Thursday, August 6, 2015

Simple forecasts of employment

Relatively recently, a Current Employment Report has been released, 1st Quarter of 2015. In February, I made a forecast of 1st Quarter 2015 headline employment. The forecast amount was 63,010. Headline employment for the period was 62,760, which means the forecast was within my margin of error of 1,050 (95% confidence). To reflect, my forecasts of the average hourly private sector wage, employment, and CPI were all overstated. The employment forecast was not so bad, but I'd like to compare it to a naive forecasting methodology (rolling average annual rates of increase for windows of different periods of years):

Actual: 62,760
VAR forecast: 63,010 (+250)
1-year window: 62,210 (-550)
2-year window: 62,320 (-440)
3-year window: 62,450 (-310)
4-year window: 62,300 (-460)
5-year window: 62,390 (-370)

Comparatively, this one period shows a better performance for the VAR. Nevertheless, for now, consider the following forecasts for 2nd Quarter 2015 (which has already past, but is not published yet, so still unknown):

1-year window: 62,920 (+160 jobs)
2-year window: 62,730 (-30 jobs)
3-year window: 62,450 (+80 jobs)
4-year window: 62,300 (-10 jobs)
5-year window: 62,720 (-40 jobs)

If these forecasts are to be believed, there will probably be little change in headline employment between March and June 2015. A major drawback to this form of forecasting is that, since it is based on lagged rates of change, once a growth trend ends, the forecast will be thrown off. If all one uses is a simple extrapolative technique, then even a rate of change which should be able to be anticipated because of known or suspected (perhaps by some type of model with structural variables) relationships between available data would not be accounted for.

In the future, I will probably add back a VAR model to forecast a number of variables together, but some simplified technique would probably be worthwhile, just to consider an alternative benchmark.

Saturday, July 4, 2015

Forecast of CPI for third quarter 2015

It has been a while since I made a forecast or reflected on the results of my prior forecasts. I have reviewed my most recent forecast for CPI for first quarter 2015 and it is far from perfect. For one thing, CPI fell by 1.6% from 117.9 in fourth quarter 2014 to 116.0 in first quarter 2015. This is in stark contrast with my projected increase of 0.6% to 118.6 in first quarter 2015. Looking across several simple forecasting methods (taking the historical average growth rate of CPI with rolling 1, 2, 3, 4, and 5 year windows), I have determined that each of these methods yielded similar results to my three-variable vector autoregression:

Actual: 116.0 (decrease of 1.6%)
1-year window: 118.4
2-year window: 118.1
3-year window: 118.1
4-year window: 118.4
5-year window: 118.5

Going forward one period to the next period for my simple extrapolating forecast methods, consider the following results:

Actual: 116.5 (increase of 0.4%)
1-year window: 115.9
2-year window: 115.9
3-year window: 116.0
4-year window: 116.4
5-year window: 116.4

The method with the lowest Mean Absolute Percentage Error between the uniform start period for the forecasts (2001 quarter 4) was the 5-year window. I still want to further refine my use of different forecasting methods, but I can at least keep this simple method in reserve, to look at alongside forecasts from a more advanced technique.

For third quarter of 2015, here are the forecasts:

1-year window: 115.4
2-year window: 115.4
3-year window: 116.5
4-year window: 116.7
5-year window: 116.9

I will try to keep up with the reports as they come out and compare my forecasts against them in the future.

Wednesday, May 27, 2015

Government bonds and borrowing

In my previous post "What is fiscal policy?", I laid out a brief explanation of what fiscal policy is and mentioned that deficit/debt financing allows a government to make current expenditures with the commitment to pay back expenditures in the future. Often, a government resorts to issuing a bond rather than a loan from a bank. A bond is basically just a loan which is to be paid to the holders of a security.

From the point of view of individuals, the bond is a security that one can buy which will yield payment(s) in the future. The demand for the bond once it is issued is determined by the rate of interest that investors require given their perception of the risk that the issuer will not repay the bond as agreed. This will be reflected in the market price of the bond which, in turn, determines its interest rate. Consider the following simple example. A government has a number of capital improvement projects that it wishes to fund with borrowing. Rather than borrowing a specific amount, policymakers decide that they are reasonably certain about their ability to pay back an amount the next year, so the government issues a 1-year zero-coupon (aka, discount) bond in the amount of $50 million.[1] Given the risk profile of the government, investors are willing to lend at an annual interest rate of 5%. Here is the equation for a 1-year maturity discount bond:

Thus, at 5%, the Current Price which would be offered in the market would be about $47.6 million for the payment of $50 million in 1 year. If the market interest rate were instead 10%, the Current Price would be about $45.5 million. It is not hard to see that this is identical to the case where a $45.5 million 1 year maturity bond is issued, investors are willing to lend at 10% interest and the ultimate amount to be paid in 1 year is $50 million.

There are several different types of bonds, but one thing to keep in mind is that the price of the bond is going to be inversely related with its rate of interest. If individuals become more willing to invest in a given bond, the interest rate will decline.

