Blessed by the Potato

Back when we talked about the gross debt service ratio, we discussed how sometimes a simple rule can scale in surprising ways.

A set percentage of your house’s value is a very common rule-of-thumb for estimating ongoing maintenance costs — 1% is typically seen these days (and what I use), though older articles from before the massive run-up in prices can be double that. It’s a tough thing to estimate accurately because there are so many line items and they’re all irregular; however, maintenance is a pretty big part of the cost of ownership so you can’t ignore it when considering the costs of shelter. A rule-of-thumb at least gets it in people’s minds — people who might otherwise only consider the costs they had to pay in the last few months (indeed, so many bad amateur rent-vs-buy considerations just use the mortgage payment). So if it’s even approximately right to +/- 50% I’m happy to continue to spread the rule-of-thumb.

In a comment this morning, Geoff says:

My issue with such articles about homeownership is that anytime they mention home maintenance, they use a percentage of the house value to estimate costs; this worked maybe when home price reflected the size of the house. But there’s no way that my 1961 built 1200 sq.ft house in toronto (worth I’d say $750,000 easy) costs more to maintain than my in-laws 3500 sq.ft house in Trenton (worth I’d say $150,000).

And that’s totally fair. The rule-of-thumb isn’t perfect, indeed even for Geoff and his immediate neighbours their maintenance costs may not be that close because of luck or the materials and workmanship they choose. To some extent size will matter, but a maintenance rule-of-thumb won’t scale perfectly with size, either: it costs so much to get a new furnace in the first place, and only a bit more for a bigger one; likewise the expensive kitchen/bathroom/etc. stuff only scales up a bit for larger houses, with most of the extra space being bedrooms/offices/living rooms that are basically simple shells. And within a jurisdiction size is usually captured by price increases anyway.

But keep in mind that there are other factors that will make a smaller place in the city more expensive to maintain than a larger one in the country, so there will be some scaling with price (though it’s never going to be a perfect set percentage).

For instance, the potentially higher labour costs in the city is a somewhat obvious one, but a much bigger factor is that the standards will be different. More expensive houses (larger or smaller) will be more expensive to maintain because they are lived in by people with more expensive tastes (or are situated in neighbourhoods with more expensive standards). For example, on PEI my parents’ kitchen is now about 20 years old, and still perfectly fine — heck, it’s probably still one of the more recent ones out there! In Toronto a 20-year-old kitchen would be scandalous and require a reno — a 10-year refresh is de rigueur — though in that case it may not be a cash cost as you can choose to not do it and just suffer relative depreciation (or reduced appreciation) relative to the trendy neighbourhood. Likewise, when a kitchen renovation does come up on PEI a whole new set of appliances would likely come in a fair bit under $5000; in a prestigious Toronto ‘hood each stainless steel, celebrity-chef-approved piece might run that much. Similarly for the rest: the standard for outdoor appearance in a less-expensive area might be a bag of grasseed and patience whereas in a million-dollar neighbourhood landscapers, sod, and shrubbery would be expected. Etc.

Indeed, it’s for this reason that I think the more common rule-of-thumb is value-based rather than size-based. And for cases of extremely expensive small properties or very cheap large ones, well, you’ll hopefully know to nudge your estimates up or down appropriately.

I want to be clear up front: I like Rob Carrick. He talks a lot of sense, and is one of the few voices out there talking about the potential dangers to homeownership-at-any-cost, and breaking the misconceptions about renting. I also think the “rent and invest the difference” message is incredibly important — it was the central idea behind creating the rent-vs-buy calculator.

That said, I had a lot of head-shaking moments at his column this weekend called “the renter’s guide to successful investing.” These are mostly quibbles to be sure, but there are a lot of quibbles for a thousand-word article.

I really get the concept of trying to make advice bite-sized and manageable: something close enough that people actually follow is better than precise advice that people tune out (indeed, I have been guilty of that often enough). However, the maxim is that things should be made as simple as possible but no simpler. I think this is unfortunately a case of over-simplifying. Similarly, I just don’t get his “real life ratio”¹.

