Consumption jumps way up at 65
In my simulation output of consumption level I get a huge increase at age 65, I assume because the software thinks I'll wait until then to start spending my tax sheltered accounts. The jump is from 40k to 300k. Wow.
In order to smooth this I tried to set an earlier date to start withdraw but the software does not let me do this.
Reading another response in the forum I changed my indebtedness to 100% to see if that would somehow allow the software to smooth my consumption. It did not. Things changed, but I still have a jump from $40k to $150k at age 65.
Any advice on how to smooth my consumption, since that is the point of this exercise?
Thanks,
Jim
RSS
This sure does sound like the usual situation of being income constrained and not allowing much borrowing. However the discrepancy is really much larger than one might expect from "saving too much" alone.
First you might try an age of first withdrawal set to something like 60 (earlier can generate penalties - though earlier is possible with certain restrictions ) even if you plan on retiring later. Also make sure time of last withdrawal is reasonable, say the date you set for end of plan.
Also I'm not sure what you mean by setting indebtedness to 100% - the setting for acceptable debt is a dollar amount not a percent - though it is easy to miss that since everything else on that side of the screen is a %.
All this assumes you are a saving machine with more retirement assets than would be expected from your current and projected income and you could spend more now by borrowing or withdrawing earlier to smooth consumption.
Hi,
Good tip about the borrowing limit. I see that it isn't labeled and I did not know what the units were. I now set it to 500,000 - I assume we're talking about the field labeled "maximum indebtedness" ?
I am a savings machine. I am 49 and want to retire next year. I'm happy to take money out of my retirement accounts early - if that's the way to smooth my consumption who cares about penalties? I prefer not to borrow, but if that's the only way this tool will work, then I'll give it a shot.
I have tried to set "age of first withdrawal" on my retirement account to less than 65, but the spinner control will not let me set it below 65. Any thoughts on that?
With these new settings I still have this huge discontinuity at age 65. Here are some other observations: I get Recommended Annual Savings of -80k (that's negative 80k) a year until I hit 65. On my income trajectories the Recommended line is in the middle of the data points until after the retirement jump up; at that point ALL the data points are above the Recommended line - how can that be?
Here's a link to an image so you can see what I mean
http://www.jimschrempp.com/esplanner/esplanner1.jpg
Thanks for any help,
Jim
[/img]
You are running into a couple of my pet peeves - linked restrictions set on different menus/pages/tabs.
Lfting restriction on age of first Withdrawal.
You can't set age of first withdrawal any sooner than age of last contribution.
Age of last contribution is under Retirement Accounts, tab Contributions. You should also make sure you don't have any contributions in the table to the right earler than last contribution (don't know what happens if you do...) - this one burned me too.
Other thoughts.
Your income trajectory looks weird alright. It appears you have a very nice income in 2007 (earned/wages I'm thinking) and then you retire in 2007/2008 and don't have any benefits income (pension or what not) and are living from non-retirement account money (taxable savings and investments0? Note that anything happening in the current year (2006) is suspect since the program assmes the values you enter are for the whole year not some portion.
I also note that the program is saying you can withdraw ~400,000 per year after age 65 for the next 20 or so years. If you really have that much invested in your retirement accounts I would run ,not walk to a financial planner and pay them a bit of money to sort things out for you - you might also check the number of zeros you put in some of your account balances.
However the program doesn't usually "ramp down" income down like this when you are short of income. It smooths for the limited income interval to some extent, even if low, value. You get a ramp-down even after 65 which is not what the program usually does either. It levels consumption within all intervals that share a common asset base which roughly levels income too. So something you have setup is confusing it.
It is kind of hard to talk about this in a public forum since it is your personal finances :^) so if you want to take it off-line you might be a little more specific about the details. You should probably use "ficticious" numbers.
Ted C.
