From The Developing Economist VOL. 1 NO. 1
The Relationship Between Monetary Policy and Asset Prices: A New Approach Analyzing U.S. M&A Activity
IN THIS ARTICLE
This article details the relationship between asset prices and monetary policy with a specific focus on the mergers and acquisitions market. The existing literature has studied extensively the link between monetary policy and stock prices and housing prices, but has not analyzed other assets, such as MA transactions. Monetary policy theory suggest that a negative shock to monetary policy that lowers interest rates increases asset prices. A lower interest rate decreases the cost of borrowing, raises investment levels (say for firms or home-buyers), and thus raises the asset price. Using a VAR methodology, the empirical evidence in this study, however, does not find this relationship between monetary policy shocks and MA activity. The response of MA activity – measured by average EBITDA multiple and the number of transactions – does not respond inversely to shocks in monetary policy.
Black Tuesday, the infamous Wall Street Crash of 1929, triggered the Great Depression, the most severe global recession since before the Industrial Revolution. The Great Depression began with this devastating drop in the asset prices of companies. Unfortunately, the Federal Reserve made critical errors in judgment and in philosophy that severely worsened the Great Depression for years. Since then, scholars have been better able to understand monetary policy, including its relationship with asset prices. In these efforts, scholars and monetary policymakers have hoped to avoid the consequences that can result from asset price crashes and even possibly prevent such crashes in the first place. On Milton Friedman's 90th birthday, former Federal Reserve Chairman Ben Bernanke famously commemorated Friedman's scholarship in this field. Bernanke (2002) concluded his remarks to Friedman and the birthday party attendees stating, "Regarding the Great Depression, you're right, we [the Federal Reserve] did it. We're very sorry. But thanks to you, we won't do it again."
Just a few years later while testifying in front of Congress during his nomination to become Chairman, Bernanke falsely observed that no housing bubble existed to burst, noting that asset price increases in the housing market "largely reflect strong economic fundamentals" (Henderson 2005). Former Chairman Alan Greenspan suggested the housing price increases were merely "froth" in local markets (Henderson 2005). To the dismay of both chairmen, there was indeed a housing bubble and it collapsed. Combined with excessive risk taken by banks and financial institutions in the subprime lending market, the Great Recession resulted. Unlike the Great Depression, this time the Federal Reserve, under Bernanke's guidance, took enormous steps to provide liquidity, be a lender of last resort, and constantly strive to stabilize financial conditions. Although not perfect, most scholars would agree that the Fed's efforts were commendable and often ingenious during the Great Recession.
I provide this brief history of the two worst economic downturns in U.S. history to exemplify the important relationship between monetary policy and asset prices. In the Great Depression, falling stock prices were the trigger; in the Great Recession, housing prices took this role. It may not be practical to expect the Fed to prevent such collapses, but in the very least, an optimal response is required to minimize the potentially disastrous outcomes. However, the academic literature has so far only studied stock prices and housing prices. I argue that, just as housing prices were far off the radar of policy makers and scholars before the 2006-2007 collapse, other assets may be equally troubling in future downturns. The core aim of this paper is to extend the literature beyond stock prices and housing prices and consider the relationship between monetary policy and a third asset class – mergers and acquisitions ("M&A"). M&A activity is an enormous market, totaling more than 14,000 transactions in 2012 alone with an average transaction size over $200 million. This is not meant to be a prediction for the next recession, although the possibility certainly exists. In the very least, understanding the relationship between monetary policy and asset prices more broadly is a critical task that can benefit scholars and policymakers.
With this aim in mind, the article will progress as follows. In Section I, I present an extensive literature review covering monetary policy as it relates to asset prices. Section II presents the Asset Price Channel, a hypothesis based on existing literature that conceptually explains the potential relationship between asset prices and monetary policy. Next, Section III applies this hypothesis toM&Aactivity specifically and then presents a discussion on M&A valuation and why this asset type is relevant to include in the monetary policy literature. Section IV and V describe the data along with the methodology for forming a model to test the Asset Price Channel. The data includes basic Taylor rule variables, the Federal Funds Rate ("FFR"), the ten-year Treasury rate, and M&A metrics including the number of transactions and the average EBITDA multiple. Section VI presents the results from the described models, showing no evidence to support the Asset Price Channel, contradicting the existing literature that studies stock prices and the housing market.
