Dynamic Adjustment: The Problem of Pro-Cyclical Behavior

As illustrated in Figure 9.2, banks adjust their lending standards over time, but, unfortunately, the dynamic adjustment of credit standards appears to impart a very pro-cyclical bias to the credit cycle. From the graph we can see that the percentage of banks that tighten credit drops dramatically in the early phase of an economic recovery (1992–1994, 2002–2004, 2010–2011) and stays easy for most of the economic expansion. Then the percentage of banks that tighten credit rises sharply just before a recession (1999–2000, 2007–2008). This credit cycle, while certainly rational from an individual bank’s point of view, becomes quite pro-cyclical when viewed in the aggregate.

Price/Rate Adjustments: Sudden and Uneven

Interest rate movements, changes in the price of credit, are both sudden and uneven—a distinct break from the assumption of smooth movements to new information.

As illustrated in Figure 9.3, the variability of the real one-year Treasury yield is dramatic. This variability reflects the impact of several forces that are part of the interest rate framework. What is intriguing is that while the framework may be steady over time, the variability of the individual forces impacts the behavior of the one-year rate. In particular, this economic expansion has witnessed a significant change in the path of anticipated economic growth, expected inflation, and the context for monetary policy actions. Indeed, in recent years, we have witnessed the emergence of a new factor, global markets, into the FOMC’s reaction function. In addition, long-term interest rates are trending downward as several members of the Organization for Economic Cooperation and Development (OECD) have experienced downward trending 10-year Treasury yields since at least the mid-1990s (Figure 9.4. One potential reason for this downward trend may be the lower inflation and subpar gross domestic product (GDP) growth rates in some developed economies.

As these forces evolve, so does our model for interest rate behavior as evidenced by the movement in the short- and long-term yield curves, which we will examine later in this chapter. In economics, the framework constantly evolves, and nowhere is this more evident than in the realm of interest rates.

What Was Our Starting Point?

As illustrated in Figure 9.5, the starting framework posited that the pattern of the funds rate set by the FOMC would be based on the distance from a perceived equilibrium unemployment rate and a target inflation rate. For the period of 1990 to 2008, this appeared to work fairly well. Unfortunately, since 2008, this framework has begun to generate serious errors.

What changed? Our previous research has noted that there was a structural break in the behavior of several series subsequent to the credit crisis/recession of 2008–2009.¹ For the unemployment rate, the decline to the perceived measure of full employment was not associated with a strong labor market and a rapid rise in labor compensation. Meanwhile, the speed of the economy’s approach to the 2 percent inflation target was far slower than anticipated. As a result, current FOMC policy is more cautious in raising the funds rate than implied by the Taylor rule using the traditional unemployment rate (Figure 9.5).

Enter a New Player: Global Developments

Global developments have always lurked off-stage and occasionally made a rude appearance in the interest rate environment, yet these developments have now added two elements of uncertainty. First, how do we measure this factor? Second, how does this factor fit our model? Since the FOMC introduced this factor into its policy statements, the importance has become evident. One measurement of the importance can be approximated by the safe-haven buying of financial assets as illustrated, in part perhaps, by private sector purchases of U.S. financial assets (Figure 9.6). In fact, when we include global developments in a forecasting model, we get a better forecast for the U.S. 10-year Treasury.

Moreover, the persistent decline in global yields in recent years has reinforced the message of a change in global growth expectations and the inability of key G7 central banks to reignite inflation toward a 2 percent target.

Choices: A New Framework and a Bullish Flatter Yield Curve

Ever since the taper tantrum discussion in 2013, there have been two distinct moves in the yield curve as illustrated in Figure 9.7. The long end of the yield curve has exhibited a bullish flattening trade with the decline in the 10-year 2-year spread. This reflects the yield pick-up for U.S. Treasury debt relative to what is available for investors in Europe and Japan while also reflecting the incentive of a stronger dollar to attract foreign inflows. Meanwhile, the short end of the yield curve reflects the anticipation of a FOMC increase in rates or at least some form of tighter policy going forward. The uncertainty in the market provides the motivation for a safe-haven move while limiting the extent of Fed tightening.

Previously, we dealt with the reality that the framework for interest rate behavior has evolved over recent years. But what happens if policy itself is inconsistent and what are the broader implications for the market?

The Price of Success

Since 1994, the PCE deflator has averaged less than 2 percent (Figure 9.8). Shall we, then, just declare victory and go on vacation? Alternatively, given this success, what happens if a policy maker wishes to push the envelope to achieve greater economic growth/employment given the current below 2 percent inflation even as the unemployment rate has reached the full employment level? This is our trade-off today.

For the policy maker, the advantage is that, with current inflation so low and inflation expectations falling over the past six months, the marginal cost of additional inflation is low. Even if inflation were to run above 2 percent for a while, the perceived damage to inflation expectations appears limited to some.

