how IFRS records PPE in our part of the world. So the comparison is not apple to apple as DCR's data center assets are at fair value.

Yes the rules and regulations governing reits listed in Sg and US are very different. That's why I qualified my post with "rough quick and dirty benchmark". 

Also AI is evolving so rapidly that it is a realtime task just to understand well the (hottest asset class) DC layer alone. It compounds assessing well the different types of DCs, including usage, orchestration, obsolescence and geographical implications.

Now back to the question of where to keep the goose, whether is it to slay or keep it for laying eggs. Pre-covid, there is a reason why we see many condos in Cambodia or Iskandar been sold in Spore. And move a few years back, they even got a few ex EPL players to advertise US/UK land banking.

Some assets are simply meant to be sold.

I like and appreciate how you meticulously "measure and weigh" the goose both before and after it is sold. 

The goose is never to be taken for granted. 

[weijian]

The goose is never to be taken for granted. 

Hi swakoo,

From what has already transpired, it seems that the goose may have been chosen, exported and slaughtered for its meat. But that doesn't mean that the plate has no braised goose meat left though.

IMHO, the Sponsor has done the right moves when its share price is trading at a big discount to NAV - mainly selling assets near NAV and doing SBB. The latter is rare for REITs because Mgt (controlled by their Sponsors) are not incentivised to do SBB since it reduces Sponsor's AUM. For example, we can just reference to Capitaland China Trust (CLCT). CLCT sold a significant asset (divested to C-REIT) and received proceeds that is ~10% of its market cap. It has been almost 1 year and it has done no SBB (with SBB been 1 of the actions it said it could take)

For DCR, the high cost of equity has no light at the end of the tunnel and so it is probably unable to do EFRs to acquire new data centers. CLCT is in a similar situation. So when DCR decides to do some decent SBB, then it may mean something.

DCR SBB spending history so far (in usd)
1H26~3.9mil
FY25~1mil
FY24~15.5mil
FY23~3.4mil
FY22~5.8mil

For all we know, the slow wind-down of DCR may be in motion, isn't it? ~60% of the proceeds in this 2nd divestment (the first divestment happened in Jan2024) will be used to retire debt, rather than for new acquisitions. So that feels like another anomaly to me. DR is the chef of this braised goose meat dish and they have intimate knowledge of the remaining assets. IF they are compelling enough, it could just be a matter of time.....