Mixed omens for Oracle Power

While the chart shows that Oracle’s shares might, before Covid-19, have been predicting riches, it’s probably an illusion, writes John Cornford.

In 2015, I first highlighted the four coal-to-power companies on AIM hoping to build power stations at their coal assets in Africa and Asia, to supply the dire need for reliable electricity. While tiny companies, their projects were very large, slated to cost up to $1bn and beyond, so some of that value, surely, would rub off on them?

With only the companies’ financial structures (coal assets, shares in issue, the funding required and estimated project profits), I arrived at a crude order of attractiveness which has always remained the same. Ncondezi Power (LON:NCCL) in Mozambique at the top and Kibo Energy (LON:KIBO) in Tanzania (and subsequently also in Mozambique and Botswana) consistently at the bottom. Edenville (LON:EDL) and Oracle Power (LON:ORCP) were in the middle.

Now, with at least two projects closer to fruition and more detailed information available from various sources (but not from the companies), it’s possible, after a lot of maths, to estimate the cash generation profile of a typical project. As the plants are almost ‘off-the-shelf’ from two Chinese builders, the estimates can be fairly reliable.

But it is their benefit to shareholders in four ‘sponsor’ companies, with different share structures and balance sheets, that is more difficult to assess. Kibo, having stuck its fingers into many pies with no clear timescales has issued too many shares already. It has therefore remained the worst choice, while Ncondezi (closer to commissioning with far fewer shares in issue) remains the best.

However, there has been no reliable ‘research’ from any broker putting all this together. As a result, investors rely on the bulletin boards. While these attract informed comment on some companies, they are often, especially for AIM, dominated by amazingly unrealistic expectations.

Inappropriate use of net present values (NPVs)* grotesquely exaggerates the value to shareholders, especially for major projects where loan repayments take away more than half the income over the first half of their lives − something NPVs don’t show.

Also, some inexperienced investors seem to think that their company’s project and its profits will all belong to them − whereas it will only belong, when built, in proportion to whoever puts up the very large amount of capital needed and which the inexperienced investors couldn’t possibly afford.

In fact, the only way to forecast the value their shares might have when operating (this is the way professional investors and banks will do it) is to calculate the year-by-year cash flows and how many shares it will be divided among. While even that isn’t the full story, any other ‘method’ will be hopelessly wrong.

So, what about Oracle after its spectacular December spike?

The company’s project for a 660MW station, on its block VI concession on Pakistan’s vast Thar Coalfield, has been in the making for over 12 years. However, it has suffered from the usual developing-country political delays as well as a change in the Chinese partners who will build and put up most of the cost under the auspices of The China Pakistan Economic Corridor. So, to get back on track, Oracle throughout 2019, backed by its broker Brandon Hill, radically changed its management and appointed an experienced Pakistan-based chief executive, Naheed Memon.

Brandon Hill supported the company’s planning work with £2m of equity funding during 2018-19 (which more than doubled issued shares and forced them steadily lower) plus a further £450,000 in loans until, in late November, a further £700,000 raise was announced at 0.25p (an almost all-time low). This included £500,000 from Sheikh al Maktoum of Dubai, who runs large natural-resource investments throughout the Gulf and, if warrants he was granted are exercised, would end up with nearly 20% of the company (Brandon Hill having halved the 22% holding it built up in 2018/19).

That prompted the shares to partly recover their previous two years’ fall, but it was not until a week later in December that they suddenly jumped to a near fivefold rise over the week − solely caused by speculation on the bulletin boards, which seemed to assume that the news meant sudden riches for shareholders, but with no attempt to work out exactly what.

Partly, that may have been because in around 2014, when issued shares were only about 1/7th of their number now (or 1/10th if warrants are exercised), the company’s then broker had estimated a price at which shares could be issued to fund Oracle’s share of its project, of 10p − a figure being bandied about even today. Obviously, it should now be only 1p, all other things being equal.

But the project is now much larger − the new chief executive recently announced that instead of in two stages of 660MW, it will comprise the full 1,320MW from the start. While a development agreement has been signed with the two Chinese builders who have agreed to fund 88% of the 25% equity component of the full cost (leaving Oracle with the ‘right’ to subscribe for 12% of that 25%) and to find 100% of the 75% debt component, there has been no clear information about its cost.

The chief executive was still recently quoting the $1.6bn slated four years ago for the 660MW mine and power station combined, although an almost exactly similar scheme on the adjacent Thar II block was commissioned last June at a reported cost of $1.9bn, and Thar VI will start at least five years after Thar II.

So the cost for the larger project must surely be $3-3.5bn, which means that if Oracle wants to take up its 12% equity share, its shareholders will have to find more than $90m (£70m) compared with its current £12m market cap. While it will be credited with planning work done, it totals less than £5m so far.

While that £65m might be raised in stages, it doesn’t include the probably substantial funding for another recent initiative which might have tempted in the Sheikh − to exploit Oracle’s coal to produce fertilisers and gas/oil.