First off the message about saving is confusing. At the start he says “Homeowners build wealth by paying their mortgage down and increasing their equity in a house that they will presumably be able to sell for more than they paid. […] A homeowner with a paid off house has the luxury of choosing to: continue living mortgage-free in the home (and rent-free, for that matter);[…]” which leaves off the issue of homeowners also needing to save. He doesn’t actually say “forced savings” at any point, but this statement brings that terrible idea to mind. Homeowners also need to save — indeed, you can’t live “rent-free” in a paid-off house because you have to pay property tax, upkeep, insurance, etc. That notion does come in at the end, off-handedly: “Just as a homeowner needs to have dedicated savings for retirement, so does the renter.” That idea really doesn’t come through in the article, unfortunately, and I’m sure many missed it.

The big issue though is this mysterious 1.5% rule-of-thumb he comes up with. I mean, the logical thing to do would be to take some average figure for the principal paid down with a mortgage payment and use that. Or to go to a rent-vs-buy calculator and find out what the actual cost difference is and invest that. Instead, he takes an estimate based on maintenance + (1/2)*(property tax). What?

This is not remotely accurate for the areas where the rent-vs-buy debate is most important. Check it out for Toronto: the so-called “renter’s dividend” is almost twice what Rob’s rule-of-thumb predicts. I get 2.7% (if you insist on putting it into a percentage of the house price instead of a dollar value for each situation), and that’s for the apples-to-apples case. If you’re a renter saving up for your first place and plan to move up from your rented accommodations (e.g. to move to a house from a small apartment), then you should be saving that difference, which might be 5% of the house’s value.

Secondly, it really irks me that he doesn’t ever suggest that you should calculate it accurately. The rule-of-thumb is all that’s presented, and that is further simplified down to 1.5%.

Now rather than just bitch and not provide a solution, here is a new rule-of-thumb derivation:

Rule-of-Thumb for Amount of “Renter’s Dividend” for Apples-to-Apples Comparisons

How much does it cost to own a house? Roughly speaking property taxes are a hair under 1% in many municipalities, combine that with insurance to make the estimate an even 1%; insurance maintenance rules-of-thumb are about 1% (as Rob has in the article); transaction fees and/or discretionary “while we’re at it…” renovations run about 0.5%/yr amortized out (~10% every 20 years); the last tricky part is figuring out the mortgage payment — remember the renter wants to save the cash flow difference, so this time we do want to include the principal repayment portion. For a 25-year mortgage with 10% down at 3.5%, the mortgage is 5.5% of the property value. The total cash cost of owning is thus 8.0% (which includes principal repayment).

So find what your rent is as a portion of that, subtract, and voila.
Toronto: price-to-rent of 240X translates to rent being 5% of the price of a house; thus save 3%.
Vancouver: price-to-rent of 330X translates to rent being 3.6% of the price of a house; thus save 4.4%
Canada: if the more typical price-to-rent is 180X in other centres in the country, then rent would be 6.7%, so save 1.3% (I guess this is kinda close to what Rob got).

[Note the rule-of-thumb figures are not quite in perfect agreement with the spreadsheet — people should also save what the homeowner would be (e.g., 10% of income for retirement, depending on your guideline of choice]

Rule-of-Thumb for Amount of “Renter’s Dividend” for Moving-Up Comparisons

Now if you’re in a small apartment but planning to move up to a house it gets a bit more complicated. If you want to budget as though you were already in that house and saving the difference, then you would be saving much more. To make the rule-of-thumb simple, assume that the move up is for double the cost of your current place. That is, if you’re in a 2-bedroom apartment and want to move up to a 3-bedroom detached house, assume that if the house is $500,000, your apartment is $250,000.

So if you’re in a Toronto apartment, saving up for your first house (and planning on living to that budget — i.e., it’s not so far in the future that you’re counting on significant wage increases to make it work), then you’d want to save the “renter’s dividend” from your rental versus a comparable condo (half the house) plus all of the cost of owning on the difference between that condo and the house (approximated as half the house). So that first half would depend on the price-to-rent in your city, say 3% in Toronto, plus 8% of the difference, for a total of 11% on the half the value of the house, or 5.5% on the full value of the house.