There are constraints to how much debt a country can issue. One is whatever restraints are put on the issuance of debt, for example, there are formal debt limits that are set by constitutional provisions or the laws of a jurisdiction. In Guam, it is 10% of the aggregate real property valuation, which is expected to rise considerably soon.

A second consideration is the total debt that the market will bear. As total debt rises, investors will tend to view the ability or willingness of a government to pay for the additional debt with increasing skepticism, which will eventually translate into rising interest rates, generally.  This will raise the burden of interest payments and make it harder for a government to service additional debt.

A third consideration is the technical sustainability of government debt.  In the long run, individuals generally need to make means and ends meet (i.e., the present discounted value of lifetime expenditures should equal the present discounted value of lifetime revenue). A  crucial difference between an individual's ability to pay and the government's is that a government is presumed to be a permanent institution. The definition of solvency for a government, then, is that the present value of revenues should equal or exceed the present value of expenditures. What this means is that taking together current revenues and future revenues, properly discounted at a given rate of interest, would need to equal or exceed the current expenditures and future expenditures, properly discounted at a given rate of interest.[2]

These, however, are constraints, not definitions of the optimal public policy with regard to the accumulation of debt. Some may argue the desirability of targeting a nominal amount of debt, a real amount of debt (discounted by the price level), or a debt-to-output ratio, as I discussed briefly in the previous post. I am personally inclined to favor targeting a long-run debt-to-output ratio.

[1] As a practical matter, the amount raised can be determined, but could not be known for certain well before the issuance.
[2] Just to clarify the issue a little for those mathematically inclined, but unfamiliar with the concept of present value, here's an expression of present value for a current value at time "t":

Wednesday, April 29, 2015

What is fiscal policy

If I were to guess, I'd suggest that when most people hear the words fiscal policy, they would think of the government's taxing, spending, and borrowing. In economics, the words have a broader meaning. Upon reflection, I would say that fiscal policies are the choices that the government collectively makes with regard to the use of purchasing power and the means by which such purchasing power is obtained. When putting this definition together, I had in my mind the need to encompass the following: purchases of goods and services, employment and retirement of government workers, transfer policies, taxes, fees, intergovernmental transfers, borrowing, and helicopter money.


The uses of puchasing power are fairly straightforward. The two main categories are transfer payments and providing goods or services. Transfer payments are simply distribution of purchasing power to a set of individuals. The best known policy of this nature is social security, which distributes money to the disabled and senior citizens. Providing goods and services involve a combination of buying goods and services from businesses and employing government workers (which also means providing for their retirement).


There are fundamentally two ways to pay for the government's use of purchasing power: taxation and printing money. The primary way is through taxation (whether through "taxes" or "fees").

A major issue in finance is when this taxation will occur. The typical "balanced budget" approach is for the government to pay for all expenditures in a period by raising an equal amount of revenue in the same period. The alternative "deficit/debt financing" approach is for the government to borrow in the current period and promise to repay lenders in the future. This latter option can employ loans or a bond, but the principle is the same (I'm ignoring the possibility of printing money, for simplicity and because nobody seriously suggests that as a sustained source of revenue).

There are two main ideas of how to conduct fiscal policies in the long run. One is to maintain a given level of nominal debt (i.e., revenues and expenditures are equal), the other is generally to maintain a stable debt-to-gdp ratio (nominal debt grows only as fast as nominal output). Think of it this way: policy may be aiming at lowering (for instance) nominal debt or the debt-to-gdp level, but there is generally a level that is going to be targeted in the long-run (even if that level is zero).

According to old keynesian and neowicksellian/new keynesian models, fiscal policy can stimulate an economy with less than full employment of productive factors (labor and capital).

Tuesday, February 24, 2015

Forecast of Guam Variables for Calendar Year 2015 First Quarter

Recently, the Current Employment Report has come out with less than stellar performance for the last quarter of 2014.  Total employment declined.  My first forecast was not as accurate as I would have liked.  In retrospect, I used too many lags of data in my VAR and univariate models.  Furthermore, I was working to reduce the root mean squared error a bit too hard.  This time, I have one model, with three (kinda four) variables.  I am, once again, employing a Vector Autoregression.  My variables are:

Acceleration of (d^2) CPI inflation,
Acceleration of (d^2) Hourly wage inflation, and
Change (d) in Employment relative to the change.

After adjusting to the figures published for CPI, Average Hourly Private Sector Wage (AHW), and Total Employment (Emp), I have the following forecast results:

2015, 1st Quarter:
CPI:       118.6    (~2)
AHW:    13.30   (~0.37)
Emp:      63,010 (~1,050)

In the future, I'd like to try to tighten up my one point ahead forecasts, perhaps by combining multiple methods of estimation.

Friday, December 12, 2014

CPI and Employment Forecasts

Out of curiosity, I took some statistical software that I have and generated the forecasts of CPI and Payroll Employment below. In the future, I would like to further refine my forecasting techniques, but I decided that, for the time being, I would begin creating one-point ahead forecasts. I hope to refine my techniques in the future. My current model is a vector autoregression based upon two variables, payroll employment as a percent of the age 16-64 population and change in the rate of inflation. 

