Physical climate change and the sovereign risk of emerging economies

Journal of Economic Structures

The Official Journal of the Pan-Pacific Association of Input-Output Studies (PAPAIOS)

Table 7 Channels of temperature–sovereign risk connection: general warmness

	(1)	(2)
	\(\Delta\)EMBI	\(\Delta\)EMBI
HistoricalTempAnomaly	0.0596 (0.0834)	0.0825 (0.117)
VeryColdCountry (percentile) \(\times\) HistoricalTempAnomaly	− 0.0731 (0.0802)
ColdCountry (percentile; base category) \(\times\) HistoricalTempAnomaly	0 (0)
MildCountry (percentile) \(\times\) HistoricalTempAnomaly	− 0.0639 (0.114)
WarmCountry (percentile) \(\times\) HistoricalTempAnomaly	− 0.224 (0.174)
VeryWarmCountry (percentile) \(\times\) HistoricalTempAnomaly	− 0.491** (0.233)
VeryColdCountry (\(\le\) 10 \(^{\circ }\)C) \(\times\) HistoricalTempAnomaly		− 0.0492 (0.110)
ColdCountry (\(> 10\) & \(\le\) 15 \(^{\circ }\)C; base category) \(\times\) HistoricalTempAnomaly		0 (0)
MildCountry (\(> 15\) & \(\le\) 20 \(^{\circ }\)C) \(\times\) HistoricalTempAnomaly		− 0.207 (0.124)
WarmCountry (\(> 20\) & \(\le\) 25 \(^{\circ }\)C) \(\times\) HistoricalTempAnomaly		− 0.125 (0.155)
VeryWarmCountry (\(> 25\) \(^{\circ }\)C) \(\times\) HistoricalTempAnomaly		− 0.547** (0.229)
Precipitation	0.0640 (0.406)	0.00819 (0.393)
Observations	9957	9957
R-squared	0.524	0.524
Country FE	Yes	Yes
Region\(\times\)MonthYear FE	Yes	Yes
Total “very warm” country effect	− 0.432	− 0.464

This table shows OLS estimation results of a panel of 54 countries from 1994m1 to 2018m12. \(\Delta\)EMBI are monthly natural log returns of a country’s EMBI index. HistoricalTempAnomaly is the difference between monthly temperature of a country and its 1901–1950 temperature average of the same month. Each country is grouped into a bin either according to percentiles or 5 \(^{\circ }\)C intervals (see Table 5 for country classification). One bin is omitted due to multicollinearity (base category). Single terms of the bins are subsumed by time fixed effects. Total “very warm” country effect is the sum of the VeryWarmCountry interaction and the single term of HistoricalTempAnomaly. Standard errors (in parentheses) are clustered at the country level and control for heteroscedasticity and serial correlation, ***, ** and * indicate statistical significance at the 1%, 5% and 10% level, respectively. See Table 18 for variable definitions and sources