Featured Post

Was it constitutional for Proposition 124 to replace PSPRS' permanent benefit increases with a capped 2% COLA?

In this blog I and multiple commenters have broached the subject of the suspect constitutionality of PSPRS' replacement of the old perma...

Thursday, February 8, 2018

PSPRS to members: Keep calm and don't think about annualized returns

As we discussed in the last post, we will not know for about another two months if PSPRS' investment strategy will show the success of its more conservative approach.  (The Russell 3000 dropped about 3.65% today and is down 8.33% for the month of February 2018.)  However, PSPRS will not wait that long to see evidence.  They came out with a press release on February 6, 2018 which boasts that "the current PSPRS portfolio is 72 percent less volatile than the Standard & Poor (S & P) 500 Index."  This kind of public relations is what you buy with the $95,000 a year taxpayers shell out to PSPRS spokesman Christian Palmer and the $72,000 paid to an outside consultant.

As PSPRS is wont to do, they only present half the picture.  While controlling volatility is part of the picture, PSPRS still has to balance this with maximizing returns.  The time frame PSPRS used to graph its volatility ran from May 2008 to November 2017.  This period starts with the market crash in 2008 and runs to the latest investment returns PSPRS reported.  I wanted to compare PSPRS' returns with those of the PSPRS Cancer Insurance Plan (CIP), Arizona State Retirement System (ASRS), the Standard & Poor (S&P) 500 Index, and a Vanguard 3 Fund Portfolio that Ben Carlson uses to measure endowment performance at his blog, A Wealth of Common Sense.  This Vanguard 3 Fund Portfolio is a  mix of 53% US stocks, 27% foreign stocks, and 20% bonds.  The time period for the following table is for the period ending June 30, 2017, which is necessary for consistency purposes.  Here are the annualized rates of return for the various portfolios over one, three, five, and 10 years:
 

1-year 3-year 5-year 10-year
PSPRS 11.85% 5.27% 7.95% 3.98%
CIP 10.12% 4.34% 7.64% 5.36%
ASRS 13.90% 5.70% 9.60% 5.60%
S&P 500 16.80% 7.72% 12.96% 4.86%
Vanguard 14.90% 7.00% 8.30% 6.50%

PSPRS' annualized returns over the 10-year period is lower than all the other funds.  It might be fairer to look at the 5-year returns since PSPRS had not fully implemented its strategy ten years ago.  During the 5-year period, PSPRS still lags all but its own CIP, a fund that consists of 50% US and foreign stocks, 45% bonds, and 5% commodities.  ASRS bested PSPRS by 1.65% over the five year period, and the S&P 500 bested PSPRS by 5%.  A 5% annual return on a $9 billion portfolio would earn nearly $2.5 billion over five years, so while the S&P 500 was more volatile, it also earned a significantly higher return for that added risk.

If we include reinvested dividends in the 5-year period that ended June 30, 2017, the S&P 500 earned 15.28%.  This extra annualized return of 7.33% would have earned PSPRS an additional $3.82 billion.  We all know that there is a correlation between risk and reward, but it is deceptive of PSPRS to point only to the volatility of the S&P 500 without acknowledging that the S&P 500 had nearly double the return of PSPRS' portfolio.  Even the less impressive additional 0.88% annualized return over ten years would have earned PSPRS an additional $824 million.  With reinvested dividends, the S&P 500 had an annualized 10-year rate of 7.10%.  This additional 3.12% per year would have earned PSPRS an additional $3.24 billion over ten years.  Unfortunately, there are no volatility numbers for the other portfolio.  Regardless, it is clear that the volatility/return tradeoff has been favorable for the S&P 500, but PSPRS does not want members to see this full, honest picture

While PSPRS likes to make absurdist claims on incomplete and cherry-picked information, we should not do that as well.  It is not realistic to solely compare PSPRS against the S&P 500, as no pension or endowment would ever be 100% invested in a single market index, or even the Vanguard 3 fund portfolio.  A fairer comparison would be against another pension fund like ASRS that deals with the same risk/reward conundrum.  Over the five and ten year periods, ASRS has bested PSPRS by 1.65% and 1.62%, respectively.  The extra 1.65% would have earned PSPRS an additional $767 million over five years; the extra 1.62% would have earned an additional $1.57 billion over ten years.

