Economic forecasting made easy

New techniques of predicting growth have proliferated recently. The Bangko Sentral ng Pilipinas (BSP), for instance, has added to its suite of forecasting models, “nowcasting” models, to estimate growth as a measure of domestic demand in formulating monetary policy. The results of these quantitative exercises are exclusively for internal use but BSP’s performance has been satisfactory. Its business cycle model, business and consumer expectations surveys, as well as high-frequency data monitoring have also supported its macroeconomic surveillance.

Other central banks have similarly embarked on nowcasting because relying on quarterly economic and financial data with long lags and frequent revisions would serve very little purpose. Nowcasting yields economic estimates in virtually real-time in lieu of the official measures. Many advanced economies have shifted to this methodology following the exponential improvements in computing power. For instance, the Reserve Bank of New Zealand is now doing nowcasting based on machine-learning algorithms with superior results relative to estimates using conventional methodologies like simple autoregressive model and factor model.

Three years ago, La Salle’s Cesar Rufino presented his interesting work “Nowcasting Philippine Economic Growth Using MIDAS Regression Modelling” which showed the viability of using MIDAS or Mixed Data Sampling Regression. This addresses the problem of mixed frequency. Based on quarterly data of output, monthly data on inflation, industrial production, and the Philippine Stock Exchange, he performed nowcasting. His results were quite revealing in that using MIDAS nowcasting, estimates turned out better than those coming from traditional models both in-sample and out-of-sample.

Given this broad context, by the time this column comes out Friday, the Philippine Statistics Authority (PSA) shall have released the 4^th quarter and whole year 2020 real GDP (gross domestic product).

Against the actual performance of the Philippine economy, we can see how economists, researchers, bank analysts, credit rating agencies, and international financial institutions assessed the Philippine macroeconomy in the absence of actual figures. We are not privy to the econometric techniques they used in churning out their estimates but if they used the traditional methods, they would have some problems with mixed frequency of explanatory variables, many of which would also not be available even at this point.

Nowcasting enables econometricians to make use of high-frequency data. Using machine-learning algorithms could further refine their estimation.

First, we survey the BusinessWorld’s poll of economists from the University of Asia and the Pacific, San Juan de Letran, De La Salle and Ateneo de Manila; bank and other financial institution analysts from Union Bank, Sun Life Financial, Philippine National Bank, BDO, ANZ, RCBC, HSBC, Nomura, Security Bank, ING and BPI; data and information business provider and research outfits IHS Markit, Capital Economics and Oxford Economics; and the International Finance Institute.

For the 4^th quarter 2020 forecast, we have a high of -5% by Union Bank an2d a low of -12.2% by BPI. Median for the last quarter last year stood at -8.5% with both Capital Economics and RCBC in the same groove. There was almost an even representation of bank analysts and research outfits and academics on both sides of the median.

Following their last-quarter take on the real GDP, those polled fell into a median of -9.5% for 2020, the lower end of -8.5% to -9.5% forecast by the Development Budget Coordination Committee. No forecaster was higher than the Government’s optimistic upper-end scenario of -8.5%.

What is notable in this range of forecasts is that in the more optimistic scenario of higher than -9.5% there seems to be more substantial weight placed on the positive impact of further easing on the lockdown and the possible rollout of anti-COVID-19 vaccines. These are expected to raise consumer spending. Forecasts also turned up because of the anticipated seasonal rise in overseas workers’ remittances.

We need more than a pinch of salt in dealing with the uncertainty surrounding the optimistic scenario. It would hardly align with the recent surge in new cases given the aggressive British variant of the coronavirus. Moody’s considered health and safety issues in the Philippines as “highly negative risk.” There are budgetary pressures to sustain spending on public health. In addition, mass vaccination to achieve herd immunity is immensely challenged by the public’s reluctance to get the jab for lack of sufficient public information. Pricing of the vaccines raises issues about possible graft. Sourcing of the vaccines remains non-devolved to local government units and private corporates. Society could very well leverage on their quick logistical capability and accelerate mass vaccination. Remittances are no savior as they dropped by nearly 1% for the first 11 months of 2020. With high unemployment, it would be difficult to expect people to spend even when prices were brought down to fire sale levels.

What about the views of foreign banks?

Their views mirror those of their counterparts. Pandemic mitigation is crucial in pitching their forecasts.

