# WP18 — Critical literature review (Phase 1)

Prepared 2026-10-20 as the basis for Section 4 of the paper (« Extensive critical literature review, organized by mechanisms and open gaps »). Every reference cited below is listed in `literature_matrix.csv` and `paper/references.bib` with its verification source (Crossref JSON in `research/verification/`, resolved DOI, or publisher page). Papers that could not be verified are listed in `inaccessible_sources.md` and are not cited. Citation keys in brackets are the BibTeX keys.

## 1. The question and the organising principle

WP18 asks under what conditions inherited wealth, mortgage leverage, nominal debt, housing supply and down-payment constraints generate persistent owner–renter wealth inequality among households with similar labour income, and what changes under unexpected inflation. The literature that bears on this question is large but fragmented along five mechanisms that are rarely combined in a single general-equilibrium (GE) model: (a) the down-payment or collateral constraint that gates entry into ownership; (b) the leveraged return on housing and the resulting divergence of owner and renter balance sheets; (c) the revaluation of nominal mortgage debt under surprise inflation; (d) the supply side, i.e. land scarcity and the elasticity that converts demand shocks into prices rather than quantities; and (e) the intergenerational transmission of wealth through bequests and inter vivos transfers. The review is organised around these mechanisms, then turns to the Canadian institutional setting, and closes with the gaps that justify the model specified in Phase 2.

A preliminary observation sets the tone for the whole exercise. Statistics Canada's 2023 Survey of Financial Security reports a median net worth of $457,100 for families under 35 who own their principal residence against $44,000 for those who do not, and, among families aged 55–64, $1.4 million for owners with an employer pension against $11,900 for renters without one [statcan2024survey]. These gaps are an order of magnitude larger than the income differences between the groups, which is precisely the pattern the model must be able to generate and, more importantly, must be able to generate *persistently* across generations before any claim about inflation or supply can be evaluated.

## 2. Collateral, leverage and the price feedback

The amplification mechanism at the heart of the paper descends from Kiyotaki and Moore [kiyotaki1997credit]: when durable assets serve as collateral, a shock that raises their price relaxes borrowing limits, which raises demand and the price again. Iacoviello [iacoviello2005house] transplanted this mechanism into a monetary DSGE model with housing collateral and *nominal* debt, obtaining two results that WP18 inherits directly: collateral constraints amplify demand shocks, and debt deflation transfers wealth from borrowers to lenders in a way that dampens the response to inflation shocks. The limits of that model are equally instructive. It has a patient and an impatient representative household, no tenure choice, no wealth distribution and no inheritance; the redistribution it identifies is between two types, not within a cross-section of households that differ in initial wealth.

The quantitative heterogeneous-household literature of the 2010s supplies the architecture WP18 needs. Favilukis, Ludvigson and Van Nieuwerburgh [favilukis2017macroeconomic] build an overlapping-generations (OLG) GE economy with aggregate and idiosyncratic risk, collateral constraints and *heterogeneous bequest motives* that produce a realistic wealth distribution; they show that relaxing financing constraints produces a large house-price boom through a fall in the housing risk premium rather than through lower interest rates. Kaplan, Mitman and Violante [kaplan2020housing] add a rental market, long-term mortgages and tenure choice, and reach a conclusion that disciplines any claim about the collateral–price feedback: in their calibration, shifts in credit conditions move homeownership, leverage and foreclosures but *do not move house prices*, because the rental sector and long-term mortgages absorb the demand shift. Greenwald and Guren [greenwald2025do] resolve the apparent conflict between this result and the earlier credit-driven-boom literature by showing that everything hinges on the degree of segmentation between the owner-occupied and rental stock: with landlords absorbing credit-driven demand (no segmentation) credit does not move prices; with perfect segmentation it does; their estimated elasticity implies credit explains 32–53 % of the 2000s boom. For WP18 this means that the price-feedback component of contribution (i) is not a free parameter: the strength of the feedback is governed by the owner/rental segmentation parameter and by supply elasticity, and the model must make both explicit and report results across them. Sommer, Sullivan and Verbrugge [sommer2013equilibrium] and Sommer and Sullivan [sommer2018implications] provide the template for a rental equilibrium with landlords and an endogenous rent–price ratio; the latter paper is a standing warning that GE price responses can reverse partial-equilibrium intuition (removing the mortgage-interest deduction *raises* ownership once prices fall).

