The Money Mustache Community

Learning, Sharing, and Teaching => Investor Alley => Topic started by: wakkowarner on December 05, 2013, 12:32:05 PM

Title: Time of month to invest
Post by: wakkowarner on December 05, 2013, 12:32:05 PM
Ok, so if you are dialed in with saving a large percentage of your earnings, wanting to invest on a regular basis, have scoped out some nice index funds with low fees, and want to set up a schedule of when to invest.

When do you invest?  Is there a better "time" to invest than others.  I'm not talking about market timing, or examining P/E ratios or anything like that.  I'm saying that, with all the historical data we have, if you were to take roughly 30 people who each invested the same amount of money but each chose their "buy" date on a different day of the month (so 1st person bought on the first, 2nd person on the 2nd, etc.) would one perform better than the others if you examined the data over the last 40 years?  I tried Googling this and it seems some call it "seasonality" though some also use that term for buying in certain months (times of the year as opposed to a particular day of a month).  Or has the cat already got out of the bag and all of this is priced in and so time of the month doesn't matter? 

What about time of day?  Is it better to time market orders in the middle of the day, to avoid the crazy fluctuations that can happen at the beginning of the day?  Is it best to avoid market orders and instead just use limit orders?

I've been investing for many years now in index funds.  I've also does some individual stock investing.  I know market timing is a trap, but does anyone know if it is generally better to time things by certain days of the week (Mon, Tues, etc.), certain days of the month (last day, first day, 15th), or certain times of the day (opening, closing, 1:00PM) regardless of the current state of the market?
Title: Re: Time of month to invest
Post by: okiedoke on December 05, 2013, 12:47:07 PM
Not to be rude, but you're overthinking this and doing it wrong. 

If you believe in largely efficient markets (which, by implication, you largely do, if you're doing index funds and say market timing is no bueno), the prices accurately reflect all the information.  There is no "right" or "better" time.  Just stick to a schedule, whatever that is.   The end.
Title: Re: Time of month to invest
Post by: matchewed on December 05, 2013, 01:02:56 PM
Doesn't matter. Day of the month or time of day for when to invest is meaningless for long term investing strategies.
Title: Re: Time of month to invest
Post by: wakkowarner on December 05, 2013, 01:09:42 PM
Not really overthinking it from my viewpoint.  Been doing it the simple way for years, haven't put much thought into it in the past.  Not sure if that lack of thought has cost me though.  Maybe a 7.5% vs 7% gain averaged over 20 years.  Maybe not.  Past performance does not guarantee future results, I've heard all of this before and know that there are no guarantees.

However this is an area I'm sure others are far more educated than myself.  Just wondering if any other like-minded people (i.e Mustachians) have looked into this.  Not sure if I believe in largely efficient markets.  I just know a lot of other early retirees have had great success with this strategy.  But like MMM says, it's always good to look for more optimization.  Just wondering if there is some further optimization that is possible by avoiding after-market orders, or avoiding typical dividend dates, or anything like that.

I've got the simple stuff down.  Ready to explore what comes after "Index Fund Investing for Dummies".

And "doing it wrong" is what I fear.  Well, it's been working, so I don't think my simple strategy is necessarily wrong, but I'm just wondering if there is an even slightly better way to do what I am already doing.
Title: Re: Time of month to invest
Post by: Russ on December 05, 2013, 01:16:15 PM
Index Investing for Dummies also happens to be Index Investing for Smarties. That's what's so difficult about it.

On average the best time to invest is "as soon as possible"
Title: Re: Time of month to invest
Post by: RaveOregon on December 05, 2013, 01:26:10 PM
If there is a way to optimize your returns more it will not be as simple as "Invest at 9:45 a.m. on the first Tuesday of the month." But hey if you analyzed the data and found a day and time of the month that appeared to be better it wouldn't hurt you any to try it. After all If Efficient Market Hypothesis is true it won't change anything:

 
Not to be rude, but you're overthinking this and doing it wrong. 

If you believe in largely efficient markets (which, by implication, you largely do, if you're doing index funds and say market timing is no bueno), the prices accurately reflect all the information.  There is no "right" or "better" time.  Just stick to a schedule, whatever that is.   The end.
Title: Re: Time of month to invest
Post by: wakkowarner on December 05, 2013, 01:47:16 PM
But hey if you analyzed the data and found a day and time of the month that appeared to be better it wouldn't hurt you any to try it.

