MARKETPSYCHSENTIMENT DATA IN THE FOREIGN EXCHANGE MARKETSIAN MACGILLIVRAY THOMSON REUTERS RESEARCH amp DEVELOPMENT

NOVEMBER 2013

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

REUTERSYUYA SHINO

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

EXECUTIVE SUMMARY

Financial markets are not entirely efficient in pricing all available information they are known to show systematic and repeating patterns that are apparently irrational ndash or ldquononoptimalrdquo in academic parlance

Behavioral economics studies have linked investor mood to this nonoptimal behavior For example recent academic research into patterns in the FOREX market has suggested that sentiment plays a role

We attempted to explain one of these patterns using Thomson Reuters MarketPsych Indices (TRMI) These multidimensional sentiment indices exclusively published by Thomson Reuters offer a proxy for the mood of investors and their perceptions of macroeconomic events in the Financial Markets

It is important to determine which signals are consistently predictive and to quantify what they offer used properly they can predict rates movement accurately enough to generate superior returns at lower risk

Our results show that a number of the TRMI indices are correlated with excess returns in the FOREX market A handful of these show power over a wide range of currencies and a 15-year time period For these few indices we find annualized 4ndash5 excess returns with attractively low levels of associated risk

For example our country-level index on financial system instability shows that this index provides a contrarian signal

Here the rightmost portfolios represent buying currencies with high levels of uncertainty negativity and fear around their financial systems The leftmost portfolios represent buying currencies from countries with low levels of financial system instability

This strategy leads to average monthly annualized returns of 4

It is one of five strategies that we look at in this paper

-100

000

100

200

300

400

500

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Country-Level Financial System Instability

REUTERSALEX GRIMM

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

THE FORWARD PREMIUM PUZZLE

The foreign exchange market ndash the worldrsquos largest financial market ndash is generally efficient but there is consistent empirical evidence of the ldquoforward premium puzzlerdquo which is that high-interest-rate currencies tend to offer superior returns compared to low-interest-rate currencies Investor irrationality is often cited as a possible explanation

BACKGROUNDFrom Famarsquos seminal 1984 paper Forward and Spot Exchange Rates through to Lustig et alrsquos 2012 Common Risk Factors in Currency Markets the apparent irrationality in FOREX markets is not a novel research problem but is nevertheless one that remains to this day

Recent work such as Yursquos 2013 paper A Sentiment-based Explanation of the Forward Premium Puzzle suggests that sentiment ndash the mood of investors and market participants ndash may help to explain this puzzle The MarketPsych Currency- and Country-level Indices are real-time proxies for sentiment when building a model of FOREX spot prices

MARKETPSYCH INDICESThe Thomson Reuters MarketPsych Indices give granular sentiment scores for currencies and countries on both mood-based topics (such as Joy or Fear) and macro- or currency-specific topics (such as Economic Uncertainty Instability or Government Anger) To create these ever-changing scores millions of documents are processed each day from news and social media sources MarketPsych uses innovative language models powered by sophisticated expert systems

Just consider the following phrase

ldquohelliptraders seem hopeful that the Yen will bounce back from its recent slumprdquo

Alone this fragment has very little influence but when thousands of news stories and social media sources are making similar predictions and observations this gets reflected in the MarketPsych indices for the Yen it can provide a real clue as to where people think the market is going and is scored on the Yen Price Forecast TRMI

We considered the full range of indices offered by MarketPsych for countries and currencies Many of these showed predictive power in certain regimes for certain currencies and currency-types within certain time periods On top of this a few stood out as generally predictive in a very wide range of conditions to offer a broad and powerful signal

We posit that investment professionals in the foreign exchange markets would do well to stay informed of general sentiment in addition to their usual arsenal of fundamentals data and that MarketPsych data can provide a real and usable proxy for this sentiment

AN EXAMPLEldquoProfitrdquo in the foreign exchange markets is generally the sum of gains made through changes in the spot price of a currency pair and the forward premium the risk-free interest rate differential between two countries

If I were to borrow $1m USD today at 2 and use it to buy an equivalent amount of Brazilian $R I could then invest that money in Brazil at 5

In three monthsrsquo time I might then wish to convert this money back to USD expecting to have made 3 profit thanks to this interest rate difference An efficient market should however have arbitraged away this opportunity lowering the USDBRL spot price to give me a net profit of 0

Over a large time period and all currencies this is not the case and this phenomenon is known as the Forward Premium Puzzle it is the subject of much academic and industry-led research

SENTIMENT EXPLAINED

MarketPsych offers 52 basic sentiment indices for currencies and countries Some of these are consistently predictive and others perform well under certain conditions While the approach and breadth offered by MarketPsych are novel this is by no means the first application of sentiment data to the markets

SENTIMENT IN THE MARKETSAny sentiment index aims at its core to represent the psyche of the aggregate investment professional whose actions will lead to price changes in the market Many neuroscience publications relate this to activation of the brainrsquos reward system or its loss avoidance system and studies have shown irrational behavior linked to weather (Krivelyova and Robotti 2003) to recent market activity (Fisher and Statman 2000) and public mood (Bollen et al 2011)

Baker amp Wurglerrsquos oft-cited Investor Sentiment in the Stock Market uses a number of market factors such as puts vs calls to calculate the implied investor sentiment in the equities market through market actions with good success Bollen et al use Twitterreg data to calculate public mood around given equities and again enjoy success

In the Forward Premium Puzzle we see constant empirical evidence of irrationality in the foreign exchange market This behavior can be linked to sentiment and sentiment has been proven predictive in the equity markets Recent research such as Yursquos 2013 paper A Sentiment-based Explanation of the Forward Premium Puzzle begins to show how with a good sentiment index similar successes could be found in the foreign exchange market

ldquoThe market can remain irrational a lot longer than you or I can remain solventrdquo

ndashJohn Maynard Keynes

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

EXECUTIVE SUMMARY

Financial markets are not entirely efficient in pricing all available information they are known to show systematic and repeating patterns that are apparently irrational ndash or ldquononoptimalrdquo in academic parlance

Behavioral economics studies have linked investor mood to this nonoptimal behavior For example recent academic research into patterns in the FOREX market has suggested that sentiment plays a role

We attempted to explain one of these patterns using Thomson Reuters MarketPsych Indices (TRMI) These multidimensional sentiment indices exclusively published by Thomson Reuters offer a proxy for the mood of investors and their perceptions of macroeconomic events in the Financial Markets

It is important to determine which signals are consistently predictive and to quantify what they offer used properly they can predict rates movement accurately enough to generate superior returns at lower risk

Our results show that a number of the TRMI indices are correlated with excess returns in the FOREX market A handful of these show power over a wide range of currencies and a 15-year time period For these few indices we find annualized 4ndash5 excess returns with attractively low levels of associated risk

For example our country-level index on financial system instability shows that this index provides a contrarian signal

Here the rightmost portfolios represent buying currencies with high levels of uncertainty negativity and fear around their financial systems The leftmost portfolios represent buying currencies from countries with low levels of financial system instability

This strategy leads to average monthly annualized returns of 4

It is one of five strategies that we look at in this paper

-100

000

100

200

300

400

500

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Country-Level Financial System Instability

REUTERSALEX GRIMM

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

THE FORWARD PREMIUM PUZZLE

The foreign exchange market ndash the worldrsquos largest financial market ndash is generally efficient but there is consistent empirical evidence of the ldquoforward premium puzzlerdquo which is that high-interest-rate currencies tend to offer superior returns compared to low-interest-rate currencies Investor irrationality is often cited as a possible explanation

BACKGROUNDFrom Famarsquos seminal 1984 paper Forward and Spot Exchange Rates through to Lustig et alrsquos 2012 Common Risk Factors in Currency Markets the apparent irrationality in FOREX markets is not a novel research problem but is nevertheless one that remains to this day

Recent work such as Yursquos 2013 paper A Sentiment-based Explanation of the Forward Premium Puzzle suggests that sentiment ndash the mood of investors and market participants ndash may help to explain this puzzle The MarketPsych Currency- and Country-level Indices are real-time proxies for sentiment when building a model of FOREX spot prices

MARKETPSYCH INDICESThe Thomson Reuters MarketPsych Indices give granular sentiment scores for currencies and countries on both mood-based topics (such as Joy or Fear) and macro- or currency-specific topics (such as Economic Uncertainty Instability or Government Anger) To create these ever-changing scores millions of documents are processed each day from news and social media sources MarketPsych uses innovative language models powered by sophisticated expert systems

Just consider the following phrase

ldquohelliptraders seem hopeful that the Yen will bounce back from its recent slumprdquo

Alone this fragment has very little influence but when thousands of news stories and social media sources are making similar predictions and observations this gets reflected in the MarketPsych indices for the Yen it can provide a real clue as to where people think the market is going and is scored on the Yen Price Forecast TRMI

We considered the full range of indices offered by MarketPsych for countries and currencies Many of these showed predictive power in certain regimes for certain currencies and currency-types within certain time periods On top of this a few stood out as generally predictive in a very wide range of conditions to offer a broad and powerful signal

We posit that investment professionals in the foreign exchange markets would do well to stay informed of general sentiment in addition to their usual arsenal of fundamentals data and that MarketPsych data can provide a real and usable proxy for this sentiment

AN EXAMPLEldquoProfitrdquo in the foreign exchange markets is generally the sum of gains made through changes in the spot price of a currency pair and the forward premium the risk-free interest rate differential between two countries

If I were to borrow $1m USD today at 2 and use it to buy an equivalent amount of Brazilian $R I could then invest that money in Brazil at 5

In three monthsrsquo time I might then wish to convert this money back to USD expecting to have made 3 profit thanks to this interest rate difference An efficient market should however have arbitraged away this opportunity lowering the USDBRL spot price to give me a net profit of 0

Over a large time period and all currencies this is not the case and this phenomenon is known as the Forward Premium Puzzle it is the subject of much academic and industry-led research

SENTIMENT EXPLAINED

MarketPsych offers 52 basic sentiment indices for currencies and countries Some of these are consistently predictive and others perform well under certain conditions While the approach and breadth offered by MarketPsych are novel this is by no means the first application of sentiment data to the markets

SENTIMENT IN THE MARKETSAny sentiment index aims at its core to represent the psyche of the aggregate investment professional whose actions will lead to price changes in the market Many neuroscience publications relate this to activation of the brainrsquos reward system or its loss avoidance system and studies have shown irrational behavior linked to weather (Krivelyova and Robotti 2003) to recent market activity (Fisher and Statman 2000) and public mood (Bollen et al 2011)

Baker amp Wurglerrsquos oft-cited Investor Sentiment in the Stock Market uses a number of market factors such as puts vs calls to calculate the implied investor sentiment in the equities market through market actions with good success Bollen et al use Twitterreg data to calculate public mood around given equities and again enjoy success

In the Forward Premium Puzzle we see constant empirical evidence of irrationality in the foreign exchange market This behavior can be linked to sentiment and sentiment has been proven predictive in the equity markets Recent research such as Yursquos 2013 paper A Sentiment-based Explanation of the Forward Premium Puzzle begins to show how with a good sentiment index similar successes could be found in the foreign exchange market

ldquoThe market can remain irrational a lot longer than you or I can remain solventrdquo

ndashJohn Maynard Keynes

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

THE FORWARD PREMIUM PUZZLE

The foreign exchange market ndash the worldrsquos largest financial market ndash is generally efficient but there is consistent empirical evidence of the ldquoforward premium puzzlerdquo which is that high-interest-rate currencies tend to offer superior returns compared to low-interest-rate currencies Investor irrationality is often cited as a possible explanation

BACKGROUNDFrom Famarsquos seminal 1984 paper Forward and Spot Exchange Rates through to Lustig et alrsquos 2012 Common Risk Factors in Currency Markets the apparent irrationality in FOREX markets is not a novel research problem but is nevertheless one that remains to this day

Recent work such as Yursquos 2013 paper A Sentiment-based Explanation of the Forward Premium Puzzle suggests that sentiment ndash the mood of investors and market participants ndash may help to explain this puzzle The MarketPsych Currency- and Country-level Indices are real-time proxies for sentiment when building a model of FOREX spot prices

MARKETPSYCH INDICESThe Thomson Reuters MarketPsych Indices give granular sentiment scores for currencies and countries on both mood-based topics (such as Joy or Fear) and macro- or currency-specific topics (such as Economic Uncertainty Instability or Government Anger) To create these ever-changing scores millions of documents are processed each day from news and social media sources MarketPsych uses innovative language models powered by sophisticated expert systems

Just consider the following phrase

ldquohelliptraders seem hopeful that the Yen will bounce back from its recent slumprdquo

Alone this fragment has very little influence but when thousands of news stories and social media sources are making similar predictions and observations this gets reflected in the MarketPsych indices for the Yen it can provide a real clue as to where people think the market is going and is scored on the Yen Price Forecast TRMI

We considered the full range of indices offered by MarketPsych for countries and currencies Many of these showed predictive power in certain regimes for certain currencies and currency-types within certain time periods On top of this a few stood out as generally predictive in a very wide range of conditions to offer a broad and powerful signal

