Noisy Proxy Research Articles

The temporal stability of children's neighborhood experiences: A follow-up from birth to age 15

Despite increasing attention being paid to the temporal dynamics of childhood disadvantage, children’s neighborhood characteristics are often measured at a single point in time. Whether such cross-sectional measures serve as reliable proxies for children’s long-run neighborhood conditions depends on the stability in children’s neighborhood experiences over time. We investigate the temporal stability in children’s neighborhood environment, focusing on two of the most commonly studied neighborhood characteristics: The ethnic composition and mean income of the neighborhood. Using Dutch population register data, we follow an entire cohort of children from birth up until age 15. We use year-toyear correlations in the percentage non-Western minorities and the mean income in the neighborhood to evaluate the temporal stability of children’s neighborhood experiences. Children’s neighborhood characteristics were found to be more stable over time with regard to ethnic composition than with regard to income. Children who had moved at least once were found to have lower stability in neighborhood characteristics than children who never moved. Finally, neighborhood experiences were found to be more stable over time for ethnic minorities than for the native Dutch, although differences were small with regard to income. Single point-in-time measurements of neighborhood characteristics are reasonable proxies for the long-run ethnic composition of children’s neighborhood environment, but rather noisy proxies for the long-run income status of their neighborhood, particularly for those who moved.

Stocks Should Be Valued with a Term Structure of Required Returns

In theory, the intrinsic value of a stock is determined by discounting the projected cash flow by a term structure of time-varying required returns. In practice, investors typically use a single discount rate, often a Treasury rate plus a risk premium. A single discount rate is a noisy proxy for the full term structure and can cause large valuation errors. If a single discount rate is used, the yield to perpetuity recommended by John Burr Williams is likely to be a better approximation to a complete term structure than are short-term rates.

Commentary on: Implied Cost of Equity Capital Estimates as Predictors of Accounting Returns and Stock Returns

The expected rate of equity returns is a central input into various managerial and investment decisions that affect the allocation of scarce resources. Research on capital markets has devoted significant effort to studying how and why expected returns vary over time and across firms. Cochrane (2011) called these questions the central organizing agenda in contemporary asset-pricing research. At the heart of this research agenda lies a longstanding measurement problem: ex-ante expected returns are unobservable and ex-post realized returns are noisy proxies (Campbell, 1991; Vuolteenaho, 2002). Since Botosan (1997), the accounting literature offered a promising solution to this measurement problem: the development of a novel class of expected-return proxies (ERPs), collectively known as the implied cost of equity capital (ICC).

Challenges and perspectives for large‐scale temperature reconstructions of the past two millennia

AbstractKnowledge of the temperature variability during the last one to two millennia is important for providing a perspective to present‐day climate excursions, for assessing the sensitivity of the climate to different forcings, and for providing a test bed for climate models. Since systematic instrumental temperature records only extend back to the nineteenth century, such knowledge mainly relies on climate‐sensitive proxy data. Here we critically assess some of the many challenges related to large‐scale multiproxy temperature reconstructions. We begin with a review of available large‐scale temperature reconstructions, focusing on the differences in low‐frequency variability and the response to natural forcings such as major volcanic eruptions and changes in total solar irradiance. Then, we discuss different proxy selection strategies, review previously used reconstruction methods, and discuss their ability to reconstruct the amplitude of the low‐frequency variability. To shed additional light on the challenges of large‐scale reconstructions, we investigate the spatial and temporal correlation structures in the observed temperature field and discuss the implications of these correlation structures regarding the required number and positions of proxies. We demonstrate how the unavoidable uncertainty related to noisy proxies will show up as bias and variance in the reconstruction and that the partition between these forms of errors depends on the reconstruction method. Pseudo‐proxy experiments are conducted to further discuss the influence of noise and the requirements regarding the geographical location and number of proxies necessary for reliably reconstructing the low‐frequency variability. We conclude with recommendations for future large‐scale temperature reconstructions.

