DCA withdrawals

[WM]

There seems to be good agreement that dollar cost averaging is a good way to put money into mutual funds - simple, can be automatic, and you generally come out ahead as the share price fluctuates. But I've seen a couple references here and other places that say when you take your money out, DCA works equally much against you.

I remember being convinced by the math when I saw it, and I admit that intuitively it makes sense that you can't have it both ways, but I guess my brain doesn't want to admit that DCA-ing out would be so bad.

What is the better strategy? Lump sums? Do you have to start trying to market-time?

 

[renferme]

The problem with mutual funds is the tax reporting of dividends, cap. gains, and
withdrawals.
Do a lot of withdrawals, like monthly for example, and that's a lot of record keeping and lot to put on the 1040 at tax time - a real nuisince in my opinion.
I'm no expert, but if you're retired and want to withdraw from a mutual fund,
perhaps the best (easiest) way to do it is to make one withdrawal per year.
Put the money in a savings account or in short term CDs with staggered maturities.
Then as the CDs mature each month, use that money to live on (or whatever you want to do with it). Then there is only one line to report at tax time.
.
Or do withdrawals from a money market fund with a stable $1 per share value.
This works well within an IRA, as you can move money from fund to fund within the
IRA without any accounting hassles until you withdraw it. Monthly withdrawals, for example, from a money market fund make it easier because each month the per share value stays the same ($1), unlike a stock/bond fund where the NAV changes daily.
.

 

[jdw_fire]

What is the better strategy? Lump sums? Do you have to start trying to market-time?

Well there is the buckets approach, the selling some of your FI when the stocket market is down approach, the selling the least looser/biggest gainer approach, living off the cash portion of your portfolio until the market goes up & then selling to replentish said cash approach, doing your reallocation by just selling where your allocation is too high approach, etc. There are a number of ways (techniques) for withdrawing and not using DCA however they all pretty much do market timing, except maybe the buying a CPI adjusted SPIA approach.

 

[a]

I don't believe in DCA: Suppose a carnival barker comes up to you and offers you this bet: you bet $100, he rolls a die, and on a 1-4 you double your money and on a 5 or 6 you lose your money. It's a winning bet, right? We'd all take it. However, the DCAer is the person who says "Hold up, just wait a minute! I don't mind a winning bet, but I want to make even MORE money! Tell you what, how about I make a hundred $1 bets, same terms. I'm gonna end up even richer than you originally thought!" The carnival barker is gonna look at you like you're nuts, and rightfully.

Here's a more interesting question. Even if there is no mathematical basis for DCA being superior to lump sum, isn't it still true that, on average, people buying into the stock market - via lumps or DCA or whatever - statistically expect to buy MORE shares at a lower price? But this seems almost contradictory. It's as if people buying stocks always get more for their money, on average, than the stock market holders as a whole...wouldn't that mean, if the entire stock market gradually sold all their shares and re-bought them, everyone would end up with more shares in the market than it originally began with?

Bizarre stuff. At any rate, if DCA is a fallacy, then you don't need to worry about withdrawals, and should just do it in the way that makes you happiest for other reasons. For example, saving on tax bookkeeping as said above, or perhaps forcing yourself to do DCW (dollar cost withdrawing) as a way of preventing you from spending your money too fast.

 

[yakers]

For me it doesn't matter whether I believe in DCA or not, it is the way my resources flow in. I get paid on a regular schedule and I save/invest on a regular schedule. It seems to haved worked very well as I kept buying since 1986 through today, up markets and down. Now if I were to get a large chunk of money I might entertain the question of how to manage it but I don't have that *problem*. I know there are studies that if you have a lump sum it is better to invest it immediately than DCA as time in the market is more important than timing the market and the market tends to go up more than down. But there is maybe a 30% risk that on a given day it would work worse than a DCA approach. again, for me DCA has worked better than saving up money in cash form and then investing it. Volitility can be your friend in saving. But not in withdrawals. There are a number of sites that address this, one good one is: http://bobsfiles.home.att.net/reverseDCA1.html