Morgan Stanley is as optimistic as PNB and UAP in predicting a 6% drop in the 4^th quarter and an 8.8% output decline for the whole year 2020. It premised its scenario on the bounce back of private consumption, milder decline in passenger vehicle sales, and recovery in exports.

United Overseas Bank’s forecast of -8.2% is also on the upper side of the median, exactly the same view of IIF (Institute of International Finance). However, it’s full-year prediction is right on the government’s worse end and median of -9.5%. It shares the whole-year view of IIF, Capital Economics, and RCBC.

Standard Chartered Bank placed the whole year 2020 real GDP at -8.9% exactly as PNB’s. The Bank believes the BSP will keep the rates steady for growth reasons while inflation is expected to remain under control. However, the Bank thinks the pre-COVID-19 growth levels would not be achieved until 2022.

Citibank did not release any forecasts of Philippine growth but it shared Standard Charter’s two-year timetable for restoring growth to pre-pandemic level. This pessimism derives from our harsh lockdown that impaired mobility and economic activity. It also noted that the Philippines lags behind in vaccines procurement in the ASEAN.

Credit rating agencies also weighed in on some aspects of the growth story of the country.

S&P, for instance, warned of the observed deterioration in banks’ asset quality. Huge exposure in real estate was pointed out to be the main risk driving the banks’ credit profile. This should alert the regulators to consider financial stability issues in pursuing aggressive monetary policy.

While Fitch was rather bullish on the country’s construction industry due to priority public spending on infrastructure, it recognized that “the damage to infrastructure and agriculture was significant in 2020…”

In the case of Moody’s, it argued that the “Philippine focus on infrastructure investment has led to a corresponding lack of funding for social services like healthcare, with the resulting inability to efficiently deal with public health…” The credit agency warned that the pandemic could still worsen and potentially strain public health facilities and disrupt business activity. In turn, the stress could constrain revenue generation and shock-absorption capacity.

The key take-away is the debt watchers’ concern about downside risks particularly the economic scar of the pandemic. Failure to establish some balance in the budget could affect the country’s ability to sustain growth. In our previous column, we talked about the failure of public policy to consider and prepare for the worst outcome. In forecasting, it would be prudent to anchor one’s model on a more realistic set of assumptions that conforms with what is unfolding today — our state of pandemic mitigation, ability to source and administer appropriate vaccines, the rapid spread of the latest mutation of the coronavirus and our policy offsets.

The World Bank (WB) and the International Monetary Fund (IMF) share broadly similar reservations about our prospects for recovery. The WB is therefore supporting the country’s proposed procurement of vaccines up to $400 million under the Emergency Response Project. This will be complemented by the Asian Development Bank’s $325-million loan under the Asia-Pacific Vaccine Access Facility.

For its part, the IMF was frank in admitting that the Philippines will likely see a slower pace of economic recovery in 2021. The damage in 2020 was most severe and for this reason, the Fund decided to reduce its real GDP forecast from 7.4% to 6.6% for 2021, close to the lower end of the Government’s 6.5-7.5% target.

For the whole year of 2020, the Fund expects a decline of 9.6%, much deeper than the original forecast of -8.3%. The Fund’s prediction is slightly south of the -9.5% median of BusinessWorld’s poll of economists and researchers. The IMF explained that this view “reflects the larger-than-expectation year-on-year contraction in the third quarter.”

Economic scars again.

Today, the PSA should have announced the growth story of the Philippines for the 4^th quarter and whole year 2020. Some of our forecasters could be spot on, some south and some north of the median for both periods. We don’t subscribe fully to John Kenneth Galbraith who said that “there are two kinds of forecasters: those who don’t know and those who don’t know they don’t know.”

We have more confidence today in our forecasters given the vast improvements in both modelling and computing power. But those consistent outliers ought to review their forecasts and assumptions so that they can avoid the cardinal sin “of faring only marginally better than a random forecast generator.”

One way of achieving this is to ensure a pretty good alignment between their assumptions and the real world.

Diwa C. Guinigundo is the former Deputy Governor for the Monetary and Economics Sector, the Bangko Sentral ng Pilipinas (BSP). He served the BSP for 41 years. In 2001-2003, he was Alternate Executive Director at the International Monetary Fund in Washington, DC. He is the senior pastor of the Fullness of Christ International Ministries in Mandaluyong.