Wednesday, September 28, 2011

Is TIAA Traditional a good deal? - The Payout Phase
The accumulation in any retirement investment is only as good as the income it buys. TIAA Traditional with the 3% guaranteed interest rate does not permit lump sum withdrawal of the accumulation – it has to be done over a period of nine years. According to a recent TIAA-CREF testimony about one third of retirees annuitize their accumulations, i.e. turn the accumulations into a stream of lifetime income. Is this income a good deal? Would one do better withdrawing the money and buying an annuity from another provider?
There are several difficulties in comparing TIAA’s annuity payments to alternatives. First, annuities come in thousands of varieties. This makes apples to apples comparisons difficult. TIAA Traditional in the annuity pay-out phase, like the accumulation phase, also has the capacity to pay additional amounts of interest, and the income payments you receive may fluctuate though never below the guaranteed amount. Historically, these fluctuations were almost always up but it has happened that payments went down. In the last 10-years or so payments were mostly flat. There is a guaranteed floor but it appears so conservative that the benefits have always been above this floor.
Another complication in comparing TIAA Traditional payments to alternatives is that TIAA payments depend on how long you have been accumulating money within TIAA. The longer you have been with TIAA, the higher the payment. The difference comes from a gradual payout of what TIAA calls “unneeded contingency reserves.” In one example, being with TIAA for 30 years lead to a 5% increase in the initial payment. Of course, there is no telling what difference this will make in the future.
Keeping in mind the difficulties in making valid comparisons, TIAA payments appear quite competitive – especially for women. A study from the year 2000 compares annuity payments from TIAA to a large number of insurers and finds TIAA payments slightly higher for men and quite dramatically higher (about 15%) for women. More recently I compared TIAA Traditional quotes to Berkshire Hathaway and MetLife’s fixed annuities. The TIAA payments purchased using recent vintages were a bit lower, but payments purchased using pre-1990 accumulations were more than 10% higher. However, TIAA payments can go up while fixed annuities from Berkshire or MetLife will not. Moreover, TIAA payments are unisex, i.e. men and women get the same payment. Other insurers pay women less than men because women are expected to live longer. This means that if there is any advantage of going with TIAA it is certainly higher for women than men.
A major issue with purchasing a fixed annuity is that inflation will eat the purchasing power of the payments. An ideal solution is inflation indexed annuities, but the market for these seems small and not very competitive. Given TIAA-CREF’s non-profit status, and that its TIAA Traditional annuities have potential for additional amounts, it is quite likely that TIAA would pass any gains associated with high inflation to its annuitants. Any other annuity issuer would pass these gains to its shareholders. TIAA annuities are certainly far from an inflation hedge, but I speculate that in this respect they beat other insurers.