Now maybe Rob went through a similar rule-of-thumb derivation, and was simply afraid that the numbers were too large — either that no one would believe the so-called “renter’s dividend” would be so enormous in this environment, or that the suggested amount to save would shock people. Moving on.

“Whether you’re a renter or an everyday investor, there’s only one way to set up a disciplined investing program. You need to have money transferred electronically into your investment account from your chequeing account every time you get paid, or once per month.” [emphasis mine] Ok, maybe it’s just hyperbole, but there are lots of ways to set up a disciplined investing program — as individual as the person. Sure, I’m a big advocate for automation: there are lots of good reasons for it to work and it’s proven². I make a strong case for it in my new book, too. But it’s not the only way. Indeed, automation is very much a do-as-I-say-not-as-I-do recommendation: for my situation, with highly seasonal spending and freelance income, I do almost all of my saving in the first half of the year. If I went automated I would just have to compensate with a larger savings account to buffer the changes in my budget. Plus a natural frugal inclination means I don’t worry about blowing my budget just because I don’t hide my money from myself. I recognize that there can be better ways: some people for instance, respond best by taking out a loan for the next year’s savings to invest, and then target paying down the loan. Others may target a few “no spend” pay periods and save in say 3-4 bi-weekly binges (even those who just save the “triple bonus” pay periods that come up twice per year on a bi-weekly paycheque system manage to hit a 7.6% savings rate, which is not terrible). Whatever works³.

A penultimate, minor quibble: the first table is near-useless. It’s a table of “if you get X% return saving $Y per month, after 30 years you’ll have $Z pile of money.” It’s completely dissociated from the message of the article. Would have been much more useful to combine the first and second tables: pick a rate of return (say 6%) and show that you would have enough to buy the average house price (or not, or buy two, whatever the case is) for each city if you rent and invest the difference.

A final complaint: he ends off with a specific mutual fund recommendation. I know he doesn’t phrase it precisely as a recommendation, instead as an example backing up what returns to expect (which is sadly needed when many readers may think investing means a HISA at 1.2% nominal), but still, the mention of one specific fund to end a column titled “…guide to successful investing” irked.

1. As an aside, if the real life ratio spreadsheet is targeted at potential home buyers I would have built-in defaults on some of the calculations based off purchase price, like mortgage payments, property tax, insurance, and maintenance, because first-time buyers may not know these numbers in advance.

3. Which yes, is usually automation/pay-yourself-first.

Thanks to some discussions with people (and Redditors) I have updated the rent-vs-buy investment method calculator (aka the ultimate rent-vs-buy comparison tool for Canadians). You can see the spreadsheet in Google Drive here (and save a copy to your own Google Drive or download in various spreadsheet formats) or click here to download it in Excel format.

Please see the original page for instructions, and the follow-up discussions: part 1 on things to consider and discussion questions, and part 2 on the sensitivity to various inputs changing.

If you haven’t seen this before, it’s a very detailed and customizable rent-vs-buy calculator. It assumes that all else being equal, you can compare apples-to-apples options for your shelter. If buying costs more, the renter will save the difference in monthly cash flow and invest it. It allows you to model a change in interest rates over time (specifying a rate for years 1-5, 5-10, and 10+), includes the effect of transaction fees, house price appreciation, taxes on the renter’s investments, and most importantly: investment returns that compound over time.

What’s new:

• The default mortgage rate is now 3.49%, the lowest big-bank 5-year fixed rate my rate-comparing friends at ratesupermarket.ca were able to find. With the move to a 5-year fixed (the most common option chosen) I’ve updated the back-end mortgage calculations to account for the bizarre 6-month compounding of fixed mortgages in Canada.
• The CMHC charges have been updated for the recently announced changes (though those won’t take effect for another month).
• The summary box (scroll over to the right) now also says how much the buying case wins by (in the event that it does) so you don’t have to look down at the full results table.
• The default comparison has been updated. I’ve just spent a quick half hour searching for comparable listings and found many exact — same unit — apples-to-apples comparisons, and Toronto’s price-to-rent is easily over 240X right now¹.

