Housing inflation in Singapore’s CPI
Housing prices have always been a hot topic for discussion amongst Singaporeans. Both academics and netizens have written extensively on the relentless rising cost of living due to housing price inflation. In a footnote to a 2008 paper on Singapore’s policy responses to ageing and retirement, Prof Mukul Asher cites a 2007 Citi memo (which I have unfortunately not been able to unearth) authored by Dr Chua Hak Bin arguing that Singapore’s inflation rate was severely understated primarily due to the way housing inflation was calculated:
Chua (2007b) has argued that the inﬂation rate in Singapore, as measured by the consumer price index, is signiﬁcantly understated, primarily because of the way the housing component is incorporated.
One of the bigger mysteries in Singapore is that although public housing prices have gone through the roof, inflation has not. Why should this be so, especially since home-ownership rates in Singapore is one of the highest in the world (88.8% as of 2009). What determines how housing prices in Singapore affects the inflation rate? To answer this question, we must examine how housing prices are factored in the consumer price index (CPI). This may come as a surprise to some, but housing prices in Singapore are not reflected in the CPI.
How then does Singapore’s Department of Statistics evaluate the housing component of the CPI? Like many other countries, Singapore utilises an indirect measure (since 1977/1978) by calculating what is known as owner’s-equivalent rent (OER). This refers to the rent a home-owner in Singapore would pay if he were renting the apartment from himself. The notion of calculating imputed rent, as it is known, may be somewhat odd a measure to use, but is used most often worldwide as a means of calculating the costs of owner-occupied housing (OOH).
Why should one prefer the OER method of calculating housing inflation? After all, Singapore is fond of boasting of its 88.8% home-ownership rate as being one of the highest in the world, yet paradoxically it uses a measure of rental costs to measure inflation due to housing costs, which effectively means that such a inflation metric would be accurate only to the remaining 11.2% of Singapore residents.
While I am not able to find any official justification in Singapore for preferring the concept of OER over housing prices and the mortgage rate for measuring housing inflation, evidence from the United States justify it on the reasoning that housing is typically considered part investment and part consumption:
Until 1983, the bureau measured housing inflation by looking at what it cost to buy and own homes, considering factors like house prices, mortgage interest costs and property taxes. But given the shifts in interest rates and housing prices, those measures could show big bounces from month to month. Besides, homes are a strange hybrid of a consumable good and a long-term investment. As part of a long-running evaluation, the bureau wanted to “separate out the investment component from the consumption component” of the housing market, said Patrick C. Jackman, an economist at the bureau.
In other words, a home isn’t just where you hang your hat or an investment in the future (or, in the case of visitors to CondoFlip.com, for two days). It’s something that is consumed, the way potato chips, gasoline or a barber’s skills are. So the bureau decided to track a measure that represents only the consumption of housing: rent. And rents had much to recommend themselves as a long-term inflationary yardstick – they don’t jump significantly from month to month – and as a proxy for home prices. Over time, after all, rents correlated very closely to home prices.
But even in the United States, analysts have long criticised the rental-equivalent approach for housing in the CPI. See for eg. this Seeking Alpha article dated 2007. If one believes that interests rates in the United States were unnecessarily low in the early 2000s leading to a housing bubble, this was likely because the Fed saw no reason to tighten because the CPI in turn showed no sign of acute inflation:
Doesn’t this suggest that OER is understating inflation? There is an argument that, rather than including observed rents, the existing price of a home should be in the consumer price index. The reasoning goes like this. A homeowner’s foregone income is based on the current sale price of the house. This means that the implicit rent, which is essentially a measure of the opportunity cost of owning rather than renting, should be based not on the rental market but on the price of the house.
Making this change in the consumer price index would make an enormous difference. To see how big, start with the fact that since 2000, the U.S. headline CPI has risen at an average annual rate of 2.75%, while the traditional core CPI has gone up 2.20% per year on average. If government statisticians had been using the price of homes sold rather than rents, consumer price inflation would have registered an annual increase of something like 4% per year – roughly one and one-quarter percentage points higher. And core CPI inflation would have been something like 3.8%; that’s more than one and one-half percentage points above the official reading. Had these been the inflation readings, it’s hard to imagine the Fed keeping their federal funds rate target below 2% for three years.
But how applicable is the US experience in justifying the Singapore adoption of OER for housing inflation? Home-ownership rates in the US are much lower (67.8% circa 2008) than that in Singapore, which means that OER in the US is a more accurate measure of housing inflation than in Singapore. Secondly, housing ownership in the United States unlike in Singapore really means home ownership ie. the owner owns both the house and the land underneath it, so housing is genuinely more of an investment compared to Singapore, where HDB flat owners typically own a 99-year lease from the HDB and whose flats may be forcefully re-acquired without compensation to flat owners if they break certain HDB rules.
