This item is only available as the following downloads:
CHANGING SLICES OF A GROWING PIE: THE DECLINE OF THE WAGE SHARE AND RISE OF THE PROFIT SHARE OF U.S. GDP BY REBECCA RYAN KEENAN A THESIS Submitted to the Division of Social Sciences New College of Florida in partial fulfillment of the requirements for the degree Bachelors of Arts in Economics Under the Sponsorship of Dr. Tarron Khemraj Sarasota, Florida May, 2012
i Acknowledgements Any achievement I have made was only possible with the support of my parents, whose time and generosity allowed me the opportunity of the New College of Florida experience I also want to thank my best friends, two of which are Lauren Keenan and Lisa Keenan It was nice to have family at NCF and even nicer to now have NCF be part of the family. This thesis would not have bee n possible without my advisor Dr. Tarron Khemraj, who encouraged my interest in the Financial Crisis I would also like to thank m y baccalaureate committee, Professors Rick Coe and Patrick VanHorn, whose classes I would have little economic foundation without. "There's always the same amount of good and evil, too...But we can never alter the ratio of good to evil. All we can do is keep things stirred up so neither good nor evil solidifies. That's when things get scary. Life is like a stew, you have to stir it fr equently, or
ii Table of Contents Acknowledgments .................................................................... .................. ........... ........ i Table of Contents ......................................................... .............................. .................... ii Figures ............................................................................................. ................... ........... iii Abstract ............................................................................................. ............................. iv Chapter 1: Introduction ...................................................................... ............ ... ..... .. 1 1.1 Mot ivation................................................................................. ................ ..............2 1.2 Literature Review....................................................................................... .... ........ 7 1.2.1 Income Inequality, Debt and Capital Account Activity............................. 7 1.2.2 The 2007 Financial Crisis and Institutional Policies ................................. 15 Chapter 2: Formulation of the Profit Ratio Model and Wage Ratio Model .... ......33 2.1 Data and Models .......................................................................... ............................ 33 2.2 Dependent Variables................................................................................................ 37 2.3 Independent Variables............................................................................................. .39 2.3 .1 Political Climate........................................................................................ 39 2.3 .2 Characteris tics of the Private Sector......................... .. ..............................43 2.3 .3 Consumerist America............................................................................... ..45 2.3 ............................................ 48 2.4 Omitted Variables ...................................................................... ........................ ........52 2.5 Methodology ................................................................. ....... ...... ............. .................55 Chapter 3 : Interpretations and Conclusion ................. ............................... ............... 58 4.1 Interpreting the Results............................................................. ............. ....... ... ...... 58 4.2 Summary and Conclusion....................................... ............................... ......... .......... 81 References..................... ................................ ................................................................... 84
iii Tables & Figures Table 1. Different Slices of t Table 3. Profit Ratio Model Improve Table 4. Coefficients and Signi Table 5. Accompanying Estimation Table 6. LM Test for Serial Correlati 76 Table 7. LM Test for Serial Correlat Table 8. Q Statistics for the Table 9. WR Model Correlatio Table 10. Wald Test for Joint Significa Figure 1. Wage Share and Profit Share of U.S. Gross Figure 2. Rising Productivity Yet a Falling Wage Figure 3. One Perspective of Growing Debt Figure 4. A Second Perspective of Growing De Figure 5. Monetary Policy Effects th Figure 6. Money Changing Figure 7 The Wage Ratio and the Labor Figure 8 Monetary Policy Trying to Steer t Figure 9 ....26 Figure 10 Corporate Activity & Figure 11 Profit Ratio M Figure 12 Wage Ratio M
iv CHANGING SLICES OF A GROWING PIE: THE DECLINE OF THE WAGE SHARE AND RISE OF THE PROFIT SHARE IN U.S. GDP Rebecca Ryan Keenan New College of Florida, 2012 ABSTRACT The purpose of this thesis is to provide an empirical examination of the wage share and profit share of Gross Domestic Product in the United States, as well as to shed light on future avenues for further research related to the topic. Comparisons of the shares are modeled as multivariate equations and estimated using Ordinary Least Squares in EViews 7.0. The determinants in each model are loosely categorized under four broad topics that reflect aspects of the federal government, the private sector, Americ an workers and monetary policy. Final results show the profit share of the U.S. economy is significantly impacted by the Net Capital Rule change of 2004, the Subprime Mortgage Crisis, Republican Presidents, growth in the cost of a barrel of oil, the lagged profit share, and the effect Federal Funds Rate. The wage share model finds the Net Capital Rule change of 2004, Republican Presidents, unemployment rate, corporate cash flow share and lagged wage share to be significant. Dr. Tarron Khemraj Div ision of Social Sciences
1 Chapter 1: Stylized Facts This thesis intends to provide an exploratory empirical examination of the wage share and profit s hare distributions of Gross Domestic Product (GDP). Avenues for potential future research are offered as well When investigating these ratios, four general areas concerning the political climate, characteristics of the private sector, American consum ers, and macroeconomic policy are presented as evidence explainin g the wage share and profit share historical trends, and what may be the implications of these relationships. I find the impact of the inertia of the profit ratio, growth in the cost of a barrel of oil, Republican Presidents, the Net Capital Rule change of 2004, and the Subprime Mortgage Crisis all hold significance relationships with the profit ratio. The wage share model finds the Net Capital Rule change of 2004, Republican Presidents, unemployment rate, corporate cash flow share and lagged wage share to be significant. Chapter one begins with a discussion of Figure 1 showing the wage and profit shares trends, followed by a variety of recent views of income inequality a topic closely tied to this thesis. Relationships tying income inequality, debt and capital account activity are presented and examined. Chapter two describes how the two mult ivariate models were developed. That is, which independent variables were chosen to be included and why. I also ta lk about how powerful I expect the coefficients to be on each dependent variable. All the variables are loosely fit into one of the above four cat egories, each reflecting broad forces that affect the aggregate economy. Omitted variables are briefly discuss ed The third chapter shows the results and analysis of the estimated models using Ordinary Least Squares (OLS) along with a brief discussion of the Gauss Markov
2 assumptions for Best Linear Unbiased Estimator. A summary and conclusion of the study follow s. 1.1 Motivation Since before 1970, the American economy has been experiencing what Berg and $5.8 trillion to $13.3 trillion in 2011. 1 During this time period, inequal ity in the United 2 Unlike the clear economic growth of the American economy in recent decades, the growth of the wage share has not kept pace. No doubt, the pub lic is cognizant of developing income inequality. The near disaster of the 2007 Financial Crisis galvanized citizens to show their discontent through Occupy Wall Street and a wave of accompanying Occupy protests occurring nationally in multiple U.S. cities. As recent as January 2012, the Chairman of the Council of Economic Advisers Alan Kruger delivered remarks pertaining to income inequality to th e Center for American Progress e United States over the last three decades has reached the point that inequality in incomes is causing an unhealthy division Krueger re 3 1 Data from the St. Louis Federal Reserve Economic Data. 2 Milanovic, B ranko. "More or Less." Finance & Development September 2011, 6 11. 3 Alan B. Krueger, "The Rise and Consequences of Inequality in the United States," Chairman, Council of Economic Advisers January (2012): 1 10.
3 Figure 1 Wage Share and Profit Share of U.S. Gross Domestic Product This chapter presents the stylized behavior of the wage share and profit share trends as a percent of United States GDP (see Figure 1). The two slices of the economic pie merit analysis as the disparity between the two reflect a sense of changing income equality in the United States. It is important to emphasize however, that this thesis i s not about income inequality, but more so an econometric exercise. By examining how GDP is dispersed to different groups in the economy such as corporations earning profits (after tax) and workers earning wages and salaries (non monetary fringe benefits excluded) the distributi on of income may be sensed. Fringe benefits were not included when computing the wage share. Although the inclusion of fringe benefits would have added another dimension to the share since good benefits are sought after by workers the inclusion could potentially misrepresent the wage share. Benefits financially support workers however, the consumer baskets offered
4 by insurance providers may not be well customized to clientele. This can leave some benefits unused or force medical c overage to be paid for out of pocket. Additionally, this study would have been enhanced if the profit share was based around total profits since after tax profits may seem distorted without knowing before tax profits. The incorporation of the two would have offered an empirical and visual estimation of corporate taxes paid out. Lastly, this thesis cannot speak clearly to income inequality since the role of the Gini coefficient (mentioned on page 2) is absent. Income in equality is traditionally measured by the Gini coefficient whose estimation is advanced in technique due to the usage of ratios. Still, examining the wage share offers a perspective on worker livelihood. 4 In the context of the growing American economy, th e decline of the wage share translates into an even greater decrease in the share of GDP than in the case of a stagnating wage share. Comparatively, the slice belonging to corporate profits has risen in the background of increasing overall economic growth. The essence of these changing proportions is a reflection of shifting income distributions. Examining Tables 1, the rewards of economic growth are increasingly being received by the growing slice of corporate profits. The wage ratios and profit ratios dis played in Table 1 are the shares of GDP from the year shown. Table 2 shows GDP from each decade as a share of 2010 GDP in order to display the growing rate of GDP, decade by decade. 4 Income received through investments as well as from government programs (old age security, unemployment insurance, medicaid, etc.) are not included.
5 Table 1. Different Slices of the Pie Table 2. GDP G rowth, Decade By Decade 1960 Wage Ratio Profit Ratio 1970 1980 1990 2000 2010 GDP(1960) as a share of GDP(2010) GDP(2010) GDP(1990) as a share of GDP(2010) GDP(2010) GDP(1970) as a share of GDP(2010) GDP(2010) GDP(2000) as a share of GDP(2010) GDP(2010) GDP(1980) as a share of GDP(2010) GDP(2010) GDP(2010) as a share of GDP(2010) GDP(2010)
6 When considering the wage share, further concern may be directed at the trend break occurring at the end of the post World War II era. Figure 1 shows relative stability in the wage share from 1947 until 1970, when a decline commences Figure 2 displays a simple comparison of the wage share of GDP (or wage ratio) and business and nonfarm business sectors of productivity Real GDP in these two sectors is the basis of the output components of the major sector labor productivity and multifa ctor productivity measures. 5 Productivity shows a clear upward trend, defined by Berg and Ostry (2011) as a growth spell: a minimum period of eight years that begins with a growth upbreak and finishes with a downbreak. Figure 2 R ising Productivity Yet a Falling Wage Share 5 BLS Handbook of Methods U.S. Department of Labor, 1997. s.v. "Chapter 10 Productivity Measures : Business Sector and Major Subsectors ." http://stats.bls.gov/opub/hom/pdf/homch11.pdf (accessed May 20, 2012).
7 1.2 Literature Review 1.2.1 Income Inequality, De bt and Capital Account activity This next section, which will pick back up on the note of productivity growth and wage share, also discusses the inter relationships between primarily three of five developments coinciding with financialization. 6 Those developments are income inequality a focus of this thesis debt and the activities pertaining to the Capital Account. The two developments not discussed, central bank policy devoted mainly to price stability and the alignment of top managerial pay to stock price movements through share options 7 may be subjects of further research relating to this thesis. The second portion of the literature review is devoted to the 2007 Financial Crisis and surrounding events. Policies and actions adopted by the Federal government as well as the Federal Reser ve Bank, as well as related economic reactions are discussed. ***** Noting how the United States has been experiencing a growth spell since the St. Louis Federal Reserve began data collection in 1947, it is interesting to observe a falling wage ratio since the 1970s. The change in trend contradicts a consequence generally 6 Note: The following information is from Dnhaupt, Petra. "Financialization and the rentier income share evidence from the USA and Germany." IMK Working Paper (2010): 3 5 A broad definition of addition to internal changes within the financial sect or. Since there does not exist a precise and agreed upon definition for financialization, the five areas in academic literature Krippner (2005) summarizes will be used. Those main stances are: shareholder value orientation, finance via capital markets rath new financial instruments and the supremacy of profit making via financial rather than real channels. 7 Dnhaupt, Petra. "Financialization and the rentier income share ev idence from the USA and Germany." IMK Working Paper (2010): 3 5
8 implicit of rising productivity; that of rising wages as a result of workers producing more per hour of labor. Economist Facundo Alv aredo considers the decline of the wage ratio to be related to the emergence of the working rich the top executives at the end of the 20 th century. When the century began the wealthiest 1 percent were primarily capital owners, but over time this group ha s also come to include a new set of high end, highly compensated wage earners. According to Alvaredo, the share of income going to the top by major shocks from the Grea t Depression and wartime, though more recently, because of jumps in executive pay. 8 Evidence further supporting this notion comes from Skott/Ryoo (2008), who finds an alignment of top managerial pay to stock price movements through share options. 9 These fi ndings add to the explanation of what factors may contribute to the declining wage share and rising profit share. Al varedo (2011), Saez (2008), and Aron Dine and Shapiro (2007) document the disparity between overall wage growth and wage growth less the 1 p ercent between the early 1990s and the 2000s. These researchers find a significant difference of growth in the wage ratio. Alvaredo, the most recent study, finds average real income grew annually at 1.3 percent between 1993 and 2008. Taking out the 1 perce nt, that average real income is nearly halved to an annual 0.75 percent; real incomes for the excluded top percentile averaged out at 3.9 percent half of the overall economic growth within the period. Economist Emmanuel Saez finds the share of total inco me going to the top 10 percent of households in the US (families with incomes of $109,600 and higher) has 8 Alvaredo, Facundo. "Inequality over the Past Century." Finance & Development no. September (2011): 28 9. 9 Skott, P., and S. Ryoo. "Macroeconomic implications of financialization." Cambridge Journal of Economics 32. no. 6 (2008): 827 862.
