Running for the Exit International Banks and Crisis

Running for the Exit: International Banks and Crisis Transmission Comments by Vedran Šošić Financial Stability Department Croatian National Bank

Goal of the paper p Explores the role of cross-border lending in crisis transmission. p Observes the impact of different forms and intensities of financial integration on the stability of capital flows during the crisis: n How did the bank involvement with the country impact on change in it’s lending during the crisis.

Main findings and conclusions p It’s important to build a relationship. The extent of cooperation with local banks positively correlated with stability of lending. n Having a subsidiary also helps. n p IF YOU’RE GOING TO GLOBALIZE, DO IT TO THE FULL EXTENT!!! n More integration is better than less integration.

Overall paper assessment Topic hugely important in today’s world. p Product of a long lasting research effort in international banking. p Quite an enjoyable reading. p Impressive presentation: p n p Provides clear and concise description of research (and there is lot of it!!!). Paper seems like an experienced conference attendee: n Huge section on robustness issues.

Specific issues – data p Available data used close to full: n p However, DEALOGIC dataset not used to the full extent: n p Data aggregated by host countries. Authors express reservations regarding BIS dataset: n p Combines different data sources in a clever way. How well the two datasets correspond on a country-pair basis? Several data definitions used for robustness, but all treat changes in small and large exposures alike: n What would results look like if change in exposures is observed (as % of host country banking assets, GDP, or some other scaling variable)? Change in loan stocks rather than loan flows? What if change in flows is scaled?

Fundamental issues (I) Paper concentrates on cross-border lending as a major channel of influence in a globalized financial setting. p What about other possible channels of influence? p E. g. reputation risks. n Bad news about parent institutions may trigger run on their subsidiaries. n

Fundamental issues (II) p Consensus in macro-finance is getting more tolerant for some forms of capital controls: n p How to reconcile it with the findings of the paper? The paper assumes away most of the demand effects: Focus is on the determinants of supply. n What was the role of relationships in building macro imbalances which fed into recessions? n What if potential destabilizing effects of excessive capital inflows due to closer relationship for some countries dominates it’s stabilizing effect during the crisis? n

Fundamental issues (III) p Paper claims that information issues are at the center of the cross-border dynamics: Results do not follow thoroughly such approach. n Lending to subsidiaries did not act as a buffer – it was rather lending to the corporate sector! n Was is because subsidiaries were able to meet depressed loan demand, on because parents did not want to stand by their subsidiaries („functional distance“)? n

Fundamental issues (IV) p What about the credit risk? Does direct cross-border lending to the corporate sector magnify or diminish the frequency of default? n If probability of default for cross-border loans is higher than for domestic loans, switch to the corporate sector lending during the crisis may deteriorate overall loan quality. n

Some remaining questions p Characteristics of banks in host countries play no role in the model: Admittedly, this is somewhat intertwined with the demand effects, but not fully. n Do the efforts to make banking systems in host countries more resilient make sense from the standpoint of stabilizing capital inflows? n Or, would more resilient domestic banks make cross-border borrowing unnecessary when loan demand is depressed? n p How to facilitate integration of the corporate sector in international banking networks? n (know some ways, but would not unconditionally advise those)

A note on pricing of loans p Explanatory variables hardly explain any variation in loan pricing. n Probably the consequence of omitting host country specific variables – in this case country risk.

Some political economy p Was it advantageous for the home countries to restrict the volume of cross border lending? n p If it was advantageous, deeper integration may eventually backfire. Would geographical proximity in the future be more important than regulatory regime in home country? The main lesson of the paper may be “chose your parents wisely”. p But can policymakers really choose between the parents?

… and some suggestions for robustness check! p In analysis of financial stability (stress tests), it is customary to look at what’s beyond the 97 th percentile, not to abstract from it. n p Did any of the parent institutions cut loose some of their subsidiaries during the crisis? n p We may learn important things by looking at the worst case scenarios – just look at stress testing exercises! To what effect? What if some of parent banks collapses?

Thank you!