Taking a deeper look at the property price models of the RBA

Welcome to Finance and Fury. This episode we’ll take a deeper look into the RBA property market models and how different inputs affect prices.

Potential models that give the ability forecast future growth of the market based around assumptions – from a study by the RBA – A Model of the Australian Housing Market

• Not set in stone – it is a best guess model – the accuracy is built around the assumptions – But the RBA built an empirical model of the Australian housing market that quantifies interrelationships between the factors that drive property price growth
• Looking back over the past 30 years – the Australian housing prices have increased on average by 7¼ per cent per year – since the inflation-targeting period since the mid-90s – by around 7 per cent per year – However, these averages mask three distinct phases:
  1. During the 1980s, annual housing price inflation was high, at nearly 10 per cent on average, but so too was general price inflation. In real terms, housing price inflation during the 1980s was relatively low, at 4 per cent per annum compared with 4.5 per cent during the period from 1990 to the mid 2000s,
  2. The 1990s until the mid 2000s were marked by quite high housing price inflation, of 7.2 per cent per annum, on average, in nominal terms – but 4.5% in real terms
  3. Since 2010 – Annual nominal housing price inflation over the past decade was lower than either of these periods, at a little over 5 per cent on average – and 2.5 per cent over the past decade in real terms
• So what are the major factors and how much do they have an effect on price

Wont go through the whole document or all the RBA papers – but give a summary

In this RBA model – the findings

• They find that low interest rates (partly reflecting lower world long-term rates) explain much of the rapid growth in housing prices and construction over the past few years.
• Another demand factor, high immigration, also helps explain the tight housing market and rapid growth in rents in the late 2000s. A large part of the effect of interest rates on dwelling investment, and hence GDP, works through housing prices.
• The Australian housing market shows strong relationships between interest rates, investment, rents and prices.
  1. The RBA paper combines these relationships in one – hopefully realistic – model. The model provides internally consistent projections for housing construction, prices and rents. It estimates responses to interest rates, allowing for feedback between quantities and prices.
• It helps explain historical developments and some of the key relationships that they found include:
  1. Interest rates, income and housing prices have strong and clear effects on residential construction.
  2. Dwelling completions and changes in population explain the rental vacancy rate.
  3. The vacancy rate has a strong and clear effect on rents.
  4. Interest rates, rents and momentum have large effects on housing prices.
  5. Housing prices and construction are mutually determined, so examining bivariate relationships in isolation can be misleading.
• Some of these observations are not new – but this paper examines these relationships

What are the model outcomes – responses to variables – bring the equations together

• Responses to Interest Rates – Shows that a cash rate change that is expected to be long-lasting feeds one-for-one into long-term interest rates and the user cost of housing – but the temporary change having much less effect.
  1. As interest rates drop – so does user costs – Changes in the user cost give rise to similar, but lagged, changes in the rental yield (lower the cost the higher the net income from property) – which involves substantial increases in housing prices
  2. The combination of lower interest rates, higher housing prices and higher income boost dwelling investments – which increases the number of dwelling stock as well as rental vacancy rate
     1. Rents initially rise due to the income boost from lower interest rates, but as extra supply builds, they begin to fall.
     2. The signs of the effects on vacancies and rents both vary with time and with the expected duration of the shock.
  3. Since 2011 – the cash rate has fallen from 4.75% to 1.5% (3.25% drop) – at the time of the report last year – since then obviously lower to 0.25% – the user cost has fallen from almost 5 per cent to around 3½ per cent (1.5%)
     1. The decline in the user cost reflects a fall in long-term real interest rates (only be half), which in turn reflects falls in global rates and expectations that the decline in rates will be persistent.
     2. The model estimates that the reduction in real interest rates accounts for most of the subsequent boom in dwelling prices and a large part of the boom in dwelling investment – increasing the supply of property
        • The increase in housing supply boosts the vacancy rate and reduces rents. However, these effects are offset by the effect of higher income, with neither the vacancy rate (bottom left) nor rents (bottom right) being much changed on net.

• Responses to Population Growth – Reflecting a surge in immigration, year-ended growth in the adult population (15 years and older) rose from 1.5 per cent in 2005 to 2.4 per cent in 2008.
  1. To assess the effects of this surge, RBA ran a simulation in which adult population growth continues to grow at its 2005 rate of 1.5 per cent.
     3. With per capita income unchanged by assumption- dwelling investment also increase by about 3.3 per cent. This raises housing supply slightly. However, the short-run boost to housing demand is much larger, leading to a fall in the rental vacancy rate to a near-record low of 1½ per cent in 2008 (3rd row, left).
     4. Rents, which were already growing quickly, accelerate to grow 4 percentage points faster than the overall rate of inflation. Without the extra population, our simulation suggests that real rents would have only grown by 2 per cent a year, as shown by the 3rd row, right panel. The cumulative result was that real rents were 9 per cent higher in 2018 than they would have been otherwise. The increase in rents gradually flows on to a similar increase in dwelling prices, although this effect is small relative to the effect of interest rates, discussed in the previous section.