Jim, I believe Ted has provided the proper answers to most of your questions. The program is working correctly. You have an unusal income and retirement plan, so you are significantly borrowing constrained. This explains the large dissaving prior to withdrawing from your retirement accounts and the large jump in your consumption when you get access to this inocme. Call me at 617 834-2148 if anything remains unclear. best, Larry
Jim, If you tell the program you are going to retire very early, also tell it you'll have a huge retirement account that you can't access until you hit, say, 62, and also tell it you can't borrow, it will do the best it can to smooth your living standard, namely recommend a low spending, but smooth spending level before retirement and a high, but smooth one after retirement. best, Larry
I really need to ask you to call me to go over your situation. My first thought is that you should start taking regular withdrawals from your retirement accounts as soon as you retire -- namely age 50. I've been told that this is legal and does not invoke the 10 percent tax penalty as long as the withdrawals are smooth through time. You'll need to check with your accountant on this. But the program allows you to set the age of first withdrawal before age 59 to deal with exactly your situation. best, Larry (617 834-2148)
Jim,
I just associated this posting with another in the forum under "Taxes". There you say you have about half your assets in taxable accounts. Since your issue here is the big jump at age of first withdrawal (you have that set at 65) I suspect you do not have all your assets correctly entered as "reqular assets" vs "retirement assets". If that is the case, then you probably don't need to worry about figuring out how to withdraw retirement assets before 59 but just use your regular assets as entered correctly.
Check to make sure all assets that are NOT in tax advantaged accounts are listed as regular assets. (Tax advantaged accounts would be regular IRA, Roth IRA, 401k, Keogh, 457, 403b, certain annuities and generally any account where contributionsa reduce your income for taxes)
Sounds like you must have bought a bunch of Google at IPO :lol:
Hi,
I have triple checked my assets and I believe I have correctly entered amounts in both the taxable (checking, mutual funds, bonds) categories. Also noting that I put the bond funds I hold in the mutual funds bucket. I have also triple checked my retirement assets as being in my 401k (a rather small amount) and my IRA (which is a much larger amount as it was a rollover from a previous 401k). I have roughly the same amount in total in the taxable and non-taxable accounts.
I did have the Age At Last Contribution set to the default, even though I had $0 for the contribution level. I set this to be 50 and then the withdrawal spinner let me start withdrawals at 51. Great to be over that problem.
I have left everything else pretty much alone. No home sale, no vacation home. I did not change inflation, etc. I wanted to see how the software would behave before I started screwing around with it.
For the Monte Carlo simulation I created two portfolios. One named "taxable" and one "non taxable". I set these to the current asset allocation structure in those accounts. (The model does not have an intermediate corporate bond bucket so I did make a little deviation from reality.) My allocations are pretty simple in govt bonds, large cap, international, small cap. The allocations I have today were recommended by Financial Engines. I applied the taxable portfolio to the taxable asset pool and non-taxable to the non-taxable asset pool. I set these allocations for this year and extended them to all years in the model. I did not change asset allocations along the way.
Since the software now lets me withdrawal from retirement earlier I also set the borrowing limit back to $0, since I don't really plan to borrow.
Now my observations are:
1. The consumption is smoothed, but much lower than I expected. And about half of what FE says I can spend.
2. Recommended Annual Savings starts at $50k this year, then minus 50k next year when I retire, then minus $5k the next year and a slow downward trend until I am recommended to save *minus* $50k per year when I am 90 years old. What does negative savings mean?
3. From looking at Net Worth and Total Annual Spending the model uses up my entire retirement savings by the time I am 52. How can that possibly be? (Withdrawals start at 51, go to 89, and 100% annuitized with 100% of non-annuitized to be spent) In fact the monte carlo report on Percentile of Distribution of Retirement Account also shows the account value crashing at 52. Perhaps I have misunderstood the settings?
4. The Income Trajectories with Dates along the X-axis now show some years where the "high" point is higher than the "very high" point. Also the Regular Assets Trajectories graph has years where the Very High is the lowest point and the Median is the highest point. Here's a graph: http://www.jimschrempp.com/esplanner/esplanner2.jpg At least the Recommended Trajectory is now within these confused data points. I could be wrong, but it seems like the monte carlo engine is broken. Under any set of assumptions it should behave better. Am I wrong about this?
Thank you all for your comments.
Larry, I don't really see how the software can be working correctly when the monte carlo output has the 95th percentile below the 50th. Are you an employee of ESP? If so, I'd be pleased to chat with you on the phone.
Ted, I don't think I've munged up with any extra zeros. The fact that the software was telling me to spend $0 for the next 15 years really gave me a bang-up post age 65 spending plan. I have used a fee based financial planner for several years. I also use Financial Engines religiously. I am hoping that ESPlanner can become my third point of planning confirmation.
