What AIG Can Teach Investors About Diversification
Imagine that back in 2007, you had $3,000,000 to invest on behalf of a private family investment partnership you were running. You decide to split this into three different companies. One of these was AIG, the insurance conglomerate.
AIG shares fell from a high of $1,459.00 each to $6.60. The Board of Directors had to do a 20-1 reverse stock split to keep the thing from trading for less than the value of bottle caps. Even with the stock at $50.84 today and $36.03 in cash dividends you received over that time period (most of which came before the destruction of the common stock), the economic losses are catastrophic. Your $1,000,000 now represents around $35,400 worth of stock and $25,000 in cash dividends that piled up in your brokerage account for a grand total of $60,400.
Take a moment to fully reflect on the gravity. If you had invested $1,000,000 in the company seven years ago, you’ve patiently waited to find yourself sitting on only $60,400, a big part of which came from dividends. Your losses were a staggering $939,600. For every dollar you contributed, you’re left with a mere 6¢ despite the better part of a decade passing. The odds are overwhelming that you will never make that money back with inflation adjustments because the AIG company is still as large as it was, with all the lumbering slowness of a behemoth, only there are exponentially more shares outstanding. The stock could quadruple and you’d barely have made a dent in your losses. That is the very definition of permanent capital impairment.
Here is the crazy part: That other $2,000,000 you invested? If you earn perfectly average rates of return on your holdings for the next 25 years, you’d end the period with around $21,670,000, despite losing one-third of your initial capital straight out of the gate. If you had one average holding, and one super-holding that turned out to be a winning lottery ticket, like Wal-Mart, you’d end up with $106,000,000.
That is the reason diversification matters. Done correctly, it can increase your odds of good results in a very real way. This is especially true when you realize that most of the time, the businesses that do fail won’t fail right away, so you get an Eastman Kodak effect where you might have years of spin-offs and dividends compounding for you outside of the initial investment itself.
Throwing all of your capital into a single company, no matter how successful it appears to be, is not an intelligent way to behave because, 1.) if you are right, you would have still grown richer if it were part of a reasonably diversified portfolio that did well in the aggregate over the measurement period and 2.) if you are wrong, you exponentially increase the chances that you lose everything. It’s a terrible way to position your finances even if the probabilities are in your favor. Unfortunately, it seems the 1 in 8 families who invest in stocks directly are doing just that by choosing to own shares of only a single enterprise or two, often in the employer of one of the spouses.
Update: For a related post on how diversification can increase returns under certain circumstances, read The Mathematics of Diversification and Wealth Building.
Reader Comments (12)
Comments are presented chronologically, with replies indented beneath the comments to which they respond.
IW
April 4, 2014
how did you go from 2 mil USD to $21,670,000 in 25 years?
Making reasonable assumptions would make the number closer to an expectation of 10$ usd ( which will be worth more like 4 mil in today's values due to inflation)
Scott McCarthy
April 4, 2014
Replying to IW
The implied IRR on the $2mm is 10% nominal. Not sure why you think ~3% real is more realistic?
IW
April 4, 2014
Replying to Scott McCarthy
Most projections as of today are at 5%, why use the long term historical average? The starting point does matter
IW
April 4, 2014
Replying to IW
Actually it could be 5% real thus ~7% nominal - In either case $21,670,000 probably is too optmimistic tho.
Gilvus
April 4, 2014
Replying to IW
Historical inflation has been between 3-4% (calculated using CPI-U and the CAGR formula), so 8% nominal is more like it.
On top of that, most "total stock market return" statistics are calculated from the index on the start and end dates, while completely ignoring dividends you'd have received for owning every company on that index. If you had reinvested your dividends throughout the 25 years, a 10% nominal return is still quite reasonable.
IW
April 5, 2014
Replying to Gilvus
The numbers are for total returns (incl dividends)
There may have been a small UK bias in what I wrote but here are my sources, from
http://www.fsa.gov.uk/static/pubs/other/projection-rates12.pdf
2% inflation: "Bank of England continues to target 2% CPI inflation in the medium-term (within a symmetric 1-3% range to allow for short-term fluctuations) and this still tends to be the basis for most medium-term forecasts" (then they go on and say they use a different number for RPI but for my personal portfolio I tend to prefer CPI so that I can compare cross-country real retuns)
In the same document you'll see an expected 4%-5.5% expected real total return for UK equities.
Having a quick look at the US projections at
http://www.rickferri.com/blog/investments/portfolio-solutions-30-year-market-forecast/
the numbers look similar for the US (both for inflation and returns).
2% inflation expectation
5% real equities return expectation.
I think the 10% rate is on the optimistic side of things but who knows we may be luckier than we expect to be.
IW
April 5, 2014
Replying to Gilvus
so nominal amount (7% nominal return ends up): 10.8 mil (2*1.07^25)
real amount (5% real return): 6.8 mil (2*1.05^25)
not bad but not 21 mil.
IW
April 5, 2014
Replying to IW
I made a post linking to the UK FCA and Rick Ferri for the 5% real total return and 2% inflation expectations - it's waiting for approval
Gilvus
April 4, 2014
Replying to IW
Compounded Annual Growth Rate = ((Future Value/Present Value)^(1/No. of Years)) - 1
CAGR = (21,670,000/2,000,000)^(1/25) - 1
CAGR = (10.835 ^ 0.04) - 1
CAGR = 1.1000 - 1
CAGR = 0.1000 = 10.00%
Like Scott wrote above, 10% nominal is very reasonable.
Gilvus
April 5, 2014
Good find. However, I still suggest you read the linked article from almost exactly one year ago (fourth paragraph in the article above) about why 10% is a magic number on this blog.
When you start projecting so far out (10-15 years for the PwC report and 30 years for the Portfolio Solutions' report), you're stuck with using qualitative magic to extrapolate into the future. Generally speaking, Joshua and most of the regulars on this site prefer using dusty tomes of historical data rather than scrying over crystal balls.
IW
April 5, 2014
Replying to Gilvus
just read it - I see a dollar cost averaging is assumed, where the long term results (10%) are asymptotically almost guaranteed to happen.
As buying 3 stocks and "letting them be" was assumed, I went with the lump sum investment scenario:The numbers I quoted are projections for investing a lump sum as of now and leaving it invested. I don't find these numbers to be crystal ball like, I think they're pretty reasonable given the current market valuation but again, dollar-cost averaging the amount (investing in both overvalued and undervalued markets) will bring it to the historical average.
Re whether historical data or projections should be used: anyone investing a lump sum at the peak of the dot com era in 1999 was almost guaranteed to perform below the historical 10% nominal, valuations at the starting point matter a lot (assuming a lump sum is invested). Not really a criticism to the blog which I think is excellent and constantly stresses the importance of valuation.
Gilvus
April 5, 2014
Replying to IW
Yeah, all-in vs. dollar-cost averaging makes a big difference because the latter normalizes your return as you stretch your time horizon. It also removes one's ability to romanticize the "what-ifs" of investing (e.g. "what if I poured X dollars in Y company on 3/9/09?") that (in part) drives speculative bubbles.
Have a good weekend.