From Record Monthly Surplus to the Largest Annual Deficit in 6 Years¶
On May 25, 2018, Judy Woodruff interviewed Marc Short, White House Director Of Legislative Affairs, on the PBS Newshour. The video and transcript can be found at this link. At 2:43 in the video, Marc Short states the following:
And, lastly, in April, we had the largest surplus on record in American history. That is a result of the growth and the revenues coming into the federal government because of the tax plan.
A Washington Times article similarly reported the record surplus. It starts:
The federal government took in a record tax haul in April en route to its biggest-ever monthly budget surplus, the Congressional Budget Office said, as a surging economy left Americans with more money in their paychecks — and this more to pay to Uncle Sam.
All told the government collected \$515 billion and spent \$297 billion, for a total monthly surplus of \$218 billion. That swamped the previous monthly record of \$190 billion, set in 2001.
However, the article went on to give more perspective than Marc Short, stating:
April is always a strong month for government finances, with taxpayers filing their returns for the previous year and settling up what they owe, even as expenditures often dip for the month.
But this year was particularly strong, with receipts jumping 13 percent compared to a year ago.
Following is Python code which reads the Monthly Treasury Statements for the current issue and previous issues and plots the monthly receipts, oulays, and surplus/deficit starting in October of 2014.
# Monthly Treasury Statement (MTS) - Current Issue:
# https://www.fiscal.treasury.gov/fsreports/rpt/mthTreasStmt/current.htm
# Monthly Treasury Statement (MTS) - Previous Issues:
# https://www.fiscal.treasury.gov/fsreports/rpt/mthTreasStmt/backissues.htm
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
import datetime
pd.set_option('display.width', 120)
#pd.set_option('max_rows', 200)
# START OF VARIABLES TO UPDATE
month_index_last = 12
yrmo_last = '1809'
moyr_last = '0918'
# (Note: manually update Jupyter URL and last month in final note below)
# END OF VARIABLES TO UPDATE
savefig_rec = "mts" + yrmo_last + "rec12m.png"
savefig_recoth = "mts" + yrmo_last + "recoth12m.png"
savefig_def = "mts" + yrmo_last + "def12m.png"
xlabel_last = "Source: Monthly Treasury Statements (see http://econdataus.com/mts" + yrmo_last + ".html"
xlsx_files_last = "https://www.fiscal.treasury.gov/fsreports/rpt/mthTreasStmt/mts" + moyr_last + ".xls"
xlsx_files_yr = [
'https://www.fiscal.treasury.gov/fsreports/rpt/mthTreasStmt/mts0915.xls',
'https://www.fiscal.treasury.gov/fsreports/rpt/mthTreasStmt/mts0916.xls',
'https://www.fiscal.treasury.gov/fsreports/rpt/mthTreasStmt/mts0917.xls',
xlsx_files_last
]
def joinyear(year):
iyr = year - 2015
xx = pd.read_excel(xlsx_files_yr[iyr], sheet_name='Table 7', index_col=0, skiprows=4)
#if year < 2018: # fix required for mts0818.xls
xx = xx.iloc[[0,1,3,4,5,6,7,8,9,10,99,102], 0:12]
#else:
# xx = xx.iloc[[0,1,3,4,5,6,7,8,9,10,98,101], 0:12]
years = [year-1,year-1,year-1,year,year,year,year,year,year,year,year,year]
months = ['10','11','12','01','02','03','04','05','06','07','08','09']
#months = [10,11,12,1,2,3,4,5,6,7,8,9]
for i in range(0,12):
years[i] = str(years[i])+"-"+months[i]+"-01"
xx.columns = pd.to_datetime(years)
xx.index = ['Individual','Corporation','Employment','Unemployment','Other Retirement',
'Excise','Estate','Customs','Miscellaneous','Total Receipts','Total Outlays','Surplus/Deficit']
if year >= 2018:
xx = xx.iloc[:, 0:month_index_last]
#print(xx)
return(xx)
def joinyears(start_year, end_year):
yy = joinyear(2015)
for year in range(start_year+1, end_year+1):
yy = yy.join(joinyear(year))
return(yy.T)
def dofilter(ff, numeric=True, rollingsum=False, normalize=False, divisor=1000):
#print(yy)
#print(yy.T)
first = 0
for i in range(0,len(ff.columns)):
#print(ff.iloc[:,i]) #DEBUG
if (numeric):
ff.iloc[:,i] = ff.iloc[:,i].str.replace(',','').astype(int)
if (rollingsum):
ff.iloc[:,i] = ff.iloc[:,i].rolling(window=12).sum()
first = 11
if (normalize):
ff.iloc[:,i] = ff.iloc[:,i] - ff.iloc[first,i]
ff.iloc[:,i] = ff.iloc[:,i]/divisor
#yy = yy.T
#yy = yy.iloc[:,9:]
return(ff)
yy0 = joinyears(2015, 2018)
yy = yy0.copy()
zz = dofilter(yy)
dd = zz.iloc[:,9:]
print('U.S. TREASURY RECEIPTS, OUTLAYS, AND DEFICITS: Monthly Amount ($billions)')
print(dd)
fig, ax = plt.subplots(1, 1, figsize=(12, 8))
ax.plot(dd)
ax.set_title('U.S. TREASURY RECEIPTS, OUTLAYS, AND DEFICITS: Monthly Amount')
ax.set_xlabel(xlabel_last)
ax.set_ylabel('Billions of Dollars')
ax.grid(zorder=0)
ax.legend(dd.columns)
The reason for starting in October of 2014 is that this date is the start of fiscal year 2015, the first year for which the Treasury has spreadsheets posted on its site. In any event, the above plot clearly shows that receipts and surpluses peak in April of every year, presumedly due to taxpayers filing their returns. The plot also shows that the prior two months appeared to have had the largest monthly deficits since at least October of 2014. In fact, a ZeroHedge article mentions that the "March budget deficit of \$208.7 billion was 18% higher than \$176.2BN deficit recorded last March, and was the biggest March budget deficit in US history."
