My column two weeks ago induced an interesting response from Greg Becker in the letters pages. Business Day normally frowns upon too much banter in this manner but the editor has allowed me a hall pass to respond.

Becker raises a range of interesting medium and long-term fiscal risks which one might easily cop-out to say are beyond the forecast horizon — but they demand some more serious thought and throw up some important pointers.

Indeed markets are often far too focused on the “medium-term expenditure framework” three-year window period of the fiscal policy outlook and not at the longer term and often more interesting zone where demographics, the full effects of small, short-run changes in policy compound to and where growing guaranteed risks can crystallise.

Moreover, the National Treasury’s strategic wins in the short term on key issues (such as demanding a better designed National Health Insurance (NHI) and comprehensive social security system than are on offer by other parts of government) presumably play out and occur (in some shape or form even if not what are now contemplated).

The first thing to say is that the Treasury is running a relatively, though not excessively, conservative stance precisely because of the range of risks on or over the horizon.

One of the most interesting points raised is that the gold & foreign exchange contingency reserve account (GFECRA) can cause perverse incentives for the government to weaken the rand to generate more windfall gains from unrealised foreign exchange gains on reserves that would be transmitted back to the fiscus.

However, to weaken the currency like this — a government policymaker would have to make (knowingly) bad decisions for little gain. To try to get even bigger gains the policymaker would have to blow the currency up. This would then create all manner of inflation and bank run problems .

We must remember that the GFECRA reforms here are aligning the SA Reserve Bank/Treasury flows of unrealised forex gains on reserves to something mirroring international norms. You simply don’t see policymakers trying to make a quick buck this way elsewhere in the world.

The state-owned enterprises (SOE) risks that are feared are much less concerning we believe — excluding Eskom and Transnet for now. After these two there simply are relatively small SOEs who’s bailouts would be a small fraction of the size of Eskom’s and of expenditure.

For other guarantees we must consider a probability weighted view of them being called in the long term. The probability of, for instance, the government guarantee on GEPF being called is minimal if sound policy around them is conducted (and their trustees remain free to not invest in Pravin Gordhan’s disastrous pet projects such as SAA).

Similarly the chances of calling the huge guarantees on the Renewable Independent Power Producer Programme (REIPPP) is small especially when we consider an unbundled Eskom with a high credit quality National Transmission Company that commands the tariff and has full pass through of REIPPP cuts from the tariff.

Transnet, of course, is an elephant in the room but a combination of tight conditionality on reforms and also probably a moderate (versus Eskom size at least) bailouts of R15bn-R20bn a year is already in our forecast — even if it is not yet in the Treasury’s.

Eskom has a generation bad bank legacy debt problem after the end of the bailout from 2026/27. About R120bn of debt will remain on the balance sheet at this point with a dwindling generation asset base.

A proper balance sheet optimisation that borrows within the envelope of declining assets and their revenue streams and is open to restructuring this debt — as will be happening widely with carbon stranded assets globally in the coming decade — for unguaranteed debt — will surprise no-one offshore for sure — who are well used to such things and understood the risk when they first invested.

The issue of cross default is also raised. It’s a common misunderstanding that cross default provisions exist in debt. This is not the case. There are cross defaults only on guaranteed SOE debt of a particular entity if the government does not honour the guarantee of another SOE if called upon.

Finally, the issue of NHI is raised and the impact of this kind of policy and their unintended consequences. Here views are required on if the thing will happen or not — we do not think this NHI will ever happen and so it is not in our fiscal framework. However, an NHI that is properly constructed and does not impale the private sector will come about eventually though arguably too late.

This will be done in collaboration with the private sector and only as and when additional growth related revenue proceeds become available so we would view it as net neutral on the fiscus. There are of course wider and more pressing issues of spending on the social wage but the recent raise to the social relief distress grant that was possible because the department of social development has tightened access so much, is one example of how these things can happen in the current framework.

All these things can go wrong of course, either though a complete shift in Treasury leadership under a particularly dramatic coalition (which is not baseline), or through errors (such as mistimed bailouts of SOEs), or through marked shifts in policy choices where the Treasury is overruled. Yet each of these seems hard to contemplate.

The far more important risk is the far more dull one. That each fiscal iteration is broadly “fine”, and a forthcoming fiscal rule mandates medium run lower debt levels. But there is always a tiny bit of slippage, interest rates remain that bit too high and growth that bit too low and so things ratchet — slowly — worse until eventually things fall over the cliff edge.

I am far more worried about this in the medium run — looking through the “positive” shorter term dynamic — than I am on any of the specific risks raised on the letters page.

• Peter Attard Montalto leads on political economy, markets and the just energy transition at Krutham, a SA research-led consulting company.