 

[yakers]

Another good analysis:

http://tinyurl.com/39x73e

 

[Nords]

Here's a more interesting question. Even if there is no mathematical basis for DCA being superior to lump sum, isn't it still true that, on average, people buying into the stock market - via lumps or DCA or whatever - statistically expect to buy MORE shares at a lower price? But this seems almost contradictory. It's as if people buying stocks always get more for their money, on average, than the stock market holders as a whole...wouldn't that mean, if the entire stock market gradually sold all their shares and re-bought them, everyone would end up with more shares in the market than it originally began with?
Bizarre stuff.

I think the board's earlier discussions have pretty well established that DCA returns less than lump-sum investing, but that investors tend to put more away with DCA than they otherwise would.

In other words it's the discipline of regularly investing small amounts before you find something to fritter them away on while you're "saving up" for the next big lump. It can be automated, too, for simplicity and less emotional investing.

But there are plenty of more complicated ways to put money in the portfolio, too, like value-average investing around one's asset allocation.

 

[Animorph]

The reason DCA returns less than lump sum is because the market, ON AVERAGE, goes up. So you miss out on that growth if you delay putting money into the market.

However, that should work to your advantage when taking money out. Doing a DCA withdrawal would leave your money in the market longer and pick up growth you would miss by taking more money out sooner.

Of course that's just the averages, so do what makes you feel good. You could look at a lump sum as tilting your asset allocation to cash a little too much at the start of the year. I'd vote for DCA.

Dan

 

[ERD50]

I don't believe in DCA: Suppose a carnival barker comes up to you and offers you this bet: you bet $100, he rolls a die, and on a 1-4 you double your money and on a 5 or 6 you lose your money. It's a winning bet, right? We'd all take it. However, the DCAer is the person who says "Hold up, just wait a minute! I don't mind a winning bet, but I want to make even MORE money! Tell you what, how about I make a hundred $1 bets, same terms. I'm gonna end up even richer than you originally thought!" The carnival barker is gonna look at you like you're nuts, and rightfully.

a, the 100 separate $1 bets is a less risky bet. I would prefer it over a single $100 bet. Why would it be nuts to take the 100 $1 bets?

With one bet, you have 2/3 chance of winning $100, but a 1/3 chance of losing $100.

With 100 separate bets, you are much more likely to win a sure $67, plus or minus a bit. It is not an attempt to get richer, it is an attempt to accept less, but with more certainty.

Review that, and maybe you *will* believe in DCA.

But, to get a bit more philosophical, since he is offering what appears to be free money, I would run the other way. Something is up. Remember, 'you can't cheat an honest man'.

-ERD50

 

[chinaco]

There have been studies on this. I did just this in 2002. I sold a substantial amount of individual stocks and moved it into MF. I looked at the results historically. I did time some of the purchases instead of doing a strict mechanical investment (e.g., 25 of each month). I netted out a little worse off than if I had done a lump sum. I also tested to see what it would have looked like if I did it each month. I would would have done a little better but less than a lump sum. I believe you results depends on the general market direction. Since I was buying into a rising market, I would ave been better off with a lump sum. Had the market dropped, I would be better of with DCA. When I was buying, the market was just choppy (i.e., generally rising). Oh, one other thing... This money was going into the S&P 500 index.

 

[WM]

Ok, so DCA in won't necessarily beat a lump sum, but it's more realistic for most of us. And the argument is that by doing that you take advantage of times when the market is down, which more than offsets the times that you buy at a higher price.

When making withdrawals, you lose more by selling low than you gain (on average) by selling high. You're better off selling at higher prices (obviously), hence the possible market timing, or as one of the links from yakers said, holding several different funds so you can sell from whichever ones are up (or less down).

And I assume that if lump sums are better on the way in, then they must be worse on the way out, since you're removing a big chunk of money at the beginning of the year rather than leaving some in to keep compounding.