We posit that investment professionals in the foreign exchange markets would do well to stay informed of general sentiment in addition to their usual arsenal of fundamentals data and that MarketPsych data can provide a real and usable proxy for this sentiment

AN EXAMPLEldquoProfitrdquo in the foreign exchange markets is generally the sum of gains made through changes in the spot price of a currency pair and the forward premium the risk-free interest rate differential between two countries

If I were to borrow $1m USD today at 2 and use it to buy an equivalent amount of Brazilian $R I could then invest that money in Brazil at 5

In three monthsrsquo time I might then wish to convert this money back to USD expecting to have made 3 profit thanks to this interest rate difference An efficient market should however have arbitraged away this opportunity lowering the USDBRL spot price to give me a net profit of 0

Over a large time period and all currencies this is not the case and this phenomenon is known as the Forward Premium Puzzle it is the subject of much academic and industry-led research

SENTIMENT EXPLAINED

MarketPsych offers 52 basic sentiment indices for currencies and countries Some of these are consistently predictive and others perform well under certain conditions While the approach and breadth offered by MarketPsych are novel this is by no means the first application of sentiment data to the markets

SENTIMENT IN THE MARKETSAny sentiment index aims at its core to represent the psyche of the aggregate investment professional whose actions will lead to price changes in the market Many neuroscience publications relate this to activation of the brainrsquos reward system or its loss avoidance system and studies have shown irrational behavior linked to weather (Krivelyova and Robotti 2003) to recent market activity (Fisher and Statman 2000) and public mood (Bollen et al 2011)

Baker amp Wurglerrsquos oft-cited Investor Sentiment in the Stock Market uses a number of market factors such as puts vs calls to calculate the implied investor sentiment in the equities market through market actions with good success Bollen et al use Twitterreg data to calculate public mood around given equities and again enjoy success

In the Forward Premium Puzzle we see constant empirical evidence of irrationality in the foreign exchange market This behavior can be linked to sentiment and sentiment has been proven predictive in the equity markets Recent research such as Yursquos 2013 paper A Sentiment-based Explanation of the Forward Premium Puzzle begins to show how with a good sentiment index similar successes could be found in the foreign exchange market

ldquoThe market can remain irrational a lot longer than you or I can remain solventrdquo

ndashJohn Maynard Keynes

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

MARKETPSYCH INDICESMood-based indices are available for all currencies and countries For concepts such as Joy Fear Optimism and Uncertainty among others these represent the aggregate mood of those writing about a given asset Uncertainty around Greece for example is extremely high around the debt crises in late 2009

Asset-class specific indices are also available for currencies and countries Positive and negative references to a given countryrsquos economic growth for example are aggregated into the ldquoEconomic Growthrdquo index

Certain indices may be sparse the Danish Kroner for example is the subject of relatively little discussion and is therefore the sparsest index we considered The currency-specific ldquoCurrency Peg Instabilityrdquo index is

as expected only used for currencies that are actually pegged Appendix B contains further information on available indices and their use

MARKETPSYCH QUrsquoEST-QUE CrsquoEST The MarketPsych Indices are calculated from news and social media text News data is collected from a host of mainstream news sources and top business publications featuring the entire Reuters news feed and high-quality Internet news from Moreover Technologies Social media data is collected from top finance-related blogs microblogs and tweets as ranked by popularity indices

Indices are calculated minutely an indication of the number of articles (the ldquobuzzrdquo) contributing to the indices for a given asset (countrycurrency) is provided for each set of data points

FINDING ALPHA

With 26 currencies and 52 sentiment indices we have a wealth of information available to us but it takes some work to determine which signals are consistently predictive and to quantify what they offer

We set out to find the indices that are most historically predictive These are shown in detail in Appendix A For completeness Appendix B shows the NA values for each index in each currency ndash how often there is insufficient source data to generate a value

MEASURING ALPHAWe measured the fit of MarketPsych Indices individually as independent variables against forward premiums and spot-price movements We used the Fama-Macbeth regression for panel analysis over all currencies and used a Newey-West estimator to correct for autocorrelation and heteroskedasticity in single-currency regressions

For the top performers listed we found p-values below 005 and high r2 values in panel regressions using 26 currencies against the USD in a 15-year time period from 1998 to 2013

Further portfolio testing showed significant trending in profit towards portfolios composed of currencies with higher sentiment values in a given month for all four top performing indices with monthly High-Minus-Low means of +4 and Sharpe Ratios from 04 to 06

PORTFOLIO TESTINGWe used MarketPsych sentiment data for 26 currencies against the USD in a 15-year time period from 1998 to 2013 in order to create six monthly portfolios (baskets) of currencies We did this for each individual high-performing sentiment index separately

We did not consider currencies in a month where they had few sentiment scores for the given index There was little news discussing instability in the Danish financial system in February 2006 for example thus the DKK did not feature in portfolios for that index in March 2006 The average monthly portfolio contained four currencies

Profit was calculated as the difference in spot prices for a given currency vs the USD between the day of investment on the start of a month and the strike date at the end of the month combined with the one-month forward premium (essentially the difference in risk-free interest rates) on the day of investment

We used only out-of-sample (ie prior) data to form portfolios The March 2011 monthly test for example only used sentiment data from February 2011 to rank currencies into six portfolios Profit (excess returns) was defined as the percentage difference between the March

31st spot price and the March 1st spot price plus the one-month forward premium at March 1st

Detailed results are presented in Appendix A

REVEALING THE TOP PERFORMERSWe found a number of sentiment indices that performed well for only a given few currencies or only in given market conditions Others were found to be highly reflective of market movements but lacked significant predictive power

Five indices however show predictive power in a wide range of conditions over all currencies considered in a long time frame

CO_ECONGRW (Economic Growth) and CO_EC_UNCR (Economic Uncertainty) measure statements in text around growth of the economy and uncertainty respectively The economy in this case is defined as all of its constituent parts from phrases about individual business entities all the way up to general text around a given countryrsquos economy

It is already widely accepted that numbers such as consumption growth are predictive of spot-price movements These indices do not simply serve as a proxy or a peek ahead at the figures Sentiment around these topics instead reflects investor beliefs and emotion around these fundamentals It is beliefs that move markets in the short term as these indices show

We further found CO_FNSYSIN (Financial System Instability) a narrower signal based only on discussion explicitly relating to a countryrsquos financial system CO_TRUST (Country-Level Trust) and CU_UNCERTN (Currency-Level Uncertainty) to be predictive in a wide range of market conditions

These indices found to be similar to and slightly predictive of market volatility show investor sentiment around a given countryrsquos overall economic health and stability ndash a natural factor in trading decisions

They represent a rational way of exploiting sentiment ndash and for finding alpha

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix A Portfolio Trading 042006 to 052013Evenly split portfolios containing all currencies with available data (average four per portfolio) for each month were rebalanced monthly based on an average of the previous monthrsquos sentiment index We then calculated the excess returns (forward premium + spot change) against the USD for the following month

In practice this would involve investing in currencies 4 5 and 6 and selling or shorting the currencies 1 2 and 3

Portfolio Number 1 2 3 4 5 6

Spot Mean 170 030 -110 -170 000 -160

SD 780 990 920 1010 870 1000

Premium Mean 180 170 170 170 120 150

SD 050 050 060 060 050 060

Excess Mean 010 140 280 340 120 310

SD 780 980 920 1000 870 1000

SR 00128 01429 03043 034 01379 031

Switching Frequency 5030 6898 8524 7741 7139 5301

Sentiment Mean -00102 -00028 00011 00057 00104 0018

HML Mean NA 138 274 333 114 300

SD NA 744 607 703 581 716

SR NA 01855 04514 04737 01962 0419

Avg Switch Freq 6772

CO_ECONGRW (Country-Level Economic Growth)

Portfolio Number 1 2 3 4 5 6

Spot Mean 020 050 -110 170 -030 -290

SD 960 930 900 1030 840 890

Premium Mean 080 130 210 210 200 130

SD 040 050 060 060 060 060

Excess Mean 060 080 320 040 230 420

SD 940 920 890 1030 840 890

SR 00638 0087 03596 00388 02738 04719

Switching Frequency 5813 7289 8765 7620 7651 6325

Sentiment Mean -00023 -00007 00001 00006 00012 00027

HML Mean NA 016 253 -026 170 356

SD NA 461 489 546 564 627

SR NA 00347 05174 -00476 03014 05678

Avg Switch Freq 7244

CO_EC_UNCR (Country-Level Economic Uncertainty)

000

100

200

300

400

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_ECONGRW

000

100

200

300

500

400

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_EC_UNCR

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Portfolio Number 1 2 3 4 5 6

Spot Mean 160 160 130 -090 -270 -300

SD 970 870 1020 940 940 990

Premium Mean 150 160 170 200 140 100

SD 060 060 060 060 060 060

Excess Mean -010 000 040 300 400 500

SD 960 870 1010 930 930 980

SR -00104 0 00396 03226 04301 04082

Switching Frequency 5030 8755 6807 7169 7410 7199

Sentiment Mean -00102 -00014 -00002 0 00001 00005

HML Mean NA 018 058 310 419 420

SD NA 718 710 690 685 758

SR NA 00251 00817 04493 06117 05541

Avg Switch Freq 7323

CO_FNSYSIN (Country-Level Financial System Instability)

Portfolio Number 1 2 3 4 5 6

Spot Mean 180 190 -310 -200 -050 -060

SD 930 780 1060 1030 910 940

Premium Mean 070 100 170 170 220 240

SD 040 040 050 080 070 060

Excess Mean -110 -090 470 370 260 300

SD 930 780 1060 1010 900 920

SR -01183 -01154 04434 03663 02889 03261

Switching Frequency 5813 4880 6687 7771 7048 6084

Sentiment Mean -00023 -00064 -00018 00016 00059 00166

HML Mean NA 013 578 477 367 402

SD NA 674 546 628 586 663

SR NA 00193 10586 07596 06263 06063

Avg Switch Freq 6235

CU_UNCERTN (Currency-Level Uncertainty)

Portfolio Number 1 2 3 4 5 6

Spot Mean 072 066 -099 188 -301 -172

SD 924 924 822 1032 841 1099

Premium Mean 145 143 135 146 149 249

SD 033 042 048 072 052 081

Excess Mean 074 078 234 -042 450 421

SD 924 915 820 1022 837 1077

SR 00798 00847 02855 -00415 05372 03913

Switching Frequency 5030 4669 6988 7982 7440 7229

Sentiment Mean -00102 -00105 -00023 00017 00047 00075

HML Mean NA 004 161 -116 376 348

SD NA 608 593 581 634 703

SR NA 00066 02715 -01997 05931 0495

Avg Switch Freq 6506

CO_TRUST (Country-Level Trust)

-100

000

100

200

300

400

500

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_FNSYSIN

-200

000

200

400

600

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CU_UNCERTN

-100

000

100

200

300

400

500

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_TRUST

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

We see here high NA values for PEGINST (currency peg instability) for currencies not involved in pegs and we can see in general that the DKK is a rarely-discussed currency

Data for indices calculated at the currency level are presented below In addition to the standard set of MarketPsych indices currency-level indices further include market forecast carry trade peg instability and market momentum

AUD 1226 0 0 0 193 217 028 121 006 006 028 0 011 6726 9089 017

BRL 3567 0 0 175 5505 6505 4009 2595 491 1585 2949 022 4274 9718 9961 4136

CAD 1475 0 0 006 458 3026 094 331 006 011 055 0 072 8818 9188 061

CHF 2904 0 011 1287 3876 7896 323 3203 58 265 1645 0 2568 8117 9393 1491

DKK 8541 0 6392 9419 9893 9959 9745 9716 8697 9538 952 6605 9639 9994 9751 9254

EGP 7114 0 2438 7909 9131 977 8531 7455 602 7623 8464 2012 8677 9989 9983 8711

EUR 878 0 0 0 0 0 0 0 0 0 0 0 0 6842 6328 0

GBP 1385 0 0 0 282 1679 028 017 0 0 011 0 017 8978 9757 011

HKD 2138 0 0 486 1552 63 1253 2921 077 552 668 0 1193 8835 7405 834

IDR 3828 0 11 2199 526 8199 4392 5298 10 2486 2757 122 2503 9796 9978 3326

ILS 5976 0 97 5432 8268 9518 801 7057 3862 634 6463 116 6452 9961 9994 616

INR 1665 0 0 055 768 3429 248 342 022 055 182 0 094 9845 9729 199

JPY 942 0 0 0 006 1137 011 011 0 0 0 0 0 328 9685 0

KRW 1825 0 0 088 834 423 872 414 011 177 188 0 409 9845 9983 32

MXN 3662 0 017 1888 323 7985 4583 4506 1022 2849 2844 039 3153 9779 9923 3114

MYR 3355 0 05 1088 4851 7796 4254 4293 602 1646 1983 055 1923 9934 9718 2133

NOK 7869 0 4226 8612 948 9886 9417 9377 8235 8766 8761 4101 8932 996 9983 8298

NZD 1977 0 0 171 1347 5831 1033 1999 055 309 42 0 718 7742 9757 271

RUB 5635 0 645 483 7788 9383 6809 6976 3613 5442 5753 1201 5931 9867 9939 6348

SEK 7328 0 2644 7457 9178 9793 877 9106 7328 8178 8317 2789 8379 9922 9939 8122

SGD 2086 0 0 204 111 6046 795 2391 083 342 502 0 806 9304 9067 646

THB 2631 0 017 563 3462 6455 2579 1911 16 707 1436 033 1198 995 9459 153

TRY 4276 0 11 2489 5679 8793 5764 6861 1519 2895 2228 135 3637 9637 100 4388

TWD 2751 0 0 629 2082 7018 2214 3981 409 1099 1441 0 1888 9398 9431 1679

USD 488 0 0 0 0 028 0 0 0 0 0 0 0 4694 2595 0

ZAR 1557 0 006 094 69 3026 204 26 017 099 171 006 348 8647 9564 221

Averages 3349 0 678 2179 3651 5995 3341 3505 1685 2345 2569 703 2801 8830 9215 2741