Collateral Damage along the Road to Inflation

However, there are economic issues associated with continued easing in the pursuit of growth, and the risk of rising inflation through increased injections of liquidity. The flow of liquidity can, like water, move anywhere. As a result, the distortions in private markets may already be appearing, not through mispriced goods (inflation), but through financial/real assets.

Persistent zero rates made it easy to justify the purchase of new equipment in an industry (energy) perceived to have a growing output potential. The result was excess capacity in an industry now undergoing retrenchment. In the meantime, low rates allowed firms with visions of limited potential to substitute capital for labor.

As illustrated in Figure 9.9, the policy of administered low short-term rates has produced an era of financial repression where the real return to short-term investments, one- and two-year Treasury yields, was negative. The proper pricing of a risk-free short-term rate is not possible given that the rate is an administered rate by policy makers and not the marketplace. For savers to gain a real return, the imperative is to extend maturity out to five years.

Buy Existing, Not New

With persistent low interest rates, firms may find it easier to purchase existing equipment rather than buy new. This helps to explain the coincidence of high levels of merger and acquisition activity this cycle along with subpar business investment. As illustrated by the Tobin Q ratio (Figure 9.10), when the Q ratio is below 1.0, then the market value is less than the recorded value of the assets of the company, intimating the market may be undervaluing the company. In this case, the incentive is to buy the company rather than buy new capital. Low interest rates are providing an incentive to invest—not in new equipment but rather in existing equipment given an environment with excess capacity.

On the Variability of Pricing Risk over the Cycle

Another means to judge the credit cycle is the pricing of high-yield debt. As illustrated in Figure 9.11, late-cycle pricing tends to show a steady rise in the risk premium even before the signs of recession are evident in the real economy. From 2000 to 2001 and again over the past year, there has been a rise in the risk premium for high yield credit. During 2007, the premium rose dramatically and spiked in 2008 with the onset of the recession. As the economic recovery takes hold, the risk premia tend to decline fairly steadily until just before the next recession starts the cycle again.

The outcome is that pricing of risk in the high-yield bond market is actually quite sensitive to changes in market perceptions. This pattern is reminiscent of the pattern of corporate profit growth we have written about before as another signal of potential economic difficulties. Both indicate that credit premia are rising with the aging of the credit cycle.

On the Variability of Credit Delinquencies over the Cycle

Household debt delinquencies (Figure 9.12) have their own interesting cyclical pattern. Delinquencies are fairly steady mid-expansion (2004–2006) but then gradually start to rise as the latter stages of the economic cycle appear (2006–2007) before they rapidly escalate into the recession. Delinquencies peak after the recession and then gradually decline as the recovery proceeds. One notable exception has been the secular rise of student loan delinquencies.

So Why Are Downturns More Predictable?

Our three patterns of the credit cycle raise the question of why recessions are not more often easily predicted or recognized. From our work, the conditions for a recession may be present but the timing and source of the spark that lights the recession lamp is seldom obvious.² At present (based on 2016:Q1 data) the probability of recession remains below 50 percent for the next six months, but we are on watch given the elevated risk premia.

A Look at Policy History in the Short Run

There also appears to be a drift in the FOMC’s assessment of the longer-term fed funds rate forecast during the short run of the current economic expansion. The FOMC has consistently lowered its expectations for the longer-term funds rate.

During this cycle, we have also witnessed a consistent drop in the estimates of potential GDP growth by the Congressional Budget Office (CBO), as illustrated in Figure 9.14, largely due to slower labor force growth and a slowdown in productivity growth. Growth estimates have consistently moved toward a 2 percent consensus as opposed to the 3 percent plus that was the standard for many forecasts—both in the private marketplace and among policy makers. Lower potential growth estimates would incentivize firms to borrow less to invest in future production and lower rates would lead households to save more, leading to a lower natural rate of interest.

Downshifting Equilibrium Rates Produces Lower Policy Rates

The persistent downshifting in policy expectations can also be observed in the dot plot (Figure 9.15). For example, the long-run fed funds rate projections by the FOMC began at 4.25 percent, and have consistently been downshifted to the current estimate of 3.25 percent. The global decline in yields indicates that the drift to lower rates is a global pattern and the forces behind these movements are likely global as well. For investors and policy makers, there is no return to the past. We have moved into new territory with new opportunities and risks.