Is TIAA Traditional a Good Deal? - The Accumulation Phase
TIAA Traditional is a guaranteed annuity that is best understood by considering its two phases: the accumulation phase and the payout phase.  In the accumulation phase, participants contribute money which grows at a guaranteed minimum rate of 3% for most contracts plus some additional amounts as declared by the company. The additional amounts are applied every year but different amounts are applied to different portions of the account. For example, additional 1% may be applied to money that was contributed in 2002 but only additional 0.5% to money that was contributed in 2010. Thus, different rates of interest are applied to different `vintages’ in each account.
The system of vintages makes calculating rates of return complicated. A few years ago TIAA began publishing ten-, five- and one-year rates of return. These are calculated consistent with the SEC’s guidelines for mutual funds as the annual growth of one dollar deposited ten, five- and one- years ago. This is helpful but it is not quite comparable to rates published for mutual funds. The problem is that the rates of return reflect the growth of only that particular vintage. For example, the ten-year return only reflects the growth of the money in the ten year old vintage. It does not apply to money contributed before the ten years or after. In the case of a mutual fund, the ten-year return applies to the entire balance as of ten years ago. Thus, a side by side comparison of TIAA Traditional returns and a comparable mutual fund is not possible.
I had a conspiracy theory. Because TIAA can assign different additional interest rates to different vintages, it has a great incentive to assign additional interest to the vintages that are `invisible’ in the ten- five- and one-year table of returns. For example, a four-year old vintage could be assigned very low additional interest, because the return on money deposited four years ago is not shown in the ten-, five- and one- year rates of return. Next year that vintage will be five-years old and so TIAA could assign high additional interest because it would be visible in the ten-, five- and one-year return table. TIAA is able to do this because they could pick and choose which vintage yields high interest.
The lack of data on historical interest rates for each vintage only fueled on my conspiratorial mind. It is impossible to find these historical rates on the TIAA-CREF website or anywhere on the internet. I requested the historical interest rate from TIAA-CREF as part of due diligence done on behalf of my institution’s retirement plan committee. A spreadsheet came back a few weeks later with all the historical rates, but it also had a message “Not for Further Distribution.” Why? I asked myself. This data should be on the internet. After all, there are 190 billion dollars invested in this product. Shouldn’t everyone have access to the historical interest rates? What are they trying to hide?
Alas, if they are hiding anything it is pretty good performance. I pitted TIAA Traditional against Vanguard Total Bond Index Fund (VBMFX) and Fidelity U.S. Bond Index Fund (FBIDX). The results are shown in the chart below. A participant putting away $100 a month starting in April 1990 would end up with almost identical amounts of money at the end of 2010 using any of the three investment options. The trajectory for TIAA Traditional is remarkably smooth compared to the two index funds. It appears that TIAA probably invests in the same types of investments that the bonds funds do, but TIAA smoothes out the fluctuations. In addition, TIAA does not seem to play any games with the additional interest rates. The “visible” ten- five- and one- year returns are no higher than the “invisible” eleven-, nine-, six-, or four-year horizons. I also compared the TIAA Traditional to Barclays Stable Income Market Index. This index begins only in July 1999. TIAA Traditional RA contracts slightly outperformed the index, while TIAA Traditional SRA contracts slightly underperformed.  
In summary, average returns on TIAA Traditional match those in the overall U.S. bond market, but TIAA does a fabulous job smoothing out short term fluctuations. It has returned well above its guaranteed 3% interest rate. Yes, past performance is no guarantee of future returns, but at least I can put my skepticism and conspiracy theories to rest.

Is TIAA Traditional a Good Deal?
Three months ago I was very skeptical of TIAA Traditional. A number of things made me nervous: the apparent arbitrariness with which crediting rates are set; the illiquidity of accumulations; the opaqueness of the vintage system; the lack of information on historical interest rates; and the entrenchment of individual contracts that makes moving TIAA-CREF assets to a different provider difficult.  I went on to investigate. The result of this investigation are the following two posts. They summarize two phases of the TIAA Traditional: the accumulation phase and the payout phase. The bottom line is that that TIAA Traditional is a pretty good deal: in the accumulation phase, the returns match returns on bond index funds; in the payout phase the payments appear competitive (particularly for women) and are probably better inflation hedge than annuities from other insurers. If you would like details with pictures, read the next two posts.
TIAA-CREF has 70% of the higher education market
TIAA-CREF continues to play its historically important role in providing retirement savings in higher education. The 2009 Form 5500 disclosure data on 40 largest private higher education defined contribution plans shows that the average share of TIAA-CREF in assets is nearly 70%. Within TIAA-CREF, the two oldest investment options, TIAA Traditional Annuity and CREF Stock variable annuity account for 47% and 30% of TIAA-CREF assets.  This means that TIAA Traditional Annuity accounts for nearly one third of all retirement assets in higher education, making it undoubtedly the single largest investment option. Even Federal Reserve Chairman Ben Bernanke, according to his financial disclosure statement, has most of his financial assets in TIAA-CREF - split about evenly between TIAA Traditional and CREF Stock.