The OER method also requires that the rental market in Singapore be broadly similar to the housing market in terms of housing composition, types and sizes. Unfortunately statistics on subletting approvals by the HDB show that the plurality of rented flats are 3-room flats rather than 4-room flats, which forms the plurality of owned HDB flats. Furthermore the rental equivalence approach implicitly assumes that housing prices mirror rental prices over the long term, but this assumption whilst reasonable at present, may not hold true in the future. And as stated above, Singapore’s extremely high home ownership rates mean that rental cost inflation or disinflation do not reflect actual costs borne by owners in OOH.
Lastly, the rental equivalence approach ignores fluctuations in the mortgage rate for home-owners which is linked to inter-banking rates in Singapore. Hence a rise in interest rates drives up the cost of servicing a mortgage, but this would not be reflected in rental costs and hence not the inflation rate; yet another troubling disconnect.
But if one does not use OER to determine housing inflation in the CPI, what are the alternatives? It turns out that there are various other methods other countries have chosen. See here, here and here for papers explaining in greater in detail the following.
Alternative ways of treating housing in the CPI
The acquisitions method treats housing as a durable good much like other goods in the CPI basket composition. It measures the cost of acquiring a new house, subsequent renovations and expansions all at market value and charges the costs right at the time of each purchase. While consistent with treatment of other goods such as clothing, food, and transport in the consumer price index, this approach ignores fluctuations in the interest rates and hence mortgage rate. Furthermore, housing is seldom purchased outright; much more often owners have to take out a mortgage which amortizes over a lengthy period, hence the mortgage rate greatly impacts the cost of living the CPI is supposed to track.
The user cost approach attempts to capture the actual costs and opportunity costs borne by home-owners such as mortgage rate, property taxes, maintenance fees and depreciation expenses as well as the initial price of the property, and the expected capital gains from selling the property at the end of a defined period. Whilst likely the most inclusive method of treating housing in the CPI, it suffers from the problem of being extremely complicated due to the consideration of so many factors. As a consequence countries which make use of this method typically do not utilise the full user cost method, but rather a simplified and modified version of it eg. Canada and Iceland.
The payments approach measures chiefly the actual cash outflows of home-owner in the OOH, such as mortgage payments, property taxes, maintenance charges, service and conservancy (S&C) charges. In some respects it may be viewed as a cash flow version of the user cost approach. However, this method generally excludes considerations of the downpayment and depreciation as well as any potential capital gains/losses.
The last approach to treating housing in the CPI is to exclude it entirely from consideration. While this approach was adopted by a surprisingly large number of countries such as Argentina, Belgium, Brazil, Russia, Korea and the whole of the Euro zone, it is hard to argue that Singapore should do so as well, especially since the average Singaporean household spend about a quarter of their monthly income on HDB mortgage payments.
If the rental equivalence approach is a flawed means of treating housing in the CPI basket, what approach should Singapore use then? While I cannot make any recommendation based on any rigorous economic reasoning (as I am not an expert on this), the user cost approach appears to me to be a leading candidate to replace the OER approach. Some of the drawbacks commonly associated with it (and the other two alternatives) are mitigated to some extent by certain unique features of Singapore’s public housing market.
For example, unlike in many other countries where housing asset can genuinely appreciate in value, in Singapore HDB flats can only depreciate because flats are leased for 99 years rather than owned. Secondly, the problem of separating the investment from consumption (which bedevils all three alternative approaches) is made simpler by the fact that HDB flat owners do not own land, but only leases. Thirdly the argument against the user cost (and other) approach(es) that any rise in interest rates would cause an increase in mortgage rates and hence inflation, thereby nullifying the effects of anti-inflationary monetary policy does not apply in Singapore since MAS does not target interest rates but rather exchange rates.
Whatever the chosen metric of measuring housing inflation, a major point of consideration is that any proposed approach, however complicated or simple should take into account housing prices, mortgage payments and other fees and taxes which home-owners actually have to pay and this itself is a strong reason to reject the rental equivalence approach which does not incorporate any such considerations.
As an example, Iceland uses a modified version of the user cost approach in calculating housing inflation. Curiously it does so for much the same reason I personally feel Singapore should strongly consider, namely that of high ownership rates:
The rental equivalence approach cannot be used in Iceland because of how small the rental market is and also because of the Icelandic market’s difference in composition from what generally applies to owner-occupied housing. The majority of Icelanders, or about 80%, live in owner-occupied housing according to the household expenditures survey.