The empirical side of the collateral channel is well established. Mian and Sufi [mian2011house] document home-equity borrowing of 25–30 cents per dollar of house-price gain, concentrated among constrained owners; Gete and Reher [gete2018mortgage] show that tighter mortgage supply pushes households into renting and raises rents; Cloyne, Ferreira and Surico [cloyne2020monetary] show that the consumption response to interest rates is driven by mortgagors, with outright owners not responding at all. None of these papers speaks to intergenerational persistence, but together they justify two modelling choices: owners must be allowed to be liquidity-constrained despite high net worth, and the rental market must be an explicit margin rather than an outside option.

The mortgage-contract literature is the second pillar of this section. Campbell and Cocco [campbell2003household] formalise the household's choice between a nominal fixed-rate mortgage (FRM), which carries inflation (wealth) risk, and an adjustable-rate mortgage (ARM), which carries income risk. Chambers, Garriga and Schlagenhauf [chambers2009accounting] embed contract choice in an OLG GE model with tenure choice and attribute 56–70 % of the 1994–2005 rise in US homeownership to mortgage innovation and at least half of the post-1940 rise to the introduction of the FRM: contract design changes *who can own*, which is the sense in which contribution (iv) treats the mortgage contract. Guren, Krishnamurthy and McQuade [guren2021mortgage] and Campbell, Clara and Cocco [campbell2021structuring] study contract design for macro stabilisation in equilibrium life-cycle models; the latter is notable for pricing mortgages off an infinitely lived risk-averse lender's consumption, a device WP18 can borrow to keep lender wealth in the accounting of inflation redistribution, as the author's specification requires.

## 3. Nominal debt and surprise inflation

The Fisher channel [fisher1933debt] is the oldest element of the paper. Its modern quantification is Doepke and Schneider [doepke2006inflation], who revalue the nominal positions of US households by age and wealth and find that a moderate inflation episode transfers wealth from rich, old bondholders and from foreigners to young middle-class households with fixed-rate mortgages and to the government. Adam and Zhu [adam2016price] replicate the exercise for the euro area and Meh and Terajima [meh2011inflation] for Canada; the Canadian study finds that the main winners from inflation are young middle-class households with mortgage debt, while the old, the rich and holders of non-indexed defined-benefit pensions lose, and that Canada–US differences are substantial even between similar economies. Pallotti et al. [pallotti2024who] bring the accounting up to the 2021–2023 shock and add a nuance that matters for a model with renters: rigid rents acted as a hedge for lower-consumption households, so the episode worked like an age-dependent tax, with retirees losing up to 14 % of triennial income and roughly half of 25–44-year-olds gaining.

These are accounting exercises. They hold prices, rents, interest rates and behaviour fixed and therefore cannot answer the paper's question, which is whether the direct revaluation survives once rents, house prices, the real interest rate and lender wealth respond. Three papers go further. Meh, Ríos-Rull and Terajima [meh2010aggregate] compute the GE response to an unexpected 1 % price-level increase in an OLG economy with nominal positions calibrated for Canada and show that the aggregate and welfare effects are larger under inflation targeting than under price-level targeting, that the youngest, poorest and the government gain, and that the government's windfall, if rebated, lowers GDP for a long time; the model has no housing market, so the owner–renter dimension is absent. Auclert [auclert2019monetary] provides the sufficient-statistic language (unhedged interest-rate exposure, the Fisher channel, earnings heterogeneity) that WP18 will use in stating its inflation proposition, but has no tenure structure or intergenerational dynamics. Garriga, Kydland and Šustek [garriga2017mortgages] is the closest precedent for contribution (iii): in a GE model with long-term nominal mortgages they show that monetary policy transmits through the pricing of new mortgages and through the real burden of existing debt, that persistent inflation benefits FRM borrowers *over time* but hurts ARM borrowers *immediately*, and that effects are stronger under ARMs. Their borrowers and lenders are, however, two representative types; there is no renter who is priced out when nominal rates rise, no ownership-entry margin and no inheritance.