I guess that's what I'm asking.  If there is anyone that knows if this analysis has been done.  What I've found so far is this:
http://www.cbn.com/finance/best-days-month-invest.aspx
http://investorplace.com/2012/05/tips-techniques-for-successful-stock-trading-may/#.UqDlhRAgqf4
http://thepatternsite.com/Seasonality.html

But I'm wondering is this practice sound?  Has this been refuted elsewhere.  I plan on continuing my research but I thought I'd hit up everyone here to see what others thought of these strategies (have they been tried and proven false, has someone had success with this, etc.).
Title: Re: Time of month to invest
Post by: nawhite on December 05, 2013, 01:58:19 PM
I've got the simple stuff down.  Ready to explore what comes after "Index Fund Investing for Dummies".

I feel like I'm in a similar point in my investing education. The next steps I've found:

1. Dollar Cost Averaging vs Dollar Value Averaging vs Lump Sum Investing. If you want to play with timing you might want to read about some of the other strategies. Vanguard just released a report saying that Lump Sum investing is better more of the time than dollar cost averaging. YMMV. Dollar Value Averaging is WAY more complicated than Dollar Cost Averaging and requires some amount of gambling with your expected rate. But... if you choose an expected rate well, and you sell in a tax efficient way, dollar Value averaging can win by around 1% depending on volatility.

2. Tax-Loss/Gain Harvesting. This is hard to do well with Vanguard but can increase your effective yields on the order of 0.5% depending on lots of factors. Check out the Mad Fientist: http://www.madfientist.com/tax-loss-harvesting/

3. Other tax optimizations. This is where an accountant can actually help you. If you go to your accountant 1x per year to do your taxes, you'll think you're getting ripped off because they aren't doing anything turbotax can't do. The benefit with an accountant comes when you see them after April 15th. They can help you with different planning strategies. Things like

"you don't have enough deductions this year to beat the standard deduction, but if you hold off on your property taxes and charitable giving until January and then do the reverse for 2014 and pay for those taxes and donations in December, then you can deduct both in one year and we'll get one year of standard deduction and one year of itemized deductions and save you a couple thousand."


Accountants are awesome for things like this if you ask for them. There are all sorts of advanced topics concerning tax loopholes and you can optimize to your heart's content.

4. Figure out how to make more money

5. Live your life

As for timing which day or what time to invest, you'll get more bang for your buck (on a risk adjusted basis) researching Tax Loss/Gain Harvesting.
Title: Re: Time of month to invest
Post by: matchewed on December 05, 2013, 02:20:52 PM
But hey if you analyzed the data and found a day and time of the month that appeared to be better it wouldn't hurt you any to try it.

I guess that's what I'm asking.  If there is anyone that knows if this analysis has been done.  What I've found so far is this:
http://www.cbn.com/finance/best-days-month-invest.aspx
http://investorplace.com/2012/05/tips-techniques-for-successful-stock-trading-may/#.UqDlhRAgqf4
http://thepatternsite.com/Seasonality.html

But I'm wondering is this practice sound?  Has this been refuted elsewhere.  I plan on continuing my research but I thought I'd hit up everyone here to see what others thought of these strategies (have they been tried and proven false, has someone had success with this, etc.).

Think of it this way. If there were an advantage to a particular day/time and it was publicized on a website, or if it was easy to determine and see a pattern (frankly with computing power and various software abilities this pattern would be easy to spot), how long would this advantage continue to be an advantage? A month? If that even?
Title: Re: Time of month to invest
Post by: dragoncar on December 05, 2013, 02:44:08 PM
I invest on the first day of my period.
Title: Re: Time of month to invest
Post by: AlanStache on December 05, 2013, 03:06:45 PM
I have read some computational articles that looked at using the cycle of moon, rain fall totals and other odd indicators of the market and they found that a few did slightly better than strict buy/hold.  Dont recall reading about day of week or day of month being tested.  Like others have said I would expect the 'get in as quick as you can' strategy to be basically best.  Also remember correlation/causation and that there are penguins that have 'predicted' the supper bowl for the past ten years running but I sure has heck would not call a bookie based on that tip.

You could pull data from yahoo finance on IBM or some sp500 etf and load it into excel to do the analysis, might make a 'fun' evening project.  Would recommend an IPA be with in arms reach.  Be careful that some fluke days dont skew the results.  I would do the test to look at holding for one year and not to today so that every thing is more evenly weighted.
Title: Re: Time of month to invest
Post by: engineerjourney on December 05, 2013, 03:45:51 PM
I invest on the first day of my period.
I literally laughed out loud.  Thank you.
Title: Re: Time of month to invest
Post by: Will on December 05, 2013, 08:35:32 PM
I invest on the first day of my period.