Concentrating on Q and Cash Flow

Investment spending by US public firms is highly concentrated. The 100 largest spenders account for 60% of total capital expenditures and drive most of the variation in aggregate US investment. This high concentration creates a disconnect between the average public firm and macroeconomic aggregates. For large firms, cash flow remains the primary driver of investment spending and has not declined in importance as it has for smaller public firms. The cash flowing to big spenders provides a better forecast of future investment opportunities than noisy proxies for Tobin’s q even though these firms are not financially constrained. Our results are inconsistent with recent models that predict higher investment-cash flow sensitivity for small young growth firms and suggest that cash flow is still the most important determinant of macroeconomic fluctuations in investment spending.

The Temporal Stability of Children's Neighborhood Experiences: A Follow-Up from Birth to Age 15

Despite increasing attention being paid to the temporal dynamics of childhood disadvantage, children's neighborhood characteristics are often measured at a single point in time. Whether such cross-sectional measures serve as reliable proxies for children's long-run neighborhood conditions depends on the stability in children's neighborhood experiences over time. We investigate the temporal stability in children's neighborhood environment, focusing on two of the most commonly studied neighborhood characteristics: The ethnic composition and mean income of the neighborhood. Using Dutch population register data, we follow an entire cohort of children from birth up until age 15. We use year-to-year correlations in the percentage non-Western minorities and the mean income in the neighborhood to evaluate the temporal stability of children's neighborhood experiences. Children's neighborhood characteristics were found to be more stable over time with regard to ethnic composition than with regard to income. Children who had moved at least once were found to have lower stability in neighborhood characteristics than children who never moved. Finally, neighborhood experiences were found to be more stable over time for ethnic minorities than for the native Dutch, although differences were small with regard to income. Single point-in-time measurements of neighborhood characteristics are reasonable proxies for the long-run ethnic composition of children's neighborhood environment, but rather noisy proxies for the long-run income status of their neighborhood, particularly for those who moved.

Discretionary Accruals: Earnings Management... Or Not?

This paper discusses some limitations of discretionary accruals measures. While discretionary accruals are acknowledged to be noisy proxies for earnings management, they are still widely used in the literature. This paper attempts to explain from basic econometrics how discretionary accruals are estimated, and in doing so why they are inappropriate measures for earnings management. It is shown that decisions of peer firms will influence the regression coefficients, and hence residuals, in accruals models which may lead to false conclusions about earnings management in other firms. This point is emphasized using an artificially constructed firm with no changes in their fundamental performance, and hence no discretion in their accruals. I also note concerns about the inferences which is commonly not acknowledged in research. Finally, using AAERs and Enron as examples, I demonstrate how discretionary accruals do not capture what literature often claims.

Skill Demands and Mismatch in U.S. Manufacturing

Recent economic events have sparked debates over the degree of structural mismatch in the U.S. economy. One of the most frequent claims is that workers lack the skills that employers demand. The existing literature, however, analyzes this potential mismatch at a high level of aggregation with abstract indices and noisy proxies that obscure the underlying mechanisms. The authors address these issues by presenting and analyzing results from a survey of U.S. manufacturing establishments. The survey is the first, to their knowledge, to directly measure concrete employer skill demands and hiring experiences in a nationally representative survey at the industry level. The findings indicate that demand for higher-level skills is generally modest, and that three-quarters of manufacturing establishments do not show signs of hiring difficulties. Among the remainder, demands for higher-level math and reading skills are significant predictors of long-term vacancies, but demands for computer skills and other critical-thinking/problem-solving skills are not. Of particular interest, high-tech plants do not experience greater levels of hiring challenges. When the authors examine the potential mechanisms that could contribute to hiring difficulties, they find that neither external regional supply conditions nor internal firm practices are predictive of hiring problems. Rather, the data show that establishments that are members of clusters or that demand highly specialized skills have the greatest probability of incurring long-term vacancies. The authors interpret these results as a sign that it is important to think about factors that complicate the interaction of supply and demand—such as disaggregation and communication/coordination failures—rather than simply focusing on inadequate labor supply.

