San José State University
Department of Economics

applet-magic.com
Thayer Watkins
Silicon Valley
& Tornado Alley
USA

A Statistical Review of
Current Economic Conditions
in the U.S. (February 2010)

Latest Statistics
from the BEA
Employment
Recession
NBER and
Recession
Gross Domestic
Product (GDP)
Private
Investment
Purchases
International
Trade
Government
Sector
Consumer
Purchases
Tax Rebates
of 2008
Nondurable
Goods
Purchases
Durable
Goods
Purchases
Consumer
Purchases
of Services
    Conclusions    

Latest Statistics

On January 29th the Bureau of Economic Analysis (BEA) published its first (advance) estimate of the real GDP for the fourth quarter of 2009. According to that estimate the real GDP of the U.S. increased 1.4 percent over its level in the third quarter of 2009. The annualized rate would therefore be 5.6 percent. If the 1.4 percent quarterly continued for four quarters the increase would be the 5.7 percent growth the BEA publicized. This 5.7 percent growth rate is quite literally an incredible rate of growth. The 1.4 percent percent rate quarterly increase is in contrast to 0.55 of 1 percent in the third quarter and the decline of 0.185 percent in real GDP (0.75 of 1 percent annual rate) for the second quarter of 2009 compared to the first quarter. While the increase in real GDP in the third and fourth quarter, if it were true, would be really good news for the economy, but there are some werious suspicions that the BEA may have succumbed to Administrative and/or Fed arm-twising to come up with a statistical end of the recession. Those suspicions are dealt with elsewhere. Here the statistics will be treated at their face value even though they are not credible. For the third quarter of 2009 the first (advance) estimate was an annual growth rate of 3.5 percent; the second estimate reduced the growth rate to 2.8 percent and the third (final) estimate reduced the growth rate to 2.2 percent. The graph below of quarterly growth rates over the past five years puts the 5.7 percent growth rate into perspective.

There are continuing weak spots in the economy. Businesses continued to sell off inventory without replacing it. In the third quarter the rate was $139.2 billion ($2005). As terrible as this is, it was an improvement from second quarter rate of 160.2 billion ($2005) per year. The investment in plant and equipment by businesses continued to decline. This quantity reflects business confidence about the future of the economy. In the fourth quarter businesses continued to sell off inventory without replacing it but the rate dropped to 33.5 billion ($2005). Most of the growth in real GDP for the fourth quarter came from the 105.7 billion ($2005) reduction in the disinvestment in inventory. The increase in real GDP between 2009III and 2009IV was 182 billion ($2005).

At this point the most instructive exercise is to examine where the 182 billion dollar increase in real GDP came from.

The Quarter-to-Quarter Changes in the Components of Real GDP in 2009
(Billions of $2005)
Component1st Q
to 2nd
2nd Q
to 3rd
3rd Q
to 4th
GDP-23.971.5182.0
Consumption-20.263.645.9
Durables-15.551.0-2.4
Nondurables-9.87.621.3
Services2.811.825.8
Investment-101.817.7127.4
Residential
Struct.
-23.515.25.0
Nonresidential
Struct.
-32.8-19.49.1
Equipment
& Software
-11.03.327.9
Inventory-46.321.0105.7
Net Exports56.1-27.016.3
Exports-15.059.362.8
Imports-71.286.446.5
Government41.416.9-1.1
Federal27.219.80.2
State & Local14.712.2-1.2
Residual0.4-8.5-1.8

The total change for a component cannot be derived by adding up the changes in the subcomponents because of the methodology used by the BEA. The real value for a component is separately computed using the current value figure and a price index. The BEA explicitly warns that the totals of the subcomponents will not necessary equal the value for the component. For example, the changes in the subcomponents of GDP add up to 188.5 billion rather than 182.0 billion. The difference could be called the residual change, but it is not equal to the change in the residual. Likewise the subcomponents of private domestic investment add up to 147.7 billion, whereas the change in private domestic investment is 127.4 billion.

Despite the fact that the changes in the subcomponents cannot be exactly related to the change in the component one still sees that a major part of the change in real GDP came from a decrease in the disinvestment in inventory. Thus supposedly the economy is booming because businesses are not selling off inventory without replacing it to the extent they were doing so. Net exports increased because exports increased more than the amount that imports increased. Thus supposedly foreign buyers are pulling the U.S. out of its recession. Those foreign buyers themselves are in recessions resulting from the decreases in imports by the U.S. Isn't that amazing? Also consumers purchases increased but not in durables goods.