9 climbed substantially during the last three decades. The share as of 2007 was 49.7% of total income. This rate surpasses the stock market bubble peak in the 1920s, the last time the wage ratio for the top 10 percent was so high. The top decile share of income between World War II and the 1970s remained fairly steady, around 33%, a feature Saez suggests is an effect of wartime policies. 10 Additional resea r ch by Kumhof and Rancire suggest that income inequality played a key role in the origins of the Great Depression of 1929 and the Great Recession of 2007. Preceding the crises, there were acute increases in both income inequality and the debt to income ra tios. 11 Where did the rise in debt come from? Looking at real personal consumption expenditures before the Great Recession, Figure 3 shows what Berg and Ostry might call a consumption spell. Sustaining high levels of consumption in the face of a falling wag e ratio indicates consumer borrowing. 10 Saez, Emmanuel. University of California Berkeley, Department of Economics, "Striking it Richer: The Evolution of Top Incomes in the United States. Last modified March 2, 2012. A ccessed April 1 5, 2012. http://elsa.berkeley.edu/~saez/saez UStopincomes 2010.pdf. 11 Kumhof, Michael, and Romain Rancire. "Unequal=Indebted." Finance & Development September 2011. http://www.imf.org/external/pubs/ft/fandd/2011/09/kumhof.htm (accessed October 28, 2011 ).
10 Figure 3 One Perspective of Growing Debt in the U.S. Economy Kumhof and Rancire find the gap between consumption of the lower and middle class households and upper class households to be smaller than the gap of income between the groups, evidence of the debt to income ratio being concentrated in the former. In 1983 at the beginning o f a measured twenty five year period, the debt to income ratio stood at 80 percent for the top 5 percent of households; 60 percent for the other 95 percent of households. In 2008 the ratios flipped, standing at 65 percent and 140 percent, respectively. The se findings are supported by Figure 4, showing a clear rise in household debt in the years following 1983, particularly so after 2000. In order to sustain their standard of living, the lower 95 percent borrowed. By purchasing loan backed assets the wealthy were essentially investing their money into borrowers living on credit. This ultimately led to an increase in the wealth inequality, a result of consumption inequality being less than income inequality.
11 Figure 4 A Second Perspec tive of Growing Debt in the U.S. Economy The case of rising debt prompts the investigation into an explanation for how Americans were able to borrow money. Some cite loans with contracted two year zero percent annual rates that allowed funds to be given to myopic borrowers unprepared for the adjustable interest rate that followed. C ut rates that promote borrowing for individuals and financial institutions were also implemented to decrease income inequality though some say this increased debt 12 Policies such as these incorporate domestic and international financial liberalization, and add downward pressure on current accounts recording imports, exports and transfers of no n financial goods and services. 13 Skott/Ryoo (2010) finds that the lifting of capital controls allowed a tremendous rise in national and cross boarder financial capital flows affecting the capital account 14 Giant corporations seeking new markets will have a higher level of corporate cash flow 12 Federal Reserve "Section 109 of the Riegle Neal Interstate Banking and Branching Efficiency Act." Accessed April 8, 2012. http://www.federalreserve.gov/boarddocs/supmanual/cch/sec109.pdf 13 Kumhof and Rancire Unequal=Indebted 6 14 The Balance of Payments is composed of the capital account and current account.
12 activity if they invest abroad/if there is an outflow of capital going abroad. If capital effects. Part of the capital a ccount financial assets and liabilities, conducte d by the private and public sectors is Foreign 15 The long held worry made worse by the 2007 Financial Crisis, describes the scenario where global markets destabilize if there is a run on the currencies of countries with giant deficits like the United States. However this is unlikely, at least for the U.S., which is mainly seen as a safe haven by lenders. Despite this unli kelihood, the authors note that the increase in income inequality factors into the deterioration of aggregate savings investment balances for wealth ier countries, since poor and middle class citizens borrow from the rich and foreign lenders. Jahan and McDo nald, cite the International Monetary Fund as finding financial globalization, particularly concerning FDI, to be associated with widening income gaps in developing countries Not only this, but there exists a risk that the small minority of elites may tak e control and direct financial liberalization in such a way that access is limited. 16 Additionally Easterly, Islam, and Stiglitz, (2001) find that output growth volatility may result from financial development, especially in cases when private sector 15 Kumhof and Rancire Unequal=Indebted 7 16 Jahan and McDonald, A Bigger Slice of a Growing Pie 18.
13 credit exceeds 100 percent of GDP. A properly function financial system can support growth, however, the quality of institutions and the regulations laid in are vital. 17 ***** As discussed earlier in this chapter, Kumhof and Rancire recognize the Great D epression and the Great Recession as two major economic crises in the past hundred years. Data showing sharp rises both in income inequality and household debt to income 18 Economists Andrew Berg and Jonathan Ostry also note the recent rise in income inequality parallels that of the rise in the 1920s. Furthermore, Berg and Ostry point out that in both scenarios there was a boom in the financial sector and voluminous levels of borrowing by the lower and middle classes, a combination whose end resulted in (another) huge financial crisis. 19 But is there enough evidence to pin down the direction of causality between the recent financial crisis and income inequality? Some economists argue factors like loose monetary policy, excessive financial liberalization, and asset price bubbles that arose when there was an unusually sharp rise in the debt to income ratios may be explanations of the 2007 crisis. However, another Fault Lines: How Hidden Fractures Still Threaten the World Economy responds that these factors, though important, may 17 Ibid.,18. 18 Kumhof, Michael, and Romain Rancire. "Leveraging Inequality." Finance & Development December 2010. http://www.imf.org/external/pubs/ft/fandd/2010/12/Kumhof.htm (accessed October 28, 2011). 19 Berg, Andrew G., and Jonathan D. Ostry. "Warning! Inequality May Be Hazardous to Your Growth." IMFdirect (blog), April 8, 2011. http://blog imfdi rect.imf.org/2011/04/08/inequality and growth/ (accessed April 17, 2012).
14 have only been the manifestations of l ong term, subsurface activities caused by a advantaged income group might have been leveraged in the political sphere to prop up easy credit, which maintained demand and job c reation even in the presence of stagnating incomes. 20 While the role of income inequality pertaining to the 2007 Financial Crisis (and even the Great Depression of 1929) is still uncertain, it may be said with confidence that income inequality is a key co mponent when it comes to growth. In a model developed by Berg and Ostry, aspects that could end long growth spells are estimated. Variables included are political institutions, health and education, macroeconomic instability, debt and trade openness, to na me a few. Among all these variables, income inequality showed equal income distribution displayed a greater probability in ending high growth spells. 21 America and emerging Asia would more than double the anticipated length of a growth spell. Moreover, the model demonstrated income inequality to be a significant aspect for growth regardless of which other variables were included, or for that matter, how a growth spell was defined. 22 20 21 Warning! Inequality May Be Hazardous to Your Growth". 22 Ibid.
15 So it seems from all the above information, income inequality matters for economic growth. Frankly speaking, a certain degree of income inequality is fundamental to the effective functioning of a market economy and the necessary incentives for investment and growth. However, like most everything else in the world, there exist negative aspects of income inequality something this thesis attempts to make clear. Risks of excessive income inequality include crises and rising obstacles that 23 growth theory; it may very well indicate the decline of a society. A n adaptive and critically thinking work force vital in a globalizing age, a broader base of human capital automatically feeds into itself and m akes education more accessible with countless positive spillover effects spurring innovation. 1.2.2 The 2007 Financial Cr isis and Institutional Policies Figure 5 presents the growth of money, data that is largely a result of and reaction to credit demanded by borrowers. The graph shows the money supply declining after 1987; the year Alan Greenspan was appointed Chairman of the Federal Reserve (the Fed) and the start of the early 1990s recession. In 1993 the money supply rose in conjunction 23 Berg, Andrew G., and Jonathan D. Ostry. "Inequality and Unsustainable Growth: Two Sides of the Same Coin?." IMF Staff Discussion Note (2011): 3 5. http://www.imf.org/external/pubs/ft/sdn/2011/sdn1108.pdf (accessed April 22, 2012).
16 with the Clinton era boom, hovering around 6 percent of GDP until 2005 when it dropped to 4 percent and then climbed again in subsequent years until 2010. Figure 5 Monetary Policy Effect s the Money Supply The growth of money supply dropped by about 5 percent in 2010, as the first round of Quantitative Easing (QE1) by the Fed ended that March. QE1 was initiated at the end of November 2008, when the Financial Crisis was several weeks in bloom. The Fed seized control of Fannie Mae and Freddie Mac, two mortgage giants; bailed out American International Group Inc. (AIG); and forced the global financial services firm Lehman Brothers Holdings Inc. to declare bankruptcy, while similar firms waited in jeopardy until President George W. Bush signed the $700 billion bailout plan into law. Shortly after the Fed cut the effective Federal Funds Rate to 1 percent and after QE1 was
17 applied, the 30 year fixed rate mortgages started to decline. 24 pl an had some success in lowering mortgage interest rates as an attempt to unfreeze credit markets across the globe. Loosening the credit freeze was the objective of QE1 and later QE2. Injections of cash into the economy aimed to stimulate spending by maint aining or lowering the mortgage interest rates in order to discourage saving. This would increase lending some such as International Monetary Research Professor Tim Co ngdon argue the money supply shrank in 2010 right after QE2 began due to foreign regulators reacting to the crisis by pressuring their banks to increase capital asset ratios and decrease risk assets. It was during this time as well that the mortgage inte rest rates began to rise for the next six months. 25 Even in the United States however, despite any command from the Fed, excess reserves of depository institutions jumped by approximately $1 trillion following the end of 2008 for security against bank runs and defaulting loans. 26 The graph in Figure 5 shows the velocities of M1 and M2 from the money stock. M1, being more liquid, is more sensitive to the number of times a dollar changes hands and is therefore more volatile. It is visible to see that both velo cities begin to plummet 24 da Costa, Polyana. Bankrate.com, "Financial Crisis Timeline : Collapse and Bailout." Last modified September 21, 2011. Accessed April 20, 2012. 25 Evans Pritchard, Ambrose. "US money supply plunges at 1930s Pace as Obama eyes fresh stimulus." The Telegraph May 26, 2010. http://www.telegraph.co.uk/finance/economic s/7769126/US money supply plunges at 1930s pace as Obama eyes fresh stimulus.html (accessed April 4, 2012). 26 "Educational Resources." Federal Reserve Bank of San Francisco (blog), March 2010. http://www.frbsf.org/education/activities/drecon/2010/0310.html (accessed April 4, 2012).
18 right around 2007 2008, indicating that households and firms were holding onto cash. Soon after QE1 takes effect there is a spike in the velocities. After about a year of the Fed buying up dirty securities from the private sector (N ov. 25, 2008 March 31, 2010), QE1 comes to a close. Again in 2010, the percent change of velocities dramatically drops. The Fed decides to initiate QE2 later that year, from Nov. 3, 2010 June 30, 2011, yet the velocity of money does not improve and for the next six months the mortgage interest rate rises by about half a percent (4.42% 5%) before finally descending again. Figure 6 Money Changing Hands Former Chair to the Council of Economic Advisers for the Obama Administr ation on to it
19 ve easing bigger ... More radically, they could go to a price level target, which would allow inflation to be higher than the target for a few years in order to compensate for the past few years, 27 Furthermore, increasing QE1 would have had a bigger effect on the economy since it was implemented when markets were functioning poorly and were not as liquid. 28 estimations for filling the output gap arrived at $1.8 trillion. The estimation was rejected by fellow economic advisors Lawrence Summers and Peter Orszag, at which point Romer compromised on the stimulus number and lowered it to $1.2 trillion. Regardless, the compromised number neve feared the financially conservative Congress would dismiss the proposal. Eventually, the Congress passes the $800 billion stimulus package (QE1). Not after much time the second stimulus followed, al so purchasing $600 billion of longer term Treasury securities, though having a smaller effect on the overall economy. As QE3 is currently have been presented on the memo with both pros and cons. 29 ***** 27 Yglesias, Matthew. "Romer on Boosting the Economy." Think Progress March 25, 2011. http://thinkprogress.org/yglesias/2011/03/25/200338/romer on boosting the economy/?mobile=nc (accessed April 2, 2012). 28 Alon, Titan, and Eric Swanson. "Operation Twist and the Effect of Large Scale Asset Purchases." FRBSF (blog), April 25, 2011. http://www.frbsf.org/publications/economics/letter/2011/el2011 13.html (accessed April 2, 2012). 29 Scheiber, Noam. "EXCLUSIVE: The Memo that Larry Summers Didn't Want Obama to See." The New Republic February 22, 2012. http://www.tnr.com/articl e/politics/100961/memo Larry Summers Obama (accessed April 5, 2012).
