Thomas Quistgaard Pedersen

Pitfalls in VAR Based Return Decompositions: A Clarification

We analyze the pitfalls involved in VAR based return decompositions. First, we show that recent criticism of such decompositions is misplaced and builds on invalid VAR models and erroneous interpretations. Second, we derive the requirements needed for VAR decompositions to be valid. A crucial – but often neglected – requirement is that the asset price needs to be included as a state variable in the VAR. In equity return decompositions this requirement is equivalent to including the dividend–price ratio in the VAR. Finally, we clarify the intriguing issue of the role of the residual component in return decompositions. In a properly specified first-order VAR, it makes no difference whether cash flow news or discount rate news is backed out residually, and it makes no difference whether both news components are computed directly or one of them is backed out residually.

The Log-Linear Return Approximation, Bubbles, and Predictability

We study in detail the log-linear return approximation introduced by Campbell and Shiller (1988a). First, we derive an upper bound for the mean approximation error, given stationarity of the log dividendprice ratio. Next, we simulate various rational bubbles which have explosive conditional expectation, and we investigate the magnitude of the approximation error in those cases. We find that surprisingly the Campbell-Shiller approximation is very accurate even in the presence of large explosive bubbles. Only in very large samples do we find evidence that bubbles generate large approximation errors. Finally, we show that a bubble model in which expected returns are constant can explain the predictability of stock returns from the dividend-price ratio that many previous studies have documented.

Predicting Returns and Rent Growth in the Housing Market Using the Rent-to-Price Ratio: Evidence from the OECD Countries

We investigate the predictive power of the rent-to-price ratio for future real estate returns and rent growth in 18 OECD countries over the period 1970 to 2011. First, we document that in most countries returns are signi?cantly predictable by the rent-price ratio. An increase (decrease) in the ratio signals a future increase (decrease) in returns. Second, there are large cross-country di¤erences in how the rent-price ratio predicts rent growth. For some countries the direction of predictability is negative, for other countries it is positive. Third, the predictive patterns are highly dependent on whether returns and rents are measured in nominal or real terms. Finally, there is some evidence of sub-sample instability in the predictive patterns, especially wrt. rent growth predictability. The predictability tests are conducted within a restricted VAR framework based on the dynamic Gordon growth model. This model implies restrictions across the VAR parameters that can be used to construct powerful tests of predictability.

Predictable Return Distributions

This paper provides detailed insights into predictability of the entire stock and bond return distribution through the use of quantile regression. This allows us to examine speci?c parts of the return distribution such as the tails or the center, and for a suf?ciently ?ne grid of quantiles we can trace out the entire distribution. A univariate quantile regression model is used to examine stock and bond return distributions individually, while a multivariate model is used to capture their joint distribution. An empirical analysis on US data shows that certain parts of the return distributions are predictable as a function of economic state variables. The results are, however, very different for stocks and bonds. The state variables primarily predict only location shifts in the stock return distribution, while they also predict changes in higher-order moments in the bond return distribution. Out-of-sample analyses show that the relative accuracy of the state variables in predicting future returns varies across the distribution. A portfolio study shows that an investor with power utility can obtain economic gains by applying the empirical return distribution in portfolio decisions instead of imposing an assumption of lognormally distributed returns.

Testing for Explosive Bubbles in the Presence of Autocorrelated Innovations

We analyze an empirically important issue with the recursive right-tailed unit root tests for bubbles in asset prices. First, we show that serially correlated innovations, which is a feature that is present in most financial series used to test for bubbles, can lead to severe size distortions when using either fixed or automatic (based on information criteria) lag-length selection in the auxiliary regressions underlying the test. Second, we propose a sieve-bootstrap version of these tests and show that this results in more or less perfectly sized test statistics even in the presence of highly autocorrelated innovations. We also find that these improvements in size come at a relatively low cost for the power of the tests. Finally, we apply the bootstrap tests on the housing market of OECD countries, and generally find less strong evidence of bubbles compared to existing evidence.

A New Index of Housing Sentiment

We propose a new measure for housing sentiment and show that it accurately tracks expectations about future house price growth rates. We construct the housing sentiment index using partial least squares on household survey responses to questions about buying conditions for houses. We ?find that housing sentiment explains a large share of the time-variation in house prices during both boom and bust cycles and it strongly outperforms several macroeconomic variables typically used to forecast house prices.