Predicting Recessions In Major Texas Metropolitan ...

University of Texas at El PasoDigitalCommons@UTEP

Open Access Theses & Dissertations

2019-01-01

Predicting Recessions In Major Texas MetropolitanEconomies Using Yield Spreads And OtherEconomic IndicatorsAaron Dodson NazarianUniversity of Texas at El Paso, [email protected]

Follow this and additional works at: https://digitalcommons.utep.edu/open_etdPart of the Economics Commons

This is brought to you for free and open access by DigitalCommons@UTEP. It has been accepted for inclusion in Open Access Theses & Dissertationsby an authorized administrator of DigitalCommons@UTEP. For more information, please contact [email protected].

Recommended CitationNazarian, Aaron Dodson, "Predicting Recessions In Major Texas Metropolitan Economies Using Yield Spreads And Other EconomicIndicators" (2019). Open Access Theses & Dissertations. 2004.https://digitalcommons.utep.edu/open_etd/2004

El Paso Water, City of El Paso Office of Management & Budget,

National Science Foundation Grant DRL-1740695, UTEP Center for the Study of Western

Hemispheric Trade, and Hunt Institute for Global Competitiveness at UTEP.

The yield spread has been found to serve as a valuable economic forecasting tool. This

research employs dynamic autoregressive probit downturn models using the United States yield

spread and other regional and macroeconomic variables. This study then inspects the predictive

power of the United States yield spread on the five largest urban economies in Texas, the four

largest metropolitan areas along the Texas-Mexico border, as well as the Texas state economy.

The other regional and macroeconomic variables are included in model specifications based on

characteristics of the economies being analyzed. Results indicate that a narrowing of the United

States Yield Spread for either country tends to increase the probability of recessions in all the

economies analyzed. Decreases in the real value of the peso are found to reduce the likelihood of

a recession in border economies and Texas. However, results for west Texas intermediate oil price

are mixed and suggest that for some economies when oil prices increase, the probability of a

recession increases- which is counter to conventional expectations.

Chapter 4: Empirical Analysis

Table 1

Economic recession prediction is an area of interest for public and private decision makers.

For national economies, the yield spread, the difference between long-term and short-term treasury

bills, is a valuable recession forecasting tool (Estrella and Mishkin, 1996; Dueker, 1997). When

the yield spread by itself is compared with other financial variables such as stock prices, and

monetary aggregates, it tends to be the only financial variable that effectively predicts recessions

after one quarter (Estrella and Mishkin, 1998). Research by Nyberg (2010) and Kauppi and

Saikkonen (2008) show that usage of the yield spread within dynamic binary response models

outperforms standard static models in predicting future recessions.

Yield spreads have also been shown to effectively predict economic recessions for state

economies (Gauger and Schuck, 2002; Shoesmith, 2003), but there is relatively little research on

this topic for metropolitan economies. That gap in the literature is somewhat puzzling.

Historically, there is much more information available regarding national and regional economies

than there is for urban economies (Klein, 1969). Given the relative paucity of information

regarding metropolitan economies, the potential predictability of business cycle downturns for

these areas by models with minimal data requirements may provide a very useful tool to

policymakers and business analysts.

While the latter is true, metropolitan business cycle indices (BCIs) are not widely available.

This study takes advantage of previously published BCIs for nine urban economies located within

the state of Texas. Those indices are compiled using a well-known methodology involving

Kalman filtering and dynamic single-factor analysis (Stock and Watson, 1991). The indices are

maintained and updated by the Federal Reserve Bank of Dallas. As coincident indicators, these

BCIs provide gauges of current economic conditions for each of the geographic areas monitored

(FRBD, 2018). This study employs those indices for the five largest urban economies in Texas,

the four largest metropolitan areas along the Texas-Mexico border, as well as the regional BCI

estimated for the Texas state economy.

To examine metropolitan BCI downturn predictability, the study uses yield spreads plus

some other economic indicators that are potentially related to business cycle developments across

Texas. Economic conditions in Mexico have been shown to affect border regions of Texas

(Fullerton et al., 2017). Accordingly, a peso/dollar real exchange rate and a yield spread for

Mexico are also included in the sample. Because energy activities influence economic conditions

in many regions of Texas (Lee, 2015), the sample also includes the prices of West Texas

Intermediate crude oil. Parameter estimation is carried out using a dynamic probit methodology

(Ng, 2012).

Subsequent sections of the paper are as follows. The next section provides a brief overview

of related studies. Section three describes the methodological framework and data employed.

Section four provides an empirical analysis. Section five summarizes principal results and

implications for future research.

Previous research examines what information the term structures for U.S. Treasury bill

interest rates contain about future economic conditions in national economies. Research indicates

that longer-term Treasury bill maturities have significant predictive power for future changes in

inflation (Mishkin 1990). The yield spread, the difference between long-term and short-term

treasury bills, has been found to serve as a valuable recession forecasting tool. The yield spread

tends to outperform other common recession indicators for a period of two to six quarters in the

future (Estrella and Mishkin 1996). In further research, the yield spread by itself tends to be the

only economic variable that reliably predicts recessions after one quarter (Estrella and Mishkin

1998).

Other research explores the ability yield spreads to predict future economic conditions in

developing economies. Gonzalez, Spencer and Walz (2000) determine that Mexican yield spreads

have significant forecasting ability for inflation and real growth. Interestingly, the U.S. and Euro

area yield curves contain information about future inflation and growth in emerging economies.

That especially holds true for countries with currency exchange rates that are pegged to the U.S.

dollar (Mehl 2009). Both studies indicate that the yield curve in emerging economies also contain

information about future inflation and growth.

A substantial volume of recession predictability utilizing yield curves has been conducted

for national economies. A smaller number of studies have been examined this topic for state and

regional economies. One such study finds that yield spreads can forecast multi-state regional

economic downturns, but the effectiveness of recession prediction varies according to regional

economic structures (Gauger and Schuck 2002). Another study successfully modeled recessions

in 34 of the 50 state economies in the United States in statistically reliable manners (Shoesmith

2003).

Forecasting economic conditions in U.S.-Mexico border regions is a unique challenge

because cross-border economic relationships affect metropolitan business cycles (Fullerton 2001).

Those commercial and industrial ties include retail sector “exports,” health sector tourism, as well

as supply chain linked manufacturing, transportation, and warehousing activities (Phillips and

2008). Similarly, energy sector fluctuations are likely to play outsized roles in the business

cycle that characterizes urban economic conditions in places like Houston. Consequently, the

inclusion of variables that reflect those types of considerations may augment the information

provided by yield spreads.

When available, BCIs provide useful means for understanding prevailing states of national,

regional, or metropolitan economies. Stock and Watson (1991) develops a widely used BCI

methodology known as dynamic single-index factor modeling that employs Kalman filters. This

methodology develops BCIs under the assumption that the co-movements of key economic

indicators are influenced by a common underlying, unobservable factor. This approach has been

used to generate BCIs for various geographic regions. Among others, the latter include Texas

(Phillips 2005), Midland-Odessa (Downs and Fullerton 2017), Lubbock (Fullerton and Subia

2017), plus border urban economies in Texas (Phillips and Cañas 2008). Regional BCIs provide

fairly up to date gauges of whether the economies analyzed are expanding or contracting.

A common approach to predicting the onset of economic contractions is to use binary

recession indicators as dependent variables. Various studies indicate that the slope of the yield

curve is the most reliable recession predictor (Dueker 1997). Incorporating lags of the binary

recession indicators in the equation specifications has been found to significantly increase the

predictive power of business downturn probit models (Kauppi and Saikkonen 2008; Nyberg 2010).

To analyze metropolitan BCI downturn predictability, this study utilizes yield spreads from the

U.S. and Mexico, plus other regionally relevant economic variables, with parameter estimation

carried out using a dynamic probit methodology. Dynamic and dynamic autoregressive probit

models have been found to perform well in this context (Ng 2012; Fullerton et al. 2017).

Regional BCI modeling efforts may benefit from the inclusion of other variables that

augment the information contained in the yield spread. For the border metropolitan economies,

Mexican yield spreads and peso/dollar currency exchange rates are likely to help predict BCI

downturns because economic condition in Mexico also affect the business conditions on the north

side of the boundary (Fullerton 2001; Fullerton et al. 2017). Oil prices are a useful indicator for

predicting business cycle downturns in economies with substantial energy activities (Lee 2015).

For example, in the petroleum driven economy of Midland-Odessa, oil price fluctuations tend to

correspond with similar shifts in local BCIs (Downs and Fullerton 2017).

The objective of this study is to develop probit downturn models for the five largest urban

economies in Texas, the four largest metropolitan areas along the Texas-Mexico border, as well as

the Texas state economy. The five largest urban economies in Texas are Austin-Round Rock,

Dallas-Plano-Irving, Fort Worth-Arlington, Houston-The Woodlands-Sugarland, and San

Antonio-New Braunfels. The four largest metropolitan areas along the Texas-Mexico border are

Brownsville-Harlingen, El Paso, Laredo, McAllen-Edinburg-Mission.

Probit analysis is used to quantify the probability of recessions in a particular time period.

This approach has been used to model business cycle contractions in multiple geographies. A

static probit model can be written as follows:

(1) Pr(Yt = 1) = F(β0 + β1Xt-k)

In Equation (1), Pr is the probability of an existing recession (Yt = 1 if a recession is underway at

time t, 0 if not), Xt−k is an explanatory variable at time t−k, β0 and β1 are parameters to be estimated,

and F represents the cumulative normal distribution function.

One drawback of the static model is that it does not take advantage of autocorrelated

information potentially embedded within the binary recession indicator. In such cases, dynamic

probit model specifications take into account prior states of the economy by including a lag of the

dependent variable as shown in Equation (2).

(2) Pr(Yt = 1) = F(β0 + β1Xt-k + β2Yt-m)

Dueker (1997) argues that the dynamic version of the probit model is better suited to

handling problems such as serial correlation that frequently arise in the context of time-series

modelling. Along those lines, Kauppi and Saikkonen (2008) find that dynamic probit models tend

to outperform static specifications for predicting national economic downturns in the United

States. The model in Equation (2) can be further augmented by introducing additional explanatory

variables. Standard selection criteria such as pseudo-R2 statistics can be used to identify which

lags of candidate explanatory variables to include in an equation (Nyberg 2010).

