Australian Property Trusts: Illustrating the possible impact of stripping the Safe Harbour from July 2023.

Things have been challenging for the Australian commercial property sector over the last few years. Cap rates are nowhere near as optimistic as they used to be. Negative carry is more widespread. WALEs and occupancy rates are down, with fewer anchor tenants willing to commit to long term leases. The double digit percentages being modelled are the cost of financing, not the cash yields which were previously commonplace. The previous shortage of investment grade CBD commercial property has swung to oversupply, thanks to the structural shift in working habits in the post Covid world. Claims that this is a temporary blip, are now viewed sceptically.

In the midst of all this doom and gloom, Treasury released a consultation paper in August 2022 seeking comment on changing the thin capitalisation regime by replacing the current asset-based tests with earnings based approaches. There are also proposed changes to the arm’s length debt test. A whole page was dedicated to these proposed thin capitalisation changes in the government’s budget papers in October 2022.

This note is not advice. It will however attempt to crudely illustrate, for non tax practitioners, how the proposals to replace the asset based safe harbour with an earnings based fixed ratio, (intended to apply to income years commencing on or after 1 July 2023) might impact a commonly used 3 tier fixed unit trust investment structure for Australian commercial property.

Commonly seen 3 tier Australian unit fixed trust structure

The typical 3 tier fixed unit trust structure I tend to come across involves an asset trust, a mid trust and a hold trust. This type of Australian fixed unit trust structure is commonly used by foreign domiciled, close ended master fund to invest in Australian real estate. The relevant entities usually qualify for either Managed Investment Trust (“MIT”) or Attribution Managed Investment Trust (“AMIT”) treatment.

[For the readers unfamiliar with Australian property trusts, fixed unit trusts are both income and capital transparent. All trust income needs to be distributed to beneficiaries for maximum efficiency. MITs and AMIT are types of REITs with concessionary features. The ability to make income distributions to non resident beneficiaries attracting only 15% withholding tax rate and tax deferred capital distributions are amongst its most attractive features.]

Hold trust and mid trust function as intermediate holding vehicles for the master fund. An asset trust typically holds a single property only. Consequently, multi property type portfolios tend to comprise multiple asset trusts owned by multiple mid trusts. Additionally, some structures use the different tiers of trust vehicles to structurally subordinate funding, to comply with financier requirements.

For example, asset trust may be funded with third party senior debt and equity from mid trust. Mid trust’s investment in the equity of asset trust may be funded via a combination of mezzanine debt and equity from hold trust. Hold trust in turn could be funded with a combination of related party debt and equity from the master fund. Not all structures follow this formula. There are many variations, such as running the various forms of funding back to back downstream or changing the levels at which the various types of debt are injected down the vertical silo of trusts.

Example for illustration purposes

• Debt covenants limit total gearing to a 65% loan to valuation ratio;
• Asset trust acquires a ready for occupancy new build CBD office tower for $100m;
• Land value accounts for $50m of consideration paid;
• Capital allowances are claimed over 40 years for the $40m cost of construction;
• Final fit out costs of $10m are depreciated over 10 years;?
• Asset trust borrows $50m via syndicated bank facility at a 2% margin over base rate;
• Mid trust provides $50m in equity to asset trust;
• Mid trust borrows $7m in mezzanine debt at a 3% margin, $7m of related party debt at a 4% margin and receives $36m of equity investment from hold trust;?
• Hold trust itself is funded with $40m in equity from the master fund;
• The 2% base rate in Year 1 drops to 1% in Year 2 and stabilises at 3% from Year 3 onwards;
• The property’s value appreciates at 2% p.a.; and
• Surplus cash funds distributions (if required). ??

Head Trust

Functions as a conduit and as such, it should just flow through distributions from Mid Trust.

The current Australian thin capitalisation regime

The maximum allowable debt a taxpayer is allowed to have, is the greatest of:

• the “safe harbour” debt amount which is broadly 60% of the adjusted accounting value of the entity’s Australian assets (for non financial entities);
• the “worldwide gearing” debt amount requires taxpayers to work out the gearing ratio of the global group as a percentage and apply this against the adjusted accounting value of Australian assets to quantify the debt amount; and?
• the “arm’s length” debt amount essentially requires an analysis of the maximum amount of debt the entity could reasonably have borrowed from commercial lenders on a stand alone basis, without the benefit of parental support. Essentially this method is used to support gearing in excess of the 60% safe harbour or the worldwide gearing ratio to satisfy thin capitalisation restrictions.?

Wholly owned Australian companies have the option of forming a consolidated group (akin to a fiscal unity in other jurisdictions). If this occurs, only a single consolidated thin capitalisation calculation is required for the group. No such concession is available to wholly owned trust groups that retain flow through tax treatment. Consequently, each of the three trusts will need to separately ascertain if they have exceeded their own maximum allowable debt amount.

The worldwide gearing amount largely replicates the safe harbour debt amount calculation with some adjustments to work out a global group gearing ratio. In terms of the illustrative example, the master fund would need to take on debt to produces a more favourable outcome (i.e. >60% gearing ratio). To date, I have not encountered this funding approach for inbound real estate investors with an exclusively Australian asset base. It is for this reason that this note skips a comparative analysis of the worldwide gearing test against the earnings based approach.