It really is that time of the month.
Title: Re: Time of month to invest
Post by: wtjbatman on December 06, 2013, 08:25:57 AM
My research has revealed that you should only invest on days that end with the letter "y".
Title: Re: Time of month to invest
Post by: the fixer on December 06, 2013, 10:21:21 AM
If it was as simple as "buying on the last Friday of each month performs .1% better over time than buying on random days" then Goldman Sachs would create a new HFT algorithm to exploit it and make more money. That would smooth out the price and this inefficiency would go away from your or my perspective.
Title: Re: Time of month to invest
Post by: FrugalSpendthrift on December 06, 2013, 10:44:53 AM
Is there a better "time" to invest than others.  I'm not talking about market timing, ....

What about time of day?  Is it better to time market orders in the middle of the day, to avoid the crazy fluctuations that can happen at the beginning of the day?

Everything you described sounds like you are trying to time the market, without calling it market timing.  I time my investments to happen on the days I get paid.
Title: Re: Time of month to invest
Post by: AlanStache on December 06, 2013, 10:49:33 AM
Quote
If it was as simple as "buying on the last Friday of each month performs .1% better over time than buying on random days" then Goldman Sachs would create a new HFT algorithm to exploit it and make more money. That would smooth out the price and this inefficiency would go away from your or my perspective.

Not necessarily worth there time (labor) or tying up 'real amounts' of capital for a small advantage, where an individual might be willing to put in the effort.   Still dont think a 'buy on the second Thursday' strategy would be worth it or maybe even profitable, just saying that there could be small advantages us small fish could use that the big guys cant. 

Warren Buffet said something to the effect he could double 1mil in a year but given his large size those opportunities are not available to him or worth the effort.
Title: Re: Time of month to invest
Post by: the fixer on December 06, 2013, 11:04:48 AM
Not necessarily worth there time (labor) or tying up 'real amounts' of capital for a small advantage, where an individual might be willing to put in the effort.
Granted I'm no expert on HFT, but as a software developer my guess is these programs are implemented fundamentally with a cost function. As soon as a share of something goes on sale in the stock exchange, an HFT algorithm would evaluate the expected rate of return it would get by placing a bid on it and selling it at a later projected time. The reverse would be true if it sees an order placed for something it holds. It would be trivial to adjust such a function to assume greater return on some days versus others. It's even possible that a cost function could organically pick up on and exploit this inefficiency without anyone trying to make it do so: for instance suppose a simple cost function "is this asset priced lower than its average price over the last week?" That function should buy more and sell less on certain days of the week if there is an optimal day of the week to purchase something.
Title: Re: Time of month to invest
Post by: gimp on December 06, 2013, 11:27:15 AM
That's actually a very interesting question.

Here's what I would recommend:

- Download stock price info for as far back as you can. DOW index would be fine for this.
- Use the spreadsheet to see what investing (for example) $1000 once a month yields, times 31 or whatever for each day.
- Calculate the means, medians, standard deviations. If you can see a big deviation, that would be your answer. If you see almost no deviation, that would be another answer.
- If there is a big difference for certain days, ask why.
- If there's no difference, just invest on whatever day is convenient. If there is a noticeable difference, you probably won't be able to divine why, and you should therefore just invest on whatever day is convenient.
Title: Re: Time of month to invest
Post by: CommonCents on December 06, 2013, 11:37:04 AM
My dad did a bunch of research on this actually.  Replying to flag this so I remember to ask him what he found.  I think for the most part it was no correlation, but there was something that did have an effect that is eluding me.
Title: Re: Time of month to invest
Post by: AlanStache on December 06, 2013, 11:52:28 AM
@ the fixer

#include AI.h
#include vanguard.h
#include Amazon.h

bool main() {
      /*1*/  StrongAI_v2_63(market_data, &system);
      /*2*/  Broker( system );
      /*3*/  Amazon_Spend( profit );
}

I wish it was that easy :-)
Title: Re: Time of month to invest
Post by: RobertBirnie on December 06, 2013, 12:09:09 PM
I've got the simple stuff down.  Ready to explore what comes after "Index Fund Investing for Dummies".

I feel like I'm in a similar point in my investing education. The next steps I've found:

1. Dollar Cost Averaging vs Dollar Value Averaging vs Lump Sum Investing. If you want to play with timing you might want to read about some of the other strategies. Vanguard just released a report saying that Lump Sum investing is better more of the time than dollar cost averaging. YMMV. Dollar Value Averaging is WAY more complicated than Dollar Cost Averaging and requires some amount of gambling with your expected rate. But... if you choose an expected rate well, and you sell in a tax efficient way, dollar Value averaging can win by around 1% depending on volatility.

