Consider a 2sls model of the price of natural gas. Natural gas is subject to many factors that affect their prices on a day-to-day basis, and it appears that natural gas prices may be subject to political activity initiated by natural gas exporter. It would be useful to isolate the political effect here, since the underlying economic effects tell us what the unfettered economic price of the commodity should be. We might be able isolate the effect through 2sls. Review the 3 models below. USGDP is used to proxy the demand side of the equation. State the conditions that surround these instruments (their required statistical properties). You might suggest alternate possible instruments that might accomplish this. Are there better ways of approaching this issue than 2sls? If so, suggest them, if not, argue the case for the 2sls application.
1. PNGt = USGDPt + OUTPUTt + e, IV: UNdays
2. PNGt = USGDPt + OUTPUTt + e, IV: Transport
3. PNGt = USGDPt + OUTPUTt + e, IV: Tdeficit
where PNGt = price of natural gas in US$/1000 cubic feet end of month t
USGDPt = world gross domestic product in month t
OUTPUTt = output of natural gas in 1000 cubic feet in month t
UNdays = number of days the United Nations discusses Iran's nuclear program in month t
Transport = physical quantity of natural gas transported in month t
Tdeficit = US trade deficit in billions of US$ at end of month t