I repeat the chart of outcomes from that post below. Note that the worst case outcome was a $37,333 loss, the best case was a $133,287 gain, and the median profit from this strategy would have been $26,937. (If you haven't read Investing the Mortgage, I suggest that you do so first. It explains the analysis in more detail.)
The first difference is that we can't really borrow money and invest it in the market — not all of it, anyway — because we have to start paying it back almost immediately from our investments. If we borrow a 4% fixed rate thirty-year mortgage of $100,000, the payments will be about $5,724 a year and we will reduce our portfolio by that amount annually. The average amount of the $100,000 that would remain invested each year over a ten-year period would be a little more than $74,000. With less money invested, of course, our portfolio will grow more slowly.
When we pay the mortgage from future paychecks, as I modeled in the last post, we have the luxury of leaving all of the borrowed funds in the market for the entire 10-year period. If the market goes up, as we hope, holding more stocks will earn more money.
The second difference, which may be less obvious, is that spending from a stock portfolio over time, to pay the mortgage or anything else, creates sequence of returns risk. (See my posts on the topic if this is unfamiliar to you.)
Here's an example.
The following table shows the historical sequence of returns a 50/50 portfolio would have experienced from 1981 through 1990. The second row shows the least advantageous sequence of those same returns (sorted smallest to largest) for an investor spending down a stock and bond portfolio. The third row shows the best case sequence for this investor (returns sorted highest to lowest).
In all three scenarios, the compound growth rate is the same, 8.7%, and that would be our return if we didn't sell stocks along the way to pay the mortgage. Without portfolio spending, there is no sequence of returns risk. The terminal portfolio value will be the same regardless of the order of the returns.
However, if we assume that the retiree has a portfolio valued at $100,000 and is spending say, 4% of the initial portfolio value each year ($4,000), then her portfolio value at the end of this 10-year period would be different in all three cases. The historical value of the portfolio would be $168,470, the best possible outcome would have been $184,820, and the worst $147,988.
This range of values that results from reordering the returns is sequence of returns (SOR) risk. It isn't "good" risk. It can't be diversified away and we aren't compensated for it by the market. And like SOR risk in retirement portfolios, early losses have a disproportionately bad impact on the mortgage-to-invest strategy outcome.
The combined impact of reducing stock exposure and adding sequence of return risk can be fairly dramatic. I altered the model from my last post such that mortgage payments are made each year by selling stocks and bonds from the portfolio, as a retiree might do, instead of paying them from salary and only selling stocks at the end of ten years, as someone still working would.
In the following chart, I compare the historical outcomes of the mortgage-to-invest strategy using a $100,000 mortgage assuming the investor is still working and paying the mortgage from salary (the blue columns from my last post) with the historical outcomes assuming a retiree pays the mortgage by spending down a portfolio of stocks and bonds (red columns).
As you can see from the chart above, mortgage-to-invest can be a fairly risky strategy. You can lose a lot of money even when you pay the mortgage from salary. Even if you're still working, a $100,000 bet could have resulted in a profit of $133,287 if you executed it from 1989 to 1998, or a loss of $37,733 if you employed it from 2000 to 2009. That's pretty risky.
After you retire and begin selling stocks to pay the bills, the outcomes are even worse. You lose more in worst cases and your expected return on this strategy dropped 65% over the studied period.
The analysis is pretty much the same, by the way, if we inherit $100,000 and decide to invest it rather than pay down the mortgage, or if we simply decide to continue to hold a stock portfolio and owe a mortgage simultaneously.
A couple of observations. First, many people will find it difficult, before or after retirement, to pay off the mortgage by selling their investments. Maintaining some reserves and liquidity are important, too. We can argue that if paying off the mortgage requires converting too much of your wealth into illiquid home equity that you should consider downsizing, but life isn't always that neat. Paying off the mortgage isn't the right answer for everyone.
Second, for many households, mortgage-to-invest isn't an intentionally chosen strategy, but one that simply developed over time. You bought a house with a mortgage. You invested in a 401(k). You retired and began paying the mortgage and other bills with portfolio spending. Your risk increased unnoticed.
In either case, it is important to understand the risks and rewards and at least consider the alternatives of downsizing or paying off the mortgage if those are options for you. The analysis isn't as simple as "I have a 4% mortgage and I can earn 8% on my investments."
Regardless, as I mentioned in Investing the Mortgage, this strategy increases the chances both before and after retirement that you will lose your home to foreclosure risk, which, in my opinion, trumps any other home financing risk.
I have another basic concern with this strategy after retirement. While it makes perfect sense to borrow a mortgage when we are young, expecting to pay it back with future job earnings, it is a riskier proposition to borrow a mortgage and expect to pay it back with future stock market earnings.
I often point out that our finances change significantly after we retire and we can't view them through the same set of guidelines as before. This is a prime example.
In my next post, I'll discuss when you might want to make this bet.