Even if the output recession has ended the unemployment rate will continue at its current high rate and may well increase.

A quick review of the current state of economy is in order at this point. As was indicated above the perception of the decline in the economy is distorted by the reporting of the rates at their annual rate rather than the quarter-to-quarter change. To emphasize that the reporting of the annual rates exaggerate the decline, note that real GDP was reported by the BEA to have declined 6.4 percent in the first quarter of 2009. However the real GDP ($2000) in the first quarter of 2009 was just 2.5 percent below what it was in the first quarter of 2008, and only 0.04 of 1 percent below what it was in the first quarter of 2007. The BEA has now switched to 2005 prices and this makes the real GDP in the first quarter of 2009 3.3 percent below the level in the first quarter of 2008 and 1.3 percent below the figure for the first quarter of 2007. The volatility of these figures as a result of a change in the base year prices indicates the margin of error in any such macroeconomic statistics.

The U.S. economy was officially in an output recession since the real GDP -- the output of goods and services produced by labor and property located in the United States measured in the year 2000 prices -- decreased significantly for two quarters in a row. Those two quarters were the fourth quarter of 2008 and the first quarter of 2009. There was a small decrease for the third quarter of 2008 over the second quarter of 2008, but for 2000 prices it was 1/8 of 1 percent and too small, given the accuracy of the data, to say whether there really was a decrease.

The BEA is now reporting the real GDP and its components in the year 2005 prices. In 2005 prices the statistical picture is somewhat different. According to the new figures on real GDP in 2005 prices the GDP peaked in the fourth quarter of 2007 then dropped in the first quarter of 2008 but rose in the second quarter. The drop from the second quarter to the third quarter was 0.0676 of 1 percent, a small but statistically significant change. The GDP in 2005 prices then decreased 1.37 percent (5.5 percent annual rate) in the fourth quarter of 2008 and 1.65 percent (6.6 percent annual rate) for the first quarter of 2009.

It may seem strange at only this late date for the official statistics to verify the prevalent notion of the U.S. economy being in a recession. However the cry of recession came long before there was justification and may have been a major contributor to an output recession actually occurring.

Remarkably, according to the estimates by the BEA the real GDP in 2000 prices for the United States in 2008 increased by 1.1 percent over the level in 2007. Yes, this was a 1.1 percent increase. In 2005 prices the increase was 0.44 of 1 percent. However that increase reflected two quarters of growth, one of near constancy and one quarter of decline.

The media in their desperate search for sensationalism usually report the annual rates of decrease as though they are quarter-to-quarter decreases. This can be attributed to simple ignorance on the part of the news people but the Bureau of Economic Analysis, through carelessness, abets the media misrepresentation. The Bureau of Economic Analysis states the change figures correctly in its first statements of its news releases but often later in the same release incorrectly state annual rates as though they are quarter-to-quarter changes. For example, consider a recent news release of the BEA on real GDP:

Real gross domestic product decreased at an annual rate of 6.3 percent in the fourth quarter of 2008, (that is, from the third quarter to the fourth quarter), according to final estimates released by the Bureau of Economic Analysis. In the third quarter, real GDP decreased 0.5 percent. (Emphasis added.)

The real GDP did not decrease 0.5 percent from the second quarter to the third quarter. It decreased 1/8 of one percent and if that rate continued for four quarters the decrease would be 0.5 of one percent.

As stated previously the U.S. economy was officially in an output recession from the fourth quarter of 2008. The previous talk in the media of a recession was really about an employment recession. It is important to distinguish between the two concepts because the cause and the appropriate policy actions for the two types of recession may be different. For more on the employment recession see EmploymentRecession.

Note that the condition of the economy is a different matter than the condition of the financial markets. Elements of the financial world are often engaged in risky practices that fail. This is akin to the thousands of gamblers who lose money in the casinos. Gamblers losses should not be mistaken for the state of the economy. The financial markets in contrast to the casino world serve an economic purpose and deserve a special attention but their condition is not the same as the condition of the overall economy which is dealt with in the following.