Despite the whining and excuses of PSPRS Board of Trustees Chairman Brian Tobin and PSPRS Chief Investment Officer (CIO) Ryan Parham, ASRS has consistently outperformed PSPRS while operating in the same market conditions.  Yet Mr. Parham, who was paid $268,000 in 2016, is the second highest paid in employee in the Arizona state government (Note: this does not include Arizona university system employees).  Who was the highest paid Arizona state government employee?  Why it was Paul Matson, the Director of ASRS, who made $285,230 in 2016.  For some perspective, Karl Polen, the CIO of ASRS, made $201,420 in 2016, while Jared Smout, PSPRS Administrator, made $210,000 in 2016.

So Mr. Matson, who has successfully run ASRS for 15 years, makes only $17,230 more than Mr. Parham.  Even more perplexing is how Mr. Parham is paid $57,000 more per year than ASRS' CIO.  The combined salary of Mr. Matson and Mr. Polen is $486,650.  The combined salary of Mr. Smout and Mr. Parham is $478,000.  Looking at the millions and billions in higher ASRS earnings, I think that the $8,650 more paid to Mr. Matson and Mr. Polen is the bargain of the century.

The difference in performance and management between ASRS and PSPRS is stark, and while ASRS continues along its steady path of stronger earnings, PSPRS has to blow smoke.  Until PSPRS has some evidence to show us that their strategy is working, they might do something useful for members, like getting their software working properly so members can get retirement estimates.  I think seven months is enough time for members to wait for this to be fixed.

Tuesday, February 6, 2018

PSPRS investment earnings through November 2017 (with some discussion about COLA's)

The following table shows PSPRS' investment returns, gross of fees*, versus the Russell 3000 through November 2017, which is the fifth month of the current fiscal year (FY), with the FY end 2014, 2015, 2016, and 2017 returns included for comparison:

Report PSPRS PSPRS Russell 3000 Russell 3000
Date Month End Fiscal YTD Month End Fiscal YTD
6/30/2014 0.78% 13.82% 2.51% 25.22%
6/30/2015 -0.73% 4.21% -1.67% 7.29%
6/30/2016 -0.32% 1.06% 0.21% 2.14%
6/30/2017 0.22% 12.48% 0.90% 18.51%





7/31/2017 0.83% 0.83% 1.89% 1.89%
8/31/2017 1.06% 1.91% 0.19% 2.08%
9/30/2017 0.80% 2.72% 2.44% 4.57%
10/31/2017 0.64% 3.38% 2.18% 6.85%
11/30/2017 1.28% 4.70% 3.04% 10.10%

There is usually about a two-month lag in PSPRS reporting its investment returns.

PSPRS' Board of Trustees published its January 2018 meeting materials last week, but in light of the market conditions over the past several days, it is nice to discuss them now, rather than earlier.  The Russell 3000 is down 6.00% for the month of February 2018 as of February 5, 2018, so this will possibly be the first real test of PSPRS' investment strategy.  I say "possibly" because we are only five days into February with quite a bit of time for the market to recoup its loss or even end with a gain.  Regardless, we will have to wait two months to find out the end-of-month (EOM) results for February 2018.

In the meantime, the November 2017 EOM results provide us with a picture of the market since the election.  For the current FY, PSPRS has earned 46.50% of the Russell 3000 (4.70% vs. 10.10%, respectively).  From December 1, 2016 through November 30, 2017, PSPRS' did a little better, earning 53.40% of the Russell 3000 (10.66% vs. 19.93%, respectively).  This tracks with what has been PSPRS' usual pattern, earning 50-60% of the Russell 3000, with last fiscal year being an exception. 

For some perspective, PSPRS' Cancer Insurance Plan (CIP) has earned 5.85% FY-to-date and 15.03% over the preceding 12 months.  These earnings were net of fees as opposed to PSPRS', which are gross of fees, making the CIP's earnings even more impressive compared to PSPRS'.  The Arizona State Retirement System (ASRS) has earned 3.5%, net of fees, through September 30, 2017, which is the latest number available on ASRS' website.  This handily beats PSPRS' return, gross of fees, through the same period.  Whether PSPRS' conservative strategy pays off remains to be seen, but through November 2017, PSPRS has lagged both the CIP and ASRS.

