Six months after purchasing our home in December, I was already in the process of refinancing. This is highly unusual. How did we decide the time was right to refinance and whether it made sense financially? What do the numbers look like? Read on to find out…
The Setup
On Tuesdays at work, I have lunch with a former coworker; we’ll call him Dave. We generally catch up on life, comment on how things would be if we ran things, and entertain each other with life’s dramas.
Near the end of May, Dave mentioned to me that he had refinanced his newly purchased home. I was highly skeptical of whether this was a wise move, but Dave is a pretty smart guy who’s opinion I trust and respect. After asking for more details, it appeared sound and I became intrigued.
Being a numbers guy with a finance background, I keep much closer tabs on interest rates and other financial data and trends then most people. Even so, I had missed a solid drop in rates. I simply wasn’t expecting to see such a dramatic change in such a short period. Fortune favors the vigilant!
The Numbers
As a general rule of thumb, rates need to drop at least a full percentage point for it to be worth refinancing. Every time you close on a loan, you have to pay transaction costs and refinancing takes some time and effort, so you need to make back what you pay as well as save enough money to make it worth it on top of that.
I did my research and it appeared rates had dropped nearly a percentage point on a 30-year term, and even further on a 20-year term. When we originally purchased our house, we considered going for a 20-year loan but went for a 30-year instead to be conservative in case something happened that impacted our income.
In the past six months, we had paid down 10% of our loan and were making an extra payment of $1,000/month in a race to get rid of PMI (Private Mortgage Insurance – extra payments you have to make until you have 20% equity). Mrs. Wallet also lost her job in April, and we were still comfortably paying the required bill plus the extra grand each month, so we were more comfortable with how the cash flows worked out for us.
Our original loan was at 4.5%, which was a good rate at the time. Now, we were being offered a 30-year at 3.75% or a 20-year at 3.375%. The opportunity to switch to a 20-year term and knock ten years minimum off our loan while keeping payments roughly the same and reducing the amount of interest we pay per month by over $300 was too good to pass up.
While the payment amount went up some, PMI was cut in half so it only increased our required payments by ~$30/month with the substantially lower interest rate. In reality, the economics changed even further, as the combination of cutting our PMI in half and reducing our interest rate reduced the financial benefit of paying the loan down early to get rid of PMI. In layman’s terms, while our required payment would go up by $30/month, our actual payment would be reduced by ~$1,000/month, as it now made much more sense to invest that money instead, and we’d still be paying less interest and less PMI.
As I mentioned earlier, there is a transactional cost to refinancing. At a very high level, it looked like refinancing would cost me roughly $1,000 and ultimately save me at least $70,000. Sounds pretty worth it, right? Let’s take a closer look.
Modeling results
Most people would probably be satisfied just knowing they were saving money, but I’m not most people! I already have models of both a top-down yearly view of my mortgage versus renting as well as a bottoms-up monthly model of the cash flows I expect, so it was easy to copy the latter and modify it for both the 30-year and 20-year refinance scenarios for comparison.
Part of the reason I did this is that summary statistics can be misleading. For example, in Statistics there is an example called Anscombe’s quartet, where four datasets have nearly identical descriptive statistics, yet have strikingly different distributions and appear very different graphically.
In this case, I modeled the scenario first that I kept extra payments flat until PMI was paid off. In the 30-year scenario, I would save less money, but more of it would occur upfront due to the lower payments. In the 20-year scenario, I would save more money, but the difference in cash flows would be negative until the last 10-13 years, where all the savings would occur.
The timing of the cash flows is extremely important for several reasons. First, there’s the time value of money, one of the foundations of modern finance. Would you rather have $1,000 today, tomorrow, or in a year? If I get $1,000 today, I can invest it and it could be worth more than $1,000 in a year’s time, whereas $1,000 in a year isn’t even worth $1,000 in today’s dollars due to inflation (you can’t buy as much with it – more intuitive if you think out 10-20+ years).
Second, this could change the economics significantly depending on what we do with the house. What if we sell in five years, before we realize any savings? Sure, we’ll have more equity, so we may get a higher return on our investment in the house, but is that superior to the return we would make investing the difference elsewhere?
Currently, our plan is to just keep the house, moving out after renovating everything and renting the main house as well so that we generate positive cash flow to support us living elsewhere and let our tenants pay the mortgage. At that point, we’d either go back to renting or buy another house, depending on where we’re at, the current economics, and how much fun we had being homeowners the first go-round.
This is important because after PMI goes away, it’s more like $100/month more that we’ll be paying, so rental income has a higher bar to clear prior to the property generating positive cash flow. It’s also interesting to consider because we’ve been building sweat equity in the property (improving it’s value by doing work cheaply on our own to create value and thus margin), which won’t be realized until we either sell or refinance again.