So, what does this mean for Oracle’s future share price? Without more information an answer has to be approximate. But from the economic studies available for other CTP (coal to power) projects, annual revenue from the 1,320MW power station should be around $1.4bn, with operating profit at the project level around $750m.

Shareholders won’t see that, however, because in the first 12 years, around $380m per annum will be used to repay the $2.6bn bank loan (if at 10% interest and 7% repayments, this would be $327m). That will meet 75% of the capital cost, while around $175m per annum will be paid in tax.

An NPV calculation won’t allow for this, so, depending on the loan interest, the profit available to shareholders would be between $195m and $248m, with Oracle’s 12% anything between $23m and $30m.

But Oracle won’t actually receive that share. As a minority shareholder, it will depend on the Chinese majority allowing a dividend, from cash they might prefer to keep to fund the fertiliser project or any problems. So Oracle’s annual cash income might be only $12-$15m (£10-12m). To obtain this it will have had to stump up at least £65m, which if raised at 1p per share would add 6.5bn shares to the 2.6bn shares in issue now if all warrants are exercised, meaning cash earnings per Oracle share would be around 0.13p.

And it is that cash inflow per share (not expected before at least five years from now) that will determine the price at which Oracle can raise its £65m at project financial close (perhaps in 2021).

As a limited life, non-growth but cash generating investment, Oracle will be rated (just on the power plant − it’s anyone’s guess how the fertiliser project might complicate matters) on a yield basis, like Pakistan’s largest energy company, Engro, whose 11% yield in Karachi is unlikely to be bettered by Oracle in London.

So, depending what Oracle might pay out to its own shareholders, that puts a lid on the price at which it can get away its cash raise, and therefore the value of the company at that point. This, I’m afraid, will be far lower than the bulletin boarders seem to expect.

At a 1p issue price, we arrived at issued shares (and warrants) of 9bn, producing cash earnings of 0.13p. Even if all paid in dividend, an 11% yield would attract a share price of 1.2p.

At a 2p issue price, total shares would be 6bn instead of 9.6bn, and cash earnings per share would be 0.2p, so that on an 11% yield the share price would be 1.9p, which fits better.

So the conclusion is that at financial close in, say, a year’s time, Oracle might not justify raising that £65m at higher than 2p – assuming no other shares are issued for the fertiliser project and to meet company costs for the next five years. But, just as they did for Sirius Minerals (and Kibo), brokers pushing the shares will no doubt use the misleadingly high NPV basis to justify a higher ‘target’ − so what could it be?

Based on what has been published and on the more accurate cash-flow profiles calculated before, Oracle’s 1,320MW project costing $3.3bn would show the 8% NPV for the whole 100% project to be about $1.7bn, and therefore the value from when income starts and capex is spent would be about $5bn.

However, that ignores the benefit to shareholders of gearing up their 25% project share through 75% loans, but ignores the cost to them of repaying those loans. Adjusting for those gives very different NPVs for the 25% shareholders than for the project as a whole – a fact missed by brokers (or inexperienced investors) who go by NPVs.

Combining the two factors leads to an 8% NPV for the 25% shareholders from revenue start, amounting to $4.3bn for a $825m cost, of which Oracle’s share is $516m, and if analysts divided that by 6bn shares in issue they would arrive at 8.6p per share. If the cost had been raised at 1p per share, they would arrive at 5.5p.

And that demonstrates just how misleading any valuation based on an NPV can be, and why in 15 years researching most junior mining companies, I’ve never found a share to even remotely approach brokers’ NPV based ‘target’ prices. And that is also because, as I’ve explained, anyone would be a fool to pay up front for 25 years of future cash flow (which is what paying the NPV entails), from even a reliable project.

This all means that Oracle’s shares don’t deserve more than a 1p price right now, although that might change. For Sheikh Mahtoum, who doesn’t have to borrow to raise funds, the returns from investing in a power project can be much better than available to stock-exchange investors, who have to pay loan interest and satisfy institutional investors’ demand for an extra-high risk return. This probably explains his interest and that he might, at some stage, offer to buy out the whole power project.

And Oracle might (and should) restructure itself to shelter its power-project shareholders from the costs and share dilution that the separate coal-to-fertiliser project will involve, by hiving off into separate companies.

But any buyout from the Sheikh might not be for much more than Oracle will have spent. It would be more if he accepts a lower return, but the ‘bulletin-board’ hopefuls should ignore predictions of vast riches. They probably won’t, but sensible investors will know the danger of chasing the shares.

*A note on NPVs

The formula for an NPV is – NPV=PV-Capex.

In more detail, it is the project income once built (PV), less the cost of building. But because costs and incomes are spread over time, each period is discounted using a discount rate (in our examples at 8% per year) and each period is added up to get the lifetime NPV.

The NPV varies with the discount rate (so brokers can choose a convenient low one to inflate it) and is the average interest rate that an investor would get over and above the return of his upfront investment over its lifetime (ie just like an annuity). So that for a risky long-term investment, a sensible investor would demand significantly more than 8%. In other words, even the NPVs I’ve calculated look much too high.

N.B. This article was prepared before the company’s Q&A session April 7^th, and nothing said there has changed the author’s view.