Rubicon — Growth to come after FY19 challenges

Reported FY19 results were broadly in line with our estimates on an IFRS basis and at the US GAAP EBITDA level. This was the first full year of reporting with a March year end and with this corporate structure. We are now able to clearly see the seasonal attributes of the business and, notwithstanding weather-related disruption in H2, a stronger year-on-year trading performance in the second half of the year. No dividend was declared as expected and Rubicon ended FY19 with US$22m net debt.

At the headline level, revenue and gross profit (at US$49.1m and US$18.8m,1 respectively) were both c US$1m lower than our pre-results estimates, which we consider to be within a reasonable range of variation given the nature of weather effects in the year. Specifically, there were two significant hurricane events around the turn of the half year and the impact on forestry customers (ie initial damage incurred and subsequent ground condition evaluation) took time to be fully evaluated. Based on management comments, we believe that the majority of the group gross profit difference versus our expectations is attributable to unsold seedlings due to weather conditions written-off, taken above the line. Nevertheless, benchmarking H219 against the prior year, we note that revenue was c 10% above the H218 level with US$0.8m higher gross profit (albeit with a 210bp lower margin yo-y). Tree seedling planting seasons dictate that H2 is the dominant trading period for ArborGen particularly from a US perspective, so to make progress against the prior year despite weather challenges is a creditable performance we feel. In reported NZ IFRS terms, the profit shortfall against our expectations was effectively compensated for by tighter central cost control and a positive change in the fair value of biological assets – which we had assumed to be a neutral item – such that EBITDA, EBIT and PBT were all in line with our estimates.

North America: Against the company supplied pro forma FY18 seedling sales figure of c 270m, more seedlings were sold in FY19 at c 280m, a c 4% y-o-y increase. Moreover, a more favourable mix – including a higher proportion of MCP2 and varietal seedlings – and, we believe, some general firmness in pricing, meant that revenue generated in this region was stronger than we had modelled. ArborGen has historically had stronger market positions with larger independent commercial forestry owners who have generally been more receptive to moving from traditional open pollinated to the latest generation mass control and varietal seedlings to increase yields. In FY19, the company noted more emphasis on private landowners and this may have been a contributory factor in average selling price progression. As noted earlier, gross margin was affected by seedling impairment charges taken above the line.

FY19 represented ArborGen’s first year of operation of the Taylor Nursery, Trenton, on behalf of the South Carolina Forestry Commission. This 10-year agreement aspires to increase production capacity to 30m seedlings per annum in FY20 (from c 5m pre-agreement), achieving c 15m maiden production for ArborGen in FY19.

In November 2018, ArborGen also announced a five-year agreement with TexMark Timber Treasury (Triple T) to lease its seed orchards and nursery in the contiguous Jasper and Tyler counties, TX. This adds a further 30m seedling capacity, commencing FY20. During this lease period, ArborGen will be the exclusive supplier of Triple T’s Texas loblolly pine seedling requirements; there is provision to extend the agreement and, at the end of the initial term, put and call options to transfer the ownership of the leased properties for US$2.5m. ArborGen now has eight seedling nurseries in its US portfolio (six owned, two operated) with an estimated total capacity of c 400m seedlings per annum. This compares with the 280m sold during FY19.

Hurricane damage to the Bellamy, FL, seed orchard was flagged in October 2018. However, ArborGen has other orchards in the Atlantic Coast, Piedmont, Arkansas and South Texas regions in which there are extensive but relatively young trees of latest generation genetics that are expected to attain their mature seed production levels in two to three years’ time. Consequently, near-term constraints in building higher-value MCP seedling volumes are expected to dissipate. So, while there should be some post-hurricane rebound in seedling demand, incremental FY20 sales (and therefore the overall mix) could potentially dilute the average selling price (ASP). This effect is expected to start to ease during FY21. For the record, our updated model contains annual seedling volume growth of 12% in FY20 and 6% in each of the following two years.

Brazil: We do not have year-on-year comparators, but, as shown in Exhibit 1, 50m seedlings were sold in FY19. Revenues were affected by demand weakness in the dominant eucalyptus forestry sector in FY19, exacerbated in the reporting currency by a c 14% depreciation of the real against the US dollar. In contrast, pine seedling sales were stated to have been better than management had anticipated. The Brazilian economy itself was seen to be decelerating in the second half of the year and this coincided with a general election process culminating in a new government coming into office on 1 January. Economic weakness was also seen in neighbouring South American countries, which could have had some effect on sentiment. That said, the weaker real may have had some positive effect on exports of downstream forestry products further afield (key sectors here being pulp in particular but also paper and wood panels), which would be supportive of replanting activity in due course. In trading terms, Brazilian operations generated a positive gross profit contribution (on US$6.1m sales) but an operating loss of US$1m, implicitly carrying an overhead base for a higher revenue stream. Looking ahead, the new Brazilian government was elected on a social reform ticket, aiming also to tackle public debt levels but it is too early in its term to assess progress, though higher GDP growth is expected in 2019 compared to 2018.

Australasia: Seedling sales grew by c 20% in FY19. This region enjoys the highest ASP in any of the areas served by ArborGen – reflecting the higher proportion of advanced seedlings sold – and pricing appears to have been firm year-on-year. The company’s national nursery coverage in New Zealand has been advantageous in securing a 12-month seedling supply contract with Crown Forestry to support the NZ government’s 10-year, One Billion Trees planting programme, getting underway in earnest in FY20. We assume that there are good prospects for follow-on agreements. Few additional details are given in the results release, but we understand that the expectation is that that the region (and both countries within it) will achieve record sales and profits in FY20.