I'll worry about Lump Sum investing after I win the lotto. As I have no choice in that though, I subscribe to the idea that dollar cost averaging is better. I get paid twice a month, setup all my bills as close to my payday as possible and as soon as bills clear invest the rest. And doing twice a month also doubles the number of sample points i'm dollar cost averaged against. Then on the 5th and 20th of the month you'll hear me muttering "fall fall fall" under my breath... :)
Title: Re: Time of month to invest
Post by: KingCoin on December 06, 2013, 03:17:00 PM
Not necessarily worth there time (labor) or tying up 'real amounts' of capital for a small advantage, where an individual might be willing to put in the effort.   Still dont think a 'buy on the second Thursday' strategy would be worth it or maybe even profitable, just saying that there could be small advantages us small fish could use that the big guys cant. 

Warren Buffet said something to the effect he could double 1mil in a year but given his large size those opportunities are not available to him or worth the effort.

The stock market is deep and liquid, so it would make an enormous amount of sense for institutional investors to arbitrage it. In fact, they can trade far faster and more cheaply than retail investors, so they'd be in a much better position to profit than Joe Main St. fiddling around on E*Trade with $100k and paying $10/trade.

What Buffet is saying is that the most lucrative opportunities are the least liquid ones. That could be a mispriced micro-cap stock, or some structured note with only $5 million notional outstanding. These opportunities don't move the needle for a large hedge fund or mutual fund, so they don't bother with them. That leaves the door open for a savvy small player to profit.
Title: Re: Time of month to invest
Post by: wakkowarner on December 10, 2013, 10:53:40 AM
Is there a better "time" to invest than others.  I'm not talking about market timing, ....

What about time of day?  Is it better to time market orders in the middle of the day, to avoid the crazy fluctuations that can happen at the beginning of the day?

Everything you described sounds like you are trying to time the market, without calling it market timing.  I time my investments to happen on the days I get paid.
I was wondering when someone would throw this accusation out there.  If investing twice a month, every month, regardless of upturns or downturns (but asking the question if a particular day or time has made any historical difference) is timing the market.... then yes, I am trying to time the market without calling it timing the market.

There are tons of things out there that I, admittedly, don't know.  With the articles that I linked I was just wondering if anyone else has seen any valid studies on this.  I'd like to see what CommonCents dad found.

I can see how day of month could be influenced on when typically get paid, since that would translate to when 401k's typically get funded.  Maybe it doesn't matter that 70% (making up this number) of salaried individuals are paid on the same day of the month.  But it occurred to me when I had to pick the day of the month of when to automatically allot my Roth contribution that I didn't get the same choice for my 401k.  Then I realized that all of the other 170,000+ employees at my company also didn't get that same choice.  And that the same was true (most likely) for millions of employees.  What Wall Street has done about this was unknown to me (hence my question).

Also what about my question regarding time of day?  Or market vs limit orders?  There is some psychology behind all of this.  And has been shown historically, human psychology drives the market more than anything else when it comes to the short term.  Sure I may be a long term investor, but telling a broker (this is regarding market orders) that every two weeks that they can get me X amount of stock for $600 makes me think that they would try to minimize X as much as possible, especially if they get limit orders that say only buy it if it reaches Y price.  Knowing that they had a bunch of orders, couldn't they swing the price one way or another with enough market orders.  Sure it will correct itself by the end of the day/week/month.  But people do day trade.  All of us are influenced by these minute by minute price fluctuations (because at any given minute our buy order could go through).  I'm not really worried about them, I don't think it's worth the stress.  But as my I invest more and more I'd like to maximize my potential.  If there is no possible gain here then woohoo!  Proceed as I have been, otherwise optimize.
Title: Re: Time of month to invest
Post by: the fixer on December 10, 2013, 11:55:17 AM
Regarding 401(k)s, in my experience they at least sometimes don't purchase shares on payday. My contributions would always show up a few days later, with a matching buy date/price. The other tricky part is that buying shares of a mutual fund is not directly a market transaction: the fund's orders don't have to occur at the same time that someone transacts the shares. Large funds also go to great lengths to be sneaky about when they execute transactions because of the profits one could make by timing them.

It's still possible that all the major plan providers build in the same delays, but there is unlikely to be as many trades on the same day as you think.

