1. The Usual Suspects: A Primer on Investment Banks’ Recommendations and Emerging Markets Javier Santiso Chief Economist and Deputy Director OECD Development Centre Conference Opening and Innovation on Financial Emerging Markets Beijing - China March 28 2007 Sebastián Nieto Parra Chaire Finances Internationales Sciences Po Paris http://www.financesinternationales.sciences-po.fr/

2. 2 Objective of the paper I Overview of the literature II Description of the data III Investments bank’s business and research publications IV Emerging markets capital flows and research publications V Conclusions VI

3. 3 Two core questions  Do recommendations given by investment banks have an impact on the allocation of portfolio flows in the emerging markets? That reveals the influence of analysts’ recommendations on investors’ behaviour.  Above all, do recommendations are related with the business of investment banks? Information provided by banks to investors could be biased depending on their own objective that sometimes could differ from those of investors.

4. 4 Objective of the paper I Overview of the literature II Description of the data III Investments bank’s business and research publications IV Emerging markets capital flows and research publications V Conclusions VI

5. 5 Investment banks’ recommendations  The impact of investments banks’ recommendations on capital markets has been concentrated in OECD countries. -Womack (1996), Jackson (2005), Boni and Womack (2002), Barber et al (2001), Asquith et al (2005), … Variety of results: -analysts are confronted by a trade-off between sending true signals and optimistic signals. -Larger number of buy recommendations than sell recommendations. -Market reaction to upgrades is less pronounced than the market reaction to downgrades by analysts. -Impact of the measures introduced by the NYSE and NASDAQ, but also the sanctions established by the SEC in 2002.

6. 6 Investment banks’ recommendations  Research literature on emerging markets is scarce and concentrated in the equity market: -Seasholes (2000), Bae et al (2005): Accuracy of local vs foreign forecast analysis. -Bacmann and Bollinger (2001): Boom of the stocks covered by analysts between 1993 and 2000.  Empirical studies of the relationship between the recommendations and underwriters are scarce and concentrated to OECD countries: -Womack and Michaely (1999). Results suggest that there is a conflict of interest between investment banking and research department.

7. 7 Capital Flows to Emerging Countries  A large body has studied the determinants of capital flows: -“Push factors” or global factors literature: First half of the 90s Fernandez-Arias (1996) and Calvo et al (1993) -“Pull and Push factors” : Taylor and Sarno (1997), World Bank (2001), Alfaro et al (2005),… Most of the results conclude that local factors combined with external factors explain capital flows (FDI, foreign banks lending, bond and equity flows,…). -In addition to pull and push factors, recent empirical literature has studied the impact of information and distance on capital flows: Ghosh and Wolf (1999), Savastano (2000), Papaioannou (2004) and Portes and Rey (2005). In particular Portes and Rey (2005) develop an empirical model in which international information flows are a significant aspect to explain international equity flows.

8. 8 Our research  By using untapped and rich dataset, the purpose of this study: First, it is an attempt to analyse the determinants of the recommendations given by investment banks in the sovereign emerging bond market. Second, it allows to determine the impact of information on capital flows. For that we take into account investment banks’ recommendations as an additional factor to explain capital flows.

9. 9 Objective of the paper I Overview of the literature II Description of the data III Investments bank’s business and research publications IV Emerging markets capital flows and research publications V Conclusions VI

10. 10 Investment banks’ recommendations  Construction of a unique database containing the recommendations given by the major investment banks to the Latin American bond markets. -Direct and strict link between financial intermediaries and investors (not public information). -First publication that studies the impact that investment banks’ recommendations may have on Latin American Capital Markets.  Period: July 1997 - July 2006.  Frequency: Monthly Reports

11. 11 Investment banks’ recommendations  We have taken the recommendations given by 10 investment banks. All of them important players in the emerging bond markets.

12. 12 Investment banks’ recommendations  Main aspects concerning the recommendations given by investment banks: -Composed only by sovereign emerging debt. -We have classified three types of recommendations: Overweight (1), neutral (0) and underweight (-1). -These recommendations are assimilated to the cases of buying, maintaining and selling with respect to a portfolio (the index EMBI+ calculated by JP Morgan) Given portfolio restrictions a buying recommendation must be compensated by a selling advice.

13. 13 Investment banks’ recommendations  Example: Average of the recommendations given to Brazil by the investment banks with respect to the weight of Brazil in the EMBI Global index.

14. 14 Investment banks’ recommendations  We have taken 11 Latin American countries that represent nearly 95% of the GDP of the region. The total number of recommendations is over 3,000.

15. 15 Objective of the paper I Overview of the literature II Description of the data III Investments bank’s business and research publications IV Emerging markets capital flows and research publications V Conclusions VI

16. 16 Investment banks’ business  Banks are faced with a trade-off concerning recommendations: -While sell side business could have the incentive to build reputation by giving accurate information in the long term …. -…. in the short term recommendations could be biased in order to obtain short term profits. -Additionally, investment banking activities could be motivated to recommend optimistically the assets which they are participating as underwriters in an IPO.

17. 17 Underwriters’ recommendations  Recommendations given by banks that have been underwriters for Latin American sovereign bond issues. - 90% of the underwriters recommend to investors at the announcement date of the issue to buy or to maintain in their portfolio the bonds issued by the countries where they are acting as underwriters.