Diving Deeper
After modeling everything out on a monthly basis, I came back with IRRs (Internal Rates of Return – similar to modeling what return or interest you’d have to receive on an annual basis to earn a similar return) in the 350-1400% range. Ludicrously good, if not too surprising when you’re only paying around $1,000 to save upwards of seventy grand. Scale comes into play when you start getting really high percentages. We’re talking less than $100,000 here.
I prefer analyzing IRRs over NPVs (Net Present Value – a measure of how much economic value a project creates over time in comparison to an alternate investment in today’s dollars) as it requires less assumptions, although I examined both and both are useful. When using NPV you have to make assumptions about discount rates (the rate you’d earn on an alternative investment or your cost of capital), so it’s helpful to compare with a range of rates. No one can predict the future.
Ultimately, it came down to this: we were going to save a lot of money either way, but we had to decide whether it was more important to save less money now or more money in the future. Ultimately, we decided to go with saving more money overall, as we (and likely future rents) could easily support the small difference in payments. I really don’t like paying interest I don’t have to.
Let’s take a look at what this all looks like!
| Cash Flow | 30 yr, no extra | 30 yr | 20 yr, no extra | 20 yr |
| payback | 0.07 | 0.18 | 0.09 | 0.21 |
| IRR | 1,372% | 468% | 1,080% | 364% |
| NPV | $36,391 | $21,331 | $23,906 | $11,612 |
| SUM | $34,126 | $17,910 | $85,005 | $71,338 |
First off, I’ve provided you with some summary statistics for four different refinance scenarios I considered and their returns versus my existing loan. I compared a thirty year at 3.75% and a twenty year at 3.375%, with and without paying the extra $1,000 I’ve been paying each month in a race to reduce PMI. For NPV, I used 8%, which is higher than I would usually use as an alternative return rate, as it provides a more conservative NPV.
This first table (above) looks at the effect on total cash flows. A few things to note: payback period (the amount of time in years it takes to earn the money I invested back) is ridiculously low aka fast for all four scenarios. IRR’s are ridiculously high, but if you look at the sum or NPV, you can see that the scale is only so large (and you have to remember this is over 24-27 years, depending on the scenario). NPV is considerably lower than sum in most cases, due to comparing the returns to investing at 8% as an alternative and much of the returns (for the twenty year scenarios in particular) occuring far out in the future. The further in the future a cash flow is, the more heavily it’s discounted. Let’s take another look:
There are several interesting things to note here. First, I was previously paying $1,000/month extra in principal on my old loan. I first evaluated each loan term based on keeping that and then realized it didn’t make as much sense anymore and compared with no extra payments or smaller extra payments (I didn’t include the latter to avoid noise).
As far as the numbers go, there is a large positive return in the first couple years due to payments being reduced as PMI has gone down. In the cases where I’m not paying extra principal (blue and yellow) this amount is even larger as I’m saving $1,000/month. After year two, returns calm down a bit, even going negative for over ten years in the 20yr scenarios due to the higher payments (remember, we’re just looking at total cash flows here, including principal; we’ll get to interest in a minute).
In the twenty year scenarios, there is a huge positive return starting in year 19 due to the loan ending much earlier. After that, all returns are the cash I would have paid in total on my original loan. In the thirty year scenarios, it’s extremely interesting that returns become strongly negative around year 24. This is due to no longer paying extra for as long to knock down PMI as quickly as possible, which actually ends up extending the life of the loan despite the lower interest rate! That result surprised me.
The reason both the twenty and thirty year scenarios converge with their no extra principal selves is due to the difference in the middle of the loans only being the difference in total payments. The no extra principal versions just last longer.
This graph shows the same information cumulatively, so that each point on the lines contains the total return up to that point in time. This does a little better job of showing how things ultimately shake out over time and doesn’t have the overlap in the chart above. This also shows how much better the twenty year scenarios pan out., returning between $70,000 and $85,000 to us over time depending on whether we pay extra principal at first to get rid of PMI faster.
But just looking at the overall cash flows doesn’t tell the whole story! One of the big reasons we wanted to refinance was to save on interest and PMI, and all the money we’re paying in principal isn’t just thrown away to the bank – it’s being invested in a real estate asset (our home). What does the story look like if we take out the principal we’re paying and just focus on the interest and PMI we’re saving?
The story, sans principal
| No Principal | 30 yr, no extra | 30 yr | 20 yr, no extra | 20 yr |
| payback | 0.34 | 0.32 | 0.27 | 0.26 |
| IRR | 8% | 275% | 334% | 358% |
| NPV | -$87 | $12,842 | $24,505 | $30,733 |
| SUM | -$18,206 | $15,540 | $53,368 | $66,068 |
A few things to notice here are that the thirty year with no extra principal has a negative sum and NPV (I explain why under the last couple charts below), and the twenty year scenarios are significantly outperforming the thirty year scenarios by all metrics. But how are these returns distributed over time? Are they all the same? Has the overall picture changed from when we examined total cash flows?