Expected Returns and Dividend Growth Rates Implied by Derivative Markets

The dividend-price ratio is a noisy proxy for expected returns when expected dividend growth is time-varying. This paper uses a new and forward-looking measure of dividend growth extracted from S&P 500 futures and options to correct the dividend-price ratio for changes in expected dividend growth. Over January 1994 through June 2011, dividend growth implied by derivative markets reliably forecasts future dividend growth, and the corrected dividend-price ratio predicts S&P 500 returns substantially better than the standard dividend-price ratio, in-sample and out-of-sample. Time-varying expected dividend growth is important to explain price movements, especially because it is highly correlated with expected returns.

ESTIMATING THE PERSISTENCE AND THE AUTOCORRELATION FUNCTION OF A TIME SERIES THAT IS MEASURED WITH ERROR

An economic time series can often be viewed as a noisy proxy for an underlying economic variable. Measurement errors will influence the dynamic properties of the observed process and may conceal the persistence of the underlying time series. In this paper we develop instrumental variable (IV) methods for extracting information about the latent process. Our framework can be used to estimate the autocorrelation function of the latent volatility process and a key persistence parameter. Our analysis is motivated by the recent literature on realized volatility measures that are imperfect estimates of actual volatility. In an empirical analysis using realized measures for the Dow Jones industrial average stocks, we find the underlying volatility to be near unit root in all cases. Although standard unit root tests are asymptotically justified, we find them to be misleading in our application despite the large sample. Unit root tests that are based on the IV estimator have better finite sample properties in this context.

What Does Investor Sentiment Reflect: Animal Spirits or Risks?

The role of investor sentiment in the stock market has attracted attentions of many economists. Previous papers show that investor sentiment has return predictability and it is more pronounced among stocks that are more difficult to value and/or to arbitrage, and emphasize the behavioral role of investor sentiment. However, it still remains unclear whether this predictability is actually due to a causal effect of autonomous animal spirits or not. Alternatively, investor sentiment may simply reflect systematic risks, which would affect stock returns. In this alternative case, the predictability would be mere coincidence, not causation. In this paper, I try to understand the meaning of innovations in investor sentiment. I use the investor sentiment index constructed by Baker and Wurgler (2006). I set up a structural model in which sentiment innovations arise from both animal spirit shocks and several risk shocks, and animal spirit shocks could affect stock returns. By matching impulse response functions from data simulated by the theoretical model to those from the actual US data, I estimate parameters in the model. The estimated model moderately replicates the historical data in the actual stock market. The estimation results show that a substantial amount of variation in investor sentiment is explained by systematic risk shocks as well as by animal spirit shocks, and that animal spirit shocks can have significant effects on stock returns. The findings suggest that investor sentiment is a noisy proxy of animal spirits and autonomous animal spirits are at least in part responsible for the apparent return predictability of investor sentiment.

Statistical framework for evaluation of climate model simulations by use of climate proxy data from the last millennium – Part 1: Theory

Abstract. A statistical framework for comparing the output of ensemble simulations from global climate models with networks of climate proxy and instrumental records has been developed, focusing on near-surface temperatures for the last millennium. This framework includes the formulation of a joint statistical model for proxy data, instrumental data and simulation data, which is used to optimize a quadratic distance measure for ranking climate model simulations. An essential underlying assumption is that the simulations and the proxy/instrumental series have a shared component of variability that is due to temporal changes in external forcing, such as volcanic aerosol load, solar irradiance or greenhouse gas concentrations. Two statistical tests have been formulated. Firstly, a preliminary test establishes whether a significant temporal correlation exists between instrumental/proxy and simulation data. Secondly, the distance measure is expressed in the form of a test statistic of whether a forced simulation is closer to the instrumental/proxy series than unforced simulations. The proposed framework allows any number of proxy locations to be used jointly, with different seasons, record lengths and statistical precision. The goal is to objectively rank several competing climate model simulations (e.g. with alternative model parameterizations or alternative forcing histories) by means of their goodness of fit to the unobservable true past climate variations, as estimated from noisy proxy data and instrumental observations.

