Discussion of Cesa‐Bianchi, Cespedes, and Rebucci

There are many possible causes of housing cycles, such as housing demand, interest rates, and credit supply—not to mention irrational exuberance. This paper's goal is to assess the contribution of one specific factor, the worldwide supply of bank credit, to house price fluctuations. This is particularly relevant to the previous decade's boom–bust cycle, which coincided with a significant increase in the volume of international lending. Motivated by the observation that much of this lending flowed to emerging market economies (EMEs), the paper investigates the possibility that changes in global lending supply had a disproportionate impact on those economies. The analysis is methodologically innovative, using a panel vector autoregression (VAR) and utilizing instruments external to the system to achieve identification. Another contribution is the assembly of an impressive global data set of property prices, pulling together data from the Bank for International Settlements (BIS), the Organisation for Economic Co-operation and Development (OECD), and a variety other public and private sources. The relationship between credit and house prices is straightforward and at some level almost mechanical. After all, financing the purchase of houses with accelerating prices will inevitably require an ever-increasing volume of credit. What is interesting is that so much of the credit growth from 2002 to 2007 came from foreign sources. The price–credit link is clearly evident in Figure 1, which plots an index of the average level of real house prices for 57 countries along with the real “global liquidity” series used in the paper and interpreted as a measure of worldwide loan supply.1 The fit is striking. House prices and “global liquidity” both accelerated sharply in the early 2000s and contracted in 2008. In addition, there appears to be a robust long-run relationship between the two, although the authors do not model it explicitly. Aggregate House Prices and Liquidity. Notes: The aggregate house price is based on the data used in Kuttner and Shim (2012, 2013), calculated as the cumulative average real growth rate of house prices for each of the 57 countries in the data set. The real liquidity series is the same data as those used in the paper. This pattern does not prove that shifts in global credit supply cause house price fluctuations, of course. The “global liquidity” variable used in the paper, which is based on data from the BIS, is simply the observed amount of cross-border bank lending. Since this is an equilibrium outcome, it will be affected by both supply and demand factors.2 In order to disentangle the two, the authors make the plausible assumption that demand (“pull”) factors are domestic in origin while supply (“push”) factors are external or global in nature. The difficulty is that two of the external variables used in the analysis—the real exchange rate and the current account balance—are also influenced by domestic developments. The distinction between internal and external variables is therefore not sufficient to isolate credit supply. The authors’ identification strategy makes clever use of the “small country” assumption, which allows them to treat foreign variables as exogenous, and the use of monetary and financial variables from the U.S. as instruments. This is an excellent idea in principle, but in practice some of these variables are influenced by non-U.S. factors. The VIX, for example, rose sharply in 1997–98 during the Asian and Russian crises, reflecting instability in emerging markets rather than conditions in the U.S. Its validity as an instrument for global loan supply is therefore questionable. The salient empirical finding is that that property markets are considerably more volatile in EMEs than in advanced economies. More importantly, the VAR analysis shows that the EMEs respond much more strongly to “global liquidity” shocks than advanced economies: the current account response is twice as large in EMEs, and the house price response is five times the size. Missing from the analysis is a quantitative assessment of the shocks’ contribution to house price volatility, something analogous to the variance decomposition from a conventional VAR model. For example, it would be interesting to know what proportion of the doubling of Poland's house prices between 2005 and 2007 was due to global credit supply, as opposed to factors specific to that country. Such an assessment would have important policy implications. If the growth in domestic demand accounted for most of the boom, then contractionary monetary policy might have been warranted. On the other hand, contractionary policy would have been counterproductive if the lending was primarily the result of “push” factors, as tighter policy would have exacerbated the capital inflow and put upward pressure on the exchange rate. In addition, the results shed little light on what explains the disparities between countries in the response to “global liquidity” shocks. Why did Estonia experience a fourfold price increase followed by a collapse, while prices in Switzerland only took off after the global bust? The domestic variables in the regression play a role, of course, but so do monetary and exchange rate policies. More generally, the paper's broad-brush dichotomy between advanced economies and EMEs conceals a great deal of interesting heterogeneity within these categories. Very different results might have been obtained had the set of countries been partitioned along different lines. Stark