CURRENCY

N

A VALS

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

PRICEUP

MKTFCST

CARYTRD

PEGINST

MKTMMTM

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

Data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below In addition to the standard set of MarketPsych indices country-level indices further include market risk trade balance budget deficit central bank and consumer sentiment

AUD 274 0 0 0 0 0 0 0 0 0 0 0 061 3793 243 006

BRL 490 0 0 0 011 011 0 0 0 0 0 0 11 677 05 403

CAD 188 0 0 0 0 0 0 0 0 0 0 0 072 2457 293 0

CHF 929 0 0 0 055 099 0 0 0 0 0 0 4147 9094 0 541

DKK 1755 0 0 024 1143 936 089 018 0 036 041 0 7636 9034 529 2079

EGP 1126 0 0 0 146 353 017 0 0 0 006 0 4888 7534 633 3318

EUR 002 0 0 0 0 0 0 0 0 0 0 0 0 033 0 0

GBP 127 0 0 0 0 0 0 0 0 0 0 0 061 1844 0 0

HKD 984 0 0 0 033 022 0 0 0 0 0 0 2419 873 2927 624

IDR 537 0 0 0 017 028 006 0 0 0 0 0 32 6282 359 105

ILS 941 0 0 0 0 0 0 0 0 0 0 0 3223 7203 2489 1194

INR 273 0 0 0 0 0 0 0 0 0 0 0 0 4075 017 0

JPY 335 0 0 0 0 0 0 0 0 0 0 0 006 5014 0 0

KRW 644 0 0 0 0 022 0 0 0 0 0 0 055 915 243 193

MXN 645 0 0 0 006 017 0 0 0 0 0 0 1226 8161 055 215

MYR 718 0 0 0 044 083 0 011 0 0 0 0 707 80 939 983

NOK 1360 0 0 0 714 1199 023 006 0 0 017 0 5511 8709 628 3598

NZD 942 0 0 0 022 022 0 006 0 006 0 0 2733 8476 2198 668

RUB 278 0 0 0 0 006 0 0 0 0 0 0 006 3908 094 15

SEK 1152 0 0 0 274 268 0 0 0 006 0 0 6512 8234 732 1252

SGD 933 0 0 0 039 182 0 0 0 0 0 0 2115 9039 1629 994

THB 705 0 0 0 006 028 0 011 0 0 006 0 762 8702 663 398

TRY 722 0 0 0 008 068 0 0 0 0 0 0 1814 7333 051 1561

TWD 771 0 0 0 061 099 0 0 0 0 0 0 607 9227 872 696

USD 001 0 0 0 0 0 0 0 0 0 0 0 0 017 0 0

ZAR 714 0 0 0 006 006 0 0 0 0 0 0 1364 6897 2021 409

Averages 675 0 0 001 099 133 005 002 000 002 003 000 1783 6451 863 782

CURRENCY

NA VALUES

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

MKTRISK

TRADBAL

BDEFCIT

CENBANK

CNSMSNT

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous page data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices credit easy vs tight economic forecast growth amp uncertainty election sentiment financial system instability fiscal policy loose vs tight government anger government corruption inflation interest rate and investment flow

AUD 055 398 0 0 0 1022 4892 3247 149 028 0 077 006 7178

BRL 817 1745 144 0 011 6748 7891 5279 4373 1496 0 265 105 5748

CAD 039 061 0 0 0 558 4992 2004 171 0 0 033 006 6322

CHF 2783 3009 1115 0 116 899 6946 8769 6659 3689 127 1469 1115 9608

DKK 5966 7263 4046 012 1339 91 9467 8649 7666 59 504 6108 689 9846

EGP 5504 6967 2287 022 1536 5998 9496 8313 2965 2416 022 4697 1581 7881

EUR 0 0 0 0 0 055 447 088 0 0 0 0 0 2501

GBP 0 006 0 0 0 359 1922 718 017 0 0 0 0 7062

HKD 1375 4462 1253 0 083 7968 8112 8697 6599 3628 061 1353 1894 8796

IDR 2536 4923 464 0 127 5359 8414 6249 2961 442 0 1083 174 6122

ILS 3526 5241 078 0 05 4131 8705 7024 303 073 0 2248 1474 8576

INR 006 237 0 0 0 591 5445 2032 039 006 0 044 05 1723

JPY 028 099 0 0 0 2115 3274 2308 497 044 0 0 011 5295

KRW 1507 2838 171 0 033 5903 698 7123 2678 845 0 254 74 3733

MXN 2181 27 326 0 022 5682 8498 7245 2551 502 0 348 1325 8029

MYR 3657 4503 1039 0 055 6939 905 7492 4331 1315 0 2851 4215 7602

NOK 5917 7036 3427 023 1593 9332 9195 9086 8104 6453 428 4215 1702 9903

NZD 723 4931 906 011 155 762 8531 7885 5533 2717 05 2065 331 9249

RUB 445 3 006 0 0 1968 6459 4058 056 028 0 278 728 5114

SEK 4667 4941 2348 011 548 8519 872 8167 7798 5461 229 3214 3231 9883

SGD 3164 5119 138 0 232 8719 8625 8741 751 3821 083 1596 3777 9045

THB 3401 5273 828 0 127 4246 7096 7195 312 1022 006 2043 2258 7151

TRY 2143 535 903 0 219 5865 8422 7148 2979 2118 0 717 65 8152

TWD 386 5809 823 0 188 6218 9128 8747 5665 2424 022 2998 2833 7874

USD 0 0 0 0 0 0 099 0 0 0 0 0 0 994

ZAR 1811 3131 26 0 077 5781 8233 6477 1513 442 006 26 287 7951

Averages 2158 3321 839 003 250 4992 6886 5875 3240 1726 059 1470 1421 6975

CURRENCY

CRDTEVT

DEFAULT

ECONCFT

ECONGRW

EC_UNCR

ELECSNT

FNSYSIN

FISCLVT

GVTANGR

GVTCORR

GVTINST

INFLATN

INTRATE

INVFLOW

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index BUZZThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous two pages data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices Monetary policy loose vs tight natural disasters regime change social inequality social unrest and unemployment

AUD 6516 0 1988 7559 011 536

BRL 7427 088 6013 7405 017 4666

CAD 651 0 1679 7311 0 442

CHF 804 1253 7521 9663 337 5693

DKK 9822 391 859 9627 877 3466

EGP 9759 1508 4669 9226 034 6637

EUR 1331 0 0 3556 0 0

GBP 1165 0 05 2601 0 077

HKD 8134 558 7786 8393 331 5621

IDR 8779 066 6956 9221 011 6232

ILS 9355 022 1659 8021 0 4137

INR 5157 0 083 4815 0 977

JPY 2568 0 2236 7537 0 768

KRW 8526 42 5577 8498 033 5312

MXN 9177 061 19 8636 006 365

MYR 8785 1077 7569 947 099 7608

NOK 9652 2479 8658 9766 919 309

NZD 8597 387 7863 9094 155 5527

RUB 8299 0 256 8705 0 4074

SEK 9223 3387 8234 9268 682 422

SGD 8056 1618 8603 9034 939 7338

THB 8609 436 3517 8901 022 8067

TRY 7519 92 3291 9468 0 6068

TWD 9304 872 8034 9359 304 4296

USD 42 0 0 674 0 0

ZAR 836 166 3418 5086 0 1706

Averages 7273 740 4467 7727 184 3854

CURRENCY

MONELVT

NATDSST

REGMCHG

SOCINEQ

SOCUNRS

UNEMPLY

THOMSON REUTERS FINANCIAL AND RISK

ABOUT THE AUTHORIan MacGillivray is a Senior Research Engineer in the Thomson Reuters Corporate Research amp Development group His current work includes an entity-centric framework to discover quantify and qualify relationships and similarities an effort to find new sources of alpha for the markets and explorations of social media and behavior His research background is in probability theory and agent-based systems He has a keen interest in physics and language and has a BSc in Computer Science from Aston University He was awarded the Thomson Reuters Inventor of the Year award for a patent filed in 2011

ABOUT THOMSON REUTERSThomson Reuters is the worldrsquos leading source of intelligent information for businesses and professionals We combine industry expertise with innovative technology to deliver critical information to leading decision makers in the financial and risk legal tax and accounting intellectual property and science and media markets powered by the worldrsquos most trusted news organization With headquarters in New York and major operations in London and Eagan Minnesota Thomson Reuters employs approximately 60000 people and operates in over 100 countries Thomson Reuters shares are listed on the Toronto and New York Stock Exchanges

For more information about Thomson Reuters please go to wwwthomsonreuterscom

Original research by Thomson Reuters Research amp Development in collaboration with the Stevens Institute of Technology With great thanks to Dr Eleni Gousgounis Matthew Bombard and William Calhoun for their significant contributions to this and related projects

Results provisional pending peer-reviewed publication in a leading financial journal

Visit thomsonreuterscom

For more information contact your representative or visit us online

copy Thomson Reuters 2013 All rights reserved 100531111-13

Thomson Reuters Machine Readable NewsAsif Alam3 Times SquareNew York NY 10036USAasifalamthomsonreuterscomhttpthomsonreuterscommachine-readable-news

Thomson Reuters RampDIan MacGillivray3 Times SquareNew York NY 10036USAianmacgillivraythomsonreuterscom httplabsthomsonreuterscom

MarketPsychRichard Peterson MD179 Post Road WWestPort CT 06880USArichardmarketpsychdatacomhttpwwwmarketpsychcom

Stevens Institute of TechnologyDr Eleni GousgounisCastle Point on HudsonHoboken NJ 07030USAelenigousgounisstevenseduhttpstevensedufsc

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix A Portfolio Trading 042006 to 052013Evenly split portfolios containing all currencies with available data (average four per portfolio) for each month were rebalanced monthly based on an average of the previous monthrsquos sentiment index We then calculated the excess returns (forward premium + spot change) against the USD for the following month

In practice this would involve investing in currencies 4 5 and 6 and selling or shorting the currencies 1 2 and 3

Portfolio Number 1 2 3 4 5 6

Spot Mean 170 030 -110 -170 000 -160

SD 780 990 920 1010 870 1000

Premium Mean 180 170 170 170 120 150

SD 050 050 060 060 050 060

Excess Mean 010 140 280 340 120 310

SD 780 980 920 1000 870 1000

SR 00128 01429 03043 034 01379 031

Switching Frequency 5030 6898 8524 7741 7139 5301

Sentiment Mean -00102 -00028 00011 00057 00104 0018

HML Mean NA 138 274 333 114 300

SD NA 744 607 703 581 716

SR NA 01855 04514 04737 01962 0419

Avg Switch Freq 6772

CO_ECONGRW (Country-Level Economic Growth)

Portfolio Number 1 2 3 4 5 6

Spot Mean 020 050 -110 170 -030 -290

SD 960 930 900 1030 840 890

Premium Mean 080 130 210 210 200 130

SD 040 050 060 060 060 060

Excess Mean 060 080 320 040 230 420

SD 940 920 890 1030 840 890

SR 00638 0087 03596 00388 02738 04719

Switching Frequency 5813 7289 8765 7620 7651 6325

Sentiment Mean -00023 -00007 00001 00006 00012 00027

HML Mean NA 016 253 -026 170 356

SD NA 461 489 546 564 627

SR NA 00347 05174 -00476 03014 05678

Avg Switch Freq 7244

CO_EC_UNCR (Country-Level Economic Uncertainty)

000

100

200

300

400

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_ECONGRW

000

100

200

300

500

400

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_EC_UNCR

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Portfolio Number 1 2 3 4 5 6

Spot Mean 160 160 130 -090 -270 -300

SD 970 870 1020 940 940 990

Premium Mean 150 160 170 200 140 100

SD 060 060 060 060 060 060

Excess Mean -010 000 040 300 400 500

SD 960 870 1010 930 930 980

SR -00104 0 00396 03226 04301 04082

Switching Frequency 5030 8755 6807 7169 7410 7199

Sentiment Mean -00102 -00014 -00002 0 00001 00005

HML Mean NA 018 058 310 419 420

SD NA 718 710 690 685 758

SR NA 00251 00817 04493 06117 05541

Avg Switch Freq 7323

CO_FNSYSIN (Country-Level Financial System Instability)

