I like Kitces, and I refer to his site regularly. But he's wrong on this point
I'm firmly in the leveraged-investing-via-mortgage camp (in case it wasn't already obvious), but I wouldn't characterize Kitces as being "wrong." Instead, like many of the people on the opposite side of the debate in this thread (and the umpteen other equivalent threads across the forum), it's just that he's focusing on certain specific risks to the exclusion of others.
(It is true that margin loans and mortgage loans have differing characteristics that make mortgage loans a patently better vehicle for leveraged investing, but, as Canuck noted, Kitces concedes that point and it doesn't contradict his central argument, which, essentially, is simply that leveraged-investing-via-mortgage is in fact a form of leveraged investing!)
Kitces is clearly correct that leveraged investing, as a matter of course, presents the risk (not present with unleveraged investing) that your portfolio can go negative (the reason we use leverage in the first place is to put more capital at stake, with the inherent risk/reward trade off that our returns will increase if things go well while our losses will increase if things go poorly). This risk unquestionably exists, and, for the 4%-rate mortgage we've been using as a benchmark, has in fact already materialized in the past (ignoring tax and related considerations) (because leveraged-investing-via-mortgage using a 4% mortgage loan has a below-100% historical success rate of coming out ahead).
But, as you pointed out, Kitces is ignoring other risks that leveraged-investing-via-mortgage mitigates, most importantly, in my view, inflation risk, but also others, like the property loss risks dragoncar cited.
In addition, the unstated corollary to Kitces' warning that leveraged investing exposes you to low expected-probability black swan events is that there is a high expected-probability of being rewarded. For me, this is the biggest consideration -- effectively, by retaining my mortgage, I'm deciding to reduce my expected likelihood of working longer than necessary (and I'm willing to accept the risk that, in the unlikely (based on my expectations) event that the world implodes, I might have been better off otherwise).
A typical home loan is 30 years. So that's our borrowing horizon. Over a 30 year timeline, the historic risk of owning the broad stock market has been zero.
A nuanced nitpick on this point: technically, the horizon that matters is the weighted average life to maturity of the mortgage (which is shorter than 30 years, because of amortization). If the loan had no required principal or interest payments until final maturity,
then your statement would be completely true (with no nits to pick). But in reality, if, say, the stock market plummeted by 99% in years one through five of the loan, then fully recovered and soared to new heights in the remaining 25 years, you'd still come out behind (suffer total portfolio depletion) by having had retained the mortgage, because the required prepayment/amortization schedule would have forced you to liquidate stocks during the initial phase of the horrendous sequence of returns.