• Responses to Completions – increase in construction as a deviation from baseline – assume that building approvals increases by 10 per cent for one year – represents about 21,400 extra approvals for new dwellings
  1. The extra ‘supply’ (as it is commonly termed) would increase the vacancy rate and hence lower rents and housing prices. The proportionate response of rents and prices (0.4 per cent) is 2.5 times as large as the increase in the number of dwellings (0.16 per cent). This ratio (2.5) represents the inverse of the elasticity of housing demand. It also applies to larger and more sustained shocks. As a rule of thumb, every 1 per cent increase in the number of dwellings (when driven by an increase in supply) lowers the cost of housing by 2½ per cent. Our estimate of the elasticity of housing demand lies well within the range of other estimates. Abelson et al (2005) estimate that a 1 per cent increase in the Australian housing stock per capita leads to an estimated decrease in real housing prices of 3.6 per cent in the long run. Girouard et al (2006) summarise ten international studies, which have an average estimate of 3.1. Two recent and arguably more thorough studies point to smaller effects. Albouy, Ehrlich and Liu (2016, Table 3) find that a 1 per cent increase in real housing expenditure in the United States is associated with less than a 2 percentage point reduction in price. Oxford Economics (2016) find that a 1 per cent increase in the number of houses in the United Kingdom would reduce house prices by 1.8 per cent and cite (their Figure 20) several other elasticities that range between 1.1 and 2.2.
• Responses to Changed Price Expectations and the User Cost – Our measure of the user cost assumes that home buyers expect real constant-quality housing prices to continue rising at their post-1955 average rate of 2½ per cent a year.
  1. This is a simple assumption that is consistent with some of the main features of the data. However, forecasts from RBA model imply that real housing prices (measured with a different quality adjustment) will grow at an annual average rate of 0.2 per cent over the next ten years.
     1. This scenario can also be interpreted as capturing the effect of an exogenous fall in housing prices if the model to be extended to include taxes on housing, this is an important channel through which they would operate. As shown in the top left panel, expected capital appreciation declines 2½ percentage points. The user cost (not shown) rises by the same amount. Housing prices (top right and middle left) then take a long time to adjust, falling gradually, but substantially, to be one-third lower after five years.
     2. housing bubbles do not ‘burst’, they gradually deflate. This scenario is extremely unlikely: nothing like it has happened in Australia before. However, in scale and duration, it resembles the largest housing collapses seen during the global financial crisis (Ireland, Spain, United States), so is relevant as a worst-case scenario to be guarded against. Falling house prices result in large falls in investment (middle right), which reduce vacancies (bottom left), boosting rents (bottom right).

• The decline in construction dampens the fall in prices. The limitations of this partial equilibrium exercise should be emphasised. In a more complicated model, falling house prices would reduce household wealth and hence consumption. The net worth of financial intermediaries would fall. Offsetting this, interest rates would fall, moderating the macroeconomic impact. That said, the housing market response would be an important part of the overall effect.

Their model quantifies some important developments:

1. The model suggests that much of the strength in housing prices and construction over the past few years can be explained by the fall in interest rates – some of this fall reflects lower world real interest rates and some is cyclical.
2. A large part of the effect of interest rates on dwelling investment (and hence real GDP) occurs through the channel of housing prices.
3. The model suggests that an increase in population growth will reduce rental vacancies, boost rents and housing prices, and increase construction. This helps to explain developments following the immigration surge of the mid 2000s.
4. The model is consistent with some important longer-run trends. Construction activity is approximately cointegrated (trends together) with income, although the housing stock is not. The rental yield is cointegrated with the user cost of housing. Rents tend to grow slightly faster than inflation but slower than income per capita

Conclusions –

• Looking ahead, it seems unlikely that there will be a return to the rather extreme conditions of the past few decades when significant increases in household debt supported high housing price growth. Nonetheless, protracted periods of changes in population growth that are not met by adjustments in dwelling supply could lead to periods of sizeable changes in housing price growth. One important factor for housing price growth is the ability of the supply of new dwellings to respond to changes in demand. The significance of this is made clear by the recent increases in higher-density housing and lower growth of those prices relative to prices of detached houses, whose supply has been less responsive.
• Varied bag – and before the recent economic downturn – But the major factors have been
  1. Interest rates declining – rates would need to continue to decline
• But not looking as good as the past 20 years worth of growth as a percentage – may become lower with either

Thank you for listening to today’s episode. If you want to get in contact you can do so here: http://financeandfury.com.au/contact/

https://www.rba.gov.au/publications/bulletin/2015/sep/3.html

https://www.rba.gov.au/publications/rdp/2019/pdf/rdp2019-01.pdf