As stated above the decline in real GDP measured in year 2000 prices from the second quarter of 2008 to the third quarter, based upon the Final estimate of the real GDP for the third quarter of 2008 by the Bureau of Economic Analysis of the U.S. Department of Commerce was about 1/8 of 1 percent. This was so small that it was well within the margin of accuracy of the GDP statistics. (If this rate of decrease continued for four quarters it would result in a decrease in real GDP of 0.5 of 1 percent.) There are politicians that that are construing the problem in financial institutions as a crisis in the economy of unprecedented proportions. This is bogus and it is being used to justified the transformation of the U.S. economy into a social welfare state. There are two industries that are suffering long term declines, residential housing construction and motor vehicle sales.

In mid-2008 those two industries, motor vehicles and residential construction, wee not the tip of the iceberg, they were the iceberg. Their problems and the problems of the financial institutions should not have been construed as a general problem of the economic system of the U.S. Ironically the subprime mortgage crisis is the result of operating Fannie Mae and Freddie Mac as social welfare agencies rather than business institutions. Having the entire economy operate as a social welfare state economy is not the solution to the problems brought about by operating key financial institutions as social welfare agencies. The nature and origin of the subprime mortgage crisis is dealt with elsewhere.

The Declaration of a Recession by
the National Bureau of Economic Research (NBER)

At the beginning of December 2008 the National Bureau of Economic Research (NBER) announced that the U.S. economy is in a recession that started a year before. The media promptly announced that it was official the economy is in recession. What the media neglected to publicize is that the NBER's definition of recession is different from the standard definition. The term recession is officially defined as a period in which the real GDP has declined for two quarters in a row. The real GDP (2000 value dollars) declined between the second quarter and the third quarter by $15.0 billion which represents a percentage decrease of about 1/8 of 1 percent based upon the Final estimate of the Bureau of Economic Analysis (BEA). This is well within the margin of error of the statistics so the decline was not statistically different from zero.

Statistics for the fourth quarter of 2008 are now available in the form of the FINAL estimates published by the Bureau of Economic Analysis of the Department of Commerce. They show the Gross Domestic Product of the U.S. in the year 2000 prices declined from $11,712.4 billion in the third quarter to $11,522.1 billion in the fourth quarter, a decrease of $190.3 billion. The percentage decrease is 1.625 percent. The Advance estimates for the real GDP for the first quarter of 2009 only became available at 8:30 AM EST on April 29th. The BEA Advance estimate shows the real GDP decreasing from $11,522.1 billion for 2008IV to $11,340.9 billion for 2009I, a decrease of 1.573 percent.

The NBER looks at a variety of statistics and what NBER identifies is not an output recession per se but some notion of an economic malaise. This is a legitimate endeavor but they should not use the term recession which has the connotations of the real GDP receding. For more on a review of the NBER's pronouncement see Recession 2008? Timing and Terminology.

The NBER assessment of a recession having started in the fourth quarter of 2007 is based upon its economists' perception that a broad-based decline in economic activity has occurred. There had been a decline in employment starting in that period but the output statistics except for residential housing construction and motor vehicle manufacturing were not showing such a decline. At least that is, until after the financial markets crisis in September of 2008. Since recessions are closely connected with business expectations of a need for increased capacity in the future any public pronouncements concerning the current and future state of the economy are likely to feed back into the future state of the economy. Media pronouncements that the economy is in recession and that it is going to be a very long recession are quite likely to discourage business investment and bring about the condition being forecast. The review of the output statistics do not support the notion that the U.S. economic output was receding in the period before the September financial crisis.

A Review of the Statistics Compiled
by the Bureau of Economic Analysis

A macroeconomic review of the economy should start with real Gross Domestic Product. This is the output of final goods and services produced with the U.S. valued in the year 2005 prices. The quarterly figures are seasonally adjusted.

The picture was remarkably smooth until the downturn in 2008IV. The variations in the growth rate from quarter to quarter are barely perceptible until the decline for 2008IV. The variations are there as shown in the graph for quarter-to-quarter growth rates which will be shown below.

The Final estimate of the real GDP for the third quarter of 2008 by the Bureau of Economic Analysis is $11,712.4 billion in 2000 prices. This is essentially the same as the value for the second quarter, $11,727.4 billion. The media ballhooed this as evidence of a recession but the data are not sufficiently accurate to say the GDP really declined. Numerically the decrease from the second quarter to the third was $15.0 billion or 0.128 of 1 percent (or roughly 1/8 of 1 percent) and if this rate of decline continued for four quarters the real GDP would have declined 0.512 of 1 percent (which the media reported as a decrease in real GDP of 0.5 of 1 percent).