20 To re cap: The debt to income ratio for the lower 95 percent of households has been growing at a greater increasing rate since the 1980s. The wage ratio has been mostly stagnating and declining since the 1970s despite growing productivity, and Americans ar e in debt as a consequence of living off credit. The Financial Crisis first touches down in 2007 and banks stop lending and hold onto cash in the form of excess reserves, freezing credit. This means there is less spending and economic activity. Unemploymen t rises as a result and Americans are drowning under debt originating from the housing bubble (when owning a home could be seen as a durable good and an asset). Figure 7 The Wage Ratio and the Labor Force Parting Ways Figure 7 shows the rate of unemployment paired with wage ratio trend. Between 1970 and 1980 unemployment is oscillating though mostly above its natural rate, while the wage ratio trends down from 52 percent to just above 44 percent. The rate of unemployment above the natural rate represents recessions or low points of the business
21 cycle, and unemployment below the natural rate signifies booms. During the Clinton era boom (1993 2000) unemployment slips below its natural rate and the wage ratio increases. Finally, du ring the most recent years of the Bush Administration and just as the Obama Administration begins (coincidentally the same timing as the Financial Crisis), unemployment shoots up to nearly 10 percent a rate not seen since the early 1980s. Furthermore, th ere is an increasing trend of labor force participation that has a natural pull on the level of employed labor force. Strikingly, there appears a divergence between the path of the wage ratio and the path of the employed labor force. Keeping in mind growi n g labor participation, Figure 7 essentially shows a reduction of the wage ratio per percent of employed labor participants. This information combined with information from Tables 1 and 2 seen earlier in the chapter, which show the changing wage ratio and p rofit ratio against the backdrop of growing GDP (and productivity for that matter), clearly illustrates a shrinking distribution of GDP going to the wage ratio. For the sake of background information, Figure 8 is included, showing the historical course of the effective Federal Funds Rate (FFR). The FFR is the average interest rate at which federal funds trade throughout a day, affecting banks and financial institutions on the basis that the listed interest rate is the cost of trading assets. The FFR also i ndicates where the Fed is trying to direct the economy. The Federal Reserve Chairman during the 1980s, Paul Volker, dramatically increased the FFR in order to combat the 1970s and early 1980s stagflation. After 1980 the FFR was set at declining rates, comp aratively. The lowest dips occur to guard against recessions in the early 1990s, (just below 4 percent), in the early 2000s (below 2%), and again after the 2007 crisis (essentially 0%) where it remains low to this day. Although the FFR does affect
22 banks, t he effect is minimal. Monetary policy is much like a two by four: it will hit the target and everything surrounding the target. One of the most interesting insights offered by Kumhof and Rancire states that the rise of indebtedness for the 95 percent of financial intermediation. Further evidence by these researchers finds support of the growing wealth inequality by showing the ratio of private credit to GDP jumping from 90 to 210 percent; that is more than twice the former ratio from 1983. To say the least, between the years of 1981 and 2007 the U.S. financial sector experienced a rapid expansion as a share of GDP from 4 to 8 percent with increased debt ultimately raising financial vulnerability. The fragility o f the financial sector was made apparent during the Financial Crisis, when 10 percent of mortgage loans became delinquent, and there was a 5 percent fall in output. It was not always the case that the American financial sector was a delicate piece of the e conomy.
23 Figure 8 Monetary Policy Trying to Steer the U.S. Economy Many deregulatory policies were passed in the 1980s ultimately leading to over 2,000 banks closing between 1985 and 1992 during the Savings and Loan Crisis. W hen the 1990s arrived laws were signed breaking down any further barriers blocking diverse financial tools or restrictions. 30 In 1994 the Riegle Neal Interstate Banking and Branching Efficiency Act (IBBEA) was passed in an effort to give nationally charted banks equality with state charted banks by reason of increased industry competition. The IBBEA repealed the McFadden Act of 1927, which prohibited interstate banking by federal charted banks via stipulation that those banks could not branch outside of the state in of that 30 Johnson, Simon, and James Kwak. 13 Bankers New York: Pantheon Books, 2010.
24 31 This made studying regional financial activity more difficult due to the rapid industry changes that followed and limitations resulting from those changes. 32 It Regardless, t he Fed Chairman A lan Greenspan strongly supported deregulation 109 and its legislative history make clear that section 109 is to be administered without imposing additional regulatory does not impose additional data reporting requirements or require banks to produce, or 33 Research has been conducted to examine the meeting of deregulatory policies a nd the financial industry. Cornett, Ors, and Tehranian (2002) find an improved performance of commercial banks surrounding the Section 20 subsidiary, as banks are allowed to diversify financial activities relative to the added risk factors. Cyree (2000) ex amines an increase from 10 percent to 25 percent of total revenues from investment banking and finds larger banks benefitting more than smaller ones. 34 Further deregulation continued under the repeal of a portion of the Glass Steagall Act of 1933, which pr ohibited banks from engaging in investment banking. Although there were loopholes that allowed banks to indulge in investing activities, such as Section 31 Federal Reserve "Section 109 of the Riegle Neal Interstate Banking and Branching Efficiency Act." Accessed April 8, 2012. http://www.federalreserve.gov/boarddocs/supmanual/cch/sec109.pdf. 32 Kozlowski, Paul J. "Financial analysis after the Riegle Neal Interstate Banking and Branching Efficiency Act." Journal of Regional Analysis and Policy (1990): 74 75. http://jrap journal.org/pastvolumes/1990/v29/29 1 5.pdf (accessed April 4, 2012). 33 Johnson and Kwak, 13 Bankers 98 103. 34 Al Mamun, Abdullah, Kabir Hassan, and Van Son Lai. "The Impact of the Gramm Leach Bliley Act on the Financial Services Industry." Journal of Economics and Finance 28. no. Fall (2004): 333 335. http://www.springerlink.com/content/n0827468451287pn/fulltext.pdf (accessed April 3, 2012).
25 20 subsidiaries allowing nonbanking activities, the Gramm Leach Bliley Act of 1999 (GLB) was the piece of legislation actually legalizing investment banking by commercial banks. The GLB also repealed the provision under the Bank Holding Company Act of 1956 that prohibited the merge of commercial banks and insurance firms. 35 The consolidation of these differe nt sectors within the financial services i ndustry, as evident in Figure 9 eventually lead to banks being labeled by citizens, lawmakers and (corporate) pillars. 35 Akhigbe, Aigbe, and Anne Marie Whyte. "The Gramm Leach Bliley Acct of 1999: Risk Impl ications for the Financial Services Industry." The Journal of Financial Research 27. no. Fall (2004): 435 434. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=498588 (accessed April 3, 2012).
26 Figure 9 From Bank Fragmentation to Consolidation 36 Studies investigating the effects of GLB generally show an increase in wealth for insurance and brokerage firms. Hendershott Lee, and Tompkins (2002) find a significantly positive wealth effect for the insurance and brokerage industries. The lack of a positive effect for commercial banks is attributed to loopholes and laws that previously the size of the firm is the only factor explaining the cross sectional variation of wealth effect. Carow and Heron (2002) have similar findings that show banks not benefitting 36 Mother Joines, "How Banks Got Too Big to Fail." Last modi fied February 2010. Accessed April 9, 2012. http://www.motherjones.com/politics/2010/01/bank merger history.
27 from GLB though brokerage and insurance companies do. They also find foreign banks, thrifts, and finance firms received negative returns. 37 Later in 2004 the Securities and Exchange Commission (SEC) approved new rules permitting major Wall Street brokerages with a minimum of $5 billion tentative net CSE, th e amount of money designated as net capital can be reduced by as much as 30 percent in some cases. Lowering net capital is a means of increasing leverage, or debt to equity ratios. Authorities claimed the $5 billion floor would survive tests under the Base l model (an in house, risk measuring model determining necessary amounts of net capital). 38 Investment banks such as Bear Stearns, Goldman Sachs, Lehman Brothers, Merrill Lynch, and Morgan Stanley all received CSE titles by applying to the SEC. Haberman ( 1987) finds the importance of government regulation ensuring properly evaluate the implications of the merger between financial firms, and 2) the collapse of these mer ged firms can be destructive to the financial system and cause major losses. 39 Despite SEC Commissioner Paul Atkins warning that monitoring the complex models used by the brokerages under the CSE rules and verifying that net capital does not 37 Mamun, Abdullah, M. Kabir Hassan, and Son Van Lai. "The impact of the Gramm Leach Bliley act on the financial services industry." Journal of Eco nomics and Finance 28. no. 3 (2004). http://www.springerlink.com/content/n0827468451287pn/ 38 Drawbaugh Kevin. Stanford Law School Securities Class Action Clearinghouse, "US SEC Clears New Net Capital Rules For Brokerages." Last modified April 28, 2004. Accessed April 2, 2012. http://securities.stanford.edu/news archive/2004/20040428_Headline08_Drawbaugh.htm 39 Haberman, Gary. "Capital Requirements of Commercial and Investment Banks: Contrasts in Regulation." FRBNY Quarterly Review no. Fall (1987): 1 2. http://www.newyorkfed.org/research/quarterly_review/1987v12/v12n3article1.pdf (accessed April 7, 2012).
28 fall too low, w steadfast on their decision. Even after another SEC Commissioner Harvey Goldschmid 40 Figure 10 shows the share of GDP going to corporate cash flow, the profit ratio and total profits after tax in billions of dollars.Corporate cash flow is the actual amount of cash corporations generate (total inflows and outflows moving through the firm) within a time frame. Total co rporate profits after tax are the net income remaining after asset and liability inflows an d outflow are settled (Chapter two gives a more explicit definition of corporate profits after tax). As can be seen, corporate cash flow does not necessarily indicat e the progress of profits, and at times is more vital to the operation of a company. Figure 10 shows the increasing prominence financial service activity comes to play in the economy. Although the trend is staggering, it steadily moves upwards from a low er point of 7.1 in 1970. From early to late 1980s the cash flows hover around 9 percent until the Savings and Loan Crisis of the late 1980s and early 1990s, at which er a boom (high was 10.2 percent share). From start to finish, corporate cash flow as a share of GDP has nearly doubled from 6.1 percent to 12.1 percent. 40 US SEC Clears New N et Capital Rules For Brokerages
29 Figure 10 Corporate Activity & Profit A possible explanation of the increase of profit ratio oversight. This issue deals with the bidding down process major regulators are forced to participate in since government regulating offices live off of fees levied by member financial firms. Government agencies like the Office of Comptroller of the Currency and the Office of Thrift Supervision each maintain a jurisdiction of regulation over a specific area of the financial sector. However, af ter GLB effectively repealed Glass Steagall Act and the Bank Holding Company Act, newly merged brokerage firms in the financial sector fell under multiple jurisdictions. Rather than being monitored by all possible regulators, these brokerages have the opti on of choosing which government agency will
30 agencies must par take in a bidding down competition a deal also including the degree of lax regulation in order to maintain s urvival. 41 ***** Arriving back at Figure 1 (located below for convenience), there appears a clear increase in the profit share of GDP and fall in the wage share following 1999, the year GLB is passed. Additionally, the profit ratio trend gravitates up after the Net Capital Rule change was passed by the SEC in 2004, however, a break in trend soon occurs as the 2007 Financial Crisis approaches. Conversely, the wage ratio trends down from a high point of 53 percent in 1970 to just above 44 percent in 2011. The two visibly notable points of decline in the wage ratio trend occur in 1970 and 2000. In the former, the wage ratio drops from a peak of 53.1 percent share to 45.8 percent in 1994. From 1994 the wage ratio begins to pick back up again until its peak in 200 0 (48.5), after which point the trend barrels down to 44.3 percent in 2011. 41 Johnson and Kwak 13 Bankers 97 100.
31 Figure 1 Wage Share and Profit Share of U.S. Gross Domestic Product Policies protecting against problems of unequal income distribution of income inequality are highly debatable among politicians and economists alike. This happens naturally, as the mixture of equal rights (evenly distributed) and unequal incomes (dispersed) cause tension between the principals of our democracy and the principles of capitalism. 42 Ho wever, this tension is what complicates determining when income inequality has gone too far. On the subject of economic equality and efficiency, economist Arthur Okun The differentials in income are meant to serve as incentives rewards and penal ties to promote efficiency in the use of resources and to generate a great, and 43 Yet, this concept did not negate the protection of social and 42 Okun, Arthur. Equality and Efficiency: The Big Tradeoff Washington, D. C.: Brookings Institution, 1975, vii. 43 Ibid., vii.
32 political rights. Okun agreed with government regulation on the grounds that the peopl morals implied here which seek the fairness and justice that the United States judicial system, legislature and executive branches pursue to protect that cases like Citizens United versus the Federal Elections Committee cause outrage. Regardless, dealing with income inequality was stifled for a long time by the taboo ambiance surrounding the subject; anti Communist sentiments that developed during the Red Scares and Cold War halted discussion on the topics. 44 It was not until the inequality and distribution became subjects widely talked about again. The heart of income inequality touches on some essential truths in American soc iety. One, that the everyday l ife of citizens is constricted almost exclusively by time and money; and two, that time is money. Therefore, when wages are the core of income 44 This despite the fact that the question of income distribution was the center stage of famed economist Principles of Political Economy (1817).
