The UNCTAD has a new book on development strategies (free pdf available). There is a section of fiscal space, which is an idea that is discussed quite a lot these days. Here is a paragraph from that section:
Fiscal space is an essential aspect of the policy space needed by the developmental State (see TDR 2014, chapter VII). Even if governments are allowed to conduct some development policies within the multilateral, regional or bilateral frameworks, they still need to finance them. To that end, strengthening public domestic revenues is key, given that they are more sustainable in the long run than relying on aid or debt, as well as being less subject to restrictions and conditions that hamper policy space.
I think that the crucial thing here is to recognise that even thought government in developing economies can create unlimited amounts of currency, the weakness of the “public domestic revenues” (=taxes) might lead to the problem that every “Peso Queso” created and spent at the margin leads to households converting them at the earliest instance into hard currencies, like USD, CHF, EUR and so on. That would mean that the exchange rate would drop as a result of increased government spending. This also what Warren Mosler, one of the founder of Modern Monetary Theory (MMT), says in the context of the Russian default of 1999:
All throughout this process, the Russian Government had the ABILITY to pay in rubles. However, due to its choice of fixing the exchange rate at level above ‘market levels’ it was not, in mid August, WILLING to make payments in rubles. In fact, even after floating the ruble, when payment could have been made without losing reserves, the Russian Government, which included the Treasury and Central Bank, continued to be UNWILLING to make payments in rubles when due, both domestically and internationally. It defaulted on ruble payment BY CHOICE, as it always possessed the ABILITY to pay simply by crediting the appropriate accounts with rubles at the Central Bank.
Why Russia made this choice is the subject of much debate. However, there is no debate over the fact that Russia had the ABILITY to meet its notional ruble obligations but was UNWILLING to pay and instead CHOSE to default.
I absolutely agree with Mosler. Now the tricky question for developing countries is: if because of free capital flows and speculators driving the exchange rate up and down the volatility of the exchange rate is too high to attract serious foreign (direct) investment, what can you do? You can protect yourself by piling up USD reserves at the central bank. If in crisis, you can defend your exchange rate from falling “too much” (“undershooting”). In a boom, you can use your own currency to buy foreign currency and hence appreciate your own anyways.
To conclude: developing countries should not have over-valued exchange rates, since a correction might lead to a steep fall in the exchange rate and create lots of inflation (via more expensive foreign goods). This reduces real purchasing power and is surely recessionary. If, in such a situation, government uses fiscal policy it might only deepen the problem. The private sector, once it gets an additional “peso queso”, instantly sells it for USD in the expectation of more depreciation to come. And come it will – this is a self-fulfilling prophecy! Government should then increase government spending and taxes in lockstep, which would mean that the fiscal multiplier would be very low. Perhaps a suspension of free capital movement would be the only solution if things are really, really bad. This is not something that never happened to developing economies.
Marc Lavoie in his new “Post-Keynesian Economics – New Foundations” has a table (1.4) with fallacies of composition: when all actors do the same thing, they defeat themselves. Perhaps another fallacy should be added?