news article background image
2017-08-31Macro `n Cheese

Weekly analyses from our chief economist, Roger Josefsson.

Welcome to the first of many! This week's topic is the Swedish housing market.

A quick introduction

What you have here is the first of many posts by Roger Josefsson, the chief economist at Macrobond. Every week Roger will be writing up an analysis on a topic of his choice, and presenting it using Macrobond from start to finish. You’ll get an idea of the data we have, the analyses you can perform and just how easy it is to present and share work in this way.

Those of you who are Macrobond users can click on any of the charts to open them in the application, and see exactly what data was used, the analyses that were applied, as well as chart styles and formatting. Not to mention Roger’s riveting analysis of current developments across markets around the globe. Speaking of Roger’s perspective…


The world’s strongest economy has one fatal flaw

Sweden’s Housing market

A couple of years ago – in the wake of the financial crisis – the Secretary-General of OECD, Angel Gurria, labelled Sweden “the Pippi Longstocking economy” referring to the beloved story of “Pippi Longstocking, the world’s strongest girl” by author Astrid Lindgren.

However, had Gurria ventured deeper into the Swedish economy he would for sure have noticed what many investors, both domestic and international, have long since been obsessing about; the Swedish housing market.


Macrobond chart

Figure 1: Booming housing prices

Note: HOX residential price indices are spliced together with SCB and BIS data. Even longer time series can be constructed with the Riksbank’s historical house price indices.


Economists like looking at house prices in real terms, controlling for the developments of prices in general. This makes the hockey-stick shape of Swedish house price developments even more eye-catching.


Macrobond chart

Figure 2: A good investment, if there ever was one

Note: HOX residential price indices are spliced together with SCB and BIS data and deflated with Statistics Sweden’s CPI-index and historical shadow CPI-index. Even longer time series can be constructed with the Riksbank’s historical house price and inflation series.


Lower Nominal Interest Rates

Lower nominal interest rates are, of course, one of the main culprits; mortgages have become cheaper (yes, money illusion) since the Swedish financial and banking crisis in the early 1990’s. As amortisation periods were gradually extended ad infinitum1. This allowed households to climb the housing ladder by levering up, while simultaneously keep their monthly mortgage costs intact (or even lowering them). As a result, the household debt ratio has been on the rise for an extended period of time.

1 The Swedish FSA re-introduced mandatory amortisation in 2016, and look to sharpen that regulation further come 2018. In addition a loan-to-value cap was introduced in 2010 and a host of measures directed towards reigning in bank lending has also been introduced.


Macrobond chart

Figure 3: Household debt ratio is troubling


A classic example of a search-market

The housing market is a classic example of a search-market, which, implies that the prices for the whole market (and thus for household residential wealth) are decided by only a few transactions. However, while prices (and wealth) may change quite rapidly, debts remain intact (in nominal terms).


Keep in mind…

Before moving on to how the above developments can be interpreted, it is probably good for you to know that for a longer time I have taken a negative (and erroneous, which my better half never fails to remind me) position on the Swedish housing market. That said, my views have never incorporated a set date for an inevitable house price collapse, but rather about if current prices are long run sustainable. Here, I take my cue from Keynes who put it rather eloquently:

“Markets can remain irrational a lot longer than you and I can remain solvent.”

After all, we cannot positively say “This is nuts, when’s the crash”. A possible way out of the predicament would be to see house prices rise less than incomes over a period of time, to restore some notion of equilibrium. Admittedly, such slowly deflating asset bubble resolutions are quite unusual. None the less, it is apparently the objective of the Swedish FSA and other concerned authorities to engineer such an outcome.


How is the housing market analysed?

Error Correction Models

The consensus among Swedish analysts is that house price valuations are stretched but, by and large, can be explained by fundamentals. To fortify these arguments, a number of statistical models have been constructed over the past few years. Most of them belong to a category labelled Error Correction Models (ECM). The advantage of such models is of course that they are quite straight forward demonstrating the relationship between house prices and other variables (interest rates etc.) closely tied to the housing market. The Riksbank has, among other things, estimated such models a couple of times2 and come to the obvious conclusion that “house prices are in line with fundamentals”.

2 One is from the Riksbank’s 2010 report on the Swedish housing market and the other from a Riksbank Working Paper draft.


Macrobond chart

Figure 4: Statistical models suggest that house prices are richly valued

Note: ‘HPR’ is house prices in levels, ‘d_p’ is house prices in logarithmic differences.The Riksbank has used an after tax interest rate, which we have replicated by multiplying the interest rate series with (1-td), where td is the tax deductibility ratios employed in Sweden over the sample period.


My tentative update of the Riksbank’s models suggests that current house prices are a tad high, albeit only a moderate 10%, which should be of no concern to any policy maker.

Of course, these models have a place but they also suffer from the (probably overly strict) assumption that correlated trends in the underlying data are constant. It’s hard to fathom, for example, how interest rates could fall much further…


Fundamental models

In more fundamental models, with long run stable equilibriums, residential prices are explained by three main variables:

  • User costs (i.e. interest rates)
  • Budget share of housing
  • Residential capital stock (in relation to the population).

When these three variables are “eternally stable” the housing market is in equilibrium and house prices are at fundamentally correct levels.

Niels Bohr is often credited for having said:

“Prediction is very difficult, especially about the future”.

Given that I have a hard time hitting GDP-growth forecasts for the current quarter, deciding eternally stable levels for three composite variables is not something I will spend my days on.