I have been asked about creating a space for fudge factors (in particular, to model the case where the owner gets a roommate or rents out a basement/secondary suite²) and I have not included that and do not feel persuaded to. Having such a field would just invite non-comparable comparisons (like comparing renting a full house to owning half of one with a call option on the rest). It’s a spreadsheet, so it’s not hard to account for such cash flows (for instance, just over-write the maintenance fee column with a combination of increased maintenance fees from being a landlord and a negative cash cost item for the rent income), and I would much prefer you think deeply about it by doing it manually than just jumping ahead to the fudge cell to justify buying.

1. I was overly fair to the buying case before and renting was still better — a point that was lost on many. The comparison now starts with one such matched pair (in North York). Renting now totally blows buying out of the water. I don’t want to belabour the Toronto housing bubble issue too much (I’d rather people focus on the usefulness of the tool and try it out for their own purposes without getting distracted by my situation), but it’s not even close guys. And I’m still being too fair by being at the bottom of the range — many of the condos that were “only” 240X had maintenance fees of ~1.4-1.5%, vs the sheet’s default of 1.1%, and those condo fees don’t even cover all maintenance/upkeep needs.
2. I already had a short post on this topic, but in brief: if it doesn’t make sense to rent out a whole house, how does renting out half of one suddenly become financial genius?

Update: Etienne (who was featured by Garth Turner recently) emailed me with the fix for a minor bug: the CMHC premium was being applied to the whole house value rather than just the loan value. Fixed as of March 15, 2014. The magnitude of the error depends on the downpayment; for example with 5% down it made the mortgage 0.15% too large, for 10% down it was 0.24% too large.

One of my shining triumphs here has been to create (with generous help from Matthew Gordon) the ultimate buy-vs-rent calculator tool (direct link to the spreadsheet).

The beautiful thing about a spreadsheet-based calculator like that is that you can follow the calculation, item-by-item, and check it for bugs if you get unexpected results. Earlier this week, B&E posted a list of calculators out on the net, and rather than linking to my supremely excellent calculator and associated post, Robb linked to a Get Smarter About Money calculator. Ok, it’s web-based and a little more user-friendly than a spreadsheet, and has graphs and sliders (though why you need house price to go up to $10M is a question left unanswered)… but was it accurate? I’ve seen many, many terrible buy-vs-rent calculators (even some seemingly excellent ones like the famous New York Times ones that just doesn’t work for Canadians due to tax differences). So I played around with it. And I quickly saw wonky results like this:

In the comparison I have there, the price-to-rent multiple is 260X ($2500 monthly rent on a $650,000 house); as we’ve learned from previous posts in realistic scenarios it should be better to rent with prices so detached from rents. Yet here the calculator is showing a rather large benefit to buying if you can only wait 7 years or more. Then, strangely and inexplicably, renting rapidly takes the lead in the final few years of the comparison, with some sort of apparent discontinuity at year 30. If you look at their “chart” you can see more errors immediately: I had entered $2500/mo in rent, which is $30,000 per year, yet the “total renting expenses” came to just $12,360 in their chart, a factor of three too low. The buying expenses were only about $31k in their chart, whereas the mortgage alone is that much, with a total cash outlay of nearly $45k each year.

Now if you instead do the same comparison in my calculator, you’ll find that renting beats buying right from the start (due to the high transaction costs), and is fairly flat in terms of net benefit for about 10 years, at which point the investment portfolio starts to get large enough that investment returns become comparable to rent and the exponential growth becomes truly noticeable. There is no big “buying is better” hump in the middle. Moreover, the magnitude of the difference is notable: in my calculator renting beats buying in such a scenario by over $600,000 in year 30, versus the nearly break-even result from this scenario in the flashy online tool, with the same assumptions regarding investment returns, inflation, mortgage interest, and other sundry costs.