In order to factor in the periodic surpluses and deficits that occur over each year, it makes sense to look at the 12-month rolling sum of these values. That is done in the following Python code and the resulting plot.
yy = yy0.copy()
zz = dofilter(yy, rollingsum=True, normalize=True)
dd = zz.iloc[:,9:]
print('U.S. TREASURY RECEIPTS, OUTLAYS, AND DEFICITS: Change in 12-Month Rolling Sum ($billions)')
print(dd.iloc[11:,:])
fig, ax = plt.subplots(1, 1, figsize=(12, 8))
ax.plot(dd)
ax.set_title('U.S. TREASURY RECEIPTS, OUTLAYS, AND DEFICITS: Change in 12-Month Rolling Sum')
ax.set_xlabel(xlabel_last)
ax.set_ylabel('Billions of Dollars')
ax.grid(zorder=0)
ax.legend(dd.columns)
As can be seen, the 12-month rolling sum of receipts has increased by about \$54 billion since Trump took office. However, yearly outlays have increased by about \$248 billion, causing the yearly deficit to increase by about \$194 billion.
Focusing on the receipts, the following Python code plots the 12-month rolling sum of the three largest contributors to receipts. Those are individual income taxes, corporation income taxes, and employment taxes. Employment taxes consist chiefly of payroll taxes.
yy = yy0.copy()
zz = dofilter(yy, rollingsum=True, normalize=True)
dd = zz.iloc[:,0:3]
oo = zz.iloc[:,3:9]
dd.is_copy = False # avoids warning
dd['Other'] = oo.sum(axis=1, skipna=False)
print('SOURCES OF U.S. TREASURY RECEIPTS: Change in 12-Month Rolling Sum ($billions)')
print(dd.iloc[11:,:])
fig, ax = plt.subplots(1, 1, figsize=(12, 8))
ax.plot(dd)
ax.set_title('SOURCES OF U.S. TREASURY RECEIPTS: Change in 12-Month Rolling Sum')
ax.set_xlabel(xlabel_last)
ax.set_ylabel('Billions of Dollars')
ax.grid(zorder=0)
ax.legend(dd.columns)
fig.savefig(savefig_rec)
As can be seen, yearly individual income tax receipts have risen about \$124 billion under Trump. Yearly employment tax receipts have risen about \$34 billion and yearly corporation tax receipts have dropped about \$94 billion. It is interesting to note that most of the drop in corporation tax receipts have occurred since the passage of the Tax Cuts and Jobs Act of 2017 so that may be the most visible effect of the tax bill on receipts. The surge in individual tax receipts likely has little to do with the tax bill since the taxes paid in April are based on the prior tax law. Regarding individual tax receipts, the Washington Times article does say the following:
Analysts said they’ll have a better idea of what’s behind the surge as more information rolls in, but for now said it looks like individual taxpayers are paying more because they have higher incomes.
“Those payments were mostly related to economic activity in 2017 and may reflect stronger-than-expected income growth in that year,” the analysts said in their monthly budget review. “Part of the strength in receipts also may reflect larger-than-anticipated payments for economic activity in 2018. The reasons for the added revenues will be better understood as more detailed information becomes available later this year.”
In any event, the following Python code shows the 12-month rolling sum of the other contributors to receipts.
yy = yy0.copy()
zz = dofilter(yy, rollingsum=True, normalize=True)
dd = zz.iloc[:,3:9]
print('OTHER SOURCES OF U.S. TREASURY RECEIPTS: Change in 12-Month Rolling Sum ($billions)')
print(dd.iloc[11:,:])
fig, ax = plt.subplots(1, 1, figsize=(12, 8))
ax.plot(dd)
ax.set_title('OTHER SOURCES OF U.S. TREASURY RECEIPTS: Change in 12-Month Rolling Sum')
ax.set_xlabel(xlabel_last)
ax.set_ylabel('Billions of Dollars')
ax.grid(zorder=0)
ax.legend(dd.columns)
fig.savefig(savefig_recoth)
As can be seen, yearly receipts from excise taxes have risen about \$3 billion, customs taxes have risen about \$7 billion, and yearly miscellaneous receipts have fallen about \$20 billion since Trump took office. Adding that to the \$124 billion gain in individual tax receipts, \$34 billion gain in employment tax receipts, and \$94 billion loss in corporation tax receipts gives a total gain of about \$54 billion for all receipts, same as was shown in the second plot above.
