At What Point Do You Make Your Own Index From Individual Stocks?

[TripleB]

There must be a certain point of personal wealth where making your own index from individual stocks is preferable to buying an index fund and paying a 0.07% ER I believe that level of wealth is well beyond anything I will ever achieve in life, but I am curious from an academic standpoint. I plan to throw some numbers together in Excel and build a decision model, and would like to hear feedback on factors to consider in the model.

My current thoughts for factors are:

1) How many stocks do you need to hold to effectively track the TSM index, or create a suitable analog? Traditional finance theory says 30 stocks are "good enough" for diversification, and the DJIA is just that—30 stocks, and it tracks the TSM pretty damn close.

2) Is it worthwhile to hold 80% of your equities in individual stocks for "free" (no expense ratio) and then hold 20% in an extended market index (at an expense ratio of about 0.15% or so)

3) Sensitivity analysis on the model based on frequency of future trades. i.e. am I contributing/withdrawing money every week that I somehow have to figure out how to disperse into individual stocks, or just maintaining a large pool of money? If I decide to sit on cash instead of stocks for a period of time to make larger block transactions, then what is the opportunity cost of losing market exposure during that time?

4) Is it worthwhile to maintain a portion of it in an actual TSM index fund, for new purchases or withdrawals, while keeping the bulk in an expense-free individual stock state?

5) How much time will this take? What is the value of your time?

6) Sensitivity analysis of upside and downside risk of failing to track the market appropriately.

7) Are the spread costs of buying individual stocks multiplicative as compared to just a single spread of buying an ETF of TSM?

8 ) What is the cost of the stock trades themselves?

As you can tell, this isn't a simple decision model I propose to create and will probably take several hours to build. My goal is to find a number value that will say "If someone a goal to invest/withdraw $Y on a basis of Z time frame periods, SOLVE for the value of $X to make it worthwhile to build your own index."

I'm pretty sure the answer will be that it's never a good idea to do this. I do have a fantasy of having a 0% Expense Ratio portfolio where I hold individual stocks, gold bullion, and treasury bonds/bills direct from the government, and have to pay no one anything (which is a fallacy because I am paying in opportunity cost of my time).

I could envision myself with $1M in assets and holding $250k in stocks, with about $200k across 30 individual stocks of about $5k to $10k in each stock, and $50k in a TSM mutual fund, using a broker that gives me free trades on those 30 stocks I hold. I'd only be saving myself about $150 per year to do this with a high likelihood of a potential deviation from the index of a few thousand dollars, so it seems silly to even consider. It's more of an academic question I have just to calculate it for the sake of calculating it.

[stone]

TripleB, I'm an index fund skeptic. I've even wondered whether the whole index fund concept gets trumpetted to create a pool of "dumb money" to be fed off of. What would be interesting would be to search for a clutch of say 10 decent companies that were in diverse sectors. Perhaps include a couple of conglomerates such as GE and Berkshire Hathaway and perhaps even a couple of good quality small cap companies. Then have them each as  2.5% of the PP and use the same rebalancing tolerances as for the 4x25% PP and see what the consequence would be including the trading costs. I wouldn't be astonished if that beat an index fund even if backtested with many different baskets of 10 companies.
Basically the whole reason the PP works is because the efficient market hypothesis is mistaken. The market overshoots what makes sense with hindsight. If it didn't, then the PP would not work anything like as well as it has done. I don't see why that same phenomenon might not apply to individual stocks just as much as to asset classes.

[fnord123]

Self-managing an equivalent of an index fund would take a lot of work - one would need to monitor the market cap of each company, figure out taxes, make decisions about when to add/remove companies, etc.  No rocket science, but a lot of tedium - why not leave it to someone else to do for a tiny 7 basis points?  If a person enough money to hold enough stocks in sufficient volume for it to make sense financially, wouldn't their time be worth so much that spending time on managing the self-run index be a net loss?

As to picking 10 decent companies - that is a very speculative play.  AIG, Fannie, and Freddie sure looked like decent companies in terms of balance sheets right up until they didn't.  There are lots of examples of really big companies imploding/failing.  Let's assume 1 in 100 decent companies fail each year - that means every year, you would be risking a 10% chance of a 10% portfolio loss - that is serious tail risk.  I just don't see the benefit of taking on this extra risk.

[melveyr]

Whether or not the markets are efficient, we live in a winner take all society. That means, you better hope you own some shares in the winning company. Technology creates massive stratification of profits in many industries, because it only takes one company to service the entire world for many tasks. Once that firm has cemented itself, it can persist for a long period of time before they become obsolete. I think this trend will continue, and it makes a subtle case for indexing. We don't know who is going to be the winners of tomorrow, but it is likely whoever the winners are, are going to win big.

I wonder if the idea of only owning 30 stocks being diversified is from empirical evidence, or extrapolation from assuming a normal distribution.

Still though, I am also curious to see how much money it would take to replicate the total market index. It would take more discipline, but I think I would enjoy seeing all of the shares. It would really bring home the concept of owning a little chunk America.

