شرایط بازار سهام و سیاست پولی در یک مدل DSGE برای ایالات متحده

This paper investigates the interactions between stock market fluctuations and monetary policy within a DSGE model for the U.S. economy. First, we design a framework in which fluctuations in households financial wealth are allowed—but not necessarily required—to exert an impact on current consumption. This is due to the interaction, in the financial markets, of long-time traders holding wealth accumulated over time with newcomers holding no wealth at all. Importantly, we introduce nominal wage stickiness to induce pro-cyclicality in real dividends. Additional nominal and real frictions are modeled to capture the pervasive macroeconomic persistence of the observables employed to estimate our model. We fit our model to post-WWII U.S. data, and report three main results. First, the data strongly support a significant role of stock prices in affecting real activity and the business cycle. Second, our estimates also identify a significant and counteractive response of the Fed to stock-price fluctuations. Third, we derive from our model a microfounded measure of financial slack, the “stock-price gap”, which we then contrast to alternative ones, currently used in empirical studies, to assess the properties of the latter to capture the dynamic and cyclical implications of our DSGE model. The behavior of our “stock-price gap” is consistent with the episodes of stock-market booms and busts occurred in the post-WWII, as reported by independent analyses, and closely correlates with the current financial meltdown. Typically employed proxies of financial slack such as detrended log-indexes or growth rates show limited capabilities of capturing the implications of our model-consistent index of financial stress. Cyclical properties of the model as well as counterfactuals regarding shocks to our measure of financial slackness and monetary policy shocks are also proposed.

“Financial and economic conditions can change quickly. Consequently, the Committee must remain exceptionally alert and flexible, prepared to act in a decisive and timely manner and, in particular, to counter any adverse dynamics that might threaten economic or financial stability”. Chairman Ben S. Bernanke, Financial Markets, the Economic Outlook, and Monetary Policy, speech held at the Women in Housing and Finance and Exchequer Club Joint Luncheon, Washington, DC, January 10, 2008 (Bernanke, 2008). Policymakers closely monitor financial market's behavior. This is due to the strict interconnections between financial and real sectors in the economy. Swings in asset prices affect real activity through several channels (households wealth, firms’ market value of collateral, Tobin's Q), and, consequently, inflation and the term structure. On the other hand, stock market fluctuations are driven by expectations on future returns, which are tightly linked to expectations on the predicted evolution of the business cycle, inflation, and monetary policy decisions.1 Of course, policy-makers need to gauge financial markets’ conditions and identify their drivers to appropriately implement monetary policy actions.2 While the supply-side interplay between stock prices and the real economy has been given some attention in the analysis of large scale, quantitative models with financial frictions, considerably less (if not zero) attention has been paid in analyzing the role of the demand-side interplay, working through wealth effects on households’ consumption, in the standard small scale Dynamic New Keynesian (DNK) model. On the other hand, such workhorse model, despite its parsimony, has been shown to have meaningful implications for the pricing of equity markets and the response of the stock market to real and monetary shocks.3 The standard new-Keynesian model of the business cycle, however, as much widely adopted in central banks as well as academic circles to perform monetary policy analysis, typically considers stock prices as redundant for the computation of the equilibrium values of inflation, output, and the policy rate.4 This is so because financial wealth fluctuations are fully smoothed out by infinitely lived agents, both at the individual and aggregate levels. This feature of the standard new-Keynesian framework effectively shuts down the demand-side channel of transmission of financial shocks and makes it ill-suited to investigate the role of stock prices in the macroeconomic environment. This paper proposes a small-scale new-Keynesian model in which stock prices are allowed to play an active role in determining the dynamics of the business cycle, through the demand side. Building on previous contributions by Nisticò (2005) and Airaudo et al. (2007), we consider a framework in which households face a constant probability of exiting the financial markets in each period and interact with a fraction of agents who enter the financial markets holding no wealth at all.5 Consequently, aggregate consumption cannot be perfectly smoothed out in reaction to swings in financial wealth, and stock-price fluctuations thereby affect aggregate demand. In order to take it to the data, we add several features to the setup in Nisticò (2005). First, we assume nominal-wage stickiness. Carlstrom and Fuerst (2007) show that this assumption makes real dividends pro-cyclical. Indeed, following a monetary policy tightening that induces a fall in firms’ labor demand, if wages were fully flexible, firms’ marginal costs would fall as well, and firms’ dividends would counter-cyclically increase. By contrast, the presence of nominal wage stickiness makes revenues fall more than marginal costs, thus delivering pro-cyclical real dividends. Second, we add price and wage indexation to past inflation and productivity growth, and external habits in consumption. These additional features enable our framework to capture the endogenous persistence in the U.S. macroeconomic data. Finally, we allow for a stochastic trend in total factor productivity, which allows us to estimate our model without pre-filtering our observables. An appealing feature of our theoretical framework is that it implies a microfounded, endogenous measure of financial slack at business cycle frequencies, that we label “stock-price gap”. In analogy with the output gap, we define the “stock-price gap” as the percentage deviation of the real stock-price index from its frictionless level—consistent with an equilibrium with no dynamic distortions—and is therefore the relevant benchmark for monetary-policy makers. Such measure of financial conditions endogenously interacts with the output gap via the IS curve and the pricing equation, and may enter the Taylor rule that describes the systematic behavior of the U.S. monetary policy authority. The microfoundation of the model enables us to identify the effect that macroeconomic shocks exert on our measure of financial stress. We fit our new-Keynesian model to U.S. data over the post WWII sample with Bayesian techniques and perform several exercises. Our main results can be summarized as follows. First. The data give strong support to our new-Keynesian model with stock prices. In particular, our estimates suggest that a significant ratio of traders in the financial markets are periodically replaced by newcomers holding zero financial assets. This makes the economy significantly non-Ricardian, and implies a finite average planning horizon for households’ financial investments. Second. The evidence shows a significant systematic response of the Fed to stock-price dynamics. Specifically, the estimated interest-rate rule displays an additional component, responding to non-zero stock-price gaps. Third. Our estimated stock-price gap is consistent with the phases of booms and busts occurred in the sample, as dated by Bordo et al. (2008).6 Moreover, our estimated stock-price gap allows us to evaluate the ability of alternative proxies, currently used in the empirical literature, to capture the dynamic and cyclical implications of a prototypical DSGE New-Keynesian model. In this respect, we show that these alternative measures can be very poor representations of such implications. Additionally, we perform several counterfactual exercises. As to the dynamic response of the economy, we estimate a 25 basis points unexpected rise in the federal funds rate to cause an on-impact negative reaction of the stock-price gap of about 20 basis points. By contrast, an unexpected 1% boom in the stock-price gap induces an on-impact interest rate hike of 12 basis points, which about doubles within a year. Two very recent papers are closely related to ours: Milani (2008) and Challe and Giannitsarou (2008). Milani (2008) estimates a purely forward looking version of Nisticò (2005), in which households make inference on the future evolution of the business cycle on the basis of the observed oscillations in the stock market. He finds that the direct effect of the stock market on the business cycle is negligible, while the expectational effect is important. By contrast, we find a significant direct effect of financial wealth's swings on the real GDP. Differences between our results and Milani (2008) may be attributed to the model structure—we model several nominal and real frictions, the most important one probably being nominal wage stickiness—and, especially, the treatment of the data. Indeed, while Milani (2008) uses HP-filtered series of output and the stock-price index as proxies for the respective gaps, we relate the observable growth rates of the relevant time series to the latent state variables of our model. Therefore, we let the internal propagation mechanism of our model construct the gaps in a theoretically consistent fashion, without resorting to any pre-estimation filtering. Challe and Giannitsarou (2008) study the asset-pricing implications of the standard New Keynesian model, in which equilibrium stock prices are consistent with the households’ optimization problem but do not have any real effect on consumption. They aim to show that a calibrated DSGE model is able to replicate the reaction of stock prices to a monetary policy shock as estimated by some VAR analysis. With respect to Challe and Giannitsarou (2008), we allow for a two-way interaction between the real and the financial part of the system, and we estimate our framework with U.S. data instead of resorting to calibration. Finally, by scrutinizing the demand-channel of transmission of financial fluctuations, our approach complements a related strand of literature (e.g. Christiano et al., 2003 and Christiano et al., 2007; Queijo von Heideken, 2009), which instead focus on the role of the banking sector and financial frictions in affecting the supply-side of an economy by working with extensions of the Bernanke et al. (1999) financial-accelerator model. The paper is structured as follows. Section 2 presents our microfounded new-Keynesian model of the business cycle in which stock prices are allowed, but not required, to affect the equilibrium values of output, inflation, and the policy rate. Section 3 discusses our estimation strategy. Section 4 presents and comments our results. Section 5 proposes further discussion, and Section 6 concludes.

