Why should removing and re-adding a canned asset causes different results?
I recently upgraded to 2.15.2 and noticed a disturbing/surprising effect. If I remove a canned asset from one of my portfolios, then re-add the exact same canned asset (in the exact same percentage, of course), my Recommended Annual Consumption changes.
When I remove and re-add the canned asset, the Average Real Return stays the same, but the Relative Risk changes. For example, the old Relative Risk for "Small Cap Stocks" is 8.9%. The new one is 231%. Interestingly enough, the old Relative Risk for "Large cap Stocks" is 6.9%, and the new one for "Large Cap Stock" (notice the difference in capitalization on the word "cap/Cap") is 100%.
I have always only used canned assets.
Is this behavior expected? Do I really need to remove/re-add each canned asset from each portfolio with each upgrade? Or should I be posting this in the "support" section?
RSS
Every time you run the Monte it's taking 500 fresh draws, so your recommended consumption will change. This relative risk display sounds buggy, though. This display has nothing to do with the calculations. But I don't really follow what you are saying is happening, so I'd appreciate your giving me a call at 617 834-2148. best, Larry
The good news is that, other than handling %ages, the UI has nothing to do with the statistical properties of the canned assets. Those statistical properties are embedded in the computation engine and are changed yearly in the February update.
So if the UI displays "odd" statistical parameters for the canned assets, it's a UI bug but it will not affect the computational aspects of the canned assets.
Best,
Dick Munroe
Hi, Please call me at 617 834-2148 so I can understand better what you are seeing. best, Larry
I've opened a support ticket on this. I'll also give you a call. I've recreated all of my portfolios, and now they're stable - the symptoms were a one time event, and may have been unrelated to the recent upgrade.
Please ignore me if I am off-base here. I have only used ESPlanner for a few months now but have seen something very similar discussed before in the forums here.
See if the following thread might explain what you are seeing with your one-time anomaly from removing and re-adding a canned asset in your Retirement Assets account.
http://www.esplanner.com/forum/future-rates-return-monte-carlo-assets
If the other thread is correct and I understand what you were seeing, then ESPlanner may be just working as designed.
Is ESPlanner just updating the canned assets to their newest defaults from the update or is something else happening in this case? It seems that if the asset attributes had been updated between the two versions of ESP being updated, then you would not normally see these changes until you removed and then re-added the asset type. Could this explain the observation?
Best Regards,
John
Our release schedule is organized so that we can put out fully updated tax and financial figures in the 01-Feb release. We grab new data from our sources, do the necessary statistical calculations and plug them into ESPlanner. Typically the differences in the statistical calculations are small year to year so differences in the output are small. This year was an exception due to the melt-down in the financial markets. For example, the Large Cap Stock (an average of data starting in, I believe, 1929) mean rate of return dropped by a full percentage point (from 9 and change to 8 and change). Differences like these are unavoidably visible in your outputs since ESPlanner basically works from "today" onwards. If you compared reports generated in 2008 and reports generated in 2009 using the same data you would see substantial differences in LCS returns.
There are two distinct types of assets in ESPlanner, "canned" and "user defined". The statistical characteristics of the canned assets are buried in the computation engine (that's the component of ESPlanner that does the heavy economic lifting). The user interface only provides %ages of canned assets to the CE and the CE ignores any statistical data that the UI may also provide for these assets. User defined assets are fully controlled by the UI. The CE eats the %ages of these assets, plus the statistical data, massages it appropriately into the form necessary for internal processing and off it goes.
In general (although due to human error this hasn't always been the case) if you don't change assets characteristics (beta, variance, relative risk) or %ages, you shouldn't see much difference in outputs (Monte Carlo is a statistical operation so there will be SOME differences from run to run).
Now this isn't to say that asset definitions are the only thing that affect your financial position. For example, we found a pretty egregious error in most of the state tax routines in the last update (discussed at length elsewhere in the forum). In general, this error should have resulted in calculating a tax burden that was too low, thus allowing you to accumulate or spend assets too aggressively. Fixing these bugs also show up in your financial position, in this case by lowering your standard of living (or disposable income) to account for the increased tax burden.
Best,
Dick Munroe