Identifying Trend: The Anchoring Bias

What has been the trend in 5- and 10-year yields since 1968, the start of rising inflation and interest rates in an activist policy era, and are those trends reliable guides for the future? How permanent is permanent? How normal is normal? Figure 9.16 highlights the problem of an anchoring bias for the period since 1968. For both the 2- and 10-year Treasury benchmark yields there are two distinct patterns. First, there is a steady rise in interest rates from 1968 to 1979 and then a distinct downtrend thereafter. This highlights the issue that there have been two different economic regimes and, indeed, we know that since 1979, Paul Volcker’s focus on inflation led to a new set of central bank objectives. Second, there is an equally apparent steady decline in nominal interest rates since 1982. These two distinct patterns indicate that neither the 2-year nor 10-year Treasury rate over this period are mean reverting, and treating the 2- and 10-year rates as a coherent series since 1968 is not the correct approach. Yet many analysts will employ the extended period since the 1960s as their sample set when developing econometric models. In prior work, we have found separately that the 5-year Treasury rate is also not mean reverting.³

In Table 9.1, we show the calculations that indicate that the average and standard deviations of interest rates are distinct between the two periods and that the average values of the 5-year and 10-year rates are statistically different. For example, for the 1968–1981 period the average value of the Aa corporate bond was 8.99 percent in the first period and 7.82 percent in the second period.⁴ Are these values significantly different such that we can proceed on the assumption that they represent two different interest rate regimes?

One counterintuitive result is that once we compare the means and standard deviations between these two periods, we note that the stability ratio—a series’s standard deviation as a percent of its mean— is actually higher in the second period than in the first period. This is true for the four interest rates examined in Table 9.1. The evidence from the stability ratios indicates that despite the recent low level of interest rates, the volatility of interest rates has actually risen in recent years.

We can also test for a structural break in these series by employing a State-Space approach.⁵ The results listed in Table 9.2 confirm the structural break. There is clear evidence of a structural break in the level of market interest rates in the fourth quarter of 1982. For example, a break occurred for the Aa corporate bond yield, 5-year Treasury yield, and 10-year Treasury yield, which are all statistically significant at the 0.01 level. The table also shows a second intriguing result: the structural break for the Baa corporate bond occurs in 2008:Q4, just after the collapse of Lehman Brothers.

These results illustrate the important role that intellectual biases can play in investing and economic forecasting. Here, the problem of the anchoring bias appears in two ways with respect to the level of rates.⁶ Based on the experience of the early post-WWII period, investors were accustomed to low inflation and were surprised by the blowout of inflation in the late 1970s. Meanwhile, the next generation of bond investors anchored their expectations of inflation on the experience of the late 1970s and failed to anticipate the drop in inflation and interest rates in the 1980s.

As noted earlier, the structural break in corporate bond yields followed the Lehman Brothers collapse. This represents a challenge to interest rate modelers since we must now recognize that since 2008 we are in a different sampling period with a new interest rate regime. This may help explain the problems for many forecasters in accounting for the continued low level of interest rates compared to the historical lineage.

The Dramatic Shift in Real Interest Rates

Real interest rates were consistently below zero during the 1970s and above zero throughout the post-1970 period until the recovery from the 2001 recession (Figure 9.17). These patterns corroborate the view that investors in Treasury debt were consistently surprised by the rise in inflation during the 1970s. Then, during the 1980s, inflation fears persisted and nominal rates did not completely adjust to the rapid drop in actual inflation from 1982 to 1992. These behaviors imply that investors exhibit a dynamic adjustment process where they gradually learn about the path of policy and policy’s implications for inflation. Therefore, real interest rates partially adjust to lower actual inflation.

The current period is also unusual since the conduct of monetary policy is aimed at keeping nominal interest rates very low while promoting a rise in inflation. The net result is that real interest rates remain negative throughout this period. Lowered expectations for economic growth and inflation may reflect, in part, the experience of this recovery but also the impact of higher taxes and underlying changes in labor force growth and productivity.

Incomplete Information: Regulation and Capital Spending

First, there is the issue of incomplete information—economic agents do not know all the facts and may delay decisions or make different decisions than what would have been made if perfect information was available. This is evident today in the full plate of regulations and capital requirements associated with the Dodd-Frank Act, the Affordable Care Act, and Basel III, since the full set of rules has yet to be spelled out. Moreover, there is the legacy of legal risks from the events of 2004–2009 that continues to leave a sense of caution in financial markets today.

Limits to Financing the Economy

Complete information on policy actions is not available, so we must recognize that current financial markets have significantly more risk embedded in them than commonly perceived. Therefore, lending and investment decisions depend on a broader set of economic and policy factors than commonly associated with those decisions. For example, business investment is commonly modeled as a function of expected final sales, market interest rates, and current cash flow/profitability. However, in an environment of uncertain tax, regulatory, and credit policy, business firms would be projected to invest less given the higher level of policy uncertainty, as discussed in Chapter 5.

Given the uncertainty of economic policy, the high level of cash on hand at nonfinancial firms and households, as well as the high level of reserves held by private banks, is more understandable. As illustrated by Townsend, there is a significant cost to verification that will impact economic activity.⁷

Imperfect Information: Altered Reality

Second, the case of imperfect information—information that does not perfectly reflect reality—is often the reality facing decision makers in the current environment of administered interest rates. Administered interest rates necessarily represent imperfect information—they do not represent the forces in the marketplace, and like wage-price controls, rent controls, and usury rates, often lead to pricing distortions. These distortions make it exceedingly difficult to make proper risk-reward trade-offs.