33 Chapter 2: Formulation of the P rofit Ratio Model and Wage Ratio Model In this chapter, the basic analytical framework behind the developed multivariate models is explained. After listing the sources of data, relative units of measurement and associated equation operations used in the model estimations, variables are discusse d. In the next section, each variable is listed under one of four labels, although the categorization is not strict as there exist overlaps. The explicit definitions of the dependent variables are then presented. R easons for the chosen independent variab les and how they are operationalized are described, and a section on omitted variables is included. Explicit definitions for the variables may be found in sources by National Income and Product Accounts (NIPA), as cited by the Bureau of Economic Analysis ( BEA). 45 2.1 Data and Models The data used to estimate determinants of wage share of U.S. GDP and profit share of U.S. GDP strictly come from government datasets. Datasets for these shares of Economic Data (FRED). The BEA defines as a measure of profits from current production of organizations such as financial institutions, nonprofit organizations, Federal Reserve banks, and federally sponsored credit agencies treated 45 Evans, Donald L., Kathleen B. Cooper, and Steven J. Landefeld. Buraeu of Economic Analysis, "Corporate Profits: Profits Before Tax, Profits Tax Liability and Dividends." Last modified September 2002. http://www.bea.gov/scb/pdf/national/nipa/methpap/methpap2.pdf
34 as corporatio ns in the NIPA (file Federal corporate tax returns). Inventory and depreciation accounting practices are reflected and used for Federal income tax returns, with dividends and undistributed corporate profits included less profits tax liability. 46 The defini NIPA as consisting of monetary remuneration of employees, including the compensation of corporate officers ; commissions, tips, and bonuses; voluntary employee contributions to certa in deferred compensation plans, such as 401 (k) plans; employee gains from exercising nonqualified stock options; receipts in kin; and miscellaneous 47 Although all data is retrievable from FRED, the Board of Governors of the Fede ral Reserve System compiled the effect Federal Funds Rate and the Household Credit Market Debt Outstanding. The original source of Corporate Profits After Tax, which is used to show the profit ratio, is released by the U.S. Department of Commerce subdivisi on: the Bureau of Economic Analysis (BEA). The BEA also compiles the data on imports and exports of U.S. goods and services, Personal Current Transfer Payments, Corporate Net Cash Flow, and Compensation of Employees: Wages & Salary Accruals. Dow Jones & Co mpany is responsible for data compilation on per barrel cost of West Texas Intermediate oil a grade of crude oil used as benchmark in oil pricing. The Bureau of Labor Statistic within the U.S. Department of Labor measures the household data of annual aver ages measuring the unemployment rate and the civilian noninstitutional population from the Current Population Survey, along with business and nonfarm business productivity measures The zerodoner dummy variable designates all 46 Ibid. 47 Bureau of Economic Analysis, s.v. "A Guide to the National Income and Product Accounts of the United States." http://www.bea.gov/national/pdf/nipaguid.pdf (acc essed May 20, 2012).
35 Zero D emocratic, One R epublican). The Net capital rule change of 2004 is abbreviated as SEC, as the Securities and Exchange Commission approved the change; years preceding 2004 a re marked by zero, and 2004 to 2010 denoted as one. The Glass Steagall dummy (GS) is annually represented with zeros until 1999 to 2010, where ones signify the passage of the Gramm Leach Bliley Act, which effectively repealed the most important stipulation s of GS, denote the years. Each data series is marked by annual dates. The wage ratio estimation dates from 1959 to 2010 and the profit ratio estimate ranges from 1955 to 2010. Listed below are the models, abbreviations of utilized data and related measur ement units, and a legend to operations of variables. (1) Profit Ratio = (2) Wage Ratio = Abbreviations : CASH = Corporate Cash Flow ($ Billions) CPI = Consumer Price Index DEBT = Household Debt ($ Billions) FFR = Effective Federal Funds Rate (%) GDP = Nominal Gross Domestic Product ($ Billions) INF = The differenced natural log of the Consumer Price Index (%) M = Imports ($ Billions)
36 OIL = Price of oil per barrel ($) PRES n PROF = Corporate Profits After Tax, financial and nonfinancial SEC = Net Capital Rule change of 2004 dummy variable SP estimation TO = Trade Openness share of GDP (%) TRAN = Perso nal Current Transfer Payments ($ Billions) UNEMP = Unemployment Rate (%) WAGE = Compensation of Employee: Wages & Salary Accruals ($ Billions) X = Exports ($ Billions) Legend : Variables scaled as a percent of GDP are denoted with an asterisk. Where, TO* A lagged dependent variable is denoted by ( 1) or t 1 and placed at the end of a variable. Where, and t 1 is equal to time less one, or in this case, year minus one. Variable growth is represented by the differenced natural log. 48 Two examples are given where, such that, 48 The natural log is used primarily because of automated application by the program EViews 7.0. However, use of the natural log and the associated value of e are frequently found throughout statistics.
37 2.2 Dependent Variables (1a) (2a) In Chapter one t he comparison of the profit ra tio (1a) and wage ratio (2a) is presented from various economic perspectives and historical events. As previously stated, the growing gap between the two shares reflects changing income distributions between financial and non financial corporations earning profits and workers earning wages. Although high income laborers are categorized under the wage sha re of GDP, studies show the richest laborers have gained significant sway of growth in said share in recent decades. The climbing influence the top percent has over the wage ratio coincides with the growing profit share of U.S. GDP. The causes of a declini ng wage ratio and swelling profit ratio are investigated in this chapter. The literature from Chapter one explains theoretical models encompassing income inequality through analysis of the Gini coefficient and formal models that primarily deal with income distribution between the top one or ten percent, and the remaining bottom percent. These models contain determinants such as debt, consumption and integrity of political institutions, among others. Alternatively, the models presented here are a combinatio n of already mentioned concepts and dummy variables testing specific policy changes by financial and government agencies. A distinct difference between models from the literature and the developed models is broadness. The profit ratio and wage ratio are b roader measures merely implying an idea of income inequality since the comparison do es not measure how
38 wealth is distributed among members of society and the units are not precise regarding incomes Rather, this study reveals the magnitude of political and economic forces affecting overall profits and wages however, the interpretations of the findings are subject to question Comments by readers are encouraged and may be sent to Rebecca.Keenan@ncf.edu. Profit Ratio As mentioned before, the profit ratio is the share of GDP devoted to corporate profits after tax in billions of dollars. To scale actual profits as a percent of U.S. GDP, the operation expressed by equation (1a) is used. Nominal GDP, also termed in billions of dollars, is used to scale profits e arned after tax from financial and non financial corporations, as opposed to real GDP. This is done to circumvent the cancellation of the price level, as depicted below. & In the event of There occurs in which case the price level is cancelled out of the denominators and the nominal share of profits out of nominal GDP remains.
39 Finally, the fraction is multiplied by one hundred to give a percent: Wage Ratio The same operations that were applied to the profit ratio are applied to the wage ratio in order to produce equation (2a): Wages and salary accruals as a form of compensation for employees define WAGE in the model and are displayed in units of billions of dollars, as is GDP. Throughout be used to define GDP adjusted for the price level. 2.3 Independent Variables The dependent variables listed in equations (1) and (2) have been loosely categorized under the following headings: Political Climate, Characteristics of the Private to re flect the mixture of forces that compose the United States macro economy. 2.3.1 Political Climate PRES, SEC Variables listed under political climate are those which were carried out by the United States Federal government or participating agency. These determinants reflect
40 ideologies and policies enacted upon, and as pieces of recent history, they contribute to economic and political debates and decisions being made today. Presidential Party dummy (PRES) The Presidential Party variable is designed as a proxy to determine the overall effect of Republican Presidents on the wage ratio. This was done by creating a se ries from 1947 2011, where all years under Republican Presidents are denoted with the number one. All years under Democratic Presidents are assigned a zero value. Historically, the Republican Party has been concerned with controlling inflation in the eco nomy, while the Democratic Party focuses more on curbing unemployment despite resulting effects on inflation. 49 As lowering inflation is an easier target to achieve via monetary policy and unemployment a more difficult issue to tackle due to its nature, the relationship between the Presidential dummy variable and the wage ratio is assumed strongly negative. This assumption is represented as Other evidence of a negative relationship resides within the Republican Party's history of union busting. Where unions are broadly defined as organized groups having wage bargaining power, the Reagan Administration took full force in fighting the trade union, Professional Air Traffic Controllers Organization in 1981, after 12,000 striking federal air traffic control lers were fired. 50 More recently, Wisconsin Governor Scott Walker (R) made public his willingness to initiate the National Guard defenses against opposition to his proposal that would eliminate collective bargaining rights for state 49 Pearson, "Economic Policymaking." Accessed April 22, 2012. http://wps.ablongman.com/long_edwards_ga_12/33/8517/2180597.cw/index.htm 50 Hirsh, Stacey. "Reagan preside ncy pivotal for unions." The Baltimore Sun June 8, 2004. http://www.baltimoresun.com/bal bz.unions08jun08,0,1073570.story (accessed April 22, 2012).
41 employees. Ohio, Iowa an d Indiana are looking to curtail or eliminate collective bargaining rights as well. 51 The Citizens United versus the Federal Elections Committee ruling by the conservative leaning Supreme Court in 2010 is another instance increasing corporate power over wag e earners. The ruling, in favor of Citizens United, overturned a ban on corporate (nonprofit and otherwise) and union spending related to political motives during elections. Campaigns to advertise for, or against, a politician by corporations or unions is was received by conservative groups while liberals received just over half that amount ($94 million), helping Republics win back the House of Representatives in the 2010 election s. 52 As a last note on the predicted negative effect of the Presidential Party dummy on the wage ratio, Political Scientist Larry Bartels book Unequal Democracy takes an displays projected income inequality under Republican and Democratic Presidents, 1947 2005. The growth of inequality (a ratio of the 80 th percentile of income earners to the 20 th percentile) is higher for Republican Presidents at all times than for Democr atic Presidents. Further, the increase in income inequality for Republican Presidents rises from just over a 3.0 income ratio of 80 th /20 th to over 5.0. The trend of the 80 th /20 th 51 Associated Press, First. "Wisconsin Governor Scott Walker Readies National Guard Against Unions." The Huffington Post February 11, 2011. http://www.huffingtonpost.com/2011/02/11/scott walker unions wisconsin national guard_n_822225.html (accessed April 4, 2012). 52 France Presse, Agence. Republicans attack unions to hurt Democrats: analysts, "The Raw Story." Last modified February 19, 2011. http://www.rawstory.com/rs/2011/02/19/republicans attack unions to hurt democrats analysts/.
42 income inequality ratio under Democratic Presidents slightly lowers from above 3.0 to around a ratio of 2.8. 53 As for the effect of PRES on the profit ratio, information and data referenced from Unequal Democracy supports the supposition of a positive correlation ( ). Bartels graphs a historical relationship bet ween partisanship, election year Republican incumbent (or successor) spent at least slightly more than his Democratic challenger, while every Democratic incumbent (or successor) s pent at least slightly less campaign increased the probably of swaying an undeci ded voter towards the candidate who spent the funds, by almost 4 percentage points. 54 Net Capital Rule change of 2004 dummy (SEC) The Net Capital Rule Change of 2004 is a tool for the major banks that survived the 1990s and early 2000s mergers and assumes a negative value for wages ( ). of tentative net capital are not exempt, and are required to value their securities at market These discount values help determine the liquidation value of a broker 53 Bartels, Larry. Unequal Democracy: The Political Econo my of the New Gilded Age New York: Princeton University Press, 2008. 54 Ibid., 118 20.
43 order to decide if the broker dealer is in possession of enough liquid assets to pay all non all obligations due to clients may be ensured in the case of a delayed liquidation of assets. 55 This change essentially allows big banks to bet bigger, a feature leading to the The Net Capital Rule dummy is assumed to hold a power ful positive effect on profits ( ) since the change was a liberalizing agent in the private sector, allowing corporations to augment more power via greater leverage. 2.3.2 Characteristics of the Private Sector CASH*, PROF* t 1 TO* The shares of U.S. GDP represented in this section are the chosen aspects of the private sector testing for income inequality on the wage ratio and profit ratio. These variables being tested have been picked in part for the comparatively rapid growth or increasing significance in globalizing economies. Net Corporate Cash Flow share (CASH*) The share of Net Corporate Cash Flow is a variable valuing the economic activity of corporations domestically and internationally. As mentioned in the last chapter, Net Corporate Cash Flow is not a measure of profit s o much as an examination of how much 55 Wannisky, Kathleen E., and Jill M. Peterson. Government Accountability Office, "Alternative Net Capital Requirements for Broker Dealers that are Part of Consolidatio n Supervised Entities." Last modified June 25, 3004. Accessed April 23, 2012. http://www.gao.gov/decisions/majrule/d04896r.pdf.