I. Literature Review Summaries
In this literature review, I discuss several topics concerning asset prices and monetary policy. First, scholars are divided on whether optimal policy rules should include asset prices. Related to this, empirical studies have examined both whether asset prices respond to monetary policy and whether monetary policy responds to asset prices. I also briefly review articles that link foreign asset prices with domestic monetary policy. Throughout this literature review, I emphasize that economists have only studies asset prices and monetary policy with housing prices and stock prices. Economists have not linked monetary policy to other asset classes, including M&A activity which is the focus of this article.2
Cecchetti et al. (2000) outlines a scenario in which asset price misalignments create undesirable instability in inflation and employment. In other words, booms cause busts, and busts are harmful to the macroeconomy. Considering historical cases of asset booms, the authors then consider what steps central banks can take to avoid these pitfalls.3 The authors advocate a "lean against the wind" strategy where central banks respond to booms by increasing the interest rate in order to counter rising asset prices and dampen boom-bust cycles.4 This strategy includes asset prices in the policy rule to best stabilize inflation and output.
By examining forward-looking structural models of G7 economies from 1972 to 1998, Goodhart and Hofmann (2000) similarly contend that a monetary policy rule excluding asset price movements increases inflation and output gap variability because the information contained in asset prices is useful in forecasting future demand conditions. Bordo and Jeanne (2002) consider a stylized boom-bust dynamic model in stock and property prices. The thought experiment discusses the role of pre-emptive monetary policy. This sort of ex ante policy differs from policy rules that respond to an asset price bust only ex post, like an inflation-targeting rule. By compare moving averages of asset prices in OECD countries from 1970 to 2001, the analysis identified twentyfour stock booms and twenty housing booms.5 The authors contend that a response to asset prices restricts monetary policy during a boom and is insurance against the risk of real disruption induced by the potential for a bust or even a moderate asset price reversal. In this way, they favor a policy rule that includes asset prices in order to yield tighter monetary policy ex ante before a boom develops.
Several scholars, however, hold the view that policy rules including asset prices yield sub-optimal results. Bernanke and Gertler (2001) evaluate a standard new-Keynesian model while also incorporating informational friction in credit markets. The model then simulates a shock of a five-period increase in the nonfundamental component of stock prices followed by a bust in the sixth period. The results show that an aggressive inflation-targeting rule dominates accommodative approaches in reducing both inflation and output variability. Placing a weight on stock prices does help marginally, but Bernanke and Gertler conclude this is not the optimal policy because of the practical difficulties in separating fundamental from non-fundamental movements in stock prices.6 Ultimately, the practical difficulties outweigh the marginal gains in policy outcomes. Carlstrom and Fuerst (2007) consider the inclusion of asset prices in monetary policy in a model with either sticky prices or sticky wages. A central bank response to share prices in the case of sticky wages does yield optimal policy because firm profits and share prices move positively with inflation. However, in a model with sticky prices, a central bank responding to share prices implicitly weakens its overall response to inflation because increases in inflation tend to lower firm profits, leading to suboptimal monetary policy. The authors conclude that, because of the sticky price model, monetary policy rules should not include asset prices.
Gilchrist and Leahy (2002) assess large movements in asset prices in the United States and Japan from the 1970s through the 1990s. Using this data, they consider various shocks to the economy, including asset price busts. They conclude that weak inflation targets produce huge swings in output. Regardless of including asset prices in the policy rule, this empirical study concludes that aggressive inflationtargeting yields the optimal outcome. Filardo (2000) employs a framework outlined by former Bank of England member Charles Goodhart that proposed policy rules that include broad measures of housing and stock prices. He dismisses this approach, primarily because of the difficulty in identifying the signs of nonfundamental movements in asset prices. Filardo illustrates that erroneous identification of price bubbles has significant unintended consequences that harm economic outcomes.7 Even without this difficulty, he concludes that including asset prices has little impact in improving policy outcomes.
The corollary question asks whether asset prices respond to monetary policy. Bernanke and Kuttner (2005) conducted an event-study analysis by looking at daily data from FOMC decisions from 1989 to 2002 and tracking the movement in stock prices in response to monetary policy shocks. Using several modeling techniques, such as VAR forecasts, Bernanke and Kuttner conclude that an unexpected 25-basispoint cut to the Federal Funds Rate leads to a 1% increase in stock indexes on that same day. Rigobon and Sack (2004) use a VAR model that employs an identification technique through heteroskedasticity. Examining the Dow Jones Industrial Average, SP 500, the Nasdaq, and the Wilshire 5000 from 1994 to 2001, these authors find very similar results to the Bernanke and Kuttner analysis. For example, an unanticipated 25-basis point increase in the short-term interest rate results in a 1.7% decline in the S&P 500.
Laevan and Tong (2012) take a deeper look at this question by examining varying responses by different types of firms. There should be variance among firms – those more dependent on external financing should have larger swings in stock prices due to a monetary policy shock. The data examines 20,121 firms across forty-four countries, with the average response of stock prices roughly 4:1 from an unexpected change in interest rates.8 Firms are then classified as either dependent or (relatively) independent on external financing, interacting this variable with the monetary policy shock. Indeed, firms more dependent on external financing are disproportionately affected.