Combining Spatial Information Sources While Accounting for Systematic Errors in Proxies

SummaryEnvironmental research increasingly uses high dimensional remote sensing and numerical model output to help to fill space–time gaps between traditional observations. Such output is often a noisy proxy for the process of interest. Thus we need to separate and assess the signal and noise (often called discrepancy) in the proxy given complicated spatiotemporal dependences. Here I extend a popular two-likelihood hierarchical model by using a more flexible representation for the discrepancy. I employ the little-used Markov random-field approximation to a thin plate spline, which can capture small-scale discrepancy in a computationally efficient manner while better modelling smooth processes than standard conditional auto-regressive models. The increased flexibility reduces identifiability, but the lack of identifiability is inherent in the scientific context. I model particulate matter air pollution by using satellite aerosol and atmospheric model output proxies. The estimated discrepancies occur at a variety of spatial scales, with small-scale discrepancy particularly important. The examples indicate little predictive improvement over modelling the observations alone. Similarly, in simulations with an informative proxy, the presence of discrepancy and resulting identifiability issues prevent improvement in prediction. The results highlight but do not resolve the critical question of how best to use proxy information while minimizing the potential for proxy-induced error.

Identification of linear relationships from noisy data using errors-in-variables models—relevance for reconstruction of past climate from tree-ring and other proxy information

Reliable paleoclimate reconstructions are needed to assess if the recent climatic changes are unusual compared to pre-industrial climate variability. Here, we focus on one important problem in climate reconstructions: Transfer functions relating proxies (predictors) and target climatic quantities (predictands) can be seriously biased when predictand and predictor noise is not adequately accounted for, resulting in biased amplitudes of reconstructed climatic time series. We argue for errors-in-variables models (EVM) for unbiased identification of linear structural relationships between noisy proxies and target climatic quantities by (1) introducing underlying statistical concepts and (2) demonstrating the potential biases of using the EVM approach, the most commonly used direct ordinary least squares (OLS) regression, inverse OLS regression, or the reduced major axis method (‘variance matching’) with a simulation example of artificial noise-disturbed sinusoidal time series. We then develop an alternative strategy for paleoclimate reconstruction from tree-ring and other proxy data, explicitly accounting for the identified problem.

Technical Note: Correcting for signal attenuation from noisy proxy data in climate reconstructions

Abstract. Regression-based climate reconstructions scale one or more noisy proxy records against a (generally) short instrumental data series. Based on that relationship, the indirect information is then used to estimate that particular measure of climate back in time. A well-calibrated proxy record(s), if stationary in its relationship to the target, should faithfully preserve the mean amplitude of the climatic variable. However, it is well established in the statistical literature that traditional regression parameter estimation can lead to substantial amplitude attenuation if the predictors carry significant amounts of noise. This issue is known as "Measurement Error" (Fuller, 1987; Carroll et al., 2006). Climate proxies derived from tree-rings, ice cores, lake sediments, etc., are inherently noisy and thus all regression-based reconstructions could suffer from this problem. Some recent applications attempt to ward off amplitude attenuation, but implementations are often complex (Lee et al., 2008) or require additional information, e.g. from climate models (Hegerl et al., 2006, 2007). Here we explain the cause of the problem and propose an easy, generally applicable, data-driven strategy to effectively correct for attenuation (Fuller, 1987; Carroll et al., 2006), even at annual resolution. The impact is illustrated in the context of a Northern Hemisphere mean temperature reconstruction. An inescapable trade-off for achieving an unbiased reconstruction is an increase in variance, but for many climate applications the change in mean is a core interest.

When Does Idiosyncratic Risk Really Matter?