 

[Mysto]

I have no study or math to back it up, but I suspect that lump sum give superior results but DCA has less downside risk - based on the fact that the market is more up than down.

 

[samclem]

If you realy wanted to make this a science, then value-averaging 9as suggested by Nords) would hold the most promise of optimizing returns. Slightly simpler but based on the same premise: instead of selling assets to get a particular number of dollars (as in DCA'ing), you could sell off the same number of shares each period. When shares have appreciated, you'd be withdrawing more money, when shares have depreciated, you'd be withdrawing less money. In this way, at least you wou;dn't be selling more shares when their price is down.
Implementation would be painful--what to do about reinvested dividends, etc.
I think I'll just stick with an annual withdrawal of 4% of the existing balance, put it in a money-market fund, and do something fun.

 

[lostcowboy]

Hi new to the board, a while back I came across a web site that was trying to sell you software that would do reverse dca for you. While the web site failed, the idea is worth looking into. Maybe someone would make a spreadsheet to test it. http://web.archive.org/web/20070205233528/http://www.saccocompany.com/index.htm

 

[mathjak107]

pulling out any long term period of 10-15 years stocks are up 2/3 of the time and down only 1/3 so statistically dca has you buying in to a market going higher and higher as time goes on and buying less and less shares as time goes on . .

 

[Independent]

But I've seen a couple references here and other places that say when you take your money out, DCA works equally much against you.

I remember being convinced by the math when I saw it, and I admit that intuitively it makes sense that you can't have it both ways, but I guess my brain doesn't want to admit that DCA-ing out would be so bad.

What is the better strategy? Lump sums? Do you have to start trying to market-time?

This board started with FireCalc. I think that one big message from programs like FireCalc is that the order of investment returns matters when you are withdrawing. One intutive reasoning for that is that if you start retirement with some down years you'll sell a lot of shares quickly at the low price. That's "kind of like" the reverse DCA reasoning.

If you really need a fixed dollar amount for spending, then I don't see anything that's consistently better than periodic dollar withdrawals. It seems that lump sums merely increase volatility.

Most people think they have enough discretionary spending to have some flexibility regarding withdrawals. One way of using that flexibility amounts to market timing (I'll take out more/less this month because I think prices will go down/up soon).

Another way doesn't. This says "I'll withdraw and spend less this month because prices are down. I have no way of knowing when they will go up, so I should simply live on less while they are down." Funny, but the two different ways of reasoning could give the same dollar amount of withdrawals in some months.

Some people go into retirement with the intention of spending a fixed percent of their assets every month, thereby avoiding DCA (if it really means anything to them) and letting their spending reflect the market. You need considerable discretionary spending to make that work, but you can also argue that it maximizes the likely payout.

Others use blended strategies. Samclem refers to the "95% of previous year in down years" rule in another thread.

 

[youbet]

pulling out any long term period of 10-15 years stocks are up 2/3 of the time and down only 1/3 so statistically dca has you buying in to a market going higher and higher as time goes on and buying less and less shares as time goes on . .

While I agree with your observation in theory, it is rather hypothetical. Who would DCA over a 10 - 15 year period? Who would have the money at the beginning of a 10 - 15 year period to lump sum at the beginning?

I think most DCA scenarios consider a much shorter time span. For example, if you came upon a sum today and wanted to invest in equities but feel that the market is likely to trend lower, you might DCA over a 12 mo period and therefore buy-in at the average price over that time. And you might be right..... or you might be wrong. But unless you are talking about long periods of time, such as the 10 - 15 years you suggest, there is no statistical foundation for lump sum investing always providing higher returns than DCA investing.

 

[youbet]

The reason DCA returns less than lump sum is because the market, ON AVERAGE, goes up.

Yes, that's true for a group of investors who, in aggregate, make many lump sum investments. But for you as an individual, your single lump sum investment may outperform or underperform a 6 - 12 mo DCA approach depending on the market trend over the period.