The degree of accuracy of the GDP estimate can be judged by the residual figure in the national income accounts. If everything were exact the residual would be zero. Instead the residual is usually more than ten billions of dollars. For example, in the fourth quarter of 2007 it was about a negative $62 billion. If the residual were constant from quarter to quarter it would reflect some systematic error, but the residual fluctuates from quarter to quarter. In 2008II it was a negative $44 billion and in 2008III it was a negative $16 billion. A $28 billion fluctuation in the residual over two quarters makes the $15.0 billion decline in real GDP in the third quarter of 2008 small and statistically insignificant.

In contrast the decrease in real GDP for the third quarter of 2008 to the fourth was $190.3 billion, a decrease of 1.6 percent. However it is notable that the residual for the national income accounts shifted from a negative $16.4 billion in the third quarter to a positive $19.3 billion, a net change of $35.7 billion.

There was a pattern of alternating high and low growth rates for several years in the early 2000's.

These fluctuations prior to 2008IV might have been due to an artifact of the statistical reporting or tabulating system rather than something that is actually happening in the economy.

A scan of the graph of real GDP for the U.S. since 2000I shows there were a number of times where the growth of the economy slowed and then went on to higher growth.

However, as it happens, there is an identifiable source for the slowing down of the economy which is revealed below.

If any macroeconomic problem is developing it should show up among the various components of aggregate demand.

Barely noticable until 2008III is a downward trend in gross domestic investment purchases, shown in green in the above graph. Almost always if there is a macroeconomic problem it is manifested in a decline in investment purchases. Therefore the levels of investment need a closer look.

Private Domestic Investment

There the problem is clearly the decline in residential construction investment. This is not just a recent development. Residential construction investment has been declining since the last quarter of 2005. There is also a problem with the decline in inventory investment culminating with a net sell-off of inventory in the last quarter of 2007. Net inventory investment goes through some radical fluctuations and the upward fluctuations offset in the recent past the decline in residential construction. Now the down fluctuations in inventory investment are coinciding with the declines in residential construction.

In previous recessions the source of the problem was a decline in the purchase of equipment and other investment goods. In these current circumstances the level of business investment in plant and equipment, shown in the graph as Nonresidential, was not declining up until 2008III. It declined only after a barrage of pronouncements that the economy was in a recession and that it was going to to a long one and hence any company that continued to increase its productive capacity would be foolish. Thus in this situation the crisis in the financial markets due to the effective bankruptcy of Fannie Mae and Freddie Mac in September of 2008 was used to induce fear on the part of businesses that they should not increase their productive capacities. The decline in the production and purchases of automobiles and residential housing was used to support that fear.

It is notable that the price level declined in the fourth quarter of 2008 by about 1.0 percent or 4.1 percent on an annual basis. This raised the real rate of interest by about 4.5 percent. This means that despite the nominal rate of interest being low the real rate increased significantly. This could account for some of the decline in investment in plant and equipment, but the price deflation was probably an effect of the output recession rather than a cause of it.

The components of nonresidential private investment show the same pattern as the total. Here is the graph of the figures along with a graph of inventory change with an expanded scale.

Whereas between 2008II and 2009II real GDP ($2005) declined by 3.83 percent, investment in plant and equipment declined by 19.7 percent. The selling off of inventory and not replacing it (negative inventory investment) increased by $123.1 billions ($2005) over this period. This $123.1 billion decline in inventory investment along with the $569.8 billion ($2005) decrease in investment in plant and equipment accounts for 135 percent of the $513.8 billion ($2005) decline in GDP over that period.

The investment in residential structures has been steadily declining since 2006I. The problems of this sector are separate from the current recession. However this continued decline contributes to the decline in demand for U.S. output.

The decrease in investment in residential housing over the period was $118.5 billion. This amount along with the declines in investment in plant, equipment and inventory account for 158 percent of the decrease in real GDP. There were other declines, such as $162 billion in consumer purchases, and some positive influences, such as an increase in net exports of $145.6 billion. (An increase from −476.0 billion to −330.4 billion.) Government purchases increased by $61.7 ($2005). The point is that the real economic problem is in private investment.