Portfolio Number 1 2 3 4 5 6

Spot Mean 180 190 -310 -200 -050 -060

SD 930 780 1060 1030 910 940

Premium Mean 070 100 170 170 220 240

SD 040 040 050 080 070 060

Excess Mean -110 -090 470 370 260 300

SD 930 780 1060 1010 900 920

SR -01183 -01154 04434 03663 02889 03261

Switching Frequency 5813 4880 6687 7771 7048 6084

Sentiment Mean -00023 -00064 -00018 00016 00059 00166

HML Mean NA 013 578 477 367 402

SD NA 674 546 628 586 663

SR NA 00193 10586 07596 06263 06063

Avg Switch Freq 6235

CU_UNCERTN (Currency-Level Uncertainty)

Portfolio Number 1 2 3 4 5 6

Spot Mean 072 066 -099 188 -301 -172

SD 924 924 822 1032 841 1099

Premium Mean 145 143 135 146 149 249

SD 033 042 048 072 052 081

Excess Mean 074 078 234 -042 450 421

SD 924 915 820 1022 837 1077

SR 00798 00847 02855 -00415 05372 03913

Switching Frequency 5030 4669 6988 7982 7440 7229

Sentiment Mean -00102 -00105 -00023 00017 00047 00075

HML Mean NA 004 161 -116 376 348

SD NA 608 593 581 634 703

SR NA 00066 02715 -01997 05931 0495

Avg Switch Freq 6506

CO_TRUST (Country-Level Trust)

-100

000

100

200

300

400

500

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_FNSYSIN

-200

000

200

400

600

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CU_UNCERTN

-100

000

100

200

300

400

500

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_TRUST

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

We see here high NA values for PEGINST (currency peg instability) for currencies not involved in pegs and we can see in general that the DKK is a rarely-discussed currency

Data for indices calculated at the currency level are presented below In addition to the standard set of MarketPsych indices currency-level indices further include market forecast carry trade peg instability and market momentum

AUD 1226 0 0 0 193 217 028 121 006 006 028 0 011 6726 9089 017

BRL 3567 0 0 175 5505 6505 4009 2595 491 1585 2949 022 4274 9718 9961 4136

CAD 1475 0 0 006 458 3026 094 331 006 011 055 0 072 8818 9188 061

CHF 2904 0 011 1287 3876 7896 323 3203 58 265 1645 0 2568 8117 9393 1491

DKK 8541 0 6392 9419 9893 9959 9745 9716 8697 9538 952 6605 9639 9994 9751 9254

EGP 7114 0 2438 7909 9131 977 8531 7455 602 7623 8464 2012 8677 9989 9983 8711

EUR 878 0 0 0 0 0 0 0 0 0 0 0 0 6842 6328 0

GBP 1385 0 0 0 282 1679 028 017 0 0 011 0 017 8978 9757 011

HKD 2138 0 0 486 1552 63 1253 2921 077 552 668 0 1193 8835 7405 834

IDR 3828 0 11 2199 526 8199 4392 5298 10 2486 2757 122 2503 9796 9978 3326

ILS 5976 0 97 5432 8268 9518 801 7057 3862 634 6463 116 6452 9961 9994 616

INR 1665 0 0 055 768 3429 248 342 022 055 182 0 094 9845 9729 199

JPY 942 0 0 0 006 1137 011 011 0 0 0 0 0 328 9685 0

KRW 1825 0 0 088 834 423 872 414 011 177 188 0 409 9845 9983 32

MXN 3662 0 017 1888 323 7985 4583 4506 1022 2849 2844 039 3153 9779 9923 3114

MYR 3355 0 05 1088 4851 7796 4254 4293 602 1646 1983 055 1923 9934 9718 2133

NOK 7869 0 4226 8612 948 9886 9417 9377 8235 8766 8761 4101 8932 996 9983 8298

NZD 1977 0 0 171 1347 5831 1033 1999 055 309 42 0 718 7742 9757 271

RUB 5635 0 645 483 7788 9383 6809 6976 3613 5442 5753 1201 5931 9867 9939 6348

SEK 7328 0 2644 7457 9178 9793 877 9106 7328 8178 8317 2789 8379 9922 9939 8122

SGD 2086 0 0 204 111 6046 795 2391 083 342 502 0 806 9304 9067 646

THB 2631 0 017 563 3462 6455 2579 1911 16 707 1436 033 1198 995 9459 153

TRY 4276 0 11 2489 5679 8793 5764 6861 1519 2895 2228 135 3637 9637 100 4388

TWD 2751 0 0 629 2082 7018 2214 3981 409 1099 1441 0 1888 9398 9431 1679

USD 488 0 0 0 0 028 0 0 0 0 0 0 0 4694 2595 0

ZAR 1557 0 006 094 69 3026 204 26 017 099 171 006 348 8647 9564 221

Averages 3349 0 678 2179 3651 5995 3341 3505 1685 2345 2569 703 2801 8830 9215 2741

CURRENCY

N

A VALS

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

PRICEUP

MKTFCST

CARYTRD

PEGINST

MKTMMTM

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

Data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below In addition to the standard set of MarketPsych indices country-level indices further include market risk trade balance budget deficit central bank and consumer sentiment

AUD 274 0 0 0 0 0 0 0 0 0 0 0 061 3793 243 006

BRL 490 0 0 0 011 011 0 0 0 0 0 0 11 677 05 403

CAD 188 0 0 0 0 0 0 0 0 0 0 0 072 2457 293 0

CHF 929 0 0 0 055 099 0 0 0 0 0 0 4147 9094 0 541

DKK 1755 0 0 024 1143 936 089 018 0 036 041 0 7636 9034 529 2079

EGP 1126 0 0 0 146 353 017 0 0 0 006 0 4888 7534 633 3318

EUR 002 0 0 0 0 0 0 0 0 0 0 0 0 033 0 0

GBP 127 0 0 0 0 0 0 0 0 0 0 0 061 1844 0 0

HKD 984 0 0 0 033 022 0 0 0 0 0 0 2419 873 2927 624

IDR 537 0 0 0 017 028 006 0 0 0 0 0 32 6282 359 105

ILS 941 0 0 0 0 0 0 0 0 0 0 0 3223 7203 2489 1194

INR 273 0 0 0 0 0 0 0 0 0 0 0 0 4075 017 0

JPY 335 0 0 0 0 0 0 0 0 0 0 0 006 5014 0 0

KRW 644 0 0 0 0 022 0 0 0 0 0 0 055 915 243 193

MXN 645 0 0 0 006 017 0 0 0 0 0 0 1226 8161 055 215

MYR 718 0 0 0 044 083 0 011 0 0 0 0 707 80 939 983

NOK 1360 0 0 0 714 1199 023 006 0 0 017 0 5511 8709 628 3598

NZD 942 0 0 0 022 022 0 006 0 006 0 0 2733 8476 2198 668

RUB 278 0 0 0 0 006 0 0 0 0 0 0 006 3908 094 15

SEK 1152 0 0 0 274 268 0 0 0 006 0 0 6512 8234 732 1252

SGD 933 0 0 0 039 182 0 0 0 0 0 0 2115 9039 1629 994

THB 705 0 0 0 006 028 0 011 0 0 006 0 762 8702 663 398

TRY 722 0 0 0 008 068 0 0 0 0 0 0 1814 7333 051 1561

TWD 771 0 0 0 061 099 0 0 0 0 0 0 607 9227 872 696

USD 001 0 0 0 0 0 0 0 0 0 0 0 0 017 0 0

ZAR 714 0 0 0 006 006 0 0 0 0 0 0 1364 6897 2021 409

Averages 675 0 0 001 099 133 005 002 000 002 003 000 1783 6451 863 782

CURRENCY

NA VALUES

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

MKTRISK

TRADBAL

BDEFCIT

CENBANK

CNSMSNT

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous page data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices credit easy vs tight economic forecast growth amp uncertainty election sentiment financial system instability fiscal policy loose vs tight government anger government corruption inflation interest rate and investment flow

AUD 055 398 0 0 0 1022 4892 3247 149 028 0 077 006 7178

BRL 817 1745 144 0 011 6748 7891 5279 4373 1496 0 265 105 5748

CAD 039 061 0 0 0 558 4992 2004 171 0 0 033 006 6322

CHF 2783 3009 1115 0 116 899 6946 8769 6659 3689 127 1469 1115 9608

DKK 5966 7263 4046 012 1339 91 9467 8649 7666 59 504 6108 689 9846

EGP 5504 6967 2287 022 1536 5998 9496 8313 2965 2416 022 4697 1581 7881

EUR 0 0 0 0 0 055 447 088 0 0 0 0 0 2501

GBP 0 006 0 0 0 359 1922 718 017 0 0 0 0 7062

HKD 1375 4462 1253 0 083 7968 8112 8697 6599 3628 061 1353 1894 8796

IDR 2536 4923 464 0 127 5359 8414 6249 2961 442 0 1083 174 6122

ILS 3526 5241 078 0 05 4131 8705 7024 303 073 0 2248 1474 8576

INR 006 237 0 0 0 591 5445 2032 039 006 0 044 05 1723

JPY 028 099 0 0 0 2115 3274 2308 497 044 0 0 011 5295

KRW 1507 2838 171 0 033 5903 698 7123 2678 845 0 254 74 3733

MXN 2181 27 326 0 022 5682 8498 7245 2551 502 0 348 1325 8029

MYR 3657 4503 1039 0 055 6939 905 7492 4331 1315 0 2851 4215 7602

NOK 5917 7036 3427 023 1593 9332 9195 9086 8104 6453 428 4215 1702 9903

NZD 723 4931 906 011 155 762 8531 7885 5533 2717 05 2065 331 9249

RUB 445 3 006 0 0 1968 6459 4058 056 028 0 278 728 5114

SEK 4667 4941 2348 011 548 8519 872 8167 7798 5461 229 3214 3231 9883

SGD 3164 5119 138 0 232 8719 8625 8741 751 3821 083 1596 3777 9045

THB 3401 5273 828 0 127 4246 7096 7195 312 1022 006 2043 2258 7151

TRY 2143 535 903 0 219 5865 8422 7148 2979 2118 0 717 65 8152

TWD 386 5809 823 0 188 6218 9128 8747 5665 2424 022 2998 2833 7874

USD 0 0 0 0 0 0 099 0 0 0 0 0 0 994

ZAR 1811 3131 26 0 077 5781 8233 6477 1513 442 006 26 287 7951

Averages 2158 3321 839 003 250 4992 6886 5875 3240 1726 059 1470 1421 6975

CURRENCY

CRDTEVT

DEFAULT

ECONCFT

ECONGRW

EC_UNCR

ELECSNT

FNSYSIN

FISCLVT

GVTANGR

GVTCORR

GVTINST

INFLATN

INTRATE

INVFLOW

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index BUZZThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous two pages data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices Monetary policy loose vs tight natural disasters regime change social inequality social unrest and unemployment

AUD 6516 0 1988 7559 011 536

BRL 7427 088 6013 7405 017 4666

CAD 651 0 1679 7311 0 442

CHF 804 1253 7521 9663 337 5693

DKK 9822 391 859 9627 877 3466

EGP 9759 1508 4669 9226 034 6637

EUR 1331 0 0 3556 0 0

GBP 1165 0 05 2601 0 077

HKD 8134 558 7786 8393 331 5621

IDR 8779 066 6956 9221 011 6232

ILS 9355 022 1659 8021 0 4137

INR 5157 0 083 4815 0 977

JPY 2568 0 2236 7537 0 768

KRW 8526 42 5577 8498 033 5312

MXN 9177 061 19 8636 006 365

MYR 8785 1077 7569 947 099 7608

NOK 9652 2479 8658 9766 919 309

NZD 8597 387 7863 9094 155 5527

RUB 8299 0 256 8705 0 4074

SEK 9223 3387 8234 9268 682 422

SGD 8056 1618 8603 9034 939 7338

THB 8609 436 3517 8901 022 8067

TRY 7519 92 3291 9468 0 6068

TWD 9304 872 8034 9359 304 4296

USD 42 0 0 674 0 0

ZAR 836 166 3418 5086 0 1706

Averages 7273 740 4467 7727 184 3854

CURRENCY

MONELVT

NATDSST

REGMCHG

SOCINEQ

SOCUNRS

UNEMPLY

THOMSON REUTERS FINANCIAL AND RISK

ABOUT THE AUTHORIan MacGillivray is a Senior Research Engineer in the Thomson Reuters Corporate Research amp Development group His current work includes an entity-centric framework to discover quantify and qualify relationships and similarities an effort to find new sources of alpha for the markets and explorations of social media and behavior His research background is in probability theory and agent-based systems He has a keen interest in physics and language and has a BSc in Computer Science from Aston University He was awarded the Thomson Reuters Inventor of the Year award for a patent filed in 2011