To help select an estimated equation functional form, the pseudo-R2 metric developed by

Estrella (1998) is employed. The metric is calculated as shown below.

(3) Adjusted Pseudo - R2 = 1 – (𝐿𝑢

𝐿𝑐)−(

2𝑛)𝐿𝑐

In Equation (3), Lu is the unconstrained maximum value of the log-likelihood, Lc is the

constrained maximum value of the log-likelihood assuming all coefficients except the constant are

zero, and n is the sample size. Standard diagnostic statistics such the t-statistic are also utilized.

The modelling framework employed in this study analyzes probabilities of BCI downturns

for selected urban economies located in Texas as functions of yield-spreads as well as other

regional and macroeconomic variables. This study employs business cycle indices for the five

largest economies in Texas, the four largest metropolitan areas along the Texas-Mexico border, as

well as a regional BCI estimated for the Texas state economy. The other regional and

macroeconomic variables are included based on characteristics of the economies being analyzed.

The five largest economies in Texas all engage in energy activities or are greatly affected

by energy prices (FRBD 2014). As noted above, oil prices can help predict business cycle

fluctuations in economies with substantial energy activities (Lee 2015). Accordingly, West Texas

Intermediate oil prices are included as part of the sample data collected for those five urban

economies.

In the four largest metropolitan economies along the Texas-Mexico border, this study

utilizes a framework similar to that outlined by Fullerton (2001). In that study, border region

economic performance is modelled as a function of both national and international variables.

Subsequent studies have confirmed that the peso/dollar exchange rate strongly influences business

activity along the border (Patrick and Renforth 1996; Coronado and Phillips 2007; Niño et al.

2015). Yield spreads for the United States and Mexico are also included in the specifications for

each of these border economies.

The dichotomous dependent variables identify downturns in each metropolitan economy.

According to Klein and Moore (1983), the binary variables are constructed using monthly

frequency regional business cycle index values. In all nine economies, the binary dependent

variable is defined by shifts in the business cycles indices. If there is a recession, the binary

dependent variable for that specific month is equal to one. If there is not a recession, this variable

is equal to zero. For purposes of this study, a recession is defined as six consecutive months (or

more) of negative growth in a business cycle index. An economic contraction ends after six

consecutive months of positive growth in a business cycle index.

The United States yield spread is calculated as the 10-year Treasury bond rate minus the

3-month Treasury bill rate. All United States interest rate data are from the Federal Reserve Bank

of St. Louis (FRED 2015). The yield spread of Mexico is calculated as the 1-year Treasury bill

rate minus the 28-day Treasury bill rate (CETES). All Mexican interest rate data are from the

central bank of Mexico (BM 2018a). This study utilizes the above Mexican yield spread and a

peso/dollar (MXN/USD) real exchange rate index because economic conditions in Mexico

sometimes have pronounced impacts on the business cycles of the United States border cities (BM

2018b; Phillips and Cañas 2008). These international economic variables are important for this

research because the cities selected for this study and their cross-border counterparts in Mexico

share a variety of commercial and industrial linkages. The dependent variable takes a lag of one

in order to capture potential autocorrelation structures of the dependent variables (Ng 2012).

Additionally, experimentation is also conducted with an alternate lag structure of three months that

Dueker (1997) posits as the minimum recognition lag time for recessions.

Three different specifications employing the dynamic probit framework are proposed.

Equation (4) is used for the five largest urban economies in Texas. Equation (5) is employed for

the four border metropolitan economies. Equation (6) is utilized for the Texas state business cycle.

(4) Pr(Yt = 1) = F(β0 + β1USSPt-k + β2WTIt-h + β3Yt-m + εt)

(5) Pr(Yt = 1) = F(β0 + β1USSPt-k + β2MXSPt-h + β3REXRt-i + β4Yt-m + εt)

(6) Pr(Yt = 1) = F(β0 + β1USSPt-k + β2MXSPt-h + β3REXRt-i + β4WTIt-j + β5Yt-m + εt)

Table 1

Variable Name Description Hypothesized Coeff. Sign USSP USA Yield Spread (-)

WTI West Texas Intermediate Oil Price, $/bbl (-)

Y Business Cycle Recession Indicator (+)

MXSP Mexico Yield Spread (-)

REXR Real Peso per Dollar Exchange Rate Index (+ or -)

Table 1 summarizes the hypothesized relationships between the recession indicator, Y, and

each of the explanatory variables. In Equations (4) through (6), USSP is the United States yield

spread, WTI is the monthly West Texas Intermediate Crude Oil Spot Price in dollars per barrel,

Yt−m is a the binary dependent variable with a lag of m months, MXSP is the yield spread for

Mexico, and REXR is the inflation adjusted peso/dollar exchange rate index. The corresponding

model is estimated for each of the metropolitan economies mentioned above in the previous

section.

Equation (4) is used to examine whether the yield spreads and the spot prices of WTI oil

can help predict recessions in the five largest Texas metropolitan economies. A decrease in the

United States yield spread, which results from higher short-term interest rates and/or lower long-

term rates, is hypothesized to increase the probability that a recession will occur in future quarters

(this is also posited for Equations (5) and (6)). That is because high short-term interest rates are

often associated with contractionary monetary policy and lower long-term rates may reflect

expectations of an economic slowdown in coming years (Dueker 1997). A decrease in the spot

prices of WTI oil is hypothesized to increase the probability a recession will occur in future

quarters. That is because low oil prices dampen growth within the energy sector which hurts the

Texas economy as a whole. A value of 1 in the binary dependent variable is hypothesized to be

associated with an increased probability that a recession will occur in future quarters.

Equation (5) is used to examine whether the yield spreads and the real exchange rate index

can help predict recessions in the four largest metropolitan economies along the Texas-Mexico

border economies comprised in the sub-sample. For similar reasons to the United States yield

spread, the yield spread for Mexico is expected to have an inverse relationship with the probability

of recession. Economic slowdowns in Mexico may coincide with downturns in cities on the north

side of the border for a variety of reasons. First, retail sectors in many United States border cities

rely on a steady influx of Mexican shoppers. Those sales tend to decline when such shoppers

reduce consumption, as typically occurs when Mexico faces a recession (Coronado and Phillips

2007; Phillips and Cañas 2008). Other border region economic sectors such as freight

transportation, wholesale trade, and financial services conduct business with manufacturers

located in Mexico (Cañas et al. 2013). Thus, a higher probability of recession in Mexico, as

signaled by a flattening or inversion of that country’s yield curve, is hypothesized to increase the

probability of recession on the north side of the border.

The impacts of real exchange rate on border city economies is ambiguous. Some prior

research suggests that peso depreciations can have strong adverse impacts on retail sectors in the

United States border cities (Patrick and Renforth 1996). However, peso depreciations also tend to

stimulate export-processing activity in Mexican border cities, which may help fuel economic

activity on the north side of the border (Niño et al. 2015). If a real depreciation of the peso lowers

the probability of recession for any of the border economies examined, then the exchange rate

coefficients will be negative. The converse will occur if peso weakness increases the likelihood

of a business cycle downturn.

Equation (6) is used to examine whether the yield spreads, the real exchange rate index,

and the spot prices of WTI oil can help predict recessions for the Texas state economy. The

impacts of fluctuations in these variables on the metropolitan economies in the sample are

discussed above. Equation (6) reflects many aspects of the modern Texas economy, regionally,

nationally, and internationally.

Chapter 4: Empirical Analysis

Equations with varying specifications are estimated for each economy. The final

specifications are selected by taking into consideration pseudo-R2 values, lag length information

criteria, coefficient statistical significance, and other statistical diagnostic tools (Gauger and

Schunk 2002; Nyberg 2010). Initial dependent variable lag specifications of one month are sub-

optimal when compared to the alternate dynamic lag specification of three months. Results for the

initial dynamic lag specification of one month are published in Appendix Tables 7.A through 8.B.

Sample data employed are from January 1991 to May 2018, a span of about 27 years. Primary

In general, Equations 4 through 6, outlined in the previous section, deliver favorable

estimation results. Given the geographic location of Laredo on the Eagle Ford shale formation, an

alternate model is specified by including the West Texas intermediate oil price as an explanatory

variable. Fort Worth, Laredo, and the Texas economies are the only economies that deviate from

the general equation specifications outlined in the previous section. The coefficient sign for West

Texas intermediate oil prices in Fort Worth and Texas were positive which runs counter to

conventional wisdom. It is, therefore, removed from the model specification for those economies.

A positive coefficient for West Texas intermediate oil prices also results when it is included

in equations estimated for Laredo. The version summarized in Table 2 exhibits much better

statistical traits than other specifications, as well as more realistic coefficients for the other

regressors. Alternative specification outcomes for all of the regions analyzed are included in the

appendix materials.

El Paso Laredo McAllen Brownsville

Coefficient 4.4528** -0.2697 1.5702 3.8379**

USSP -3.1534*** -0.3832*** -1.1136*** -1.8656***

MXSP -0.1853*** -0.2122*** -0.8429*** -0.9804***

REX -0.0118** -0.0015 -0.0065* -0.0104***

WTI 0.0098***

Yt-3 3.3155*** 1.9310*** 6.5922*** 2.3203***

Akaike Inf. Crit. 0.1972 0.7265 0.1981 0.183

Hannan Quinn Crit. 0.2218 0.7556 0.2226 0.2075

Schwartz Inf. Crit. 0.2586 0.7993 0.2594 0.2444

Log-likelihood -24.779 -105.52 -25.008 -22.628

Restr. Log-likelihood -119.95 -193.75 -169.56 -99.921

Total Obs. 302 307 303 302

Pseudo R-squared 0.7143 0.5356 0.8826 0.6257

The sample period analyzed is January 1991 to May 2018.

* Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

El Paso Laredo McAllen Brownsville

USSP -26 -21 -25 -26

MXSP -6 0 0 0

REX -8 -8 -8 -5

WTI -20

Yt-3 -3 -3 -3 -3

Austin Dallas Fort Worth

Coefficient 0.3581 0.6132 -1.5565***

USSP -2.3122*** -2.5788*** -3.9404***

MXSP

REX

WTI -0.0169** -0.0178**

Yt-3 3.6108*** 3.3606*** 9.0217***

Akaike Inf. Crit. 0.1882 0.1947 0.0946

Hannan Quinn Crit. 0.2077 0.2144 0.1091

Schwartz Inf. Crit. 0.2369 0.244 0.1309

Log-likelihood -24.694 -25.307 -11.569

Restr. Log-likelihood -122.24 -126.98 -93.828

Total Obs 305 301 308

Pseudo R-squared 0.7225 0.7435 0.7206

The sample period analyzed is January 1991 to May 2018.

* Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

Austin Dallas Fort Worth

USSP -23 -27 -20

MXSP

REX

WTI -17 -17

Yt-3 -3 -3 -3

Houston San Antonio Texas

Coefficient -1.0262** -0.1913 -1.1515***

USSP -2.9101*** -1.9445*** -0.8373***

MXSP -0.0967**

REX

WTI -0.0091 -0.0170*

Yt-3 10.397*** 7.2924*** 3.2774***

Akaike Inf. Crit. 0.1386 0.197 0.2237

Hannan Quinn Crit. 0.158 0.2164 0.2427

Schwartz Inf. Crit. 0.1872 0.2456 0.2713

Log-likelihood -17.278 -26.238 -31.231

Restr. Log-likelihood -114.94 -120.68 -82.332

Total Obs. 307 307 315

Pseudo R-squared 0.758 0.6987 0.3975

The sample period analyzed is January 1991 to May 2018.

* Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

Houston San Antonio Texas

USSP -21 -15 -13

MXSP -6

REX

WTI -21 -21

Yt-3 -3 -3 -3

1-

year Treasury bill rate minus the 28-day Treasury bill rate (CETES). That varies substantially

from the USSP measure calculated as the difference between the 10-year Treasury bond rate and

the 3-month Treasury bill rate. The lead times for MXSP align closely with those reported in other

studies (Fullerton 2017; and Reyna-Cerecero et al. 2008).

The results in Table 2 indicate that real depreciation of the peso against the dollar decreases

the probability of a recession in all four of the border economies. As previously stated, peso

depreciations tend to stimulate export-processing activity in northern border municipalities in

Mexico (Coronado et al. 2004; Fullerton et al. 2004; Canas et al. 2007; Niño et al. 2015). That

generally leads to increased economic activity in the adjacent metropolitan areas on the northern

side of the international boundary (Hanson 1996; Varella-Mollick et al. 2006; Canas et al. 2013).

The negative REX coefficients provide additional evidence along those same lines.

Nearly all of the West Texas intermediate spot oil price parameter estimates for the

metropolitan economies are negative. The exception is the WTI coefficient estimated for Laredo.

As stated at the beginning of this section, oil prices are included in the specification of this border

economy because of its presence on the Eagle Ford Shale formation. The positive parameter is

puzzling. As will be discussed below, Laredo has a very high concentration of employment in

transportation and warehousing. Across the border, and closely linked to that segment of the

Laredo metropolitan economy, are large manufacturing sectors in both Monterrey and Nuevo

Laredo. Transportation and manufacturing are energy intensive sectors and that may be what leads

to the positive correlation between oil price hikes and recessions in the former Rio Grande

Republic. As more data become available, additional research appears warranted.

Prior research indicates that the yield spread has relatively high predictive power in regions

that are heavily dominated by interest-sensitive employment sectors such as manufacturing and

construction (Gauger and Schunk 2002). Shoesmith (2003) confirms that the use of interest-rate

spreads to predict future state recessions is more effective in manufacturing-intensive states.

Fullerton et al. (2017) also finds evidence that the yield spread helps predict employment losses in

southern border economies in the United States. Those regions typically have relatively high

concentrations of interest-rate sensitive sectors such as construction, manufacturing,

transportation, and warehousing.

In order to see if this pattern holds true for the economies analyzed in this study, formal

log-likelihood ratio test outcomes are juxtaposed against corresponding percentages of

employment in construction, manufacturing, and transportation. To calculate the log-likelihood

statistic, restricted versions of Equations (4) through (6) are specified without either of the yield

spread variables. Then, using the unrestricted log-likelihoods displayed in Tables 2, 4, and 6 plus

the restricted equation log-likelihoods, log likelihood ratio test statistics are calculated. Table 8

reports those results. When the log-likelihood ratio statistic is larger than the corresponding chi-

squared critical value, the null hypothesis is rejected and model performance is improved by the

inclusion of the yield spread variables. If the pattern described in the preceding paragraph is true,

economies with the largest market shares of the interest-rate sensitive jobs should correspond with

the largest log likelihood ratio test statistics.

Industry Share of Total Employment*

Construction and

Manufacturing

Construction, Manufacturing,

Transportation and

Warehousing

Log Likelihood Ratio

Test Statistic**

El Paso 12.2% 17% 70.925

Laredo 6.1% 19.7% 20.668

McAllen 9.4% 13.2% 45.274

Brownsville 9.8% 13.7% 66.297

Austin 12.1% 14% 40.133

Dallas 12.9% 16.5% 40.373

Fort Worth 15% 21.5% 63.945

Houston 15% 19.3% 55.093

San Antonio 11.1% 13.8% 37.677

Texas 14.1% 17.8% 22.437

* Employment data are for the period 2001–2017; the data are from the United States Bureau of Economic Analysis.

** The 5% critical value of the chi-squared distribution is 5.99; the null hypothesis is that excluding the yield spread

variables does not significantly reduce the fit of the model.

0.0

0.2

0.4

0.6

0.8

1.0

19

93

19

94

19

95

19

96

19

97

19

98

19

99

20

00

20

01

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

20

16

20

17

20

18

Predicted Probability of Recession

Pro

ba

bilit

y

Figure 1: El Paso (MSA) Actual and Fitted Values

0.0

0.2

0.4

0.6

0.8

1.0

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recession

Pro

babil

ity

Figure 2: Laredo (MSA) Actual and Fitted Values

0.0

0.2

0.4

0.6

0.8

1.0

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recession

Pro

babil

ity

Figure 3: Mcallen-Edinburg-Mission (MSA) Actual and Fitted Values

0.0

0.2

0.4

0.6

0.8

1.0

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recession

Pro

babil

ity

Figure 4: Brownsville–Harlingen (MSA) Actual and Fitted Values

0.0

0.2

0.4

0.6

0.8

1.0

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recession

Figure 5: Austin-Round Rock (MSA) Actual and Fitted Values

Pro

babil

ity

0.0

0.2

0.4

0.6

0.8

1.0

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recessions

Figure 6: Dallas-Plano-Irving (Metropolitan Division) Actual and Fitted Values

Pro

babil

ity

0.0

0.2

0.4

0.6

0.8

1.0

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recession

Figure 7: Fort Worth Arlington (Metropolitan Division) Actual and Fitted Values

Pro

babil

ity

0.0

0.2

0.4

0.6

0.8

1.0

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recession

Pro

ba

bilit

y

Figure 8: Houston-The Woodlands-Sugar Land (MSA) Actual and Fitted Values

0.0

0.2

0.4

0.6

0.8

1.0

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recessions

Figure 9: San Antonio-New Braunfels (MSA) Actual and Fitted Values

Pro

babil

ity

0.0

0.2

0.4

0.6

0.8

1.0

199

2

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

Predicted Probability of Recession

Pro

babil

ity

Figure 10: Texas (State) Actual and Fitted Values

The dynamic probit models estimated for each Texas economy in the sample provide

generally adequate information regarding the probability of business cycle downturns. In most

cases, when the predicted probabilities exceed 0.5 (50 percent), recessions occur. In contrast, the

predicted probabilities tend to stay below 0.5 when the economies are growing. The recession

forecasts do have some missteps. In 2000, two of the equation simulations generated false signals

for slumps that never materialized. Those false signals are for the Brownsville border economy

and the Dallas metropolitan economy.

Regional and metropolitan business cycle indices are not widely available, but several

variables are estimated for Texas and nine of the largest urban economies in that state. Modeling

and predicting recessions in Texas, however, offers a special challenge. The state is so large that

the various urban economies located in Texas exhibit unique business cycle idiosyncrasies. Given

the importance of international trade in Texas, many urban economies in the state are affected by

domestic and international factors. This study attempts to allow for these factors using domestic

and foreign yield spreads, a real peso per dollar exchange rate, West Texas Intermediate oil prices,

and a dynamic lag variable. To date, there have been few economic downturn prediction efforts

conducted for metropolitan economies. Among the regional business cycle studies that have been

completed, yield spreads have been shown to reliably predict recessions.

The United States yield spread is also found to predict slumps in all of the economies

analyzed in this study. Confirming relatively important cross-border industrial and commercial

linkages, Mexico yield spread is found to help anticipate economic contractions for all four border

economies as well as for the Texas state economy. Unexpectedly, the real peso per dollar exchange

rate index is a reliable predictor of business cycle downturns for only two of the four border

economies. Somewhat surprisingly, West Texas Intermediate oil price declines help predict

economic slumps for four of the large urban economies, but not for the state as a whole. Finally,

the 3-month dynamic lag specification of three months perform better than the 1-month dynamic

lag specification, offering regional evidence that three months is the minimum recognition lag time

for recessions. In-sample simulations indicate that the estimated models exhibit good predictive

behavior with only minimal false signal emissions.

Future research may benefit from more Mexico yield spread data. Although Mexico does

have a term structure, yields on government bonds with maturities of longer than three years only

date from 2000 forward. As more yield spread observations become available, that may contribute

better information regarding the onset of regional downturns in Texas. More broadly, metropolitan

business cycle index estimation has fairly minimal data requirements. These indices provide

useful information to policymakers and business analysts. As the procedure is extended to analyze

more regions, further research on business cycle predictability will become feasible. That should

allow clarifying the conditions under which apparent coefficient sign anomalies occur.

E.D. Saenz-Rojo, and A.G. Walke. (2017). Yield Spreads, Currency

Movements, and Recession Predictability for Southern Border Economies in the United

States. Applied Economics 49 (30): 2910-2921, doi: 10.1080/00036846.2016.1251556.

Fullerton, T.M., Jr., and Subia, M.Z. (2017). Metropolitan Business Cycle Analysis for Lubbock.

Journal of Economics & Political Economy, 4 (1), 33-52.