In contrast to the finding of the current literature, we provide new evidence supporting a positive relation between idiosyncratic risk and the expected future market return. Since a large part of the idiosyncratic risk can be diversified away easily, the conventional aggregate idiosyncratic risk measures can only be noisy proxies for the undiversified idiosyncratic risk, which may be priced according to Merton (1987). We thus propose a simple noise reduction method that includes two different noisy measures (dual-predictor) in the same predictive regression to reexamine the relation. We show that the noise effect tends to cancel out in our framework due to the highly correlated noise components of the two measures. In particular, our dual-predictor alone explains future market returns with an adjusted R 2 of 4%. Such a large predictive power is economically significant and robust to the inclusion of other popular predictors and to the choices of sample periods and additional market indices.

Estimating the Persistence and the Autocorrelation Function of a Time Series that is Measured with Error

An economic time series can often be viewed as a noisy proxy for an underlying economic variable. Measurement errors will influence the dynamic properties of the observed process and may conceal the persistence of the underlying time series. In this paper we develop instrumental variable (IV) methods for extracting information about the latent process.Our framework can be used to estimate the autocorrelation function of the latent volatility process and a key persistence parameter. Our analysis is motivated by the recent literature on realized (volatility) measures, such as the realized variance, that are imperfect estimates of actual volatility. In an empirical analysis using realized measures for the DJIA stocks we find the underlying volatility to be near unit root in all cases. Although standard unit root tests are asymptotically justified, we find them to be misleading in our application despite the large sample. Unit root tests based on the IV estimator have better finite sample properties in this context.

Expected Returns and Dividend Growth Rates Implied in Derivative Markets

The dividend-price ratio is a noisy proxy for expected returns when expected dividend growth is time-varying. This paper uses a new and forward-looking measure of dividend growth extracted from S&P 500 futures and options to correct the dividend-price ratio for changes in expected dividend growth. Over January 1994 through June 2011, dividend growth implied by derivative markets reliably forecast future dividend growth, and the corrected dividend-price ratio predicts S&P 500 returns substantially better than the standard dividend-price ratio, in-sample and out-of-sample. Time-varying expected dividend growth is important to explain price movements, especially because it is highly correlated with expected returns.

Noise, Capm and the Size and Value Effects

We model a continuous time one factor economy where stock prices are noisy proxies of the informationally efficient values. The pricing error process is modeled as a mean-reverting process; this gives us a well defined notion of over-pricing and under-pricing in the market, and stocks fluctuates between being over-priced and under-priced. We show that in this economy, cap-weighting is a sub-optimal portfolio strategy. A non-cap-weighted portfolio outperforms a cap-weighted portfolio with the same factor exposure. This is because, in a cap-weighting scheme, portfolio weights are driven by market prices; as such, more weights are allocated to overvalued stocks and less weight to undervalued stocks. Following the same intuition, we show that in the cross-section, large cap stocks tend to underperform small cap stocks and high price-to-book stocks (growth stocks) tend to underperform low price-to-book stocks (value stocks). In this economy, if we proxy the hidden risk factor with the market clearing portfolio and regress stock returns against market returns (a CAPM regression), a non-cap-weighted portfolio exhibits significant alpha! However, when we regress the same non-cap-weighted portfolio against market, size and value, the Fama-French style regression exhibits factor loadings on value and size with insignificant alpha. Clearly, value and size are not risk factors in our economy! Value and size shows up in our regressions because the market portfolio is a poorly constructed proxy for the hidden risk factor. The paper contributes to the anomalies literature by showing that mean-reversion in stock returns and the Fama-French size and value effects are driven by the same market defect - pricing noise! This suggests that models, such as disposition effects and information herding, which can generate stock price over-reaction and therefore mean-reversion in stock prices, can also explain the value and size puzzle.

Spin-offs, Divestitures, and Conglomerate Investment

We examine whether spin-offs or divestitures cause improvements in conglomerate investment efficiency. At issue are endogeneity of these restructuring decisions and correct measurement of investment efficiency. Endogeneity is a problem because the factors that induce firms to spin off or divest divisions may also improve investment efficiency; measurement error is a problem because efficiency measures employ Tobin’s q as a noisy proxy for investment opportunities. We find important differences between firms that divest or spin off and a control sample. After accounting for these differences and for measurement error in q, we find no evidence of improvements in investment efficiency. (JEL G31, G34)