18. 18 Size of the market and recommendations Objective of the sell side business: to sell portfolios to a large variety of financial intermediaries. The percentage invested in these portfolios increases relative to the size of each country. High correlation between recommendations and size of the market: “too big to underweight” Credit risk is not a relevant variable to determine the recommendations

19. 19 Objective of the paper I Overview of the literature II Description of the data III Investments bank’s business and research publications IV Emerging markets capital flows and research publications V Conclusions VI

20. 20 Determinants of capital flows  In order to test the impact of recommendations on capital flows (Bond flows and Equity flows respectively), we have used the following two panel data regressions models: (i) (ii) where and : percentage allocated by funds in country i with respect to the total amount invested in emerging economies. : the average of the investment banks’ recommendations given to country i. : Pull variables defined by capital markets (exchange rate, spread of sovereign bonds and rate of return of equity).

21. 21 Determinants of capital flows : Pull variables that are strongly influenced by real sector (economic activity, inflation rate and interest rate). : country invariant variables which capture global factors (US nominal rates and US industrial production).  Period of the analyses: 1997-2005 for equity flows 2002-2006 for bond flows  Frequency: Monthly  Countries: Argentina, Brazil, Chile, Colombia, Mexico, Peru and Venezuela  OLS and FE estimation. Since OLS estimation are known to deal inadequately with time series and cross-section heterogeneity, we reported also Fixed Effects estimates (FEM estimators).

22. 22 Determinants of capital flows  In order to determine if a Random Effects Model (REM) was an adequate econometric model for this analysis we realised the Hausman Test. The null hypothesis underlying the Hausman Test (FEM and REM estimators do not differ substantially) was rejected.  In order to avoid problems of endogeneity between independent and dependent variables we have also taken into account the first lag of each of the explanatory variables in the regressions. In fact, by taking the lagged explanatory variable we could solve causality problems which are common to capital flows analysis.  We present only the results of FE estimators with lagged explanatory variables (see annex for the others results)

23. 23 Determinants of bond flows

24. 24 Determinants of equity flows

25. 25 Objective of the paper I Overview of the literature II Description of the data III Investments bank’s business and research publications IV Emerging markets capital flows and research publications V Conclusions VI

26. 26 Determinants of capital flows 3 conclusions concerning the determinants of capital flows: 1.The impact of investment banks’ recommendations on capital flows is positive and significant. 2.The impact of the recommendations given to external public debt goes beyond sovereign bond flows. Indeed, although their influence is minor, these recommendations also affect private equity flows. 3. This new microeconomic variable improves the fit of capital flows regressions more than some traditional macroeconomic variables such as interest rates, economic growth and inflation rate.

27. 27 Business and Investment banks’ recommendations We can not reject the hypothesis that the information transmitted to investors could be biased with the purpose to obtain short term profits and to recommend optimistically the assets which banks are underwriters in an IPO.  Further research: The results are preliminary. We have in part neglected the role of the recommendations in the sell side long term business. Indeed, further research must be done concerning the performance of these recommendations in terms of investment value and to contrast them with the underwriting activity.

28. 28 Policy Lessons 1.There is a need for more detailed information disclosure by investment banks in order to determine if past recommendations are related to macroeconomic variables and financial variables or whether they are associated with their business in emerging economies. 2.Given that banks’ recommendations and portfolio flows are related, an international co-operation must be established in order to encourage investment banks to cover more countries.

29. 29 ANNEXES

30. 30 Investment banks’ recommendations  An example of the recommendation given by one of the most important actor in the Latin American Bond Market.

31. 31 Investment banks’ recommendations  An example of the weight of some Latin American countries in the EMBI index (depends on the amount outstanding of sovereign debt).

32. 32 Underwriters’ recommendations  415 underwriters or Lead managers and 215 sovereign issues. Almost 75% of the underwriters are located in Brazil, Argentina, Colombia and Mexico.

33. 33 Underwriters’ recommendations: Argentinean case  The Argentinean case is very useful, interesting and special case -67 per cent of the recommendations were to maintain the positions in Argentinean External Debt (prior 2001) -Some of the comments given by banks months before the crisis were unrealistic: Morgan Stanley: “We are maintaining our Market Perform recommendation on Argentine bonds….Relaxation of fiscal targets and an innovative IMF- led financial package from creditors both improve Argentina’s credit outlook. Argentina needs to raise an estimated $2.6 billion to fulfil its first quarter financing requirements. New issues are expected to total $5.6 billion in 2001. Growth and fiscal performance are becoming the focus of investors’ attention.” January 26, 2001. Salomon Smith Barney (Citigroup): “The successful implementation of the IMF support package — with the associated debt management transactions — and the change in the global outlook probably increases the chances that economic activity will pick up in the second half of the year. We therefore recommend a neutral position in external bonds and local currency instruments.” January 17, 2001

34. 34 Underwriters’ recommendations: Argentinean case Source: Nieto Parra (2006) from Bloomberg

35. 35 Underwriters’ recommendations  Underwriter’s recommendations vs. Recommendations given by other investment banks - 75% of the Lead mangers’ advice was higher than or equal to that made by other investment banks.

36. 36 Underwriters’ recommendations  Underwriter’s recommendations vs. Recommendations given by other investment banks during the announcement date of the issue of a bond. - On average underwriter’s recommendations are more favourable than no-underwriters’ recommendations

37. 37 Underwriters’ recommendations  What is the incentive that no-underwriters could have to give an equal or higher recommendation than underwriters? For most of the Latin American countries 90% of the issues were realised by 10 investment banks

38. 38 Underwriters’ recommendations  By calculating the HHI we obtained the same results: Theoretically, this market could be characterised by an imperfect competitive market in which underwriters are playing a repeated game. By taking investment banks’ recommendations as a marketing product, it is then advantageous to investment banks to recommend a country even if at that period they have not been underwriters.

39. 39 Determinants of bond flows ..

40. 40 Determinants of bond flows ..

41. 41 Determinants of bond flows ..

42. 42 Determinants of equity flows ..

43. 43 Determinants of equity flows ..

44. 44 Determinants of equity flows ..