ABOUT THOMSON REUTERSThomson Reuters is the worldrsquos leading source of intelligent information for businesses and professionals We combine industry expertise with innovative technology to deliver critical information to leading decision makers in the financial and risk legal tax and accounting intellectual property and science and media markets powered by the worldrsquos most trusted news organization With headquarters in New York and major operations in London and Eagan Minnesota Thomson Reuters employs approximately 60000 people and operates in over 100 countries Thomson Reuters shares are listed on the Toronto and New York Stock Exchanges

For more information about Thomson Reuters please go to wwwthomsonreuterscom

Original research by Thomson Reuters Research amp Development in collaboration with the Stevens Institute of Technology With great thanks to Dr Eleni Gousgounis Matthew Bombard and William Calhoun for their significant contributions to this and related projects

Results provisional pending peer-reviewed publication in a leading financial journal

Visit thomsonreuterscom

For more information contact your representative or visit us online

copy Thomson Reuters 2013 All rights reserved 100531111-13

Thomson Reuters Machine Readable NewsAsif Alam3 Times SquareNew York NY 10036USAasifalamthomsonreuterscomhttpthomsonreuterscommachine-readable-news

Thomson Reuters RampDIan MacGillivray3 Times SquareNew York NY 10036USAianmacgillivraythomsonreuterscom httplabsthomsonreuterscom

MarketPsychRichard Peterson MD179 Post Road WWestPort CT 06880USArichardmarketpsychdatacomhttpwwwmarketpsychcom

Stevens Institute of TechnologyDr Eleni GousgounisCastle Point on HudsonHoboken NJ 07030USAelenigousgounisstevenseduhttpstevensedufsc

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Portfolio Number 1 2 3 4 5 6

Spot Mean 160 160 130 -090 -270 -300

SD 970 870 1020 940 940 990

Premium Mean 150 160 170 200 140 100

SD 060 060 060 060 060 060

Excess Mean -010 000 040 300 400 500

SD 960 870 1010 930 930 980

SR -00104 0 00396 03226 04301 04082

Switching Frequency 5030 8755 6807 7169 7410 7199

Sentiment Mean -00102 -00014 -00002 0 00001 00005

HML Mean NA 018 058 310 419 420

SD NA 718 710 690 685 758

SR NA 00251 00817 04493 06117 05541

Avg Switch Freq 7323

CO_FNSYSIN (Country-Level Financial System Instability)

Portfolio Number 1 2 3 4 5 6

Spot Mean 180 190 -310 -200 -050 -060

SD 930 780 1060 1030 910 940

Premium Mean 070 100 170 170 220 240

SD 040 040 050 080 070 060

Excess Mean -110 -090 470 370 260 300

SD 930 780 1060 1010 900 920

SR -01183 -01154 04434 03663 02889 03261

Switching Frequency 5813 4880 6687 7771 7048 6084

Sentiment Mean -00023 -00064 -00018 00016 00059 00166

HML Mean NA 013 578 477 367 402

SD NA 674 546 628 586 663

SR NA 00193 10586 07596 06263 06063

Avg Switch Freq 6235

CU_UNCERTN (Currency-Level Uncertainty)

Portfolio Number 1 2 3 4 5 6

Spot Mean 072 066 -099 188 -301 -172

SD 924 924 822 1032 841 1099

Premium Mean 145 143 135 146 149 249

SD 033 042 048 072 052 081

Excess Mean 074 078 234 -042 450 421

SD 924 915 820 1022 837 1077

SR 00798 00847 02855 -00415 05372 03913

Switching Frequency 5030 4669 6988 7982 7440 7229

Sentiment Mean -00102 -00105 -00023 00017 00047 00075

HML Mean NA 004 161 -116 376 348

SD NA 608 593 581 634 703

SR NA 00066 02715 -01997 05931 0495

Avg Switch Freq 6506

CO_TRUST (Country-Level Trust)

-100

000

100

200

300

400

500

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_FNSYSIN

-200

000

200

400

600

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CU_UNCERTN

-100

000

100

200

300

400

500

1 2 3 4 5 6

Exce

ss R

etur

ns

Portfolio

Rebalancing Monthly on CO_TRUST

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

We see here high NA values for PEGINST (currency peg instability) for currencies not involved in pegs and we can see in general that the DKK is a rarely-discussed currency

Data for indices calculated at the currency level are presented below In addition to the standard set of MarketPsych indices currency-level indices further include market forecast carry trade peg instability and market momentum

AUD 1226 0 0 0 193 217 028 121 006 006 028 0 011 6726 9089 017

BRL 3567 0 0 175 5505 6505 4009 2595 491 1585 2949 022 4274 9718 9961 4136

CAD 1475 0 0 006 458 3026 094 331 006 011 055 0 072 8818 9188 061

CHF 2904 0 011 1287 3876 7896 323 3203 58 265 1645 0 2568 8117 9393 1491

DKK 8541 0 6392 9419 9893 9959 9745 9716 8697 9538 952 6605 9639 9994 9751 9254

EGP 7114 0 2438 7909 9131 977 8531 7455 602 7623 8464 2012 8677 9989 9983 8711

EUR 878 0 0 0 0 0 0 0 0 0 0 0 0 6842 6328 0

GBP 1385 0 0 0 282 1679 028 017 0 0 011 0 017 8978 9757 011

HKD 2138 0 0 486 1552 63 1253 2921 077 552 668 0 1193 8835 7405 834

IDR 3828 0 11 2199 526 8199 4392 5298 10 2486 2757 122 2503 9796 9978 3326

ILS 5976 0 97 5432 8268 9518 801 7057 3862 634 6463 116 6452 9961 9994 616

INR 1665 0 0 055 768 3429 248 342 022 055 182 0 094 9845 9729 199

JPY 942 0 0 0 006 1137 011 011 0 0 0 0 0 328 9685 0

KRW 1825 0 0 088 834 423 872 414 011 177 188 0 409 9845 9983 32

MXN 3662 0 017 1888 323 7985 4583 4506 1022 2849 2844 039 3153 9779 9923 3114

MYR 3355 0 05 1088 4851 7796 4254 4293 602 1646 1983 055 1923 9934 9718 2133

NOK 7869 0 4226 8612 948 9886 9417 9377 8235 8766 8761 4101 8932 996 9983 8298

NZD 1977 0 0 171 1347 5831 1033 1999 055 309 42 0 718 7742 9757 271

RUB 5635 0 645 483 7788 9383 6809 6976 3613 5442 5753 1201 5931 9867 9939 6348

SEK 7328 0 2644 7457 9178 9793 877 9106 7328 8178 8317 2789 8379 9922 9939 8122

SGD 2086 0 0 204 111 6046 795 2391 083 342 502 0 806 9304 9067 646

THB 2631 0 017 563 3462 6455 2579 1911 16 707 1436 033 1198 995 9459 153

TRY 4276 0 11 2489 5679 8793 5764 6861 1519 2895 2228 135 3637 9637 100 4388

TWD 2751 0 0 629 2082 7018 2214 3981 409 1099 1441 0 1888 9398 9431 1679

USD 488 0 0 0 0 028 0 0 0 0 0 0 0 4694 2595 0

ZAR 1557 0 006 094 69 3026 204 26 017 099 171 006 348 8647 9564 221

Averages 3349 0 678 2179 3651 5995 3341 3505 1685 2345 2569 703 2801 8830 9215 2741

CURRENCY

N

A VALS

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

PRICEUP

MKTFCST

CARYTRD

PEGINST

MKTMMTM

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

Data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below In addition to the standard set of MarketPsych indices country-level indices further include market risk trade balance budget deficit central bank and consumer sentiment

AUD 274 0 0 0 0 0 0 0 0 0 0 0 061 3793 243 006

BRL 490 0 0 0 011 011 0 0 0 0 0 0 11 677 05 403

CAD 188 0 0 0 0 0 0 0 0 0 0 0 072 2457 293 0

CHF 929 0 0 0 055 099 0 0 0 0 0 0 4147 9094 0 541

DKK 1755 0 0 024 1143 936 089 018 0 036 041 0 7636 9034 529 2079

EGP 1126 0 0 0 146 353 017 0 0 0 006 0 4888 7534 633 3318

EUR 002 0 0 0 0 0 0 0 0 0 0 0 0 033 0 0

GBP 127 0 0 0 0 0 0 0 0 0 0 0 061 1844 0 0

HKD 984 0 0 0 033 022 0 0 0 0 0 0 2419 873 2927 624

IDR 537 0 0 0 017 028 006 0 0 0 0 0 32 6282 359 105

ILS 941 0 0 0 0 0 0 0 0 0 0 0 3223 7203 2489 1194

INR 273 0 0 0 0 0 0 0 0 0 0 0 0 4075 017 0

JPY 335 0 0 0 0 0 0 0 0 0 0 0 006 5014 0 0

KRW 644 0 0 0 0 022 0 0 0 0 0 0 055 915 243 193

MXN 645 0 0 0 006 017 0 0 0 0 0 0 1226 8161 055 215

MYR 718 0 0 0 044 083 0 011 0 0 0 0 707 80 939 983

NOK 1360 0 0 0 714 1199 023 006 0 0 017 0 5511 8709 628 3598

NZD 942 0 0 0 022 022 0 006 0 006 0 0 2733 8476 2198 668

RUB 278 0 0 0 0 006 0 0 0 0 0 0 006 3908 094 15

SEK 1152 0 0 0 274 268 0 0 0 006 0 0 6512 8234 732 1252

SGD 933 0 0 0 039 182 0 0 0 0 0 0 2115 9039 1629 994

THB 705 0 0 0 006 028 0 011 0 0 006 0 762 8702 663 398

TRY 722 0 0 0 008 068 0 0 0 0 0 0 1814 7333 051 1561

TWD 771 0 0 0 061 099 0 0 0 0 0 0 607 9227 872 696

USD 001 0 0 0 0 0 0 0 0 0 0 0 0 017 0 0

ZAR 714 0 0 0 006 006 0 0 0 0 0 0 1364 6897 2021 409

Averages 675 0 0 001 099 133 005 002 000 002 003 000 1783 6451 863 782

CURRENCY

NA VALUES

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

MKTRISK

TRADBAL

BDEFCIT

CENBANK

CNSMSNT

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous page data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices credit easy vs tight economic forecast growth amp uncertainty election sentiment financial system instability fiscal policy loose vs tight government anger government corruption inflation interest rate and investment flow

AUD 055 398 0 0 0 1022 4892 3247 149 028 0 077 006 7178

BRL 817 1745 144 0 011 6748 7891 5279 4373 1496 0 265 105 5748

CAD 039 061 0 0 0 558 4992 2004 171 0 0 033 006 6322

CHF 2783 3009 1115 0 116 899 6946 8769 6659 3689 127 1469 1115 9608

DKK 5966 7263 4046 012 1339 91 9467 8649 7666 59 504 6108 689 9846

EGP 5504 6967 2287 022 1536 5998 9496 8313 2965 2416 022 4697 1581 7881

EUR 0 0 0 0 0 055 447 088 0 0 0 0 0 2501

GBP 0 006 0 0 0 359 1922 718 017 0 0 0 0 7062

HKD 1375 4462 1253 0 083 7968 8112 8697 6599 3628 061 1353 1894 8796

IDR 2536 4923 464 0 127 5359 8414 6249 2961 442 0 1083 174 6122

ILS 3526 5241 078 0 05 4131 8705 7024 303 073 0 2248 1474 8576

INR 006 237 0 0 0 591 5445 2032 039 006 0 044 05 1723

JPY 028 099 0 0 0 2115 3274 2308 497 044 0 0 011 5295

KRW 1507 2838 171 0 033 5903 698 7123 2678 845 0 254 74 3733

MXN 2181 27 326 0 022 5682 8498 7245 2551 502 0 348 1325 8029

MYR 3657 4503 1039 0 055 6939 905 7492 4331 1315 0 2851 4215 7602

NOK 5917 7036 3427 023 1593 9332 9195 9086 8104 6453 428 4215 1702 9903

NZD 723 4931 906 011 155 762 8531 7885 5533 2717 05 2065 331 9249

RUB 445 3 006 0 0 1968 6459 4058 056 028 0 278 728 5114

SEK 4667 4941 2348 011 548 8519 872 8167 7798 5461 229 3214 3231 9883

SGD 3164 5119 138 0 232 8719 8625 8741 751 3821 083 1596 3777 9045

THB 3401 5273 828 0 127 4246 7096 7195 312 1022 006 2043 2258 7151

TRY 2143 535 903 0 219 5865 8422 7148 2979 2118 0 717 65 8152

TWD 386 5809 823 0 188 6218 9128 8747 5665 2424 022 2998 2833 7874

USD 0 0 0 0 0 0 099 0 0 0 0 0 0 994

ZAR 1811 3131 26 0 077 5781 8233 6477 1513 442 006 26 287 7951

Averages 2158 3321 839 003 250 4992 6886 5875 3240 1726 059 1470 1421 6975

CURRENCY

CRDTEVT

DEFAULT

ECONCFT

ECONGRW

EC_UNCR

ELECSNT

FNSYSIN

FISCLVT

GVTANGR

GVTCORR

GVTINST

INFLATN

INTRATE

INVFLOW

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index BUZZThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous two pages data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices Monetary policy loose vs tight natural disasters regime change social inequality social unrest and unemployment