The gap is therefore clear. No existing paper combines (a) surprise inflation with fixed nominal mortgages, (b) an explicit rental sector with endogenous rents, (c) an ownership-entry margin gated by a down-payment constraint, and (d) a wealth distribution with inherited heterogeneity, in a model where the real interest rate, house prices and lender wealth are determined in equilibrium. The author's injunction not to assume that inflation raises real house prices or enriches every owner is well founded in this literature: Garriga et al. show that the sign for ARM borrowers is negative on impact; Pallotti et al. show that rent rigidity can protect renters; and Kaplan et al. show that the rental sector can absorb demand so that prices barely move. Contribution (iii) must therefore be stated as a set of *conditions* (contract type, share of debt revalued, rent indexation, supply elasticity) rather than as a sign.

## 4. Land scarcity, supply elasticity and the rent–price ratio

The empirical case for treating land as a scarce factor is now overwhelming. Knoll, Schularick and Steger [knoll2017no] show that real house prices in fourteen countries were roughly flat until the mid-twentieth century and then tripled, with about 80 % of the rise attributable to land prices; Davis and Heathcote [davis2007price] show that the land share of US house value rose from 35 % to over 45 % and that land drives price volatility; Rognlie [rognlie2015deciphering] shows that the post-war rise in the net capital share comes entirely from housing and reflects scarcity rather than accumulation; Piketty and Zucman [piketty2014capital] document the rise of wealth–income ratios in which housing is the main component. On the cross-section, Saiz [saiz2010geographic] and Glaeser and Gyourko [glaeser2018economic] establish that geography and regulation determine supply elasticities and that inelastic metros see persistently higher prices; Gyourko, Mayer and Sinai [gyourko2013superstar] show that inelastic supply with rising high-income demand produces *persistent* price divergence and sorting.

Two theoretical papers connect supply scarcity to the owner–renter distribution. Kiyotaki, Michaelides and Nikolov [kiyotaki2011winners] build a life-cycle GE production economy in which land and capital produce real estate and find that when the land share is large, house prices react strongly to productivity and world-interest-rate shocks, redistributing between net buyers (the young) and net sellers (the old), while financing constraints have limited price effects. Grossmann, Larin, Löfflad and Steger [grossmann2021distributional] model land as a fixed factor in a growing economy and show that surging rents redistribute from renters to landowners and raise wealth inequality. Borri and Reichlin [borri2021optimal] analyse the normative side, optimal taxation when house-price appreciation widens the owner–renter gap. Favilukis, Mabille and Van Nieuwerburgh [favilukis2023affordable] offer the most complete architecture (owners, renters, construction, incomplete markets, endogenous wealth inequality) for evaluating affordability policies in a single city.

What is missing is the interaction of supply elasticity with the other three mechanisms. Kiyotaki et al. have no bequests and no nominal debt; Grossmann et al. have representative owner and renter types; neither models the ownership-entry threshold. The natural conjecture, that lower supply elasticity raises the price response to any demand shock (credit relaxation, inheritance flows, inflation-driven demand) and therefore both the leveraged return of incumbent owners and the entry threshold faced by renters, is plausible but not established in a model that contains all the pieces. This is the content of contribution (iv), and it should be treated as a comparative-statics proposition to be proved under explicit assumptions, with the Kaplan–Mitman–Violante caveat that an elastic rental sector may neutralise much of it.