intracategory differences are visible in Figure 2, which plots real house prices for four distinct sets of countries. Panels (a) and (b) in the top row show prices and “global liquidity” for advanced economies, distinguished according to the state of the housing market post-2007. The countries plotted in (a) are those with pronounced price declines, a group that includes the U.S. and Great Britain. Not all countries experienced busts, however. Panel (b) shows the aggregate house price for a distinct set of countries, including Switzerland and Canada, whose prices continued to rise, despite the pullback in cross-border lending. Another observation is that in both sets of countries, rapid house price growth preceded the 2002 surge in cross-border lending, and did not accelerate once lending took off. These patterns suggest that “global liquidity” did not play much of a role in these countries’ housing markets. This inference is consistent with the paper's findings, and not surprising in light of the depth of these countries’ financial markets. House Prices and Liquidity for Four Alternative Country Groupings. Notes:The “advanced economies, boom and bust” group (panel (a)) includes Spain, France, Great Britain, Greece, Ireland, Iceland, Italy, Malta, the Netherlands, New Zealand, and the U.S. The “advanced economies, boom but no bust” group (panel (b)) includes Austria, Australia, Belgium, Canada, Switzerland, Finland, Norway, and Sweden. The “emerging Asia + Latin America” group (panel (c)) includes China, Indonesia, Hong Kong, India, Korea, Malaysia, Thailand, Singapore, Taiwan, Chile, Colombia, Mexico, and Peru. The “transition economy” group (panel (d)) includes Bulgaria, the Czech Republic, Estonia, Croatia, Hungary, Latvia, Lithuania, Poland, Romania, Serbia, Russia, Slovenia, Slovakia and the Ukraine. See also notes to Figure 1. Panel (c) in the bottom row displays the relationship between house prices and “global liquidity” for the emerging economies in Asia and Latin America.3 Prices fell steeply in 1997–98 due to the Asian financial crisis. (The magnitude of the decline is attenuated by the inclusion of Latin America.) House price growth remained relatively restrained throughout the early- to mid-2000s, however, even as cross-border lending accelerated. And after a relatively mild downturn in 2008, prices continued their upward trend in spite of the drying up of “global liquidity.” The overall real appreciation over the 2000–12 period was approximately 40%, similar to house price growth in the advanced economies. The story is very different for the transition economies of Central and Eastern Europe, shown in Panel (d). The housing booms in some of these countries were nothing short of spectacular. In Estonia, for example, prices shot up by a factor of 4. As a group, those countries’ housing prices peaked in 2007 at 180% of their 2000 level and fell sharply during the global financial crisis. Interestingly, the relationship between house prices and “global liquidity” for this group appears to be much tighter than for the other three. Taken together, these observations suggest that the volatility the authors attribute to EMEs generally is actually specific to the transition economies of Central and Eastern Europe. And we know that many factors unique to the region were at least partly responsible to the countries’ boom–bust cycles. The prevalence of currency boards and hard pegs is one. In addition, having just gone through a period of radical economic liberalization, there was an explosion in housing demand and a seemingly limitless appetite for borrowing. With very little domestic saving and poorly developed financial systems, the lion's share of the funding naturally came from abroad—much in the form of loans denominated in Swiss francs and other foreign currencies. Overall, the paper contributes a great deal to our understanding of housing cycles. Its carefully crafted empirical analysis provides compelling evidence on the impact on house prices of shocks in the global supply of bank credit. The underlying economic sources of the shocks remain unclear, however. Was the previous decade's increase in lending due to the global savings glut? Financial innovation? Expansionary monetary policy? Formulating an appropriate policy response requires a better grasp of what exactly drives global loan supply. We also need to know more about the country-specific characteristics that affect the transmission of global credit supply shocks. Besides the monetary and exchange rate policies mentioned earlier, other relevant attributes include things like the structure of the mortgage market and the degree of financial development. Loan supply and demand will also be affected by regulatory measures, including the non–interest rate policies examined in Kuttner and Shim (2012, 2013). The emerging market/advanced economy breakdown is not informative in these respects. For example, Hong Kong is similar to Bulgaria in that both have currency boards, but by any measure Hong Kong's financial system is much more highly developed than Bulgaria's. Similarly, the financial systems of Bulgaria and Serbia are probably roughly comparable in terms of depth and development, but unlike Bulgaria, Serbia had a managed float. Distinguishing between economies along some of these other dimensions is likely to yield a better understanding of the transmission mechanism. Kenneth N. Kuttner is at the Department of Economics at Williams College (E-mail: [email protected]).

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.