AUD 6516 0 1988 7559 011 536

BRL 7427 088 6013 7405 017 4666

CAD 651 0 1679 7311 0 442

CHF 804 1253 7521 9663 337 5693

DKK 9822 391 859 9627 877 3466

EGP 9759 1508 4669 9226 034 6637

EUR 1331 0 0 3556 0 0

GBP 1165 0 05 2601 0 077

HKD 8134 558 7786 8393 331 5621

IDR 8779 066 6956 9221 011 6232

ILS 9355 022 1659 8021 0 4137

INR 5157 0 083 4815 0 977

JPY 2568 0 2236 7537 0 768

KRW 8526 42 5577 8498 033 5312

MXN 9177 061 19 8636 006 365

MYR 8785 1077 7569 947 099 7608

NOK 9652 2479 8658 9766 919 309

NZD 8597 387 7863 9094 155 5527

RUB 8299 0 256 8705 0 4074

SEK 9223 3387 8234 9268 682 422

SGD 8056 1618 8603 9034 939 7338

THB 8609 436 3517 8901 022 8067

TRY 7519 92 3291 9468 0 6068

TWD 9304 872 8034 9359 304 4296

USD 42 0 0 674 0 0

ZAR 836 166 3418 5086 0 1706

Averages 7273 740 4467 7727 184 3854

CURRENCY

MONELVT

NATDSST

REGMCHG

SOCINEQ

SOCUNRS

UNEMPLY

THOMSON REUTERS FINANCIAL AND RISK

ABOUT THE AUTHORIan MacGillivray is a Senior Research Engineer in the Thomson Reuters Corporate Research amp Development group His current work includes an entity-centric framework to discover quantify and qualify relationships and similarities an effort to find new sources of alpha for the markets and explorations of social media and behavior His research background is in probability theory and agent-based systems He has a keen interest in physics and language and has a BSc in Computer Science from Aston University He was awarded the Thomson Reuters Inventor of the Year award for a patent filed in 2011

ABOUT THOMSON REUTERSThomson Reuters is the worldrsquos leading source of intelligent information for businesses and professionals We combine industry expertise with innovative technology to deliver critical information to leading decision makers in the financial and risk legal tax and accounting intellectual property and science and media markets powered by the worldrsquos most trusted news organization With headquarters in New York and major operations in London and Eagan Minnesota Thomson Reuters employs approximately 60000 people and operates in over 100 countries Thomson Reuters shares are listed on the Toronto and New York Stock Exchanges

For more information about Thomson Reuters please go to wwwthomsonreuterscom

Original research by Thomson Reuters Research amp Development in collaboration with the Stevens Institute of Technology With great thanks to Dr Eleni Gousgounis Matthew Bombard and William Calhoun for their significant contributions to this and related projects

Results provisional pending peer-reviewed publication in a leading financial journal

Visit thomsonreuterscom

For more information contact your representative or visit us online

copy Thomson Reuters 2013 All rights reserved 100531111-13

Thomson Reuters Machine Readable NewsAsif Alam3 Times SquareNew York NY 10036USAasifalamthomsonreuterscomhttpthomsonreuterscommachine-readable-news

Thomson Reuters RampDIan MacGillivray3 Times SquareNew York NY 10036USAianmacgillivraythomsonreuterscom httplabsthomsonreuterscom

MarketPsychRichard Peterson MD179 Post Road WWestPort CT 06880USArichardmarketpsychdatacomhttpwwwmarketpsychcom

Stevens Institute of TechnologyDr Eleni GousgounisCastle Point on HudsonHoboken NJ 07030USAelenigousgounisstevenseduhttpstevensedufsc

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

We see here high NA values for PEGINST (currency peg instability) for currencies not involved in pegs and we can see in general that the DKK is a rarely-discussed currency

Data for indices calculated at the currency level are presented below In addition to the standard set of MarketPsych indices currency-level indices further include market forecast carry trade peg instability and market momentum

AUD 1226 0 0 0 193 217 028 121 006 006 028 0 011 6726 9089 017

BRL 3567 0 0 175 5505 6505 4009 2595 491 1585 2949 022 4274 9718 9961 4136

CAD 1475 0 0 006 458 3026 094 331 006 011 055 0 072 8818 9188 061

CHF 2904 0 011 1287 3876 7896 323 3203 58 265 1645 0 2568 8117 9393 1491

DKK 8541 0 6392 9419 9893 9959 9745 9716 8697 9538 952 6605 9639 9994 9751 9254

EGP 7114 0 2438 7909 9131 977 8531 7455 602 7623 8464 2012 8677 9989 9983 8711

EUR 878 0 0 0 0 0 0 0 0 0 0 0 0 6842 6328 0

GBP 1385 0 0 0 282 1679 028 017 0 0 011 0 017 8978 9757 011

HKD 2138 0 0 486 1552 63 1253 2921 077 552 668 0 1193 8835 7405 834

IDR 3828 0 11 2199 526 8199 4392 5298 10 2486 2757 122 2503 9796 9978 3326

ILS 5976 0 97 5432 8268 9518 801 7057 3862 634 6463 116 6452 9961 9994 616

INR 1665 0 0 055 768 3429 248 342 022 055 182 0 094 9845 9729 199

JPY 942 0 0 0 006 1137 011 011 0 0 0 0 0 328 9685 0

KRW 1825 0 0 088 834 423 872 414 011 177 188 0 409 9845 9983 32

MXN 3662 0 017 1888 323 7985 4583 4506 1022 2849 2844 039 3153 9779 9923 3114

MYR 3355 0 05 1088 4851 7796 4254 4293 602 1646 1983 055 1923 9934 9718 2133

NOK 7869 0 4226 8612 948 9886 9417 9377 8235 8766 8761 4101 8932 996 9983 8298

NZD 1977 0 0 171 1347 5831 1033 1999 055 309 42 0 718 7742 9757 271

RUB 5635 0 645 483 7788 9383 6809 6976 3613 5442 5753 1201 5931 9867 9939 6348

SEK 7328 0 2644 7457 9178 9793 877 9106 7328 8178 8317 2789 8379 9922 9939 8122

SGD 2086 0 0 204 111 6046 795 2391 083 342 502 0 806 9304 9067 646

THB 2631 0 017 563 3462 6455 2579 1911 16 707 1436 033 1198 995 9459 153

TRY 4276 0 11 2489 5679 8793 5764 6861 1519 2895 2228 135 3637 9637 100 4388

TWD 2751 0 0 629 2082 7018 2214 3981 409 1099 1441 0 1888 9398 9431 1679

USD 488 0 0 0 0 028 0 0 0 0 0 0 0 4694 2595 0

ZAR 1557 0 006 094 69 3026 204 26 017 099 171 006 348 8647 9564 221

Averages 3349 0 678 2179 3651 5995 3341 3505 1685 2345 2569 703 2801 8830 9215 2741

CURRENCY

N

A VALS

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

PRICEUP

MKTFCST

CARYTRD

PEGINST

MKTMMTM

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

Data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below In addition to the standard set of MarketPsych indices country-level indices further include market risk trade balance budget deficit central bank and consumer sentiment

AUD 274 0 0 0 0 0 0 0 0 0 0 0 061 3793 243 006

BRL 490 0 0 0 011 011 0 0 0 0 0 0 11 677 05 403

CAD 188 0 0 0 0 0 0 0 0 0 0 0 072 2457 293 0

CHF 929 0 0 0 055 099 0 0 0 0 0 0 4147 9094 0 541

DKK 1755 0 0 024 1143 936 089 018 0 036 041 0 7636 9034 529 2079

EGP 1126 0 0 0 146 353 017 0 0 0 006 0 4888 7534 633 3318

EUR 002 0 0 0 0 0 0 0 0 0 0 0 0 033 0 0

GBP 127 0 0 0 0 0 0 0 0 0 0 0 061 1844 0 0

HKD 984 0 0 0 033 022 0 0 0 0 0 0 2419 873 2927 624

IDR 537 0 0 0 017 028 006 0 0 0 0 0 32 6282 359 105

ILS 941 0 0 0 0 0 0 0 0 0 0 0 3223 7203 2489 1194

INR 273 0 0 0 0 0 0 0 0 0 0 0 0 4075 017 0

JPY 335 0 0 0 0 0 0 0 0 0 0 0 006 5014 0 0

KRW 644 0 0 0 0 022 0 0 0 0 0 0 055 915 243 193

MXN 645 0 0 0 006 017 0 0 0 0 0 0 1226 8161 055 215

MYR 718 0 0 0 044 083 0 011 0 0 0 0 707 80 939 983

NOK 1360 0 0 0 714 1199 023 006 0 0 017 0 5511 8709 628 3598

NZD 942 0 0 0 022 022 0 006 0 006 0 0 2733 8476 2198 668

RUB 278 0 0 0 0 006 0 0 0 0 0 0 006 3908 094 15

SEK 1152 0 0 0 274 268 0 0 0 006 0 0 6512 8234 732 1252

SGD 933 0 0 0 039 182 0 0 0 0 0 0 2115 9039 1629 994

THB 705 0 0 0 006 028 0 011 0 0 006 0 762 8702 663 398

TRY 722 0 0 0 008 068 0 0 0 0 0 0 1814 7333 051 1561

TWD 771 0 0 0 061 099 0 0 0 0 0 0 607 9227 872 696

USD 001 0 0 0 0 0 0 0 0 0 0 0 0 017 0 0

ZAR 714 0 0 0 006 006 0 0 0 0 0 0 1364 6897 2021 409

Averages 675 0 0 001 099 133 005 002 000 002 003 000 1783 6451 863 782

CURRENCY

NA VALUES

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

MKTRISK

TRADBAL

BDEFCIT

CENBANK

CNSMSNT

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous page data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices credit easy vs tight economic forecast growth amp uncertainty election sentiment financial system instability fiscal policy loose vs tight government anger government corruption inflation interest rate and investment flow

AUD 055 398 0 0 0 1022 4892 3247 149 028 0 077 006 7178

BRL 817 1745 144 0 011 6748 7891 5279 4373 1496 0 265 105 5748

CAD 039 061 0 0 0 558 4992 2004 171 0 0 033 006 6322

CHF 2783 3009 1115 0 116 899 6946 8769 6659 3689 127 1469 1115 9608

DKK 5966 7263 4046 012 1339 91 9467 8649 7666 59 504 6108 689 9846

EGP 5504 6967 2287 022 1536 5998 9496 8313 2965 2416 022 4697 1581 7881

EUR 0 0 0 0 0 055 447 088 0 0 0 0 0 2501

GBP 0 006 0 0 0 359 1922 718 017 0 0 0 0 7062

HKD 1375 4462 1253 0 083 7968 8112 8697 6599 3628 061 1353 1894 8796

IDR 2536 4923 464 0 127 5359 8414 6249 2961 442 0 1083 174 6122

ILS 3526 5241 078 0 05 4131 8705 7024 303 073 0 2248 1474 8576

INR 006 237 0 0 0 591 5445 2032 039 006 0 044 05 1723

JPY 028 099 0 0 0 2115 3274 2308 497 044 0 0 011 5295

KRW 1507 2838 171 0 033 5903 698 7123 2678 845 0 254 74 3733

MXN 2181 27 326 0 022 5682 8498 7245 2551 502 0 348 1325 8029

MYR 3657 4503 1039 0 055 6939 905 7492 4331 1315 0 2851 4215 7602

NOK 5917 7036 3427 023 1593 9332 9195 9086 8104 6453 428 4215 1702 9903

NZD 723 4931 906 011 155 762 8531 7885 5533 2717 05 2065 331 9249

RUB 445 3 006 0 0 1968 6459 4058 056 028 0 278 728 5114

SEK 4667 4941 2348 011 548 8519 872 8167 7798 5461 229 3214 3231 9883

SGD 3164 5119 138 0 232 8719 8625 8741 751 3821 083 1596 3777 9045

THB 3401 5273 828 0 127 4246 7096 7195 312 1022 006 2043 2258 7151

TRY 2143 535 903 0 219 5865 8422 7148 2979 2118 0 717 65 8152

TWD 386 5809 823 0 188 6218 9128 8747 5665 2424 022 2998 2833 7874

USD 0 0 0 0 0 0 099 0 0 0 0 0 0 994

ZAR 1811 3131 26 0 077 5781 8233 6477 1513 442 006 26 287 7951

Averages 2158 3321 839 003 250 4992 6886 5875 3240 1726 059 1470 1421 6975

CURRENCY

CRDTEVT

DEFAULT

ECONCFT

ECONGRW

EC_UNCR

ELECSNT

FNSYSIN

FISCLVT

GVTANGR

GVTCORR

GVTINST

INFLATN

INTRATE

INVFLOW

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index BUZZThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous two pages data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices Monetary policy loose vs tight natural disasters regime change social inequality social unrest and unemployment