Things now look pretty different! There’s still the spike in returns at the beginning, but it’s lower. Returns are now positive for many more years in the middle, especially in the twenty year scenarios, and the thirty year scenarios don’t look as good.
This graph also clearly depicts that the thirty year scenarios end up in the negative between years 10 – 20 and stay there. The reason for this is actually due to the lower interest rate, oddly enough! Since the interest rate is lower, it takes less time to pay the mortgage down to the point that PMI goes away.
That sounds like a good thing, but since we only modeled the scenario to pay extra until PMI goes away, we still end up paying less principal down at the beginning compared to our previous loan, which causes interest payments to be higher due to the higher balance later on (between years 10 and 20). This eventually also extends the total life of the loan. I found this result counter-intuitive and had to double-check I hadn’t messed up the calculations somewhere. So what does this story look like cumulatively?
Once again, the cumulative returns give us a much clearer and smoother view of how each of these scenarios performs. There is a big gap in total performance versus the thirty and twenty year scenarios, with the latter saving us between $53,000 and $66,000 in interest over the life of the loans. And neither this graph nor the previous cumulative graphs even assumes any appreciation in the value of the house or that I’ll invest the freed up cash flow, so they are extremely conservative! Pretty crazy.
So which is true? Should we look at overall summary statistics? Should we look at return distributions over time? Should we include principal or not? Ultimately, there’s no perfect right answer; each piece of information adds something new and interesting for you to evaluate. Did I need to do all of this to understand the trade-offs I was making between options? Probably not. Did I have fun doing it? Hell yes! Aren’t you glad you have me to do all this work for you?
Really though, how should I look at this?
If I were concerned with the theoretical accuracy and getting the perfect, complete picture, here’s what I’d do: I’d model the returns in total cash flows under distributions of different market returns with associated probabilities. Then, I’d take the principal portion of my payments and model distributions of different possible outcomes for appreciation or depreciation of the house. I’d then model many different scenarios in terms of when the best time to sell the house would be, as well as different scenarios in terms of us moving out and renting the whole property out, investing the resulting positive cash flows.
Ultimately, I don’t think I would gain much from this exercise. There are much more interesting things to model! Anyhoo…
The Process
So while it’s a win financially, we still had to go through the actual process of refinancing, and I thought some of you might be interested in what that was like, particularly if you have a home and are curious about refinancing yourself. I’ll cover how to buy a home and refinance in more detail later, but I’ve written about our initial process of searching for a home in the past.
After my coworker mentioned rates had dropped, I headed over to nerdwallet and put in some basic details about our loan and my credit. This returned a number of different loan offers with different rates and terms from different companies. Perhaps unsurprisingly, the company we closed on our first mortgage with had the lowest rate again.
I reached out to the loan officer I’d closed our first mortgage with (we’ll call him Jim) and asked how things were looking and where rates were at. At the same time, I went to a couple other sites that allow you to search for and compare loans, did a Google search on mortgage deals, and checked with a few companies individually.
As an example, I primarily bank with Ally, and I reeeally wanted to give them a shot after they were slow to respond about my first mortgage. As it were, none of the other companies were able to beat the deal I was getting with my lender. In addition to the low rate, they had a solid lender credit, which meant we wouldn’t have to pay as much out of pocket in transaction fees.
Jim confirmed that rates looked great but asked if we could wait two weeks. At the time, it had been less than six months since the original sale, which would have invalidated his commission on the first deal. We didn’t have a problem with that and luckily, rates didn’t climb back up (until after we locked).
After two weeks had passed, I discussed the details and whether we wanted to do a 30-year or a 20-year with Jim. He was pretty excited and told us he didn’t think we’d ever need to refinance again if we took the 20-year, which sounded pretty good to me.
Closing (Again)
Closing the second time around was a breeze compared to the first. The first time we closed was extremely stressful, featuring a plethora of last-minute changes, extra document requests, and close calls. While we still had to provide and sign a large number of documents (bank statements, paychecks, tax returns, etc), we didn’t have to update or explain anything, there weren’t any last-minute changes, and the notary came to our house. Overall, it was a much more seamless and enjoyable process.
One interesting decision we made was whether to have our house appraised (where a professional estimates the current value of your house, which is important for PMI, how much they’ll lend you, and other mortgage things). Normally, you don’t have an option. You just pay the $500+ and deal with it, hoping you get a good value. In our case, since we had closed so recently, the lender waived the requirement for an appraisal.
It’s possible we could have saved even more money, as it’s likely the value of our house appreciated ~10% since we bought it, but at this point there was a lot going on and I didn’t want to take on any additional risk. It could always turn out against us as well. We decided to simply save the $500 and move on, saving the money we knew for sure we’d save.
In the future, I’ll continue to share our initial home purchase process in more detail, as well as share guides on how to close on or refinance a home. In the meantime, what would you like to hear or learn about? Do you have a home closing or refinance story you want to share? Comment and share below!
Much Love,
(Your) Wallet