Gauger, J., and D. Schunk (2002). Predicting Regional Recessions via the Yield Spread. Review

Klein, L.R. (1969). The Specification of Regional Econometric Models. Papers of the Regional

Science Association 23 (1): 105–115.

Klein, P. A., and G. H. Moore (1983). The Leading Indicator Approach to Economic Forecasting—

Retrospect and Prospect. Journal of Forecasting 2 (2): 119–135. doi:10.1002/(ISSN)1099-

131X.

Lee, J. (2015). The Regional Economic Impact of Oil and Gas Extraction in Texas. Energy Policy

87 (SI): 60-71, doi: 10.1016/j.enpol.2015.08.032.

Mehl, A (2009). The Yield Curve as a Predictor and Emerging Economies. Open Economies

Review 20 (5): 683–716. doi:10.1007/s11079-007-9077-x.

Mishkin, F. S. (1990). The Information in the Longer Maturity Term Structure about Future

Inflation. Quarterly Journal of Economics 105 (3): 815–828. doi:10.2307/2937901.

Ng, E. C. (2012). Forecasting US Recessions with various Risk Factors and Dynamic Probit

Models. Journal of Macroeconomics 34 (1): 112–125. doi:10.1016/j.jmacro.2011.11.001.

Table 9.A Estimation Results for Texas Border Metropolitan Economies and Texas State El Paso Laredo McAllen Brownsville Texas

Coefficient 3.128 -

1.1918*** 1.3029 2.5793 -1.7608***

USSP -2.4675*** -0.285** -0.9424* -1.3707*** -0.5981**

MXSP -0.1448** -0.0819 -0.7792*** -0.7841*** -0.0789

REX -0.0103* -0.0015 -0.0077 -0.0091**

WTI 0.0113**

Yt-1 3.9796*** 3.2662*** 7.3176*** 3.2411*** 4.0355***

AIC 0.1257 0.3728 0.1155 0.1299 0.1224

HQ 0.1503 0.4019 0.14 0.1544 0.1415

SIC 0.1871 0.4456 0.1767 0.1913 0.1701

Log-likelihood -13.9798 -51.2245 -12.4915 -14.6081 -15.2799

R. Log-likelihood -119.9527 -193.7524 -169.5586 -99.9212 -82.3319

Total Obs 302 307 303 302 315

Pseudo R2 0.8187 0.8135 0.946 0.7198 0.5854

The sample period analyzed is January 1991 to May 2018.

* Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

Table 9.B Lag Selection for Texas Border Metropolitan Economies and Texas State El Paso Laredo McAllen Brownsville Texas

USSP -26 -21 -25 -26 -13

MXSP -6 0 0 0 -6

REX -8 -8 -8 -5

WTI -13

Yt -1 -1 -1 -1 -1

Table 10.A Estimation Results for the Largest Texas Metropolitan Economies

Austin Dallas Fort Worth Houston San Antonio

Coefficient 0.0338 0.9222 -1.6864*** -1.0849 1.4225

USSP -2.778** -4.7252* -2.6649** -8.1775** -5.7005**

MXSP

REX

WTI -0.0209* -0.0327* -0.0582* -0.075**

Yt-1 5.0642*** 6.2575*** 7.1665*** 30.3601** 20.5224**

AIC 0.1016 0.0988 0.0734 0.0634 0.0863

HQ 0.1211 0.1186 0.0879 0.0828 0.1057

SIC 0.1503 0.1481 0.1097 0.112 0.1349

Log-likelihood -11.4862 -10.8758 -8.3046 -5.73 -9.2491

R. Log-likelihood -122.2365 -126.975 -93.8279 -114.9366 -120.6758

Total Obs 305 301 308 307 307

Pseudo R2 0.8498 0.8742 0.7717 0.8941 0.8673

The sample period analyzed is January 1991 to May 2018.

* Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

Table 10.B Lag Selection for the Largest Texas Metropolitan Economies

Austin Dallas Fort Worth Houston San Antonio

USSP -23 -27 -20 -21 -15

MXSP

REX

WTI -17 -17 -21 -21

Yt -1 -1 -1 -1 -1

Table 11.A Alternate Estimation Results for Fort Worth and Laredo Economies

Dynamic lag of 3 Dynamic Lag of 1 Fort Worth Laredo Fort Worth Laredo

Coefficient -2.8965*** -0.3032 -6.3108** -1.0602**

USSP -4.8876*** -0.3403*** -8.6295** -0.2506**

MXSP -0.2613*** -0.146

REX -0.0002 -0.0004

WTI 0.0207* 0.0415*

Yt-p 11.7088*** 1.9616*** 21.4024** 3.204***

AIC 0.1048 0.7409 0.0607 0.3799

HQ 0.1241 0.7652 0.08 0.4041

SIC 0.1531 0.8016 0.1090 0.4405

Log-likelihood -12.1931 -108.7299 -5.3787 -53.3064

R. Log-likelihood -93.9231 -193.7524 -93.9231 -193.7524

Total Obs 309 307 309 307

Pseudo R2 0.7109 0.5177 0.8242 0.8039

The sample period analyzed is January 1991 to May 2018.

* Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

Table 11.B Alternate Lag Selection for Fort Worth and laredo Economies Fortworth Laredo Fortworth Laredo

USSP -19 -21 -19 -21

MXSP 0 0

REX -9 -9

WTI -15 -15

Yt -3 -3 -1 -1

Table 12.A Estimation Results for Texas State (Dynamic Lag of Three Months)

Coefficient 0.2006 -2.2964*** 0.2924 -2.8605*** -1.27*** -2.5912***

USSP -1.1213*** -1.0087*** -1.058*** -1.1471*** -0.753*** -0.9248***

MXSP -0.1017 -0.1179*

REX -0.0106** -0.0117** 0.0036*

WTI 0.0459*** 0.0232*** 0.0501*** 0.0308***

Yt-3 3.9775*** 2.7707*** 3.8004*** 3.4979*** 3.2561*** 3.1783***

AIC 0.1752 0.2 0.1699 0.1774 0.2266 0.2251

HQ 0.2037 0.2239 0.1936 0.1964 0.2408 0.2442

SIC 0.2465 0.2597 0.2293 0.225 0.2622 0.2728

Log-

likelihood -21.6771 -13.3454 -21.8396 -24.0303 -32.7961 -31.4554

R. Log-

likelihood -82.4076 -82.256 -82.4076 -82.4076 -82.4076 -82.3319

Total Obs 316 314 316 316 316 315

Pseudo R2 0.5017 0.6144 0.4997 0.4742 0.3816 0.3953

The sample period analyzed is January 1991 to May 2018. * Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

Table 12.B Lag Selection for Texas State (Dynamic Lag of Three Months)

USSP -14 -14 -12 -12 -12 -13

MXSP -6 -6

REX -4 -2 0

WTI -12 -13 -12 -12

Yt -3 -3 -3 -3 -3 -3

Table 13.A Estimation Results for Texas State (Dynamic Lag of One Months)

Coefficient -1.1289 -2.7582*** -0.6726 -3.0442***

USSP -0.745** -0.9168** -0.789*** -0.8829***

MXSP -0.0875 -0.1222

REX -0.006 -0.009

WTI 0.03* 0.0211* 0.0401** 0.0242*

Yt-3 3.7704*** 3.6914*** 4.2772*** 4.1139***

AIC 0.1199 0.1169 0.1054 0.1034

HQ 0.1485 0.1407 0.1291 0.1224

SIC 0.1915 0.1766 0.1648 0.1509

Log-likelihood -12.8197 -13.3454 -11.6446 -12.3308

R. Log-likelihood -82.256 -82.256 -82.4076 -82.4076

Total Obs 314 314 316 316

Pseudo R2 0.6224 0.6144 0.6396 0.6287

The sample period analyzed is January 1991 to May 2018.

* Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

Table 13.B Lag Selection for Texas State (Dynamic Lag of One Month)

USSP -14 -14 -12 -12

MXSP -6 -6

REX -4 -2

WTI -12 -13 -12 -12

Yt -1 -1 -1 -1

Table 14.A Estimation Results for Texas State (Dynamic Lag of One Months)

Coefficient -1.799*** -2.5754** -1.7608***

USSP -0.6241** -0.7363** -0.5981**

MXSP -0.0789

REX 0.0022

WTI

Yt-3 4.2148*** 4.2661*** 4.0355***

AIC 0.1148 0.1195 0.1224

HQ 0.1291 0.1385 0.1415

SIC 0.1505 0.167 0.1701

Log-likelihood -15.1369 -14.8791 -15.2799

R. Log-likelihood -82.4076 -82.4076 -82.3319

Total Obs 316 316 315

Pseudo R2 0.5868 0.5905 0.5854

The sample period analyzed is January 1991 to May 2018.

* Statistically significant at 10%.