Here is the CPI estimate:
3rd Quarter 2014 (actual): 117.2
4th Quarter 2014 (forecast): 118.7 (and increase of 1.5 points, or by 1.3% [annualized to 5.2%])

Here is the Payroll Employment estimate:
3rd Quarter 2014 (actual): 62,480
4th Quarter 2014 (forecast): 63,350 (increase of 870 in payroll employment, or by 1.4%)

Using an alternative univariate method for one-period ahead forecast for CPI, I was able to get an equivalent forecast of 118.7.  Using another univariate method for one-period ahead forecast of Payroll Employment, I was able to get a forecast of 62,760, an increase of 280 in payroll employment, or by 0.4%. Regardless of the differences in these forecasting methods, the indication is that Payroll Employment would be expected to rise from September 2014 to December 2014, in both cases more than my estimate of the rise of the age 16-64 population, which could rise by approximately 166, or by about 0.2%.

I hope that these forecasts are true, as it may indicate that we may be experiencing an economic recovery. In the future, I would like to work on improving my forecasting methods and developing more elaborate and accurate models which may give a better indication of where we are headed in the near future, and hopefully into the more distant future, too.

Thursday, October 9, 2014

All indications point to prolonged slump

As November approaches, it may be important to understand where Guam's economy stands (from those indicators which are readily available). As a student of economics, I am most interested in knowing how big the output gap is. There are a number of lines of evidence that one can use to tell how far we are from full employment, which I will reproduce below:
  • Employment as a Percent of the 16-64 Population
  • Average Hourly Wages in the Private Sector
  • The Employment-Population Ratio
  • Changes in the Rate of Inflation
The sources of these pieces of information are the Current Employment Reports and the Unemployment Situation, both released by the Guam Department of Labor Bureau of Labor Statistics, and the Guam Consumer Price Index, which is released by the Bureau of Statistics and Plans.

Current Employment Reports

Data Sources: Current Employment Reports (GDOL), Census Demographic Database

The above graph shows the 5-quarter centered moving averages of employment as a percent of the 16-64 population. The centered average begins declining just as the current administration is about to take office. I should point out that the peak period was actually December 2010 (just before the Calvo-Tenorio Administration took office), which is masked in this graph (looking like it is September 2010), since I'm using a centered moving average, which would give a pretty good indication of the trend, but turning points tend to be distorted. Another example is that the low point is actually located on in the June 2012, but the centered moving average low point is at the December 2011. If there is a takeaway from this data, it is that, adjusted for demographics, there appears to be a large shortfall in employment since shortly after the Calvo-Tenorio Administration took office and there has been little progress in bringing back full employment.

Data Sources: Current Employment Reports (GDOL)

The graph above indicates that a similar "depression" has occurred in the average hourly wage of private sector workers. As one can see, prior to 2011, the progress of wages tracked a trend line, showing signs of movement back toward that line during practically every deviation.  In 2011, wages slumped and basically haven't come back to trend.

Unemployment Situation

Source: Unemployment Situations (GDOL)

The above graph represents the employment-population from the first quarter of 2011 to the third quarter of 2013, the latest report which this administration has released. For most of the data, I used a three-quarter centered moving average, but for the first quarter of 2011, there was not any data immediately before or after, so the number is not an average. In the first quarter of 2012 and the third quarter of 2013, I took an average of that quarter and the nearest quarter. If you compare this result with the graph of employment as a percent of the 16-64 population, you'll find that these are very similar, which is an encouraging sign. The employment-population ratio is simply taking the percent of the civilian non-institutional population which is employed, while the other is doing something very similar with the current employment report data. The numbers won't line up for various reasons, but they can give a complementary look at the position of the labor market.

Guam Consumer Price Index

Data Source: Guam Consumer Price Index

The Guam Consumer Price Index data is very interesting because by economic theory (specifically the new keynesian phillips curve), we would expect changes in the rate of inflation to be negatively correlated with the unemployment rate or, if one uses the measures shown earlier, we would expect changes to be positively correlated with the employment ratio or employment as a percent of the 16-64 population.  What we find is that the 2-year trailing average inflation rate begins declining in 2012, which, if I chose to center this upon the middle period where the rate of inflation peaked, would have been the beginning of 2011. Core inflation is the better measure of inflation because it usually better represents the longer-term trend of inflation, rather than bouncing around as much as headline inflation does. What this graph illustrates is that Guam's economy may be headed for outright deflation, which demonstrates the weakness of our economy and, potentially, hints that the economy could remain depressed.


To conclude, in brief, Guam's economy, by the lines of evidence I have examined, seems to have suffered from a prolonged slump since just after the Calvo-Tenorio Adminstration began and may remain depressed for a while longer. I would not say this is destined because (a) things can be done to address shortfalls in aggregate demand (which is at the center of the depressed economy) and (b) things may change without a specific intervention (or as some say, "The trend is your friend till the end," or, alternately, "stuff happens.").