[stone]

Fnord, if as you say, there is a 1% chance of a company going bust in any given year, then remember each company is only 2.5% of the "10company 4x25 PP". To some extent many companies shares look like 30 year amortizing (if companies only last 30years), zero coupon bonds (since dividends are so low). Doesn't all the potential for harvesting the earnings in such a situation come from rebalancing benefits as the price zig zags up and down? From what I can see, index funds let most of that potential slip by unharvested.
I guess the very hard thing to do with such a 10 company portfolio would be to know when to pull the plug on a given holding. Perhaps every 5 years decide which 10 companies to hold making a special effort not to be influenced by what was held up to that point (ie start totally afresh)? I have to stress that I'm just spouting my ignorant musings:)

Melvin R, do index funds actually do such a good job of catching profits from innovations? The vast bulk of what you own with an index fund is the 30 biggest companies anyway. I thought that it was very hard to profit from catching new technologies early because generally a few companies are all trying at the same time and only one will prevail.

[MediumTex]

Whether or not the markets are efficient, we live in a winner take all society.

When I look at the financial institutions around the world it looks more like we live in a "loser take all" society.

It's hard to predict what sector or companies are going to do well (i.e., be "winners").  Right now, one would think the gold miners would be booming, but they're not.

Perhaps people found Charlie Sheen's rants earlier this year entertaining because he kept proclaiming he was "winning" when it looked overwhelmingly as if he was losing.  Maybe people recognized this pattern from recent events in the world economy and it resonated with them as an absurd (but true) commentary on the world we are living in.

[SmallPotatoes]

I plan to hold the few stocks I do own for some time (AAPL, CLX, CVX, etc.) and use a TSM fund for rebalancing.  If you keep an eye on your individual equities and buy/sell every year when you rebalance it works well enough.  However, I don't have enough time to micromanage my own TSM portfolio.

[melveyr]

MT:

Point well taken. I guess what I am getting at is that a small minority of stocks can play a large role in how the indices perform. To hold less than the indices you are likely to achieve a much higher degree of variability over a long time frame. With the worst case scenario being missing the colossal winners and holding the colossal losers.

Indices insure that you will get your fair share of the action. If our economy is function properly, than the economic gains of the winners will offset the losses from the losers.

[craigr]

The only good argument I've seen for running your own index is for extreme tax loss harvesting reasons. So trading stocks within the index of winners vs. losers to offset losses with gains. I know of one company that was doing this for very wealthy individuals. BUT, they charged a very high annual expense ratio. So ultimately it was hard to say whether the tax savings were going to offset the added expenses.

But other than that, I can't imagine an individual being able to get the volume trade costs of someone like Vanguard or iShares down to a level low enough to beat their current expense ratios.

[Jan Van]

Maybe the point at which you have enough money to make this strategy feasible is also the point at which you don't need to be in the stock market anymore?

[TripleB]

Maybe the point at which you have enough money to make this strategy feasible is also the point at which you don't need to be in the stock market anymore?

I subscribe to the PP idea fully, and believe that I always need to be in the stock market. What's my alternative to protect my wealth? 1/3 Bonds, 1/3 cash, 1/3 gold? Get rid of stocks from the PP? What if the economy is great for 20 years. Then my money is tanking.

How about 100% cash? That's silly because we could hit hyper inflation.

How about 100% TIPS? Inflation adjusted so I cant lose money, right? Unless the government's fake CPI numbers don't match my true cost of living.

I could probably come up with a non-stock based portfolio that is something like 75% cash, 15% gold, and 10% long term treasuries but that would be silly. I'll just stick with the PP because I know it works.

[TripleB]

The only good argument I've seen for running your own index is for extreme tax loss harvesting reasons. So trading stocks within the index of winners vs. losers to offset losses with gains. I know of one company that was doing this for very wealthy individuals. BUT, they charged a very high annual expense ratio. So ultimately it was hard to say whether the tax savings were going to offset the added expenses.

Craig, can you be more specific on how this would work? If one sells an individual stock because it has a loss, and you want to realize an immediate tax benefit, then you will be out of that position for 30 days to avoid a wash sale. Perhaps you could do options trading to minimize downside risk of loss of market exposure while the wash sale timeframe expires?

Also, to be truly tax efficient, one would want to only sell winners at long-term >365 day periods to get the lower capital gains rate, but only sell losers <365 days to get the marginal tax deduction. However, if you have preferred long-term rate gains and short-term rate losses, the losses count against the gains (rather than your other income) and you lose some tax benefit.

I could imagine some strategy where in alternating years you do long-term losses and long-term gains in one set of years (i.e. all even years), and in another set of years (i.e. odd years) you only realize short-term losses and write them off against marginally-taxed investment or earned income.

Then if you have any short-term losses left over, that would otherwise be wasted since you dont have enough other income to write them off against, then either do something like a partial Roth IRA conversion, or hold the losses until the following year, make them long-term, and write off against long-term cap gains.

This could be an interesting plan of action, if you have a brokerage like WF that gives you a free number of annual trades each year. Then just hold the individual DJIA 30 stocks, and some TSM mutual fund for ease of rebalancing.

For the poster above who questioned the time involved in replicating the index and calculating market caps, my response is that I would just log into VG, look in their TSM fund prospectus, and copy their ratios of the top 30 stocks held. I might be lagging the true index by a few months, but in all reality an index is just an artificially constructed entity. There's nothing magical about it. My minor deviations might cause a minor win against the index, or a minor loss against it. As long as I am not market timing and trying to actively pick winners, I should net out even over the long term.

Technically I won't be able to perfectly replicate the index anyway, because if we're talking $50k to $250k of total stocks, you would have to buy fractional units of stocks to replicate the index perfectly since it's only a few thousand dollars in each stock and not possible to get perfect percentages.

I'd like to explore Craig's comment about tax efficiency. That may make some modified version of this idea worthwhile.