Investing lump sums

[ALongRoad]

I currently have a job that is VERY bonus heavy (think 5-8 times base salary). I spend almost my full base salary monthly on mortgage and other expense so I don't have much free cash to invest outside of the bonus.

Question: should I dump the whole bonus into the market when received or put 1/12th in each month to average the cost over the year? What would you do? Thanks!

[SeattleCPA]

I think you'd want first to read a book that explains asset allocation benefits and costs... E.g., "Unconventional Success" by David Swensen.

And then, assuming you're good on the asset allocation stuff, I'd think either lump sum or dollar cost averaging is fine. Again, though, I base this ambivalence on assumption that you've got a good, appropriate asset allocation plan in place.

[MDM]

Question: should I dump the whole bonus into the market when received or put 1/12th in each month to average the cost over the year? What would you do? Thanks!

Either might be better.  No way to know in advance.

See http://www.schwab.com/public/schwab/nn/articles/Does-Market-Timing-Work for historical results.  In short, "all at once" has been statistically better - but that doesn't mean always.

[GreatLaker]

A study done by Vanguard in 2012 concluded that investing a lump sum immediately outperformed dollar-cost averaging 67% of the time. But if it lets you sleep better by spreading out the investments over months, by all means do so. The important thing if you DCA is to have a specific plan to invest the funds so you don't end up months or years later with the money still in cash. It also depends on your timeline for accessing the funds. If it is retirement money not to be used for decades then current market fluctuations are not likely to make a huge difference in the long-term results.

Here is a link to the study:
https://pressroom.vanguard.com/nonindexed/7.23.2012_Dollar-cost_Averaging.pdf

And some comments on it by Portfolio Solutions:
https://portfoliosolutions.com/latest-learnings/topic-guides/what-you-need-know-about-dollar-cost-averaging

[nereo]

Either Dollar-Cost-Averaging or Lump-sum will work fine, and MDM/GreatLaker have already covered how lump-sum typically (though not always) wins out.

Personally I would lump-sum and forget.  I've found once I pull the trigger I can stop thinking about it.  If I try to DCA over 12 monhts I find myself monitoring weekly changes in the share price, which is not something I can control and winds up causing more stress and wastes more time.

YMMV

[maizefolk]

I agree with what everyone before me has said. The average outcome for a lump sum investment will tend to be higher than the average outcome from dollar cost averaging over the course of a year.

The reason people DCA is that their preference to avoid losing money is greater than their desire to make the maximum amount of money possible, and dollar cost averaging reduces the variability of their outcome. Variability is hard to visualize from just medians or means, so here is a histogram of possible outcomes from either lump sum investing $12,000 in stock market and waiting a year, or investing $1,000 month for 12 months for 1,725 twelve month periods between 1871 and the present



As you can see, the distribution of outcomes for lump sum investing is broader with both more great outcomes, and more terrible outcomes.  But on the other hand, in the median scenario a lump sum of $12,000 grew to $13,037 over the course of a year, and twelve months of $1,000 month investing was only worth $12,592 at the end of the year.

[GreatLaker]

I received a lump sum in Aug 2015 (about 25% of my total portfolio at the time) and invested it in my balanced portfolio in one lump sum. The market went through a lot of volatility in the 2nd half of the year, and was generally down through Feb 2016. That hurt my 2015 returns, but by mid-2016 I was back above water.

In May of 2016 I also received another large lump sum, and again invested it all in one go. That either took a lot of guts or a lot of stupidity. Especially when the Brexit vote happened shortly after. But the market performance in 2nd half 2016 was strong, so the lump sum approach helped me. Despite Brexit, Trumpit, rising rates, and a lot of global uncertainty I am definitely ahead through lump sum investments.

Both of those were stressful, but maybe trying to spread it out over 24 months, making buys every month amid all the market and global turmoil would have been as bad or worse.

IMO the important point is have an Investment Policy Statement you review regularly and follow, avoid impulses and market timing, diversify and minimize costs for a long term winning strategy.

[Radagast]

For your case, 100% lump sum. If you are doing it every year you will basically be dollar cost averaging anyway. DCA will add unnecessary complications and opportunities to waste the money, in addition to having lower expected returns. Really the only time lump sum vs. DCA should even come up is for "I won $100,000,000 in the lottery!!! What do I do???" situations, where the lump sum is a significant amount of all the money you will ever earn. For ordinary repeating income invest ASAP.