** Statistically significant at 5%.

*** Statistically significant at 1%.

Table 14.B Lag Selection for Texas State (Dynamic Lag of One Month)

USSP -12 -12 -13

MXSP -6

REX -3

WTI

Yt -1 -1 -1

Date Texas Austin Brownsville Dallas El Paso Fort

Worth Houston

Jan-91 0 0 0 0 0 0 0

Feb-91 0 0 0 0 0 0 0

Mar-91 0 0 0 0 0 0 0

Apr-91 0 0 0 0 0 0 0

May-91 0 0 0 0 0 0 0

Jun-91 0 0 0 0 0 0 0

Jul-91 0 0 0 0 0 0 0

Aug-91 0 0 0 0 0 0 0

Sep-91 0 0 0 0 0 0 0

Oct-91 0 0 0 0 0 0 0

Nov-91 0 0 0 0 0 0 0

Dec-91 0 0 0 0 0 0 0

Jan-92 0 0 0 0 0 0 0

Feb-92 0 0 0 0 0 0 0

Mar-92 0 0 0 0 0 0 0

Apr-92 0 0 0 0 0 0 0

May-92 0 0 0 0 0 0 0

Jun-92 0 0 0 0 0 0 0

Jul-92 0 0 0 0 0 0 0

Aug-92 0 0 0 0 0 0 0

Sep-92 0 0 0 0 0 0 0

Oct-92 0 0 0 0 0 0 0

Nov-92 0 0 0 0 0 0 0

Dec-92 0 0 0 0 0 0 0

Jan-93 0 0 0 0 0 0 0

Feb-93 0 0 0 0 0 0 0

Mar-93 0 0 0 0 0 0 0

Apr-93 0 0 0 0 0 0 0

May-93 0 0 0 0 0 0 0

Jun-93 0 0 0 0 0 0 0

Jul-93 0 0 0 0 0 0 0

Aug-93 0 0 0 0 0 0 0

Sep-93 0 0 0 0 0 0 0

Oct-93 0 0 0 0 0 0 0

Nov-93 0 0 0 0 0 0 0

Dec-93 0 0 0 0 0 0 0

Jan-94 0 0 0 0 0 0 0

Feb-94 0 0 0 0 0 0 0

Mar-94 0 0 0 0 0 0 0

Apr-94 0 0 0 0 0 0 0

May-94 0 0 0 0 0 0 0

Jun-94 0 0 0 0 0 0 0

Jul-94 0 0 0 0 0 0 0

Aug-94 0 0 0 0 0 0 0

Sep-94 0 0 0 0 0 0 0

Oct-94 0 0 0 0 0 0 0

Nov-94 0 0 0 0 0 0 0

Dec-94 0 0 0 0 0 0 0

Jan-95 0 0 1 0 0 0 0

Feb-95 0 0 1 0 0 0 0

Mar-95 0 0 1 0 0 0 0

Apr-95 0 0 1 0 0 0 0

May-95 0 0 1 0 0 0 0

Jun-95 0 0 1 0 0 0 0

Jul-95 0 0 1 0 0 0 0

Aug-95 0 0 0 0 0 0 0

Sep-95 0 0 0 0 0 0 0

Oct-95 0 0 0 0 0 0 0

Nov-95 0 0 0 0 0 0 0

Dec-95 0 0 0 0 0 0 0

Jan-96 0 0 0 0 0 0 0

Feb-96 0 0 0 0 0 0 0

Mar-96 0 0 0 0 0 0 0

Apr-96 0 0 0 0 0 0 0

May-96 0 0 0 0 0 0 0

Jun-96 0 0 0 0 0 0 0

Jul-96 0 0 0 0 0 0 0

Aug-96 0 0 0 0 0 0 0

Sep-96 0 0 0 0 0 0 0

Oct-96 0 0 0 0 0 0 0

Nov-96 0 0 0 0 0 0 0

Dec-96 0 0 0 0 0 0 0

Jan-97 0 0 0 0 0 0 0

Feb-97 0 0 0 0 0 0 0

Mar-97 0 0 0 0 0 0 0

Apr-97 0 0 0 0 0 0 0

May-97 0 0 0 0 0 0 0

Jun-97 0 0 0 0 0 0 0

Jul-97 0 0 0 0 0 0 0

Aug-97 0 0 0 0 0 0 0

Sep-97 0 0 0 0 0 0 0

Oct-97 0 0 0 0 0 0 0

Nov-97 0 0 0 0 0 0 0

Dec-97 0 0 0 0 0 0 0

Jan-98 0 0 0 0 0 0 0

Feb-98 0 0 0 0 0 0 0

Mar-98 0 0 0 0 0 0 0

Apr-98 0 0 0 0 0 0 0

May-98 0 0 0 0 0 0 0

Jun-98 0 0 0 0 0 0 0

Jul-98 0 0 0 0 0 0 0

Aug-98 0 0 0 0 0 0 0

Sep-98 0 0 0 0 0 0 0

Oct-98 0 0 0 0 0 0 0

Nov-98 0 0 0 0 0 0 0

Dec-98 0 0 0 0 0 0 0

Jan-99 0 0 0 0 0 0 0

Feb-99 0 0 0 0 0 0 0

Mar-99 0 0 0 0 0 0 0

Apr-99 0 0 0 0 0 0 0

May-99 0 0 0 0 0 0 0

Jun-99 0 0 0 0 0 0 0

Jul-99 0 0 0 0 0 0 0

Aug-99 0 0 0 0 0 0 0

Sep-99 0 0 0 0 0 0 0

Oct-99 0 0 0 0 0 0 0

Nov-99 0 0 0 0 0 0 0

Dec-99 0 0 0 0 0 0 0

Jan-00 0 0 0 0 0 0 0

Feb-00 0 0 0 0 0 0 0

Mar-00 0 0 0 0 0 0 0

Apr-00 0 0 0 0 0 0 0

May-00 0 0 0 0 0 0 0

Jun-00 0 0 0 0 0 0 0

Jul-00 0 0 0 0 0 0 0

Aug-00 0 0 0 0 0 0 0

Sep-00 0 0 0 0 0 0 0

Oct-00 0 0 0 0 0 0 0

Nov-00 0 0 0 0 0 0 0

Dec-00 0 0 0 0 1 0 0

Jan-01 0 0 0 0 1 0 0

Feb-01 0 1 0 0 1 0 0

Mar-01 0 1 0 0 1 0 0

Apr-01 0 1 0 1 1 0 0

May-01 0 1 0 1 1 0 0

Jun-01 0 1 0 1 1 0 0

Jul-01 1 1 0 1 1 0 0

Aug-01 1 1 0 1 1 0 0

Sep-01 1 1 0 1 1 0 0

Oct-01 1 1 0 1 1 0 0

Nov-01 1 1 0 1 1 0 0

Dec-01 1 1 0 1 1 0 0

Jan-02 1 1 0 1 0 0 0

Feb-02 1 1 0 1 0 0 0

Mar-02 1 1 0 1 0 0 0

Apr-02 0 1 0 1 0 0 0

May-02 0 1 0 1 0 0 0

Jun-02 0 1 0 1 0 1 1

Jul-02 0 1 0 1 0 1 1

Aug-02 0 1 0 1 0 1 1

Sep-02 0 1 0 1 0 1 1

Oct-02 0 1 0 1 1 1 1

Nov-02 0 1 1 1 1 1 1

Dec-02 0 1 1 1 1 1 1

Jan-03 0 1 1 1 1 1 1

Feb-03 0 1 1 1 1 1 1

Mar-03 0 1 1 1 1 1 1

Apr-03 0 0 1 1 1 0 1

May-03 0 0 1 1 1 0 1

Jun-03 0 0 0 1 1 0 1

Jul-03 0 0 0 0 0 0 1

Aug-03 0 0 0 0 0 0 1

Sep-03 0 0 0 0 0 0 1

Oct-03 0 0 0 0 0 0 1

Nov-03 0 0 0 0 0 0 1

Dec-03 0 0 0 0 0 0 0

Jan-04 0 0 0 0 0 0 0

Feb-04 0 0 0 0 0 0 0

Mar-04 0 0 0 0 0 0 0

Apr-04 0 0 0 0 0 0 0

May-04 0 0 0 0 0 0 0

Jun-04 0 0 0 0 0 0 0

Jul-04 0 0 0 0 0 0 0

Aug-04 0 0 0 0 0 0 0

Sep-04 0 0 0 0 0 0 0

Oct-04 0 0 0 0 0 0 0

Nov-04 0 0 0 0 0 0 0

Dec-04 0 0 0 0 0 0 0

Jan-05 0 0 0 0 0 0 0

Feb-05 0 0 0 0 0 0 0

Mar-05 0 0 0 0 0 0 0

Apr-05 0 0 0 0 0 0 0

May-05 0 0 0 0 0 0 0

Jun-05 0 0 0 0 0 0 0

Jul-05 0 0 0 0 0 0 0

Aug-05 0 0 0 0 0 0 0

Sep-05 0 0 0 0 0 0 0

Oct-05 0 0 0 0 0 0 0

Nov-05 0 0 0 0 0 0 0

Dec-05 0 0 0 0 0 0 0

Jan-06 0 0 0 0 0 0 0

Feb-06 0 0 0 0 0 0 0

Mar-06 0 0 0 0 0 0 0

Apr-06 0 0 0 0 0 0 0

May-06 0 0 0 0 0 0 0

Jun-06 0 0 0 0 0 0 0

Jul-06 0 0 0 0 0 0 0

Aug-06 0 0 0 0 0 0 0

Sep-06 0 0 0 0 0 0 0

Oct-06 0 0 0 0 0 0 0

Nov-06 0 0 0 0 0 0 0

Dec-06 0 0 0 0 0 0 0

Jan-07 0 0 0 0 0 0 0

Feb-07 0 0 0 0 0 0 0

Mar-07 0 0 0 0 0 0 0

Apr-07 0 0 0 0 0 0 0

May-07 0 0 0 0 0 0 0

Jun-07 0 0 0 0 0 0 0

Jul-07 0 0 0 0 0 0 0

Aug-07 0 0 0 0 0 0 0

Sep-07 0 0 0 0 0 0 0

Oct-07 0 0 0 0 0 0 0

Nov-07 0 0 0 0 0 0 0

Dec-07 0 0 0 0 0 0 0

Jan-08 0 0 0 0 0 0 0

Feb-08 0 0 0 0 0 0 0

Mar-08 0 0 1 0 1 0 0

Apr-08 0 0 1 0 1 0 0

May-08 0 0 1 0 1 0 0

Jun-08 0 1 1 1 1 1 0

Jul-08 0 1 1 1 1 1 0

Aug-08 0 1 1 1 1 1 0

Sep-08 1 1 1 1 1 1 0

Oct-08 1 1 1 1 1 1 0

Nov-08 1 1 1 1 1 1 1

Dec-08 1 1 1 1 1 1 1

Jan-09 1 1 1 1 1 1 1

Feb-09 1 1 1 1 1 1 1

Mar-09 1 1 1 1 1 1 1

Apr-09 1 1 1 1 1 1 1

May-09 1 1 1 1 1 1 1

Jun-09 1 1 1 1 1 1 1

Jul-09 1 1 1 1 1 1 1

Aug-09 1 1 0 1 1 1 1

Sep-09 1 1 0 1 1 1 1

Oct-09 1 0 0 1 0 1 1

Nov-09 0 0 0 1 0 1 1

Dec-09 0 0 0 0 0 0 1

Jan-10 0 0 0 0 0 0 1

Feb-10 0 0 0 0 0 0 1

Mar-10 0 0 0 0 0 0 1

Apr-10 0 0 0 0 0 0 1

May-10 0 0 0 0 0 0 1

Jun-10 0 0 0 0 0 0 1

Jul-10 0 0 0 0 0 0 0

Aug-10 0 0 0 0 0 0 0

Sep-10 0 0 0 0 0 0 0

Oct-10 0 0 0 0 0 0 0

Nov-10 0 0 0 0 0 0 0

Dec-10 0 0 0 0 0 0 0

Jan-11 0 0 0 0 0 0 0

Feb-11 0 0 0 0 0 0 0

Mar-11 0 0 0 0 0 0 0