In credit, there are the issues of adverse selection, moral hazard, and questions on the quality of bank capital in the United States, Europe, and especially China under the conditions of administered—not market-set—interest rates. Central banks around the globe are influencing short-term interest rates to remain below levels that would be present in an open, private capital market. Moreover, unconventional monetary policies can further distort market pricing.⁸ Finally, there are always questions on the quality of corporate profits and market valuations under a regime of administered rates. How can an investor properly assess the benchmark risk-free interest rate, reinvestment risk, or the refinancing risk associated with any financial commitment today, given that interest rates may adjust dramatically once the era of administered rates has ended?

Economic variables that we model are presumed to influence, or at least represent, the actual economy. However, imperfect information or unanticipated economic factors drive real-time decisions. Imperfect information on interest rates under the current monetary policy regimes—whether the current level of interest rates or the spread along the yield curve between corporate and Treasury debt—will lead a decision maker to be uncertain as to how much to attribute any interest rate change to either:

1. A change in the relative pricing of financial assets (bonds, cash, or equities),
2. A general decline in the pricing of all assets, or
3. A change as the result of a policy action to alter administered interest rates.

Therefore, this uncertain pricing will lead to a departure from output, credit, and exchange-rate pricing in an open market setting and generate changes in economic behavior that are an increasing function of the distortion and persistence of administered interest rates. For decision makers, the crucial problem is to determine how much of our current macro data may reflect activity that is the result of administered interest rates and would not occur, or in some cases would occur quite differently with different pricing, in an open market setting.

Backlash to Administered Interest Rates: False Information Cannot Deliver True Outcomes

Policy-manipulated prices cannot provide a true gauge of consumer demand nor producer supply in the marketplace. The same is true for administered interest rates. For private economic agents, if there is a sense that the current set of market interest rates, labor costs, or input prices do not accurately reflect the real cost of an economic asset, especially over time, then economic agents will not actively pursue any activity based on imperfect pricing schemes. For example, one lesson of the teaser rates of the past decade is that low interest rates today should rise, and when they do there is significant refinancing risk. Second, temporary tax cuts meant to spur hiring are short term and therefore not an incentive to hire a worker for the long term. The hiring firm must still assess the return on the worker compared to the cost once the temporary tax credit expires. Temporary tax programs for inputs, such as for energy, will not alter the long-run cost trends, since firms that invest long-term capital cannot do so wisely based on short-term tax breaks.

Administered interest rates, independent from market fundamentals, are most evident in sovereign debt rates in two ways. First, many central banks are buying their own country’s debt and holding sovereign debt at levels out of proportion with their historical position (Figure 9.18). Second, increased capital requirements, in the attempt to create financial stability, create a demand for sovereign debt on the part of private financial institutions, which assists in the buying of public debt at the risk of crowding out lending in the private market. This helps explain the continued low level of sovereign debt yields despite the outsized deficit and debt levels in many industrialized and emerging-market countries.

In policy making, the problem of imperfect information arises in two distinct paths. First, while there is one monetary policy, we often hear from several different Fed speakers that create different impressions for the direction of policy. Second, the Fed is currently following a broad set of labor market indicators to determine the direction of policy. While this may make good policy, a central bank that follows multiple labor market indicators can send a confusing signal to private-sector investors. This policy-making problem is further complicated by the replacement of the initial unemployment rate guidepost of 6.5 percent by a discussion of a lower guideline and an expansion of labor market indicators. This problem is compounded even further when several different inflation guidelines are offered (core and headline inflation, consumer price index [CPI], and personal consumption expenditures [PCE]). As the adage goes, a man with one clock will know what time it is, but a man who has two clocks is never sure.

Information: Future Uncertain

Information is also dynamic over the business cycle and over time itself. For example, information on the pace of economic growth and employment growth changes over time. This changed perspective leads to alternative or regretted decisions that would have been different if economic agents did indeed have perfect foresight or perfect models of the economy. It is said that in war, battle plans change after the first shot. In recent months, the significant revisions to U.S. GDP have certainly altered perceptions of the economy’s momentum and of monetary policy.

In credit markets, we have witnessed how policy benchmarks (such as the unemployment rate) have been rebenchmarked by policy makers at the Fed and the Bank of England, when the target rate was achieved more quickly in their respective countries than was initially forecasted. Moreover, the unemployment rate benchmark in the United States has been replaced with a broader set of labor market conditions due to changing underlying fundamentals, such as the participation rate. Whatever the intention of policy makers by rebenchmarking, the impact on investors and financial asset pricing will be, first, a reassessment of policy goals and, second, less specific guidance on the actions of policy makers. In both cases, information from policy makers is less than perfect in its forward guidance and, therefore, investors’ mistakes are bound to occur given the fog of policy making.