There was one other interesting bit of information in the meeting materials which will be especially pertinent to retirees. This is what it says about calculating the new cost of living allowance (COLA), which will be paid for the first time next fiscal year:
QUESTION: WHICH YEAR SHOULD WE USE AS THE BASE YEAR?
Since 2001, the Phoenix-Mesa CPI has been updated by the Bureau of Labor Statistics every 6 months.  Beginning later this year, the Phoenix-Mesa CPI will be updated every 2 months after each even month  (February, April, June, etc.). The results are typically available about two weeks after the end of the  month. For instance, the June, 2018 results will be available July 12, 2018. There are advantages and disadvantages of using the fiscal year vs. the calendar year as the basis for determining the COLA that
will be granted to benefit recipients each July 1st.
The biggest advantage of using the fiscal year as the base year (in other words, measure CPI from June 30 to June 30) is that the COLA is granted closer to the period of time the price of goods is being measured.  For instance, if the price of oil increases during the January to June timeframe, the COLA granted on July 1st will better reflect that increase.
The biggest advantage of using the calendar year is that staff would have enough time to calculate, communicate and process COLAs by July 1, as is required by statute. Any delays in the Department of Labor calculating and publishing the CPI would not have an effect on getting the COLAs processed timely. While we can’t comment on the intent of the legislature, it is important to note that typically when the statute refers to a fiscal year, it specifically states “fiscal year.” In the two sections of statute that refer to how the new COLA is calculated (§38-856.05 and §38-856.06), the statute does not specify “fiscal year” but rather states “year.” That may be an indication that the legislature intended us to measure the CPI on a calendar year basis. 
Recommendation: Staff recommends that we use the calendar year that precedes July 1 as the base year when calculating the amount of the COLA.
QUESTION: WHICH CPI SHOULD WE USE?
The Bureau of Labor Statistics publishes two measures of the Consumer Price Index—the CPI-U and the CPI-W. The CPI-U is the Consumer Price Index for all urban consumers. Approximately 89% of the population in the Phoenix metro area is included in this measurement. The CPI-W is the Consumer Price Index for a subset of the CPI-U population. It includes wage earners and clerical workers only, and represents approximately 29% of the population. Retirees who are not earning a “wage” are included in the CPI-U, but not in the CPI-W. Likewise, unemployed people are included in the CPI-U, but not in the CPI-W. In general, the CPI-W is more volatile because it uses a smaller sample population. According to the Bureau of Labor Statistics, in most cases across the country the CPI-W increases more than the CPI-U. In looking at the increases over the past 14 years in Phoenix, however, the CPI-U has outpaced the
CPI-W in 9 of those years. Because the CPI-U population coverage is more comprehensive, it is typically the measurement used in most escalation agreements.
Recommendation: Staff recommends that we use the CPI-U as the measurement when calculating COLAs since retirees are included in that population and it is less volatile over time.
Remember that the new COLA will be the lower of either the Phoenix-Mesa CPI or 2%.  The staff recommendations would have to be approved by the Board of Trustees, and they will almost certainly go along with their own staff's recommendations.  For retirees, the main issue is the rate of inflation.  The Phoenix-Mesa CPI-U annual average rate for the 2017 calendar year was 2.5%; for the FY that ended June 30, 2017, it was 2.25%.  Under either base year rate, retirees would be limited to the 2.0% COLA.  Historical rates for the Phoenix-Mesa CPI-U can be found on PDF page 212 of the meeting materials. 

If the Board of Trustees accept the staff recommendation on the base year, retirees will lag inflation by 0.5% in fiscal year 2019.  This may not seem like a lot, and for a retiree on a $50,000/year retirement, it would only amount to a "loss" of $250.  The "loss" is the amount of goods and services the retiree will no longer be able to afford in the second year.  However, the compounding effect over ten years of a 0.5% difference between the COLA and actual inflation will leave the retiree with only 95.7% of the purchasing power he had in the first year and a "loss" of  $2,689.  A 1.0% difference will leave the retiree with only 91.6% of purchasing power and a "loss" of $5,485 in the tenth year.  If one has another source of retirement income, such as a 457, Social Security, or an IRA, there will be some cushion against inflation.  This makes it all the more urgent that current workers utilize any and all retirement savings vehicles available to them.  I have bookmarked the Phoenix-Mesa CPI-U page at the US Bureau of Labor Statistics (BLS) for future reference.  

* Returns, gross of fees, are used because PSPRS usually does not report returns, net of fees paid to outside agencies, except on the final report of the fiscal year.  Returns, gross of fees, are used in the table for consistency.  The past two years fees have reduced the final annual reported return by about a half percent.  Returns, net of fees, were 13.28% in FY 2014, 3.68% in FY 2015, 0.63% in FY 2016, and 11.85% in FY 2017.