How does state borrowing enter into this argument of money created and destroyed?

Reasonable people can disagree on that.

One view is that sovereign bonds are a way to take money out of circulation. In this view, issuing sovereign bonds is similar to instating a tax, and repaying sovereign bonds is similar to abolishing a tax.

Another view (and the one I subscribe to) is that sovereign bonds are a form of money - just a less liquid one. So when the sovereign deficit spends, it is printing money whether it "funds" its spending with sovereign bonds or by simply crediting the accounts of the recipients with legal tender. The difference is only in the liquidity (or lack thereof) of the new money, with legal tender being the most liquid and bonds being progressively less liquid the greater their time to maturity.

Which of those two you adopt is a matter of intellectual convenience - they are indistinguishable for the purpose of macroeconomic planning.

How is the argument affected if the sovereign can't print money (i.e. in eurozone countries)?

Well, the argument hangs on the sovereign's ability to enforce legal tender laws. So as long as the French sovereign can enforce French law in defiance of the Bundesbank, it doesn't change... in principle. Restrictions such as the Maastricht rules are voluntary restrictions on the behaviour of the sovereign, not fundamental facts of economic life.

In practise, this pressure towards deficit errorism is a major problem with the € as currently constructed. But the € as currently constructed is not sustainable on the time scales involved in these population projections - for precisely that reason. So the € will either change to accommodate a pan-European economic policy (in which case macroeconomic stabilisation will become a federal task), or the € will break down and we will return to floating sovereign currencies.

How is it affected if the sovereign has pretty much all of its debt in its own currency (i.e. the US)?

It isn't. There are many possible complications to having debt in other people's currency, but those really aren't in the realm of pensions and labour market regulation - which are almost always exclusively accomplished with domestic legal tender.

This means that various announced projected "costs of pensions by the year 20XX" measured as percentage of projected GDP, are questionable?

Yes and no. It depends on what they mean by that.

It is, obviously, possible to input some assumptions about GDP growth, pension systems and demographics, and get a scenario for the share of pensions in GDP in 20XX. What is dishonest is to discount sovereign outlays and taxes in order to obtain a present value - sovereign budgets just don't work that way.

Now, if the taxpayers of 20XX believe that they are paying too generous pensions, then that is a wholly legitimate political position... in 20XX. But demographic trends are slow-moving, so there will be plenty of time for the people who are actually alive and paying taxes in 20XX to reduce pension benefits if they find their taxes onerous. Arguing that we need to make this decision for them is to presume to speak for the as yet unborn and uncontemplated.

How is this argument affected by demographic decline?

It isn't. See bullet immediately above.

A note: I looked up the numbers a few years ago on Greece (below population replacement rate since 1980) and found that in the mid 2000s the percentage of working age population employed was at an all time high and dependence ratios at an all time low. This was due to a steady increase in female participation in the workforce, of course, but also due to the fact that the under 18s shrank almost as fast as the over 65s. This, I would wager, is pretty much par for "infertile" countries.

That is my experience from looking at the data as well.

Thus the dependence ratio doesn't seem to be too much affected in projections even of the most demographically declining countries - and thus dependents' costs unless old people are much more costly to society than children,

Yes and no. Most projections do show an increase in dependence ratios, albeit a modest one.

[That old people are more expensive than children] is indeed probable.

I wouldn't know about that. I'd have to pull out some consolidated public sector budgets to be sure, but for Denmark I'm pretty sure that they cost the same, or close enough as makes no matter.

One could offer a counterargument to

it makes no difference to future taxpayers what this generation does with retirement age

that any changes in retirement age necessarily unfold over a long period (you can't make someone planning to retire next year, retire in 5 years time and any cuts in pensions should be rolled out over many years so as to not influence life plans and expectations too much).

One could make that argument, but one would have to make the case that the funding problems are going to show up faster than the possible rate of adjustment. Which just isn't the case for any of the major [1] EU member states except Germany. And Germany starts out with large current account and trade surpluses against the rest of the Union, and mercantilist inflation rates; they could print the money to fund their retirement obligations without corresponding tax increases and still be a net exporter and on the low end of Eurozone inflation rates.

Generally as per the "slash pensions argument": aren't, pretty much linear, projections of current trends to 50 years' time some sort of quackery?

Yes, that would be rank quackery. Fortunately, population projections aren't simply linear projections. All reasonably literate societies have very good fertility statistics, and societies with (near-)universal health care have uniformly excellent fertility statistics. We know how many children people have, we know when they have them, we know what the age and gender distribution is today and we know how many people immigrate and emigrate.

Of course you have to make assumptions about how those figures are or are not going to change over the course of your projection. But the assumptions that underpin the population projections of European states are usually not wholly unreasonable.

Especially in a time of unfolding multiple global crises and geopolitical shifts? How is it even rational to project that far into the future any current trend and to implement policy now

The usual justification is that if we have a shock that's big enough to force us to discard our projections, things will be so fucked up that we can't plan for it anyway. So we make plans based on things not going catastrophically wrong - no serious wars, no exceedingly deadly pandemics, no agricultural collapse, etc., because it's silly to get caught with your pants down just because the sky didn't fall.

Your mileage may vary on whether that's sufficient reason or not. That's a political decision that reasonable people can disagree about.

- Jake

[1] Minor EU member states are not worth doing 20-year macroeconomic projections for - their economic well-being is more dependent on what the major EU members do than on any policy of their own. (This should not be construed as an excuse to be negligent in policy planning - just as a statement on the inherent uncertainty of long-term planning for states that are not even halfway sovereign.)

Friends come and go. Enemies accumulate.