Real Return of Monte Carlo Cash Asset Class

When I go to Planning Method -> Monte Carlo -> Build Portfolios tab and select the Cash asset class, it shows a mean real return of -3.01%. I've always just assumed this is because Cash earns practically nothing these days and so that would explain the negative real return assuming around a 3% inflation rate.

But I noticed there is a Hint that says you can right click the asset class to get additional info. So I did that on the Cash asset class and saw the returns for Cash each year since 1926. For 2014, the return was -0.76%. For 2007 it was -4.08%. For 1980, it was -13.31% !!

-13.31% ???

Now I am not understanding this. Can you explain?? Is it saying if you had your cash in a regular savings account it would earn 13% below the inflation rate (for 1980)?

One reason I am interested is that although your basic savings account yields next to nothing, you can get close to 1% at internet banks. And my understanding is that the inflation rate has been around 1% in recent years. So I am thinking that if you have your cash in an internet bank, should the real return be closer to 0% ??




dan royer's picture

Yes, you are thinking right.

Inflation in 1980 was at 13.5 on average for the year.

In 2014 it was 1.6 and this year, 2015, it's been negative for a few months, 0, but not above .5 I don't think. So, yes, if we had .8% inflation in 2015 and you got .8% from your Capitol One 360 account, this would be a real return of 0%.

What I've been reading is 10-20 year inflation forecasts around 2-2.3%. Historical is around 3% or perhaps 3.2%.

But the Federal Reserve is targeting 2% and they seem to have a hard time getting there. It could be awhile I suppose before we see 2%.

Thanks Dan. But where I am confused is it appears the MC Cash returns assume there is hardly nothing earned on cash in any year whether its 1980 or 2014. But this is not true historically. For example, at it shows that taxable money market funds earned 12.68% in 1980 and 3 month CDs 13.07%. So why doesn't the Monte Carlo Cash asset class reflect this ?

The reason I ask is all portfolios are going to have Cash, of course, particularly in Regular Assets because at various times one may need a large percentage of Cash there. But with the seemingly unrealistically low negative returns shown for Cash, this distorts the overall returns of a MC portfolio if one is going to try to accurately represent the Cash allocation. To get around it, I have used Short Term Govt Bonds which shows a real return of 0.57%. If one were to use internet banks this is probably much closer than the -3 % Cash currently shows.

I just thought I'd bring this up, though, because it seems to be wrong. I don't know what your provider of this data is using for Cash returns. Do you think they are just assuming cash returns nothing regardless of the year? If yes, isn't this misleading?



Hi Tom,

The return data for cash is from Ibbotson. They have published data on returns (real and nominal) for multiple asset classes.


dan royer's picture

I believe what is being used is the average real rate of return over the years. The program creates random runs and uses rates relative to the upside and downside possibilities. Those rates for particular years are used to calculate that variation. Although there may be years when cash beats inflation, my guess is that on average the real return on cash is negative.

Remember too that you are setting the inflation rate in the program. Inflation is assumed to be constant at 3% or whatever you have indicated in Assumptions.

I'm no expert on Monte Carlo or for that matter investments. But I do recall having a money market fund in 1980 and getting high returns. I don't know if they were over or under inflation at the time, but I suspect that the program is right in looking at long-term averages.

I don't know how the MC works exactly in terms of the stochastic or the analysis. I think the assumption is the average cash return regardless of the year but that the randomness--the upside and downside--generates a range of possible returns for a given year where the inflation rate is held constant by your entry in Assumptions.

I'm generally not very comfortable with MC analysis because of the reliance on the historical data. Even when that data is substantial--say 75 years--I'm not ready to assume the next 75 years are going to be like that. And much of the data may cover only ten years or 15. Is the historical data for REITS reliable going forward? And I'm more concerned about the next ten years than I am the next 75 since I don't have 75 years to appreciate the historical average. I use nominal returns and bracket everything from extremely conservative to optimistic.

My answer may not be very satisfying and I'm happy to direct Kotlikoff to this question if you like. :)

dan royer's picture

This might be interesting to you Tom.

If that link doesn't work, just backtest a portfolio using Cash position.

In 1979 we see inflation at 13.29 and a Money Market return of 10.38% producing a real return of -2.57. Looking down the column we see some positive real returns, but it appears that the average real return is under 1%.

Thanks Dan. I am not an expert either, just a user who tries to make sense of things that seem out of whack now and then.

That link was interesting. Thanks. To a novice like me it seems to contradict the MC Cash Asset real return value. The inflation adjusted CAGR in that test (when you hover over the little icon next to the 5.04%) number is 0.86%, and when you click on the Annual Returns tab and look at the inflation adjusted returns I counted 26 of the 43 years as having positive numbers. I didn't average them, but it I doubt if it would be negative. Of course, that only goes back to 1972.

Your comments on Monte Carlo are interesting. It seems the MC advocates have declined since I bought ESP back in 2011. Back then there was so much discussion and now I see very little. I wonder if the general consensus is that we are in a new normal and historical data is not as relevant or pertinent.

Anyway, I just hope that something isn't amiss with the Cash return number being used.


dan royer's picture

I will double check with Darryl on that dataset.

Yes, I'm a little skeptical about the MC. There's too much data between the inputs and the results that I can't see. I'd rather just ask: what is my discretionary spending if I average 5% nominal? How about 4%? How about 6%? I'm reasonably confident in those sort of numbers because of my very conservative asset allocation. The MC does provide that distribution of living standard over time--I can see the upside and downside possibility--but I live year-to-year and I reenter my balances each year as we should. Perhaps that doesn't hold water, but I just feel better specifying my nominal return.

What I have done is take my retirement account balance at retirement age per ESPlanner and then go over to POrtfolio Visualizer where I linked to above, and then use the Monte Carlo there. Enter your balance at retirement age, enter per ESPlanner the amount of inflation-adjusted withdrawals, choose your asset classes and a few other parameters, and run the MC simulation there and see how your assets hold up. This does not go down to the spending lever per adult the way that ESPlanner does, but it gives you a lot of transparent control, and perhaps it accomplishes the same thing you are after--the assurance that your withdrawal rate is reasonable.

Here's an example for what it's worth:

We use cookies to deliver the best user experience and improve our site.