AUD 6516 0 1988 7559 011 536

BRL 7427 088 6013 7405 017 4666

CAD 651 0 1679 7311 0 442

CHF 804 1253 7521 9663 337 5693

DKK 9822 391 859 9627 877 3466

EGP 9759 1508 4669 9226 034 6637

EUR 1331 0 0 3556 0 0

GBP 1165 0 05 2601 0 077

HKD 8134 558 7786 8393 331 5621

IDR 8779 066 6956 9221 011 6232

ILS 9355 022 1659 8021 0 4137

INR 5157 0 083 4815 0 977

JPY 2568 0 2236 7537 0 768

KRW 8526 42 5577 8498 033 5312

MXN 9177 061 19 8636 006 365

MYR 8785 1077 7569 947 099 7608

NOK 9652 2479 8658 9766 919 309

NZD 8597 387 7863 9094 155 5527

RUB 8299 0 256 8705 0 4074

SEK 9223 3387 8234 9268 682 422

SGD 8056 1618 8603 9034 939 7338

THB 8609 436 3517 8901 022 8067

TRY 7519 92 3291 9468 0 6068

TWD 9304 872 8034 9359 304 4296

USD 42 0 0 674 0 0

ZAR 836 166 3418 5086 0 1706

Averages 7273 740 4467 7727 184 3854

CURRENCY

MONELVT

NATDSST

REGMCHG

SOCINEQ

SOCUNRS

UNEMPLY

THOMSON REUTERS FINANCIAL AND RISK

ABOUT THE AUTHORIan MacGillivray is a Senior Research Engineer in the Thomson Reuters Corporate Research amp Development group His current work includes an entity-centric framework to discover quantify and qualify relationships and similarities an effort to find new sources of alpha for the markets and explorations of social media and behavior His research background is in probability theory and agent-based systems He has a keen interest in physics and language and has a BSc in Computer Science from Aston University He was awarded the Thomson Reuters Inventor of the Year award for a patent filed in 2011

ABOUT THOMSON REUTERSThomson Reuters is the worldrsquos leading source of intelligent information for businesses and professionals We combine industry expertise with innovative technology to deliver critical information to leading decision makers in the financial and risk legal tax and accounting intellectual property and science and media markets powered by the worldrsquos most trusted news organization With headquarters in New York and major operations in London and Eagan Minnesota Thomson Reuters employs approximately 60000 people and operates in over 100 countries Thomson Reuters shares are listed on the Toronto and New York Stock Exchanges

For more information about Thomson Reuters please go to wwwthomsonreuterscom

Original research by Thomson Reuters Research amp Development in collaboration with the Stevens Institute of Technology With great thanks to Dr Eleni Gousgounis Matthew Bombard and William Calhoun for their significant contributions to this and related projects

Results provisional pending peer-reviewed publication in a leading financial journal

Visit thomsonreuterscom

For more information contact your representative or visit us online

copy Thomson Reuters 2013 All rights reserved 100531111-13

Thomson Reuters Machine Readable NewsAsif Alam3 Times SquareNew York NY 10036USAasifalamthomsonreuterscomhttpthomsonreuterscommachine-readable-news

Thomson Reuters RampDIan MacGillivray3 Times SquareNew York NY 10036USAianmacgillivraythomsonreuterscom httplabsthomsonreuterscom

MarketPsychRichard Peterson MD179 Post Road WWestPort CT 06880USArichardmarketpsychdatacomhttpwwwmarketpsychcom

Stevens Institute of TechnologyDr Eleni GousgounisCastle Point on HudsonHoboken NJ 07030USAelenigousgounisstevenseduhttpstevensedufsc

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

Data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below In addition to the standard set of MarketPsych indices country-level indices further include market risk trade balance budget deficit central bank and consumer sentiment

AUD 274 0 0 0 0 0 0 0 0 0 0 0 061 3793 243 006

BRL 490 0 0 0 011 011 0 0 0 0 0 0 11 677 05 403

CAD 188 0 0 0 0 0 0 0 0 0 0 0 072 2457 293 0

CHF 929 0 0 0 055 099 0 0 0 0 0 0 4147 9094 0 541

DKK 1755 0 0 024 1143 936 089 018 0 036 041 0 7636 9034 529 2079

EGP 1126 0 0 0 146 353 017 0 0 0 006 0 4888 7534 633 3318

EUR 002 0 0 0 0 0 0 0 0 0 0 0 0 033 0 0

GBP 127 0 0 0 0 0 0 0 0 0 0 0 061 1844 0 0

HKD 984 0 0 0 033 022 0 0 0 0 0 0 2419 873 2927 624

IDR 537 0 0 0 017 028 006 0 0 0 0 0 32 6282 359 105

ILS 941 0 0 0 0 0 0 0 0 0 0 0 3223 7203 2489 1194

INR 273 0 0 0 0 0 0 0 0 0 0 0 0 4075 017 0

JPY 335 0 0 0 0 0 0 0 0 0 0 0 006 5014 0 0

KRW 644 0 0 0 0 022 0 0 0 0 0 0 055 915 243 193

MXN 645 0 0 0 006 017 0 0 0 0 0 0 1226 8161 055 215

MYR 718 0 0 0 044 083 0 011 0 0 0 0 707 80 939 983

NOK 1360 0 0 0 714 1199 023 006 0 0 017 0 5511 8709 628 3598

NZD 942 0 0 0 022 022 0 006 0 006 0 0 2733 8476 2198 668

RUB 278 0 0 0 0 006 0 0 0 0 0 0 006 3908 094 15

SEK 1152 0 0 0 274 268 0 0 0 006 0 0 6512 8234 732 1252

SGD 933 0 0 0 039 182 0 0 0 0 0 0 2115 9039 1629 994

THB 705 0 0 0 006 028 0 011 0 0 006 0 762 8702 663 398

TRY 722 0 0 0 008 068 0 0 0 0 0 0 1814 7333 051 1561

TWD 771 0 0 0 061 099 0 0 0 0 0 0 607 9227 872 696

USD 001 0 0 0 0 0 0 0 0 0 0 0 0 017 0 0

ZAR 714 0 0 0 006 006 0 0 0 0 0 0 1364 6897 2021 409

Averages 675 0 0 001 099 133 005 002 000 002 003 000 1783 6451 863 782

CURRENCY

NA VALUES

BUZZSNTMENT

OPTIMSM

FEARJO

YTRUST

VIOLENC

CONFLCT

URGENCY

UNCERTN

MKTRISK

TRADBAL

BDEFCIT

CENBANK

CNSMSNT

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous page data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices credit easy vs tight economic forecast growth amp uncertainty election sentiment financial system instability fiscal policy loose vs tight government anger government corruption inflation interest rate and investment flow

AUD 055 398 0 0 0 1022 4892 3247 149 028 0 077 006 7178

BRL 817 1745 144 0 011 6748 7891 5279 4373 1496 0 265 105 5748

CAD 039 061 0 0 0 558 4992 2004 171 0 0 033 006 6322

CHF 2783 3009 1115 0 116 899 6946 8769 6659 3689 127 1469 1115 9608

DKK 5966 7263 4046 012 1339 91 9467 8649 7666 59 504 6108 689 9846

EGP 5504 6967 2287 022 1536 5998 9496 8313 2965 2416 022 4697 1581 7881

EUR 0 0 0 0 0 055 447 088 0 0 0 0 0 2501

GBP 0 006 0 0 0 359 1922 718 017 0 0 0 0 7062

HKD 1375 4462 1253 0 083 7968 8112 8697 6599 3628 061 1353 1894 8796

IDR 2536 4923 464 0 127 5359 8414 6249 2961 442 0 1083 174 6122

ILS 3526 5241 078 0 05 4131 8705 7024 303 073 0 2248 1474 8576

INR 006 237 0 0 0 591 5445 2032 039 006 0 044 05 1723

JPY 028 099 0 0 0 2115 3274 2308 497 044 0 0 011 5295

KRW 1507 2838 171 0 033 5903 698 7123 2678 845 0 254 74 3733

MXN 2181 27 326 0 022 5682 8498 7245 2551 502 0 348 1325 8029

MYR 3657 4503 1039 0 055 6939 905 7492 4331 1315 0 2851 4215 7602

NOK 5917 7036 3427 023 1593 9332 9195 9086 8104 6453 428 4215 1702 9903

NZD 723 4931 906 011 155 762 8531 7885 5533 2717 05 2065 331 9249

RUB 445 3 006 0 0 1968 6459 4058 056 028 0 278 728 5114

SEK 4667 4941 2348 011 548 8519 872 8167 7798 5461 229 3214 3231 9883

SGD 3164 5119 138 0 232 8719 8625 8741 751 3821 083 1596 3777 9045

THB 3401 5273 828 0 127 4246 7096 7195 312 1022 006 2043 2258 7151

TRY 2143 535 903 0 219 5865 8422 7148 2979 2118 0 717 65 8152

TWD 386 5809 823 0 188 6218 9128 8747 5665 2424 022 2998 2833 7874

USD 0 0 0 0 0 0 099 0 0 0 0 0 0 994

ZAR 1811 3131 26 0 077 5781 8233 6477 1513 442 006 26 287 7951

Averages 2158 3321 839 003 250 4992 6886 5875 3240 1726 059 1470 1421 6975

CURRENCY

CRDTEVT

DEFAULT

ECONCFT

ECONGRW

EC_UNCR

ELECSNT

FNSYSIN

FISCLVT

GVTANGR

GVTCORR

GVTINST

INFLATN

INTRATE

INVFLOW

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index BUZZThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous two pages data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices Monetary policy loose vs tight natural disasters regime change social inequality social unrest and unemployment

AUD 6516 0 1988 7559 011 536

BRL 7427 088 6013 7405 017 4666

CAD 651 0 1679 7311 0 442

CHF 804 1253 7521 9663 337 5693

DKK 9822 391 859 9627 877 3466

EGP 9759 1508 4669 9226 034 6637

EUR 1331 0 0 3556 0 0

GBP 1165 0 05 2601 0 077

HKD 8134 558 7786 8393 331 5621

IDR 8779 066 6956 9221 011 6232

ILS 9355 022 1659 8021 0 4137

INR 5157 0 083 4815 0 977

JPY 2568 0 2236 7537 0 768

KRW 8526 42 5577 8498 033 5312

MXN 9177 061 19 8636 006 365

MYR 8785 1077 7569 947 099 7608

NOK 9652 2479 8658 9766 919 309

NZD 8597 387 7863 9094 155 5527

RUB 8299 0 256 8705 0 4074

SEK 9223 3387 8234 9268 682 422

SGD 8056 1618 8603 9034 939 7338

THB 8609 436 3517 8901 022 8067

TRY 7519 92 3291 9468 0 6068

TWD 9304 872 8034 9359 304 4296

USD 42 0 0 674 0 0

ZAR 836 166 3418 5086 0 1706

Averages 7273 740 4467 7727 184 3854

CURRENCY

MONELVT

NATDSST

REGMCHG

SOCINEQ

SOCUNRS

UNEMPLY

THOMSON REUTERS FINANCIAL AND RISK

ABOUT THE AUTHORIan MacGillivray is a Senior Research Engineer in the Thomson Reuters Corporate Research amp Development group His current work includes an entity-centric framework to discover quantify and qualify relationships and similarities an effort to find new sources of alpha for the markets and explorations of social media and behavior His research background is in probability theory and agent-based systems He has a keen interest in physics and language and has a BSc in Computer Science from Aston University He was awarded the Thomson Reuters Inventor of the Year award for a patent filed in 2011

ABOUT THOMSON REUTERSThomson Reuters is the worldrsquos leading source of intelligent information for businesses and professionals We combine industry expertise with innovative technology to deliver critical information to leading decision makers in the financial and risk legal tax and accounting intellectual property and science and media markets powered by the worldrsquos most trusted news organization With headquarters in New York and major operations in London and Eagan Minnesota Thomson Reuters employs approximately 60000 people and operates in over 100 countries Thomson Reuters shares are listed on the Toronto and New York Stock Exchanges

For more information about Thomson Reuters please go to wwwthomsonreuterscom

Original research by Thomson Reuters Research amp Development in collaboration with the Stevens Institute of Technology With great thanks to Dr Eleni Gousgounis Matthew Bombard and William Calhoun for their significant contributions to this and related projects

Results provisional pending peer-reviewed publication in a leading financial journal

Visit thomsonreuterscom

For more information contact your representative or visit us online

copy Thomson Reuters 2013 All rights reserved 100531111-13

Thomson Reuters Machine Readable NewsAsif Alam3 Times SquareNew York NY 10036USAasifalamthomsonreuterscomhttpthomsonreuterscommachine-readable-news

Thomson Reuters RampDIan MacGillivray3 Times SquareNew York NY 10036USAianmacgillivraythomsonreuterscom httplabsthomsonreuterscom

MarketPsychRichard Peterson MD179 Post Road WWestPort CT 06880USArichardmarketpsychdatacomhttpwwwmarketpsychcom