During and after the Great Recession, regulatory policy making has altered the information set for investors, creditors, and borrowers. Regulators suffer from the reality that implementing new laws takes time and interpretation. Many firms exhibit a framing bias away from the risk the new rules and regulations introduce, and could put investment and credit decisions on hold in the short term.⁹ This bias for risk avoidance could lead to a reduction in the supply of credit, and, thereby, a slower pace of economic growth. Indeed, this outcome came to pass in the current economic recovery as firms adjusted to new regulations like the Dodd-Frank Act, which reversed some common market behavior and standards.

Credit decisions in recent years have also been impacted by court decisions on municipal bankruptcies, corporate bankruptcies, sovereign debt and federal mandates, or simple rewrites of federal/state laws that upset the previously understood relationship between creditor and debtor. Once again, we witness the shift in preferences away from risk taking, resulting in less credit supplied and slower economic growth. Finally, recent years have witnessed shifts in sovereign government commitments to exchange rate regimes and trading agreements as one political party assumes leadership from another in a given nation. Once again, changes in exchange rate regimes or trading relationships will impact credit markets in a way that is outside the traditional fundamentals of economic growth and inflation expectations.

Asymmetric Information—The Reality of Setting Interest Rate Benchmarks

Asymmetric information is a driving force of financial market imperfections and is the product of a marketplace where parties do not have the same information. In the traditional sense, asymmetric information arises when one party to a transaction is better informed on their future actions than the other party.¹⁰ Credit card issuers do not have as clear a picture of the credit quality or the intentions of the credit card borrower as the borrowers themselves. While generally associated with private borrowers and lenders, recent years have witnessed significant surprises from public-policy makers that have impacted the realized returns to private investors in public debt. For example, in recent years the true credit quality of sovereign debt in several countries was not well known in the marketplace.

Private investors have an incentive to acquire all relevant information on future economic conditions and policy actions. However, the literature is clear that unanticipated policy actions, not already anticipated actions, will lead to changes in economic activity.¹¹ Therefore, policy makers have an incentive to surprise the marketplace to generate the desired responses to public policy.

Louis XVI’s continued devaluation of the public debt during his reign set the pattern of public policy surprises and reflects the power of the sovereign to alter the rules of the investment game. At first glance, sovereign debt in most countries is considered fairly safe from nominal default. However, the real value of the public debt can be devalued by inflation, currency depreciation, and new taxes on the future returns to public debt. In effect, financial repression by public-policy makers act to keep nominal interest rates low while increasing inflation, such that the real value of the debt continues to decline. For example, in the United States, inflation in much of this recovery has outpaced the level of short-term interest rates such that real returns have been negative for investors at the short end of the yield curve.

Credit Rationing: Role of Nonmarket Factors

Credit rationing represents another imperfection in the financial marketplace that often results from incomplete information—particularly in business and housing finance. When the payoff to the borrower/ entrepreneur on some project exceeds a critical level, then the borrower/entrepreneur is able to pay the creditor/investor his expected return. Then, however, the creditor/investor has little incentive to monitor the ultimate payoff for the investment by the entrepreneur since she has already received her required return.¹² For example, if a lender makes a loan at 5 percent interest but the project actually returns 12 percent, the lender has little incentive to monitor the returns to the excess return. The lender has made her 5 percent required to return.

In contrast, if the project is not projected to yield the required minimum return of 5 percent, then credit is rationed and lending does not occur at any interest rate such as 6, 8, or 10 percent. Unfortunately, when regulatory policy is changing rapidly or is suspected to change in an uncertain way, regulatory risk will inhibit lending at any interest rate, especially to marginal credits, as witnessed during the early period of the current economic expansion.

Credit rationing leads to discontinuity in the marketplace, which is often the result of nonmarket forces acting to inhibit the proper assessment of risk/reward. This is another example where the lack of a perfectly competitive marketplace engenders interest rates to allocate capital inefficiently.

We also have the example when sovereign debt is rated as Tier 1 capital by regulatory institutions and yet the credit quality of all sovereign debt is noticeably not equal. Arbitrary ratings on sovereign debt lead to narrower spreads than would be assigned in an open, free market, which results in a misallocation of capital in global markets. We witnessed this pattern prior to the 2007–2009 recession. Sovereign debt traded at interest rate levels below what would be achieved in the private marketplace. The pricing of corporate and mortgage debt, which is often priced off Treasury debt rates, would also exhibit signs of mispricing.

Our review, so far, indicates that agency costs arising from asymmetric information raise the cost of external financing and further discourage real economic investment. In addition, financial market imperfections create agency costs that affect investment, altering the impact of anticipated final sales, profitability, and interest rate moves on investment. Policy uncertainty further alters private agents’ ability to make judgments on internal/external financial options.

Information Complications for Policy Effectiveness

These incomplete information issues further complicate the assessment of fiscal, monetary, and regulatory policies on economic growth, inflation, interest rates, and exchange rates. Economic actors will be hesitant to react completely to new policy actions or economic information given the higher degree of uncertainty associated with incomplete information. This gives rise to the dynamic adjustment issues discussed in an earlier chapter. The result, therefore, is that predictions made on the effectiveness of policy actions, with the assumptions of a perfectly competitive marketplace, will not come to fruition in the world of incomplete, incorrect, and uncertain information.