Exceptional items: The year also included US$4.1m of separately disclosed, non-underlying charges. Just over three-quarters related to the One Company restructuring programme (including severance in NZ and the US) and the remainder from transaction costs arising from Rubicon’s acquisition of the ArborGen minority interests. After taking a US$0.5m property disposal gain the net exceptional charge for the year was US$3.6m.

It is worth commenting here on business seasonality and the working capital cycle as reflected through the cash flow statement. Reflecting the key US planting and therefore seedling selling season, ArborGen’s profitability is strongly weighted towards its second half trading period.

Seed orchard harvesting takes place during H2 of the financial year so seed inventory rises during the period and peaks around the year end. Meanwhile, higher-value seedling inventory value in the US peaks early in Q4 (financial year) just prior to the US lifting season, thereafter moving in the opposite direction to the year end. The net profile at group level is inventory absorption in H1, release in H2. Receivables obviously follow the seedling sales cycle, so broadly peak at the year end, with a positive collections effect in H1 cash flow. The caveats to these observations are that they depend on local climate and weather conditions in each of the regions served and also noting that we only have one full year of reporting and that was affected by adverse weather conditions in the company’s largest market. That said, we would expect the pattern to be broadly similar in most normal trading years.

FY19 cash flow performance: Rubicon ended FY19 with c US$22m net debt, an increase of c US$12m over the course of the year (and c US$3m above our estimate). Underlying free cash flow was neutral in the period so the outflow can be attributed to the exceptional items noted above and net corporate activity (ie US$10m final deferred consideration paid to ArborGen minority shareholders less US$2.4m received from prior year non-core disposals).

Adjusted EBITDA of US$9m (NZ IFRS basis less the favourable fair value of biological assets effect) was substantially retained as the overall working capital outflow was only just over US$1m. Within this, we note that seed investment (wip and finished goods) – most probably the net result of increased seed orchard capacity constrained partly by adverse weather – rose but cash absorption here was effectively matched by an uplift in payables. Net interest payments at just above US$2m were in line with the P&L and no corporation tax was paid in the year.

Under NZ IFRS, R&D is capitalised and expenditure here approached US$5m in the year overall. Other capex of US$2.7m was in line with depreciation at the gross level, prior to the receipt of US$0.8m disposal proceeds. There was a small bias towards higher capex (tangible and intangible) in H2.

Cash flow outlook: With a significant revenue uplift modelled for FY20 (part rebound from a weather-affected FY19, part new nursery and supply agreements), we have assumed that increased working capital investment is required to support it. In broad terms, this is financed by increased profitability and, before c US$1m exceptional cash outflows (relating to final US retention payments), we again expect to see a neutral free cash flow outturn in FY20. Thereafter, we expect that higher levels of profitability will fund working capital and maintained capex spend and generate positive free cash flow, which will enable net debt to trend down to c US$15m (or 0.9x EBITDA) by the end of FY22 versus c US$22m (2.2x EBITDA) at the end of March 2019.

In the four countries in which Rubicon operates, real GDP estimates are expected to exceed 2% in each of the next two years in all but one of them, the exception being Brazil (in 2019 only). The core industries that ArborGen serves (including housebuilding, pulp and paper) are considered to be cyclical, so we take this data to be broadly illustrative of macroeconomic growth.

As mentioned above, the need for forestry owners to replace damaged treestocks and the new US nursery facilities feeding in during FY20 onwards underpin expectations of progress there. Additionally, in New Zealand, 2019 is the first year of the 10-year Crown Forestry tree planting programme and Rubicon has secured a supply contract that will have a material impact on local seedling sales volumes and profitability. The primary changes to our estimates relate to the mix in the US (lower proportion of MCP seedlings following seed orchard damage, on broadly unchanged volumes) and we have also lowered volume expectations in Brazil to reflect a lower base year and recent eucalyptus demand. As a result, our group revenue estimates have been trimmed modestly for the next two years, but a 400–500bp gross margin reduction flows through to materially lower PBT projections. Nevertheless, expected growth from FY19 is significant and this continues into FY22, which we publish for the first time. Over this three-year period, our EPS CAGR exceeds 30%.

We do not currently expect that Rubicon will make dividend payments over our estimate horizon.

Updating our DCF analysis following the latest set of results, using the same methodology as in our September initiation note, the core assumption drivers are:3

Using the above inputs generates a revised gross equity value of US$225m compared with US$240m previously. This variance is due to a slightly lower profit trajectory by 2029 and a higher net debt position. At the per share level, this value equates to NZ$0.71 with a small benefit here from a relatively firm US dollar translation effect. Investors should also be aware of the following adjustments to this headline valuation:

Since we initiated coverage, the US risk-free rates and market risk premium have reduced. For illustration, if we used these inputs it would generate a company WACC of 7.4% and a gross equity value of NZ$1.33 per share. Given that Rubicon has operations outside of the US, this WACC is probably too flattering, but it does serve to illustrate valuation upside in lower-risk/lower group cost of finance scenarios.

We have also updated our sensitivity table to demonstrate the effects of a range of US gross margin and average selling price (ASP) assumptions on the above valuation analysis, as shown in Exhibit 4. The range shown is slightly narrower than the previously presented version but broadly similar overall. (Note that our core analysis above generates ASP growth approaching 11% and a gross margin slightly over 51% (c 60bp and 200bp higher than before, respectively) on average for US operations beyond our explicit estimate years.)