Why Infrastructure Bonds Default

Promoters of infrastructure funds argue that investments in infrastructure projects offer less risk to investors because they provide necessary services to local and national economies. Consistent demand to make use of infrastructure projects makes for stable cash flows during long investment horizons, a quality that is unique to the set of opportunities that investors consider when seeking returns on their capital. To support this assumption, promoters often point to the experience of the municipal bond market in the United States, which is the broadest and deepest capital market that supports infrastructure financing. Over the course of decades, the default rates of U.S. municipal bonds has been extraordinarily low, giving investors a sense of comfort that their capital is not be heavily exposed to credit risk. My experience in the municipal bond market has led me to conclude that the above argument is a fair one if placed in the appropriate context. Defaults or restructurings of infrastructure projects are more prevalent than investors realize but they receive less attention by those who track and report on the progress of investor capital. During a time like the present in which interest rates are low and credit spreads narrow, investors gravitate toward higher yielding investment opportunities and they struggle to detect when a yield level signifies an elevated danger of loss of capital. In the infrastructure debt market, such projects have several characteristics that I will highlight. While at Nuveen Investments, I directed a study of the bond defaults and restructurings that we experienced in the bond funds that we managed. Our study reviewed our investment experience between 2001 and 2003, a period during which companies like Enron, WorldCom and United Airlines entered bankruptcy and Arthur Anderson dissolved. Preceding this period was a six year span in which interest rates declined, the equity market appreciated significantly, investors were stretching for yield and borrowers were levering balance sheets. A credit crisis commenced when the stock market crashed and corporate bond default rates eventually rose to a peaked of 15% in 2002. With approximately $35 billion in infrastructure debt in its various funds, Nuveen had a unique population of infrastructure projects that we could examine to understand why some infrastructure projects failed to cover the costs of financing. At that time, we had invested in over 2000 infrastructure project bonds that were rated BBB or lower and 200 that were not rated by either Moody’s or Standard & Poor’s. The results of the study yielded the below conclusions.
Read More

What Really Drives the Performance of Infrastructure Funds?

Written By: William M. Fitzgerald
Researched by Kelsey Smith

Download the pdf of this report
Written: October 19, 2012


Why We Are Curious

Earlier this year, I participated in a conference call with a member of the investment staff of one of the large corporate pension funds in the United States.  We asked him to describe their approach toward investing in infrastructure.  He stated that the investment staff is interested in infrastructure because it expects the performance of the asset class to diversify their equity market risk and to hedge against the risk of inflation.  One of the vehicles that the staff chose to make this investment is a listed infrastructure equity strategy sponsored and managed by a prominent U.S. fund company and he stated that the staff was quite pleased with the results.  As the call continued, I pulled up the returns of the infrastructure fund on my Bloomberg terminal and compared its investment performance to that of the S&P 500 Index.  Curiously, the infrastructure strategy performed nearly identically to the S&P 500 index since the inception of the infrastructure fund in 2007.  At a quick glance, the infrastructure strategy didn’t appear to meet one of the objectives – diversifying equity risk.  We decided to conduct a more thorough investigation.

The Context

We examined the marketing materials for listed infrastructure funds of two prominent investment companies that are leaders in the infrastructure investment marketplace: Fund A is managed by a large U.S. fund company and the largest investor in infrastructure in the United States; Fund B is sub-advised to the largest private equity investor in infrastructure globally.  Each promoted the following characteristics of their investment strategies:

Consistent Demand.  Infrastructure assets and services are necessary and their use remains consistent even when prices change.

High Barriers to Entry. Because they are capital intensive and often encouraged by governments that must approve the projects, infrastructure investments face little threat of competition.

Stable Cash Flows.  Because they enjoy consistent demand and stable or rising pricing, infrastructure investments offers stable cash flows that are less risky that investments in typical companies.

Inflation Hedge.  The contracts that determine the pricing of infrastructure assets commonly tie prices to changes in aggregate pricing indices such as the Consumer Price Index.  Therefore, as prices rise, cash flow rises.

The promoters of these funds distill the above stated characteristics into the following expected investment outcomes:

a)      Low correlation with equities

b)      High correlation with inflation


The Analysis

Using regression analysis, we explored the extent to which the returns of each of these funds reflected these investment outcomes.  As the dependent variable, we collected rolling one-year total returns for each of the funds and modeled the following independent variables to these return series:

  • Rolling one-year returns of the S&P 500 Index
  • Year-over-year Consumer Price Index for All Urban Consumers
  • Rolling one-year returns of the Barclays Capital Investment Grade Corporate Bond Index

From this analysis, we can estimate and model the degree to which variability of these factors explains the variability of the investment returns of each of the funds.  If the marketing messages of these funds is genuine, then we would expect to see a low correlation and poor fit of the of the S&P 500 returns to the fund returns and a high correlation and strong fit of CPI data to the fund returns.