44 money is being handled by corporations. The anticipated relationship between the cash flow and wage ratio is moderately negative ( ). Reasons for this are associated with the passage of GBL and change in the Net Capital Rule, which give more leverage to corporations (some of which used to be banks, like JPMorgan Chase) rather than other groups such as unions. Lagged Profit Ratio (PROF* t 1 ) The lagged profit ratio, like the lagged wage ratio, represents the inertia eff ect from levels of profit of the previous year on current profit levels. This variable is assumed to have a large positive effect ( ) on the profit ratio since corporations exist in a much more liberalized economy, and profit ratio has increased by almost 7 percent since a low point of 3 percent in 1986. Additionally, it is expected year t 1 is closely related to year t and that increasing domestic productivity influence positivity. Trade Openness (TO*) The next variable, trade openness share, is the sum of the imports and exports of goods and services scaled as a percent of GDP. The variable effectively measures globalization in the U.S. economy. Because the dollar is very strong compared to other curren cies it can be expected that imports exceed exports since it is cheaper for Americans to buy foreign goods and more expensive for foreign countries to buy domestic goods. In this respect, the trade openness wage share correlation is anticipated as
45 Amer ican firms want to buy cheaper foreign goods to keep costs down, thereby reducing demand and affecting wages. Furthermore, as technology enhances the ability to export services, the American economy whose service sector has been increasing within the past 30 years, may find greater competition in a global market. The added pressure of foreign competition lowers wages. Unlike the effect trade openness has on wages, the effect on the profit ratio is predicted as As domestic regulations on the financial in dustry decline, corporations have a greater ability to throw their weight around in the country and around the world. Giant corporate entities recognize potential revenue streams from abroad and seek to invest in new markets. Moreover, corporations take ad vantage of the cheaper cost of resources found abroad in order to aid growth. 2.3.3 Consumerist America DEBT*, [ln( OIL )] WAGE*t 1, TRAN*, UNEMP This section aims to give credence to the actions of American consumers, the backbone of the U.S. economy. The above variables consider fundamental sources of income and payments for a majority of Americans. Household Debt share (DEBT*) Household debt a s a share of GDP presumes in the wage ratio. The direction of the relationship is determined on the foundation that debt makes wage earners poorer when in excess. Excessive debt is termed as an interruption of
46 lifetime consumption smoothing. As debt rise s, consumers must set aside a larger amount of their salaries to pay back debt, which includes bank determined interest rates on loans. Between household debt as a share of GDP and the profit ratio, is assumed. Evidence previously mentioned shows volumi nous debt occurred as a result of climbing consumer spending despite a falling wage share. This indicates borrowing, which implies that consumers as a whole are paying increasing amounts of interest on mounting borrowed funds. As debt holders pay back loan s received by financial intermediaries, bank determined interest rates are received as revenue to lenders. Growth in Price of Oil per Barrel ( [ln( OIL )] The coefficient of growth in the price of a barrel of oil is figured to be a large positive ( ). As the cost of oil per barrel rises, revenues for oil corporations also rise, which implies an increase in profits. Additionally, while the Obama Administration has been focusing on green energy projects, the U.S. economy remains highly dependent on oil to produce and transport goods and services. Therefore, it could be assumed that even small increases in the cost of oil increase corporate profits. Wells Fargo Chief Portfolio Strategist Brian Jacobsen reports his findings for a 1 percent increase in oil prices to lead to a 0.83 percent increase in corporate profits. 56 Lagged Wage Ratio (WAGE* t 1 ) The once lagged wage ratio predicts a positive correlation with wage share 56 Jacobsen, Brian. "Oil prices and profits: Three ways to look at energy prices." Market Insights (blog), April 12, 2012. http ://www.wellsfargoadvantagefunds.com/wfweb/wf/mobile/tl/ena_20120412.jsp (accessed April 20, 2012).
47 ( ) This is based on the idea of inertia discussed in above in lagged profit ratio. In this case, wage ratio is being predicted based on past values of itself (the wage ratio). The implementation of the lagged variable categorizes the model as autoregressive or dynamic. Personal Current Transfer Payments (TRAN*) These transfer payments, distributed by the Federal government, assume relating to the wage share. This presumption is based off the belief that social welfare programs such as unemployment insurance and social security provide aid to those Americans seeking employment or medical attention. A countering point of view would c laim a negative correlation between this independent variable and the dependent variable on the basis that welfare programs are disincentives that promote freeloading. The Unemployment Rate (UNEMP) In view of the underconsumptionist concept, the demand side approach as defined in Bhaduri and Marglin (1990), the unemployment rate is predicted to be negative ( ) Underconsumption dictates low wages as a causality of depressed effective demand, which negatively effects unemployment and output in the short r un. As wages decrease, there is a decrease in consumption that drives down the demand for goods, ultimately provoking firms to try and be more cost efficient via layoffs or freezes in hiring.
48 FFR, INF, SP The value of variables predicted here echoes monetary policy by the Federal Reserve Bank (Fed) over the past half century. The quasi governmental agency plays a significant role in the financial industry though it lacks the ability to target specific areas of the U.S. economy. Monetary policy was pivotal during the 2007 Financial Crisis, where some pundits claim the $800 billion bailout approved by Congress and operated by the Fed prevented an outcome comparable to the Great Depression. The Effective Federal Funds Rate (FFR) As the first variable listed in the profit, the effective Federal Funds Rate (FFR) is not deemed powerful though it is thought to maintain a negative relationship ( ). The predicted negative beta value assumed between the dependent and independent sums of money in the federal funds market. Therefore, a higher FFR is more costly to banks and a lower FFR allows for less expensive trading. Additionally, the associati on between interest rates and the return on investments/assets is made known. Although the current FFR is essentially 0 percent, thereby have almost no effect on profits, the high interest rates set during the 1980s may skew the relationship to be farthe r below zero. Despite historical instances of high interest rates however, the coefficient to this beta is still predicted to be weak.
49 Inflation (INF) Inflation, measured as the differenced natural log of the Consumer Price Index, is predicted to be negatively correlated with wage share ( ) As inflation increases, the purchasing power of nominal wages declines, thus decreasing the real wage. The Fed is able to more or less accurately target inflation via monetary policy. The Fed will tend to raise t he interest rate if inflation is above target. This encourages consumers to hold less cash and keep a high amount of bank deposits, as the return dictated by the interest rate is greater. The deposit of cash into bank accounts contracts the money supply, t hereby decreasing inflation. To raise the interest rate, the Fed contracts the supply of U.S. treasury bills and other specific assets via buying them off the market, which is how the money supply is contracted. Subprime Mortgage Crisis dummy (SP) Prof its are recognized as being highly susceptible to fluctuations in the economy and consequently are called volatile. Therefore, when a wave of subprime mortgage delinquencies and foreclosures occurred in 2007, profits dramatically dropped. Examining Figure 1 from the first chapter, it is seen that profits quickly rise following the sharp drop. The sharp movement has everything to do with the $800 billion bailout and following rounds of quantitative easing that gave banks the capital necessary to continue fun ctioning. Due to the volatility of profits, there is some ambiguity in predicting the value of the coefficient. The model tests the estimation of the values assigned the number one;
50 from 2007, when the crisis occurred, until present. All preceding years h old a zero value. However, studying the profit trends seen in Figure 1 and other figures presented in Chapter one, the SP dummy is assumed The addition of the SP dummy arose after the first round of testing upon consideration of the wave of fo reclosure characterizing the burst of the housing bubble. The dummy is highly insignificant in the wage ratio model (90.49 percent), however its addition improved the profit ratio model. The differences are highlighted below in Table 3
51 Table 3 Profit Ratio Model Improvement ( SP dummy) Adding the SP dummy improved the probability values for FFR, DEBT*, TO*, PRES and SEC variables. There is also a decrease in the residual sum of squares (RSS), which indicates a smaller range of variance in the data that is not explained by the Variable Coef ficien t (w/o) Std. Error (w/o) t Stat (w/o) Prob. (w/o) Coefficien t (w/sp) Std. Error (w/sp) t Stat (w/sp) Prob. (w/sp) Constant 3.741005 0.82800 2 4.51810 8 0.000 0 3.480337 0.78047 5 4.45925 5 0.000 1 FFR 0.075079 0.03386 0 2.21735 2 0.031 4 0.085307 0.03189 8 2.67435 6 0.010 3 DEBT* 0.013936 0.01557 5 0.89472 8 0.375 4 0.017537 0.01463 3 1.19846 3 0.236 7 PROF*( 1) 0.528666 0.09726 4 5.43539 0 0.000 0 0.572433 0.09236 1 6.19775 1 0.000 0 log (OIL) 0.837273 0.37635 1 2.22471 6 0.030 8 0.618969 0.36076 6 1.71570 6 0.092 8 TO* 0.027513 0.08374 8 0.32852 0 0.743 9 0.064857 0.07950 6 0.81575 5 0.418 8 PRES 0.396504 0.18913 2 2.09644 3 0.041 3 0.418837 0.17717 8 2.36393 8 0.022 3 SEC 2.078189 0.59751 2 3.47807 2 0.001 1 2.678330 0.59900 9 4.47126 6 0.000 0 SP -----------------------------------1.273483 0.45577 4 2.79411 0 0.007 5 RSS 17.12973 0.853534 0.832175 39.96024 48 0.597385 1.458230 1.664288 14.68967 0.874398 0.853019 40.89963 47 0.559058 1.458230 1.952682 R2 Adjusted R2 F stat d.f. S.E. Regressio n S.D. dependen t var Durbin Watson stat
52 regression. For this reason it is logical that the R squared and adjusted R squared are higher. T he increase in R squared shows a better fit to the actual trend and the increase in adjusted R squared expresses that the model has been improved by the added variable. The decrease in the standard error term is a good sign showing the estimated standard d eviation has been reduced. Lastly, the Durbin Watson statistic is closer to 2, from a range of 0 4. A value close to 2 indicates no serial correlation. 2.4 Omitted Variables When creating the models, a Glass Steagall dummy, Education, Union Coverage and the Exchange Rate were variables initially included. The Glass Steagall dummy measured the effects of the passage of the Gramm Leach Bliley Act of 1999 with a one value assigned to 1999 until present and a zero value a ll years before; Education was to be measured by the percent of the population attending post secondary schools; Union Coverage, which is different than membership, to be measured as a percent of all employed labor force participants; and the Exchange Rate of the U.S. dollar to the British pound. Up until the night of April 27, 2012, a Glass Steagall dummy variable ( GS ) was included in the wage ratio model. The estimated variable proved to be highly significant at essentially 0 percent probability and with a strong coefficient of 1.060130. However, upon examining the dummy, where all years prior to the passage of the Gramm Leach Bliley Act of 1999 were supposed to be assigned a zero value it was recognized that the data was input incorrectly. Instead of a one value being assigned to 1999 until the present
53 time, a one value was assigned from 1989 until 2011. When correcting this error and properly assigning the values, the significance shrunk to 60.27 percent with a coefficient of 0.164468. Fortunately, thi s did not greatly alter the significance or coefficients of the other independent variables listed in the wage ratio model. Although correcting for this mistake resulted in an insignificant econometric statistic, there may remain some economic value in t he incorrect dummy. It was mentioned in the first chapter that the Gramm Leach Bliley Act was passed in 1999. However, according to 13 Bankers, this legislation merely legalized certain loopholes on 20 subsidiaries in the Glass Steagall Act. It would therefore make sense that the dummy testing for years following 1989 would capture more significance. Still, in this scenario the issue that remains is when banks began taking advantage of the Section 20 subsidiaries loophole, or if the practice had been commonplace since the passage of Glass Steagall in 1933. The Education variable was to be exclusively added to the wage ratio model. Over the decades, job opportunities have become increasingly tied to the level of education a degree in order to be qualified for jobs that do not require critical thinking or amount to challenges on par to temp work. Regardless of to pay for school, in order to obtain an average paying job, higher education is a must. The relationship between the education and the wage ratio would have been assumed to be ambiguous because of the conflict na ture of the cultural aspect and the financial aspect, of student loan debt.
54 Unfortunately, despite the cultural relevance of education and the implication the variable has on a workforce entering a globalized age, the probability value for this variable co nsistently showed to be insignificant. Recognizing the lack of importance the variable apparently played in the model, Education was removed. Union Coverage was initially applied to both wage ratio and profit ratio models. Unions provide certain health an d pay benefits private companies and corporations do not. Union coverage implies a lower cost to the worker albeit no voice in union decisions such as contract proposals. Union membership does grant this right, as signified by the higher dues imposed. By reducing healthcare costs and maintaining certain standards of wages, union coverage would have been assumed to hold a positive relationship with the wage ratio. However, as the probability value consistently turned up insignificant, as Education had, the variable was excluded. Finally, the Exchange Rate was a desired variable however finding a unit of to the pound, another comparable currency highly valued throughout the world, the measurement of effects would have been too narrowly defined for broad economic aggregates like the wage ratio and profit ratio. Therefore, this variable was eliminated.
55 2.5 Methodology One of the first considerations when estimating a time series regression is transforming variables that exhibit linear or exponential trends into stationary series. The non stationary variables in this thesis include: government transfer payments, imports, CPI, corporate cash flow, total financial and non financial profits, wages and salary accruals, and household debt. One method of transforming datasets into stationary series is to take the first difference. Taking the natural log of a series is another m eans of obtaining stationarity. In order to account for inflation, the CPI was transformed using the natural log, then multiplied by 100 to give a percentage. A benefit of measuring variables in percentages is the basic context provided of 10 (or 100) pe rcent. Variables such as household debt, government transfer payments, wages and salary accruals, total profits, corporate cash flows, and the sum of imports and exports, were all divided by GDP and multiplied by 100 to make the data trend stationary, an d in percentiles to show the rate of change. Identifying the independent variables in terms of the rate of change allows for continuous time variables to be represented and calculated as discrete variables. This is practical since economists are unable to instantly measure continuous variables such as GDP. 57 The rate of unemployment was left in the original form, as a percent of the labor force, since the trend is stochastic. The same idea was applied to the FFR. 57 Bar, Michael. San Francisco State University, "Growth Rates." http://bss.sfsu.edu/mbar/ECON560/Growth Rates.pdf
56 Lagging the profit ratio produces a statio nary trend, where the correlation between the lagged independent variable and the dependent variable relies on the length of the lag (year minus 1 in this case) and not when the series started. 58 Although the wage ratio is lagged, the series trends down and therefore holds an initial stationary state. Lastly, the growth in the price of oil per barrel is determined through two mathematical processes. First, differencing the series, which displays the rate of change in price. Second, taking the natural log of the differenced series renders the growth in price. The econometrics program Eviews 7.0 used to regress both multivariate equations is set to automatically use the natural log, opposed to log base ten. The natural log is preferential in econometrics and statistics since the value of e is often found throughout various aspects of statistics. Ordinary Least Squares (OLS) is the method for estimating the unknown parameters in the multivariate equations. OLS will be used because it is an efficient (least variance), unbiased estimator. Gauss Markov assumptions stipulate the conditions for efficiency and unbiasedness. OLS works only when the Gauss Markov assumptions satisfied. These assumptions, whose definitions are borrowed heavily from Running Regressions 59 include: A zero value mean of error terms. No autocorrelation, which specifies the covariance between error terms is zero. 58 Ramu Ramanathan, Introductory Econom etrics with Applications, 5th Edition (United States of America: Michael P. Roche, 2002), 472. 59 M.C. Baddeley, and D.V. Barrowclough, Running Regressions, (New York: Cambridge University Press, 2009), 25.