The relative certainty of the Safe Harbour Debt Amount

The debt amount tested at the asset trust level is the $50m senior debt. The safe harbour debt amount is, crudely speaking, 60% of the average accounting value of the asset trust’s asset adjusted for non debt liabilities (which tend to be working capital items). As the asset trust is geared well within the 60% threshold, no interest deductions are denied.

You will note that there is excess borrowing capacity in the asset trust in the example. The debt capacity increases in step with underlying property value growth. Property groups regularly revalue properties and reflect this in their financial statements in accordance with accounting standards. This increases debt capacity in advance of refinancing events.

For thin capitalisation purposes, parent entities can access the excess debt capacity via the associate entity excess amount. Mid trust relies on the excess debt capacity of asset trust to support its debt load as it has no other assets other than its investment in asset trust. Although mid trust is over geared from a tax perspective, it is still within typically acceptable commercial LVR ranges. In any case, this position reverses from Year 3 onwards due the underlying property value growth.

The asset based safe harbour model suits investments that are capital intensive. It is especially forgiving for ventures with long construction periods requiring extended lead times to full earnings ramp up. The safe harbour is critical at the evaluation stage as it enables investors to target and model for a particular debt to equity ratio. Additionally, it weathers earnings volatility better than the earnings model, as evidenced in the example above where the falling rental yield did not adversely affect debt carrying capacity, as long as asset values remain stable or increase.

Finally, the asset based model provides more flexibility for remedial action. Taxpayers have the option of making capital contributions to top up the equity in the structure if a diminution in asset values reduces debt capacity. From a tax integrity perspective, the current system works in that debt capacity is driven entirely by the underlying value of the Australian asset base. Also, there is no ability to achieve arbitrage by double counting assets. The method statements have been designed to only take into account an Australian asset’s value once.

The earnings method, a fixed ratio over the uncontrollable

No specific details have been released in relation to how the earning based fixed ratio is to be quantified, consequently my illustrative example is largely driven off accounting. This should be a reasonable proxy, given the relatively limited types of income streams likely to arise in real estate investment but it may not be suitable for other sectors.

Earnings and the cost of funding are the two key attributes that affect the quantum of interest deductions denied under the proposed earnings approach. The events of the last few years have shown that landlords cannot maintain or increase earnings if tenants are unable to pay rent and occupancy rates continue to decline. Similarly, central banks and financial markets set the cost of financing. As illustrated in the example, any change in these two uncontrollable attributes resulted in a punitive outcome despite the debt levels being within commercial norms; i.e. a 65% LVR.

The example shows mid trust in receipt of income distributions exceeding accounting profit. This reflects increased taxable income from the denial of interest deductions. Tax transparency of trusts is achieved only if all taxable income is distributed to beneficiaries. This in turn, creates a funding shortfall for payment of trust income distributions. Cash that could otherwise fund tax deferred distributions would need to be diverted to fund the additional income distributions.

If the debt funding is sourced from non resident lenders, things get worse. Interest withholding tax needs to be remitted even if deductions are denied pursuant to thin capitalisation. Additionally, the denied interest deduction is treated as an income distribution subject to 15% withholding tax when repatriated to non resident beneficiaries by hold trust. This has the effect of raising the effective tax rate on the non deductible interest to 25%. The total cash return on investment has been dually compromised by the higher effective tax rate due the denial of interest deductions and reduced capacity to fund tax deferred distributions.

If the proposed 1 July 2023 start date stands, there is little that can be done in terms of remedial action and no time to undertake any such action. Also, the 15 year time limit on the carry forward of excess interest deductions could be problematic for investments with long construction periods. Shifting to this earnings based model without allowing for time to transition out of current arrangements will adversely impact an already embattled sector.

The task of reducing the amount of third party sourced debt is a complicated exercise. Also, capitalising related party debt may have little or no impact on the outcome. Whilst the revenue authority would be happy to have debt levels reduced, there will undoubtedly be many roadblocks in attempting to increase gearing when things turn around. This is despite the debt creation rules having been repealed over two decades ago.

Both the consultation paper and budget paper are silent on what form of grouping type relief is likely to be made available, if any. Ideally, some form of ability to access excess debt capacity in a similar vein to the current associate entity excess amount rules should be considered. There would also need to be clarity around what the definition of debt to be tested under the model. For instance, some form of exclusion for on lent third party debt should be introduced to prevent punitive double counting.

The other slightly strange feature of this model is that the same rental income stream is being repeatedly used to ascertain the amount of interest deductible under the fixed ratio for the different tiers of trusts. This contrasts with the asset based model where the key attribute that drives debt capacity is only taken into account once at all tiers of the structure.

Arm’s Length Debt Test

The current legislation applies the arm’s length debt test on an entity by entity basis. For a measure that purports to attempt to mirror real world credit analysis to determine a genuinely commercial level of debt for the business from independent commercial lenders, the approach adopted to date diverts substantially from commercial reality. This is especially the case in dealing with debt facilities provided to multi property portfolios. This has been a sticking point for some time, including the classification of external vis-à-vis related party debt where third party debt is on lent on a back to back basis within the vertical silo. It remains to be seen as to whether there will be any favourable changes to what may now become the preferred method of working out maximum allowable debt if the asset based model is replaced.