Apr-11 0 0 0 0 0 0 0

May-11 0 0 0 0 0 0 0

Jun-11 0 0 0 0 0 0 0

Jul-11 0 0 0 0 0 0 0

Aug-11 0 0 0 0 0 0 0

Sep-11 0 0 0 0 0 0 0

Oct-11 0 0 0 0 0 0 0

Nov-11 0 0 0 0 0 0 0

Dec-11 0 0 0 0 0 0 0

Jan-12 0 0 0 0 0 0 0

Feb-12 0 0 0 0 0 0 0

Mar-12 0 0 0 0 0 0 0

Apr-12 0 0 0 0 0 0 0

May-12 0 0 0 0 0 0 0

Jun-12 0 0 0 0 0 0 0

Jul-12 0 0 0 0 0 0 0

Aug-12 0 0 0 0 0 0 0

Sep-12 0 0 0 0 0 0 0

Oct-12 0 0 0 0 0 0 0

Nov-12 0 0 0 0 0 0 0

Dec-12 0 0 0 0 0 0 0

Jan-13 0 0 0 0 0 0 0

Feb-13 0 0 0 0 0 0 0

Mar-13 0 0 0 0 0 0 0

Apr-13 0 0 0 0 0 0 0

May-13 0 0 0 0 0 0 0

Jun-13 0 0 0 0 0 0 0

Jul-13 0 0 0 0 0 0 0

Aug-13 0 0 0 0 0 0 0

Sep-13 0 0 0 0 0 0 0

Oct-13 0 0 0 0 0 0 0

Nov-13 0 0 0 0 0 0 0

Dec-13 0 0 0 0 0 0 0

Jan-14 0 0 0 0 0 0 0

Feb-14 0 0 0 0 0 0 0

Mar-14 0 0 0 0 0 0 0

Apr-14 0 0 0 0 0 0 0

May-14 0 0 0 0 0 0 0

Jun-14 0 0 0 0 0 0 0

Jul-14 0 0 0 0 0 0 0

Aug-14 0 0 0 0 0 0 0

Sep-14 0 0 0 0 0 0 0

Oct-14 0 0 0 0 0 0 0

Nov-14 0 0 0 0 0 0 0

Dec-14 0 0 0 0 0 0 0

Jan-15 0 0 0 0 0 0 0

Feb-15 0 0 0 0 0 0 0

Mar-15 0 0 0 0 0 0 0

Apr-15 0 0 0 0 0 0 0

May-15 0 0 0 0 0 0 0

Jun-15 0 0 0 0 0 0 0

Jul-15 0 0 0 0 0 0 0

Aug-15 0 0 0 0 0 0 0

Sep-15 0 0 0 0 0 0 0

Oct-15 0 0 0 0 0 0 0

Nov-15 0 0 0 0 0 0 0

Dec-15 0 0 0 0 0 0 0

Jan-16 0 0 0 0 0 0 0

Feb-16 0 0 0 0 0 0 0

Mar-16 0 0 0 0 0 0 0

Apr-16 0 0 0 0 0 0 0

May-16 0 0 0 0 0 0 0

Jun-16 0 0 0 0 0 0 0

Jul-16 0 0 0 0 0 0 0

Aug-16 0 0 0 0 0 0 0

Sep-16 0 0 0 0 0 0 0

Oct-16 0 0 0 0 0 0 0

Nov-16 0 0 0 0 0 0 0

Dec-16 0 0 0 0 0 0 0

Jan-17 0 0 0 0 0 0 0

Feb-17 0 0 0 0 0 0 0

Mar-17 0 0 0 0 0 0 0

Apr-17 0 0 0 0 0 0 0

May-17 0 0 0 0 0 0 0

Jun-17 0 0 0 0 0 0 0

Jul-17 0 0 0 0 0 0 0

Aug-17 0 0 0 0 0 0 0

Sep-17 0 0 0 0 0 0 0

Oct-17 0 0 0 0 0 0 0

Nov-17 0 0 0 0 0 0 0

Dec-17 0 0 0 0 0 0 0

Jan-18 0 0 0 0 0 0 0

Feb-18 0 0 0 0 0 0 0

Mar-18 0 0 0 0 0 0 0

Apr-18 0 0 0 0 0 0 0

Date Laredo Mcallen San

Antonio

Yield

Spread

US

Yield

Spread

MX

Real P/$

Exchange

Rate Index

WTI Spot

Crude Oil

Price

Jan-91 0 0 0 1.6795 0.9700 104.9979 24.9590

Feb-91 0 0 0 1.7337 1.2500 105.4182 20.5230

Mar-91 0 0 0 2.0165 0.9600 105.8416 19.8600

Apr-91 0 0 0 2.2145 0.8100 106.2704 20.8230

May-91 0 0 0 2.4364 0.2900 106.7045 21.2400

Jun-91 0 0 0 2.5290 0.7800 107.1393 20.1950

Jul-91 0 0 0 2.5191 -0.0900 107.5713 21.4200

Aug-91 0 0 0 2.4023 0.8400 108.0133 21.6880

Sep-91 0 0 0 2.2765 0.4000 108.4481 21.8570

Oct-91 0 0 0 2.3859 -0.6200 108.8797 23.2280

Nov-91 0 0 0 2.7305 0.0000 109.1450 22.4650

Dec-91 0 0 0 2.9057 -0.2900 109.1524 19.5170

Jan-92 0 0 0 3.1238 0.4000 109.0970 18.8200

Feb-92 0 0 0 3.3900 0.7100 108.9256 18.9950

Mar-92 0 0 0 3.4014 0.5200 109.0223 18.9160

Apr-92 0 0 0 3.6410 -0.6200 109.0810 20.2430

May-92 0 0 0 3.6710 -0.6100 110.1455 20.9400

Jun-92 0 0 0 3.5150 -0.6100 110.8768 22.3750

Jul-92 0 0 0 3.5605 0.2700 110.8060 21.7590

Aug-92 0 0 0 3.3910 1.4600 109.9076 21.3500

Sep-92 0 0 0 3.4514 1.7200 109.7284 21.9020

Oct-92 0 0 0 3.6571 -0.0500 110.8772 21.6880

Nov-92 0 0 0 3.6642 0.5500 110.9205 20.3420

Dec-92 0 0 0 3.4786 2.0800 110.8644 19.4070

Jan-93 0 0 0 3.5289 2.2700 110.5739 19.0750

Feb-93 0 0 0 3.2674 1.9800 110.1792 20.0530

Mar-93 0 0 0 2.9657 1.3700 110.5134 20.3470

Apr-93 0 0 0 3.0395 0.9900 110.0583 20.2700

May-93 0 0 0 3.0150 1.7000 111.0254 19.9400

Jun-93 0 0 0 2.8155 0.9500 110.9756 19.0700

Jul-93 0 0 0 2.6990 0.8900 111.0574 17.8660

Aug-93 0 0 0 2.5895 0.3100 110.6663 18.0090

Sep-93 0 0 0 2.3514 -0.3000 110.6699 17.5140

Oct-93 0 0 0 2.2380 0.0700 110.7232 18.1450

Nov-93 0 0 0 2.5405 -2.0100 112.1845 16.6990

Dec-93 0 0 0 2.6423 -0.7500 110.4921 14.5100

Jan-94 0 0 0 2.7055 0.2000 110.4850 15.0000

Feb-94 0 0 0 2.6395 1.0800 110.6272 14.7800

Mar-94 0 0 0 2.8891 1.6400 116.7639 14.6600

Apr-94 0 0 0 3.1895 -0.6300 119.2349 16.3800

May-94 0 0 0 2.9124 0.0500 117.7558 17.8800

Jun-94 0 0 0 2.8509 -0.6100 119.4873 19.0700

Jul-94 0 0 0 2.8435 -0.7400 120.9166 19.6500

Aug-94 0 0 0 2.6283 -1.0500 120.2482 18.3800

Sep-94 0 0 0 2.7086 -0.3700 120.8775 17.4600

Oct-94 1 0 0 2.6380 0.0300 121.4464 17.7100

Nov-94 1 0 0 2.5110 0.5400 122.3992 18.1000

Dec-94 1 0 0 2.0452 -3.3900 139.7569 17.1600

Jan-95 1 1 0 1.8785 -1.7700 196.0215 17.9900

Feb-95 1 1 0 1.5305 -5.5000 202.1404 18.5300

Mar-95 1 1 0 1.2896 -21.5200 238.2814 18.5500

Apr-95 1 1 0 1.2221 -26.5500 223.9779 19.8700

May-95 1 1 0 0.7823 -24.1700 211.9996 19.7400

Jun-95 1 1 0 0.5309 -11.7500 221.2615 18.4200

Jul-95 1 1 0 0.6870 -3.8300 218.2821 17.3000

Aug-95 1 1 0 0.9243 1.4600 220.1131 18.0300

Sep-95 1 1 0 0.7660 2.6500 224.0810 18.2300

Oct-95 0 1 0 0.5957 0.1400 237.8974 17.4400

Nov-95 0 1 0 0.4119 1.3400 272.2891 17.9900

Dec-95 0 1 0 0.4160 -1.1200 272.3353 19.0400

Jan-96 0 1 0 0.4995 -0.0800 266.8279 18.8800

Feb-96 0 1 0 0.8455 4.5300 266.8066 19.0700

Mar-96 0 1 0 1.1724 2.3200 269.2741 21.3600

Apr-96 0 1 0 1.4223 2.8800 265.6369 23.5700

May-96 0 0 0 1.5864 6.0500 264.3285 21.2500

Jun-96 0 0 0 1.6785 5.7500 268.1683 20.4500

Jul-96 0 0 0 1.5718 3.5900 271.0269 21.3200

Aug-96 0 0 0 1.4509 6.2300 267.1586 21.9600

Sep-96 0 0 0 1.5925 6.3500 268.2465 23.9900

Oct-96 0 0 0 1.4068 2.3800 273.2384 24.9000

Nov-96 0 0 0 1.0305 -2.6000 281.5510 23.7100

Dec-96 0 0 0 1.2595 -1.5200 280.0506 25.3900

Jan-97 0 0 0 1.4143 2.3900 278.3866 25.1700

Feb-97 0 0 0 1.2779 4.3400 277.0605 22.2100

Mar-97 0 0 0 1.4055 2.4400 283.1118 20.9900

Apr-97 0 0 0 1.5855 2.7900 281.0105 19.7200

May-97 0 0 0 1.5133 4.6400 281.0816 20.8300

Jun-97 0 0 0 1.4190 2.4400 282.5323 19.1700

Jul-97 0 0 0 1.0255 1.9300 280.3706 19.6300

Aug-97 0 0 0 1.0205 2.2500 276.7654 19.9300