## 5. Tenure choice, the down-payment constraint and the property ladder

The microeconomics of tenure choice begins with Henderson and Ioannides [henderson1983model], whose rental externality rationalises an ownership premium and hence a wedge between rents and user costs, and with Linneman and Wachter [linneman1989impacts], who show empirically that it is the *wealth* (down-payment) constraint rather than the income constraint that blocks ownership. Stein [stein1995prices] shows that down-payment effects generate multiple equilibria and price–volume comovement in a static setting. The paper closest to contribution (i) analytically is Ortalo-Magné and Rady [ortalomagne2006housing]: in a life-cycle model with a property ladder and a down-payment constraint, the ability of young households to finance the down payment on a starter home is the key driver of prices, income shocks cause overshooting, and capital gains on starter homes relax the constraints of existing owners. This is an ownership-entry threshold with price feedback. It has, however, no bequests, no nominal debt, no inflation and no supply side; the income process is stylised and the distribution of wealth across households of equal income is not an object of analysis.

The quantitative life-cycle literature fills in the behavioural detail: Gervais [gervais2002housing] (OLG GE with tenure and a down payment), Iacoviello and Pavan [iacoviello2013housing] (age profiles of ownership and debt in GE), Halket and Vasudev [halket2014saving] (saving for the down payment versus mobility), Díaz and Luengo-Prado [diaz2010wealth] (housing and the shape of the wealth distribution), Kaas, Kocharkov, Preugschat and Siassi [kaas2021low] (policy determinants of the German–US ownership gap, with the welfare caution that closing the gap would lower welfare), Floetotto, Kirker and Stroebel [floetotto2016government] (winners and losers along transition paths), and Halket and Pignatti [halket2015homeownership] (rental scarcity as a microfounded driver of ownership). Paz-Pardo [pazpardo2024homeownership] is especially relevant because it identifies a *competing* explanation for falling young ownership and wealth: riskier and more unequal earnings for later cohorts. WP18's claim that inheritance generates owner–renter gaps *among households with similar labour income* must therefore be isolated from the earnings-risk channel, which argues for holding the earnings process fixed across inheritance quantiles in the baseline experiments. Nakajima and Telyukova [nakajima2020home] close the life cycle: retired owners dissave slowly because they prefer to stay and cannot borrow against the house, so housing is the main asset bequeathed, which ties the tenure block to the bequest block.

## 6. Incomplete markets, bequests and intergenerational wealth

The incomplete-markets tradition (Samuelson [samuelson1958exact], Diamond [diamond1965national], Aiyagari [aiyagari1994uninsured], Huggett [huggett1996wealth]) provides the equilibrium concept. De Nardi [denardi2004wealth] is the canonical bequest specification: voluntary bequests modelled as a luxury good together with intergenerational earnings links are required to match the top of the wealth distribution; WP18 adopts this warm-glow form and lets the bequeathed asset include housing. Benhabib, Bisin and Zhu [benhabib2011distribution] and Gabaix, Lasry, Lions and Moll [gabaix2016dynamics] supply the analytical benchmarks for persistence: Pareto tails arise from capital-income risk, and standard random-growth processes converge too slowly to explain fast changes in inequality, which is a useful yardstick when the model's persistence results are stated in terms of half-lives of the ownership gap.

The empirical literature on intergenerational wealth has converged on three findings that the model must respect. First, parental transfers at the time of purchase are a first-order determinant of entry into ownership: Engelhardt and Mayer [engelhardt1998intergenerational] find that about a fifth of US first-time buyers receive transfers covering half the down payment; Guiso and Jappelli [guiso2002private] find transfers advance first purchase by about two years in Italy; Blickle and Brown [blickle2019borrowing] find that a transfer raises the renter-to-owner transition probability by 6–8 percentage points (35 % of the mean) through relaxed constraints rather than wealth effects; Charles and Hurst [charles2002transition] find that parental wealth and transfers explain much of the Black–White ownership gap in the US at equal income. Second, parental *housing* wealth specifically transmits to children: Daysal, Lovenheim and Wasser [daysal2023intergenerational] estimate on Danish registers that about 25 % of a parental housing-wealth gain in early childhood passes through to the child's adult housing wealth and 27 % to total wealth, with only modest effects on the extensive ownership margin and with earnings and education explaining only 20–30 % of the transmission; Benetton, Kudlyak and Mondragon [benetton2022dynastic] study the US parent–child link through home equity. Third, transfers are not the whole story: Fagereng, Mogstad and Rønning [fagereng2021why] show with Korean-born adoptees in Norway that family environment raises child wealth and risk-taking causally, and that the effect is *not* mainly through gifts or inheritances; Nekoei and Seim [nekoei2023how] show that inheritances reduce inequality on receipt but raise it within a decade because wealthy heirs earn higher returns; Boar [boar2021dynastic] shows that parents save to insure their children's income risk, so the timing of transfers is endogenous.