Households and firms are uncertain on the current economic environment and suspected future policy options. Therefore, any good news would only draw a partial reaction, and, hence, economic activity delivers less than the full anticipated outcome to policy initiatives than would otherwise be projected. This has certainly been the case in the United States during the current economic recovery/expansion.

One intriguing note is that many variables that do not affect investment in perfect capital markets matter very much in an imperfect capital market. Average tax rates, idiosyncratic risk, and policy uncertainty enter the picture. The average corporate tax rate, as well as the marginal rate, impact investment by reducing the firms’ ability to exploit internal finance. Risk impacts agency costs. Policy uncertainty impacts household and firms’ assessments on the anticipated future returns on any contemplated economic activity.

Finally, the combination of dynamic adjustment and incomplete information gives rise to a broader set of possible economic outcomes that would not have been considered within the confines of the perfectly competitive model.

Incorrect Information—The Price of Inaccurate Information

What is the price of making economic decisions based on incorrect information? When we assume perfectly competitive markets with efficient and accurate information, we are idealizing the decision-making environment for households, firms, and investors. Unfortunately, inaccurate information on growth, inflation, and credit quality abound and may lead to decisions and economic outcomes that would not occur if we had perfectly accurate information.

Decision makers must make investments based on their reading of the economic releases that are frequently revised. In fact, the variability of the employment release, arguably the single most important monthly economic release, is notable for the significant revisions that occur in subsequent months. Measures of inflation reflect the well-meaning but arbitrary adjustments for quality changes, as well as a number of implicit price assumptions. Moreover, the existence of administered prices and direct price controls, as well as sales and excise tax changes, will alter perceptions of price trends for many economic agents. Unemployment rates come in several varieties, from U-3 to U-6, presenting the decision maker with numerous alternatives to judging the labor market, each with its own measurement problems.

Our financial system is an indicator of real investment performance and the efficient allocation of capital. However, there are many issues with judging the accuracy of information of individual and corporate credit. The experience of the 2007–2009 recession period drew out the visible flaws in credit ratings for corporations, states, and mortgage securities and denotes that credit quality differences were not accurately reflected in the credit ratings utilized by lenders and investors. Moreover, public-policy makers, through incentives and credits, alter the risk profiles of selected investments. Sometimes the proper risk is not accurately reflected in interest rate spreads. Therefore, the real market risk of a project is not recognized until too late, as we have observed with several municipal issues in recent years.

Over the long run, incorrect signals on the real risk of an investment will lead to the persistent misallocation of capital over time. Questions on the pace of real economic growth, for example, are aimed at several emerging markets today. Inflation may also be understated in many countries, including several advanced nations, since many prices are controlled by governments and accurate comparisons between nations on the pace of inflation are difficult to judge. Moreover, comparisons on real exchange rates and trade competitiveness are also difficult. Government’s role in housing policy and the secondary market will cloud the assessment of the real return on housing and mortgage finance. McKinnon provides a starting point to address the relationship between economic growth and financial markets.¹³

Yield Curve Term Premium in an Administered Interest Rate Marketplace

Interest rates on long-term bonds should reflect the expected average of the interest rates on short-term bonds over its lifetime—but what happens when selected parts of the yield curve are manipulated for public policy objectives?

The committee is prepared to increase or reduce the pace of its asset purchases to ensure that the stance of monetary policy remains appropriate as the outlook for the labor market or inflation changes.

—FOMC Statement, May 1, 2013

A reduction in Fed-provided liquidity would be projected to raise short-term interest rates due to reduced liquidity in the short run and lower long-term interest rates over time as expectations for the pace of growth and inflation would decline. Yet from May to October, the yield curve steepened as 10-year Treasury yields rose relative to 2-year Treasury yields. In an environment where growth and inflation changed very little, there was a clear rise in the term premium for Treasury notes.

The Committee decided to await more evidence that progress will be sustained before adjusting the pace of its purchases.

—FOMC Statement, October 30, 2013

This halt to the threat of tapering would signal a reversal of a move to tighten policy immediately and thereby generate a flatter yield curve, which indeed is what happened initially, and yet the impact was very short-lived.

The Committee decided to modestly reduce the pace of its asset purchase.

—FOMC Statement, December 18, 2013

The action to reduce liquidity over time by the FOMC would have been foreseen to flatten the yield curve as short-term rates would increase relative to long-term rates. This is what happened during the December 2013–April 2014 period. There was a clear decline in the term premium in direct contrast to the experience of May 2013–December 2013.

This peculiar period highlights the problem of many interest rate models that solely employ economic factors such as expectations and estimates for growth, inflation, deficits, and the federal funds rate as inputs. In an era of administered interest rates by the monetary authority and forward-looking financial agents, changes in interest rates will reflect changes in sentiment and Fed commentary that are difficult to catch within the traditional modeling process.