57 Homoskedasticity: the variance of the error is constant across all observations. Correct model specification, where there are no omitted or excess variables and proper functional form. Exogeneity: the error term and explanatory variables have no correlation. Lineaity in the parameters such that, the model capturing the data generating process is linear in the parameters because OLS is a linear estimation method. Additionally, it should be noted the coefficients being estimated by the model are the and that 0 is an estimated constant term representing the intercept. The symbol is the error term. ed to identify functional form and heteroskedasticity. In order to compute the heteroskedasticity and autocorrelation consistent (HAC) standard errors, the Newey West test is conducted. Newey West is favorable to the White test since the standard error of computation of Newey West standard errors are robust to arbitrary departures from homoskedasticity. 60 For the reason that standard tests for heteroskedasticity suppose independence of the errors, it is important to test for serial correlation. If the error terms are serially correlated, the White test and Newey West test will not generally be accurate. 61 Therefore, the Breusch Godfrey test will be applied. 60 Jeffrey M. Woolridge, Introductory Econometrics: A Modern Approach, (United States of America: South Western College Pub, 2006) http://fmwww.bc.edu/ec c/F2007/228/EC228.f2005.nn12.pdf 61 Ibid.
58 Chap ter 3: Interpretations and Conclusions 4.1 Interpreting the Results Table 4 lists the coefficients for each independent variable along with the accompanying p values in the Profit Ratio model and Wage Ratio model, bother put through the HAC filter. The HAC filter is preferred over the White test for serial correlation in this stu dy since White assumes the residuals of the estimated equation are serially uncorrelated. The LM test proved otherwise of this assumption for the Wage Ratio model, supporting the decision to apply HAC. The more general covariance estimator proposed by Newe y and West (1987) is consistent in the existence of heteroskedasticity and autocorrelation of unidentified form via interpretation of the standard errors (S.E. of regression) 62 Furthermore, using time series data the S.E.s are robust for random autocorrela tion (up to the order of the chosen lag) and random heteroskedasticity. Additionally, although HAC has no effect on the estimation of coefficients, the S.E.s of the equations will change 63 62 The standa rd error of the regression is a summary measure based on the estimated variance of the residuals around the dependent variable. 63 EViews 7 User's Guide United States of America: Quantitative Micro Software, LLC, 2009 ; 33 5
59 Table 4 Coefficients and Significances Variable Profit Ratio Model Wage Ratio Model Constant 3.480337 (0.0003) 19.81839 (0.0002) FFR 0.085307*** (0.0018) DEBT 0.017537 (0.2563) 0.0 34201 (0.1 106 ) PROF( 1) 0.572433*** (0.0000) log(OIL) 0.618969* (0.0697) TO 0.064857 (0.3469) 0.114070 (0.3436 ) PRES 0.418837* (0.0563) 0.329617* (0.0903) SEC 2.678330*** (0.0004) 0. 675078 (0.0467 ) ** SP 1.273483*** (0.0023) UNEMP 0.194718 *** (0.0049 ) log(CPI) 3.729617 (0.4014 ) CASH 0. 389244** (0.0196 ) TRAN 1.078896 (0.4821 ) WAGE( 1) 0. 709365 *** (0.0000) = 10% p value significance, ** = %5 p value significance, *** = 1% p value significance. P value is in parentheses.
60 Table 5 Accompanying Estimation Statistics Profit Ratio Model Sample (adjusted): 1955 2010 Included observations: 56 after adjustments HAC standard errors & covariance (Bartlett kernel, Newey West fixed bandwidth = 4.0000 ) Wage Ratio Model Sample (adjusted): 1959 2010 Included observations: 52 after adjustments HAC standard errors & covariance (Bartlett kernel, Newey West fixed bandwidth = 4.0000) S.E. of regression 0.559058 0. 420489 RSS 14.68967 7.426056 R 2 0.874398 0.973960 Adjusted R 2 0.853019 0.968380 F stat 40.89963 174.5442 d.f 48 43 S.D. dependent var 1.458230 2.364682 Durbin Watson stat 1.952682 2.241289 With the HAC correction filter in place, I would next like to claim 10.00% as the significance level of p values for this study. Although 5.00% is the standard benchmark in most statistic and many econometric models, a 10.00% significance level lends to mo re variables being captured by the model and reduces omitted variable bias. Beyond reasons of leniency, the sample size in both models is not very large. Although the sample size may not be considered low, a more forgiving p value will be utilized since st andard errors are less likely to be small, making fewer coefficients significant.
61 Actual and Fitted Models with Residuals Figure 1 1 Profit Ratio Model Figure 12 Wage Ratio Model
62 The PR and WR trends and residuals can be seen above. The fitted WR trend line lies more tightly around the actual trend line, than is seen in the PR model. However, this may likely be attributed to the evidence of serial correlation in the WR model. The HAC filter was used since I did not want to include the assumption that the residuals of the estimated equations are serially uncorrelated. Furthermore and more importantly, HAC is consistent in the existence of heteroskedasticity and autocorrelation of u nknown format through interpretation of the standard errors. Residuals are also presented in the graphs, helping to explain the estimates of experimental error found by the measured distances from the observed responses to the predicted responses given by the PR and WR models. Well behaved residuals give the appearance of being about normal and more or less independently distributed with a mean of 0 and some constant variance. The residuals in each graph appear noisy, a desirable trait in residuals as it represents the regression model to predict a response by some kind of structure that should be accounted for in the model. The standard error in the WR is 0.377017, a r ange smaller than in the PR model, 0.559058. Both residuals appear noisy, a good sign that the models exhibit constant variance around the mean of 0.
63 Effect Federal Funds Rate Beginning with the first variable shown, the FFR is highly significant in the PR 64 (0.18 percent) albeit showing a relatively small negative effect ( 0.09 rounded) on the dependent variable. This result is about what I expected, perhaps a little bigger, since the FFR is an interest rate banks pay to exchange funds if and when the condition of borrowing funds occurs. Therefore, banks may not always need to use the Federal Funds market and even so, will only pay an interest rate whose variance depends upon the motives of the Federal Reserve. Household Debt share of GDP The househo ld debt share of GDP (DEBT) is not significant for either the PR or WR (25.63 percent and 11.06 percent), though shows small coefficients of 0.02 and 0.03 (rounded to the hundredth), respectively. Although the effect of debt in the WR is not significant, the p value does come close to the significance standards I applied, closing in on 11 percent. I found the values of these coefficients surprising for both models. Before estimating the model I correctly predicted the household debt share would have a neg ative relationship with the wage ratio. However, I assumed the numerical value would be higher based on the work of Kumhof and Ranci re, which shows U.S. debt rising as a result of the growing gap between increasing consumption and declining wages. 64 From here on out, I will
64 I was further surprised by the small, yet negative DEBT coefficient for the WR. Initially I though some noticeable portion of profit share came from interest paid to banks by Americans living on credit. Upon further consideration, it is possible that the estimat ed coefficient has captured the long term effects of excessive bank lending (that which includes the wave of foreclosures in the housing market following the Financial Crisis) on the overall PR. Despite these possibilities, the statistical insignificance o f the coefficients shown by the p values jeopardizes my predications and result interpretations. The Inertia of the Profit Ratio Moving onto the lagged profit ratio variable, the coefficient is strong at 0.57 and the p value highly significant at ne arly 0.00 percent. This result is in accord to my prediction that the inertia offered by a lagged dependent variable in the right side of the of profits in the last ye ar contributes to how profits grow in following years. I particularly assumed this to be the case since several modifications to laws limiting financial liberalization were removed in recent decades by Congress and the Securities and Exchange Commission, t hereby allowing big banks and corporations more freedom to integrate themselves into every part of the economy in attempt to raise profits. Growth in Oil Prices The next variable to be discussed, growth in the price of oil, shows particularly interesting resul ts in this study. The variable is highly significant (6.97 percent) in this
65 context. What is more exciting is the size of the coefficient, standing at 0.62 (rounded). This implies a one dollar change in the price of oil has a 0.62 effect on the PR. Part of what makes this result interesting is the relationship between the gas and credit card companies. When an individual pays at the pump by credit card the most common way people pay in current times they are also paying a fee to credit card companies The fee is essentially the cost of convenience for gas stations to be able to accept payments by customers using credit cards. One example given in a video by KhanAcademcy.org shows customers paying $0.05 to credit card companies for every gallon of gas bought. However, oil producers are the biggest gainers from the growth in oil prices since oil is not only used to produce gas but other sorts of oils found to lubricate car engines and in a multitude of other products. 65 One of the most important features noted is the graph of short run supply and demand of the pr ice of oil versus the quantity. As seen in the graph below, price may change substantially within a given range due to limited supply while the demand will remain the same since oil is so an integ 65 Khan, Salman. "Breakdown of Gas Prices." Macroeconomics Khan Academy. 2012. Web, http://www.khanacademy.org/finance economics/macroeconomics/v/breakdown of gas prices.
66 Trade Openness share of GDP Trade openness as a variable in the PR model came up with a coefficient of approximately 0.06 and a p value of 34.69 percent. In the WR model, approximately 0.11 coefficient and 34.36 percent. Although the econometric significance (or rather, lack thereof as the p value indicate) of the trade openness variable allows a discard of the in both models, I believe the economic significance should not be tossed out as well. As mentioned in the second chapter, trade openness is a variable highly representative of globalization as a process, showing the economic activity via inflows and outflows of goods and services. Throughout different stages of the process Americans can exp ect to see many sectors in the economy adjusting, and therefore having sometimes positive and other times negative reactions to globalization. These reactions will depend on what stage of globalization the sector is experiencing.
67 Economist Tyler Cowen believes there are three forces that will combine to make the United States a global export powerhouse, as it had been for the latter half of the 20 th century. First, despite the large dent in American manufacturing employment by the mobility of capital se eking lower production costs abroad, the revolution of artificial intelligence and computing power will (eventually) save the day. Currently, factories are seeing more software driven machines replacing factory worker jobs. This effect of globalization wil l hurt employment levels and ultimately wages, as technology makes capital more affordable. This effect will be relatively brief however, as this same technology is helping developing countries rise to an economic level closer to that of the US. Economic growth of foreign countries will shift demand away from construction and other goods related to infrastructure development to higher quality outputs that American exporters are geared toward producing. Secondly, Cowen states that it is important to reco gnize that a growing reliance on smart machines is largely why and how economies are globalizing. A striking insight shared by this observation is that as reliance grows, more domestic wage rates become labor in manufacturing, but China is too, even as its manufacturing output is rising. The fact that Chinese manufacturing employment is falling along with ours means that both our higher wages and their lower wages are becoming less relevant for the locat ion of manufacturing decisions. The less manufacturing has to do with labor costs and relative wage levels, the greater the 66 66 Cowen, Tyler. "What Export Oriented America Means." The American Interest May/June 2012. http://www.the american interest.com/article.cfm?piece=1227 (accessed April 22, 2012 ).
68 are the recent discoveries of very large shale oil and natural gas deposits that will generate foreign demand. 67 This process will take time, and the positive effects may not be seen immediately. But for the initial period of time, wages will decrease as compute rs become even more powerful and take a bigger role in manufacturing. Conversely, the increased use and reliance on technology will aid corporations and companies looking to enter into new markets. Even the President of the Atlanta Federal Reserve Dennis Lockhart spoke to me personally in February 2012, about how the structure of the American job economy is shifting. Lockhart said point tradable sector 68 69 This means that the tradable goods and services produ ced at home which can be sold abroad, creates a gap between rising high wages in the tradable sector and low wages in the non tradable sector. Clearly, if the American economy wants to fulfill these predictions, it is of the upmost importance to fund educa tion, as many of the exports being produce by our country (now and in the future) are mainly services requiring a high degree of education and knowledge. 67 Unfortunately for environmentalists fracking is the method used to account for 20 percent of domestic natural gas production, a number predicted to rise over coming decades, possibly accounting for half of all US natural gas output. Fracking can leak gas into water tables, a hazardous threat. Cowen briefly discusses likely outcomes and realities of this debate in s Oriented America 68 The non tradable sector includes education, construction and food service. 69 Keenan, Rebecca. "Distribution's out of the dog house." The Catalyst March 2, 2012. ncfcatalyst.org?s=distribution's out of the dog house&x=0&y=0 (accessed April 22, 2012).