As to Google, I credit my current asset portfolio to the fact that I did *not* invest in Google; read "Cashing in on the American Dream" in 1985; and fired every financial adviser who tried to coach me on how to "beat the market" - they clearly do not know what financial planning is about. If you're younger and willing to be a net saver instead of a net spender, then later life will be good to you too. ;-)
Jim
Well I wish I were younger but no such luck. I'm already retired but did put a fair amount away - my problem was too many years in graduate school not earning much but having a great time :)
BTW,(Dr.) Larry Kotlikoff is a bit more than an employee of ESPlanner - he is the originator of the algorithms used for consumption smoothing and I'm sure much else in the program. Part of the appeal of ESP is the people you get to consult with :)
Negative savings mean the program is suggesting you withdraw from your regular assets - which I'd expect it to do early on. Positive values of savings mean you are putting money into your regular assets. This is different from a negative regular asset balance which means you should be barrowing - I suspect the program no longer is suggesting that.
You can look under the Monte Carlo topic for more on what the various MC reports "mean". The tutorial has a section on the Monte Carlo methods use here that I found helpful but I still have to think hard about what's being reported. It is not a program bug that in some years the 5% is higher than the 95th percentile - the sum over all plan years for the standard of living is what the percentile ranking is based on (correct ESP folks if I'm wrong.).
Your quick synopsis sounds like you have things setup appropriately. In all the cases I've run for my personal scenarios, the withdrawl from retirement funds is constant over time after retirement with any excess over recommended spending going into savings. I'm not sure that is the most tax efficient approach but it doesn't seem to make a lot of difference in my case (did some hand calculations to check).
I've setup ESPlanner without Monte Carlo enabled to use the same "expected average return" as I use in Quicken retirement planner. I use the ESPlanner tax rate average in Quicken and then pretty much set the two to use the same other assumptions. They agree closely enough for me (within 3-4% in yearly consumption and asset values).
I don't know much about Financial Engines but the first place I look to sort out planning differences is in return on asset assumptions. Are you sure the FE return assumptions are the same as the ESP assumptions? A percent or two difference can cause big differences in results.
I have found ESP to provide pretty much the same guidance as Quicken retirement planner, Fidelity's planner (appropriately fudged), and the results my own CFP provides. I find ESP to be my "first of the year" choice and then finagle Quicken to match - Quicken prorates expenses, income, and asset growth per month while ESP is only accurate on yearly boundaries. (another pet peeve...)
Personally I would trust ESP results - why they differ from your FE results is not something I'm really competent to analyize.
Finally, the graph you sent does not look wrong but is pretty much what I'd expect assuming a plan that ends in 2048 or so. ESP will spend every dollar you have by end of plan unless you setup something to prevent that . There are couple of ways to do this latter if you care.
Ted C
Hi Ted,
Thanks for all the comments.
On the MC percentile... I thought that the output of the simulation would be a series of confidence intervals at each year. Essentially at each year the simulation would predict a range of possible results. The 5th percentile value would mean that 5% of the time the model showed less than that value. The 95th percentile would indicate that 95% of the time the model predicted a value of less than that point. In this case I don't think it makes sense for the 5th to ever be higher than the 95th %ile. In other tools I've seen nice curves for the 95th, 75th, 50th, 25th and 5th; the lines might touch, but never cross.
Maybe the ESP folks can educate us?
As to Negative Savings... I see now that the documentation says, "Annual saving is the increase over a year in the level of regular assets. " So a negative savings is to be expected as I spend down my assets. I think my question now becomes, since I want to die broke, why do my savings go positive in the later years? I've sent my file on to Larry for analysis.
As to Financial Engines... It also gives a maximum smoothed income, but it does not give detailed spending plans over each year. However, it does recommend buying and selling specific mutual funds in order to maintain your desired level of risk (which equates to a specific return). I like it because it tells me exactly what to buy and sell. ESP requires me to input my asset allocation - FE tells me what asset allocation to use. But again, FE does not show it over time.
Thanks again for the comments.
Jim
Jim,
We've taken a close look at your case and looking at the reports, you're borrowning constrained until you start withdrawing at age 70 from your retirement accounts, then you get hit with minimum distribution requirements and pay about a gazillion dollars in taxes. Unfortunately, you told the user interface to use 55 for the year of first withdrawal and, for some reason we haven't figured out yet (that's today's job by the way), we're using 70 internally.
