For me, I’d like to be able to set AA on all brokerage-type accounts with custom investment growth rate and dividend yield per asset. This is similar to https://changemap.co/projectifi/projectifi/task/5785-manually-entered-portfolio-for-progress/ but a bit different.

Instead of stock tickers, it’d be nice to just list: * Domestic stock * International stock * Domestic bonds * International bonds

or whatever, really… those could be short-text fields. Have it in a table where, for each text field, we can set growth and dividend rates differently, and potentially frequency of dividend payment.

This will help account for the tax drag of bonds in a taxable account as compared to tax-deferred. It would also make it a bit easier to project growth because different assets grow at different rates on average. Right now, we have to average out the whole portfolio.

To piggyback on this idea. Assuming you’ve specified your detailed asset allocations on each account, you’d now be able to derive your global asset allocation, and asset locations (taxable, tax-deferred, etc.)

As a consequence of building this feature, you might then want the ability to model rebalancing via Cash-flow priorities. Currently the app assumes 100% tax-efficient rebalancing just happens, which I think means “the least tax is paid to rebalance above some drift threshold”, but that would imply buy/sell rebalancing is used. Cash-flow rebalancing is probably more common / powerful for most people?

The idea: Create a coordinated set of Cash-flow priorities that let the user specify Asset-allocation and preferred asset-location. For example, you could specify a target global asset allocation, and preferred asset-location, which the app would strive to satisfy across all account types. This could also take into account Cash and Crypto.

I think this may have to be an alternative / advanced set of cash-flow priorities that would not play nice with the existing ones. But for folks who’ve gone to the trouble of setting asset allocations on each account, I would think they’d like the option of specifying a target, and having cash-flows attempt to adjust for drift, since that’s at least part (if not most) of how they will rebalance.

I may be of a different opinion, but I think there’s not a huge amount of value in allowing the user to set particular asset allocations in particular account types. More important is simply the user’s ability to choose whatever overall anticipated growth and dividend percentage he/she chooses in each account. If users can do that (which PL already allows us to do), the AA is somewhat irrelevant for purposes of PL. Sure, if we’re shooting for extreme amounts of user flexibility, AA could play a minor role, but far more important is simply setting the anticipated growth and dividend in each account based on whatever AA someone has chosen, combined with their own anticipations of how the market will do in the future. (Same with inflation, BTW Ultimately, I think there are probably more important edits that could be made by the PL team than user-specified AAs.

I tend to agree there are more important things right now. The issue with account-level growth rates is they make setting a glide path pretty cumbersome when you have many accounts. But if there were a way to define reusable account growth rate profiles (using the advanced editor) and apply the same profile to multiple accounts you’d have a good compromise.

I’m wondering if there is any plan to incorporate historical data from other asset classes into the Monte Carlo simulation. Is it true that it only uses S&P 500 and bond returns since 1928? I know that other asset class data isn’t as readily available and would limit the simulation to maybe the last 50 years, but international stocks in particular would be helpful. Not sure if there is also good data for small cap value, for example, which has historically had higher returns, but has been more risky.

This is a good point. As a natural consequence of adding more asset classes, PL would need to find historical data to match the asset class, otherwise, it wouldn’t work in monte carlo.

Definitely like this suggestion. Currently it’s possible to add a custom “fixed” return for each account, but it would be much easier for me as a user to just specify the asset allocation for each account and have PL calculate the return instead of having to manually calculate it for each account based on the ratio of stocks/bonds. PL could include default assets of stocks, bonds, and cash (with historical returns) and allow the user to add more assets and specify their own returns for those assets (since historical data may be sparse for some).

I used to be in favor of this, but I’m now less sure of the benefit, and it may open a big can of worms and lead to feature bloat.

Adding my two cents to this one. I personally think this is one of the biggest areas where PL is coming up short. I think the target audience for the long-term PL user is people that spend a lot of time thinking and planning their financial future. I know I do and my asset allocation and expected growth rates by asset class are central to my plan. I would very much like something similar to what Mathew described above. I think there should be say five categories that the user defines at the plan level. Default them to something like US stocks, Ex-US stocks, Bonds and Money. You could then default them to a standard 60/40 allocation and conservative default growth rates with historical standard deviations for each one. Users could then change their category labels and specify their expected growth rate and standard deviation for each group. I would have 4 or 5 columns for the values defaulting to 10 year increments out from the plan start but allowing the user to change the years or use milestones (i.e., 10 years after retirement etc.). Then at the account level the user would specify the allocation based on these defined groups. The user should also have a similar 4 column 10-year type grid to specify how the allocation changes over time. I would think with this structure it would be fairly simple for PL to calculate the return to use for each account by year then do the normal calculations. When I say simple, I mean for Kyle, definitely not for me. This model would also allow for monte carlo simulations to be run using the user defined growth rates.

I think a very useful side effect of this feature is the ability to classify assets as FI and non-FI.