Both monetary policy and financial regulatory policy will alter the term premium in the marketplace (as well as spreads between corporate and federal debt, for example) all along the yield curve. As a result, investors cannot be certain of the real, market-setting term premium. Nor can investors be certain of how the market will react when the era of administered interest rates ends.

For example, in theory, a “permanent” increase in money growth eventually increases the short-term nominal interest rate—but investors are unsure about how long “permanent” will last. For example, investors debate today to what extent the Fed will maintain its enlarged balance sheet. The liquidity effect of easier monetary policy has reduced short-term interest rates and assisted in liquidity for the equity and bond markets as well as several emerging-market currencies. However, this effect disappears when “permanent” becomes something less. This impact of less-than-permanent monetary policy ease became evident in the sell-off in bonds and emerging-market currencies in mid-2013.

The Policy Maker’s Catch-22—The Dynamic Inconsistency of Current Policy and Implications for Private Decision Makers

One of the catch-22s of policy is that anticipated, well-announced policies have little economic impact upon announcement. Kuttner finds that since 1989, there is no evidence that anticipated monetary policy moves have any impact on interest rates.¹⁴ Therefore, there is a question of whether forward guidance has any impact on markets at all. Moreover, well-telegraphed policies, such as the Fed’s tapering program, should also have little to no market impact since that policy is fully anticipated once that policy is announced.

Today, the challenge for policy makers and private-sector decision makers is that a consistent, well-telegraphed monetary policy will not move the markets. Instead, policy must surprise the marketplace to have a market impact. We witnessed this in May 2013, when Chairman Bernanke surprised markets when he indicated a tapering of large-scale asset purchases was in the offing. As a result, interest rates rose rapidly.

Moreover, Kydland and Prescott show that policy makers have an incentive not to stay with a commitment to a consistent low-inflation policy since the additional cost of slightly more inflation is perceived to be very low. Policy makers will therefore pursue an expansionary policy to achieve a greater pace of growth (and lower rate of unemployment).¹⁵ In recent months, we have witnessed a tendency on the part of some U.S. monetary policy makers to allow for the possibility that inflation could drift above the formerly perceived 2 percent target for a while in order to achieve a higher pace of growth and a lower unemployment rate. These recent comments fit the pattern that Kydland and Prescott analyzed, where policy makers would seek a little more growth at the price of a little more inflation, but also at the expense of lowered credibility of a set 2 percent inflation target. For private-sector agents, however, there is also a problem of judging just how much and how long policy makers will accept an inflation rate above the 2 percent target.

Once economic agents anticipate a given inflation target, the policy makers, given discretion, have an incentive to allow just a bit more inflation to achieve just a bit more growth/lower unemployment rate. However, given the long-run neutrality of money, higher inflation will, over time, be met by marginally diminishing improvements in employment and real growth. In addition, higher inflation also has the advantage of depreciating the real value of federal government debt. This pattern of financial repression, where inflation exceeds the administered level of nominal interest rates, will lower the real value of government debt and thereby lead to a loss of wealth by government bond holders.¹⁶

This dynamic inconsistency problem persists in many other areas of policy, such as financial regulation, housing, and fiscal policy, where there is an incentive for policy makers to give the appearance of a set of goals in the short run and then renege on those goals to achieve a bit more success in achieving another particular goal. In reality, the public does not know in advance whether policy makers truly share their preferences on inflation and growth—witnessed in the grand debate surrounding the initial creation of the euro and the European Central Bank, where so many Germans were skeptical that their strong anti-inflation preferences would be honored.

For the U.S.-based decision maker today, there is a great deal of uncertainty about the policy makers’ trade-off between inflation and growth and, for now, there appears to be an acceptance on the part of policy makers that a little more inflation will be acceptable in order to reduce unemployment, but at the cost of debt depreciation of U.S. sovereign debt. Yet, over time, there is no long-run trade-off between inflation and unemployment.¹⁷ Moreover, there is a worrying pattern in recent years that the unemployment rate gives the appearance of being relatively steady while the rates of inflation in consumer prices and producer prices have started to rise on a year-over-year basis. Although inflation rates may be perceived as low by some, inflation is rising and that change will lead to an uncertain degree of market response in a world of administered interest rates.

Predicting Yields—Multivariate Analysis

As U.S. Treasury yields have remained at historically low levels, it is important to consider whether the tools applied to predict yields have changed. Using historical benchmarks to perform analysis on yields may be misleading, given the permanent shift in the relationship between asset classes since the past recession. This shift can be attributed to changes in the anticipated pace of economic growth and inflation, future tax changes, and changes in the balance of and demand for Treasury debt.

When performing analysis on the yield curve, it may be more practical to look at the direction of change rather than the yield levels compared to historical norms. In addition, looking at the pattern of yield movements in other countries (specifically, the G7 countries) may be a practical tool in understanding movements in the U.S. Treasury yields. In fact, the two-way relationship between global yields and the U.S. Treasury yields implies that changes in U.S. Treasury yields can also have predictive power over global yields.