While writing this paper, financial markets are experiencing extraordinary events. The quest for a modern modeling of the interaction between the financial and the real side of the economy is more compelling than ever. We make a first effort along this line by constructing a new-Keynesian model of the business cycle that allows for financial wealth effects to play an active role for the dynamics of output, inflation, and interest rates. This is due to the turnover and interaction between agents holding positive financial wealth and newcomers not having cumulated such wealth yet. When fitting our new-Keynesian model to US data over the post WWII sample, we find remarkable support for the role played by financial market frictions in this economy. We estimate the average rate of replacement of old traders with newcomers to range between 7% and 20%, which implies an effective average planning horizon for US households’ financial investments between 5 and 15 quarters. Moreover, we detect a significant, counteractive and systematic response of the Fed to stock price fluctuations as captured by non-zero stock-price gaps, possibly instrumental to the stabilization of inflation and output. Our model-consistent measure of financial slack, labeled “stock price gap”, captures remarkably well the phases of booms and busts occurred in the post-WWII period. Commonly employed empirical proxies of the financial slack such as growth rates or statistically de-trended stock price indices correlate to our microfounded financial slack measure just mildly. Therefore, they do not seem to extensively capture the dynamic and cyclical implications that the Dynamic New Keynesian model has suggest the stock market. In terms of counterfactual dynamic responses, we estimate a 25 basis points unexpected rise in the federal funds rate to cause an on impact negative and significant reaction of the stock-price gap of about 0.2%. By contrast, an unexpected 1% boom in the stock-price gap induces an interest rate hike of about 12 basis points, on impact, which about doubles within a year and it is remains significant for some quarters. We believe our framework can represent a first modeling step towards the construction of a more complete model of the business cycle able to deal with financial market frictions. We view the introduction other important features, like supply-side effects via firms’ balance sheet, durable goods capturing housing services, and a non-trivial role for financial intermediaries and their interaction with households and the monetary policy authority, as an interesting avenue for future research in the field, and as part of our agenda