 

The Results

For the returns of each of the funds, the S&P 500 proved to have significant explanatory power, the CPI had weak explanatory power and the Barclays Aggregate Bond Index had insignificant explanatory power.  The specific results are as follows:

 

Fund A

Fund A is an open-ended mutual fund sponsored, marketed and managed by a U.S. fund company whose municipal bond funds as a group hold the greatest dollar value of infrastructure investments in the world.  We tested the total returns of this fund from December 31, 2007 when the track record begins to May 31, 2012 during which the fund produced an annualized return of -0.24% while the S&P 500 Index produced an annualized return of +0.28%.  The regression model with the most explanatory value for this fund is:

 

Fund A total return = [1.05 * S&P 500 total return] - [1.34 * CPI] + 2.82

 

When the S&P 500 returns were tested against Fund A returns, the correlation coefficient was .95, which indicates very high correlation.  When CPI data were tested, the correlation coefficient was .34 which suggests a relatively weak correlation.  The standard error of the test indicates that the S&P 500 accounts for most of the difference between the model estimate of fund returns and the actual fund returns but that including CPI reduces these difference slightly and therefore plays some modest role in explaining returns.

The model indicates that the best estimate of the total return of Fund A is based on a point-by-point tracking with the S&P 500 and that inflation measured by CPI plays a minor role in determining total returns.

 

Fund B

Fund B is a closed-end fund listed on the NYSE and is sponsored and marketed by a U.S. fund company and sub-advised by a firm that specializes in making private equity investments in infrastructure projects and companies.  We tested the total returns of this fund from April 30, 2004 when its track record begins to May 31, 2012 during which the fund produced an annualized total return of +5.89% compared to the annualized total return of the S&P 500 of +3.92%. The regression model with the most explanatory value is:

 

Fund B total return = [1.38 * S&P 500 total return] + [2.84 * CPI] – 1.81

 

The correlation between Fund B returns and the S&P 500 was .86, a bit lower than for Fund A but still suggests a strong correlation with the stock market.  The correlation between Fund B returns and CPI was .18, a very week correlation.  As with Fund A, the standard error indicates that the S&P 500 accounts for nearly all of the difference between the model estimates of returns and the actual returns but the CPI estimator slightly improved these differences.

 

Conclusions

Based on this analysis, we conclude that each of these listed infrastructure funds fails to deliver what their marketing material suggests – diversification of risk of the equity market and a hedge on the risk of inflation.  Instead, the investment strategies that these two funds employ actually provides direct “beta” exposure to the S&P 500 Index – Fund A on a point-by-point basis and Fund B with a multiplier to the S&P 500 returns.  In addition, investors in these strategies should not expect much in terms of a hedge against inflation.

As for why the strategies of these infrastructure funds so badly misses the investment outcomes that their sponsors advertise, this is a question that we will investigate in a future paper.

 

Appendix: Model Results

 

Regression Statistics


Multiple R

0.979

R Square

0.959

Adjusted R Square

0.957

Standard Error

5.137

Observations

43

 

ANOVA

         

 

df

SS

MS

F

Significance F


Regression

2

24562.77542

12281.39

465.4089

1.98814E-28

Residual

40

1055.535352

26.38838

   

Total

42

25618.31078

 

 

 

 

 

Coefficients

Standard Error

t Stat

P-value


Intercept

2.816

1.169

2.409017

0.020692

S&P 500 Index

1.052

0.043

24.51265

1.04E-25

CPI

-1.341

0.634

-2.11402

0.040796

 

FUND B: SUMMARY OUTPUT

Regression Statistics


Multiple R

0.940

R Square

0.883

Adjusted R Square

0.880

Standard Error

10.297

Observations

86

 

ANOVA

 

df

SS

MS

F

Significance F


Regression

2

66226.81181

33113.41

312.3309

2.37E-39

Residual

83

8799.682323

106.0203

   

Total

85

75026.49413

 

 

 

 

 

Coefficients

Standard Error

t Stat

P-value

Lower 95%

Upper 95%

Lower 95.0%

Upper 95.0%


Intercept

-1.811

2.042

-0.88679

0.377753

-5.87175

2.250422

-5.87175

2.250422

S&P 500 Index

1.376

0.062

22.11266

1.53E-36

1.252295

1.499841

1.252295

1.499841

CPI

2.836

0.711

3.986877

0.000143

1.420955

4.250125

1.420955

4.250125

The Journey of Austerity in Europe

Written By: Xue Han
Luxembourg Garden Fellow, Global Infrastructure Asset Management LLC


Quick Note

Following on Xue Han’s paper that estimated impact of the “multiplier effect” on the U.S. economy of the automatic budget cuts that the U.S. Congress has imposed (Deficit Reductions and Multiplier Effects on the U.S. Economy, January 2012), I asked her to estimate the impact of the austerity programs that several members of the European Community have agreed to implement in order to confirm their commitments to remain members of the EC.  In The Journey of Austerity in Europe, she provides her forecasts for economic growth in each of the countries that make up the bulk of the EC.  In this paper, Xue Han describes and compares the austerity measures of each subject country and applies an estimate of the multiplier to the available forecasts of GDP.  Her research demonstrates that policy decisions to “front load” or “back load” budget cuts or tax increases are critical in determining the paths of economies in the region and provides estimates for the recoveries of these economies.