By way of example; I have a house, and a live it it. It in now way affect my FI number from an asset allocation point of view.

However, if I plan to sell it, then the cash it turns in to DOES impact my FI number. If I then funnel some of this cash into another home, that portion is now non-FI again.

I think the current system approaches this idea with the concept of liquid and illiquid, but my home and my stocks are both illiquid, but I want them considered differently for calculating my “Net worth required for FI”.

From my reading of the feature request, I wouldn’t expect this ability you seek as part of its implementation.

I know in Discord some others have requested the ability to mark things as “not included for FI”, but since FI really only exists as a Milestone that you can define however you like, I don’t believe there’s a special way to mark things up as “for FI” and not for “FI”.

The simplest thing to do today is to just not include things that aren’t in your FI plan in Projection Lab at all. However, the example of your house is a tough one to reason about. Whether or not you DO sell your house, the potential to sell it, always exists. But as it is an illiquid asset PL would never sell it automatically. So you perhaps you can use the illiquid vs liquid definition to get what you want. But then your comment about your stocks also being illiquid is head scratcher for me as well.

You can make certain accounts illiquid (and put an end date on that illiquid status), this can give you more fine-grained control, such that PL will tell you if your plan is likely to end in bankruptcy while still respecting whatever vesting rules or holding requirements you may want to impose.

Interesting ideas though. Feel free to open up a discussion on the discord, I’d like to hear more about what you’re doing with the tool.

jumping in to add to the initial request: one place/reason I see this as very important is because of how Roth conversions are handled. I got my accounts all set up the way I have/like them where all of my bonds are in pre-tax at 50% and Roth is 100% equities. In my case that turns out to be 25% bonds for my whole portfolio.

Then bring in Roth conversions. Next thing I know I have finished converting everything to Roth and boom I’m at 100% equities. Well THAT’s not real. Sure I could play around with ramping my Roth’s bond % over time but it’s not practical try to twist two knobs at once (aka Roth conversion levels and Bond % ramps).

I set up a dummy Roth account with $0 and managed to get PL to do the conversions into there; and that was set at 50% so that my pre-tax to Roth conversions maintained their relative ratio but it still was a bit of a hack that did end up requiring me to play around with the bond ramp to maintain my portfolio bond ratio; but of course THAT was dynamic based on living expenses, etc.

So bottom line, just another vote for a means to control 1) portfolio bond allocation and 2) per account allocation preference. So, “only put bonds into account X until that’s not enough to maintain the portfolio allocation and then put bonds into account y; rinse repeat”

How about using providing something similar to what ‘Portfolio Visualizer’ provides. Has anyone used this service??

I use Portfolio Visualizer alongside ProjectionLab. I feel like they serve different purposes and don’t see the need for that information within PL.

This request is similar to one I just made. I didn’t see this request before I made mine. One important thing nobody mentioned is that the correlation between asset classes is very important for accurate projections. If assets have low correlation, they can reduce the volatility of the portfolio even if the asset has a high volatility itself. www.paulmerriman.com/portfolios-2023-updates

I agree this is an important feature to add. I don’t think support for individual tickers is necessary but being able to specify asset classes and their percentages is critical to moving PL closer to being a complete modeling solution for the individual investor. At present, I need to use PL in conjunction with other tools that do provide historical returns data for asset classes.

Being able to specify asset allocation for each account rather than for just the whole portfolio would also create the possibility for future features in PL like modeling drawdowns from accounts in different sequences or optimizing asset location.

Ideally popular alternative investments like gold, commodities, REITs could be included. As a reference, the tools at Portfolio Visualizer have a well fleshed out list of 38 asset classes across 4 categories: US equity, international equity, fixed income and alternatives.

Agree this would be one of the most important features to have. I’m not clear/convinced simply using one blended portfolio mean/sd allows for meaningful Monte Carlo modelling. As mentioned by Gary Kline above correlation between asset classes is very important for accurate projections.

I’ve been considering purchasing the Lifetime license, but this feature being missing is the only thing that’s making me hesitate as other software products offer setting up your asset allocation by account and for the asset allocation to be projected over time

I can’t add anything new to the conversation. I just want to add another user request to include the ability to allocate stocks/bonds on the account level.

I think this would be helpful. Many have significantly different allocations for tax-deferred accounts vs investment accounts given the different tax treatments. I’ve tried using the custom return profile for different accounts, but these return customizations are really focused on equity not bonds. This is a bit of an issue on estimating RMDs, marginal tax rate/IRMAA when trying to do more granular Roth conversion projections.

I’ll add a couple more notes on how I would use this feature. I may or may not explore Projection lab with out it, as it seems like it will be very manual. I have some investments in debt funds which are dividend bearing, but have a different risk profile than bonds, and also have some investments in a qualified opportunity zone fund. The QOZ fund adds more complexitiy, as appreciation is not taxable, but distributions (dividends) may be taxable.