Some economists are, nonetheless, undeterred by the difficulties in deciding equilibriums. And, even if their essays have attained a certain amount of patina, the Riksbank, and the Fiscal Policy Council’s studies3 concluded that Swedish house prices could well be up to 40% too high. Unfortunately, minute changes to the assumptions in these models result in house prices being anything from correctly valued to dramatically overvalued.

3 Fiscal Policy Council report, Riksbank report on the housing market, and box in “The Swedish Economy Report” from National Institute for Economic Research (only available in Swedish).


Getting back to basics

For reasons already disclosed, I tend to prefer simpler, more robust measures. In the simplest fundamental model, asset prices are explained by the future utility (income) over the asset’s life-span. These future incomes are discounted to present day with what I could have earned had I put my life savings elsewhere (the alternative cost is expressed as an interest rate, in this case). To some extent this interest rate is also a product of how much general incomes are expected to rise in the coming years.

It is, honestly speaking, all but impossible to calculate future incomes or interests with any degree of reliability. Due to this, and much else, economists often use historical data – and not just a couple of years or decades worth of data – but many decades worth of data (not devoid of challenges either). Because, over time and on average, we may be more accurate when judging how high asset prices should be in relation to our incomes. I take some reassurance in the existence of a similar indicator for the stock market, named for another elderly gentleman, Warren Buffett.


Macrobond chart

Figure 5: The Buffett Indicator

Note: In this graph, we have used a spliced house price index and a spliced per-capita-income measure. In addition, we have added a similar Euroland measure for reference.


While house prices in Euroland are only marginally (ca 10%) above their long-term average, Swedish house prices are almost 50% higher when making the same comparison. Either interest rates (risk premia) should have structurally decreased or our preferences have changed decidedly for a higher share of budget for housing. If not, the alternative is that income growth will exceed house price growth for the foreseeable future (let’s hope).


The other side of the coin

So, is this it? Are Sweden and Swedish housing markets heading towards the precipice?

No, not necessarily.

Despite my pessimistic inclination, I willingly admit that there are valid arguments that this time, in Sweden, is different.


Good old supply and demand

Demand still seems to be strong, given the high population growth, and despite supply increasing rapidly.  As a proxy for the supply-demand situation, we can look at the number of house completions in relation to demographic developments. Using this, admittedly simple, indicator it seems that supply is not yet keeping up with demand.


Macrobond chart

Figure 6: Housing is still in demand


Household savings

Considering the ongoing extension of average amortisation periods, low inflation and interest costs, discussed earlier, it would be quite alarming to see households’ savings decreasing. However, the savings ratio has been on a clear upward trend and is currently near historically high levels.


Macrobond chart

Figure 7: The savings ratio is at historically high levels…


Household wealth

High savings and positive asset price developments at large, have contributed to a strong increase in household net wealth. And even if the asset side is price sensitive, Swedish households’ net wealth is considerably higher than in most (all) countries that experienced a large drop in house prices during the financial crisis.


Macrobond chart

Figure 8: …and net wealth is pretty awesome!

Note: Analysts often put assets and liabilities in relation to disposable incomes. However, I’m on a one-man quest to change analysts into using GNI (per capita) since it is a much better measure of household welfare. Putting Swedish net wealth in relation to disposable income would push household net wealth up to around 800%.


The government’s financial situation

Not only that, but the Swedish government’s financial situation is likewise considerably stronger than in most comparable countries, providing an important shock absorber (via very strong automatic stabilisers) should the housing market and economy fall into a recession.

Okay, but what about household debt?

The last of the, in my opinion, really strong arguments for Sweden being different runs contrary to the Riksbank and Governor Ingves’ main contention; The “worryingly” high household debt levels. To support this view, the Riksbank often shows a graph of the household debt ratio, usually in relation to many other countries with lower debt ratios.


Macrobond chart

Figure 9: The Riksbank has an opinion. But is it balanced?


I maintain that such a graph is quite misleading. Disposable incomes are measured after tax and in Sweden (as in the other Nordic countries) most welfare services are tax financed4. To address these shortcomings, Statistics Sweden is actually producing a measure called adjusted disposable income5. Interchanging these two measures in the denominator and superficially pondering the concept of debt (don’t try this at home) we can actually create a rather large span of debt ratios that all have something to say for them. Importantly, no matter how we create them they are always lower than the Riksbank’s measure of the debt ratio (the black line) and puts Sweden more towards the midst of the pack in terms of debt ratios. – Go figure!

4 OECD has some great data on this.

5 It is a cautious estimate of how much welfare services that is individually consumed.


Get these posts delivered to your inbox

If you like what you read, why not subscribe to have these analyses delivered to your inbox every Friday?

Sign up


Disclaimer

We don’t usually have views and opinions about economic and financial states of affairs, (not ones that we express publicly as a company, anyway). We do believe, however, that people can and do appreciate a variety of perspectives. What you’ve just read is the perspective of our resident chief economist. While we think he’s very smart, Macrobond Financial does not expressly endorse the views he presents here. And, as the old adage goes, you shouldn’t believe everything you read (not without finding the data, performing a few analyses and presenting it in a nice chart). We want to make it clear that we are not offering this information as investment advice. That being said, if you have the application you can easily check everything that’s mentioned here, and decide for yourself. If you don’t have the application, now you have a great reason to get it.