They also subjectively ignored the post-WW2 period for most countries; coincidentally when countries had high debt and grew like crazy.
I bet that if you started with the desired results in mind for a study like this, you could create a list of variables to try and pick the combination that maximizes your metric.
Then provide a post-hoc justification for each choice.
For example:
* debt-to-gdp 80%, 85%, 90%, 95%
* Ignore years after ww2 (the war was an externality)
* Don't ignore years after ww2 (why would we? they're years.)
* Ignore years after ww2 for countries that got aid from US (removing an externality)
* Ignore years after ww2 for countries that did not get aid from US (most countries did, don't want to mix the sample)
* Ignore years after ww2 for countries that were in the European theater (removing an externality)
* Ignore years after ww2 for countries that were not in the European theater (not affected by war like rest of globe)
* Ignore years after any war (externality)
* Weigh by population of country at time of study (lazy way to account for size of industry)
* Weigh by population of country by year (better way to accounts for size of industry)
* Equal weight for all countries (easy to explain)
* For each country, take periods of high debt-to-gdp and compare against periods without high debt-to-gdp; then compare countries equally (easy to explain-ish)
* Take each period of consecutive years with a high debt-to-gdp ratio and call each period one sample point. Average these inequal-length points together and compare against all years of low debt-to-gdp (Come on! Nobody would believe-- oh no.)
They also subjectively ignored the post-WW2 period for most countries; coincidentally when countries had high debt and grew like crazy.
I bet that if you started with the desired results in mind for a study like this, you could create a list of variables to try and pick the combination that maximizes your metric.
Then provide a post-hoc justification for each choice.
For example: * debt-to-gdp 80%, 85%, 90%, 95%
* Ignore years after ww2 (the war was an externality)
* Don't ignore years after ww2 (why would we? they're years.)
* Ignore years after ww2 for countries that got aid from US (removing an externality)
* Ignore years after ww2 for countries that did not get aid from US (most countries did, don't want to mix the sample)
* Ignore years after ww2 for countries that were in the European theater (removing an externality)
* Ignore years after ww2 for countries that were not in the European theater (not affected by war like rest of globe)
* Ignore years after any war (externality)
* Weigh by population of country at time of study (lazy way to account for size of industry)
* Weigh by population of country by year (better way to accounts for size of industry)
* Equal weight for all countries (easy to explain)
* For each country, take periods of high debt-to-gdp and compare against periods without high debt-to-gdp; then compare countries equally (easy to explain-ish)
* Take each period of consecutive years with a high debt-to-gdp ratio and call each period one sample point. Average these inequal-length points together and compare against all years of low debt-to-gdp (Come on! Nobody would believe-- oh no.)