Correlations rise and fall but the likely pattern of recovery across Europe suggests that international diversification will continue to deliver superior long-term returns for the risk taken than a purely domestic-based strategy. Mark Callender reports

The traditional argument for cross-border property investment is that local occupier markets are subject to their own peculiar demand and supply forces and tend to peak and trough at slightly different times. This lack of synchronisation means that returns on an international property portfolio should be less volatile than those for a purely domestic portfolio. In theory, adding international property shifts the efficient frontier for a portfolio to the left, enabling an investor either to achieve the same level of return at lower risk, or to achieve a higher return for the same level of risk.

Unfortunately, the global financial crisis which began in 2007 smashed this logic. Banks across Europe (and indeed the world) simultaneously stopped lending on property, the commercial mortgage-backed security (CMBS) market froze and investors lost confidence, occupiers postponed signing new leases and tenant insolvencies climbed. As a result, property yields rose in virtually every European market in 2008 and rents fell, or were at best flat in almost every market in 2009. Property markets, which in the past had followed different cycles, were suddenly highly synchronised and all moving in the wrong direction.

Figure 2 shows the simultaneous nature of the downturn in office markets by measuring the average correlation in prime office rents across 33 European cities. Prime office rents fell in all 33 cities in 2009 and the average correlation coefficient increased from 0.4 in 2006 to 0.7 last year. The figure also shows the corresponding correlation in prime retail rents across 28 European cities. While the average correlation coefficient for prime retail rents also rose after 2006, in general, retail markets were more diverse and, whereas prime retail rents fell in 16 of the 28 cities in 2009, they were stable in nine cities and rose marginally in three German cities.

It would appear that retail markets are more insulated from the global economy than office markets and therefore offer greater diversification benefits. This greater diversity is probably due partly to variations in regulations on retail prices and opening hours, as well as to retail planning policies which differ between countries and regions. In addition, the degree of structural change in retail, due to the impact of internet shopping and the entry of new discount retailers, varies significantly across countries.

The other conclusion we can draw from the correlation analysis is that the degree of synchronisation in European property markets is cyclical and tends to rise during recessions - the early 1990s and late 2000s - and fall during periods of economic recovery, as different markets revive at different times. The key issue now for cross-border investors is whether the future upturn in European property markets will be highly synchronised, or resemble the staggered recovery of the mid-1990s? There are a number of reasons why some markets will recover quite quickly while others will probably remain in the doldrums for several years.

The first obvious point is that future economic growth, and therefore tenant demand, is likely to vary widely across Europe. According to the latest forecasts high levels of inward investment should mean that the Czech Republic and Poland lead the economic recovery over the next four years with growth of over 3% per year. Norway, Sweden and the UK should also see above-average growth, thanks to the depreciation of their currencies against the euro, and Norway has the added advantage of oil. Germany, too, is forecast to outperform the euro-zone average with growth of 2% per year, its best relative economic performance since reunification. At the other extreme, economic growth in Italy and Spain is forecast to average just 1% a year over the next four years, and the recoveries in the Netherlands, Portugal and Switzerland are expected to be relatively sluggish.

In part this range in forecast growth is simply a legacy of the credit boom that preceded the financial crisis and the damage that emergency measures have inflicted on government finances. Thus in Ireland, Spain and the UK, and to a lesser extent in the Netherlands and Switzerland, the pressure on households to repay debt and raise savings is likely to restrain consumer spending for a number of years. Likewise, while virtually all European governments will this year be in breach of the Maastricht criteria limiting fiscal deficits to 3% of GDP, the problem is most acute in Greece, Ireland, Spain and the UK. In these countries the need to reduce the government's fiscal deficit through tax increases and spending cuts is likely to act as a significant drag on economic growth for several years.

Another crucial factor sometimes overlooked is international competitiveness. In Germany and to a lesser extent France, wage restraint and improvements in productivity have more or less flattened unit labour costs over the past 10 years. In Italy and Spain unit labour costs have risen by 25-50% over the past 10 years, reflecting the automatic indexation of wages to inflation in Spain and poor labour mobility in both countries, so that there has been relatively little migration from regions of high unemployment to areas of low unemployment.

