“In the CEM universe it [the Fund] looks great, but in TUCS it's closer to the bottom 5%... TUCS [measures] performance ranking and CEM [measures] cost ranking… If you don't survive short term, long term does not matter. “
In response to an annual report almost devoid of information on investment performance coupled with consistently rosy reports on the topic emanating from the Office of Investment Management, I sent two open letters to the Representative of the Secretary-General for Investments, Pedro Gauzo, on 15 October 2023 http://unpension.blogspot.com/2023/10/open-letter-to-pedro-guazo.htmland on 8 November 2023. http://unpension.blogspot.com/2023/11/second-open-letter-to-representative-of.html
Neither letter has received a direct response. Rather, they appear to have triggered a flurry of postings of investment reports on the Fund’s website and ever more strenuous statements about the Fund’s positive investment performance vis a vis its peers, the latest in the Fund’s New Year’s message on 3 January 2023.
Questions on the topic are coming from other sources as well. In its latest resolution (A/78/662, para. 3), the General Assembly endorsed the recommendation of the Advisory Committee on Administrative and Budgetary Questions (ACABQ) (A/78/7.Add.1, para. 12) that it requests information from the Secretary-General about the Fund’s performance in comparison with its peers.
An analysis of the Fund’s investment performance by its Former Chief Risk Officer in October 2023 formed the basis of the questions posed to the RSG in the two open letters last year.
Here’s his latest very informative posting. Notable quotes: “In the CEM universe it [the Fund] looks great, but in TUCS it's closer to the bottom 5%... TUCS [measures] performance ranking and CEM [measures] cost ranking… If you don't survive short term, long term does not matter. “ Stay tuned.
(Note: There has also been no response from the RSG to my question about the cause(s) of the OIOS investigation of 15 Fund staff (paragraph 109 of UN report A/78/301) including, reportedly, senior investment officers, and the possible impact on staff morale and performance, in the past, currently, and in future.)
______________________
By Ajit Singh, Former Chief Risk Officer, United Nations Joint Staff Pension Fund UNJSPF)
9 January 2024.
Folks, here is the most recent TUCS report comparing the UN Pension Fund with a peer group of US Public Pension Funds over 5 billion as of third quarter 2023. For the 1 year, the Fund is slightly above the median, second quartile. However, for 3, 5 and 10 years, the Fund is solidly in the bottom 4th quartile, closer to bottom 5 percent. Outside of the UN, a chart like that would be a career ending event for the CIOs.
The Fund asserts that it compares very well with its peers. The Fund did not use TUCS; rather, it used CEM ranking. In the CEM universe it looks great, but in TUCS it's closer to the bottom 5%.
The question then is: What is in the universe? I have had multiple meetings with TUCS, CEM and some of the largest custodians to find that out. It’s good information to compare and contrast both. I have decided to present to my Board both TUCS and CEM rankings to make sure they are fully informed: TUCS for performance ranking and CEM for cost ranking.
Annual reports clearly show that the Fund has not made its real return objectives for over 15 years (3.43% for 10 years and 2.18% for 15 years, both below stated IPS objective of 3.5% real return). Recent multi-years subpar performance seems to have dominated excellent previous performance.
The Fund is leaning on past excellent performance to assert that it’s meeting long term objectives. If the current trend of subpar performance continues, it will wipe out the old excellent performance. Not meeting a real return objective for 15 years should trigger multiple alarms. An urgent diagnostic is needed. Excellent funding ratio seems more an actuarial artifact than due to investment performance.
If you don't survive short term, long term does not matter. Fifteen years is too long to not meet the stated objectives and simultaneously claim that the Fund is the best performing fund! TUCS reports clearly show that is not the case.
The most important decision the Fund makes is asset allocation. The Risk group many years ago tested this. The group calculated with precision the impact of asset allocation on each quarter and even calculated the impact of delay in implementation of the asset allocation decision. Math for these calculations is well defined and the methodology was fully documented in the presentation. Perhaps, the Fund can repeat that exercise again to see if the asset allocation process is still effective.