Credit Spreads: A Break with History

Traditionally, credit spreads vary over the business cycle. Spreads tend to widen during periods of economic weakness or uncertainty and narrow during periods of economic prosperity. Therefore, periods of optimism can be represented by a tightening in credit spreads, while pessimism is associated with increases in spreads. These patterns reflect the dominance of cyclical forces—not secular change—and yet, secular forces may indeed be the more important driving force in interest rates since 2007.

Tradition may be taking a backseat. For analysts, the challenge is to recognize when credit spreads are at extremes and when such spreads provide a signal of a possible change in the economy, or at least sentiment on the economy. Behind the utilization of any cyclical pattern as a guideline is an implicit assumption that spreads may vary and that they will vary around the same mean value over time and over different cycles. However, how might we assess changes in sentiment as represented by credit spreads if, in fact, the average values and their volatilities vary over time?

Table 9.8 shows that credit spreads between corporate bonds and Treasury bonds have, on average, risen during the post-1982 period. In addition, standard deviations have also risen, while stability ratios have actually declined. The smaller stability ratio for the 1982–present period implies that the volatility of spreads has declined in recent years. How can we measure changes in benchmark credit spreads as a signal of possible change in the economy? Moreover, can we identify a structural change in credit spreads since 1968?

Testing for a structural break in credit spreads is crucial, as a positive finding indicates that a series has changed for a specific time period when compared to its historical norm. A break implies that a benchmark for a series—for example, the average level of volatility for a given period—has shifted compared to historical standards. In this example, we can test to determine if there has been a structural break in a credit spread. Given a structural break, it would be misleading to employ a historical benchmark in analysis. Following a break, a benchmark may be higher or lower than the historical average.

We test for a structural break in credit spreads using the State-Space approach, with the results presented in Table 9.9. Using the results from the table, we are able to determine that yield spreads between corporate and Treasury bonds did experience a shift during the past recession (2008:Q4) that was significant at the 1 percent level. For the Aa corporate 5-year spread, there is evidence of a structural break during the Volcker period (1980:Q2) and again in 2008:Q4, the Lehman shock.

Patterns since the Great Recession: Corporate Debt Yields and Equity Earnings—Case against the Central Wisdom of Low Volatility

There may have been a shift in the relationship of returns between asset classes since the Great Recession. Typically, an increase in economic growth is associated with an improvement in earnings and a rise in interest rates. Alternatively, weak economic growth is associated with weaker earnings and a decline in bond yields. In the expansion of the 1990s, S&P 500 earnings declined along with declining Baa bond yields (Figure 9.19) as would be anticipated. However, in more recent cycles, the pattern has not always held. Earnings yields rose, while Baa yields fell in the early parts of the 2001 and 2009 expansions (2001–2003 and again during the 2009–2013 period). The Aa corporate 10-year spread repeats this pattern. In contrast, the Lehman shock appears the dominant factor in the Baa corporate 5-year spread and Baa 10-year spread. Has there been a change in the relationship between bond yields and the S&P 500 earnings yield? Is there evidence of a structural break in this relationship, particularly since the past recession?

One simple way to identify a possible shift in the relationship is to calculate the mean, standard deviation, and the stability ratio during several economic expansions. The mean for the S&P earnings yield in the most recent period (2007–2014 [Table 9.10) exceeds the Baa corporate bond yield mean, which is different than in the first three periods. While there appears to be a shift in the mean, is there a change in volatility?

	S&P 5OO Forward Earnings Yield			Baa Corporate Bond Yield
Period	Mean	Std. Dev.	Stability Ratio	Mean	Std. Dev.	Stability Ratio
1992–2014	6.55	1.24	18.9	7.11	1.20	16.9
1992–2000	6.28	1.18	18.8	8.09	0 61	7.6
2000–2007	5.77	0.92	15.9	7.12	0.87	12.2
2007–2014	7.55	0.82	10.9	6.06	1.06	17.5

Source: Bloomberg LP and Federal Reserve Board

To gauge how volatility among series may have changed over time, we can compare the stability ratios of different time periods. If the ratios of the recent period are smaller than the past, then we can conclude that volatility has declined over time. As shown in Table 9.10, the mean of the S&P 500 earnings yield was highest in the 2007–2014 period, while its standard deviation and stability ratio were both lower when compared to the other periods.

These data imply that earnings have behaved differently since the start of the Great Recession when compared to the past. The Baa corporate bond yield has the smallest mean along with a fairly large standard deviation, leading to the largest stability ratio for the 2007–2014 period compared to the past two subperiods. That is an indication of different behavior in the Baa series as well. Curiously, with a large stability ratio, this argues against the case that the recent period is one of low volatility. There appears to be confusion between a low mean value of Baa rates and their volatility.