The implication for the model is twofold. The inheritance channel WP18 formalises is a real and measurable channel, but the paper must not claim that it is the only or the dominant source of intergenerational wealth persistence; the model should be presented as isolating one mechanism (transfers and bequests interacting with the down-payment constraint and leveraged housing returns) while holding other channels (preferences, financial literacy, earnings transmission) fixed. And the Nekoei–Seim finding that *returns* on inherited wealth drive the reversal from equalising to disequalising suggests that leveraged housing returns, which are exactly what the inheritance buys access to, are a candidate explanation that the model can test but should not assume.

Kuhn, Schularick and Steins [kuhn2020income] provide the macro-historical motivation: over 1949–2016 the portfolio composition of US households (leveraged housing for the middle class, equity for the top) makes asset-price movements the main driver of wealth-inequality dynamics. This is the leverage-portfolio channel that WP18 formalises in a model with explicit tenure.

## 7. The Canadian institutional setting

Canada differs from the US in ways that matter for every mechanism above. Mortgages are typically amortised over 25 years but *renewed* every five years or less at the prevailing rate, so the fixed-rate contract is fixed only for the term: Kartashova and Zhou [kartashova2020how] exploit these scheduled resets to show asymmetric spending and deleveraging responses; Bounajm and McWhirter [bounajm2024modelling] document how the Bank of Canada's LENS model represents the term structure of mortgage debt and notes that longer fixed terms slow the pass-through of rate changes; Godbout, Su and Xu [godbout2025how] estimate that about 60 % of outstanding mortgages renew in 2025–2026 and that five-year fixed-rate borrowers face average payment increases of 15–20 % relative to December 2024. CMHC's Spring 2026 report [cmhc2026residential] shows that only 11 % of mortgages extended at chartered banks in February 2026 were traditional five-year fixed contracts, with 42 % variable. The consequence for contribution (iii) is that the Canadian FRM is an *intermediate* contract between the US 30-year FRM of Doepke–Schneider and the ARM of Garriga et al.: the inflation gain accrues only for the remaining term and is partly clawed back at renewal through higher nominal rates, unless the surprise is fully transitory. This is a genuinely open question for a Canadian calibration and is the clearest institutional contribution the paper can make. Allen, Clark and Houde [allen2019search] document search frictions and rate dispersion in the Canadian insured-mortgage market, which can be noted but need not be modelled.

On the household side, Meh and Terajima [meh2011inflation] provide the nominal-exposure profiles by age and tenure, Hou [hou2010homeownership] the age profile of ownership (steep before 40, plateau near 65 at about three quarters, rising across cohorts, closely tied to income), Mirdamadi and Khalid [mirdamadi2023parents] the parental-ownership channel (children of owners more likely to own; family gifts to first-time buyers rose from 20 % to 28 % of buyers over 2015–2021 with average gifts rising from about $50,000 to $80,000, figures the article attributes to a CIBC analysis), and the Survey of Financial Security [statcan2024survey] the owner–renter wealth gaps quoted in Section 1 (39 % of families hold a mortgage). On the supply side, no Canadian analogue of Saiz's elasticity estimates was found in the verified sources; supply elasticity will have to be treated as an illustrative parameter with a range, not a Canadian calibration target, unless a verified source is located in Phase 3.