TUCS vs CEM
Who is in the peer universe? The Fund does not tell us that. It does say: “A peer group of 19 global funds with assets under management ranging from $48.3 to$127.8 billion, with a median size of $82.4 billion.”
While I have no way to know who is in the peer universe, a brief search shows who is not. It only uses 19 funds with AUM from 48.3 to 127 billion. This means some of the very well managed funds like Texas Teachers, CalSTERS, CalPERS are not being compared with.
This is not a true peer comparison as most of the best managed public pension funds have been excluded. The Fund should disclose who are these 19 funds it considers its peers.
Inclusion of foreign funds as peers can lead to false comparison: The site says “A global universe of 277 pension funds, with assets ranging from $0.1 to $1,740.3 billion, representing a total of USD $11.6 trillion. The median fund size is $8.6 billion”. Impressively big numbers.
However, international plans are not true peers. The Fund’s base currency is USD. It prepares its financial statements in USD. International plans use different base currencies. Translating that into USD brings FX conversion risk. (On June 30th, EURUSD was1.1024. On Sep 30 EURUSD was 1.0615. A European plan with stronger performance in EUR will look weaker when base currency is changed to USD).
Then there is the issue of home country bias which cannot be equalized. This makes asset allocation incompatible with the Fund. The Fund has not explained how it adjusts for these risks. The real peers of the Fund are US Public Pension Plans over 5 billion with USD as base currency.
All US Public Pension plans invest globally, just like the Fund does. Moreover, a 5 billion plan can do everything a 50 billion plan can do. A 5 billion plan can invest globally, implement various investment strategies, access public and private asset classes. Any US Public Pension Plan over 5 billion should be compared with, hence the case for TUCS. If the Fund insists on being compared with foreign plans, it should report performance both in base currency and in USD.
TUCS universe data is reported by custodians, many of whom include TUCS in their reporting package. As part of the agreement, the Fund can opt-out or opt-in. The Fund had opted-in as it used to produce TUCS reports. Someone must have gone to the custodian and explicitly opted-out. This materially important decision must have a well-documented rationale. Did the Fund share such a significant decision with the Board or the IC (Investment Committee)?
·
TUCS has contractual relationships with the custodians. CEM data is self-reported and it does not have any such agreements in place. Most custodians automatically add a TUCS report as part of the reporting package (process differs from custodian to custodian-- some default to opt-in, some default to opt-out).
As part of the agreement with the custodian, the Fund can opt-out or opt-in for TUCS reports. In contrast, CEM Universe comprises data from its self-reporting clients with whom it has consulting relationships. In addition, they send cold emails to non-clients to report their performance if they want. These emails are more likely to be ignored.
Please see the table below attached showing the data from the largest 4 custodians in the US who report the data to TUCS. One can see that TUCS has well developed pipes with custodians and the data is reported with no intervention from the clients who opted-in. This makes TUCS data a lot more persistent as barriers to exit from TUCS are a lot higher than CEM.
TUCS universe is persistent, CEM may not be so much. TUCS universe data is reported from the custodians directly, not self-reported. Those who reported 5 years ago are still reporting. If one wants to get out, they need to call the custodian, modify the agreement to opt-out. In contrast, most CEM peers have a consultant relationship with CEM. These are the plans which hired CEM to do cost benchmarking. Most ignore their cold emails to report performance unless they have a client relationship with them. In addition, CEM has some external consultants who report their client’s data to CEM.
·
Direct, contractual relationships with custodians are a big deal. It creates a wider universe, fewer dropouts and facilitates consistent reporting.
TUCS is well accepted in the pension industry and all publications which rank pension fund performance refer to TUCS. It’s the most sourced performance ranking universe.
Example of the cold email blast from CEM I get every quarter, emphasizing that CEM data is self-reported. It does not have independent pipes built with custodians. CEM is very good in what they do for cost benchmarking, but they are not known for performance benchmarking. I would certainly hire CEM to do cost benchmarking for the fund I manage, and I hope I come last. The biggest driver of higher cost are staff salaries, and here I do not want to win.
Folks, NYC Marathon ranks its participants based on when they crossed the finish line, not by how many calories they burned to finish.