The fundamental point is that the creation of the euro and the introduction of a single interest rate regime have not achieved convergence in economic growth across the euro-zone, because they have not been accompanied by labour market reforms. As a result, because Italy and Spain cannot devalue, the euro is likely to act as a wedge between northern and southern Europe over the next few years, rather than a force for economic integration.

It should also be remembered that national forecasts only ever tell part of the story and that the future range in economic growth at the city level will be significantly wider, due to the dominance of certain industries in certain locations. Thus Frankfurt and London are major financial centres, Berlin, Brussels, Rome and The Hague are primarily administrative centres, high-tech and luxury goods manufacturing is clustered around southern Germany, Switzerland and northern Italy, and tourism is a major industry in many parts of southern Europe. Arguably, globalisation and the transfer of low-value-add assembly plants to Asia and central Europe has increased the degree of specialisation in western Europe's regional economies, rather than made them more homogeneous.

In short, there is a strong case for believing that the strength of the economic recovery and hence tenant demand will vary widely across Europe over the next few years. And, of course, tenant demand is only part of the rental equation.

The other key element is how much empty or under-utilised space is currently sitting on the market. According to PMA the average vacancy rate in shopping centres ranges from around 2% in France to 12% in the UK, and vacancy rates in the office sector range from 5%-8% in London's West End, Paris and Vienna to 18-19% in Amsterdam and Frankfurt.

Those markets with low vacancy rates have an important lead; French shopping centres and London, Paris and Vienna offices will be among the first markets to see a recovery in rents in 2011, or 2012. By contrast, office markets such as Amsterdam and Frankfurt that have chronically high vacancy rates are likely significantly to lag behind. One of the challenges, or perhaps opportunities, in these cities is what to do with the stock of empty, ageing office buildings.

Finally, the other potential source of differences in the timing of cycles in property capital values and total returns is local variations in investor sentiment, which in turn drive changes in property yields. While the creation of the euro has not led to convergence in economic growth, it almost certainly has led to greater uniformity in the movement in property yields within the euro-zone in two ways.

First, to the extent that short-term interest rates influence property yields, there are no longer national differences, at least in nominal terms. Second, by eliminating exchange rate risk, the euro has encouraged property investors to enter new countries where they perceive that yields are attractive, given the prospects for future rental growth. Inevitably, this creates something of a paradox, similar to tourists seeking unspoiled destinations. While one of the aims of cross-border investment is to achieve diversification, international flows of capital will begin to dilute the benefits.

Yet it would be a mistake to assume that there is now just one large unified pan-European property investment market and that there are no longer any national differences in investor sentiment towards property. Figure 4 shows the increase in prime office gross yields in 29 cities between their low point before the global financial crisis (generally in the second half of 2007) and March 2010. Undoubtedly, there is some evidence of rational pricing on a European scale. Thus, office yields in some of those cities most affected by the financial crisis, such as Budapest, Dublin and Madrid, jumped by 200bps or more, whereas office yields in some of the main administrative centres, such as Brussels and Rome, increased by only 50-75bps.*

However, it is also apparent that there was a strong national pattern and that yields in Denmark, Germany and Switzerland rose by much less than yields in France, the Netherlands and Sweden.

For example, whereas prime gross yields on Hamburg and Munich offices were 50-100bps higher than those on Paris and Stockholm offices at the end of 2007, they were 25-50bps lower in March 2010. This is a reminder that, despite the significant increase in cross-border property investment over the past decade, pricing in most markets - offices in London, Paris and central Europe are probably the exceptions - is still generally dictated by local investors.

In Denmark, Germany and Switzerland it appears that investors regard property as a store of value in uncertain times, possibly because it is a physical asset and vacant space can usually be relet if a tenant defaults. By contrast, in France, Sweden and the UK, the consensus appears to be that property is not a safe haven, but a high-risk, cyclical asset.

In conclusion, there is still a strong case for a well-diversified pan-European property portfolio. Local variations in economic growth and in vacancy rates will mean that rents will recover at different speeds in different European markets, leading to a repeat of the staggered upswing seen in the mid-1990s. In addition, capital values and total returns will continue to be influenced to some extent by national variations in investor sentiment towards property and by diverse movements in yields.

As a result, unless there is a new shock to the global financial system, an international property portfolio should carry significantly less risk than a purely domestic portfolio.

* Note: The figure understates the full extent of the increase, because in some cities prime office yields peaked in mid-2009 and have since fallen.