Sep-97 0 0 0 1.1252 3.2000 276.5840 19.7900

Oct-97 0 0 0 0.9173 2.2100 277.7289 21.2600

Nov-97 0 0 0 0.5989 1.8900 294.5212 20.1700

Dec-97 0 0 0 0.5055 1.1900 289.2698 18.3200

Jan-98 0 0 0 0.3610 2.0000 290.8342 16.7100

Feb-98 0 0 0 0.3400 1.7300 301.9627 16.0600

Mar-98 0 0 0 0.4882 1.5800 304.6648 15.0200

Apr-98 0 0 0 0.5571 1.4900 302.2116 15.4400

May-98 0 0 0 0.5100 3.0900 304.3804 14.8600

Jun-98 0 0 0 0.3759 3.7000 316.2555 13.6600

Jul-98 1 0 0 0.3655 4.4100 316.5755 14.0800

Aug-98 1 0 0 0.2990 5.3500 329.2186 13.3600

Sep-98 1 0 0 0.0710 2.2000 363.2014 14.9500

Oct-98 1 0 0 0.4595 3.6400 360.9572 14.3900

Nov-98 1 0 0 0.2989 2.8800 355.0968 12.8500

Dec-98 1 0 0 0.1532 0.9400 352.4022 11.2800

Jan-99 1 0 0 0.2737 -0.8600 359.4700 12.4700

Feb-99 1 0 0 0.4447 -0.4700 356.0774 12.0100

Mar-99 0 0 0 0.6635 1.6300 347.3424 14.6600

Apr-99 0 0 0 0.7709 2.9700 335.8491 17.3400

May-99 0 0 0 0.9090 2.8500 332.8715 17.7500

Jun-99 0 0 0 1.1845 2.4700 339.2503 17.8900

Jul-99 0 0 0 1.1043 3.8000 333.0403 20.0700

Aug-99 0 0 0 1.0664 4.9200 334.1418 21.2600

Sep-99 0 0 0 1.0976 4.3000 332.0864 23.8800

Oct-99 0 0 0 1.0865 5.0100 339.1987 22.6400

Nov-99 0 0 0 0.8025 3.6700 334.9386 24.9700

Dec-99 0 0 0 0.9232 2.3100 334.7477 26.0800

Jan-00 0 0 0 1.1610 2.6300 337.0298 27.1800

Feb-00 0 0 0 0.7930 1.4600 335.8316 29.3500

Mar-00 0 0 0 0.3961 2.2300 330.5091 29.8900

Apr-00 0 0 0 0.1684 4.1100 333.3144 25.7400

May-00 0 0 0 0.4455 3.4300 338.0538 28.7800

Jun-00 0 0 0 0.2382 1.2300 348.3538 31.8300

Jul-00 0 0 0 -0.0925 2.1100 336.6565 29.7700

Aug-00 0 0 0 -0.4474 1.7200 330.1074 31.2200

Sep-00 1 0 0 -0.3750 1.0500 331.7891 33.8800

Oct-00 1 0 0 -0.5548 0.7800 338.4129 33.0800

Nov-00 1 0 0 -0.6362 -0.4400 338.4022 34.4000

Dec-00 1 0 0 -0.6970 0.0700 335.7712 28.4600

Jan-01 1 0 0 -0.1252 0.0000 347.3690 29.5800

Feb-01 1 0 0 0.0947 0.7100 344.9726 29.6100

Mar-01 1 0 0 0.3545 0.8800 341.9683 27.2400

Apr-01 1 0 0 1.1740 0.5600 332.4611 27.4100

May-01 1 0 0 1.6868 2.2900 325.2044 28.6400

Jun-01 0 0 0 1.7148 2.5600 323.3912 27.6000

Jul-01 0 0 1 1.6467 3.4300 325.5351 26.4500

Aug-01 0 0 1 1.5322 3.9800 324.5111 27.4700

Sep-01 0 0 1 2.0376 3.1700 333.6450 25.8800

Oct-01 0 0 1 2.3723 4.4800 333.0904 22.2100

Nov-01 0 0 1 2.7435 4.0000 327.8923 19.6700

Dec-01 0 0 1 3.3700 3.7900 325.9333 19.3300

Jan-02 0 0 1 3.3552 2.1300 325.7271 19.6700

Feb-02 0 0 1 3.1532 1.2000 323.7645 20.7400

Mar-02 0 0 1 3.4550 0.7900 322.8650 24.4200

Apr-02 0 0 1 3.4650 1.5800 324.6711 26.2700

May-02 0 0 1 3.3995 1.1200 337.4067 27.0200

Jun-02 0 0 1 3.1980 0.8200 346.2206 25.5200

Jul-02 0 0 1 2.9364 1.4100 348.3538 26.9400

Aug-02 0 0 1 2.6127 1.8800 349.3494 28.3800

Sep-02 0 0 1 2.2110 1.1700 357.0540 29.6700

Oct-02 0 0 1 2.3300 2.0100 358.9597 28.8500

Nov-02 1 0 1 2.7953 1.8700 362.7675 26.2700

Dec-02 1 0 1 2.8190 1.5300 362.5898 29.4200

Jan-03 1 0 1 2.8610 1.5700 376.0293 32.9400

Feb-03 1 0 1 2.7147 0.1500 388.3097 35.8700

Mar-03 1 0 1 2.6590 -0.0500 389.0599 33.5500

Apr-03 1 0 1 2.8076 0.4700 378.0274 28.2500

May-03 1 0 1 2.4819 1.3500 364.4528 28.1400

Jun-03 1 0 1 2.3938 1.3100 373.1530 30.7200

Jul-03 0 0 0 3.0595 2.0200 371.3077 30.7600

Aug-03 0 0 0 3.4810 2.1800 381.5935 31.5900

Sep-03 0 0 0 3.3148 1.5800 388.4484 28.2900

Oct-03 0 0 0 3.3482 0.8100 397.1557 30.3300

Nov-03 0 0 0 3.3478 1.4300 395.1682 31.0900

Dec-03 0 0 0 3.3555 0.8900 400.4444 32.1500

Jan-04 0 0 0 3.2495 0.8100 388.6368 34.2700

Feb-04 0 0 0 3.1353 0.6900 391.5523 34.7400

Mar-04 0 0 0 2.8765 0.3800 391.0403 36.7600

Apr-04 0 0 0 3.3919 0.2000 400.1102 36.6900

May-04 0 0 0 3.6805 1.4900 409.2974 40.2800

Jun-04 0 0 0 3.4419 1.6600 404.5723 38.0200

Jul-04 0 0 0 3.1424 1.4400 407.9322 40.6900

Aug-04 0 0 0 2.7768 1.3400 405.1660 44.9400

Sep-04 0 0 0 2.4486 1.2000 408.3695 45.9500

Oct-04 0 0 0 2.3055 0.9400 404.8354 53.1300

Nov-04 0 0 0 2.0825 0.8800 405.0985 48.4600

Dec-04 0 0 0 2.0073 0.1100 398.5636 43.3300

Jan-05 0 0 0 1.8490 0.0100 400.1849 46.8400

Feb-05 0 0 0 1.5905 0.0900 396.4375 47.9700

Mar-05 0 0 0 1.7041 0.4700 395.8117 54.3100

Apr-05 0 0 0 1.5033 0.6700 395.5842 53.0400

May-05 0 0 0 1.2381 0.4300 390.8128 49.8300

Jun-05 0 0 0 0.9636 0.0800 385.1952 56.2600

Jul-05 0 0 0 0.8925 -0.0200 380.1856 58.7000

Aug-05 0 0 0 0.7417 0.0300 379.3750 64.9700

Sep-05 0 0 0 0.7057 -0.2700 383.2433 65.5700

Oct-05 0 0 0 0.6675 -0.1000 385.0956 62.3700

Nov-05 0 0 0 0.5655 -0.1600 380.0860 58.3000

Dec-05 0 0 0 0.4990 -0.3000 377.5901 59.4300

Jan-06 0 0 0 0.0840 -0.1800 375.7342 65.5100

Feb-06 0 0 0 0.0305 -0.0900 372.6552 61.6300

Mar-06 0 0 0 0.0922 0.2000 380.6478 62.9000

Apr-06 0 0 0 0.2684 0.5400 391.8296 69.6900

May-06 0 0 0 0.2736 0.4200 393.7922 70.9400

Jun-06 0 0 0 0.1923 0.5400 404.8354 70.9600

Jul-06 0 0 0 0.0135 0.4800 392.0501 74.4100

Aug-06 0 0 0 -0.2104 0.4500 386.6138 73.0500

Sep-06 0 0 0 -0.2100 0.4700 390.1088 63.8700

Oct-06 0 0 0 -0.3162 0.4500 388.2564 58.8800

Nov-06 0 0 0 -0.4733 0.2400 387.3569 59.3700

Dec-06 0 0 0 -0.4130 0.2300 386.2974 62.0300

Jan-07 0 0 0 -0.3452 0.4200 388.7648 54.5700

Feb-07 0 0 0 -0.4432 0.3400 390.6706 59.2600

Mar-07 0 0 0 -0.5200 0.4800 395.5415 60.5600

Apr-07 0 0 0 -0.3167 0.3700 390.8270 63.9700

May-07 0 0 0 -0.1177 0.3600 385.0565 63.4600

Jun-07 0 0 0 0.3581 0.3800 385.1881 67.4800

Jul-07 0 0 0 0.0390 0.3200 383.8548 74.1800

Aug-07 1 0 0 0.3530 0.4300 392.3878 72.3900

Sep-07 1 0 0 0.5258 0.4900 392.6971 79.9300

Oct-07 1 0 0 0.5273 0.4500 385.4725 86.2000

Nov-07 1 0 0 0.7955 0.4400 386.3258 94.6200

Dec-07 1 0 0 1.0335 0.4100 385.7427 91.7300