This should be pretty easy to straighten out. Expect an update sometime in the next business week.
Best,
Dick Munroe
jim at jimschrempp.com wrote:As to Financial Engines... It also gives a maximum smoothed income, but it does not give detailed spending plans over each year. However, it does recommend buying and selling specific mutual funds in order to maintain your desired level of risk (which equates to a specific return). I like it because it tells me exactly what to buy and sell. ESP requires me to input my asset allocation - FE tells me what asset allocation to use. But again, FE does not show it over time.
Jim,
We're actually discussing, in general terms, building in "risk tolerance" as one of the factors in the Monte Carlo simulation. However, as the number of assets increase, it becomes more and more difficult to actually make suggestions. From a computational aspect it appears that the problem is what we in the software business refer to as order n squared (which means that the amount of time required to solve the problem is proportional to the square of the number of items in the inputs). It would be possible to provide reasonable suggestions for, say, large cap stocks versus bonds, but to provide reasonable suggestions for portfolios containing 13 or 14 different assets could easily get to be hours.
This sort of thing is being done academically but the amount of computes to do it for large numbers of assets is astronomical.
I suspect that FE and other folks aren't actually solving your problem in detail, they're just taking a look at the return, then picking some fund giving a higher return and recommending that. I don't know if FE has any relationships with the various funds. If they do, then their suggestions are automatically suspect due to conflict of interest (same for Morningstar, Fidelity, and any other planner that recommends action on your portfolio)
Interestingly enough the tradition financial wisdom about risk may be flawed as well. Larry wrote an interesting piece for Consumer Reports that deserves some reading and thought.
Sorry for the rambling, just letting you know that portfolio recommends are under discussion, but it's a major piece of work and there are other major pieces of work on our plates that need to get done as well.
Best,
Dick Munroe
Thanks. I'm looking forward to the new release.
I think you are way underestimating what FE does - from what I know.
FE makes recommendations based on efficient portfolio theory. They don't recommend individual stocks, but mutual funds that represent the asset classes they model. They claim that their advice is independent (who doesn't?) but they do offer a subset of their product for free to employees of large companies that use Fidelity (and maybe others) for their 401ks. OTOH, I can tell you that FE has not reliably recommended Fidelity funds for my portfolio.
They claim to recommend funds by monitoring their reported holdings over time - the idea is to make sure that the fund actually invests in the asset class and does not drift over time. They also have a couple of measures of how a fund tracks the overall index for that asset class. I think the idea is that funds which track the class index are better than ones that vary highly from the index - since their asset class correlations are based on these indexes (isn't that how EPT works?).
I remember one time FE suggested that I sell one international fund and buy another. I wondered why. Looking into it I found that the fund I was holding had switched to a 20% cash position. I was "paying" the fund to invest in companies - my cash position was taken care of with real cash. That made me feel that FE was keeping a good watch on the funds. (I remember back in the '80s, when I was using the old Wealth Builder software, that a large cap Fidelity Fund had moved to hold 50% bonds. That's when I first became aware that you have to watch those fund managers pretty closely.)
FE then does a Monte Carlo simulation and recommends a risk level that will meet your personal consumption goals. For myself, I set my consumption goal to what I reasonably need and FE gives me the lowest risk allocation to meet that goal. Once FE started telling me to spend more, I thought I should consider retiring.
Jim
I just updated to 2.8.5 and - as recommended - set my percent of non-annuitized assets to be spent to 0.
I was thrown off at first by a big bulge in consumption after age 80, but looking at it more closely I see that the graph Y axis does not start at 0. When I rescaled the axis manually in Excel the mountain becomes a little wave in the line at age 80.
At first I thought this bug was fixed, but I increased my retirement holdings by about 30% and my consumption line did not change until age 69, even though I said I'd start spending it at age 55. I also note that the model has me die with 12 years of spending in the bank - that's not my plan. As I increase the amount in my retirement holdings my consumption level after age 65 goes up higher and I die with even more in the bank.
I think there's still a bug in here.
Jim
Great. I downloaded an Update today. While the update still says 2.8.5 r1, it appears to be an update and now I die with $0 in the bank.
In addition, my consumption is smoothed. The reports seem to behave as expected and I can now run a few different scenarios.
Thanks,
Jim
Jim, Sorry it took a while to sort out that bug re your pre-age 59 withdrawal of retirement account assets.
best, Larry