It’s a bit distressing, as other online calculators have recently been found to have serious errors as well. For instance, Michael James uncovered one on the Globe & Mail’s site, and just today news broke on retirehappy about the government’s CPP calculator over-estimating your future CPP benefits and not at all handling early retirement scenarios correctly.

Footnote: let’s say you’re not convinced that my spreadsheet is the gold standard to which all other rent-vs-buy calculators should be held. To then check the accuracy of the online calculator, let’s run the first year’s numbers manually:
Buying: mortgage $31k, property tax $6k, insurance $1.7k, maintenance $6.5k; total cash cost: $45.2k. Principal paid down: $16.5k. Net cost of owning: $28.7k.
Renting: rent $30k, insurance $0.4k; total cash cost: $30.4k. Investment portfolio gains: $10.5k. Net cost of renting: $19.9k. Cost to sell house: $39k. Gain on house: $15k. After year 1, renting ahead by: $49.3k. Online tool says: $16k.
If you notice any errors let me know.

Fact: an “investor” buying a condo in Toronto today quite likely faces negative cashflow and poor projected investment returns based on current rents. The three reasons for buying in such a situation are that rents will increase — and increase fast enough to matter/before interest rates rise; that there will be price appreciation; or that the investor is making a mistake. Well there certainly was price appreciation in the past, and appreciation can beget appreciation until a bubble pops. Current and future buyers may not be counting on appreciation as much though, as expectations temper and we see flat or negative price growth (year-over-year price changes were negative for 416 condos in 5 of the last 12 months, and under 2% in all but 4).

That leaves rent increases. Urbanation puts out a quarterly report claiming staggering amounts of rent increases (figures that look even higher as they report price per square foot in a market with ever-smaller units). I mean, they don’t come close to keeping up with the kind of price appreciation seen over the past decade, but staggering compared to our expectation of rent increases from CPI inflation and the Ontario rent control increases. This news often gets picked up by the Star (with breathless headlines) and these high inflation figures get stuck in the minds of some. Personally, I’ve doubted these figures, as they don’t jive with my anecdotal experience or other data sources on rent inflation.

A new report from Veritas has ignited a bit of a spat, as they report a small amount of rent deflation. They also explicitly track and project the negative cash flow for an investor buying a condo in Toronto today. Urbanation took to their blog to try to debunk the new report and reinforce why we should listen to them instead.

What I found particularly shocking in Urbanation’s post is that they themselves don’t have a good handle on what proportion of the market their data is sampling. As far as I can tell, nobody does. They claim to capture 2/3 of the [condo] rental market. I back-of-the-envelope it and get more like a quarter of it (and a non-randomly selected quarter at that). And I think that says just how bad the data is: the biggest, splashiest report on the matter uses a non-random subset of data with several potential selection and reporting biases, and no one can even say what proportion of the population their sample represents or how representative it is. Smaller spot-check “mystery shopping” samples are potentially not as accurate, but nevertheless the fact that they disagree should call into question how much weight we give either figure. Nobody can even say whether the average rental tenancy lasts 3 years, 5, 7, or something in-between. Urbanation just takes an unsupported stab at 5 years below (20% turnover), and assumes that’s the same for owner-occupied and rental units:

Who’s [sic] research should be taken with a “grain of salt”?
We recognize that transactions through the MLS system don’t represent all activity in the condo rental market, but we do believe it represents the strong majority. […] The ‘true’ annual turnover rate of all condos is likely closer to 20%

[Emphasis mine]

Yes, whatever slice Urbanation gets from MLS will be bigger than the slice Veritas sampled on Craigslist and their mystery shopping adventure, but that’s not the point*. The take-home message is that they are both imperfect samples saying very different things about the market. We need to view the whole thing as a lot fuzzier than we have been: both reports likely have large margins error, and it looks like all we can say about rents on condos is that they’re basically flat, plus or minus 5%. Heck, the Ontario rent control rate may be the best estimate of rent inflation out there.

* – Even though bigger is usually better in sampling and stats, a bigger biased sample is still a biased sample.