The one-shot gimmicks to increase demand will be accepted gladly by business but they will not encourage businesses to investment in increased capacity. Businesses realize that the temporary stimuli will not be there in the future. They may only help them sell off existing inventory. The only measures that have a chance of encouraging businesses to replace inventory and increase productive capacity are the ones which are permanent.

In the past there were decreases in investment in plant and equipment which led to recessions and the recessions were then declared. In 2008 a recession was declared and as a result of the financial chaos of September 2008 businesses accepted this declaration and began to reduce their investment in plant and equipment and sell off inventory without replacing it. Now it is necessary to unconvince businesses of there being a recession which will continue for the foreseeable future. As pointed out above the reduced value of investment in plant, equipment and structures along with the disinvestment in inventory is enough to keep the economy in serious recession.

The International Sector

There is a persistent worry among politicians and the general public about international trade. In particular the general public interprets the decrease in the value of the U.S. dollar with respect to other currencies as evidence of a deterioration in the U.S. economy. On the contrary the decreased value of the dollar has reduced the U.S. balance of trade deficit. Countries often devalue their currencies to stimulate their economies. China has maintained about a 400 percent undervaluation of its currency to ensure that no one in the world will be able to underprice its products.

The recent past the level of U.S. imports has exceeded its exports by about a half trillion dollars per year but since the third quarter of 2006 this deficit has been declining. The figure for 2009II indicates a decrease in the trade deficit of 55 percent compared its value in 2006I.

Government Sector Purchases

The levels of purchases by the Federal and the State & Local governments after having continued what seemed to be an inexorable rise have declined slightly in 2009I. This was largely in purchases by state and locall governments, probably due to a decline in their tax revenues. This is the next stage in a recession. Federal defense expenditures also declined slightly, due to factors unrelated to the recession.

Consumer Purchases

Consumer purchases also tended to have a steady, regular rise, until after the financial crisis of September of 2008. However, as shown below, consumer purchases of durable goods, nondurable goods and services went up slightly in 2009I, down in 2009II and up slightly in 2009III.

It is very significant that the decline in demand for U.S. production is now primarily in private investment.

Consumer Nondurable Purchases

It must be noted that these statistics, although stated in dollar amounts, reflect the volume of consumer purchases rather than the actual expenditures on the items. There is a general concern about particular items such as gasoline. The following is a graph of the nondurable purchases shows, among other things, that purchases of energy products such as gasoline and fuel oil have been declining in recent quarters, due to higher prices, but in 2008IV and 2009I went up slightly, due to lower prices.

Generally consumer nondurable purchases, in real terms, went up in 2009I, back down in 2009II and then up in 2009III.

Consumer Durable Purchases

Consumer durable purchases generally have been increasing but there was a significant decline in the fourth quarter of 2005 and a downward trend in 2007 which continued through the third quarter of 2008. This downward trend is primarily due to the decline in motor vehicle purchases but there has also been a slight decline the purchase of household equipment such as television sets, furniture and electronic devices but this decline was only from 2008II to 2008III, after the impact of the 2008 tax rebate was felt in the second quarter. The impact of the tax rebate will be considered in more detail later. However there was an upturn in durable goods purchases in 2009I, even in motor vehicle purchases.

This points up the fact that the current recession is now based up on the decline in businesses investment purchases which was brought about by the public promotion of pessimism about the future of the economy.

Consumer Purchases of Services

The purchases of services now involves a greater dollar amount than the purchases of durable and nondurable goods combined. The trends in the levels of some services are of special interest. Note once again that the trends are in the quantities rather than the levels of expenditures.

The nature of housing services requires some explanation. Housing services includes the expenditures for rent but also an imputed value for the owner occupied housing.

The amount of medical services consumed continues to increase at an extraordinary rate. The graph displays the trend in the quantity of medical services. The trend in the cost of medical services is not shown in the above graph. It is notable that Americans spend about one fifth more on medical services than on food and about one tenth more than on housing. Clearly something is amiss in the matter of medical care in America and the public subsidy of medical care is not solving the problem but instead making it worse. Around 1900 America allowed the medical profession to create a cartel arrangement in which the production of doctors was artifically restricted. Medical schools were bullied with a threat of the loss of their accreditation into cutting their admissions to a fraction of what they had been. The reduced supply of doctors resulted in substantially increased income for doctors. However even though the medical profession created the opportunity for economic rents those rents did not entirely stay with the doctors. Medical schools increased the prices of their services so that doctors typically begin their practice with enormous debts from their medical education.