The CBO Monthly Budget Review for April 2018, referenced by the Washington Times article as mentioned above, starts as follows:
The federal budget deficit was \$382 billion for the first seven months of fiscal year 2018, the Congressional Budget Office estimates, \$37 billion more than the shortfall recorded during the same period last year. Revenues and outlays were higher, by 4 percent and 5 percent, respectively, than they were during the first seven months of fiscal year 2017.
Hence, despite the increase in receipts, the deficit is continuing to grow. The following Python code plots the increase the 12-month rolling sum of the deficit.
yy = yy0.copy()
zz = dofilter(yy, rollingsum=True)
dd = zz.iloc[:,[11]]
print('U.S. TREASURY SURPLUS/DEFICIT(-): 12-Month Rolling Sum ($billions)')
print(dd.iloc[11:,:])
fig, ax = plt.subplots(1, 1, figsize=(12, 8))
ax.plot(dd)
ax.set_title('U.S. TREASURY SURPLUS/DEFICIT(-): 12-Month Rolling Sum')
ax.set_xlabel(xlabel_last)
ax.set_ylabel('Billions of Dollars')
ax.grid(zorder=0)
ax.legend(dd.columns)
fig.savefig(savefig_def)
As can be seen, the annual deficit has increased from just under \$600 billion per year when Trump took office to nearly \$800 billion per year now. In fact, a CNBC article describes it as follows:
The deficit jumped to \$779 billion, \$113 billion or 17 percent higher than the previous fiscal period, according to a statement from Treasury Secretary Steven Mnuchin and Office of Management and Budget Director Mick Mulvaney. It was larger than any year since 2012, when it topped \$1 trillion. The budget shortfall rose to 3.9 percent of U.S. gross domestic product.
The referenced statement from Steven Mnuchin and Mick Mulvaney contains the following table:
| Budget in billions of $ | Receipts | Outlays | Deficit |
|---|---|---|---|
| FY 2017 Actual | 3,315 | 3,981 | -666 |
| Percentage of GDP | 17.2% | 20.7% | 3.5% |
| FY 2018 Estimates: | |||
| 2019 Budget | 3,340 | 4,173 | -833 |
| 2019 Mid-Session Review | 3,322 | 4,171 | -849 |
| FY 2018 Actual | 3,329 | 4,108 | -779 |
| Percentage of GDP | 16.5% | 20.3% | 3.9% |
As can be seen, both receipts and outlays increased from the prior year though receipts just increased slightly. However, the table shows that both receipts and outlays decreased as a percentage of GDP. In fact, the statement itself states:
As a percentage of GDP, receipts equaled 16.5 percent, 0.7 percentage point lower than in FY 2017 and 0.9 percentage point below the average over the last 40 years.
Also, the Congressional Budget Office's Monthly Budget Review for September 2018 states the following:
The federal budget deficit was \$782 billion in fiscal year 2018, the Congressional Budget Office estimates, \$116 billion more than the shortfall recorded in fiscal year 2017. As was the case last year, this year’s outlays were affected by shifts in the timing of certain payments that otherwise would have been due on a weekend. If not for those shifts, the deficit for the year would have been \$826 billion—\$162 billion larger than last year’s amount.
This shifting of payments was caused by the fact that October 1, 2017 was on a Sunday. For this reason, certain payments scheduled for that day had to be shifted to Friday, September 29 of the prior fiscal year. On the other hand, October 1, 2018 was on a Monday so no payments had to be shifted this year. In fact, this is the reason that the 12-month rolling sum for the deficit decreased sharply from \$890 billion to \$779 billion. For this month, September of 2017 (which had those additional shifted outlays) was dropped and September of 2018 (which had normal outlays) was added. As seen in the second graph above, this caused a drop in annual outlays and deficits. However, since outlays were shifted from FY2018 to FY2017 (but not from FY2019 to FY2018) this causes the FY2018 deficit to be lower by the amount of the shifted payments.
In any case, the record surplus in April does not appear to signify anything more than the fact that 2017 had relatively strong economic growth. Contrary to Marc Short's statement, it doesn't appear to have much, if anything, to do with the tax bill that was passed at the end of 2016. On the contrary, the fact that receipts are below their 40-year average, even in a good economy, suggests that the tax cuts have much to do with why the deficit is continuing to grow.
Note: The Jupyter Notebook from which this post is generated can be found at http://econdataus.com/mts1809.ipynb. It is identical to the one at http://econdataus.com/mts1804.ipynb except that it has been updated to include September 2018. Links to additional Jupyter Notebooks can be found at http://econdataus.com/jupyter.html.