[Retire-Canada]

Question: should I dump the whole bonus into the market when received or put 1/12th in each month to average the cost over the year? What would you do? Thanks!

All at once.

[seattlecyclone]

I agree with what everyone before me has said. The average outcome for a lump sum investment will tend to be higher than the average outcome from dollar cost averaging over the course of a year.

The reason people DCA is that their preference to avoid losing money is greater than their desire to make the maximum amount of money possible, and dollar cost averaging reduces the variability of their outcome. Variability is hard to visualize from just medians or means, so here is a histogram of possible outcomes from either lump sum investing $12,000 in stock market and waiting a year, or investing $1,000 month for 12 months for 1,725 twelve month periods between 1871 and the present



As you can see, the distribution of outcomes for lump sum investing is broader with both more great outcomes, and more terrible outcomes.  But on the other hand, in the median scenario a lump sum of $12,000 grew to $13,037 over the course of a year, and twelve months of $1,000 month investing was only worth $12,592 at the end of the year.

Nice histogram. It shows the range of possibilities for a single lump sum. If you're only having one lump sum of this magnitude for your whole life, it can be perfectly defensible to do the dollar-cost averaging thing. Even though the median result on the red curve is a bit lower, the distribution is much tighter so you're getting more of a sure thing that way.

However this won't be the only such bonus you ever receive. You will likely be getting one each year until you retire. The central limit theorem tells us that the aggregate distribution of independent random events will approach a normal curve centered around the average of the individual events. Furthermore, the more trials you have, the tighter the standard distribution will be on this normal curve.

In other words, it's reasonably likely for dollar-cost averaging to be better than investing all at once if you have only one bonus. If you have two bonuses, the odds of doing better with dollar-cost averaging go down a bit. If you have ten bonuses, it's nearly impossible for dollar-cost averaging each bonus to turn out better than lump-sum investing each one.

You can apply this principle more broadly across your life. Anytime you have a chance to make a bet where the odds are in your favor and the cost of losing won't be too terrible, take the bet. Skipping the extended warranty or increasing your insurance deductible a bit will certainly turn out to be a bad decision once in a while, but over the course of a lifetime you'll do better by making a habit out of doing these things.

[maizefolk]

@seattlecyclone

That's a good point, and one I hadn't considered at all.

Yes, looking at, say, a decades worth of bonuses the distribution of outcomes would substantially reduce the standard deviations of both the lump sum and dollar cost averaging strategies, which would definitely favor investing the whole bonus as soon as it arrives (essentially dollar cost averaging on a yearly scale instead of a monthly one) instead of spreading it out over the course of the year. I'd like to rerun the code to look at how much that would shift the shape of the two distributions, but am on the wrong computer this morning.

[nereo]

@maizeman - awesome graph. 

@seattlecyclone - learned something new today re: "Roth'. I'll stop capitalizing the last three letters. Thanks.

[SeattleCPA]

@seattlecyclone

That's a good point, and one I hadn't considered at all.

Yes, looking at, say, a decades worth of bonuses the distribution of outcomes would substantially reduce the standard deviations of both the lump sum and dollar cost averaging strategies, which would definitely favor investing the whole bonus as soon as it arrives (essentially dollar cost averaging on a yearly scale instead of a monthly one) instead of spreading it out over the course of the year. I'd like to rerun the code to look at how much that would shift the shape of the two distributions, but am on the wrong computer this morning.

I think another, easy way and quite intuitive way to look at the question is this: If OP has (say) $10,000 in cash and he's trying to decide whether or not to invest in market and I have $10,000 in market and am trying to decide whether or not to liquidate to cash, those are essentially the same decisions.

Deciding to stay in cash or return to cash are both attempts to time the market, aren't they?

[NoStacheOhio]

@seattlecyclone

That's a good point, and one I hadn't considered at all.

Yes, looking at, say, a decades worth of bonuses the distribution of outcomes would substantially reduce the standard deviations of both the lump sum and dollar cost averaging strategies, which would definitely favor investing the whole bonus as soon as it arrives (essentially dollar cost averaging on a yearly scale instead of a monthly one) instead of spreading it out over the course of the year. I'd like to rerun the code to look at how much that would shift the shape of the two distributions, but am on the wrong computer this morning.

This was my first thought. Investing regularly-paid (or even irregular, but predictable) bonuses in when you receive the money isn't the same as lump-summing a windfall. It's basically what all the rest of us do with our regular paychecks; we invest as soon as the money becomes available, and our income is spread out over time.