Prior to the housing price collapse beginning in 2006 that triggered the Great Recession, economists did not consider the damage that could be caused or triggered by a housing bubble. Several scholars and commentators have criticized that then-Chairman Alan Greenspan kept interest rates too low for too long leading up to the collapse of the bubble, allowing for easy lending and an increased demand for housing. According to this reason, the low interest rates fueled the bubble and allowed the housing market to overheat before eventually collapsing. From 2002 to 2006, the Federal Funds Rate was roughly 200 basis points below what the Taylor rule would have prescribed for policy makers. However, Bernanke (2010) has since argued that this thinking is flawed for several reasons. First, he states that the applicable Taylor rule looks at expected future inflation, not current inflation. The interest rates were on par with this revised monetary policy rule and were not too low. Bernanke also observes that the surge in housing prices began in 1998, implying that the timing of the start of the housing bubble rules out the period when interest rates were arguably too low (first in 2002 through 2006). Iacoviello (2005) similarly estimated a monetary business cycle that includes the housing market. By imposing collateral and borrowing constraints and simulating demand shocks on the housing market of nominal loans, he finds that "allowing the monetary authority to respond to asset prices yields negligible gains in terms of output and inflation stabilization." Other scholars disagree and believe the Fed should have acted otherwise. Taylor (2007) observes that monetary policy responded more effectively to inflation in the 1980s and 1990s and reduced boom-bust cycles in the housing market. He then claims that the Federal Reserve deviated from this previous action beginning in 2002. Using a counterfactual model of the housing market, he contends that the loose monetary policy failed to minimize the housing bubble and may have been a causal force in the rise of the housing bubble.
Just as they answered whether asset prices respond to monetary policy, Rigobon and Sack (2003) also study the reverse – the reaction of monetary policy to stock markets. According to the authors, stock markets have a significant impact on the macroeconomy primarily through the influence on aggregate consumption and the cost of financing to businesses. These effects play into the calculus of central bankers. Using the same VAR model from before, Rigobon and Sack establish an identification technique based on the heteroskedasticity of stock market returns. They conclude that a five percent rise in the S&P 500 increases the likelihood of a 25 basis point tightening by about one half. Bohl et al. (2007) study this same question by looking at the Bundesbank, tracking stock prices and interest rates in Germany from 1985 to 1998. Contrary to the evidence that Rigobon and Sack found in the U.S., the results in this study show that the Bundesbank did not respond to movements in stock prices, with one possible exception to the stock market crash of 1987. Bohl et al. states that "the theoretical rationale linking central bank reactions to asset prices is not yet sufficiently well developed to provide definite guidance."
Erler et al. (2013) analyze the real estate boom leading up to the Great Recession to determine if monetary policy responds to real estate asset prices. They set up a GMM model using real estate market data from 1980 to 2007 and then approximate both a Taylor rule and a Taylor-type rule with asset prices as possible monetary policy responses. The authors found a statistically significant negative response to real estate asset prices including a real estate dummy variable. In other words, the Fed actually lowered interest rates in the presence of a real estate boom, contrary to a "lean against the wind" strategy.
A related topic that several scholars have addressed is the relationship between domestic monetary policy and foreign asset prices, both if foreign asset prices respond to domestic policy and vice versa. Ida (2011) examines a theoretical New Keynesian model to determine optimal monetary policy rules in an open economy. For simplicity, the model illustrates a two-country sticky price world. In this scenario, a positive foreign productivity shock leads to an increase in foreign asset prices. Assuming an open economy, this leads to increases in both foreign and domestic consumption. Ida argues that this increased consumption raises domestic asset prices despite no change to the fundamental values of domestic producers, creating a price bubble. This creates an opportunity for central bankers to consider this type of bubble when setting interest rates. Wongswan (2008) addresses this question using empirical evidence from fifteen foreign equity indexes in Asia, Europe, and Latin America with respect to movements in U.S. monetary policy. By observing high-frequency intra-day data on dates of FOMC announcements, he employs a model similar to that of Bernanke and Kuttner. The stock indexes increase between 0.5% and 2.5% with a 25-basis-point cut in the federal funds target rate. This reinforces the inverse relationship between asset prices movements and monetary policy shocks
II. Theoretical Outline of Monetary Policy Effects on Asset Prices
The Fed sets the money supply to a level that achieves a certain interest rate. But, how does the Fed determine the optimal interest rate? According to the Federal Reserve Act, the Fed has a dual mandate to stabilize prices and minimize unemployment (Carlstrom and Fuerst 2012). This simplifies to the objective of limiting the variability of inflation and output. John Taylor famously proposed an econometric model where the interest rate is a function of changes in the price level and changes in output. This has led to the development of various monetary policy rules, known as "Taylor rules." The most basic Taylor rule is an OLS regression depicted by Equation 1 below (Ball 2011):