Jan-08 1 0 0 0.9200 0.4000 388.1498 92.9500

Feb-08 1 0 0 1.5660 0.0000 383.2539 95.3500

Mar-08 1 1 0 2.2265 0.0700 381.6611 105.5600

Apr-08 1 1 0 2.3682 0.1600 374.3689 112.5700

May-08 1 1 0 2.1157 0.4100 371.6917 125.3900

Jun-08 1 1 0 2.2095 0.3300 367.2936 133.9300

Jul-08 1 1 1 2.3550 0.7200 364.0404 133.4400

Aug-08 1 1 1 2.1371 0.5400 358.7641 116.6100

Sep-08 1 1 1 2.5433 0.4900 375.9653 103.9000

Oct-08 1 1 1 3.1241 0.6300 443.4971 76.6500

Nov-08 1 1 1 3.3361 1.0800 464.3710 57.4400

Dec-08 1 1 1 2.3859 0.0500 475.4533 41.0200

Jan-09 1 1 1 2.3905 -0.3500 492.3985 41.7400

Feb-09 1 1 1 2.5747 0.0400 516.1772 39.1600

Mar-09 1 1 1 2.6041 0.4500 524.0454 47.9800

Apr-09 1 1 1 2.7719 0.0800 479.5918 49.7900

May-09 1 1 1 3.1130 0.1200 469.9104 59.1600

Jun-09 1 1 1 3.5414 0.2200 474.4329 69.6800

Jul-09 1 1 1 3.3764 0.5300 475.0729 64.0900

Aug-09 1 1 1 3.4181 0.6400 462.2591 71.0600

Sep-09 1 1 1 3.2767 0.8800 476.3813 69.4600

Oct-09 1 1 1 3.3157 0.8800 471.5423 75.8200

Nov-09 1 1 1 3.3479 0.7200 466.8456 78.0800

Dec-09 0 1 0 3.5355 0.5700 456.8869 74.3000

Jan-10 0 1 0 3.6684 0.5000 455.3438 78.2200

Feb-10 0 1 0 3.5811 0.5100 460.8689 76.4200

Mar-10 0 1 0 3.5796 0.6300 448.4747 81.2400

Apr-10 0 1 0 3.6877 0.6500 435.2556 84.4800

May-10 0 1 0 3.2600 0.4400 450.9209 73.8400

Jun-10 0 1 0 3.0764 0.2700 452.3715 75.3500

Jul-10 0 1 0 2.8524 0.2000 456.3073 76.3700

Aug-10 0 1 0 2.5450 0.2600 452.4675 76.8200

Sep-10 0 1 0 2.4981 0.3100 457.2637 75.3100

Oct-10 0 1 0 2.4055 0.6300 442.7754 81.9000

Nov-10 0 1 0 2.6170 0.5500 438.2102 84.1400

Dec-10 0 1 0 3.1491 0.5600 440.9123 89.0400

Jan-11 0 1 0 3.2380 0.6200 431.9029 89.4200

Feb-11 0 1 0 3.4479 0.7800 429.2327 89.5800

Mar-11 0 1 0 3.3096 0.4800 427.2666 102.9400

Apr-11 0 1 0 3.4015 0.6500 418.0616 110.0400

May-11 0 1 0 3.1290 0.5300 414.1328 101.3300

Jun-11 0 1 0 2.9627 0.3500 419.2562 96.2900

Jul-11 0 1 0 2.9625 0.5100 415.1319 97.1900

Aug-11 0 1 0 2.2757 0.4700 432.7206 86.3300

Sep-11 0 1 0 1.9662 0.2300 459.6850 85.6100

Oct-11 0 1 0 2.1310 0.0400 479.3003 86.4100

Nov-11 0 1 0 1.9960 0.1200 484.8112 97.2100

Dec-11 0 1 0 1.9686 0.2300 488.9284 98.5700

Jan-12 0 1 0 1.9355 0.4400 480.1500 100.2400

Feb-12 0 1 0 1.8780 0.2800 455.1447 102.2500

Mar-12 0 1 0 2.0859 0.4000 453.5341 106.1900

Apr-12 0 1 0 1.9662 0.2800 464.0262 103.3300

May-12 0 1 0 1.7105 0.2200 481.9598 94.7000

Jun-12 0 1 0 1.5286 0.2700 497.1201 82.4100

Jul-12 0 1 0 1.4329 0.4300 476.0506 87.9300

Aug-12 0 1 0 1.5774 0.4600 468.5700 94.1600

Sep-12 0 1 0 1.6147 0.4200 461.7471 94.7200

Oct-12 0 1 0 1.6452 0.4300 457.6833 89.5700

Nov-12 0 1 0 1.5565 0.4700 465.3061 86.6600

Dec-12 0 1 0 1.6500 0.5700 457.4771 88.2500

Jan-13 1 1 0 1.8357 0.4400 452.3181 94.6900

Feb-13 1 0 0 1.8811 0.1300 452.0515 95.3200

Mar-13 1 0 0 1.8730 0.2400 447.0774 93.0500

Apr-13 1 0 0 1.7000 0.2900 434.6477 92.0700

May-13 1 0 0 1.8859 0.2600 435.6183 94.8000

Jun-13 1 0 0 2.2495 0.3000 459.9339 95.8000

Jul-13 1 0 0 2.5445 0.1600 454.5652 104.6100

Aug-13 1 0 0 2.6964 0.1600 457.5980 106.5700

Sep-13 1 0 0 2.7940 0.0700 465.4946 106.2900

Oct-13 1 0 0 2.5727 0.1400 462.8707 100.5400

Nov-13 1 0 0 2.6521 0.2100 464.4599 93.8600

Dec-13 0 0 0 2.8333 0.3500 462.5542 97.6300

Jan-14 0 0 0 2.8167 0.5200 469.2491 94.6200

Feb-14 0 0 0 2.6574 0.6200 472.4739 100.8200

Mar-14 0 0 0 2.6676 0.5100 469.8642 100.8000

Apr-14 0 0 0 2.6790 0.4300 464.6270 102.0700

May-14 0 0 0 2.5276 0.3300 460.3534 102.1800

Jun-14 0 0 0 2.5643 0.0800 461.6085 105.7900

Jul-14 0 0 0 2.5136 0.2000 461.2600 103.5900

Aug-14 0 0 0 2.3871 0.2400 467.5034 96.5400

Sep-14 0 0 0 2.5100 0.2600 469.3238 93.2100

Oct-14 0 0 0 2.2832 0.2700 479.1581 84.4000

Nov-14 0 0 0 2.3083 0.3200 482.8948 75.7900

Dec-14 0 0 0 2.1809 0.4100 512.9275 59.2900

Jan-15 0 0 0 1.8525 0.5600 521.7841 47.2200

Feb-15 0 0 0 1.9621 0.4000 530.3527 50.5800

Mar-15 1 0 0 2.0123 0.4900 540.4359 47.8200

Apr-15 1 0 0 1.9168 0.5300 541.2359 54.4500

May-15 1 0 0 2.1835 0.5300 542.3985 59.2700

Jun-15 1 0 0 2.3450 0.5800 549.5342 59.8200

Jul-15 1 0 0 2.2877 0.6400 564.8901 50.9000

Aug-15 1 0 0 2.0981 0.6600 586.2192 42.8700

Sep-15 1 0 0 2.1457 0.6200 598.6347 45.4800

Oct-15 1 0 0 2.0533 0.5100 590.2723 46.2200

Nov-15 1 0 0 2.1337 0.6800 591.4385 42.4400

Dec-15 1 0 0 2.0114 0.5400 604.4905 37.1900

Jan-16 1 0 0 1.8326 0.5000 639.1951 31.6800

Feb-16 1 0 0 1.4655 0.1700 657.1749 30.3200

Mar-16 1 0 0 1.5927 0.3100 630.6727 37.5500

Apr-16 1 0 0 1.5819 0.3000 621.9299 40.7500

May-16 1 0 0 1.5333 0.6600 641.4172 46.7100

Jun-16 1 0 0 1.3673 0.6600 662.9844 48.7600

Jul-16 1 0 0 1.1990 0.5800 660.2396 44.6500

Aug-16 1 0 0 1.2604 0.5200 656.9011 44.7200

Sep-16 1 0 0 1.3357 0.5900 680.4594 45.1800

Oct-16 1 0 0 1.4285 0.5400 673.6827 49.7800

Nov-16 1 0 0 1.6875 0.1600 709.0415 45.6600

Dec-16 1 0 0 1.9781 0.6000 729.3501 51.9700

Jan-17 1 0 0 1.9110 0.8500 759.9090 52.5000

Feb-17 1 0 0 1.8937 0.8800 724.6391 53.4700

Mar-17 0 0 0 1.7322 0.6000 689.9915 49.3300

Apr-17 0 0 0 1.4916 0.4900 666.9416 51.0600

May-17 0 0 0 1.3968 0.6900 667.9300 48.4800

Jun-17 0 0 0 1.1909 0.3800 646.7361 45.1800

Jul-17 0 0 0 1.2335 0.2600 634.6903 46.6300

Aug-17 0 0 0 1.1778 0.2800 633.1437 48.0400

Sep-17 0 0 0 1.1535 0.2100 632.8344 49.8200

Oct-17 0 0 0 1.2667 0.1500 665.7434 51.5800

Nov-17 0 0 0 1.0995 0.2700 674.7138 56.6400

Dec-17 0 0 0 1.0585 0.1600 677.7537 57.8800

Jan-18 0 0 0 1.1476 0.5500 675.6204 63.7000

Feb-18 0 0 0 1.2721 0.3000 662.3125 62.2300

Mar-18 0 0 0 1.1133 0.3700 664.2928 62.7300

Apr-18 0 0 0 1.0819 0.2700 652.2933 66.2500

WTI Spot Crude Oil Price