Don't make the mistake of assuming that each time you lump sum invest you'll outperform DCA'ing. That's only true of the performance of a number of lump sum investment vs DCA scenarios on average and over a long enough period of time that you can assume a market upward trend.

I DCA when I'm too chicken-sh*t to make a commitment. For example, I'm short in my REIT allocation but I'm really unsure about making a significant move in right now. So I'm DCA'ing in over the next few months. Could it turn out that REIT's head to the moon shortly and I regret not lump summing right now? Yep! Sure could! But I'm going to wade in over the next few months. My choice. And it could turn out to be smarter than lump summing it all today. We'll see.....

 

[figner]

When making withdrawals, you lose more by selling low than you gain (on average) by selling high. You're better off selling at higher prices (obviously), hence the possible market timing, or as one of the links from yakers said, holding several different funds so you can sell from whichever ones are up (or less down).

This is one of the reasons I like slice and dice: so that in withdrawal phase, I can just sell whichever funds are doing best (or least badly!), without having to sell as many losers with the winners. If my assets were all in a TSM type fund, this would be impossible. Of course, since the funds I buy are generally index funds, they will include both losers and winners. But at least if enough of the slices are uncorrelated enough, I should have some control over avoiding selling the worst losers.

 

[headingout]

That's my opinion as well. Seems like the whole point of owning a basket of uncorrelated assets in retirement is so that you can live off what's in favor and wait for the others to come back. The maxim is "buy low / sell high," not "buy at a mix of prices / sell at a mix of prices." I don't see this as "market timing" in the conventional negative sense of emotional trading.

It just makes good sense to me to sell what's up for filling my retirement bucket(s). I'm going to call this "active" safe withdrawals, suggesting it makes more sense to take an active rather than passive approach when it comes time to harvest your portfolio.

Does anybody know of a study comparing the approaches: say pulling your 4% out of a single balanced fund, VS. pulling out of the best-performing assets each year in a basket of uncorrelated funds, over time? It seems intuitive to me that the latter approach would be superior, but perhaps somebody with more math or experience on their side can offer guidance?

 

[FIREd]

Right now I DCA into my 401K, but I use value cost averaging (VCA) in my taxable account. To me the latter makes more sense intuitively. I'd like to implement it in my 401K as well, but trading limits prevent me from doing so. I think that I will use a reverse VCA approach in retirement, i.e. sell more of the winners and less of the losers. Right now I use a spreadsheet to determine each month what and how much of it to buy, I guess in retirement I can set a a similar spreadsheet to determine what and how much of it to sell.

 

[Independent]

This is one of the reasons I like slice and dice: so that in withdrawal phase, I can just sell whichever funds are doing best (or least badly!), without having to sell as many losers with the winners. If my assets were all in a TSM type fund, this would be impossible. Of course, since the funds I buy are generally index funds, they will include both losers and winners. But at least if enough of the slices are uncorrelated enough, I should have some control over avoiding selling the worst losers.

Good point. It seems that this strategy is closely related to periodic rebalancing.

If your money is in taxable funds, rebalancing may not be tax efficient, so I think you're saying that:
When you're in the saving phase, you buy the classes that are under weighted (because of recent poor performance).
When you are in the withdrawal phase, you sell the over weighted classes.

It seem that, if your assets are in tax qualified accounts, then you can usually rebalance as much as you want and don't need to pay as much attention to exactly where you are buying or selling.

 

[FIREd]

If your money is in taxable funds, rebalancing may not be tax efficient, so I think you're saying that:
When you're in the saving phase, you buy the classes that are under weighted (because of recent poor performance).

That's the principle behind value cost averaging and that's what I use in my taxable account. I basically rebalance my account every month using new money. I buy the asset classes that were beaten down the most during the previous month. That way I don't have to sell anything to bring back my asset allocation into line and therefore I minimize my tax liability. Plus I get to buy more of what is relatively cheaper at the time.