William M. Fitzgerald


Executive Summary

The task of this research paper is to present factual details of the austerity packages adopted by multiple EU countries and provide theoretical analysis of the controversial outcomes they will bring. This report has six major sections and is structured as follows. Section 1 briefly describes the economic and political situations of the ongoing debt crisis within the EU countries and the external support they received from the International Monetary Fund (IMF) with conditionalities. The second section, after formally defining austerity package and introducing its typical measures and fundamental goals, paints a big picture of the size and timing of the packages for all those countries on which we have information; also included, are detailed demonstrations of the austerity stories of ten countries, namely Germany, France, Italy, Spain, Austria, Greece, Latvia, Portugal and Ireland. These countries are considered as having significant impacts on the overall economy in the EU area and have implemented or planned for austerity programs. This section also provides a map on the composition of each country’s austerity measures, divided between spending cuts and tax hikes.  The results are described in the table below.

Based on certain assumptions, our estimates show that by 2016, austerity programs in these nine countries studied will negatively affect the regional economy to downsize by 3.13%, or 278.7 billion euros in total, compared to the scenario without austerity measures. In a simple average term, that is an average 4.29% lower nominal GDP than originally projection for each country. Yet fortunately, the harm of austerity does not linger long over their economic growth rates to be observed in 2016. On the budgetary side, pictures will become harsher with a total of 160.6 billion euros, or 1.87% increase in government budget deficit in these nine countries by 2016, when most countries are targeting at balanced budgets; in a simple average term, each country will see a budget deficit that is 2.2% higher than what they currently plan for.

With the nine EU countries that are considered crucial to the future directions of EU and the single currency studied, these detailed quantitative medium-term projections could provide precious insights for investors and policy makers worldwide on the complicated situation of the European debt crisis and its austerity progress.

Deficit Reductions and Multiplier Effects on the U.S. Economy

Written By: Xue Han
Luxembourg Garden Fellow, Global Infrastructure Asset Management LLC

Download the pdf of this report
Written: January 30, 2012


A Quick Note

Xue Han's paper investigates how the "multiplier" will magnify the contraction of the U.S. economy due to the automatic budget cuts that the U.S. Congress has imposed. Voters in the U.S. and the leaders that they elect are engaged in a debate on how to allocate the public's resources that are increasingly constrained. Those who argue that investments in infrastructure will be critical to setting the economy on a more favorable trajectory will need to demonstrate that the multiplier for any public investment in infrastructure will multiply at a higher rate than the other alternatives for government expenditure.

-William Fitzgerlad

Preface

This research studies the effects of budget deficits reductions - the Automatic Budget Enforcement Procedures - on the overall economy in the U.S. over the next decade and report its detailed findings in the appended report.

As the Joint Select Committee ("super committee") announced failure in late November last year, the Automatic Budget Enforcement Procedures ("trigger cuts") of $1.2 trillion in total are expected to take place, imposing evenly distributed cuts of roughly $110 billion per year for nine fiscal years starting on January 2013.

Conclusion

The effects of provisions related to the Automatic Budget Enforcement Procedures on the overall economy are considerable. Over the 11-year horizon from 2013 to 2023, the budget cuts will reduce Gross Domestic Product by $1.86 trillion, or -5.2% on average and -7.6% at the most extreme point in 2021, when comparing to projections excluding such effects. The table below demonstrates the forecasted impact both annually and cumulatively during the time horizon of the budget cuts.

Summary of Methodology

Assuming for the Automatic Budget Enforcement Procedures, the report employs the IS/LM model to examine the effects of such fiscal spending cuts on each component of GDP, which enables us to adjust for decrease in government purchase, rather than an increase in most studies, and to account for the interest rate at the "zero-lower bound". The derived formula for estimating the fiscal multiplier is

 

 

With estimated marginal propensity to consume (c=0.85838), marginal propensity to import (n=0.2641) and average tax rate (t=0.1475), the calculated multiplier is 1.88.

With appropriate adjustments, the report takes advantage of the model by the Council of Economic Advisors (Romer-Bernstein model) for studying the lagging effects of the fiscal multiplier. Applying the accumulated multiplier effects of cuts in nine consecutive years featured by the Automatic Budget Enforcement Procedures and their compound impacts to the projections of GDP by the Congressional Budget Office, the report arrives at the adjusted projections as shown in the table below.