69 For all these reasons and more, I decided to retain the trade openness variable. The definition I prov ided for trade openness is a vague measure, possibly providing an explanation for the poor p values. The broad character of the variable can be contributed to current statistics overstating the size of the American trade deficit. Items counted as imports m ay include sales of the iPads or iPods since these devices are shipped in from China. However, much of the value in the products is developed by design and retail skills that current measures do not recognize. 70 Presidential Party dummy Unequal Democracy was the inspiration one of the most influential factors of economic performance is implemented policy a n aspect economists tend to indirectly or subtly recognize in research. It was exciting to find the PR model generated a highly significant coefficient for this dummy that tests the effects of Republican Presidents. Crossing into the 5 percent range of s ignificance, the coefficient stands at 0.418837, or simply 0.4. The value of the coefficient implies that for every year a Republican President was in office, profit share declined by 0.4 percent. For good measure I exchanged this dummy variable with a dummy testing for effects of Democratic Presidents. The results showed virtually no changes except that the value of the coefficient under Democratic Presidents is positive. I found this convincing 70
70 as Bartels shows income inequality increasing under six R epublican Presidents during the second half of the 20 th century: Eisenhower (1953 1961), Nixon (1969 1974), Ford (1974 1977), Reagan (1981 1989), George H.W. Bush (1989 1993), and George W. Bush (2001 2009). Income inequality under four of the five Democra tic Presidents (Carter not included) Truman (1945 1953), Kennedy (1961 1963), Johnson (1963 1969), and Clinton (1993 2001) decreased. General intuition dictates stronger profits if at disparity of income between the majority of Americans and a small percent of Americans, like there 71 However, as this coefficient represents the average effect over a period of years curiosity unfolds around the Clinton Administration. Legislation for the Gramm Leach Bliley Act was signed into law by President Clinton in 1999. To many, this act effectively started the path of banks gambled bigger and won bigger, eventually b 2001 there was an unorthodox, bond market friendly relationship established with the Democratic Party by Robert Rubin, Director of the newly formed National Economic Council whose cred entials at the time were 26 years working at Goldman Sachs. 72 In order to truly decipher the meaning behind the dummy coefficient, it is necessary to understand when the growth of profits signals actual growth or underlying economic problems, as profit rate s did preceding the Great Recession. 71 Bartels, Larry. Unequal Democracy: The Political Economy of the New Gilded Age New York: Princeton University Press, 2008, 36. 72 Simon Johnson and James Kwak, 13 Bankers (New York: Pantheon Books, 2010), 97 100.
71 The WR model finds also finds the PRES dummy to be significant at 9.03 percent and with a relatively powerful coefficient of roughly 0.33. This indicates that the overall effect of Republican Presidents on the wage share of U.S. GDP. I was expecting this coefficient to be closer to weight in reality were this high, I would assume there would have been a greater focus in recent decades on the declining wage share. Net Capital Rule Change of 2004 dummy Again, it was intriguing to find the SEC dummy as highly significant (nearly 0 percent) and with the strongest coefficient in the PR model. Eviews estimated a coefficient of 2.678330 or 2.7, rounded. This means that on average, every year after the Securities and Exchange Commission passed the Net Capital Rule Change, the profit share of GDP rose by 2.7 percent. These results can easily be tied in with the conclusion of the Presidential Party dummy variable, and although I am not in line to decided causality, the correlation is clear. The rule change had a very positive effect on profit share, results that support my assumpti on s in Chapter two The SEC coefficient in the WR model also significant, lying just below the 5 percent range. The coefficient value stands roughly at 0.68. This came as a surprise to me, as I was expecting a negative value for SEC in the WR model yet thi s coefficient is positive and strongly so. A possible conclusion to this result is that perhaps the average
72 bank accounts and assets before the Financial Crisis, and that these returns outweigh the effects on the wage ratio after the 2007 crisis. Subprime Crisis dummy The third and final dummy in the PR model did not fail expectations. The estimation shows a highly significant, almost 0 percent probability value, as we ll as a strong, impactful coefficient of 1.273483 (approximately 1.3). The story given here says that for every year following the start of the 2007 Subprime Crisis, the profit share of GDP decreased on average 1.3 percent. Looking at the profit ratio tr end in Figure 1, it is visible that the negative effect of the Subprime Crisis took an initially larger impact than a 1.3 percent decline in the profit share, however this impact was short lived. It is plausible that the profit share can jump (at it does with the SEC dummy) and drop (as with the Subprime dummy) since profits are more subject to volatility than wages. Therefore these large coefficients, which are larger than any coefficients in the WR model, appear appropriate, particularly since these coef ficie nts represent yearly averages. Unemployment Rate As predicted, the unemployment variable is highly significant at nearly 0 percent, with a negative coefficient value in the WR model. The obvious correlation between the unemployment rate and wages does warrant an inverse relationship, yet it does not refl ect
73 a powerful effect ( 0.194718). The results offered here are about exactly what I expected; some sort of strong foundational relationship exists for sure, though not the most powerful one. Information that would be welcome however, is if this relationsh ip is consistent within the framework of the economy and history. Inflation The correlation between inflation and the WR model is rendered highly insignificant, with a p value of 40 percent. This result came as a surprise since in macroeconomics there is a clear association between inflation and wages. As inflation rises, real wages and thus, consumer purchasing power falls. The positive power of the coefficient (3.729024) appears as though it is too strong and in the wrong direction. If the probability v alue were significant, the estimation would dictate a 1 percent increase of inflation corresponds to a 3.7 percent increase in the wage share. Corporate Cash Flow share of GDP According to the WR model, Economic activity measured by the inflow and outfl ow of money in financial and non financial corporations is significant and negative. At approximately 0.5, the coefficient of the CASH variable makes a noticeable impact on the wage share. This estimation aligns with my previous prediction. As GLB allow ed different financial and non financial companies to merge, more money from the hands of citizens was to be handled by the new giants. Although CASH
74 does not have the most powerful coefficient in the WR model, 0.5 is not a number to be ignored. Governm ent Transfer Payments According to the significant p values standards set, government transfer payments are not a significant variable. The coefficient, although not highly insignificant, is a high value of 1.1, rounded. Were this coefficient within the b oundaries of significance, it would said that transfer payments have a highly positive effect on the wage ratio, which was the prediction I made before testing the model. The numerical value of the coefficient is higher than I predicted, supporting my be lief that a cushion for workers out of employment is beneficial to the wage share. Perhaps the coefficient, despite it being statistically insignificant, is truly reflecting the actual economic value of transfer payments in relation to the wage share. Unde r current stressful economic times, where food stamp participation has risen from 30 million before the recession to 46 million, a large positive value of government transfer payments on wages is sensical. 73 The Inertia of the Wage Ratio As expected, t he lagged dependent variable is highly significant and carries with it a high coefficient. The relationship told here shows an average 0.7 percent change in 73 National Public Radio April 22, 2012. www.n pr.org/2012/04/22/151166529/poverty in america defining the new poor/?sc=fb&cc=fp
75 the wage share of GDP positively correlates to a 1 percent change in the inertia of the wage ratio. These figures put into context of a historically declining wage share solidly confirm my expectations, as there is a strong correlation between the wage ratio of time t 1 to the wage ratio of time t Unless the economy were in a disastrous state, one woul d hope previous values of the wage ratio positively relate to current values. Diagnostic Tests and Corrections The Breusch Godfrey LM test (LM test) is initially used on the two multivariate equations to detect the inclusion of omitted variables and ser ial correlation. The F statistic (F stat) is an omitted variable test for joint significance of all lagged residuals, while Observations*R2 (Obs*R2) represents the LM test stat. The LM stat is simply the number of observations multiplied by the (uncentered ) R2 from the model regression 74 74 EViews 7 User's Guide United States of America: Quantitative Micro Software, LLC, 2009 ; 161
76 Table 6. LM Test for Serial Correlation, PR Model Breusch Godfrey Serial Correlation LM Test: PROFIT RATIO F statistic 0.011418 Prob. F(1,46) 0.9154 Obs*R squared 0.013896 Prob. Chi Square(1) 0.9062 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/29/12 Time: 03:27 Sample: 1955 2010 Included observations: 56 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t Statistic Prob. C 0.028011 0.831232 0.033698 0.9733 FFR 0.000311 0.032370 0.009621 0.9924 (DEBT_H/GDP)*100 0.000212 0.014921 0.014175 0.9888 PROF_RATIO( 1) 0.004566 0.102664 0.044475 0.9647 D(LOG(OIL)) 0.012112 0.381833 0.031720 0.9748 ((X+M)/GDP)*100 0.001604 0.081746 0.019626 0.9844 ZERODONER 0.003843 0.182647 0.021042 0.9833 SEC 0.007615 0.609591 0.012492 0.9901 SP_CRISIS 0.008563 0.467564 0.018315 0.9855 RESID( 1) 0.020129 0.188376 0.106854 0.9154 R squared 0.000248 Mean dependent var 2.82E 16 Adjusted R squared 0.195355 S.D. dependent var 0.516803 S.E. of regression 0.565032 Akaike info criterion 1.856565 Sum squared resid 14.68602 Schwarz criterion 2.218234 Log likelihood 41.98381 Hannan Quinn criter. 1.996783 F statistic 0.001269 Durbin Watson stat 1.927790 Prob(F statistic) 1.000000 The LM test does not show first order serial correlation or omission of a variable for the profit ratio. The Obs*R2 is 0.013896, the F stat is 0.011418 and the probability value (p value or Prob.) of the F stat (1, 46) is 91.54%. Therefore, the test accept s the null hypothesis of no serial correlation or omitted variables. However, the LM test does ratio equation, where the F stat is 3.546766, the Prob. F (1, 40) is 6.69% and the Obs*R2 is 4.235259.
77 Table 7. LM Test for Serial Correlation, WR Model Breusch Godfrey Serial Correlation LM Test: WAGE RATIO F statistic 0.812442 Prob. F(1,41) 0.3727 Obs*R squared 1.010393 Prob. Chi Square(1) 0.3148 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 04/29/12 Time: 03:28 Sample: 1959 2010 Included observations: 52 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t Statistic Prob. C 0.990689 4.686659 0.211385 0.8336 UNEMP 0.006583 0.051048 0.128951 0.8980 D(LOG(CPI)) 0.137533 4.840656 0.028412 0.9775 ZERODONER 0.010621 0.153889 0.069018 0.9453 SEC 0.029224 0.436326 0.066977 0.9469 ((X+M)/GDP)*100 0.000492 0.112434 0.004378 0.9965 (CASH/GDP)*100 0.000491 0.122573 0.004005 0.9968 (TRAN/GDP)*100 0.257032 1.446184 0.177731 0.8598 (DEBT_H/GDP)*100 0.002860 0.023767 0.120343 0.9048 WAGE_RATIO( 1) 0.020010 0.071870 0.278426 0.7821 RESID( 1) 0.154511 0.171421 0.901356 0.3727 R squared 0.019431 Mean dependent var 2.58E 14 Adjusted R squared 0.219733 S.D. dependent var 0.381587 S.E. of regression 0.421431 Akaike info criterion 1.295083 Sum squared resid 7.281763 Schwarz criterion 1.707846 Log likelihood 22.67216 Hannan Quinn criter. 1.453327 F statistic 0.081244 Durbin Watson stat 1.992236 Prob(F statistic) 0.999904 In the case of significant p values, I looked up the Ljung Box Q statistics for high order serial correlation that included 28 lags, which show the autocorrelation and partial autocorrelation functions of the residuals. The Q based on a number of lags 75 75 EViews 7 User's Guide United States of America: Quantitative Micro Software, LLC, 2009 ; 86 7
78 Table 8. Q statistics for the WR Model AC PAC Q Stat Prob 1 0.126 0.126 0.8707 0.351 2 0.003 0.019 0.8711 0.647 3 0.081 0.080 1.2474 0.742 4 0.146 0.129 2.4931 0.646 5 0.093 0.130 3.0099 0.698 6 0.060 0.099 3.2272 0.780 7 0.340 0.364 10.421 0.166 8 0.031 0.105 10.481 0.233 9 0.010 0.084 10.488 0.312 10 0.269 0.301 15.339 0.120 11 0.110 0.167 16.174 0.135 12 0.043 0.092 16.301 0.178 13 0.188 0.072 18.848 0.128 14 0.102 0.140 19.616 0.143 15 0.028 0.025 19.676 0.185 16 0.013 0.005 19.688 0.235 17 0.223 0.007 23.683 0.128 18 0.206 0.121 27.189 0.076 19 0.017 0.071 27.213 0.100 20 0.184 0.181 30.182 0.067 21 0.066 0.034 30.577 0.081 22 0.042 0.081 30.743 0.101 23 0.075 0.116 31.285 0.116 24 0.075 0.144 31.848 0.131
79 The Q stats increase in size with each lag, and all except for the second through ninth p autocorrelations and partial autocorrelations, further indicate serial correlation. A s impler method used to look for relationships between independent variables was the group statistic correlation, shown below. All relationships showing significant correlation (values above 0.5) are shaded. Clearly, the lagged wage ratio value is correlatin g with several of the other variables and for obvious reason. Viewing the coefficient correlation matrix there appears strong correlation between DEBT*, TRAN*, CASH* and TO*. Table 9. WR Model Correlation Matrix The case of multicollinearity is suspect to independent variable correlation (where one independent variable may be calculated using another variable from the right side) in addition to improper use of dummy variables. In the presence of multicollinearity even in extreme cases, Ordinary Least Sq uares (OLS) is still unbiased and BLUE (Best Linear Unbiased Estimator), though the standard errors will be inflated to some proportion. Albeit evident correlations seen in Table 6, it is more accurate to be skeptical of multicollinearity when there is a high R square and low t statistics, rather than by looking
80 at a correlation matrix. Looking at Tables 1 and 2, the standard error of the model is not especially high, though not very low either. Additionally, the t statistics and R squares in both models do not appear to strongly indicate multicollinearity. Another test to consider when searching for multicollinearity is the Variance Inflation Factor. Yet, in Table 6 above, it seems there may be multicollinearity in regards to the SEC dummy variable, as th ere is high correlation between TO*, CASH*, TRAN* and DEBT*. Despite evidence in Table 6 however, the R squared in the WR model and t statistics do not strike me as showing a high degree of multicollinearity. Regardless, multicollinearity is attended to vi a the Newey West filter. Table 10. Wald Test for Joint Significance in WR Model Wald Test: Equation: A_WAGES Test Statistic Value df Probability F statistic 8.023782 (4, 42) 0.0001 Chi square 32.09513 4 0.0000 Null Hypothesis: C(5)=C(6)=C(7)=C(8)=C(9) Null Hypothesis Summary: Normalized Restriction (= 0) Value Std. Err. C(5) C(9) 0.709279 0.347415 C(6) C(9) 0.079869 0.109306 C(7) C(9) 0.355044 0.151689 C(8) C(9) 1.113097 1.539893 Restrictions are linear in coefficients. Table Y above is a Wald test showing joint significance of the SEC coefficient with the coefficients of TO*, CASH*, TRAN* and DEBT*. The probability value is highly significant which confirms suspicions of multicollinearity. A couple of methods used when d ealing with multicollinearity include increasing sample size (this generally
81 decreases the standard errors) or dropping a variable. Neither of these approaches will be implemented as their impact would be too drastic. Rather, I will leave the offending SEC variable in the model and keep in mind the consequences of multicollinearity (inflated standard errors) 76 Under the Newey West or Heteroskedasticity and Autocorrelation Consistent Covariance (HAC) filter, which automatically corrects for serial correlation and heteroskedasticity together, the F stats, Q stats and their p values remain the same though the standard coefficients are effected. As far as m y knowledge extends on the subject, these results would seem to imply model re specification before the testing of hypotheses. However, I am skeptical of the HAC procedure as none of the statistics changed values. Yet in the residuals in Figures 1 and 2 ar e well behaved with confirming test results. 4.2 Summary and Conclusion This thesis models the wage share and profit share of U.S. GDP us ing two multivariate equations Ordinary Least Squares estimation finds a strong impact on the profit ratio from a once lagged profit ratio (0.57 77 ), growth in the cost of a barrel of oil (0.62), Republican Presidents ( 0.42), the Net Capital Rule change of 2004 (2.68), and the Subprime Mortgage Crisis ( 1.27). The effective Federal Funds Rate is also 76 Willians, R. University of Notre Dame, "Multicollinearity." Accessed April 28, 2012. http://www.nd.edu/~rwilliam/stats2/l11.pdf. 77 All coefficients listed are rounded to the hund redths place.