Stevens Institute of TechnologyDr Eleni GousgounisCastle Point on HudsonHoboken NJ 07030USAelenigousgounisstevenseduhttpstevensedufsc

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index CoverageThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous page data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices credit easy vs tight economic forecast growth amp uncertainty election sentiment financial system instability fiscal policy loose vs tight government anger government corruption inflation interest rate and investment flow

AUD 055 398 0 0 0 1022 4892 3247 149 028 0 077 006 7178

BRL 817 1745 144 0 011 6748 7891 5279 4373 1496 0 265 105 5748

CAD 039 061 0 0 0 558 4992 2004 171 0 0 033 006 6322

CHF 2783 3009 1115 0 116 899 6946 8769 6659 3689 127 1469 1115 9608

DKK 5966 7263 4046 012 1339 91 9467 8649 7666 59 504 6108 689 9846

EGP 5504 6967 2287 022 1536 5998 9496 8313 2965 2416 022 4697 1581 7881

EUR 0 0 0 0 0 055 447 088 0 0 0 0 0 2501

GBP 0 006 0 0 0 359 1922 718 017 0 0 0 0 7062

HKD 1375 4462 1253 0 083 7968 8112 8697 6599 3628 061 1353 1894 8796

IDR 2536 4923 464 0 127 5359 8414 6249 2961 442 0 1083 174 6122

ILS 3526 5241 078 0 05 4131 8705 7024 303 073 0 2248 1474 8576

INR 006 237 0 0 0 591 5445 2032 039 006 0 044 05 1723

JPY 028 099 0 0 0 2115 3274 2308 497 044 0 0 011 5295

KRW 1507 2838 171 0 033 5903 698 7123 2678 845 0 254 74 3733

MXN 2181 27 326 0 022 5682 8498 7245 2551 502 0 348 1325 8029

MYR 3657 4503 1039 0 055 6939 905 7492 4331 1315 0 2851 4215 7602

NOK 5917 7036 3427 023 1593 9332 9195 9086 8104 6453 428 4215 1702 9903

NZD 723 4931 906 011 155 762 8531 7885 5533 2717 05 2065 331 9249

RUB 445 3 006 0 0 1968 6459 4058 056 028 0 278 728 5114

SEK 4667 4941 2348 011 548 8519 872 8167 7798 5461 229 3214 3231 9883

SGD 3164 5119 138 0 232 8719 8625 8741 751 3821 083 1596 3777 9045

THB 3401 5273 828 0 127 4246 7096 7195 312 1022 006 2043 2258 7151

TRY 2143 535 903 0 219 5865 8422 7148 2979 2118 0 717 65 8152

TWD 386 5809 823 0 188 6218 9128 8747 5665 2424 022 2998 2833 7874

USD 0 0 0 0 0 0 099 0 0 0 0 0 0 994

ZAR 1811 3131 26 0 077 5781 8233 6477 1513 442 006 26 287 7951

Averages 2158 3321 839 003 250 4992 6886 5875 3240 1726 059 1470 1421 6975

CURRENCY

CRDTEVT

DEFAULT

ECONCFT

ECONGRW

EC_UNCR

ELECSNT

FNSYSIN

FISCLVT

GVTANGR

GVTCORR

GVTINST

INFLATN

INTRATE

INVFLOW

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index BUZZThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous two pages data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices Monetary policy loose vs tight natural disasters regime change social inequality social unrest and unemployment

AUD 6516 0 1988 7559 011 536

BRL 7427 088 6013 7405 017 4666

CAD 651 0 1679 7311 0 442

CHF 804 1253 7521 9663 337 5693

DKK 9822 391 859 9627 877 3466

EGP 9759 1508 4669 9226 034 6637

EUR 1331 0 0 3556 0 0

GBP 1165 0 05 2601 0 077

HKD 8134 558 7786 8393 331 5621

IDR 8779 066 6956 9221 011 6232

ILS 9355 022 1659 8021 0 4137

INR 5157 0 083 4815 0 977

JPY 2568 0 2236 7537 0 768

KRW 8526 42 5577 8498 033 5312

MXN 9177 061 19 8636 006 365

MYR 8785 1077 7569 947 099 7608

NOK 9652 2479 8658 9766 919 309

NZD 8597 387 7863 9094 155 5527

RUB 8299 0 256 8705 0 4074

SEK 9223 3387 8234 9268 682 422

SGD 8056 1618 8603 9034 939 7338

THB 8609 436 3517 8901 022 8067

TRY 7519 92 3291 9468 0 6068

TWD 9304 872 8034 9359 304 4296

USD 42 0 0 674 0 0

ZAR 836 166 3418 5086 0 1706

Averages 7273 740 4467 7727 184 3854

CURRENCY

MONELVT

NATDSST

REGMCHG

SOCINEQ

SOCUNRS

UNEMPLY

THOMSON REUTERS FINANCIAL AND RISK

ABOUT THE AUTHORIan MacGillivray is a Senior Research Engineer in the Thomson Reuters Corporate Research amp Development group His current work includes an entity-centric framework to discover quantify and qualify relationships and similarities an effort to find new sources of alpha for the markets and explorations of social media and behavior His research background is in probability theory and agent-based systems He has a keen interest in physics and language and has a BSc in Computer Science from Aston University He was awarded the Thomson Reuters Inventor of the Year award for a patent filed in 2011

ABOUT THOMSON REUTERSThomson Reuters is the worldrsquos leading source of intelligent information for businesses and professionals We combine industry expertise with innovative technology to deliver critical information to leading decision makers in the financial and risk legal tax and accounting intellectual property and science and media markets powered by the worldrsquos most trusted news organization With headquarters in New York and major operations in London and Eagan Minnesota Thomson Reuters employs approximately 60000 people and operates in over 100 countries Thomson Reuters shares are listed on the Toronto and New York Stock Exchanges

For more information about Thomson Reuters please go to wwwthomsonreuterscom

Original research by Thomson Reuters Research amp Development in collaboration with the Stevens Institute of Technology With great thanks to Dr Eleni Gousgounis Matthew Bombard and William Calhoun for their significant contributions to this and related projects

Results provisional pending peer-reviewed publication in a leading financial journal

Visit thomsonreuterscom

For more information contact your representative or visit us online

copy Thomson Reuters 2013 All rights reserved 100531111-13

Thomson Reuters Machine Readable NewsAsif Alam3 Times SquareNew York NY 10036USAasifalamthomsonreuterscomhttpthomsonreuterscommachine-readable-news

Thomson Reuters RampDIan MacGillivray3 Times SquareNew York NY 10036USAianmacgillivraythomsonreuterscom httplabsthomsonreuterscom

MarketPsychRichard Peterson MD179 Post Road WWestPort CT 06880USArichardmarketpsychdatacomhttpwwwmarketpsychcom

Stevens Institute of TechnologyDr Eleni GousgounisCastle Point on HudsonHoboken NJ 07030USAelenigousgounisstevenseduhttpstevensedufsc

THOMSON REUTERS FINANCIAL AND RISK MANAGEMENT

Appendix B MarketPsych Index BUZZThis shows the percentage of Not Available (NA) values for each indexcurrency pairing This is the number of times out of a hundred that the system is not able to generate a value from the source data

As on the previous two pages data for indices calculated at the country level (for the home country of a currency or all Eurozone countries for the Euro) are presented below Here are displayed the following country-level indices Monetary policy loose vs tight natural disasters regime change social inequality social unrest and unemployment

AUD 6516 0 1988 7559 011 536

BRL 7427 088 6013 7405 017 4666

CAD 651 0 1679 7311 0 442

CHF 804 1253 7521 9663 337 5693

DKK 9822 391 859 9627 877 3466

EGP 9759 1508 4669 9226 034 6637

EUR 1331 0 0 3556 0 0

GBP 1165 0 05 2601 0 077

HKD 8134 558 7786 8393 331 5621

IDR 8779 066 6956 9221 011 6232

ILS 9355 022 1659 8021 0 4137

INR 5157 0 083 4815 0 977

JPY 2568 0 2236 7537 0 768

KRW 8526 42 5577 8498 033 5312

MXN 9177 061 19 8636 006 365

MYR 8785 1077 7569 947 099 7608

NOK 9652 2479 8658 9766 919 309

NZD 8597 387 7863 9094 155 5527

RUB 8299 0 256 8705 0 4074

SEK 9223 3387 8234 9268 682 422

SGD 8056 1618 8603 9034 939 7338

THB 8609 436 3517 8901 022 8067

TRY 7519 92 3291 9468 0 6068

TWD 9304 872 8034 9359 304 4296

USD 42 0 0 674 0 0

ZAR 836 166 3418 5086 0 1706

Averages 7273 740 4467 7727 184 3854

CURRENCY

MONELVT

NATDSST

REGMCHG

SOCINEQ

SOCUNRS

UNEMPLY

THOMSON REUTERS FINANCIAL AND RISK

ABOUT THE AUTHORIan MacGillivray is a Senior Research Engineer in the Thomson Reuters Corporate Research amp Development group His current work includes an entity-centric framework to discover quantify and qualify relationships and similarities an effort to find new sources of alpha for the markets and explorations of social media and behavior His research background is in probability theory and agent-based systems He has a keen interest in physics and language and has a BSc in Computer Science from Aston University He was awarded the Thomson Reuters Inventor of the Year award for a patent filed in 2011

ABOUT THOMSON REUTERSThomson Reuters is the worldrsquos leading source of intelligent information for businesses and professionals We combine industry expertise with innovative technology to deliver critical information to leading decision makers in the financial and risk legal tax and accounting intellectual property and science and media markets powered by the worldrsquos most trusted news organization With headquarters in New York and major operations in London and Eagan Minnesota Thomson Reuters employs approximately 60000 people and operates in over 100 countries Thomson Reuters shares are listed on the Toronto and New York Stock Exchanges

For more information about Thomson Reuters please go to wwwthomsonreuterscom

Original research by Thomson Reuters Research amp Development in collaboration with the Stevens Institute of Technology With great thanks to Dr Eleni Gousgounis Matthew Bombard and William Calhoun for their significant contributions to this and related projects

Results provisional pending peer-reviewed publication in a leading financial journal

Visit thomsonreuterscom

For more information contact your representative or visit us online

copy Thomson Reuters 2013 All rights reserved 100531111-13

Thomson Reuters Machine Readable NewsAsif Alam3 Times SquareNew York NY 10036USAasifalamthomsonreuterscomhttpthomsonreuterscommachine-readable-news

Thomson Reuters RampDIan MacGillivray3 Times SquareNew York NY 10036USAianmacgillivraythomsonreuterscom httplabsthomsonreuterscom

MarketPsychRichard Peterson MD179 Post Road WWestPort CT 06880USArichardmarketpsychdatacomhttpwwwmarketpsychcom

Stevens Institute of TechnologyDr Eleni GousgounisCastle Point on HudsonHoboken NJ 07030USAelenigousgounisstevenseduhttpstevensedufsc

THOMSON REUTERS FINANCIAL AND RISK

ABOUT THE AUTHORIan MacGillivray is a Senior Research Engineer in the Thomson Reuters Corporate Research amp Development group His current work includes an entity-centric framework to discover quantify and qualify relationships and similarities an effort to find new sources of alpha for the markets and explorations of social media and behavior His research background is in probability theory and agent-based systems He has a keen interest in physics and language and has a BSc in Computer Science from Aston University He was awarded the Thomson Reuters Inventor of the Year award for a patent filed in 2011

ABOUT THOMSON REUTERSThomson Reuters is the worldrsquos leading source of intelligent information for businesses and professionals We combine industry expertise with innovative technology to deliver critical information to leading decision makers in the financial and risk legal tax and accounting intellectual property and science and media markets powered by the worldrsquos most trusted news organization With headquarters in New York and major operations in London and Eagan Minnesota Thomson Reuters employs approximately 60000 people and operates in over 100 countries Thomson Reuters shares are listed on the Toronto and New York Stock Exchanges

For more information about Thomson Reuters please go to wwwthomsonreuterscom

Original research by Thomson Reuters Research amp Development in collaboration with the Stevens Institute of Technology With great thanks to Dr Eleni Gousgounis Matthew Bombard and William Calhoun for their significant contributions to this and related projects

Results provisional pending peer-reviewed publication in a leading financial journal

Visit thomsonreuterscom

For more information contact your representative or visit us online

copy Thomson Reuters 2013 All rights reserved 100531111-13

Thomson Reuters Machine Readable NewsAsif Alam3 Times SquareNew York NY 10036USAasifalamthomsonreuterscomhttpthomsonreuterscommachine-readable-news

Thomson Reuters RampDIan MacGillivray3 Times SquareNew York NY 10036USAianmacgillivraythomsonreuterscom httplabsthomsonreuterscom

MarketPsychRichard Peterson MD179 Post Road WWestPort CT 06880USArichardmarketpsychdatacomhttpwwwmarketpsychcom

Stevens Institute of TechnologyDr Eleni GousgounisCastle Point on HudsonHoboken NJ 07030USAelenigousgounisstevenseduhttpstevensedufsc