Two items in the author's brief could not be identified. Searches for a Bank of Canada housing or mortgage paper by Uzeda returned only his profile (empirical macroeconomics, Bayesian time series) with no housing paper, and no paper by « Alexander and Bazel » on Canadian housing or intergenerational wealth was found in Crossref, Google Scholar-indexed sources or the Bank of Canada site. They are logged in `inaccessible_sources.md` and are not cited.

## 8. Gaps and what they imply for the model

Reading the five strands together, the following gaps are not filled by any verified paper and define the space in which WP18 can contribute.

1. **No GE model combines bequest heterogeneity, a down-payment-gated ownership margin, an explicit rental sector and price feedback, and then characterises the entry threshold analytically.** Ortalo-Magné and Rady have the threshold and the feedback but not bequests or a rental sector; Favilukis et al. have bequest heterogeneity and feedback but no tenure margin in the baseline; Kaplan et al. have everything except inheritance heterogeneity and do not characterise the threshold. Contribution (i) is defensible if it is stated as a characterisation (existence, monotonicity in bequests and LTV, dependence on the price level) in a tractable version of the model, with the feedback's strength tied to the segmentation and supply parameters.

2. **Persistence of the ownership gap across generations has been documented empirically (Daysal et al., Charles–Hurst, Mirdamadi–Khalid) but not derived.** The theoretical persistence results that exist (Benhabib–Bisin–Zhu, Gabaix et al.) concern the tail of the wealth distribution, not a tenure-based bimodality. Contribution (ii) should aim at conditions under which the owner–renter gap among equal-income dynasties has a positive long-run mass (a non-degenerate stationary distribution with two tenure classes) and, only if established, multiplicity in the Stein sense. Multiplicity should not be promised.

3. **Surprise inflation with fixed nominal mortgages has been studied in GE (Garriga et al.; Meh et al.) but never with renters, an ownership-entry margin and inheritance.** The literature already shows that the sign for owners depends on contract type and that renters can be protected by rent rigidity; what is open is whether inflation shifts the *entry threshold* (through nominal rates at renewal and through house-price responses) and therefore the next generation's tenure, and whether a Canadian five-year term changes the answer relative to the US 30-year benchmark. Contribution (iii) is incremental on the direct revaluation and potentially new on the entry-margin and term-structure dimensions.

4. **Supply elasticity has been linked to owner–renter redistribution (Kiyotaki et al.; Grossmann et al.) and mortgage contracts to ownership rates (Chambers et al.), but not to intergenerational mobility in tenure or wealth.** Contribution (iv) should be formulated as comparative statics of a mobility measure (probability that a renter's child owns; rank–rank slope of wealth) with respect to supply elasticity and contract type, with the caveat that the rental sector's elasticity may dominate.

Three cross-cutting lessons follow for the Phase 2 specification. The rental sector and the owner/rental segmentation parameter must be explicit because they govern whether credit, inheritance or inflation move prices at all. The earnings process must be held fixed across inheritance groups in the baseline so that the equal-income comparison is clean, with Paz-Pardo's earnings-risk channel treated as a separate robustness experiment. And every inflation result must be reported for at least three contract structures (long FRM, five-year renewable FRM, ARM) and two rent-indexation regimes, because the verified literature already shows that the sign of the redistribution is not invariant to these choices.

## 9. Scope and limits of this review

Seventy-two references were verified and retained; twenty-six further references were verified but set aside as peripheral (they are listed with their verification status in `inaccessible_sources.md`, section B, and remain available if a section of the paper later needs them). Abstracts were read for 38 of the 72 (Crossref abstract field, publisher page, NBER or IDEAS page, Bank of Canada or Statistics Canada page); for the remainder the bibliographic metadata were verified through Crossref and the contribution summarised from the title and the reviewer's knowledge of the paper, which is flagged in the matrix where the summary rests on metadata only (Grossmann et al., Borri–Reichlin, Benetton et al.). No OpenAlex access was available during this session (MCP connection failure), so the coverage of very recent (2025–2026) working papers on the specific intersection studied here is weaker than for published work; this is the main uncertainty in the novelty assessment and should be revisited before the manuscript is drafted.