If a bank loans an individual money, say to buy a car, the terms are clear and legally enforceable - so much money per month over so many months, otherwise we repossess the car. The bank has legal recourse, meaning that they can pursue civil and perhaps criminal legal enforcement of their claims. Loans between sovereign states are often worded this way, but they don’t work that way.
Naive lenders will put a loan to a sovereign government on the books just like a business down the street: principal outstanding, total interest due, etc. This instrument is valued via a combination of factors, among them payment history, current global interest rates, global demand for credit, political initiatives within the country, and the vague idea of ‘global prosperity and security’. A loan to the UK is affected by an uprising in Syria and fracking in the US - Syria could involve a major expense if the UK becomes militarily involved, US fracking is reducing demand for petroleum, which is good for the UK economy, but bad for that part of the UK that extracts and sells oil. The naive lender believes that these things are the material determinates of repayment, although just this level of exploration suggests how volatile global events can be.
Banks that have been doing this successfully for a few hundred years (and there aren’t very many of them left) have a different perspective. The deal is not between the bank and the sovereign state, it’s between the bank and the wealthy interests within the country. Understood another way, I don’t loan money to a non-profit food bank, I loan money to a food bank run by friends that I know and understand. I don’t have to trust them, but I do have to know their culture. Therefore, the currently elected leader of country XYZ has personal assets in the $millions. The opposition leader and his up-and-coming replacement also have significant personal wealth, or the eventual capacity to obtain it. Therefore, the money I assign to the government is basically under their stewardship, and it’s their responsibility to see that it’s invested in ways that I approve of. Note that if my economic literacy is no better than theirs, I will never see the money again. This happens with clockwork regularity.
The US has, for example, about 140 million workers, of which about 125 million are currently working. The US is largely a service economy, therefore the workforce spends a lot of it’s time seated at desks pushing bytes around in computers. One might also find auto technicians, landscapers (yard care), teachers, and truck drivers. One finds that these workers earn on average about $45,000 per year, and the assets required to keep them equipped is about $20,000. $20,000 x 140 million workers is about $3 trillion.
In comparison, the US has about 60 million people over the age of 65, the age we describe as ‘retirement’. If one presumes that it is necessary to live on $2000 per month, or $24,000 per year, times the 15 years one lives between 65 and the life expectancy of 80. The amount of cash that needs to ‘saved’ to meet these retirement objectives is about $10 trillion (($24,000 x 60 million x 15 years) / 2, since the ‘average’ retiree is halfway through their retirement). Therefore retirees need $7 trillion more in assets than workers need in assets to keep the economy running.
The landscaper owns the truck and equipment himself - any loans are likely to be solely for the vehicle, at perhaps $10,000. The teacher is working in a classroom financed by taxpayer bonds - unless the school has been around long enough the bonds are paid off. In short, it’s possible that half the business capital in the US is in the form of loans or stock, at say $1.5 trillion. The US stock market is valued at roughly ten times that ($16 trillion), however most valuation has little to do with cash flow. Therefore, nearly all of the cash people have saved for retirement is ‘frustrated’ - it has no productive use. This means it gets parked in what are in effect non-interest bearing securities (T-Bills), cash accounts, overpriced stocks, or non-performing ‘real’ assets such as largely vacant commercial real estate. Thus if someone offers to pay a ‘real rate of return’ there are plenty of people that see the interest rate but pay no attention to the risks: their current situation is hardly any improvement.
Thus money is loaned to, in effect, charismatic politicians half a world away that are accustomed to, and skilled at, promising everything to everyone. Now that the money is in their hands, the question becomes ‘What incentive exists to keep them making their interest payments?’. Part of it is fresh loans - in that sense the entire scheme is Ponzi, pay off old creditors to suck in new ones. Another is actions on the global stage - in certain cases, for instance, the US wants an ‘international consensus’ as a justification for enforcing something in Iraq or Afghanistan. A country in debt is far more pliable toward such requests than one that isn’t. This only goes so far, if the local political will is dead set against it it won’t happen. If the country produces something the lender needs, such as oil, it’s possible to freeze overseas assets or exports, sometimes more for embarrassment value than anything else. Certain creditors confiscated Argentina’s cadet training ship in such an enforcement action.
When a country gets in serious trouble the outside creditors begin to second-guess the government, thus we find the IMF instructing Greece to implement ‘austerity’. In short, if the government is making promises it can’t afford, the outside creditors are the ones that direct the government to change course. Usually this involves massive civil disturbance - however at some point the citizens realize that they are living unsustainably. In theory (at least in democracies) this should result in a complete flush of the current governing elites, realistically it puts a new government in power, however one has noticed recently that such governments may only last a few weeks - or not form at all.
If someone puts $1000 into Government A in 1980 at 7%, then they make (in theory) $70 (or the interest on the remaining outstanding principal) in interest per year from 1980 until the debt is repaid, say 30 years later. $70 x 30 years is $2100 in interest alone. Thus from the investor perspective they may have ‘gotten their money back’ even if the principal was never repaid. The deal basically boils down to two factors: taxpayer tolerance (as the government pays debt via taxation), or ‘greater fool’ as new suckers replace old ones.
Germany and Japan have lots of money sitting around in banks, and export oriented economies. Therefore if the Germans want to sell a high-speed rail line to China or India they offer to finance it - thus in effect the loan is predicated on the revenue earned from the completed system. In the short term, the loaning country has ‘created jobs’ in it’s homeland, which may or may not positively affect overall income and tax revenues, and has a somewhat tenuous ‘IOU’ from the buyer. If the project goes well and the countries get along, the loan gets repaid - if no one takes the train and the countries get in a spat it’s all gone. Generally by that time the loaning country has moved on to other issues, and considers the fallout to be a minor distraction.
In short, there are a lot of ‘ifs, and’s, but’s’, and other nuances. This is almost always wrapped up in other international relations and geopolitics. Thus it is simply one piece of a hugely complicated chessboard.