An easy way to investigate this is to look at trading volume data. If 401(k) plan purchases from salaried employees are having a significant impact on the market, the easiest place to see it is in the daily volume. Without a statistically significant higher trading volume on certain days of the month, I don't see how this effect could influence prices.
Title: Re: Time of month to invest
Post by: wakkowarner on December 10, 2013, 12:15:00 PM
Regarding 401(k)s, in my experience they at least sometimes don't purchase shares on payday. My contributions would always show up a few days later, with a matching buy date/price. The other tricky part is that buying shares of a mutual fund is not directly a market transaction: the fund's orders don't have to occur at the same time that someone transacts the shares. Large funds also go to great lengths to be sneaky about when they execute transactions because of the profits one could make by timing them.

It's still possible that all the major plan providers build in the same delays, but there is unlikely to be as many trades on the same day as you think.

An easy way to investigate this is to look at trading volume data. If 401(k) plan purchases from salaried employees are having a significant impact on the market, the easiest place to see it is in the daily volume. Without a statistically significant higher trading volume on certain days of the month, I don't see how this effect could influence prices.

This sounds very reasonable.  If they are in fact sneaky (which makes sense) to keep others from timing it that would make sense.  Also if they were to buy in chunks.  Though my understanding is that mutual fund buys are only done a the end of day price.  If that's true, then I guess it would only matter if you weren't dealing with mutual funds (which is entirely possible to still do indexes with if you were to deal with ETFs). 

I thought the articles I linked to sounded like BS compared to the other information I have learned.  Seemed like too much of a free lunch to me.  Perhaps it worked, or would have worked, in the past.  Most likely any benefit that might remain regarding this wouldn't be information available to the normal investor like myself.
Title: Re: Time of month to invest
Post by: aclarridge on December 10, 2013, 12:24:55 PM
I think it's important not to buy at the very beginning or very end of the trading session. Spreads are often larger at the beginning of the day especially for illiquid stuff so you'll pay more if you just pay the ask. Other than that it should be pretty much irrelevant.
Title: Re: Time of month to invest
Post by: CommonCents on December 10, 2013, 12:31:33 PM
I can see how day of month could be influenced on when typically get paid, since that would translate to when 401k's typically get funded.  Maybe it doesn't matter that 70% (making up this number) of salaried individuals are paid on the same day of the month.  But it occurred to me when I had to pick the day of the month of when to automatically allot my Roth contribution that I didn't get the same choice for my 401k.  Then I realized that all of the other 170,000+ employees at my company also didn't get that same choice.  And that the same was true (most likely) for millions of employees.  What Wall Street has done about this was unknown to me (hence my question).

Yep, so as said I'd do above, I did ask my dad about his prior research on this and that's what he found out.  A slight uptick at the end/beg of the month due to a bunch of fund managers being required to purchase with the influx of scheduled 401k money from first of the month paydays.  Also he speculates it's from fund managers required reporting on performance and wanting to massage the numbers positively (I'm not sure how that works, and I didn't ask him what he meant because I didn't have time to talk longer just then).  It's not every month, but it happens more often than not.  So if you wanted to go to a lot of work, you could buy near the end of the month and sell a few days into the next.  Not really advisable though.
Title: Re: Time of month to invest
Post by: prestojx on December 28, 2013, 11:12:45 AM
I think it's important not to buy at the very beginning or very end of the trading session. Spreads are often larger at the beginning of the day especially for illiquid stuff so you'll pay more if you just pay the ask. Other than that it should be pretty much irrelevant.

I agree and other than avoiding placing orders overnight, there is really no way to exploit these minute variations and is a complete waste of time.

It is important to understand that there are legions of very smart people in the financial industry looking for some way to achieve a slight edge. Everything that is commonly known and discussed on the internet has already been priced into the market or is impossible to exploit due to the friction of the transactions.

Time is much better invested in learning how to build a diversified, passive, portfolio and how to  minimize risks, costs and taxes.
Title: Re: Time of month to invest
Post by: DrTesla on December 30, 2013, 12:48:26 PM
When watching the stock market all day long for fun, you'll notice more volume and stock price volatility during the morning "Rush" from 9:30 till about 11am. This is when the Rich Investor class play with usually bigger trade volumes. They typically may wake up around 9am, get on their labtop, play for an hour and a half or make one big trade, and then go to lunch and enjoy the rest of their day. Those in the middle class who still work, usually peak their trading at around lunch time or after hours when they are "allowed" to have free time.



Title: Re: Time of month to invest
Post by: fmzip on December 30, 2013, 01:53:53 PM
Foolish, but this is what I do:

Whenever I buy, I buy on a down day. Makes me feel like I got a discount :)