What are the problems to evaluate the performance of afps as assets managers? What do we know about their return? What problems have the papers done on this issue?


Download 461 b.
Sana22.03.2017
Hajmi461 b.





What are the problems to evaluate the performance of AFPs as assets managers? What do we know about their return? What problems have the papers done on this issue?

  • What are the problems to evaluate the performance of AFPs as assets managers? What do we know about their return? What problems have the papers done on this issue?

  • Conclusions about AFPs as assets managers and policy reccomendations



AFPs face regulations on their investment decisions (draconian regulations) that limit their performance as asset managers

  • AFPs face regulations on their investment decisions (draconian regulations) that limit their performance as asset managers

  • Srinivas & Yermo, 1999 for Argentina, Chile and Perú conclude that the costs of regulation have not been important.

  • However, Berstein & Chumacero, 2003 find that the Chilean regulations are costly (a 5% tax on the value of the fund). Pereda, 2007 for Perú, Carredo 2007 for México and Jara et al 2008 for Colombia also find that regulations are costly.

  • But AFPs can undo the regulation by reallocating assets inside each asset class. Chisari & Dal Bó, 1996 show some evidence for Argentina. However, size and deepness of local markets may limit some options. Regulations also constrain the ability of AFPs to replicate market indexes.



Tapia 2008 develops Markowitz efficient frontiers for a hypothetical portfolio and finds that regulations are costly in Chile and Mexico (not in Argentina) and in Hungary (not in Poland and Czeck Republic).

  • Tapia 2008 develops Markowitz efficient frontiers for a hypothetical portfolio and finds that regulations are costly in Chile and Mexico (not in Argentina) and in Hungary (not in Poland and Czeck Republic).

  • Auguste and Artana (2006) find for Chile and Peru that the limits on asset classes have not produced costs when the AFPs performance is compared with that of mutual funds. Most of previuos papers create ideal portfolios when it is better to compare with other actors in the same markets like mutual funds because there are liquidity constrains, low frequency bias, operating costs.



Problems with the data (besides the restrictions on data):

  • Problems with the data (besides the restrictions on data):

    • Not all countries ask AFPs to value their portfolios using market prices for all assets.
    • AFPs returns are gross of fees and gross or net of operating expenses depending on the country
    • Mutual funds: returns are net of both
    • Passive indexes: gross of both


How to evaluate performance

  • How to evaluate performance



Walker & Iglesias (2007) look at the excess return per unit of risk in relation to risk-free assets. They make a joint test of performance and regulations

  • Walker & Iglesias (2007) look at the excess return per unit of risk in relation to risk-free assets. They make a joint test of performance and regulations

    • For Latin-American countries AFPs have better returns than the short-term risk-free asset, but not for the long-term risk-free asset.
    • There is some evidence for Argentina, Chile, México and Perú that AFPs were able to produce positive Alphas (comparing with a weighted average portfolio of indexes, with the AFPs’ weights)
    • But they can not distinguish if this was because of better efficiency or because the regulations helped the AFPs to achieve those Alphas.
  • Tapia (2008) finds poorer performance in Argentina, Chile and México when the efficient portfolio is unrestricted



Option: Characteristic based studies. In the alternative portfolio the return of each sub asset class is replaced with the return of mutual funds specialized in the asset class (e.g. equities)

  • Option: Characteristic based studies. In the alternative portfolio the return of each sub asset class is replaced with the return of mutual funds specialized in the asset class (e.g. equities)

  • Advantages: comparison with mutuals that take financial decisions in the same market (i.e. no theoretical decision).

  • Limitations: AFPs’ future liabilities are different than Mutual funds’ and sometimes mutual funds are forced to maintain some liquidity.



Srinivas & Yermo (1999) replace the return on some asset classes with the returns obtained by mutual funds or with indexes. AFPs better performance adjusted by risk for several countries in the region and there is no evidence of herding.

  • Srinivas & Yermo (1999) replace the return on some asset classes with the returns obtained by mutual funds or with indexes. AFPs better performance adjusted by risk for several countries in the region and there is no evidence of herding.

    • But some problems. It is a joint test of regulation and performance. They do not address valuation problems in Argentina. And they ignore that indexes can not always be replicated.
  • Auguste & Artana (2006) . In Perú AFPs had better performance either because they rebalanced their portfolios properly or because they could undo the regulation limit inside each asset class. In Chile AFPs achieved similar return but better adjusted-by-risk return, even when they were compared with the best mutual funds.



Affiliates worse because of limits to invest abroad (Yermo, 2004 for several countries in Latam).

  • Affiliates worse because of limits to invest abroad (Yermo, 2004 for several countries in Latam).

    • But some problems. There is no proper handling of different valuation methods and in some cases returns are not adjusted by risk
  • Evidence of persistence in Chile. Medium size AFPs do better (Walker, 1993; Arrau & Chumacero, 1997 and Zurita and Jara, 1999). Auguste & Artana (2006) find persistence, more in Argentina and Perú, less in Chile, and in Colombia it is explained by one AFP

  • None of the previous papers take into consideration that workers have other assets (real state, human capital)

    • Optimal investment decision is different for each individual. This supports multifunds.








Auguste & Artana (2006). In Argentina, Chile, Colombia and Perú there is some evidence that differences in AFPs’ fees are not explained by differences in performance.

  • Auguste & Artana (2006). In Argentina, Chile, Colombia and Perú there is some evidence that differences in AFPs’ fees are not explained by differences in performance.

  • Information on real state prices is scarce (Argentina and with some problems in Peru) but where it is available there is a negative correlation between changes in real state prices and AFPs’ returns (-0.18 in Argentina and –0.04 in Perú).

  • There is some evidence that the ranking in Argentine AFPs is modified when workers have some assets like real state.



AFPs as assets managers did well.

  • AFPs as assets managers did well.

  • AFPs exploit the possibilities inside each asset class to reduce the burden of the draconian regulations

  • There are regulations that are costly (in terms of lower risk-adjusted returns) but others may help

  • What is important is the cumulated return at the time of retirement and this is not properly addressed by standard estimates of past performance

  • It is important to study deeper the correlation with other assets and the returns net of fees

  • Both theory and practice suggest that multifunds are a good idea




Do'stlaringiz bilan baham:


Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2017
ma'muriyatiga murojaat qiling