where (Y – Y*) is the output gap with Y being actual output and Y* is potential output and (π–πT) is the inflation gap with πT being the target inflation.9 An important component of the Taylor rule is the Taylor Principle, which states that the coefficient απ should be greater than 1.0. This changes the nominal interest by more than the inflation rate, ensuring that the real interest rate actually adjusts to affect the real economy (David and Leeper 2007).10
One additional feature regarding the interest rate worth noting is the zero-lower bound on the nominal interest rate set by the Fed. That is, no person would save in exchange for a negative nominal return, but would rather simply hold money. So, the Fed cannot lower the nominal interest rate below zero. The Taylor rule, however, may still imply a negative interest rate. Consider monetary policy with the Taylor rule from Equation 1. Say, actual inflation equals the inflation target so the inflation gap is zero. Then, take rn = 1.0 and γ = 0.5. If the output gap is large enough (say -3.0), then the Taylor rule will suggest a negative nominal interest rate. Once a central bank reaches the ZLB in this scenario, it may lead to a liquidity trap. The model computes that the interest rate should be further lowered, but this is impossible due to the ZLB. Even worse, monetary policy is now too tight given the optimal response according to the Taylor rule. This further fuels a lack of liquidity and slows down the economy. A vicious circle– known as a liquidity trap – can develop, characterized by low levels of nominal interest rates, economic stagnation and potential deflationary periods (Bullard 2013).
Several examples exist of this ZLB scenario. Japan has been in a liquidity trap at the ZLB for most of the 1990s and is still facing this issue today. Since 2008, the U.S. and several other countries reached the ZLB during the Great Recession and are still challenged by strategies to exist these liquidity traps. As will be discussed later, this makes the FFR irrelevant because an econometric model based on a Taylor rule does not understand the ZLB constraint. Several policies are available to central banks to escape a liquidity trap. These policies including quantitative easing, purchasing long-term assets, and fiscal expansion (Bullard 2007). As an example of recent U.S. policy, the Fed has practiced quantitative easing, or buying long-term assets like mortgagebacked securities, at a rate of $85B per month. These policy options are often aimed at lowering the long-term real interest rate to provide greater liquidity and induce a robust recovery when the Fed can no longer lower short-term interest rates. For this reason, I contend that including long-term interest rates in empirical analyses is relevant because the Fed's policy is no longer solely aimed at the FFR but is also targeting long-term rates such as the 10-year Treasury rate.
Before analyzing the potential effects of monetary policy on asset prices, it is necessary to understand how asset prices are determined. The classical theory of asset prices states that the price of an asset equals the present value of expected asset income (Ball 2011). The "expected" income derives from the rational expectations assumption, that is, people's expectations of future variables are the best possible forecasts based on all available information. Thus, two variables determine the present value: forecasts of future income and the interest rate to determine present values. Looking at stock price valuation, one can better understand the valuation method of asset prices. The future earnings of a firm flow to stockholders through dividends. Thus, the price of a stock is given by:
If the dividends are assumed to be constant, then this becomes a perpetuity valuation where the present value of the stock is:
Or, as proposed by Myron Gordon, the Gordon growth model theorizes that a stock is determined by an initial expected dividend that is then expected to grow at a constant rate. In this case, the price of a stock is given by:
Finally, it is important to understand the relevant interest rate, as it does not necessarily match the FFR, or the interest rate set by the Fed. Rather, i = isafe + φ where isa f e is the risk-free rate, such as the rate on a ten-year Treasury bond, and f is the risk premium of the asset that the owner receives as compensation for baring the additional risk. Together, i is known as the risk-adjusted interest rate.
Continuing with the valuation of stock prices, it is clear how monetary policy could affect asset prices. Using the Gordon Growth Model, say D = $2, i = 0.05, and g = 0.01. The price of the stock equals $50. Now, let's say the Federal Reserve lowers the interest rate. This can have several transmission effects on the price of this stock. For one, the risk-free rate may decrease. As discussed earlier, the Federal Reserve controls the short-term, nominal rate. However, according to the expectations theory of the term structure, the long-term nominal rate is just the average of expected short term rates. Thus, assume the Fed lowers the interest rate such that the risk-adjusted i decreases to 0.04. In this case, the stock price would rise from $50 to $66.67.
Monetary policy could also affect the actual prospects of the firm's future earnings as well. The function for forecasting a firm's future earnings can take on several forms. Parameters may include management ability (M), historical performance (H), projected competitors (C), investments (I), and any number of other factors influencing production (P). Think of this forecast function in the general form of Equation 5 where any number of parameters could be used, but certainly investment is a critical variable.