By cutting back the supply of doctors the cartel arrangement was able to create monopoly prices. The public subsidy of medical care just meant that the monopoly prices went even higher. The inadequate supply of doctors resulted in doctors being overworked and hence having little time to mull over the conditions of their patients. For more on this topic see Medical Cartel. For material on the high cost of pharmaceuticals see Pharmaceuticals.

The Impact of the 2008 Tax Rebate

At the second quarter of 2008 economic conditions were definitely not in the nature of a recession in output. In the spring of 2008 there was the possibility of an incipient economic malaise but by the second quarter that worry has disappeared. The tax rebates which were sent out in 2008 could have been saved, used to pay down debt, spent on foreign goods and services or spent on domestic goods and services. It was only this last category that was of any benefit for stimulating the economy. Only a fraction of the total rebate will go for domestic goods and services. This was probably from the rebates that went to the relatively young consumers. The effectiveness of the tax rebates could be estimated from the figures for the third quarter of 2008. A rough estimate of the impact of the tax rebate is that there was an increase in effective demand of about $37 billion in year 2000 value dollars for rebates of about $74 billion 2000 value dollars. Once the tax rebate dollars were spent that was the end of it and the decline in real GDP of $15.0 billion for the third quarter could easily be accounted for by the fading away of the effect of the tax rebate. For more on the topic of the impact of the 2008 tax rebate see Tax Rebate 2008.

However, although there are theoretical reasons to question the effectiveness of tax cuts as an economic stimulus measure there is the empirical evidence of the case in 1975 when a tax cut brought an end to the 1974-1975 recession. For a similar case of an economic malaise that turned into a recession see the 1980-1983 recession.

Conclusions

Up until the fourth quarter of 2008 the only major component of consumer demand that was faltering was the purchases of motor vehicles. However there were components of investment demand which are closely connected with consumer demand which were faltering. These are residential housing construction and net investment in inventories. Up until the financial crisis prompted by the bankruptcy of Fannie Mae and Freddie Mac in September 2008 there is no evidence of an output recession. Even the slight and not statistically significant decline in real GDP from the second quarter of 2008 to the third quarter is easily accounted for by the fading out of the effect of the 2008II tax rebate. There was an employment recession but that subject to different causes than an output recession. See Employment Recession.

The output recession began with the effective bankruptcy of Fannie Mae and Freddie Mac in September of 2008. That led to the collapse of some financial enterprises and the bailout of others. The collapse did not have to go beyond those firms. However because the declared bankruptcies of Fannie Mae and Freddie Mac by their managements were unexpected by the markets there ensued a widespread panic among stock market investors which led to sharp declines in stock prices. This in turn brought further panic and a loss of consumer and business confidence in the future of the U.S. economy. Business investment in increased capacity is highly volatile. If businesses perceive no need for increased capacity then their investment purchases can go to zero. When consumers become pessimistic about the economic future they may reduce their purchases a few percent; business investors faced with the same circumstances eliminate all of their purchases.

In 2008IV business investment in plant and equipment to increase productive capacity dropped 5.9 percent compared to 2008III, this is an annual rate of decrease of nearly 24 percent. In 2009I this investment dropped 10.9 percent compared to the level in 2008IV, an annual rate of 43.5percent decrease. This could be characterized as the level of investment in plant and equipment being in freefall.

In contrast, after the financial crisis of September 2008, consumer purchases decreased only 1.1 percent between 2008III and 2008IV and went up 0.5 of 1 percent in 2009I. Clearly the problem is business confidence. Politicians have focused on saving financial institutions so that they can lend funds to those who want to make investments in increased capacity. But who would want to borrow to increase capacity in a recession. The media and politicians with their proclamations of RECESSION!! in early and mid-2008 have destroyed business confidence and brought about a real recession. It remains to be seen whether business confidence can be revived. Business confidence is like a balloon and it is pretty difficult to unpuncture a balloon.

(To be continued.)


HOME PAGE OF applet-magic
HOME PAGE OF Thayer Watkins