82 significant though less impactful ( 0.09). Household debt as a share of GDP was not found to be significant, same as trade openness. The wage ratio model is significantly affected by several of the listed independent variables such as the rate of unemployment ( 0.19), Repu blican Presidents ( 0.33), the Net Capital Rule change of 2004 (0.68), corporate cash flow ( 0.40), and the lagged wage ratio representing the dependent variable inertia (0.71). Despite both models being filtered through the Newey West test to correct fo r autocorrelation and heteroskedasticity, there appears to be a high degree of multicollinearity between the SEC dummy and GDP shares of trade openness, corporate cash flow, personal current transfer payments by the government and household debt, as indic ated by the Wald test. Furthermore, the wage ratio model underwent critical change when the Glass Steagall dummy variable was removed due to incorrect binary input. Overall, the PR estimation is more accurate. The coefficients are valued in the expected di rections, and with the exception of the trade openness variable and household debt variable, all independents are significant. Although the statistical significance of these two variables is null, the economic significance is still important. Reasons for a lame p value ascribed to the trade openness variable may be associated with vagueness of the measure and the relationship between imports and household debt. Multicollinearity between trade openness and household debt was tested using the Wald procedure, and showed 31.94 percent (insignificant). Despite the correlation coefficient between the profit ratio and household debt not being weak (0.42, rounded), the p value for debt is insignificant (25.63 percent).
83 The effects of public policy as measured by the Presidential dummy variable can be improved upon by the inclusion of a Congress dummy variable, as the controlling parties in Congress also has a large impact on the economy. Further research may follow the direction of improving the wage ratio model and determining the distribution of wages to the top percent or ten percent of society. A study may also follow which classes of income account for what percentage of financial bor rowing before and after the Financial Crisis of 2007, along with the size of those loans. Inquiry into the distribution of corporate profits is of interest, as is an estimation of which pieces of government legislation are highly impactful on the profit ra tio.
84 References Akhigbe, Aigbe, and Anne Marie Whyte. "The Gramm Leach Bliley Acct of 1999: Risk Implications for the Financial Services Industry." The Journal of Financial Research 27. no. Fall (2004): 435 434. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=498588 Alan B. Krueger, "The Rise and Consequences of Inequality in the United States," Chairman, Council of Economic Advisers January (2012). Al Mamun, Abdullah, Kabir Hassan, and Van Son Lai. "The Impact of the Gramm Leach Bliley Act on the Financial Services Industry." Journal of Economics and Finance 28. no. Fall (2004). http://www.springerlink.com/content/n0827468451287pn/fulltext.pdf Alon, Titan, and Eric Swanson. "Operation Twist and the Effect of Large Scale Asset Purchases." FRBSF (blog), April 25, 2011. http://www.frbsf.org/publications/economics/letter/2011/el2011 13.html Alvaredo, Facundo. "Inequality over the Past Century." Finance & Development no. September (20 11). Associated Press, First. "Wisconsin Governor Scott Walker Readies National Guard Against Unions." The Huffington Post February 11, 2011. http://www.huffingtonpost.com/2011/02/11/scott walker unions wisconsin national guard_n_822225.html Associated Press, First. "Wisconsin Governor Scott Walker Readies National Guard Against Unions." The Huffington Post February 11, 2011. http:// www.huffingtonpost.com/2011/02/11/scott walker unions wisconsin national guard_n_822225.html Bar, Michael. San Francisco State University, "Growth Rates." http://bss.sfsu.edu/mbar/ECON560/Growth Rates.pdf Bartels, Larry. Unequal Democracy: The Political Economy of the New Gilded Age New York: Princeton University Press, 2008. Berg, Andrew G., and Jonathan D. Ostry. "Inequality and Unsustainable Growth: Two Sides of the Same Coin?." IMF Staff Discussion Note (2011) http://ww w.imf.org/external/pubs/ft/sdn/2011/sdn1108.pdf Berg, Andrew G., and Jonathan D. Ostry. "Warning! Inequality May Be Hazardous to Your Growth." IMFdirect (blog), April 8, 2011. http://blog imfdirect.imf.org/2011/04/08/inequality and growth/
85 Branko Milanovic, The Haves and Have Nots (New York: Basic Books, 2011). Bureau of Economic Analysis, s.v. "A Guide to the National Income and Product Accounts of the United States." http://www.bea.gov/national/pdf/nipaguid.pdf Chang, Ha Joon, and Erik Reinert. How Rich Countries Got Rich, and Why Poor Countries Stay Poor New York: Carroll & Graf Publishers, 2007. Cowen, Tyler. "What Export Oriented America Means." The American Interest May/June 2012. http://www.the american interest.com/article.cfm?piece=1227 da Costa, Polyana. Bankrate.com, "Fina ncial Crisis Timeline: Collapse and Bailout." Last modified September 21, 2011. DiSalvo, David. "How Governor Rick Scott is Sabotaging Florida's Universities." Forbes April 23, 2012. http://www.forbes.com/sites/daviddisalvo/2012/04/23/how governor rick scott is sabotaging floridas universities/. Drawbaugh Kevin. Stanford Law School Securities Class Action Clearinghouse, "US SE C Clears New Net Capital Rules For Brokerages." Last modified April 28, 2004. http://securities.stanford.edu/news archive/2004/20040428_Headline0 8_Drawbaugh.htm Dnhaupt, Petra. "Financialization and the rentier income share evidence from the USA and Germany." IMK Working Paper (2010). Educational Resources." Federal Reserve Bank of San Francisco March 2010. http://www.frbsf.org/education/activities/drecon/2010/0310.html Evans, Donald L., Kathleen B. Cooper, and Steven J. Landefeld. Buraeu of Economic Analysis, "Corporate Profit s: Profits Before Tax, Profits Tax Liability and Dividends." Last modified September 2002. http://www.bea.gov/scb/pdf/national/nipa/methpap/methpap2.pdf Evans Pritchard, Ambrose. "US money supply plunges at 1930s Pace as Obama eyes fresh stimulus." The Telegraph May 26, 2010. http://www.telegraph.co.uk/finance/e conomics/7769126/US money supply plunges at 1930s pace as Obama eyes fresh stimulus.html Federal Reserve "Section 109 of the Riegle Neal Interstate Banking and Branching Efficiency Act." Accessed April 8, 2012. http://www.federalreserve.gov/boarddocs/supmanual/cch/sec109.pdf France Presse, Agence. Republicans attack unions to hurt Democrats: analysts, "The Raw Story." Las t modified February 19, 2011.
86 http://www.rawstory.com/rs/2011/02/19/republicans attack unions to hurt democrats analysts/. Haberman, Gary. "Capital Requirements of Commercial and Investment Banks: Contrasts in R egulation." FRBNY Quarter ly Review no. Fall (1987). http://www.newyorkfed.org/research/quarterly_review/1987v12/v12n3article1.pdf Hirsh, Stacey. "Reagan presidency pivotal for unions." The Baltimore Sun June 8, 2004. http://www.baltimoresun.com/bal bz.unions08jun08,0,1073570.story Finance & Development September 2011. http://www.imf.org/external/pubs/ft/fandd/2011/09/jahan.htm Jeffrey M. Woolridge, Introductory Econometrics: A Modern Approach, (United States of America: South Western College Pub, 2006) http://fmwww.bc.edu/ec c/F2007/228/EC228.f2005.nn12.pdf Keena n, Rebecca. "Distribution's out of the dog house." The Catalyst March 2, 2012. ncfcatalyst.org?s=distribution's out of the dog house&x=0&y=0 Khan, Salman. "Breakdown of Gas Prices." Macroeconomics Khan Academy. 2012. Web, http://www.khanacademy.org/finance economics/macroeconomics/v/breakdown of gas prices Kumhof, Michael, and Romain Rancire. "Leveraging Inequality." Finance & Development December 2010. http://www.imf.org/external/pubs/ft/fandd/2010/12/Kumhof.htm Kumhof, Michael, and Romain Rancire. "Unequal=Indebted." Finance & Development September 2011. http://www.imf.org/external/pubs/ft/fandd/2011/09/kumhof.htm Kozlowski, Paul J. "Financial analysis after the Riegle Neal Interstate Banking and Branching Efficiency Act." Journal of Regional Analysis and Policy (199 0) http://jrap journal.org/pastvolumes/1990/v29/29 1 5.pdf Mamun, Abdullah, M. Kabir Hassan, and Son Van Lai. "The impact of the Gramm Leach Bliley act on the financial services industry." Journal of Economics a nd Finance 28. no. 3 (2004). http://www.springerlink.com/content/n0827468451287pn/ M.C. Baddeley, and D.V. Barrowclough, Running Regressions, (New York: Cambridge University Press, 2009). Milanovic, Branko. "More or Less." Finance & Development September 2011.
87 Mother Joines, "How Banks Got Too Big to Fail." Last modified February 2010. http://www.motherjones.com/politics/2010/01/bank merger history National Public Radio www.npr.org/2012/04/22/151166529/poverty in america defining the new poor/?sc=fb&cc=fp Pearson, "Economic Policymaking." http://wps.ablongman.com/long_edwards_ga_12/33/8517/2180597.cw/index.htm Ramu Ramanathan, Introductory Econometrics with Applications, 5th Edition (United States o f America: Michael P. Roche, 2002). Saez, Emmanuel. University of California Berkeley, Department of Economics, "Striking it Richer: The Evolution of Top Incomes in the United States." Last modified March 2, 2012. http://elsa.berkeley.edu/~saez/saez UStopincomes 2010.pdf Scheiber, Noam. "EXCLUSIVE: The Memo that Larry Summers Didn't Want Obama to See." The New Republic February 22, 2012. http://www.tnr.com/article/politics/100961/memo Larry Summers Obama Simon Johnson and James Kwak, 13 Bankers (New York: Pantheon Books, 2010). Skott, P., and S. Ryoo. "Macroeconomic implicat ions of financialization." Cambridge Journal of Economics 32. no. 6 (2008). Terkel, Amanda. "Scott Walker Quietly Repeals Wisconsin Equal Pay Law." The Huffington Post April 6, 2012. http://www.huffingtonpost.com/2012/04/06/scott walker wisconsin equal pay law_n_1407329.html Walker, Douglas O. "Patterns of income distribution among world regions." Journal of Policy Modeling 29. no. 4 (2007): 643 655. http://www.sciencedirect.com/science/article/pii/S016189380700049X Wannisky, Kathleen E., and Jill M. Peterson. Government Accountability Office, "Alternative Net Capital Requirements for Broker Dealers that are Part of Consolidation Supervised Entities." Last modified June 25, 3004. http://www.gao.gov/decisions/majrule/ d04896r.pdf Willians, R. University of Notre Dame, "Multicollinearity." http://www.nd.edu/~rwilliam/stats2/l11.pdf Yglesias, Matthew. "Romer on Boosting the Economy." Think Progress March 25, 2011. http://thinkprogress.org/yglesias/2011/03/25/200338/romer on boosting the economy/?mobile=nc