The Log of Gravity at 15

Silva, João Santos; Tenreyro, Silvana

doi:10.1007/s10258-021-00203-w

Cited by 46 publications

(25 citation statements)

References 76 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the PPML, possible biases due to the amount of zero observations in the dependent variable are corrected while also accounting for heteroscedasticity. Further, Monte‐Carlo simulations show that the estimator performs well in spite a large proportion of zeros (also see Yotov et al, 2016, p. 20) and the validity of the estimator does not depend on very strong assumptions of the distribution of the data as for example would be the case for a zero‐inflated model (Silva & Tenreyro, 2021). Overall, the PPML estimator is widely accepted to be very well suited for gravity estimations.…”

Section: Methodology and Empirical Modelingmentioning

confidence: 96%

Digital gravity? Firm birth and relocation patterns of young digital firms in Germany

Hellwig

2022

Journal of Regional Science

View full text Add to dashboard Cite

This paper analyses the spatial patterns of young (<10 years) digital firms in Germany between 2008 and 2017 on county level. Determinants of firm birth locations as well as relocations are considered jointly to understand differences in location choices within firms' life cycles. I match commercial register data of 107,321 firms with countylevel administrative data to capture local characteristics. Using an OLS model with fixed effects, I find that the local knowledge base-that is, universities, research institutes, and colocated incumbents-are significant key determinants of digital firm birth when controlling for a host of local characteristics. My results indicate that for five firms per 1000 inhabitants, there is around one firm birth.Second, using a fixed effects gravity model for the analysis of relocations, I find that the most dominant explanatory factor for firm relocation across specifications is distance, that is, relocation costs. Relocation flows are more than twice as high to neighboring counties relative to other locations which shows that digital firms are not as footloose as their business model may suggest. Jointly, my results reflect economic activity's regional persistence, particularly for new firms. My paper provides evidence for policies targeting homogenous digital clusters based on strong colocation and that digital economic activity is not

show abstract

Section: Methodology and Empirical Modelingmentioning

confidence: 96%

Digital gravity? Firm birth and relocation patterns of young digital firms in Germany

Hellwig

2022

Journal of Regional Science

View full text Add to dashboard Cite

show abstract

“…The regressions are estimated using the Poisson Pseudo‐Maximum Likelihood (PPML) estimator developed by Gourieroux et al (1984a) and popularised by Santos Silva and Tenreyro (2006, 2011). This estimation method is preferred by Blackburn (2007), Fally (2015), Santos Silva and Tenreyro (2022) and Kren and Lawless (2022) among others to overcome the potential bias related to heterogeneity arising from log‐linearizing (Mayer et al, 2019) as well to handle the problem of zeros. In PPML regression, only a correct specification of the conditional mean of the dependent variable is required for consistency of the regression estimator (Gourieroux et al, 1984b), and hence can be applied to dependent variables with nonnegative values without the need to explicitly specify a distribution for the dependent variable (Correia et al, 2020).…”

Section: Methodsmentioning

confidence: 99%

How did Brexit impact EU trade? Evidence from real data

Buigut

Kapar

2023

World Economy

View full text Add to dashboard Cite

At the time it voted to exit in 2016, the UK was a leading economy within the EU. It contributed about 16 percent of the EU GDP, while the other EU countries accounted for almost half of UK's total trade. This study attempts to answer two research questions: First, how Brexit affected the EU-UK trade and second, how it affected the trade between remaining 27 EU members. To answer these questions, quarterly data are exploited for the period from 2005Q1 to 2022Q3 covering a total of 53 trading partners including the EU members. A gravity model that controls for unobserved bilateral heterogeneity and multilateral resistance is estimated by PPML. Three phases of Brexit (the referendum, transition, and post transition [under the TCA]) are analysed. The results indicate that the Brexit referendum phase depressed UK-EU trade by around 10.5%, and transition phase by around 15%. In both cases,particularly for the transition phase, the effect is greater on the UK imports from EU than the UK exports to EU.We do not find a significant effect due to the post transition (TCA) phase. Estimates show some mild but positive effect on intra-EU trade of about 1.5% and 4.6% due to Brexit referendum and post Brexit respectively, but no significant effect from the transition phase. This suggests that some EU trade with the UK was redirected to other EU members. Hence UK should aggressively seek out new trade agreements with other countries and trade blocs as

show abstract

“…One cannot tell whether the sample selection depends on the covariates, but the sensitivity of the sample selection problem can be tested by treating the missing observations as zeros in our gravity equation. Theoretically, the disparity should be limited because the observations where the conditional mean is close to zero have low variance and therefore have a minimal contribution to the value of the pseudo log‐likelihood function (Silva & Tenreyro, 2021).…”

Section: Methodsmentioning

confidence: 99%

Effects of international trade on income revisited

2023

Rev International Economics

View full text Add to dashboard Cite

This paper revisits the causal impact of international trade on income using a time‐variant instrument in a panel of counties between 1950 and 1995. In particular, in constructing the time‐varying instrument, it estimates the gravity model using the Poisson pseudo maximum likelihood (PPML) estimator. Trade has a significantly positive impact on income over the 1950–1970 period ‐the income elasticity of trade is between 0.35 and 0.50. However, air distance has not had any increasing effects on international trade since the 1970s. Therefore, the imputed trade from 1970 to 1995, used as an instrument, has little variance and should not have driven the change in income.

show abstract

The Log of Gravity at 15

Cited by 46 publications

References 76 publications

Digital gravity? Firm birth and relocation patterns of young digital firms in Germany

Digital gravity? Firm birth and relocation patterns of young digital firms in Germany

How did Brexit impact EU trade? Evidence from real data

Effects of international trade on income revisited

Contact Info

Product

Resources

About